Abaqus Example Problems Manual

ABAQUS Example Problems Manual
0-1
ADAMS is a registered United States trademark of Mechanical Dynamics, Inc.
ADAMS/Flex and ADAMS/View are trademarks of Mechanical Dynamics, Inc.
CATIA is a registered trademark of Dassault Systémes.
C-MOLD is a registered trademark of Advanced CAE Technology, Inc., doing business as C-MOLD.
Compaq Alpha is registered in the U.S. Patent and Trademark Office.
FE-SAFE is a trademark of Safe Technology, Ltd.
Fujitsu, UXP, and VPP are registered trademarks of Fujitsu Limited.
Hewlett-Packard, HP-GL, and HP-GL/2 are registered trademarks of Hewlett-Packard Co.
Hitachi is a registered trademark of Hitachi, Ltd.
IBM RS/6000 is a trademark of IBM.
Intel is a registered trademark of the Intel Corporation.
NEC is a trademark of the NEC Corporation.
PostScript is a registered trademark of Adobe Systems, Inc.
Silicon Graphics is a registered trademark of Silicon Graphics, Inc.
SUN is a registered trademark of Sun Microsystems, Inc.
TEX is a trademark of the American Mathematical Society.
UNIX and Motif are registered trademarks and X Window System is a trademark of The Open Group
in the U.S. and other countries.
Windows NT is a registered trademark of the Microsoft Corporation.
ABAQUS/CAE incorporates portions of the ACIS software by SPATIAL TECHNOLOGY INC. ACIS
is a registered trademark of SPATIAL TECHNOLOGY INC.
This release of ABAQUS on Windows NT includes the diff program obtained from the Free Software
Foundation. You may freely distribute the diff program and/or modify it under the terms of the GNU
Library General Public License as published by the Free Software Foundation, Inc., 59 Temple Place,
Suite 330, Boston, MA 02111-1307 USA.
This release of ABAQUS/CAE includes lp_solve, a simplex-based code for linear and integer
programming problems by Michel Berkelaar of Eindhoven University of Technology, Eindhoven, the
Netherlands.
Python, copyright 1991-1995 by Stichting Mathematisch Centrum, Amsterdam, The Netherlands. All
Rights Reserved. Permission to use, copy, modify, and distribute the Python software and its
documentation for any purpose and without fee is hereby granted, provided that the above copyright
notice appear in all copies and that both that copyright notice and this permission notice appear in
supporting documentation, and that the names of Stichting Mathematisch Centrum or CWI or
0-2
Corporation for National Research Initiatives or CNRI not be used in advertising or publicity
pertaining to distribution of the software without specific, written prior permission.
All other brand or product names are trademarks or registered trademarks of their respective
companies or organizations.
0-3
General conversion factors (to five significant digits)
Quantity U.S. unit SI equivalent
Length 1 in 0.025400 m
1 ft 0.30480 m
1 mile 1609.3 m
2
Area 1 in 0.64516 ´ 10-3 m2
1 ft2 0.092903 m 2
1 acre 4046.9 m2
Volume 1 in3 0.016387 ´ 10-3 m3
1 ft 3 0.028317 m 3
1 US gallon 3.7854 ´ 10-3 m3
Conversion factors for stress analysis
Density 1 slug/ft3 = 1 lbf s2/ft4 515.38 kg/m3
1 lbf s2/in4 10.687 ´ 106 kg/m3
Energy 1 ft lbf 1.3558 J (N m)
Force 1 lbf 4.4482 N (kg m/s2)
2
Mass 1 slug = 1 lbf s /ft 14.594 kg (N s2/m)
1 lbf s2/in 175.13 kg
Power 1 ft lbf/s 1.3558 W (N m/s)
Pressure, Stress 2
1 psi (lbf/in ) 6894.8 Pa (N/m2)
Conversion factors for heat transfer analysis
Conductivity 1 Btu/ft hr °F 1.7307 W/m °C
1 Btu/in hr °F 20.769 W/m °C
Density 1 lbm/in3 27680. kg/m3
Energy 1 Btu 1055.1 J
Heat flux density 1 Btu/in 2 hr 454.26 W/m2
Power 1 Btu/hr 0.29307 W
Specific heat 1 Btu/lbm °F 4186.8 J/kg °C
Temperature 1 °F 5/9 °C
Temp °F 9/5 ´ Temp °C + 32°
9/5 ´ Temp °K - 459.67°
Important constants
Constant U.S. unit SI unit
Absolute zero -459.67 °F -273.15 °C
Acceleration of gravity 32.174 ft/s 2 9.8066 m/s2
Atmospheric pressure 14.694 psi 0.10132 ´ 106 Pa
Stefan-Boltzmann 0.1714 ´ 10-8 Btu/hr ft2
5.669 ´ 10-8 W/m2 °K4
constant °R4
where °R = °F + 459.67 where °K = °C + 273.15
Approximate properties of mild steel at room temperature
Quantity U.S. unit SI unit
Conductivity 28.9 Btu/ft hr °F 50 W/m °C
2.4 Btu/in hr °F
Density 15.13 slug/ft3 (lbf s2/ft4) 7800 kg/m3
0.730 ´ 10-3 lbf s2/in4
0-4
0.282 lbm/in 3
Elastic modulus 30 ´ 106 psi 207 ´ 109 Pa
Specific heat 0.11 Btu/lbm °F 460 J/kg °C
Yield stress 30 ´ 103 psi 207 ´ 106 Pa
0-5
UNITED STATES
Hibbitt, Karlsson & Sorensen, Inc. Hibbitt, Karlsson & Sorensen (Michigan),
Inc.
1080 Main Street 14500 Sheldon Road, Suite 160
Pawtucket, RI 02860-4847 Plymouth, MI 48170-2408
Tel: 401 727 4200 Tel: 734 451 0217
Fax: 401 727 4208 Fax: 734 451 0458
E-mail: info@abaqus.com, E-mail: hksmi@abaqus.com
support@abaqus.com
http://www.abaqus.com
Hibbitt, Karlsson & Sorensen (West), ABAQUS Solutions Northeast, LLC
Inc.
39221 Paseo Padre Parkway, Suite F Summit Office Park, West Building
Fremont, CA 94538-1611 300 Centerville Road, Suite 209W
Tel: 510 794 5891 Warwick, RI 02886-0201
Fax: 510 794 1194 Tel: 401 739 3637
E-mail: hkswest@abaqus.com Fax: 401 739 3302
E-mail: support@abaqus-sn.com
AC Engineering, Inc.
1440 Innovation Place
West Lafayette, IN 47906-1000
Tel: 765 497 1373
Fax: 765 497 4444
E-mail: info@aceng.com
ARGENTINA AUSTRALIA
KB Engineering S. R. L. Compumod Pty. Ltd.
Florida 274, Of. 37 Level 13, 309 Pitt Street
(1005) Buenos Aires, Argentina Sydney 2000
Tel: +54 11 4393 8444 P.O. Box A807
Fax: +54 11 4326 2424 Sydney South 1235
E-mail: sanchezsarmiento@arnet.com.ar Tel: 02 9283 2577
Fax: 02 9283 2585
E-mail: support@compumod.com.au
http://www.compumod.com.au
AUSTRIA BENELUX
VOEST-ALPINE STAHL LINZ GmbH ABAQUS Benelux BV
Department WFE Huizermaatweg 576
Postfach 3 1276 LN Huizen
A-4031 Linz The Netherlands
Tel: 0732 6585 9919 Tel: +31 35 52 58 424
Fax: 0732 6980 4338 Fax: +31 35 52 44 257
E-mail: edwin.till@voest.co.at E-mail: support@abaqus.nl
CHINA CZECH REPUBLIC AND SLOVAK
REPUBLIC
Advanced Finite Element Services ASATTE
Department of Engineering Mechanics Technická 4, 166 07 Praha 6
Tsinghua University Czech Republic
Beijing 100084, P. R. China Tel: 420 2 24352654
Tel: 010 62783986 Fax: 420 2 33322482
0-6
Fax: 010 62771163 E-mail: asatte@biomed.fsid.cvut.cz
E-mail: zhuangz@mail.tsinghua.edu.cn
FRANCE GERMANY
ABAQUS Software, s.a.r.l. ABACOM Software GmbH
7, rue de la Patte d'Oie Theaterstraße 30-32
78000 Versailles D-52062 Aachen
Tel: 01 39 24 15 40 Tel: 0241 474010
Fax: 01 39 24 15 45 Fax: 0241 4090963
E-mail: support@abaqus.fr E-mail: abacom@abacom.de
ITALY JAPAN
Hibbitt, Karlsson & Sorensen Italia, Hibbitt, Karlsson & Sorensen, Inc.
s.r.l.
Viale Certosa, 1 3rd Floor, Akasaka Nihon Building
20149 Milano 5-24, Akasaka 9-chome
Tel: 02 39211211 Minato-ku
Fax: 02 39211210 Tokyo, 107-0052
E-mail: infohks@abaqus.it Tel: 03 5474 5817
Fax: 03 5474 5818
E-mail: hksj@hksj.co.jp
KOREA MALAYSIA
Hibbitt, Karlsson & Sorensen Korea, Inc. Compumod Sdn Bhd
Suite 306, Sambo Building #33.03 Menara Lion
13-2 Yoido-Dong, Youngdeungpo-ku 165 Jalan Ampang
Seoul, 150-010 50450 Kuala Lumpur
Tel: 02 785 6707/8 Tel: 3 466 2122
Fax: 02 785 6709 Fax: 3 466 2123
E-mail: hotline@abaqus.co.kr E-mail: hotline@compumod.com.my
NEW ZEALAND POLAND
Matrix Applied Computing Ltd. BudSoft Sp. z o.o.
P.O. Box 56-316, Auckland 61-807 Pozna
Courier: Unit 2-5, 72 Dominion Road, Sw. Marcin 58/64
Mt Eden,
Auckland Tel: 61 852 31 19
Tel: +64 9 623 1223 Fax: 61 852 31 19
Fax: +64 9 623 1134 E-mail: budsoft@man.poznan.pl
E-mail: hks-support@matrix.co.nz
SINGAPORE SOUTH AFRICA
Compumod (Singapore) Pte Ltd Finite Element Analysis Services (Pty) Ltd.
#17-05 Asia Chambers Suite 20-303C, The Waverley
20 McCallum Street Wyecroft Road
Singapore 069046 Mowbray 7700
Tel: 223 2996 Tel: 021 448 7608
Fax: 226 0336 Fax: 021 448 7679
E-mail: E-mail: abaqus@feas.co.za
compumod@mbox2.singnet.com.sg
SPAIN SWEDEN
Principia Ingenieros Consultores, S.A. FEM-Tech AB
Velázquez, 94 Pilgatan 8
28006 Madrid SE-721 30 Västerås
0-7
Tel: 91 209 1482 Tel: 021 12 64 10
Fax: 91 575 1026 Fax: 021 18 12 44
E-mail: abaqus@principia.es E-mail: femtech@femtech.se
TAIWAN UNITED KINGDOM
APIC Hibbitt, Karlsson & Sorensen (UK) Ltd.
7th Fl., 131 Sung Chiang Road The Genesis Centre
Taipei, 10428 Science Park South, Birchwood
Tel: 02 25083066 Warrington, Cheshire WA3 7BH
Fax: 02 25077185 Tel: 01925 810166
E-mail: cae@apic.com.tw Fax: 01925 810178
E-mail: hotline@hks.co.uk
0-8
This section lists various resources that are available for help with using ABAQUS, including
technical and systems support, training seminars, and documentation.
Support
HKS offers both technical (engineering) support and systems support for ABAQUS. Technical and
systems support are provided through the nearest local support office. You can contact our offices by
telephone, fax, electronic mail, or regular mail. Information on how to contact each office is listed in
the front of each ABAQUS manual. Support information is also available by visiting the ABAQUS
Home Page on the World Wide Web (details are given below). When contacting your local support
office, please specify whether you would like technical support (you have encountered problems
performing an ABAQUS analysis) or systems support (ABAQUS will not install correctly, licensing
does not work correctly, or other hardware-related issues have arisen).
We welcome any suggestions for improvements to the support program or documentation. We will
ensure that any enhancement requests you make are considered for future releases. If you wish to file a
complaint about the service or products provided by HKS, refer to the ABAQUS Home Page.
Technical support
HKS technical support engineers can assist in clarifying ABAQUS features and checking errors by
giving both general information on using ABAQUS and information on its application to specific
analyses. If you have concerns about an analysis, we suggest that you contact us at an early stage, since
it is usually easier to solve problems at the beginning of a project rather than trying to correct an
analysis at the end.
Please have the following information ready before calling the technical support hotline, and include it
in any written contacts:
· The version of ABAQUS that are you using.
- The version numbers for ABAQUS/Standard and ABAQUS/Explicit are given at the top of the
data (.dat) file.
- The version numbers for ABAQUS/CAE and ABAQUS/Viewer can be found by selecting
Help->On version from the main menu bar.
- The version number for ABAQUS/CAT is given at the top of the input ( .inp) file as well as
the data file.
- The version numbers for ABAQUS/ADAMS and ABAQUS/C-MOLD are output to the
screen.
- The version number for ABAQUS/Safe is given under the ABAQUS logo in the main
window.
· The type of computer on which you are running ABAQUS.
0-9
· The symptoms of any problems, including the exact error messages, if any.
· Workarounds or tests that you have already tried.
When calling for support about a specific problem, any available ABAQUS output files may be helpful
in answering questions that the support engineer may ask you.
The support engineer will try to diagnose your problem from the model description and a description
of the difficulties you are having. Frequently, the support engineer will need model sketches, which
can be faxed to HKS or sent in the mail. Plots of the final results or the results near the point that the
analysis terminated may also be needed to understand what may have caused the problem.
If the support engineer cannot diagnose your problem from this information, you may be asked to send
the input data. The data can be sent by means of e-mail, tape, or disk. Please check the ABAQUS
Home Page at www.abaqus.com for the media formats that are currently accepted.
All support calls are logged into a database, which enables us to monitor the progress of a particular
problem and to check that we are resolving support issues efficiently. If you would like to know the log
number of your particular call for future reference, please ask the support engineer. If you are calling to
discuss an existing support problem and you know the log number, please mention it so that we can
consult the database to see what the latest action has been and, thus, avoid duplication of effort. In
addition, please give the receptionist the support engineer's name (or include it at the top of any e-mail
correspondence).
Systems support
HKS systems support engineers can help you resolve issues related to the installation and running of
ABAQUS, including licensing difficulties, that are not covered by technical support.
You should install ABAQUS by carefully following the instructions in the ABAQUS Site Guide. If
you encounter problems with the installation or licensing, first review the instructions in the ABAQUS
Site Guide to ensure that they have been followed correctly. If this does not resolve the problems, look
on the ABAQUS Home Page under Technical Support for information about known installation
problems. If this does not address your situation, please contact your local support office. Send
whatever information is available to define the problem: error messages from an aborted analysis or a
detailed explanation of the problems encountered. Whenever possible, please send the output from the
abaqus info=env and abaqus info=sys commands.
ABAQUS Web server

For users connected to the Internet, many questions can be answered by visiting the ABAQUS Home
Page on the World Wide Web at
http://www.abaqus.com
The information available on the ABAQUS Home Page includes:
· Frequently asked questions
· ABAQUS systems information and machine requirements
0-10
· Benchmark timing documents
· Error status reports
· ABAQUS documentation price list
· Training seminar schedule
· Newsletters
Anonymous ftp site

For users connected to the Internet, HKS maintains useful documents on an anonymous ftp account on
the computer ftp.abaqus.com. Simply ftp to ftp.abaqus.com. Login as user anonymous, and type your
e-mail address as your password. Directions will come up automatically upon login.
Writing to technical support

Address of HKS Headquarters:
Hibbitt, Karlsson & Sorensen, Inc.
1080 Main Street
Pawtucket, RI 02860-4847, USA
Attention: Technical Support
Addresses for other offices and representatives are listed in the front of each manual.
Support for academic institutions

Under the terms of the Academic License Agreement we do not provide support to users at academic
institutions unless the institution has also purchased technical support. Please see the ABAQUS Home
Page, or contact us for more information.
Training
All HKS offices offer regularly scheduled public training classes.
The Introduction to ABAQUS/Standard and ABAQUS/Explicit seminar covers basic usage and
nonlinear applications, such as large deformation, plasticity, contact, and dynamics. Workshops
provide as much practical experience with ABAQUS as possible.
The Introduction to ABAQUS/CAE seminar discusses modeling, managing simulations, and viewing
results with ABAQUS/CAE. "Hands-on" workshops are complemented by lectures.
Advanced seminars cover topics of interest to customers with experience using ABAQUS, such as
engine analysis, metal forming, fracture mechanics, and heat transfer.
We also provide training seminars at customer sites. On-site training seminars can be one or more days
in duration, depending on customer requirements. The training topics can include a combination of
material from our introductory and advanced seminars. Workshops allow customers to exercise
ABAQUS on their own computers.
0-11
For a schedule of seminars see the ABAQUS Home Page, or call HKS or your local HKS
representative.
Documentation
The following documentation and publications are available from HKS, unless otherwise specified, in
printed form and through our online documentation server. For more information on accessing the
online books, refer to the discussion of execution procedures in the user's manuals.
In addition to the documentation listed below, HKS publishes two newsletters on a regular schedule:
ABAQUS/News and ABAQUS/Answers. ABAQUS/News includes topical information about program
releases, training seminars, etc. ABAQUS/Answers includes technical articles on particular topics
related to ABAQUS usage. These newsletters are distributed at no cost to users who wish to subscribe.
Please contact your local ABAQUS support office if you wish to be added to the mailing list for these
publications. They are also archived in the Reference Shelf on the ABAQUS Home Page.
Training Manuals
Getting Started with ABAQUS/Standard: This document is a self-paced tutorial designed to

help new users become familiar with using ABAQUS/Standard for static and dynamic stress
analysis simulations. It contains a number of fully worked examples that provide practical
guidelines for performing structural analyses with ABAQUS.
Getting Started with ABAQUS/Explicit: This document is a self-paced tutorial designed to help
new users become familiar with using ABAQUS/Explicit. It begins with the basics of modeling in
ABAQUS, so no prior knowledge of ABAQUS is required. A number of fully worked examples
provide practical guidelines for performing explicit dynamic analyses, such as drop tests and metal
forming simulations, with ABAQUS/Explicit.
Lecture Notes: These notes are available on many topics to which ABAQUS is applied. They are
used in the technical seminars that HKS presents to help users improve their understanding and
usage of ABAQUS (see the "Training" section above for more information about these seminars).
While not intended as stand-alone tutorial material, they are sufficiently comprehensive that they
can usually be used in that mode. The list of available lecture notes is included in the
Documentation Price List.
User's Manuals
ABAQUS/Standard User's Manual: This volume contains a complete description of the

elements, material models, procedures, input specifications, etc. It is the basic reference document
for ABAQUS/Standard.
ABAQUS/Explicit User's Manual: This volume contains a complete description of the elements,
material models, procedures, input specifications, etc. It is the basic reference document for
ABAQUS/Explicit.
0-12
ABAQUS/CAE User's Manual: This reference document for ABAQUS/CAE includes three
comprehensive tutorials as well as detailed descriptions of how to use ABAQUS/CAE for model
generation, analysis, and results evaluation.
ABAQUS/Viewer User's Manual: This basic reference document for ABAQUS/Viewer includes
an introductory tutorial as well as a complete description of how to use ABAQUS/Viewer to
display your model and results.
ABAQUS/ADAMS User's Manual: This document describes how to install and how to use
ABAQUS/ADAMS, an interface program that creates ABAQUS models of ADAMS components
and converts the ABAQUS results into an ADAMS modal neutral file that can be used by the
ADAMS/Flex program. It is the basic reference document for the ABAQUS/ADAMS program.
ABAQUS/CAT User's Manual: This document describes how to install and how to use
ABAQUS/CAT, an interface program that creates an ABAQUS input file from a CATIA model
and postprocesses the analysis results in CATIA. It is the basic reference document for the
ABAQUS/CAT program.
ABAQUS/C-MOLD User's Manual: This document describes how to install and how to use
ABAQUS/C-MOLD, an interface program that translates finite element mesh, material property,
and initial stress data from a C-MOLD analysis to an ABAQUS input file.
ABAQUS/Safe User's Manual: This document describes how to install and how to use
ABAQUS/Safe, an interface program that calculates fatigue lives and fatigue strength reserve
factors from finite element models. It is the basic reference document for the ABAQUS/Safe
program. The theoretical background to fatigue analysis is contained in the Modern Metal Fatigue
Analysis manual (available only in print).
Using ABAQUS Online Documentation: This online manual contains instructions on using the
ABAQUS online documentation server to read the manuals that are available online.
ABAQUS Release Notes: This document contains brief descriptions of the new features available
in the latest release of the ABAQUS product line.
ABAQUS Site Guide: This document describes how to install ABAQUS and how to configure
the installation for particular circumstances. Some of this information, of most relevance to users,
is also provided in the user's manuals.
Examples Manuals
ABAQUS Example Problems Manual: This volume contains more than 75 detailed examples
designed to illustrate the approaches and decisions needed to perform meaningful linear and
nonlinear analysis. Typical cases are large motion of an elastic-plastic pipe hitting a rigid wall;
inelastic buckling collapse of a thin-walled elbow; explosive loading of an elastic, viscoplastic thin
ring; consolidation under a footing; buckling of a composite shell with a hole; and deep drawing of
a metal sheet. It is generally useful to look for relevant examples in this manual and to review
them when embarking on a new class of problem.
0-13
ABAQUS Benchmarks Manual: This volume (available online and, if requested, in print)
contains over 200 benchmark problems and standard analyses used to evaluate the performance of
ABAQUS; the tests are multiple element tests of simple geometries or simplified versions of real
problems. The NAFEMS benchmark problems are included in this manual.
ABAQUS Verification Manual: This online-only volume contains more than 5000 basic test
cases, providing verification of each individual program feature (procedures, output options,
MPCs, etc.) against exact calculations and other published results. It may be useful to run these
problems when learning to use a new capability. In addition, the supplied input data files provide
good starting points to check the behavior of elements, materials, etc.
Reference Manuals
ABAQUS Keywords Manual: This volume contains a complete description of all the input
options that are available in ABAQUS/Standard and ABAQUS/Explicit.
ABAQUS Theory Manual: This volume (available online and, if requested, in print) contains
detailed, precise discussions of all theoretical aspects of ABAQUS. It is written to be understood
by users with an engineering background.
ABAQUS Command Language Manual: This online manual provides a description of the
ABAQUS Command Language and a command reference that lists the syntax of each command.
The manual describes how commands can be used to create and analyze ABAQUS/CAE models,
to view the results of the analysis, and to automate repetitive tasks. It also contains information on
using the ABAQUS Command Language or C++ as an application programming interface (API).
ABAQUS Input Files: This online manual contains all the input files that are included with the
ABAQUS release and referred to in the ABAQUS Example Problems Manual, the ABAQUS
Benchmarks Manual, and the ABAQUS Verification Manual. They are listed in the order in which
they appear in the manuals, under the title of the problem that refers to them. The input file
references in the manuals hyperlink directly to this book.
Quality Assurance Plan: This document describes HKS's QA procedures. It is a controlled

document, provided to customers who subscribe to either HKS's Nuclear QA Program or the
Quality Monitoring Service.
Introduction
This is the Example Problems Manual for ABAQUS. It contains many solved examples that illustrate
the use of the program for common types of problems. Some of the problems are quite difficult and
require combinations of the capabilities in the code.
The problems have been chosen to serve two purposes: to verify the capabilities in ABAQUS by
exercising the code on nontrivial cases and to provide guidance to users who must work on a class of
problems with which they are relatively unfamiliar. In each worked example the discussion in the
manual states why the example is included and leads the reader through the standard approach to an
0-14
analysis: element and mesh selection, material model, and a discussion of the results. Input data files
are provided for all of these cases. Many of these problems are worked with different element types,
mesh densities, and other variations. This results in a relatively large number of input data files for
some of the problems. Only a few of the input files are listed in the printed manual. The selection has
been made to provide the most guidance to the user.
All input files, both the ones that are listed in the printed manual and the ones that are referenced, are
included with the ABAQUS release. The ABAQUS/Fetch utility is used to extract these input files
from the compressed archive files provided with the ABAQUS release. For example, to fetch input file
boltpipeflange_3d_cyclsym.inp, type
abaqus fetch job=boltpipeflange_3d_cyclsym.inp
Parametric study script (.psf) and user subroutine ( .f) files can be fetched in the same manner. All
files for a particular problem can be obtained by leaving off the file extension. The ABAQUS/Fetch
execution procedure is explained in detail in `Èxecution procedure for ABAQUS/Fetch,'' Section 3.2.9
of the ABAQUS/Standard User's Manual and the ABAQUS/Explicit User's Manual.
It is sometimes useful to search the input files. The findkeyword utility is used to locate input files
that contain user-specified input. This utility is defined in `Èxecution procedure for querying the
keyword/problem database,'' Section 3.2.8 of the ABAQUS/Standard User's Manual and the
ABAQUS/Explicit User's Manual.
In addition, all the input files included with the ABAQUS release can be accessed through the
ABAQUS Input Files electronic book. This book is part of the ABAQUS online documentation
collection and, as such, is fully searchable (with the exception of numeric strings and
ABAQUS-specific terms). When reading the online version of the ABAQUS Benchmarks Manual, the
ABAQUS Example Problems Manual, or the ABAQUS Verification Manual, the user can click on an
input file name; the ABAQUS Input Files book will open to that file in a separate window.
To reproduce the graphical representation of the solution reported in some of the examples, the output
frequency used in the input files may need to be increased. For example, in ``Linear analysis of the
Indian Point reactor feedwater line,'' Section 2.2.2, the figures that appear in the manual can be
obtained only if the solution is written to the results file every increment; that is, if the input files are
changed to read
*NODE FILE, ..., FREQUENCY=1
instead of FREQUENCY=100 as appears now.
In addition to the Example Problems Manual, there are two other manuals that contain worked
problems. The ABAQUS Benchmarks Manual contains benchmark problems (including the NAFEMS
suite of test problems) and standard analyses used to evaluate the performance of ABAQUS. The tests
in this manual are multiple element tests of simple geometries or simplified versions of real problems.
The ABAQUS Verification Manual contains a large number of examples that are intended as
elementary verification of the basic modeling capabilities.
The verification of ABAQUS consists of running the problems in the ABAQUS Example Problems
Manual, the ABAQUS Benchmarks Manual, and the ABAQUS Verification Manual. Before a version
0-15
of ABAQUS is released, it must run all verification, benchmark, and example problems correctly.
0-16
Static Stress/Displacement Analyses
1. Static Stress/Displacement Analyses

1.1 Static and quasi-static stress analyses
1.1.1 Axisymmetric analysis of bolted pipe flange connections
Product: ABAQUS/Standard
A bolted pipe flange connection is a common and important part of many piping systems. Such
connections are typically composed of hubs of pipes, pipe flanges with bolt holes, sets of bolts and
nuts, and a gasket. These components interact with each other in the tightening process and when
operation loads such as internal pressure and temperature are applied. Experimental and numerical
studies on different types of interaction among these components are frequently reported. The studies
include analysis of the bolt-up procedure that yields uniform bolt stress (Bibel and Ezell, 1992),
contact analysis of screw threads (Fukuoka, 1992; Chaaban and Muzzo, 1991), and full stress analysis
of the entire pipe joint assembly (Sawa et al., 1991). To establish an optimal design, a full stress
analysis determines factors such as the contact stresses that govern the sealing performance, the
relationship between bolt force and internal pressure, the effective gasket seating width, and the
bending moment produced in the bolts. This example shows how to perform such a design analysis by
using an economical axisymmetric model and how to assess the accuracy of the axisymmetric solution
by comparing the results to those obtained from a simulation using a three-dimensional segment
model. In addition, several three-dimensional models that use multiple levels of superelements are
analyzed to demonstrate the use of superelements with a large number of retained degrees of freedom.
Geometry and model

The bolted joint assembly being analyzed is depicted in Figure 1.1.1-1. The geometry and dimensions
of the various parts are taken from Sawa et al. (1991), modified slightly to simplify the modeling. The
inner wall radius of both the hub and the gasket is 25 mm. The outer wall radii of the pipe flange and
the gasket are 82.5 mm and 52.5 mm, respectively. The thickness of the gasket is 2.5 mm. The pipe
flange has eight bolt holes that are equally spaced in the pitch circle of radius 65 mm. The radius of the
bolt hole is modified in this analysis to be the same as that of the bolt: 8 mm. The bolt head (bearing
surface) is assumed to be circular, and its radius is 12 mm.
The Young's modulus is 206 GPa and the Poisson's ratio is 0.3 for both the bolt and the pipe
hub/flange. The gasket is modeled with either solid continuum or gasket elements. When continuum
elements are used, the gasket's Young's modulus, E, equals 68.7 GPa and its Poisson's ratio, º, equals
0.3.
When gasket elements are used, a linear gasket pressure/closure relationship is used with the effective
"normal stiffness," Sn , equal to the material Young's modulus divided by the thickness so that Sn =
27.48 GPa/mm. Similarly a linear shear stress/shear motion relationship is used with an effective shear
stiffness, St , equal to the material shear modulus divided by the thickness so that St = 10.57 GPa/mm.
The membrane behavior is specified with a Young's modulus of 68.7 GPa and a Poisson's ratio of 0.3.
Sticking contact conditions are assumed in all contact areas: between the bearing surface and the
flange and between the gasket and the hub. Contact between the bolt shank and the bolt hole is
1-17
ignored.
The finite element idealizations of the symmetric half of the pipe joint are shown in Figure 1.1.1-2 and
Figure 1.1.1-3, corresponding to the axisymmetric and three-dimensional analyses, respectively. The
mesh used for the axisymmetric analysis consists of a mesh for the pipe hub/flange and gasket and a
separate mesh for the bolts. In Figure 1.1.1-2the top figure shows the mesh of the pipe hub and flange,
with the bolt hole area shown in a lighter shade; and the bottom figure shows the overall mesh with the
gasket and the bolt in place.
For the axisymmetric model second-order elements with reduced integration, CAX8R, are used
throughout the mesh of the pipe hub/flange. The gasket is modeled with either CAX8R solid
continuum elements or GKAX6 gasket elements. Contact between the gasket and the pipe hub/flange
is modeled with contact pairs between surfaces defined on the faces of elements in the contact region
or between such element-based surfaces and node-based surfaces. In an axisymmetric analysis the
bolts and the perforated flange must be modeled properly. The bolts are modeled as plane stress
elements since they do not carry hoop stress. Second-order plane stress elements with reduced
integration, CPS8R, are employed for this purpose. The contact surface definitions, which are
associated with the faces of the elements, account for the plane stress condition automatically. To
account for all eight bolts used in the joint, the combined cross-sectional areas of the shank and the
head of the bolts must be calculated and redistributed to the bolt mesh appropriately using the area
attributes for the solid elements. The contact area is adjusted automatically.
Figure 1.1.1-4 illustrates the cross-sectional views of the bolt head and the shank. Each plane stress
element represents a volume that extends out of the x-y plane. For example, element A represents a
volume calculated as (HA ) ´ (AreaA ). Likewise, element B represents a volume calculated as (HB ) ´
(AreaB ). The sectional area in the x-z plane pertaining to a given element can be calculated as
Z X2 1 1 x ¯¯X2
Area = 2 [(R2 ¡ x2 ) 2 ]dx = [x(R2 ¡ x2 ) 2 + R2 arcsin ( )] ;
X1 jRj X1
where R is the bolt head radius, Rbolthead , or the shank radius, Rshank (depending on the element
location), and X1 and X2 are x-coordinates of the left and right side of the given element,
respectively.
If the sectional areas are divided by the respective element widths, WA and WB , we obtain
representative element thicknesses. Multiplying each element thickness by eight (the number of bolts
in the model) produces the thickness values that are found in the *SOLID SECTION options.
Sectional areas that are associated with bolt head elements located on the model's contact surfaces are
used to calculate the surface areas of the nodes used in defining the node-based surfaces of the model.
Referring again to Figure 1.1.1-4, nodal contact areas for a single bolt are calculated as follows:
AC AF
A1 = ; A9 = ;
4 4
1-18
AC AD AE AF
A2 = ; A4 = ; A6 = ; A8 = ;
2 2 2 2
A3 = (AC + AD )=4; A5 = (AD + AE )=4; A7 = (AE + AF )=4;
where A1 through A9 are contact areas that are associated with contact nodes 1-9 and Ac through AF
are sectional areas that are associated with bolt head elements C-F . Multiplying the above areas by
eight (the number of bolts in the model) provides the nodal contact areas found under the *SURFACE
INTERACTION options.
A common way of handling the presence of the bolt holes in the pipe flange in axisymmetric analyses
is to smear the material properties used in the bolt hole area of the mesh and to use inhomogeneous
material properties that correspond to a weaker material in this region. General guidelines for
determining the effective material properties for perforated flat plates are found in ASME Section VIII
Div 2 Article 4-9. For the type of structure under study, which is not a flat plate, a common approach
to determining the effective material properties is to calculate the elasticity moduli reduction factor,
which is the ratio of the ligament area in the pitch circle to the annular area of the pitch circle. In this
model the annular area of the pitch circle is given by AA = 6534.51 mm2, and the total area of the bolt
holes is given by AH = 8¼ 82 = 1608.5 mm2. Hence, the reduction factor is simply 1 ¡ AH=AA =
0.754. The effective in-plane moduli of elasticity, E 10 and E 20 , are obtained by multiplying the
respective moduli, E 1 and E 2, by this factor. We assume material isotropy in the r-z plane; thus,
E 10 = E 20 = E 0 : The modulus in the hoop direction, E 30 , should be very small and is chosen such
that E 0 =E 30 = 106. The in-plane shear modulus is then calculated based on the effective elasticity
modulus: G012 = E 0 =2(1 + º ): The shear moduli in the hoop direction are also calculated similarly but
with º set to zero (they are not used in an axisymmetric model). Hence, we have E 10 = E 20 = 155292
MPa, E 30 = 0.155292 MPa, G012 = 59728 MPa, and G013 = G023 = 0.07765 MPa. These elasticity
moduli are specified using *ELASTIC, TYPE=ENGINEERING CONSTANTS for the bolt hole part of
the mesh.
The mesh for the three-dimensional analysis without superelements, shown in Figure 1.1.1-3,
represents a 22.5° segment of the pipe joint and employs second-order brick elements with reduced
integration, C3D20R, for the pipe hub/flange and bolts. The gasket is modeled with C3D20R elements
or GK3D18 elements. The top figure shows the mesh of the pipe hub and flange, and the bottom figure
shows both the gasket and bolt (in the lighter color). Contact is modeled by the interaction of contact
surfaces defined by grouping specific faces of the elements in the contacting regions. For
three-dimensional contact where both the master and slave surfaces are deformable, the SMALL
SLIDING parameter must be used on the *CONTACT PAIR option to indicate that small relative
sliding occurs between contacting surfaces. No special adjustments need be made for the material
properties used in the three-dimensional model because all parts are modeled appropriately.
Four different meshes that use superelements to model the flange are tested. A first-level superelement
is created for the entire 22.5° segment of the flange shown in Figure 1.1.1-3, while the gasket and the
bolt are meshed as before. The nodes on the flange in contact with the bolt cap form a node-based
surface, while the nodes on the flange in contact with the gasket form another node-based surface.
1-19
These node-based surfaces will form contact pairs with the master surfaces on the bolt cap and on the
gasket, which are defined with *SURFACE as before. The retained degrees of freedom on the
superelement include all three degrees of freedom for the nodes in these node-based surfaces as well as
for the nodes on the 0° and 22.5° faces of the flange. Appropriate boundary conditions are specified at
the superelement usage level.
A second-level superelement of 45° is created by reflecting the first-level superelement with respect to
the 22.5° plane. The nodes on the 22.5° face belonging to the reflected superelement are constrained in
all three degrees of freedom to the corresponding nodes on the 22.5° face belonging to the original
first-level superelement. The half-bolt and the gasket sector corresponding to the reflected
superelement are also constructed by reflection. The retained degrees of freedom include all three
degrees of freedom of all contact nodes sets and of the nodes on the 0° and 45° faces of the flange.
MPC-type CYCLSYM is used to impose cyclic symmetric boundary conditions on these two faces.
A third-level superelement of 90° is created by reflecting the original 45° second-level superelement
with respect to the 45° plane and by connecting it to the original 45° superelement. The remaining part
of the gasket and the bolts corresponding to the 45° - 90° sector of the model is created by reflection
and appropriate constraints. In this case it is not necessary to retain any degrees of freedom on the 0°
and 90° faces of the flange because this 90° superelement will not be connected to other superelements
and appropriate boundary conditions can be specified at the superelement creation level.
The final model is set up by mirroring the 90° mesh with respect to the symmetry plane of the gasket
perpendicular to the y-axis. Thus, an otherwise large analysis (¼ 750,000 unknowns) when no
superelements are used can be solved conveniently ( ¼ 80,000 unknowns) by using the third-level
superelement twice. The sparse solver is used because it significantly reduces the run time for this
model.
Loading and boundary conditions

The only boundary conditions are symmetry boundary conditions. In the axisymmetric model uz = 0 is
applied to the symmetry plane of the gasket and to the bottom of the bolts. In the three-dimensional
model uy = 0 is applied to the symmetry plane of the gasket as well as to the bottom of the bolt. The
µ =0° and µ =22.5° planes are also symmetry planes. On the µ =22.5° plane, symmetry boundary
conditions are enforced by invoking suitable nodal transformations and applying boundary conditions
to local directions in this symmetry plane. These transformations are implemented using the
*TRANSFORM option. On both the symmetry planes, the symmetry boundary conditions uz = 0 are
imposed everywhere except for the dependent nodes associated with the C BIQUAD MPC and nodes
on one side of the contact surface. The second exception is made to avoid overconstraining problems,
which arise if there is a boundary condition in the same direction as a Lagrange multiplier constraint
associated with the *FRICTION, ROUGH option.
In the models where superelements are used, the boundary conditions are specified depending on what
superelement is used. For the first-level 22.5° superelement the boundary conditions and constraint
equations were the same as for the three-dimensional model shown in Figure 1.1.1-3. For the 45°
second-level superelement the symmetry boundary conditions are enforced on the µ =45° plane with
the constraint equation uz + ux = 0. A transform could have been used as well. For the 90° third-level
1-20
superelement the face µ =90° is constrained with the boundary condition ux = 0.

A clamping force of 15 kN is applied to each bolt by using the *PRE-TENSION SECTION option. The
pre-tension section is identified by means of the *SURFACE option. The pre-tension is then
prescribed by applying a concentrated load to the pre-tension node. In the axisymmetric analysis the
actual load applied is 120 kN since there are eight bolts. In the three-dimensional model with no
superelements the actual load applied is 7.5 kN since only half of a bolt is modeled. In the models
using superelements all half-bolts are loaded with a 7.5 kN force. For all of the models the pre-tension
section is specified about half-way down the bolt shank.
Sticking contact conditions are assumed in all surface interactions in all analyses and are simulated
with the *FRICTION, ROUGH and *SURFACE BEHAVIOR, NO SEPARATION options.
Results and discussion

All analyses are performed as small-displacement analyses.
Figure 1.1.1-5 shows a top view of the normal stress distributions in the gasket at the interface between
the gasket and the pipe hub/flange predicted by the axisymmetric (bottom) and three-dimensional (top)
analyses when solid continuum elements are used to model the gasket. The figure shows that the
compressive normal stress is highest at the outer edge of the gasket, decreases radially inward, and
changes from compression to tension at a radius of about 35 mm, which is consistent with findings
reported by Sawa et al. (1991). The close agreement in the overall solution between axisymmetric and
three-dimensional analyses is quite apparent, indicating that, for such problems, axisymmetric analysis
offers a simple yet reasonably accurate alternative to three-dimensional analysis.
Figure 1.1.1-6 shows a top view of the normal stress distributions in the gasket at the interface between
the gasket and the pipe hub/flange predicted by the axisymmetric (bottom) and three-dimensional (top)
analyses when gasket elements are used to model the gasket. Close agreement in the overall solution
between the axisymmetric and three-dimensional analyses is also seen in this case. The gasket starts
carrying compressive load at a radius of about 40 mm, a difference of 5 mm with the previous result.
This difference is the result of the gasket elements being unable to carry tensile loads in their thickness
direction. This solution is physically more realistic since, in most cases, gaskets separate from their
neighboring parts when subjected to tensile loading. Removing the *SURFACE BEHAVIOR, NO
SEPARATION option from the gasket/flange contact surface definition in the input files that model the
gasket with continuum elements yields good agreement with the results obtained in Figure 1.1.1-6
(since, in that case, the solid continuum elements in the gasket cannot carry tensile loading in the
gasket thickness direction).
The models in this example can be modified to study other factors, such as the effective seating width
of the gasket or the sealing performance of the gasket under operating loads. The gasket elements offer
the advantage of allowing very complex behavior to be defined in the gasket thickness direction.
Gasket elements can also use any of the small-strain material models provided in ABAQUS including
user-defined material models. Figure 1.1.1-7shows a comparison of the normal stress distributions in
the gasket at the interface between the gasket and the pipe hub/flange predicted by the axisymmetric
(bottom) and three-dimensional (top) analyses when isotropic material properties are prescribed for
gasket elements. The results in Figure 1.1.1-7compare well with the results in Figure 1.1.1-5 from
1-21
analyses in which solid and axisymmetric elements are used to simulate the gasket.
Figure 1.1.1-8 shows the distribution of the normal stresses in the gasket at the interface in the plane
z = 0. The results are plotted for the three-dimensional model containing only solid continuum
elements and no superelements and for the four models containing the superelements described above.
Input files
boltpipeflange_axi_solidgask.inp
Axisymmetric analysis containing a gasket modeled with solid continuum elements.
boltpipeflange_axi_node.inp
Node definitions for boltpipeflange_axi_solidgask.inp and boltpipeflange_axi_gkax6.inp.
boltpipeflange_axi_element.inp
Element definitions for boltpipeflange_axi_solidgask.inp.
boltpipeflange_3d_solidgask.inp
Three-dimensional analysis containing a gasket modeled with solid continuum elements.
boltpipeflange_axi_gkax6.inp
Axisymmetric analysis containing a gasket modeled with gasket elements.
boltpipeflange_3d_gk3d18.inp
Three-dimensional analysis containing a gasket modeled with gasket elements.
boltpipeflange_3d_super1.inp
Three-dimensional analysis using the first-level superelement (22.5° model).
boltpipeflange_3d_super2.inp
Three-dimensional analysis using the second-level superelement (45° model).
boltpipeflange_3d_super3_1.inp
Three-dimensional analysis using the third-level superelement once (90° model).
boltpipeflange_3d_super3_2.inp
Three-dimensional analysis using the third-level superelement twice (90° mirrored model).
boltpipeflange_3d_gen1.inp
First-level superelement generation data referenced by boltpipeflange_3d_super1.inp and
boltpipeflange_3d_gen2.inp.
Second-level superelement generation data referenced by boltpipeflange_3d_super2.inp and
1-22
Third-level superelement generation data referenced by boltpipeflange_3d_super3_1.inp and

boltpipeflange_3d_super3_2.inp.
boltpipeflange_3d_node.inp
Nodal coordinates used in boltpipeflange_3d_super1.inp, boltpipeflange_3d_super2.inp,
boltpipeflange_3d_super3_1.inp, boltpipeflange_3d_super3_2.inp,
boltpipeflange_3d_cyclsym.inp, boltpipeflange_3d_gen1.inp, boltpipeflange_3d_gen2.inp, and
boltpipeflange_3d_cyclsym.inp
Same as file boltpipeflange_3d_super2.inp except that CYCLSYM type MPCs are used.
boltpipeflange_3d_missnode.inp
Same as file boltpipeflange_3d_gk3d18.inp except that the option to generate missing nodes is
used for gasket elements.
boltpipeflange_3d_isomat.inp
Same as file boltpipeflange_3d_gk3d18.inp except that gasket elements are modeled as isotropic
using the *MATERIAL option.
boltpipeflange_3d_ortho.inp
Same as file boltpipeflange_3d_gk3d18.inp except that gasket elements are modeled as orthotropic
and the *ORIENTATION option is used.
boltpipeflange_axi_isomat.inp
Same as file boltpipeflange_axi_gkax6.inp except that gasket elements are modeled as isotropic
using the *MATERIAL option.
boltpipeflange_3d_usr_umat.inp
User subroutine UMAT used in boltpipeflange_3d_usr_umat.inp.
boltpipeflange_3d_usr_umat.f
Same as file boltpipeflange_3d_gk3d18.inp except that gasket elements are modeled as isotropic
with user subroutine UMAT.
boltpipeflange_3d_solidnum.inp
Same as file boltpipeflange_3d_gk3d18.inp except that solid element numbering is used for gasket
elements.
References
· Bibel, G. D., and R. M. Ezell, `Àn Improved Flange Bolt-Up Procedure Using Experimentally
Determined Elastic Interaction Coefficients,'' Journal of Pressure Vessel Technology, vol. 114,
pp. 439-443, 1992.
· Chaaban, A., and U. Muzzo, ``Finite Element Analysis of Residual Stresses in Threaded End
1-23
Closures,'' Transactions of ASME, vol. 113, pp. 398-401, 1991.
· Fukuoka, T., ``Finite Element Simulation of Tightening Process of Bolted Joint with a
Tensioner,'' Journal of Pressure Vessel Technology, vol. 114, pp. 433-438, 1992.
· Sawa, T., N. Higurashi, and H. Akagawa, `À Stress Analysis of Pipe Flange Connections,''
Journal of Pressure Vessel Technology, vol. 113, pp. 497-503, 1991.
Figures
Figure 1.1.1-1 Schematic of the bolted joint. All dimensions in mm.
Figure 1.1.1-2 Axisymmetric model of the bolted joint.
1-24
Figure 1.1.1-3 22.5° segment three-dimensional model of the bolted joint.
1-25
Figure 1.1.1-4 Cross-sectional views of the bolt head and the shank.
1-26
Figure 1.1.1-5 Normal stress distribution in the gasket contact surface when solid elements are used
to model the gasket: three-dimensional versus axisymmetric results.
1-27
Figure 1.1.1-6 Normal stress distribution in the gasket contact surface when gasket elements are used
with direct specification of the gasket behavior: three-dimensional versus axisymmetric results.
1-28
Figure 1.1.1-7 Normal stress distribution in the gasket contact surface when gasket elements are used
with isotropic material properties: three-dimensional versus axisymmetric results.
1-29
Figure 1.1.1-8 Normal stress distribution in the gasket contact surface along the line z = 0 for the
models with and without superelements.
Sample listings
1-30
Listing 1.1.1-1
*HEADING
BOLTED PIPE JOINT: AXISYMMETRIC MODEL
*RESTART, WRITE, FREQUENCY=1
*WAVEFRONT MINIMIZATION
*NODE, INPUT=boltpipeflange_axi_node.inp
*ELEMENT, TYPE=CAX8R, INPUT=boltpipeflange_axi_element.inp
*ELSET, ELSET=PID1, GENERATE
609,640
42,48
50,56
58,64
66,168
193,216
477,484
577,608
641,704
**
** Contact Between Gasket and Hub
**
485,492
*SURFACE,NAME=HUB_BOT
477,S3
478,S1
479,S3
480,S1
481,S3
482,S1
483,S3
484,S1
*SURFACE,NAME=GASKET
PID3,S4
*CONTACT PAIR, INTERACTION=ROUGH, SMALL SLIDING, ADJUST=.1
HUB_BOT,GASKET
*SURFACE INTERACTION,NAME=ROUGH
*FRICTION,ROUGH
*SURFACE BEHAVIOR, NO SEPARATION
**
** Contact Between Bolt and Hub
**
1-31
** Note: Areas associated with contact nodes are determined

** by first calculating out-of-plane surface areas represented
** by the contact faces of bottom-side bolthead elements, and
** then assigning ratios of these areas to the relative contact
** nodes which lie on the faces.
**
*SURFACE,type=node,NAME=NBOLT1
5008,
5007,
5006,
5005,
5004,
5003,
5002,
5001,
5000,
*SURFACE,type=node,NAME=NBOLT1B
5017,
5016,
5015,
5014,
5013,
5012,
5011,
5010,
5009,
1-32
577,580
*ELSET, ELSET=PID36B, GENERATE
641,644
*SURFACE,NAME=HUB_BOLT
PID36,S4
*SURFACE,NAME=HUBBOLTB
PID36B,S4
*CONTACT PAIR, INTERACTION=BLT_HUB1, SMALL SLIDING, HCRIT=1.1
NBOLT1,HUB_BOLT
NBOLT1B,HUBBOLTB
*SURFACE INTERACTION,NAME=BLT_HUB1
12.899,
*FRICTION,ROUGH
NBOLT2,HUB_BOLT
NBOLT2B,HUBBOLTB
25.799,
*FRICTION,ROUGH
NBOLT3,HUB_BOLT
NBOLT3B,HUBBOLTB
36.012,
*FRICTION,ROUGH
NBOLT4,HUB_BOLT
NBOLT4B,HUBBOLTB
46.226,
*FRICTION,ROUGH
NBOLT5,HUB_BOLT
NBOLT5B,HUBBOLTB
52.378,
*FRICTION,ROUGH
1-33
NBOLT6,HUB_BOLT
NBOLT6B,HUBBOLTB
58.529,
*FRICTION,ROUGH
NBOLT7,HUB_BOLT
NBOLT7B,HUBBOLTB
63.107,
*FRICTION,ROUGH
NBOLT8,HUB_BOLT
NBOLT8B,HUBBOLTB
67.685,
*FRICTION,ROUGH
NBOLT9,HUB_BOLT
NBOLT9B,HUBBOLTB
33.842,
*FRICTION,ROUGH
**
** Mesh Refinement
**
*MPC
QUADR, 415, 406, 432, 458
QUADR, 441, 406, 432, 458
QUADR, 467, 458, 484, 510
QUADR, 493, 458, 484, 510
QUADR, 519, 510, 536, 562
QUADR, 545, 510, 536, 562
QUADR, 571, 562, 588, 614
QUADR, 597, 562, 588, 614
QUADR, 1966, 150, 166, 176
QUADR, 1953, 150, 166, 176
QUADR, 1950, 176, 192, 202
QUADR, 1937, 176, 192, 202
1-34
QUADR, 1934, 202, 218, 228

QUADR, 1921, 202, 218, 228
QUADR, 1918, 228, 244, 254
QUADR, 1905, 228, 244, 254
*ELEMENT, TYPE=CPS8R, ELSET=PID7
496, 2016, 2018, 2032, 2030, 2017, 2023, 2031, 2022
500, 2030, 2032, 2046, 2044, 2031, 2037, 2045, 2036
504, 2044, 2046, 2060, 2058, 2045, 2051, 2059, 2050
508, 2060, 5008, 5006, 2058, 2065, 5007, 2064, 2059
495, 2014, 2016, 2030, 2028, 2015, 2022, 2029, 2021
499, 2028, 2030, 2044, 2042, 2029, 2036, 2043, 2035
503, 2042, 2044, 2058, 2056, 2043, 2050, 2057, 2049
507, 2058, 5006, 5004, 2056, 2064, 5005, 2063, 2057
494, 2012, 2014, 2028, 2026, 2013, 2021, 2027, 2020
498, 2026, 2028, 2042, 2040, 2027, 2035, 2041, 2034
502, 2040, 2042, 2056, 2054, 2041, 2049, 2055, 2048
506, 2056, 5004, 5002, 2054, 2063, 5003, 2062, 2055
493, 2010, 2012, 2026, 2024, 2011, 2020, 2025, 2019
497, 2024, 2026, 2040, 2038, 2025, 2034, 2039, 2033
501, 2038, 2040, 2054, 2052, 2039, 2048, 2053, 2047
505, 2054, 5002, 5000, 2052, 2062, 5001, 2061, 2053
509, 2010, 2024, 2088, 2076, 2019, 2087, 2083, 2075
513, 2024, 2038, 2100, 2088, 2033, 2099, 2095, 2087
517, 2038, 2052, 2112, 2100, 2047, 2111, 2107, 2099
521, 2052, 5000, 2124, 2112, 2061, 2123, 2119, 2111
510, 2076, 2088, 2090, 2078, 2083, 2089, 2084, 2077
514, 2088, 2100, 2102, 2090, 2095, 2101, 2096, 2089
518, 2100, 2112, 2114, 2102, 2107, 2113, 2108, 2101
522, 2112, 2124, 2126, 2114, 2119, 2125, 2120, 2113
511, 2078, 2090, 2092, 2080, 2084, 2091, 2085, 2079
515, 2090, 2102, 2104, 2092, 2096, 2103, 2097, 2091
519, 2102, 2114, 2116, 2104, 2108, 2115, 2109, 2103
523, 2114, 2126, 2128, 2116, 2120, 2127, 2121, 2115
512, 2080, 2092, 2094, 2082, 2085, 2093, 2086, 2081
516, 2092, 2104, 2106, 2094, 2097, 2105, 2098, 2093
520, 2104, 2116, 2118, 2106, 2109, 2117, 2110, 2105
1-35
524, 2116, 2128, 5009, 2118, 2121, 2129, 2122, 2117

525, 2082, 2094, 2144, 2132, 2086, 2143, 2139, 2131
529, 2094, 2106, 2156, 2144, 2098, 2155, 2151, 2143
533, 2106, 2118, 2168, 2156, 2110, 2167, 2163, 2155
537, 2118, 5009, 5011, 2168, 2122, 5010, 2175, 2167
526, 2132, 2144, 2146, 2134, 2139, 2145, 2140, 2133
530, 2144, 2156, 2158, 2146, 2151, 2157, 2152, 2145
534, 2156, 2168, 2170, 2158, 2163, 2169, 2164, 2157
538, 2168, 5011, 5013, 2170, 2175, 5012, 2176, 2169
527, 2134, 2146, 2148, 2136, 2140, 2147, 2141, 2135
531, 2146, 2158, 2160, 2148, 2152, 2159, 2153, 2147
535, 2158, 2170, 2172, 2160, 2164, 2171, 2165, 2159
539, 2170, 5013, 5015, 2172, 2176, 5014, 2177, 2171
528, 2136, 2148, 2150, 2138, 2141, 2149, 2142, 2137
532, 2148, 2160, 2162, 2150, 2153, 2161, 2154, 2149
536, 2160, 2172, 2174, 2162, 2165, 2173, 2166, 2161
540, 2172, 5015, 5017, 2174, 2177, 5016, 2178, 2173
541, 5000, 2192, 2194, 2124, 2187, 2193, 2188, 2123
545, 2192, 2206, 2208, 2194, 2201, 2207, 2202, 2193
549, 2206, 2220, 2222, 2208, 2215, 2221, 2216, 2207
553, 2220, 2234, 2236, 2222, 2229, 2235, 2230, 2221
557, 2234, 2248, 2250, 2236, 2243, 2249, 2244, 2235
561, 2248, 2262, 2264, 2250, 2257, 2263, 2258, 2249
565, 2262, 2276, 2278, 2264, 2271, 2277, 2272, 2263
569, 2276, 2290, 2292, 2278, 2285, 2291, 2286, 2277
573, 2290, 2304, 2306, 2292, 2299, 2305, 2300, 2291
542, 2124, 2194, 2196, 2126, 2188, 2195, 2189, 2125
546, 2194, 2208, 2210, 2196, 2202, 2209, 2203, 2195
550, 2208, 2222, 2224, 2210, 2216, 2223, 2217, 2209
554, 2222, 2236, 2238, 2224, 2230, 2237, 2231, 2223
558, 2236, 2250, 2252, 2238, 2244, 2251, 2245, 2237
562, 2250, 2264, 2266, 2252, 2258, 2265, 2259, 2251
566, 2264, 2278, 2280, 2266, 2272, 2279, 2273, 2265
570, 2278, 2292, 2294, 2280, 2286, 2293, 2287, 2279
574, 2292, 2306, 2308, 2294, 2300, 2307, 2301, 2293
543, 2126, 2196, 2198, 2128, 2189, 2197, 2190, 2127
1-36
547, 2196, 2210, 2212, 2198, 2203, 2211, 2204, 2197

551, 2210, 2224, 2226, 2212, 2217, 2225, 2218, 2211
555, 2224, 2238, 2240, 2226, 2231, 2239, 2232, 2225
559, 2238, 2252, 2254, 2240, 2245, 2253, 2246, 2239
563, 2252, 2266, 2268, 2254, 2259, 2267, 2260, 2253
567, 2266, 2280, 2282, 2268, 2273, 2281, 2274, 2267
571, 2280, 2294, 2296, 2282, 2287, 2295, 2288, 2281
575, 2294, 2308, 2310, 2296, 2301, 2309, 2302, 2295
544, 2128, 2198, 2200, 5009, 2190, 2199, 2191, 2129
548, 2198, 2212, 2214, 2200, 2204, 2213, 2205, 2199
552, 2212, 2226, 2228, 2214, 2218, 2227, 2219, 2213
556, 2226, 2240, 2242, 2228, 2232, 2241, 2233, 2227
560, 2240, 2254, 2256, 2242, 2246, 2255, 2247, 2241
564, 2254, 2268, 2270, 2256, 2260, 2269, 2261, 2255
568, 2268, 2282, 2284, 2270, 2274, 2283, 2275, 2269
572, 2282, 2296, 2298, 2284, 2288, 2297, 2289, 2283
576, 2296, 2310, 2312, 2298, 2302, 2311, 2303, 2297
*ELSET,ELSET=PIN
PID7,PID8,PID9,PID10,PID11,PID12,PID13,PID14,PID15,PID16,PID17,
PID18,PID19,PID20,PID21,PID22
** flange_elements
*SOLID SECTION, ELSET=PID2, MATERIAL=MID2
1.,
**
** Note: Thicknesses of plane stress bolthead elements are
** determined by first calculating the areas which the 3-D
** volumes represented by the 2-D elements project onto the
** 1-3 plane, and then dividing the areas by the respective
** element widths.
**
** bolthead_1
51.5976,
** bolthead_12
51.5976,
** bolthead_2
92.4522,
** bolthead_11
92.4522,
1-37
** bolthead_3
117.058,
** bolthead_10
117.058,
** bolthead_4
135.37,
** bolthead_9
135.37,
** bolthead_5
164.887,
** bolthead_8
164.887,
** bolthead_6
188.382,
** bolthead_7
188.382,
** gasket_elements
1.,
** bolttrunk_1
78.6157,
** bolttrunk_4
78.6157,
** bolttrunk_2
122.446,
** bolttrunk_3
122.446,
** hole_elements
*SOLID SECTION, ELSET=PID1, MATERIAL=MID4,ORIENT=RECT
1.,
**local orientation matching global system
1-38
*ORIENTATION,NAME=RECT
1.0, 0.0, 0.0, 0.0, 1.0, 0.0
1, 0.0
** bolt_material
*MATERIAL, NAME=MID1
*ELASTIC, TYPE=ISOTROPIC
2.06E+5, 0.3
** BOLTS ASSUME PLANE STRESS CONDITIONS.
** THERE ARE 8 BOLTS SO A=8* BOLT CROSS-SECTION AREA
** flange_material
2.06E+5, 0.3
** gasket_material
.687E+5, 0.3
** hole_material
*ELASTIC, TYPE=ENGINEERING CONSTANTS
155.29E3, 155.29E3, 155.29E-3, 0.3, 0.0, 0.0, 59.728E3, 77.65E-3,
77.65E-3,
** HOOP/INPLANE DIRECT MODULI RATIO = 1.E6
*NSET,NSET=GASKTEND,GENERATE
1993,2009
*NSET,NSET=BOLTEND,GENERATE
2304,2312
*NSET,NSET=BOLTMID
2308,
*NSET,NSET=NOUT
2078,5008,2126,5017,2308
747,876,5100,2438,5117,2684,2796,9,5,1,34,26,18
*BOUNDARY
GASKTEND,2
BOLTEND,2
**
** Pre-tension section
**
*NSET,NSET=NSECT
6001,
*ELSET,ELSET=ESECT,GEN
557,560,1
*SURFACE,NAME=PSECT
1-39
ESECT,S2
*PRE-TENSION SECTION,SURFACE=PSECT,NODE=6001
**
*STEP, AMPLITUDE=RAMP, INC=20
*STATIC
1.,1.
*CLOAD
6001,1,120000.
*NODE PRINT,FREQ=1,NSET=NSECT
U,RF
*NODE PRINT, FREQUENCY=1,NSET=BOLTEND,TOTALS=YES
U,RF
*NODE PRINT, FREQUENCY=1,NSET=GASKTEND,TOTALS=YES
U,RF
*NODE PRINT, FREQUENCY=1,NSET=NOUT
U,RF
*NODE FILE, NSET=NOUT, FREQUENCY=1
U, RF
*CONTACT PRINT,SLAVE=HUB_BOT, MASTER=GASKET
*CONTACT FILE,SLAVE=HUB_BOT, MASTER=GASKET
*CONTACT PRINT,MASTER=HUB_BOLT
*CONTACT PRINT,MASTER=HUBBOLTB
*PRINT, CONTACT=YES
*END STEP
1.1.2 Elastic-plastic collapse of a thin-walled elbow under

in-plane bending and internal pressure
Elbows are used in piping systems because they ovalize more readily than straight pipes and, thus,
provide flexibility in response to thermal expansion and other loadings that impose significant
displacements on the system. Ovalization is the bending of the pipe wall into an oval, noncircular
configuration. The elbow is, thus, behaving as a shell rather than as a beam. Straight pipe runs do not
ovalize easily, so they behave essentially as beams. Thus, even under pure bending, complex
interaction occurs between an elbow and the adjacent straight pipe segments; the elbow causes some
ovalization in the straight pipe runs, which in turn tend to stiffen the elbow. This interaction can create
significant axial gradients of bending strain in the elbow, especially in cases where the elbow is very
flexible. This example provides verification of shell and elbow element modeling of such effects,
through an analysis of a test elbow for which experimental results have been reported by Sobel and
Newman (1979). An analysis is also included with elements of type ELBOW31B (which includes
1-40
ovalization but neglects axial gradients of strain) for the elbow itself, and beam elements for the
straight pipe segments. This provides a comparative solution in which the interaction between the
elbow and the adjacent straight pipes is neglected. The analyses predict the response up to quite large
rotations across the elbow, so as to investigate possible collapse of the pipe and, particularly, the effect
of internal pressure on that collapse.
Geometry and model

The elbow configuration used in the study is shown in Figure 1.1.2-1. It is a thin walled elbow with
elbow factor
Rt
¸= p = 0:167
r2 1 ¡ º2
and radius ratio R=r =3.07, so the flexibility factor from Dodge and Moore (1972) is 10.3. (The
flexibility factor for an elbow is the ratio of the bending flexibility of an elbow segment to a straight
pipe of the same dimensions, for small displacements and elastic response.) This is an extremely
flexible case because the pipe wall is so thin.
To demonstrate convergence of the overall moment-rotation behavior with respect to meshing, the two
shell element meshes shown in Figure 1.1.2-2 are analyzed. Since the loading concerns in-plane
bending only, it is assumed that the response is symmetric about the midplane of the system so that, in
the shell element model, only one-half of the system need be modeled. Element type S8R5 is used,
since tests have shown this to be the most cost-effective shell element in ABAQUS (input files using
element types S9R5, STRI65, and S8R for this example are included with the ABAQUS release). The
elbow element meshes replace each axial division in the more coarse shell element model with one
ELBOW32 or two ELBOW31 elements and use 4 or 6 Fourier modes to model the deformation around
the pipe. Seven integration points are used through the pipe wall in all the analyses. This is usually
adequate to provide accurate modeling of the progress of yielding through the section in such cases as
these, where essentially monotonic straining is expected.
The ends of the system are rigidly attached to stiff plates in the experiments. These boundary
conditions are easily modeled for the ELBOW elements and for the fixed end in the shell element
model. For the rotating end of the shell element model the shell nodes must be constrained to a beam
node that represents the motion of the end plate. This is done using the *KINEMATIC COUPLING
option as described below.
The material is assumed to be isotropic and elastic-plastic, following the measured response of type
304 stainless steel at room temperature, as reported by Sobel and Newman (1979). Since all the
analyses give results that are stiffer than the experimentally measured response, and the mesh
convergence tests (results are discussed below) demonstrate that the meshes are convergent with
respect to overall response of the system, it seems that this stress-strain model may overestimate the
material's actual strength.
Loading
The load on the pipe has two components: a "dead" load, consisting of internal pressure (with a closed
1-41
end condition), and a "live" in-plane bending moment applied to the end of the system. The pressure is
applied to the model in an initial step and then held constant in the second analysis step while the
bending moment is increased. The pressure values range from 0.0 to 3.45 MPa (500 lb/in 2), which is
the range of interest for design purposes. The equivalent end force associated with the closed-end
condition is applied as a follower force because it rotates with the motion of the end plane.
Kinematic boundary conditions

The fixed end of the system is assumed to be fully built-in. The loaded end is fixed into a very stiff
plate. For the ELBOW element models this condition is represented the NODEFORM boundary
condition applied at this node. In the shell element model this rigid plate is represented by a single
node, and the shell nodes at the end of the pipe are attached to it by using the kinematic coupling
constraint and specifying that all degrees of freedom at the shell nodes are constrained to the motion of
the single node.

The moment-rotation responses predicted by the various analysis models and measured in the
experiment, all taken at zero internal pressure, are compared in Figure 1.1.2-3. The figure shows that
the two shell models give very similar results, overestimating the experimentally measured collapse
moment by about 15%. The 6-mode ELBOW element models are somewhat stiffer than the shell
models, and those with 4 Fourier modes are much too stiff. This clearly shows that, for this very
flexible system, the ovalization of the elbow is too localized for even the 6-mode ELBOW
representation to provide accurate results.
Since we know that the shell models are convergent with respect to discretization, the most likely
explanation for the excessive stiffness in comparison to the experimentally measured response is that
the material model used in the analyses is too strong. Sobel and Newman (1979) point out that the
stress-strain curve measured and used in this analysis, shown in Figure 1.1.2-1, has a 0.2% offset yield
that is 20% higher than the Nuclear Systems Materials Handbook value for type 304 stainless steel at
room temperature, which suggests the possibility that the billets taken for the stress-strain curve
measurement may have been from stronger parts of the fabrication. If this is the case, it points out the
likelihood that the elbow tested is rather nonuniform in strength properties in spite of the care taken in
its manufacture. We are left with the conclusion that discrepancies of this magnitude cannot be
eliminated in practical cases, and the design use of such analysis results must allow for them.
Figure 1.1.2-4 compares the moment-rotation response for opening and closing moments, under 0 and
3.45 MPa (500 lb/in 2) internal pressure, and shows the strong influence of large-displacement effects.
If large-displacement effects were not important, the opening and closing moments would produce the
same response. However, even with a 1° relative rotation across the elbow assembly, the opening and
closing moments differ by about 12%; with a 2° relative rotation, the difference is about 17%. Such
magnitudes of relative rotation would not normally be considered large; in this case it is the coupling
into ovalization that makes geometric nonlinearity significant. As the rotation increases, the cases with
closing moment loading show collapse, while the opening moment curves do not. In both cases
internal pressure shows a strong effect on the results, which is to be expected in such a thin-walled
pipeline. The level of interaction between the straight pipe and the elbows is well illustrated by the
1-42
strain distribution on the outside wall, shown in Figure 1.1.2-5. The strain contours are slightly
discontinuous at the ends of the curved elbow section because the shell thickness changes at those
sections.
Figure 1.1.2-6 shows a summary of the results of this and the previous example. The plot shows the
collapse value of the closing moment under in-plane bending as a function of internal pressure. The
strong influence of pressure on collapse is apparent. In addition, the effect of analyzing the elbow by
neglecting interaction between the straight and curved segments is shown: the "uniform bending"
results are obtained by using elements of type ELBOW31B in the bend and beams (element type B31)
for the straight segments. The importance of the straight/elbow interaction is apparent. In this case the
simpler analysis neglecting the interaction is conservative (in that it gives consistently lower values for
the collapse moment), but this conservatism cannot be taken for granted. The analysis of Sobel and
Newman (1979) also neglects interaction and agrees quite well with the results obtained here.
For comparison the small-displacement limit analysis results of Goodall (1978), as well as his
large-displacement, elastic-plastic lower bound (Goodall, 1978a), are also shown in this figure. Again,
the importance of large-displacement effects is apparent from that comparison.
Detailed results obtained with the model that uses ELBOW31 elements are shown in the following
figures. Figure 1.1.2-7 shows the variation of the Mises stress along the length of the piping system.
The length is measured along the centerline of the pipe starting at the loaded end. The figure compares
the stress distribution at the intrados (integration point 1) on the inner and outer surfaces of the
elements (section points 1 and 7, respectively). Figure 1.1.2-8 shows the variation of the Mises stress
around the circumference of two elements (451 and 751) that are located in the bend section of the
model; the results are for the inner surface of the elements (section point 1). Figure 1.1.2-9 shows the
ovalization of elements 451 and 751. A nonovalized, circular cross-section is included in the figure for
comparison. From the figure it is seen that element 751, located at the center of the bend section,
experiences the most severe ovalization. These three figures were produced with the elbow element
postprocessing program FELBOWFOR (``Creation of a data file to facilitate the postprocessing of elbow
element results: FELBOW,'' Section 11.1.6).
Shell-to-solid submodeling
One particular case was analyzed using the shell-to-solid submodeling technique. The problem was
created for verification purposes to check the interpolation scheme in the case of double curved
surfaces. A solid submodel using C3D27R elements was created around the elbow part of the pipe,
spanning an angle of 40°. The finer submodel mesh has three elements through the thickness, 10
elements around half of the circumference of the cylinder, and 10 elements along the length of the
elbow. Both ends are driven from the global shell model made of S8R elements. The submodel results
agree closely with the shell model. The *SECTION FILE option is used to output the total force and
the total moment in a cross-section through the submodel.
Input files
In all of the following input files, with the exception of elbowcollapse_elbow31b_b31.inp and
elbowcollapse_s8r5_fine.inp, the step concerning the application of the pressure load is commented
1-43
out. To include the effects of the internal pressure in any given job, uncomment the step definition in
the appropriate input file.
elbowcollapse_elbow31b_b31.inp
ELBOW31B and B31 element model.
elbowcollapse_elbow31_6four.inp
ELBOW31 model with 6 Fourier modes.
elbowcollapse_elbow32_6four.inp
elbowcollapse_s8r.inp
S8R element model.
elbowcollapse_s8r5.inp
S8R5 element model.
elbowcollapse_s8r5_fine.inp
Finer S8R5 element model.
elbowcollapse_s9r5.inp
S9R5 element model.
elbowcollapse_stri65.inp
STRI65 element model.
elbowcollapse_submod.inp
Submodel using C3D27R elements.
References
· Dodge, W. G., and S. E. Moore, "Stress Indices and Flexibility Factors for Moment Loadings on
Elbows and Curved Pipes," Welding Research Council Bulletin, no. 179, 1972.
· Goodall, I. W., "Lower Bound Limit Analysis of Curved Tubes Loaded by Combined Internal
Pressure and In-Plane Bending Moment," Research Division Report RD/B/N4360, Central
Electricity Generating Board, England, 1978.
· Goodall, I. W., "Large Deformations in Plastically Deforming Curved Tubes Subjected to

In-Plane Bending," Research Division Report RD/B/N4312, Central Electricity Generating Board,
England, 1978a.
· Sobel, L. H. and S. Z. Newman, "Elastic-Plastic In-Plane Bending and Buckling of an Elbow:

Comparison of Experimental and Simplified Analysis Results," Westinghouse Advanced
Reactors Division, Report WARD-HT-94000-2, 1979.
1-44
Figures
Figure 1.1.2-1 MLTF elbow: geometry and measured material response.
Figure 1.1.2-2 Models for elbow/pipe interaction study.
1-45
Figure 1.1.2-3 Moment-rotation response: mesh convergence studies.
1-46
Figure 1.1.2-4 Moment-rotation response: pressure dependence.
Figure 1.1.2-5 Strain distribution on the outside surface: closing moment case.
1-47
Figure 1.1.2-6 In-plane bending of an elbow, elastic-plastic collapse moment results.
1-48
Figure 1.1.2-7 Mises stress distribution along the length of the piping system.
Figure 1.1.2-8 Mises stress distribution around the circumference of elements 451 and 751.
1-49
Figure 1.1.2-9 Ovalization of elements 451 and 751.
Sample listings
1-50
Listing 1.1.2-1
*HEADING
MLTF ELBOW: IN-PLANE BENDING. SYM.
HALF S8R5 MODEL,LAMDA=.167
*NSET,NSET=N100
100,
*NODE
100,0.,24.,-24.
101,0.,31.808,-24.
113,0.,16.192,-24.
900,0.,24.0,0.
901,0.,31.808,0.
913,0.,16.192,0.
2100,0.,0.,24.
2101,0.,0.,31.808
2113,0.,0.,16.192
4500,0.,-72.,24.
4501,0.,-72.,31.808
4513,0.,-72.,16.192
*NGEN,NSET=PYZ
100,900,100
101,901,100
113,913,100
2100,4500,100
2101,4501,100
2113,4513,100
*NGEN,LINE=C,NSET=PYZ
900,2100,100,0,0.,0.,0.,1.
901,2101,100,0,0.,0.,0.,1.
913,2113,100,0,0.,0.,0.,1.
*NGEN,LINE=C,NSET=LOADEND
101,113,1,0,0.,24.,-24.,0.,0.,-1.
*NGEN,LINE=C ,NSET=END2
901,913,1,0,0.,24., 0.0,0.,0.,-1.
*NGEN,LINE=C,NSET=END3
2101,2113,1,0,0.,0.,24., 0.,1.,0.
*NGEN,LINE=C ,NSET=FIXEDEND
4501,4513,1,0,0.,-72.,24., 0.,1.,0.
*NFILL
LOADEND,END2,8,100
END3,FIXEDEND,24,100
*NGEN,LINE=C
1-51
1001,1013,1,1000,,,,0.,.130526192,-.9914448614
1101,1113,1,1100,,,,0.,.258819045,-.965925826
1201,1213,1,1200,,,,0.,.382683432,-.923879532
1301,1313,1,1300,,,,0.,.5,-.866025404
1401,1413,1,1400,,,,0.,.60876143,-.79335334
1501,1513,1,1500,,,,0.,.707106781,-.707106781
1601,1613,1,1600,,,,0.,.79335334,-.60876143
1701,1713,1,1700,,,,0.,.866025404,-.5
1801,1813,1,1800,,,,0.,.923879532,-.382683432
1901,1913,1,1900,,,,0.,.965925826,-.258819045
2001,2013,1,2000,,,,0.,.9914448614,-.130526192
*ELEMENT,TYPE=S8R5
101,101,301,303,103,201,302,203,102
1101,2101,2301,2303,2103,2201,2302,2203,2102
*ELGEN,ELSET=LEG
101,6,2,1,4,200,100
1101,6,2,1,12,200,100
*ELEMENT,TYPE=S8R5
501,901,1101,1103,903,1001,1102,1003,902
*ELGEN,ELSET=LBOW
501,6,2,1,6,200,100
*ELSET,ELSET=ALL
LEG,LBOW
*MATERIAL ,NAME=PIPE
*ELASTIC
28.1E6,.2642
*PLASTIC
39440.,0.
50170.,.00473
54950.,.01264
58540.,.02836
61520.,.0491
76520.,.105
*SHELL SECTION,ELSET=LEG,MATERIAL=PIPE
.37,7
*SHELL SECTION,ELSET=LBOW,MATERIAL=PIPE
.41,7
*BOUNDARY
PYZ,1
PYZ,5,6
FIXEDEND,1,6
*NSET,NSET=COUPLED_END,GENERATE
101,113,1
1-52
*KINEMATIC COUPLING, REF NODE = 100

COUPLED_END,
*NSET,NSET=NOUT
N100,LOADEND,END2,END3,FIXEDEND
*RESTART,WRITE,FREQUENCY=5
** *STEP,INC=1,NLGEOM
** APPLY PRESSURE
** *STATIC
** *DLOAD,OP=NEW
** ALL,P,500.
** *CLOAD,OP=NEW,FOLLOWER
** 100,3,-4.7882E4
** *EL PRINT,FREQUENCY=0
** PE
** *NODE PRINT,NSET=N100
** U
** RF
** *NODE FILE,NSET=NOUT
** U,RF
** *END STEP
*STEP,INC=25,NLGEOM
HOLD PRESSURE CONSTANT, APPLY BENDING ACTION
*STATIC,RIKS
.05, 1.,.01,.5,2.0,100,4,-.2
*BOUNDARY
100,4,4,-.2
*EL PRINT,FREQUENCY=0
PE,
*NODE PRINT,NSET=N100
U,
RF,
*NODE FILE,NSET=NOUT
U,RF
*END STEP
1-53
Listing 1.1.2-2
*HEADING
MESH WITH 4 B31 + 6 ELBOW31B + 12 B31.
*NSET,NSET=NPRT
100,4500
*NODE,NSET=LOADEND
100,0.,24.,-24.
*NODE,NSET=END2
900,0.,24.0,0.
*NODE,NSET=END3
2100,0.,0.,24.
*NODE,NSET=FIXEDEND
4500,0.,-72.,24.
*NGEN,NSET=PYZ
100,900,100
2100,4500,100
900,2100,100,0,0.,0.,0.,1.
*ELEMENT,TYPE=B31
101,100,300
1101,2100,2300
*ELGEN,ELSET=LEG1
101,4,200,100
*ELGEN,ELSET=LEG2
1101,12,200,100
*ELEMENT,TYPE=ELBOW31B
501,900,1100
*ELGEN,ELSET=LBOW
501,6,200,100
*ELSET,ELSET=ALL
LEG1,LEG2,LBOW
*ELASTIC
28.1E6,.2642
*PLASTIC
39440.,0.
50170.,.00473
54950.,.01264
58540.,.02836
61520.,.0491
76520.,.105
*BEAM SECTION,SECTION=PIPE,ELSET=LEG1,
1-54
MATERIAL=PIPE
8.01,.37
1.,0.,0.
18,
*BEAM SECTION,SECTION=PIPE,ELSET=LEG2,
MATERIAL=PIPE
8.01,.37
1.,0.,0.
18,
*BEAM SECTION,SECTION=ELBOW,ELSET=LBOW,
MATERIAL=PIPE
8.01,.41,24.
0.,24.,24.
7,18,6
*BOUNDARY
PYZ,1
PYZ,5,6
FIXEDEND,1,6
*NSET,NSET=NOUT,GENERATE
100,4500,200
*NSET,NSET=NOUTRF
100,4500
*STEP,INC=11,NLGEOM
APPLY BENDING ACTION
*STATIC,RIKS
.05, 1.,.005,.5,2.0,100,4,-.2
*BOUNDARY
100,4,4,-.2
*MONITOR,NODE=100,DOF=4
PE,
*NODE PRINT,NSET=NPRT
U,
RF,
U,
*NODE FILE,NSET=NOUTRF
RF,
*OUTPUT,FIELD
*NODE OUTPUT,NSET=NOUT
U,
*OUTPUT,HISTORY
1-55
U,
*OUTPUT,FIELD
*NODE OUTPUT,NSET=NOUTRF
RF,
*OUTPUT,HISTORY
*NODE OUTPUT,NSET=NOUTRF
RF,
*END STEP
1-56
Listing 1.1.2-3
*HEADING
MLTF ELBOW TEST, FULL MODEL USING ELBOW31,
LAMDA=.167 FINE MESH WITH 8+12+24 ELEMENTS.
SECTION INTEGRATION: 7,18,6
*NSET,NSET=NPRT
100,4500
*NODE,NSET=LOADEND
100,0.,24.,-24.
*NODE,NSET=END2
900,0.,24.0,0.
*NODE,NSET=END3
2100,0.,0.,24.
*NODE,NSET=FIXEDEND
4500,0.,-72.,24.
*NGEN,NSET=PYZ
100,900,100
2100,4500,100
900,2100,100,0,0.,0.,0.,1.
*ELEMENT,TYPE=ELBOW31
101,100,200
1101,2100,2200
*ELGEN,ELSET=LEG1
101,8,100,50
*ELGEN,ELSET=LEG2
1101,24,100,50
*ELEMENT,TYPE=ELBOW31
501,900,1000
*ELGEN,ELSET=LBOW
501,12,100,50
*ELSET,ELSET=ALL
LEG1,LEG2,LBOW
*ELASTIC
28.1E6,.2642
*PLASTIC
39440.,0.
50170.,.00473
54950.,.01264
58540.,.02836
61520.,.0491
1-57
76520.,.105
*BEAM SECTION,SECTION=ELBOW,ELSET=LEG1,
MATERIAL=PIPE
8.01,.37
0.,100.,-12.
7,18,6
*BEAM SECTION,SECTION=ELBOW,ELSET=LEG2,
MATERIAL=PIPE
8.01,.37
0.,-36.,100.
7,18,6
*BEAM SECTION,SECTION=ELBOW,ELSET=LBOW,
MATERIAL=PIPE
8.01,.41,24.
0.,24.,24.
7,18,6
*BOUNDARY
PYZ,1
PYZ,5,6
FIXEDEND,1,6
FIXEDEND,NODEFORM
LOADEND,NODEFORM
*NSET,NSET=NOUT,GENERATE
100,4500,100
** *STEP,INC=1,NLGEOM
** APPLY PRESSURE
** *STATIC
** *DLOAD,OP=NEW
** ALL,PI,500.,15.65,OPEN
** *CLOAD,OP=NEW,FOLLOWER
** 100,3,-9.5764E4
** *CONTROLS,PARAMETERS=FIELD,FIELD=DISPLACEMENT
** 1.,1.,,1000.
** *CONTROLS,PARAMETERS=FIELD,FIELD=ROTATION
** 1.,1.,,6000.
** *EL PRINT,FREQUENCY=0
** PE
** *NODE PRINT,NSET=N100
** U
** RF
** *NODE FILE,NSET=NOUT
** U,RF
1-58
** *END STEP
*STEP,INC=15,NLGEOM
HOLD PRESSURE CONSTANT, APPLY BENDING ACTION
*STATIC,RIKS
.05, 1.,.001,.5,2.0,100,4,-.2
*BOUNDARY
100,4,4,-.2
*CONTROLS,PARAMETERS=FIELD,FIELD=DISPLACEMENT
1.,1.,,10000.
*CONTROLS,PARAMETERS=FIELD,FIELD=ROTATION
1.,1.,,61000.
PE,
*NODE PRINT,NSET=NPRT
U,
RF,
U,RF
*OUTPUT,FIELD
U,RF
*OUTPUT,HISTORY
U,RF
*END STEP
1.1.3 Parametric study of a linear elastic pipeline under in-plane

bending
Elbows are used in piping systems because they ovalize more readily than straight pipes and, thus,
provide flexibility in response to thermal expansion and other loadings that impose significant
displacements on the system. Ovalization is the bending of the pipe wall into an oval--i.e.,
noncircular--configuration. The elbow is, thus, behaving as a shell rather than as a beam. This example
demonstrates the ability of elbow elements (``Pipes and pipebends with deforming cross-sections:
elbow elements,'' Section 15.5.1 of the ABAQUS/Standard User's Manual) to model the nonlinear
response of initially circular pipes and pipebends accurately when the distortion of the cross-section by
ovalization is significant. It also provides some guidelines on the importance of including a sufficient
number of Fourier modes in the elbow elements to capture the ovalization accurately. In addition, this
example illustrates the shortcomings of using "flexibility knockdown factors" with simple beam
elements in an attempt to capture the effects of ovalization in an ad hoc manner for large-displacement
analyses.
1-59
Geometry and model

The pipeline configuration used in the study is shown in Figure 1.1.3-1. It is a simple model with two
straight pipe sections connected by a 90° elbow. The straight pipes are 25.4 cm (10.0 inches) in length,
the radius of the curved section is 10.16 cm (4.0 inches), and the outer radius of the pipe section is
1.27 cm (0.5 inches). The wall thickness of the pipe is varied from 0.03175 cm to .2032 cm (0.0125
inches to 0.08 inches) in a parametric study, as discussed below. The pipe material is assumed to be
isotropic linear elastic with a Young's modulus of 194 GPa ( 28:1 £ 106 psi) and a Poisson's ratio of
0.0. The straight portions of the pipeline are assumed to be long enough so that warping at the ends of
the structure is negligible.
Two loading conditions are analyzed. The first case is shown in Figure 1.1.3-1 with unit inward
displacements imposed on both ends of the structure. This loading condition has the effect of closing
the pipeline in on itself. In the second case the sense of the applied unit displacements is
outward--opening the pipeline. Both cases are considered to be large-displacement/small-strain
analyses.
A parametric study comparing the results obtained with different element types (shells, elbows, and
pipes) over a range of flexibility factors, k, is performed. As defined in Dodge and Moore (1972), the
flexibility factor for an elbow is the ratio of the bending flexibility of the elbow segment to that of a
straight pipe of the same dimensions, assuming small displacements and an elastic response. When the
internal (gauge) pressure is zero, as is assumed in this study, k can be approximated as
1:66
k= ;
¸
where
Rt
¸= p ;
r2 1 ¡ º 2
R is the bend radius of the curved section, r is the mean radius of the pipe, t is the wall thickness of
the pipe, and º is Poisson's ratio. Changes in the flexibility factor are introduced by varying the wall
thickness of the pipe.
The pipeline is modeled with three different element types: S4 shell elements, ELBOW31 elbow
elements, and PIPE31 pipe elements. The S4 shell element model consists of a relatively fine mesh of
40 elements about the circumference and 75 elements along the length. This mesh is deemed fine
enough to capture the true response of the pipeline accurately, although no mesh convergence studies
are performed. The pipe and elbow element meshes consist of 75 elements along the length. The
results of the shell element model are taken as the reference solution. The reaction force at the tip of
the pipeline is used to evaluate the effectiveness of the pipe and elbow elements. In addition, the
ovalization values of the pipeline cross-section predicted by the elbow element models are compared.
The elbow elements are tested with 0, 3, and 6 Fourier modes, respectively. In general, elbow element
accuracy improves as more modes are used, although the computational cost increases accordingly. In
1-60
addition to standard pipe elements, tests are performed on pipe elements with a special flexibility
knockdown factor. Flexibility knockdown factors (Dodge and Moore, 1972) are corrections to the
bending stiffness based upon linear semianalytical results. They are applied to simple beam elements
in an attempt to capture the global effects of ovalization. The knockdown factor is implemented in the
PIPE31 elements by scaling the true thickness by the flexibility factor; this is equivalent to scaling the
moment of inertia of the pipe element by 1=k.

In the following the results obtained with the shell element model are taken as the reference solution.
The tip reaction forces due to the inward prescribed displacements for the various analysis models are
shown in Figure 1.1.3-2. The results are normalized with respect to those obtained with the shell
model. The results obtained with ELBOW31 element with 6 Fourier modes show excellent agreement
with the reference solution over the entire range of flexibility factors considered in this study. The
remaining four models generally exhibit excessively stiff response for all values of k. The PIPE31
element model, which uses the flexibility knockdown factor, showed a relatively constant error of
about 20% over the entire range of flexibility factors. The 0-mode ELBOW31 element model and the
PIPE31 element model without the knockdown factor produce very similar results for all values of k.
The normalized tip reaction forces due to the outward unit displacement for the various analysis
models are shown in Figure 1.1.3-3. Again, the results obtained with the 6-mode ELBOW31 element
compare well with the reference shell solution. The 0-mode and 3-mode ELBOW31 and the PIPE31
(without the flexibility knockdown factor) element models exhibit overly stiff response. The PIPE31
element model with the knockdown factor has a transition region near k=1.5, where the response
changes from being too stiff to being too soft.
Figure 1.1.3-4 and Figure 1.1.3-5 illustrate the effect of the number of included Fourier modes (0, 3,
and 6) on the ability of the elbow elements to model the ovalization in the pipebend accurately in both
load cases considered in this study. By definition, the 0-mode model cannot ovalize, which accounts
for its stiff response. The 3-mode and the 6-mode models show significant ovalization in both loading
cases. Figure 1.1.3-6 compares the ovalization of the 6-mode model in the opened and closed
deformation states. It clearly illustrates that when the ends of the pipe are displaced inward (closing
mode), the height of the pipe's cross-section gets smaller, thereby reducing the overall stiffness of the
pipe; the reverse is true when the pipe ends are displaced outward: the height of the pipe's
cross-section gets larger, thereby increasing the pipe stiffness. These three figures were produced with
the aid of the elbow element postprocessing program FELBOWFOR (``Creation of a data file to
facilitate the postprocessing of elbow element results: FELBOW,'' Section 11.1.6).
Parametric study
The performance of the pipe and elbow elements investigated in this example is analyzed conveniently
in a parametric study using the Python scripting capabilities of ABAQUS (``Scripting parametric
studies,'' Section 25.1.1 of the ABAQUS/Standard User's Manual). We perform a parametric study in
which eight analyses are executed automatically for each of the three different element types (S4,
ELBOW31, and PIPE31) discussed above; these parametric studies correspond to wall thickness
1-61
values ranging from 0.03175 cm to .2032 cm (0.0125 inches to 0.08 inches).

The Python script file elbowtest.psf is used to perform the parametric study. The function
customTable (shown below) is an example of advanced Python scripting (Lutz and Ascher, 1999),
which is used in elbowtest.psf. customTable is designed to take an XYPLOT file from the
parametric study and convert it into a new file of reaction forces versus flexibility factors ( k). Such
advanced scripting is not routinely needed. It is needed in this case to overcome the limitation that a
dependent variable such as k cannot be included as a column of data in an XYPLOT file.
###############################################################
#
def customTable(file1, file2):
for line in file1.readlines():
print line
nl = string.split(line,',')
disp = float(nl[0])
bend_radius = float(nl[1])
wall_thick = float(nl[2])
outer_pipe_radius = float(nl[3])
poisson = float(nl[4])
rf = float(nl[6])
mean_rad = outer_pipe_radius - wall_thick/2.0

k = bend_radius*wall_thick/mean_rad**2
k = k/sqrt(1.e0 - poisson**2)
k = 1.66e0/k
outputstring = str(k) + ', ' + str(rf) + '\n'

file2.write(outputstring)
#
#############################################################
Input files
elbowtest_shell.inp
S4 model.
elbowtest_elbow0.inp
1-62

elbowtest_pipek.inp
PIPE31 model with the flexibility knockdown factor.
elbowtest_pipe.inp
PIPE31 model without the flexibility knockdown factor.
elbowtest.psf
Python script file for the parametric study.
References
· Dodge, W. G., and S. E. Moore, ``Stress Indices and Flexibility Factors for Moment Loadings on
Elbows and Curved Pipes,'' Welding Research Council Bulletin, no. 179, 1972.
· Lutz, M., and D. Ascher, Learning Python, O'Reilly, 1999.
Figures
Figure 1.1.3-1 Pipeline geometry with inward prescribed tip displacements.
Figure 1.1.3-2 Normalized tip reaction force: closing displacement case.
1-63
Figure 1.1.3-3 Normalized tip reaction force: opening displacement case.
Figure 1.1.3-4 Ovalization of the ELBOW31 cross-sections for 0, 3, and 6 Fourier modes: closing
displacement case.
1-64
Figure 1.1.3-5 Ovalization of the ELBOW31 cross-sections for 0, 3, and 6 Fourier modes: opening
displacement case.
Figure 1.1.3-6 Ovalization of the ELBOW31 cross-sections for 6 Fourier modes: opening and closing
displacement cases.
1-65
Sample listings
1-66
Listing 1.1.3-1
*HEADING
ELBOW31 elements with 6 ovalization modes
*restart,write
************************
*** PARAMETER DEFINITION
************************
*PARAMETER
# length a of the straight sections
a = 10.0
# radius of the curved section (bend)
bend_radius = 4.0
# outer radius of the pipe
outer_pipe_radius = 0.5
# wall thickness of the pipe
wall_thick = 0.08
# number of elements along the straight sections
# of the pipe
num_elem_s = 25
# number of elements around the bend
num_elem_c = 25
# displacement at the end of the pipe
disp = 1.0
# Young's modulus
young = 28.1E6
# Poisson's ratio
poisson = 0.0
# number of integration points through the
# thickness
nip_thru_thick = 5
# number of integration points around the pipe
nip_around_pipe = 20
# number of ovalization modes
numoval = 6
#
# total number of elements along the length of
# the pipe
#
num_elem = num_elem_c + 2*num_elem_s
#
###############################################
#
1-67
# geometrical properties of the problem:

# points along pipe c/l
#
x1 = a + bend_radius
y1 = 0.
x2 = a + bend_radius
y2 = a
x3 = a
y3 = a + bend_radius
x4 = 0.
y4 = a + bend_radius
#
x1_plus_100 = x1 + 100.0
y4_plus_100 = y4 + 100.0
#
# node control data
#
ninc = 1
n1 = 1
n1_plus_inc = n1 + ninc
n2 = n1 + num_elem_s*ninc
n3 = n2 + num_elem_c*ninc
n4 = n3 + num_elem_s*ninc
#
ndummy = n4 + 1000
#
# element control data
#
einc = 1
e1 = 1
e2 = e1 + (num_elem_s-1)*einc
e2_plus_inc = e2 + einc
e3 = e2_plus_inc + (num_elem_c-1)*einc
e3_plus_inc = e3 + einc
e4 = e3_plus_inc + (num_elem_s-1)*einc
#
# END OF PARAMETER SECTION
#
#*******************
#** Node definitions
#*******************
*NODE
<ndummy>,<a>,<a>,0.
1-68
*NODE
<n1>,<x1>,<y1>
<n2>,<x2>,<y2>
<n3>,<x3>,<y3>
<n4>,<x4>,<y4>
*NGEN,NSET=straight_1
<n1>,<n2>,<ninc>
****
*NGEN,line=c,nset=curved
<n2>,<n3>,<ninc>,<ndummy>,<a>,<a>,0.
****
*NGEN,NSET=straight_2
<n3>,<n4>,<ninc>
*NSET,NSET=all_nodes
straight_1,curved,straight_2
**
**********************
** Material definition
**********************
*MATERIAL,NAME=pipe
*ELASTIC
<young>,<poisson>
**************************
** Element data definition
**************************
*ELEMENT,TYPE=elbow31
<e1>,<n1>,<n1_plus_inc>
*ELGEN,ELSET=all_elem
<e1>,<num_elem>,<ninc>,<einc>
*ELSET,ELSET=straight1,GENERATE
<e1>,<e2>,<einc>
*ELSET,ELSET=curved,GENERATE
<e2_plus_inc>,<e3>,<einc>
*ELSET,ELSET=straight2,GENERATE
<e3_plus_inc>,<e4>,<einc>
**************************
** Element definition
**************************
*BEAM SECTION,SECTION=elbow,ELSET=straight1,
MATERIAL=PIPE
<outer_pipe_radius>, <wall_thick>, 0.
<x1_plus_100>,<y1>
<nip_thru_thick>,<nip_around_pipe>,<numoval>
1-69
*BEAM SECTION,SECTION=elbow,ELSET=straight2,
MATERIAL=PIPE
<outer_pipe_radius>, <wall_thick>, 0.
<y4>,<y4_plus_100>
*BEAM SECTION,SECTION=elbow,ELSET=curved,
MATERIAL=PIPE
<outer_pipe_radius>, <wall_thick>, <bend_radius>
<x1>,<y4>
**********************
** Boundary conditions
**********************
*BOUNDARY
<n1>,ysymm
<n1>,nowarp
<n1>,nooval
<n1>,3
<n4>,xsymm
<n4>,nowarp
<n4>,nooval
<n4>,3
**********************
*STEP,nlgeom
Apply displacements at the ends of the structure
*STATIC
1.,1.
*BOUNDARY
<n1>,1,1,<disp>
<n4>,2,2,<disp>
*NODE PRINT,NSET=all_nodes, FREQ=9999
u,
rf,
*NODE FILE,NSET=all_nodes, FREQ=9999
u,
rf,
*el file,elset=curved,f=999
coord,
*END STEP
1-70
Listing 1.1.3-2
1.1.4 Indentation of an elastomeric foam specimen with a

hemispherical punch
Products: ABAQUS/Standard ABAQUS/Explicit ABAQUS/Design
In this example we consider a cylindrical specimen of an elastomeric foam, indented by a rough, rigid,
hemispherical punch. Examples of elastomeric foam materials are cellular polymers such as cushions,
padding, and packaging materials. This problem illustrates a typical application of elastomeric foam
materials when used in energy absorption devices. The same geometry as the crushable foam model of
`Ìndentation of a crushable foam specimen with a hemispherical punch, '' Section 1.1.7, is used but
with a slightly different mesh. Design sensitivity analysis is carried out for a shape design parameter
and a material design parameter to illustrate the usage of design sensitivity analysis for a problem
involving contact.
Geometry and model

The axisymmetric model (135 linear 4-node elements) analyzed is shown in Figure 1.1.4-1. The mesh
refinement is biased toward the center of the foam specimen where the largest deformation is expected.
The foam specimen has a radius of 600 mm and a thickness of 300 mm. The punch has a radius of 200
mm. The bottom nodes of the mesh are fixed, while the outer boundary is free to move.
In ABAQUS/Standard a contact pair is defined between the punch, which is modeled by a rough
spherical rigid surface, and a slave surface composed of the faces of the axisymmetric elements in the
contact region. A point mass of 200 kg representing the weight of the punch is attached to the rigid
body reference node. In ABAQUS/Explicit the punch is modeled as either an analytical rigid surface or
a rigid surface defined with RAX2 elements. In ABAQUS/Explicit the friction coefficient between the
punch and the foam is 0.8.
Material
The elastomeric foam material is defined through the *HYPERFOAM option using experimental test
data. The uniaxial compression and simple shear data whose stress-strain curves are shown in Figure
1.1.4-2 are defined with the *UNIAXIAL TEST DATA and *SIMPLE SHEAR TEST DATA options.
Other available test data options are *BIAXIAL TEST DATA, *PLANAR TEST DATA and
*VOLUMETRIC TEST DATA. The test data are defined in terms of nominal stress and nominal strain
values. ABAQUS performs a nonlinear least-squares fit of the test data to determine the hyperfoam
coefficients ¹i ; ®i ; and ¯i .
Details of the formulation and usage of the hyperfoam model are given in `Èlastomeric foam
behavior,'' Section 10.5.2 of the ABAQUS/Standard User's Manual and Section 9.3.2 of the
ABAQUS/Explicit User's Manual; ``Hyperelastic material behavior,'' Section 4.6.1 of the ABAQUS
Theory Manual; and ``Fitting of hyperelastic and hyperfoam constants,'' Section 4.6.2 of the ABAQUS
Theory Manual. ``Fitting of elastomeric foam test data,'' Section 3.1.5 of the ABAQUS Benchmarks
1-71
Manual, illustrates the fitting of elastomeric foam test data to derive the hyperfoam coefficients.
For the material used in this example, ¯i is zero, since the effective Poisson's ratio, º, is zero, as
specified through the POISSON parameter. The order of the series expansion is chosen to be N = 2
since this fits the test data with sufficient accuracy. N = 2 also provides a more stable model than the
N = 3 case.
The viscoelastic properties in ABAQUS/Standard are specified in terms of a relaxation curve (shown
in Figure 1.1.4-3) of the normalized modulus M (t)=M0 , where M (t) is the shear or bulk modulus as a
function of time and M0 is the instantaneous modulus, as determined from the hyperfoam model. This
requires the use of the TIME=RELAXATION TEST DATA parameter of the *VISCOELASTIC
option. The relaxation data are specified through the *SHEAR TEST DATA option but actually apply
to both shear and bulk moduli when used in conjunction with the hyperfoam model.
ABAQUS/Standard performs a nonlinear least-squares fit of the relaxation data to a Prony series to
determine the coefficients, g Pi , and the relaxation periods, ¿i . A maximum order of NMAX=2 for
fitting the Prony series is used. If creep data are available, the TIME=CREEP TEST DATA parameter
is set to specify normalized creep compliance data. The ABAQUS/Explicit analysis is purely elastic.
A rectangular material orientation is defined for the foam specimen, so stress and strain are reported in
material axes that rotate with the element deformation. This is especially useful when looking at the
stress and strain values in the region of the foam in contact with the punch in the direction normal to
the punch (direction "22").
The rough surface of the punch is modeled by specifying a friction coefficient of 0.80 for the contact
surface interaction through the *FRICTION option under the *SURFACE INTERACTION definition.
Because of the unsymmetric nature of the friction material model, set UNSYMM=YES on the *STEP
option in ABAQUS/Standard.
Procedure and loading definitions in ABAQUS/Standard

The ABAQUS/Standard example is composed of two analyses. In the first case the punch is displaced
statically to indent the foam, and the reaction force-displacement relation is measured for both the
purely elastic and viscoelastic cases. In the second case the punch statically indents the foam through
gravity loading and is then subjected to a dynamic impulsive loading. The dynamic response of the
punch is sought as it interacts with the viscoelastic foam.
Case 1
During the first step the punch is displaced downward by a prescribed *BOUNDARY condition,
indenting the foam specimen by a distance of 250 mm. The NLGEOM parameter is specified on the
*STEP option, since the response involves large deformation. In the second step the punch is displaced
back to its original position. Two analyses are performed--one using the *STATIC procedure for both
steps and the other using the *VISCO procedure for both steps.
During a *STATIC step the material behaves purely elastically, using the properties specified with the
hyperfoam model. The *VISCO, *DYNAMIC, or *COUPLED TEMPERATURE-DISPLACEMENT
procedure must be used to activate the viscoelastic behavior. In this case the punch is pushed down in
a period of one second and then moved back up again in one second. The accuracy of the creep
1-72
integration in the *VISCO procedure is controlled by the CETOL parameter and is typically calculated
by dividing an acceptable stress error tolerance by a typical elastic modulus. In this problem, we
estimate a stress error tolerance of about 0.005 MPa and use the initial elastic modulus, E 0 = 2
P
¹i = 0.34, to determine a CETOL of 0.01.
Case 2
This analysis is composed of three steps. The first step is a *VISCO step, where gravity loading is
applied to the point mass of the punch. The gravity loading is ramped up in two seconds, and the step
is run for a total of five seconds to allow the foam to relax fully. In the second step, which is a
*DYNAMIC step, an impulsive load in the form of a half sine wave amplitude with a peak magnitude
of 5000 N is applied to the punch over a period of one second. In the third step (also a *DYNAMIC
step) the punch is allowed to move freely until the vibration is damped out by the viscoelastic foam.
For a dynamic analysis with automatic time stepping, the value of the HAFTOL (half-step residual
tolerance) parameter of the *DYNAMIC procedure controls the accuracy of the time integration. For
systems that have significant energy dissipation, such as this heavily damped model, a relatively high
value of HAFTOL can be chosen. We choose HAFTOL to be 100 times a typical average force that we
estimate (and later confirm from the analysis results) to be on the order of 50 N. Thus, HAFTOL is
5000 N. For the second *DYNAMIC step we set INITIAL=NO to bypass calculation of initial
accelerations at the beginning of the step, since there is no sudden change in load to create a
discontinuity in the accelerations.
Procedure and loading definitions in ABAQUS/Explicit

In ABAQUS/Explicit the punch is displaced downward by a prescribed boundary condition, indenting
the foam by a depth of 250 mm. The punch is then lifted back to its original position. The whole
analysis runs for 0.12 seconds. This analysis is comparable to the Case 1 ABAQUS/Standard analysis
performed with the *STATIC procedure.
Design sensitivity analysis

For the design sensitivity analysis (DSA) carried out with static steps in ABAQUS/Standard, the
hyperfoam material properties are given using direct input of coefficients based on the test data given
above. For N = 2, the coeficients are ¹1 = 0.16245, ¹2 = 3.59734E-05, ®1 = 8.89239, ®2 =
-4.52156, and ¯i = 0.0. Due to the current limitations in the DSA capability, the viscoelastic
properties are not included and the interaction between the rigid punch and the foam is assumed to be
frictionless with small sliding (the small sliding assumption is necessary since finite sliding is not
currently supported for DSA).
The material parameter ¹1 is chosen as one of the design parameters. The other (shape) design
parameter used for design sensitivity analysis, L, represents the thickness of the foam at the free end
(see Figure 1.1.4-1). The z-coordinates of the nodes on the top surface are assumed to depend on L via
the equation z = 300 + (L ¡ 300)r=600 . The r-coordinates are considered to be independent of L. To
define this dependency in ABAQUS, the gradients of the coordinates with respect to L :
1-73
dr
=0
dL
dz r
=
dL 600
are given under the *PARAMETER SHAPE VARIATION option.

This problem tests the hyperfoam material model in ABAQUS but does not provide independent
verification of the model. The results for all analyses are discussed in the following paragraphs.
ABAQUS/Standard: Case 1
Deformation and contour plots for oriented S22 stress and E22 strain are shown for the viscoelastic
foam in Figure 1.1.4-4. Even though the foam has been subjected to large strains, only moderate
distortions occur because of the zero Poisson's ratio. The maximum (logarithmic) total strain is of the
order of -1.8, which is equivalent to a stretch of ¸ = e-1.8= 0.17, or a nominal compressive strain of
83% indicating severe compression of the foam.
In the viscoelastic case the stresses relax during loading and, consequently, lead to a softer response
than in the purely elastic case, as shown in Figure 1.1.4-5. The force-displacement responses are
shown in Figure 1.1.4-6. The purely elastic material is reversible, while the viscoelastic material shows
hysteresis.
ABAQUS/Standard: Case 2
Various displaced configurations during the Case 2 analysis are shown in Figure 1.1.4-7.
Displacement, velocity, and acceleration histories for the punch are shown in Figure 1.1.4-8. The
displacement is shown to reach a steady value at the stress relaxation stage, followed by a severe drop
due to the impulsive dynamic load. This is followed by a rebound and then finally by a rapid decay of
the subsequent oscillations due to the strong damping provided by the viscoelasticity of the foam.
ABAQUS/Explicit
Figure 1.1.4-9 shows a plot of the initial configurations. Figure 1.1.4-10 shows a contour plot of stress
in the y-direction. Figure 1.1.4-11 shows a contour plot of logarithmic strain in the y-direction at 0.06
seconds when the maximum indentation is reached. The elements underneath the punch are seen to be
subject to large strains. The history of the punch reaction force (reference node 1000) is shown in
Figure 1.1.4-12. A plot of the punch displacement versus the punch reaction force is shown in Figure
1.1.4-13. Because of the purely elastic material behavior, there is no hysteresis and the punch reaction
force in the unloading stage follows the same curve as during loading.
ABAQUS/Design
Figure 1.1.4-14 shows the variation of the normalized sensitivity of CPRESS with respect to the design
1-74
variables L and ¹1 on the contact surface. The sensitivities have been normalized by multiplying with
the value of the design parameter and dividing by the maximum value of CPRESS. Figure 1.1.4-15 and
Figure 1.1.4-16 show the contours of sensitivity of the displacement in the z-direction to the design
paramaters L and ¹1 , respectively. Figure 1.1.4-17 and Figure 1.1.4-18 show the contours of
sensitivity of S22 to the design parameters L and ¹1 , respectively. To provide an independent
assessment of the results provided by ABAQUS, sensitivities were computed using the overall finite
difference (OFD) technique. The central difference method with a perturbation size of 0.1% of the
value of the design parameter was used to obtain the OFD results. Table 1.1.4-1 shows that the
sensitivities computed using ABAQUS compare well with the overall finite difference results.
Input files
indentfoamhemipunch_case1.inp
Case 1 of the ABAQUS/Standard example using test data for both elastic and viscoelastic
properties of the foam, which is statically deformed in two *VISCO steps.
indentfoamhemipunch_case2.inp
Case 2 in which the ABAQUS/Standard analysis is performed in three steps subjecting the punch
to both static and dynamic loading.
hyperfoam_anl.inp
ABAQUS/Explicit analysis using an analytical rigid surface.
hyperfoam.inp
ABAQUS/Explicit model using a faceted rigid surface.
indentfoamhemipunch_dsa.inp
Design sensitivity analysis.
Table
Table 1.1.4-1 Comparison of normalized sensitivities computed using ABAQUS and overall finite
difference.
Normalized³ sensitivity
´ ABAQUS OFD
¹1 dS22 0.5024 0.5018
S22 max d¹1
L
¡ dS22
¢ max
S22 max
³ dL´ max -0.1134 -0.1107
¹1 du2 -0.0075 -0.0075
umax
2 ¡ d¹
du2
1
¢ max
L
max 0.2918 0.2922
u2 ³ dL max ´
¹1
CPRESS max
dCPRESS
d¹1
0.5578 0.5582
¡ dCPRESS ¢ max
L
CPRESS max
dL max
0.8012 0.8038
Figures
1-75
Figure 1.1.4-1 Model for foam indentation by a spherical punch.
Figure 1.1.4-2 Elastomeric foam stress-strain curves.
Figure 1.1.4-3 Elastic modulus relaxation curve.
1-76
Figure 1.1.4-4 Deformation and contour plots of viscoelastic foam, ABAQUS/Standard.
1-77
Figure 1.1.4-5 Punch reaction force history: static and viscoelastic cases, ABAQUS/Standard.
1-78
Figure 1.1.4-6 Punch reaction force versus displacement response: static and viscoelastic cases,
(loading-unloading curves); ABAQUS/Standard.
Figure 1.1.4-7 Deformation plots for the visco and dynamic steps, ABAQUS/Standard.
1-79
Figure 1.1.4-8 Displacement, velocity, and acceleration histories of the punch; ABAQUS/Standard.
1-80
Figure 1.1.4-9 Initial (undeformed) configuration, ABAQUS/Explicit.
1-81
Figure 1.1.4-10 Stress S22 contour plot at 0.06 s, ABAQUS/Explicit.
Figure 1.1.4-11 Logarithmic strain in the z-direction at 0.06 s, ABAQUS/Explicit.
1-82
Figure 1.1.4-12 Punch reaction force, ABAQUS/Explicit.
Figure 1.1.4-13 Punch reaction force versus indentation, ABAQUS/Explicit.
Figure 1.1.4-14 Normalized sensitivities of contact pressure at the end of the analysis.
1-83
Figure 1.1.4-15 Sensitivities at the end of the analysis for displacement in the z-direction with respect
to L.
Figure 1.1.4-16 Sensitivities at the end of the analysis for displacement in the z-direction with respect
to ¹1 .
1-84
Figure 1.1.4-17 Sensitivities at the end of the analysis for stress S22 with respect to L.
Figure 1.1.4-18 Sensitivities at the end of the analysis for stress S22 with respect to ¹1 .
1-85
Sample listings
1-86
Listing 1.1.4-1
*HEADING
STATIC INDENTATION OF A VISCOELASTIC,
ELASTOMERIC FOAM WITH A HEMISPHERICAL PUNCH
Measure the punch reaction force during the
following 2 steps:
Step 1: Displace punch 250 mm downwards.
Step 2: Return punch to original position.
Units: N, mm, sec
*NODE,NSET=ALLN
1,0.,300.
19,0.,0.
481,300.,300.
499,300.,0.
601,600.,300.
619,600.,0.
*NODE,NSET=N9999
9999,0.,500.
*NSET,NSET=N1
1,
*NSET,NSET=N19
19,
*NSET,NSET=N481
481,
*NSET,NSET=N499
499,
*NFILL,NSET=TOP,BIAS=.85
N1,N481,12,40
*NGEN,NSET=TOP
481,601,40
*NFILL,NSET=BOT,BIAS=.85
N19,N499,12,40
*NGEN,NSET=BOT
499,619,40
*NFILL,NSET=ALLN,BIAS=.9
TOP,BOT,9,2
*NSET,NSET=CENTER,GENERATE
3,19
*ELEMENT,TYPE=CAX4,ELSET=FOAM
1,3,43,41,1
*ELGEN,ELSET=FOAM
1-87
1,15,40,10,9,2,1
*ELSET,ELSET=CENT,GENERATE
1,141,10
61,69
*ELSET,ELSET=ETOP,GENERATE
1,101,10
*RIGID BODY,ANALYTICAL SURFACE=BSURF,REF NODE=9999
*SURFACE,TYPE=SEGMENTS,NAME=BSURF
START,141.42,641.42
CIRCL,-1.,300.,0.,500.
*SURFACE,NAME=ASURF
ETOP,S3
*CONTACT PAIR,INTERACTION=ROUGH
ASURF,BSURF
*FRICTION
0.8,
*SOLID SECTION,ELSET=FOAM,MATERIAL=FOAM,
ORIENTATION=RECT
*ORIENTATION,NAME=RECT,SYSTEM=RECTANGULAR
1.,0.,0., 0.,1.,0.
3,0.
*MATERIAL,NAME=FOAM
*HYPERFOAM,N=2,TEST DATA INPUT,POISSON=0.0,
MODULI=INSTANTANEOUS
*UNIAXIAL TEST DATA
-.0217, -.05
-.0317, -.10
-.0367, -.15
-.0402, -.20
-.0433, -.25
-.0467, -.30
-.0504, -.35
-.0542, -.40
-.0604, -.45
-.0668, -.50
-.0759, -.55
-.0909, -.60
-.1083, -.65
-.1410, -.70
-.1933, -.75
-.2896, -.80
*SIMPLE SHEAR TEST DATA
1-88
.0107, .08, .0030

.0373, .16, .0166
.0533, .24, .0366
.0853, .32, .0573
.1280, .40, .0817
.1653, .48, .1098
.2080, .56, .1394
.2560, .64, .1666
.2987, .72, .1904
*VISCOELASTIC,TIME=RELAXATION TEST DATA
*SHEAR TEST DATA,SHRINF=0.5000
1.0000, 0.0001
0.9695, 0.001
0.9417, 0.002
0.8722, 0.005
0.7913, 0.010
0.7043, 0.020
0.6233, 0.050
0.5736, 0.100
0.5271, 0.200
0.5013, 0.500
0.5000, 1.000
*BOUNDARY
BOT,1,2
CENTER,1
N9999,1
N9999,6
*STEP,NLGEOM,INC=200,AMPLITUDE=RAMP, UNSYMM=YES
** Step 1: Move down punch.
*VISCO,CETOL=0.01
.0015,1.,,.05
*BOUNDARY
9999,2,,-250.
*PRINT,CONTACT=YES
*CONTACT CONTROLS,FRICTION ONSET=DELAY
*CONTACT PRINT,SLAVE=ASURF
*CONTACT FILE,SLAVE=ASURF,FREQUENCY=10
*ENERGY PRINT,FREQUENCY=5
*ENERGY PRINT,ELSET=FOAM,FREQ=5
*ENERGY PRINT,ELSET=CENT,FREQ=5
*EL PRINT,FREQUENCY=50,ELSET=CENT
S,
SINV,
1-89
E,
EE,
CE,
*NODE PRINT,FREQUENCY=25
U,RF
*EL FILE,FREQUENCY=100,ELSET=CENT
S,SINV,E,EE,CE
*NODE FILE,NSET=N9999,FREQUENCY=10
U,RF
**
** ODB OUTPUT REQUESTS
**
*OUTPUT, FIELD, FREQUENCY=4
*CONTACT OUTPUT, NSET=TOP, VARIABLE=PRESELECT
*END STEP
*STEP,NLGEOM,INC=200,AMPLITUDE=RAMP, UNSYMM=YES
** Step 2: Return punch to original position.
*VISCO,CETOL=0.01
.0015,1.,,.05
*BOUNDARY,OP=MOD
9999,2,,0.0
*END STEP
1-90
Listing 1.1.4-2
*HEADING
DYNAMIC LOADING OF AN ELASTOMERIC, VISCOELASTIC
FOAM BLOCK
3 steps:
1. Apply a gravity load on the point mass
of the punch in a visco step and
let the foam relax fully.
2. Apply a sinusoidal half-wave force on the
punch and measure the displacement, velocity
and acceleration
response in a dynamic step.
3. After removing the force, continue measuring
the dynamic response.
**
UNITS: N, mm, sec
The unit of mass is N-sec^2/mm.
Observe that 1 kg = 1 N-sec^2/m =
0.001 N-sec^2/mm.
Therefore, Force = Mass x Acceleration
is consistently calculated.
**
*NODE,NSET=ALLN
1,0.,300.
19,0.,0.
481,300.,300.
499,300.,0.
601,600.,300.
619,600.,0.
*NODE,NSET=N9999
9999,0.,500.
*NSET,NSET=N1
1,
*NSET,NSET=N19
19,
*NSET,NSET=N481
481,
*NSET,NSET=N499
499,
N1,N481,12,40
1-91
*NGEN,NSET=TOP
481,601,40
N19,N499,12,40
*NGEN,NSET=BOT
499,619,40
TOP,BOT,9,2
3,19
1,3,43,41,1
*ELGEN,ELSET=FOAM
1,15,40,10,9,2,1
1,141,10
61,69
1,111,10
*ELSET,ELSET=CORNER
1,
*ELEMENT,TYPE=MASS,ELSET=PMASS
1001,9999
*MASS,ELSET=PMASS
** 200 kg = 200 N-sec^2/m = 0.2 N-sec^2/mm
0.2,
START,141.42,641.42
CIRCL,-1.,300.,0.,500.
*SURFACE,NAME=ASURF
ETOP,S3
ASURF,BSURF
*FRICTION
0.8,
ORIENTATION=RECT
*ORIENTATION,NAME=RECT,SYSTEM=RECTANGULA
1.,0.,0., 0.,1.,0.
3,0.
*MATERIAL,NAME=FOAM
1-92
*HYPERFOAM,N=2,TEST DATA INPUT,POISSON=0.0,

MODULI=INSTANTANEOUS
** Stress: MPa = N/mm^2
*UNIAXIAL TEST DATA
-.0217, -.05
-.0317, -.10
-.0367, -.15
-.0402, -.20
-.0433, -.25
-.0467, -.30
-.0504, -.35
-.0542, -.40
-.0604, -.45
-.0668, -.50
-.0759, -.55
-.0909, -.60
-.1083, -.65
-.1410, -.70
-.1933, -.75
-.2896, -.80
.0140, .08, .0046
.0334, .16, .0166
.0533, .24, .0366
.0853, .32, .0573
.1280, .40, .0817
.1653, .48, .1098
.2080, .56, .1394
.2560, .64, .1666
.2987, .72, .1904
*VISCOELASTIC,TIME=RELAXATION TEST DATA
*SHEAR TEST DATA,SHRINF=0.5000
1.0000, 0.0001
0.9695, 0.001
0.9417, 0.002
0.8722, 0.005
0.7913, 0.010
0.7043, 0.020
0.6233, 0.050
0.5736, 0.100
0.5271, 0.200
0.5013, 0.500
0.5000, 1.000
1-93
*DENSITY
** 10 kg/m^3 = 10 N-sec^2/m^4 =
** 1.E-11 N-sec^2/mm^4
1.E-11,
*BOUNDARY
BOT,1,2
CENTER,1
N9999,1
N9999,6
*AMPLITUDE,NAME=RAMP1,VALUE=RELATIVE
** Ramp to full load in 2 sec
0.,0., 2.,1., 10.,1.
*AMPLITUDE,NAME=SINE,DEFINITION=PERIODIC,
VALUE=ABSOLUTE
** Force amplitude = 5000 N of a
** half-sine wave for a 1 sec period
1,3.1416,0.,0.
0.,-5000.
*STEP,NLGEOM,INC=100, UNSYMM=YES
** Step 1: Apply gravity force on the mass
** of the punch.
1 Apply gravity force
*VISCO,CETOL=0.01
.01,5.
*DLOAD,AMPLITUDE=RAMP1
** g = 9810 mm/s^2 used in F = Mg
PMASS,GRAV,9810.,0.,-1.,0.
** New output statements to generate ODB for the
** Visualizer tutorial
*OUTPUT,FIELD,FREQUENCY=10
*CONTACT OUTPUT,SLAVE=ASURF,MASTER=BSURF,
VARIABLE=PRESELECT
*NODE OUTPUT
U,
*ELEMENT OUTPUT,ELSET=FOAM
S,E
*OUTPUT,HISTORY,FREQUENCY=1
*NODE OUTPUT,NSET=N9999
U2,V2,A2
*ELEMENT OUTPUT,ELSET=CORNER
MISES,E22,S22
**
*PRINT,CONTACT=YES
1-94
*ENERGY PRINT,ELSET=FOAM,FREQ=5
*ENERGY PRINT,ELSET=CENT,FREQ=5
S,
SINV,
E,
EE,
CE,
LOADS,
*EL FILE,FREQUENCY=50,ELSET=CENT
S,SINV,E,EE,CE
U,
CF,RF
U,CF,RF
*END STEP
** Step 2: Apply dynamic load (half sine wave)
** to the punch
2 Apply dynamic load
*DYNAMIC,HAFTOL=5000.
.01,1.
*DLOAD
PMASS,GRAV,9810.,0.,-1.,0.
*CLOAD,AMPLITUDE=SINE
9999,2
U,V,
A,
CF,RF
U,V,A,CF,RF
*END STEP
** Step 3: Remove load and let punch/foam
** system vibrate freely
3 Remove load
*DYNAMIC,HAFTOL=5000.,INITIAL=NO
.01,10.
1-95
*DLOAD
PMASS,GRAV,9810.,0.,-1.,0.
*CLOAD,OP=NEW
*END STEP
1-96
Listing 1.1.4-3
*HEADING
DYNAMIC LOADING OF AN ELASTOMERIC FOAM WITH A
HEMISPHERICAL PUNCH
(Analytical rigid surface definition)
*RESTART,TIMEMARKS=YES,WRITE,NUM=6
*NODE,NSET=ALLN
1,0.,300.
19,0.,0.
481,300.,300.
499,300.,0.
601,600.,300.
619,600.,0.
*NSET,NSET=N1
1,
*NSET,NSET=N19
19,
*NSET,NSET=N481
481,
*NSET,NSET=N499
499,
*NFILL,NSET=TOP
N1,N481,12,40
*NGEN,NSET=TOP
481,601,40
*NFILL,NSET=BOT
N19,N499,12,40
*NGEN,NSET=BOT
499,619,40
*NFILL,NSET=ALLN
TOP,BOT,9,2
3,19
*ELEMENT,TYPE=CAX4R,ELSET=FOAM
1,3,43,41,1
*ELGEN,ELSET=FOAM
1,15,40,10,9,2,1
1,141,10
61,69
*NODE,NSET=PUNCH
1000,0.,510.
1-97
*NSET,NSET=NOUT
1000,
*ELSET,ELSET=UPPER,GEN
1,141,10
*ELSET,ELSET=LOWER,GEN
9,149,10
*SOLID SECTION,ELSET=FOAM,MATERIAL=FOAM
*MATERIAL,NAME=FOAM
*HYPERFOAM,N=2,TEST DATA INPUT,POISSON=0.0
** Stress: MPa = N/mm^2
*UNIAXIAL TEST DATA
-.0217, -.05
-.0317, -.10
-.0367, -.15
-.0402, -.20
-.0433, -.25
-.0467, -.30
-.0504, -.35
-.0542, -.40
-.0604, -.45
-.0668, -.50
-.0759, -.55
-.0909, -.60
-.1083, -.65
-.1410, -.70
-.1933, -.75
-.2896, -.80
.0140, .08, .0046
.0334, .16, .0166
.0533, .24, .0366
.0853, .32, .0573
.1280, .40, .0817
.1653, .48, .1098
.2080, .56, .1394
.2560, .64, .1666
.2987, .72, .1904
*DENSITY
** 10 kg/m^3 = 10 N-sec^2/m^4 =
**1.E-11 N-sec^2/mm^4
1.E-11,
*BOUNDARY
BOT,1,2
1-98
CENTER,1
1,1
1000,1
1000,6
*AMPLITUDE,NAME=SMOOTH,DEFINITION=SMOOTH STEP,
TIME=TOTAL TIME
0.0,0.0,0.06,-250.,0.12,0.
*SURFACE,TYPE=ELEMENT,NAME=TARGET
UPPER,S3
*SURFACE, NAME=IMPACTOR,TYPE=SEGMENTS
START, 200.,510.
CIRCL,0.,310., 0.,510.
*RIGID BODY, REF NODE=1000,
ANALYTICAL SURFACE =IMPACTOR
*STEP
*DYNAMIC,EXPLICIT
,0.12
*BOUNDARY,AMPLITUDE=SMOOTH
1000,2,2,1.
*SURFACE INTERACTION,NAME=IMP_TARG
*FRICTION
0.8,
*CONTACT PAIR,INTERACTION=IMP_TARG
IMPACTOR,TARGET
*HISTORY OUTPUT,TIME=0.0008
*EL HISTORY,ELSET=UPPER
S,LE,LEP,NE,NEP
*NODE HISTORY,NSET=NOUT
RF,U
*ENERGY HISTORY
ALLKE,ALLIE,ALLAE,ALLVD,ALLWK,ETOTAL,
DT,
*FILE OUTPUT,NUMBER INTERVAL=6, TIMEMARKS=YES
*EL FILE,ELSET=FOAM
S,LE,LEP,NE,NEP
*NODE FILE
U,RF
*ENERGY FILE
*END STEP
1-99
Listing 1.1.4-4
*HEADING
STATIC INDENTATION OF AN ELASTOMERIC
FOAM WITH A HEMISPHERICAL PUNCH
DSA
Step: Displace punch 250 mm downwards.
*PARAMETER
L=0.0
MU1=0.162449
*DESIGN PARAMETER
L,MU1
*NODE,NSET=ALLN
1,0.,300.
19,0.,0.
481,300.,300.
499,300.,0.
601,600.,300.
619,600.,0.
*NODE,NSET=N9999
9999,0.,500.
*NSET,NSET=N1
1
*NSET,NSET=N19
19
*NSET,NSET=N481
481
*NSET,NSET=N499
499
N1,N481,12,40
*NGEN,NSET=TOP
481,601,40
N19,N499,12,40
*NGEN,NSET=BOT
499,619,40
TOP,BOT,9,2
3,19
*PARAMETER SHAPE VARIATION, PARAMETER=L
1-100
1, 0.000, 0.00000, 0.000

41, 0.000, 0.01463, 0.000
81, 0.000, 0.03185, 0.000
121, 0.000, 0.05210, 0.000
161, 0.000, 0.07592, 0.000
201, 0.000, 0.10396, 0.000
241, 0.000, 0.13693, 0.000
281, 0.000, 0.17573, 0.000
321, 0.000, 0.22137, 0.000
361, 0.000, 0.27507, 0.000
401, 0.000, 0.33823, 0.000
441, 0.000, 0.41257, 0.000
481, 0.000, 0.50000, 0.000
521, 0.000, 0.66667, 0.000
561, 0.000, 0.83333, 0.000
601, 0.000, 1.00000, 0.000
1,3,43,41,1
*ELGEN,ELSET=FOAM
1,15,40,10,9,2,1
1,141,10
61,69
1,101,10
START,141.42,641.42
CIRCL,-1.,300.,0.,500.
*SURFACE,NAME=ASURF
ETOP,S3
*CONTACT PAIR,INTERACTION=SMOOTH, SMALL SLIDING
ASURF,BSURF
*SURFACE INTERACTION,NAME=SMOOTH
ORIENTATION=RECT
*ORIENTATION,NAME=RECT,SYSTEM=RECTANGULAR
1.,0.,0., 0.,1.,0.
3,0.
*MATERIAL,NAME=FOAM
*HYPERFOAM,N=2
<MU1>, 8.89239, 3.597339e-05, -4.52156, 0.0, 0.0
*BOUNDARY
1-101
BOT,1,2
CENTER,1
N9999,1
N9999,6
*ELSET, ELSET=ER, GENERATE
1, 59, 1
*NSET, NSET=NU, GENERATE
41,601,40
*NSET, NSET=NR, GENERATE
59,619,40
*STEP,NLGEOM,INC=200,AMPLITUDE=RAMP, DSA
** Step 1: Move down punch.
*STATIC
.0015,1.,,.05
*BOUNDARY
9999,2,,-250.
**
** ODB OUTPUT REQUESTS
**
*NODE OUTPUT
U,
*ELEMENT OUTPUT
S,
*CONTACT OUTPUT, NSET=TOP, VARIABLE=PRESELECT
*DESIGN RESPONSE, FREQUENCY=999
*ELEMENT RESPONSE
S,
*NODE RESPONSE
U,
*CONTACT RESPONSE
CSTRESS,
CDISP,
*EL FILE, ELSET=ER, FREQUENCY=999
S,
E,
*NODE FILE, NSET=NU, FREQUENCY=999
U,
*NODE FILE, NSET=NR, FREQUENCY=999
RF,
*END STEP
1.1.5 Collapse of a concrete slab
1-102
Products: ABAQUS/Standard ABAQUS/Explicit

This problem examines the use of cracking models for the analysis of reinforced concrete structures.
The geometry of the problem is defined in Figure 1.1.5-1. A square slab is supported in the transverse
direction at its four corners and loaded by a point load at its center. The slab is reinforced in two
directions at 75% of its depth. The reinforcement ratio (volume of steel/volume of concrete) is 8.5 ´
10-3 in each direction. The slab was tested experimentally by McNeice (1967) and has been analyzed
by a number of workers, including Hand et al. (1973), Lin and Scordelis (1975), Gilbert and Warner
(1978), Hinton et al. (1981), and Crisfield (1982).
Geometric modeling
Symmetry conditions allow us to model one-quarter of the slab. A 3 ´ 3 mesh of 8-node shell elements
is used for the ABAQUS/Standard analysis. No mesh convergence studies have been performed, but
the reasonable agreement between the analysis results and the experimental data suggests that the mesh
is adequate to predict overall response parameters with usable accuracy. Three different meshes are
used in ABAQUS/Explicit to assess the sensitivity of the results to mesh refinement: a coarse 6 ´ 6
mesh, a medium 12 ´ 12 mesh, and a fine 24 ´ 24 mesh of S4R elements. Nine integration points are
used through the thickness of the concrete to ensure that the development of plasticity and failure are
modeled adequately. The two-way reinforcement is modeled using the *REBAR option. Symmetry
boundary conditions are applied on the two edges of the mesh, and the corner point is restrained in the
transverse direction.
Material properties
The material data are given in Table 1.1.5-1. The material properties of concrete are taken from Gilbert
and Warner (1978). Some of these data are assumed values, because they are not available for the
concrete used in the experiment. The assumed values are taken from typical concrete data. The
compressive behavior of concrete in the cracking model in ABAQUS/Explicit is assumed to be linear
elastic. This is a reasonable assumption for a case such as this problem, where the behavior of the
structure is dominated by cracking resulting from tension in the slab under bending.
The modeling of the concrete-reinforcement interaction and the energy release at cracking are of
critical importance to the response of a structure such as this once the concrete starts to crack. These
effects are modeled in an indirect way by adding "tension stiffening" to the plain concrete model. This
approach is described in `À cracking model for concrete and other brittle materials, '' Section 4.5.2 of
the ABAQUS Theory Manual; ``Concrete,'' Section 11.5.1 of the ABAQUS/Standard User's Manual;
and ``Cracking model for concrete,'' Section 10.4.1 of the ABAQUS/Explicit User's Manual. The
simplest tension stiffening model defines a linear loss of strength beyond the cracking failure of the
concrete. In this example three different values for the strain beyond failure at which all strength is lost
(5 ´ 10-4, 1 ´ 10-3, and 2 ´ 10-3) are used to illustrate the effect of the tension stiffening parameters on
the response.
Since the response is dominated by bending, it is controlled by the material behavior normal to the
crack planes. The material's shear behavior in the plane of the cracks is not important. Consequently,
the choice of shear retention has no significant influence on the results. In ABAQUS/Explicit the shear
1-103
retention chosen is exhausted at the same value of the crack opening at which tension stiffening is
exhausted. In ABAQUS/Standard full shear retention is used because it provides a more efficient
numerical solution.
Solution control
Since considerable nonlinearity is expected in the response, including the possibility of unstable
regimes as the concrete cracks, the modified Riks method is used with automatic incrementation in the
ABAQUS/Standard analysis. With the Riks method the load data and solution parameters serve only to
give an estimate of the initial increment of load. In this case it seems reasonable to apply an initial load
of 1112 N (250 lb) on the quarter-model for a total initial load on the structure of 4448 N (1000 lb).
This can be accomplished by specifying a load of 22241 N (5000 lb) and an initial time increment of
0.05 out of a total time period of 1.0. The analysis is terminated when the central displacement reaches
25.4 mm (1 in).
Since ABAQUS/Explicit is a dynamic analysis program and in this case we are interested in static
solutions, the slab must be loaded slowly enough to eliminate any significant inertia effects. The slab is
loaded in its center by applying a velocity that increases linearly from 0 to 2.0 in/second such that the
center displaces a total of 1 inches in 1 second. This very slow loading rate ensures quasi-static
solutions; however, it is computationally expensive. The CPU time required for this analysis can be
reduced in one of two ways: the loading rate can be increased incrementally until it is judged that any
further increase in loading rate would no longer result in a quasi-static solution, or mass scaling can be
used (see `Èxplicit dynamic analysis,'' Section 6.2.1 of the ABAQUS/Explicit User's Manual). These
two approaches are equivalent. Mass scaling is used here to demonstrate the validity of such an
approach when it is used in conjunction with the concrete model. Mass scaling is done by increasing
the density of the concrete and the reinforcement by a factor of 100, thereby increasing the stable time
increment for the analysis by a factor of 10 and reducing the computation time by the same amount
while using the original slow loading rate. Figure 1.1.5-4 shows the load-deflection response of the
slab for analyses using the 12 ´ 12 mesh with and without mass scaling. The mass scaling used does
not affect the results significantly; therefore, all subsequent analyses are performed using mass
scaling.

Results for each analysis are discussed in the following sections.
ABAQUS/Standard results
The numerical and experimental results are compared in Figure 1.1.5-2 on the basis of load versus
deflection at the center of the slab. The strong effect of the tension stiffening assumption is very clear
in that plot. The analysis with tension stiffening, such that the tensile strength is lost at a strain of 10 -3
beyond failure, shows the best agreement with the experiment. This analysis provides useful
information from a design viewpoint. The failure pattern in the concrete is illustrated in Figure 1.1.5-3,
which shows the predicted crack pattern on the lower surface of the slab at a central deflection of 7.6
mm (0.3 in).
1-104
ABAQUS/Explicit results
Figure 1.1.5-5 shows the load-deflection response of the slab for the three different mesh densities
using a tension stiffening value of 2 ´ 10-3. Since the coarse mesh predicts a slightly higher limit load
than the medium and fine meshes do and the limit loads for the medium and fine mesh analyses are
very close, the tension stiffening study is performed using the medium mesh only.
The numerical (12 ´ 12 mesh) results are compared to the experimental results in Figure 1.1.5-6 for the
three different values of tension stiffening. It is clear that the less tension stiffening used, the softer the
load-deflection response is. A value of tension stiffening somewhere between the highest and middle
values appears to match the experimental results best. The lowest tension stiffening value causes more
sudden cracking in the concrete and, as a result, the response tends to be more dynamic than that
obtained with the higher tension stiffening values.
Figure 1.1.5-7 shows the numerically predicted crack pattern on the lower surface of the slab for the
medium mesh.
Input files
ABAQUS/Standard input files

collapseconcslab_s8r.inp
S8R elements.
collapseconcslab_s9r5.inp
S9R5 elements.
collapseconcslab_postoutput.inp
*POST OUTPUT analysis.
ABAQUS/Explicit input files

mcneice_1.inp
Coarse (6 ´ 6) mesh; tension stiffening = 2 ´ 10-3.

mcneice_2.inp
Medium (12 ´ 12) mesh; tension stiffening = 2 ´ 10-3.

mcneice_3.inp
Fine (24 ´ 24) mesh; tension stiffening = 2 ´ 10-3.

mcneice_4.inp

mcneice_5.inp
1-105
mcneice_6.inp
Medium (12 ´ 12) mesh; tension stiffening = 2 ´ 10-3; no mass scaling.
References
· Crisfield, M. A, "Variable Step-Length for Nonlinear Structural Analysis," Report 1049,
Transport and Road Research Lab., Crowthorne, England, 1982.
· Gilbert, R. I. and R. F. Warner, "Tension Stiffening in Reinforced Concrete Slabs," Journal of the
Structural Division, American Society of Civil Engineers, vol. 104, ST12, pp. 1885-1900, 1978.
· Hand, F. D., D. A. Pecknold, and W. C. Schnobrich, "Nonlinear Analysis of Reinforced Concrete

Plates and Shells," Journal of the Structural Division, American Society of Civil Engineers, vol.
99, ST7, pp. 1491-1505, 1973.
· Hinton, E., H. H. Abdel Rahman, and O. C. Zienkiewicz, "Computational Strategies for

Reinforced Concrete Slab Systems," International Association of Bridge and Structural
Engineering Colloquium on Advanced Mechanics of Reinforced Concrete, pp. 303-313, Delft,
1981.
· Lin, C. S. and A. C. Scordelis, "Nonlinear Analysis of Reinforced Concrete Shells of General

Form," Journal of the Structural Division, American Society of Civil Engineers, vol. 101, pp.
523-238, 1975.
· McNeice, A. M., "Elastic-Plastic Bending of Plates and Slabs by the Finite Element Method, " Ph.
D. Thesis, London University, 1967.
Table
Table 1.1.5-1 Material properties for McNeice slab.

Concrete properties:
Properties are taken from Gilbert and Warner (1978) if available in that paper.
Properties marked with a * are not available, and are assumed values.
Young's modulus 28.6 GPa (4.15´106 lb/in2)
Poisson's ratio 0.15
Uniaxial compression values:
Yield stress 20.68 MPa (3000 lb/in 2)*
Failure stress 37.92 MPa (5500 lb/in 2)
Plastic strain at failure 1.5´10-3*
Ratio of uniaxial tension
to compression failure stress 8.36´10-2
Ratio of biaxial to uniaxial
compression failure stress 1.16*
Cracking failure stress 459.8 lb/in 2 (3.17 MPa)
Density (before mass scaling) 2.246 ´ 10-4 lb s2/in4 (2400 kg/m3)
"Tension stiffening" is assumed as a linear decrease of the stress to zero stress, at a strain
of 5´10-4, at a strain of 10´10-4, or at a strain of 20´10-4.
1-106
Steel (rebar) properties:

Young's modulus 200 GPa (29´106 lb/in2)
Yield stress 345 MPa (50´103 lb/in2)
Density (before mass scaling) 7.3 ´ 10-4 lb s2/in4 (7800 kg/m3)
Figures
Figure 1.1.5-1 McNeice slab.
Figure 1.1.5-2 Load-deflection response of McNeice slab, ABAQUS/Standard.
1-107
Figure 1.1.5-3 Crack pattern on lower surface of slab, ABAQUS/Standard.
Figure 1.1.5-4 Load-deflection response of McNeice slab, ABAQUS/Explicit; influence of mass

scaling.
1-108
Figure 1.1.5-5 Load-deflection response of McNeice slab, ABAQUS/Explicit; influence of mesh

refinement.
Figure 1.1.5-6 Load-deflection response of McNeice slab, ABAQUS/Explicit; influence of tension

stiffening.
1-109
Figure 1.1.5-7 Crack pattern on lower surface of slab, ABAQUS/Explicit.
Sample listings
1-110
Listing 1.1.5-1
*HEADING
CORNER SUPPORTED TWO-WAY SLAB TESTED
BY MCNEICE
*NODE
1,0.,0.
7,18.,0.
61,0.,18.
67,18.,18.
*NGEN,NSET=Y-SYM
1,7
*NGEN,NSET=X-SYM
1,61,10
*NGEN,NSET=LX2
61,67
*NGEN,NSET=LY2
7,67,10
*NSET,NSET=ONE
1,
*NFILL
X-SYM,LY2,6,1
*ELEMENT,TYPE=S8R,ELSET=SLAB
1,1,3,23,21,2,13,22,11
*ELGEN,ELSET=SLAB
1,3,2,1,3,20,3
*SHELL SECTION,ELSET=SLAB,MATERIAL=A1
1.75,9
*MATERIAL,NAME=A1
*ELASTIC
4.15E6,.15
*CONCRETE
3000.,0.
5500.,.0015
*FAILURE RATIOS
1.16 , .0836
*TENSION STIFFENING
1.,0.
0.,2.E-3
*REBAR,ELEMENT=SHELL,MATERIAL=SLABMT,
GEOMETRY=ISOPARAMETRIC,NAME=YY
SLAB,.014875,1.,-.435,4
1-111
GEOMETRY=ISOPARAMETRIC,NAME=XX
SLAB,.014875,1.,-.435,1
*MATERIAL,NAME=SLABMT
*ELASTIC
29.E6,
*PLASTIC
50.E3,
*BOUNDARY
Y-SYM,YSYMM
X-SYM,XSYMM
67,3
*STEP,INC=30
*STATIC,RIKS
.05,1.,,,,1,3,-1.
*CLOAD
1,3,-5000.
S,
SINV,
E,
PE,
CRACK,
*EL FILE,FREQUENCY=10
S,
SINV,
E,
PE,
CRACK,
*OUTPUT,FIELD,FREQ=10
*ELEMENT OUTPUT
S,
SINV,
E,
PE,
CRACK,
*NODE FILE,NSET=ONE
U,
*OUTPUT,FIELD
*NODE OUTPUT,NSET=ONE
U,
*END STEP
1-112
Listing 1.1.5-2
*HEADING
CORNER SUPPORTED TWO-WAY SLAB TESTED
BY MCNEICE
*NODE
1,0.,0.
7,18.,0.
61,0.,18.
67,18.,18.
*NGEN,NSET=Y-SYM
1,7
*NGEN,NSET=X-SYM
1,61,10
*NGEN,NSET=LX2
61,67
*NGEN,NSET=LY2
7,67,10
*NSET,NSET=ONE
1,
*NFILL
X-SYM,LY2,6,1
*ELEMENT,TYPE=S4R,ELSET=SLAB
1,1,2,12,11
*ELGEN,ELSET=SLAB
1,6,1,1,6,10,6
*SHELL SECTION,ELSET=SLAB,MATERIAL=CONC
1.75,9
GEOMETRY=ISOPARAMETRIC,NAME=YY
SLAB,.014875,1.,-.435,4
GEOMETRY=ISOPARAMETRIC,NAME=XX
SLAB,.014875,1.,-.435,1
*MATERIAL,NAME=SLABMT
*ELASTIC
29.E6,
*PLASTIC
50.E3,
*DENSITY
7.3e-2,
*MATERIAL,NAME=CONC
*ELASTIC
1-113
4.15E6,.15
*DENSITY
2.246e-2,
*BRITTLE CRACKING
459.8,0.
0.,2.e-3
*BRITTLE SHEAR
1.,0.
0.,2.e-3
*BOUNDARY
Y-SYM,YSYMM
X-SYM,XSYMM
67,3
*AMPLITUDE,NAME=RAMP,DEFINITION=SMOOTH STEP
0.,0.,1.,1.
*RESTART,WRITE,NUM=1
*STEP
*DYNAMIC,EXPLICIT
,1
*BOUNDARY,AMP=RAMP
1,3,3,-1
*HISTORY OUTPUT, TIME= 5.00E-4
*ENERGY HISTORY
ALLKE, ALLSE, ALLIE, ALLWK, ALLAE, ETOTAL
*NODE HISTORY,NSET=ONE
RF,
U,
*FILE OUTPUT,TIMEMARKS=YES,NUMBER=1
*NODEFILE
RF,
U,
*ELFILE
S,
LE,
CKE,
CKLE,
CKLS,
CKEMAG,
CKSTAT,
CRACK,
*END STEP
1-114
1.1.6 Jointed rock slope stability

This example illustrates the use of the jointed material model in the context of geotechnical
applications. We examine the stability of the excavation of part of a jointed rock mass, leaving a
sloped embankment. This problem is chosen mainly as a verification case because it has been studied
previously by Barton (1971) and Hoek (1970), who used limit equilibrium methods, and by
Zienkiewicz and Pande (1977), who used a finite element model.
Geometry and model

The plane strain model analyzed is shown in Figure 1.1.6-1together with the excavation geometry and
material properties. The rock mass contains two sets of planes of weakness: one vertical set of joints
and one set of inclined joints. We begin from a nonzero state of stress. In this problem this consists of
a vertical stress that increases linearly with depth to equilibrate the weight of the rock and horizontal
stresses caused by tectonic effects: such stress is quite commonly encountered in geotechnical
engineering. The active "loading" consists of removal of material to represent the excavation. It is clear
that, with a different initial stress state, the response of the system would be different. This illustrates
the need of nonlinear analysis in geotechnical applications--the response of a system to external
"loading" depends on the state of the system when that loading sequence begins (and, by extension, to
the sequence of loading). We can no longer think of superposing load cases, as is done in a linear
analysis.
Practical geotechnical excavations involve a sequence of steps, in each of which some part of the
material mass is removed. Liners or retaining walls can be inserted during this process. Thus,
geotechnical problems require generality in creating and using a finite element model: the model itself,
and not just its response, changes with time--parts of the original model disappear, while other
components that were not originally present are added. This example is somewhat academic, in that we
do not encounter this level of complexity. Instead, following the previous authors' use of the example,
we assume that the entire excavation occurs simultaneously.
Solution controls
The jointed material model includes a joint opening/closing capability. When a joint opens, the
material is assumed to have no elastic stiffness with respect to direct strain across the joint system.
Because of this, and also as a result of the fact that different combinations of joints may be yielding at
any one time, the overall convergence of the solution is expected to be nonmonotonic. In such cases
the use of *CONTROLS, ANALYSIS=DISCONTINUOUS is generally recommended to prevent
premature termination of the equilibrium iteration process because the solution may appear to be
diverging.
As the end of the excavation process is approached, the automatic incrementation algorithm reduces
the load increment significantly, indicating the onset of failure of the slope. In such analyses it is useful
to specify a minimum time step to avoid unproductive iteration.
For the nonassociated flow case UNSYMM=YES is used on the *STEP option. This is essential for
1-115
obtaining an acceptable rate of convergence since nonassociated flow plasticity has a nonsymmetric
stiffness matrix.

In this problem we examine the effect of joint cohesion on slope collapse through a sequence of
solutions with different values of joint cohesion, with all other parameters kept fixed. Figure 1.1.6-2
shows the variation of horizontal displacements as cohesion is reduced at the crest of the slope (point
A in Figure 1.1.6-1) and at a point one-third of the way up the slope (point B in Figure 1.1.6-1). This
plot suggests that the slope collapses if the cohesion is less than 24 kPa for the case of associated flow
or less than 26 kPa for the case of nonassociated flow. These compare well with the value calculated
by Barton (26 kPa) using a planar failure assumption in his limit equilibrium calculations. Barton's
calculations also include "tension cracking" (akin to joint opening with no tension strength) as we do
in our calculation. Hoek calculates a cohesion value of 24 kPa for collapse of the slope. Although he
also makes the planar failure assumption, he does not include tension cracking. This is, presumably,
the reason why his calculated value is lower than Barton's. Zienkiewicz and Pande assume the joints
have a tension strength of one-tenth of the cohesion and calculate the cohesion value necessary for
collapse as 23 kPa for associated flow and 25 kPa for nondilatant flow.
Figure 1.1.6-3 shows the deformed configuration after excavation for the nonassociated flow case and
clearly illustrates the manner in which the collapse is expected to occur. Figure 1.1.6-4shows the
magnitude of the frictional slip on each joint system for the nonassociated flow case. A few joints open
near the crest of the slope.
Input files
jointrockstabil_nonassoc_30pka.inp
Nonassociated flow case problem; cohesion = 30 kPa.
jointrockstabil_assoc_25kpa.inp
Associated flow case; cohesion = 25 kPa.
References
· Barton, N., "Progressive Failure of Excavated Rock Slopes," Stability of Rock Slopes,
Proceedings of the 13th Symposium on Rock Mechanics, Illinois, pp. 139-170, 1971.
· Hoek, E., "Estimating the Stability of Excavated Slopes in Open Cast Mines, " Trans. Inst. Min.
and Metal., vol. 79, pp. 109-132, 1970.
· Zienkiewicz, O. C. and G. N. Pande, "Time-Dependent Multilaminate Model of Rocks - A

Numerical Study of Deformation and Failure of Rock Masses," International Journal for
Numerical and Analytical Methods in Geomechanics, vol. 1, pp. 219-247, 1977.
Figures
1-116
Figure 1.1.6-1 Jointed rock slope problem.
Figure 1.1.6-2 Horizontal displacements with varying cohesion.
1-117
Figure 1.1.6-3 Deformed configuration (nonassociated flow).
Figure 1.1.6-4 Contours of frictional slip magnitudes (nonassociated flow).
1-118
Sample listings
1-119
Listing 1.1.6-1
*HEADING
JOINTED ROCK SLOPE, 2 JOINTS, C=30, NONASSOC FLOW,
.5 SH RET
*NODE
1,0.,0.
11,100.,0.
23,272.9,0.
241,0.,74.
251,100.,74.
263,272.9,74.
731,140.4,144.
743,272.9,144.
*NGEN,NSET=BLHS
1,241,40
*NGEN,NSET=BCEN
11,251,40
*NGEN,NSET=BRHS
23,263,40
*NGEN,NSET=TCEN
251,731,40
*NGEN,NSET=TRHS
263,743,40
*NFILL,BIAS=1.5,TWO STEP
BLHS,BCEN,10
*NFILL,BIAS=.66666666,TWO STEP
BCEN,BRHS,12
*NFILL,BIAS=.66666666,TWO STEP
TCEN,TRHS,12
*NSET,NSET=SLHS,GENERATE
241,251
*NSET,NSET=SRHS,GENERATE
731,743
*NSET,NSET=BOT,GENERATE
1,23
*NSET,NSET=FILN
251,411,731
*ELEMENT,TYPE=CPE4
1,1,2,42,41
101,11,12,52,51
*ELGEN,ELSET=ALLE
1,6,40,1,10,1,10
1-120
101,18,40,1,12,1,20
*SOLID SECTION,ELSET=ALLE,MATERIAL=ALLE
*MATERIAL,NAME=ALLE
*ELASTIC
2.8E7,.2
*JOINTED MATERIAL,JOINT DIRECTION=JOINT1
45.,22.5,30.
*JOINTED MATERIAL,JOINT DIRECTION=JOINT2
45.,22.5,30.
*JOINTED MATERIAL,SHEAR RETENTION
.5,
*ORIENTATION,NAME=JOINT1
1.,0.,0.,0.,1.,0.
*ORIENTATION,NAME=JOINT2
.7934,-.6088,0.,.6088,.7934,0.
*INITIAL CONDITIONS,TYPE=STRESS,GEOSTATIC
ALLE,0.,144.,-3600.,0.,.333333
*STEP, UNSYMM=YES
*GEOSTATIC
1.,1.
*DLOAD
ALLE,BY,-25.
*BOUNDARY
BOT,2,2,0.
BLHS,1,1,0.
BRHS,1,1,0.
TRHS,1,1,0.
SLHS,1,2,0.
TCEN,1,2,0.
SRHS,1,2,0.
*EL PRINT
S,MISES,PRESS
E,
PE,
PEQC,
*NODE PRINT
U,RF
*OUTPUT,FIELD
*NODE OUTPUT,NSET=FILN
U,
*ELEMENT OUTPUT
1-121
PEQC,
*OUTPUT,HISTORY
*NODE OUTPUT,NSET=FILN
U,
*END STEP
*STEP,INC=20
*STATIC
.1,1.,.001,.1
*CONTROLS,ANALYSIS=DISCONTINUOUS
*BOUNDARY,OP=NEW
BOT,2,2,0.
BLHS,1,1,0.
BRHS,1,1,0.
TRHS,1,1,0.
*END STEP
1.1.7 Indentation of a crushable foam specimen with a

hemispherical punch
In this example we consider a cylindrical specimen of crushable foam indented by a rough, rigid,
hemispherical punch. The example illustrates a typical application of crushable foam materials when
used as energy absorption devices. The effect of rate dependence of the foam is shown. Results are also
presented for both the linear and porous elastic models in ABAQUS/Standard. This problem tests the
crushable foam model but does not provide independent verification of it.
Geometry and model

The axisymmetric model (135 CAX4 elements in ABAQUS/Standard, 135 CAX4R elements in
ABAQUS/Explicit) analyzed is shown in Figure 1.1.7-1. The foam specimen has a radius of 600 mm
and a thickness of 300 mm. The punch has a radius of 200 mm. The bottom nodes of the mesh are
fixed, while the outer boundary is free to move.
Material
The material's behavior is modeled using linear elasticity and piecewise linear strain hardening.
Results from ABAQUS/Standard using porous elasticity and exponential strain hardening are also
presented. The elastic material parameters are chosen so that the elastic stiffnesses of the linear model
and of the porous elastic model are similar at the end of the first loading step, and the piecewise linear
and exponential hardening curves are similar. The material parameters are
E = 3.0 MPa (for the linear elastic model)
∙ = 0.09 (for the porous elastic model)
º = 0.0
p0 = 0.2 MPa (This is the initial value of pc .)
1-122
pt = 0.02 MPa
¾0 = 0.22 MPa
¸ = 1.7
K =1
½ = 500
In addition, we use the following material properties for the rate-dependent case:
D = 80 per sec
p=1
Contact interaction
The contact between the top exterior surface of the foam specimen and the rigid punch is modeled with
the *CONTACT PAIR option. The specimen's surface is defined by means of the *SURFACE option.
The spherical rigid punch is modeled as an analytical rigid surface with the *SURFACE option in
conjunction with the *RIGID BODY option. An ABAQUS/Explicit model using RAX2 elements
instead of an analytical rigid surface is also available. However, an analytical rigid surface provides a
more accurate representation of curved geometries. Results for the analytical rigid surface case are
presented here.
The mechanical interaction between the contact surfaces is assumed to be rough frictional contact in
ABAQUS/Standard. Therefore, the *FRICTION, ROUGH option--which enforces a no slip constraint
between the two surfaces--is used as a suboption of the *SURFACE INTERACTION property option.
In ABAQUS/Explicit the friction coefficient between the punch and the foam is 0.95. The maximum
p
shear traction due to friction is assumed to be ¾0 = 3, or 0.127 MPa.
Loading and controls

The loading is applied by first moving the rigid surface punch into the foam specimen. This step is
followed by a second unloading step in which the punch is returned to the original position. Very large
deformations will take place, so the NLGEOM parameter is needed on the *STEP option in the
ABAQUS/Standard analysis. For nonassociated flow cases UNSYMM=YES is used on the *STEP
option. This is important to obtain an acceptable rate of convergence during the equilibrium iterations,
since the nonassociated flow plasticity model used for the foam has a nonsymmetric stiffness matrix.
In the rate-independent case the punch is moved down at a constant velocity within 0.6 seconds, while
in the fast indentation problem the punch indents half of the thickness of the specimen over a period of
15 ms to demonstrate the plastic strain rate effect. In the latter case the velocity of the punch is ramped
from zero to 20 m/s.

The overall load-deflection response of the foam specimen for all four ABAQUS/Standard models is
shown in Figure 1.1.7-2. As expected, the rate-dependent cases predict higher punch forces. The
deformed configuration of the mesh at the end of the first loading step (showing actual displacements)
is shown in Figure 1.1.7-3. The magnitude of the plastic strain is given in the contour plot of Figure
1.1.7-4. The figure shows that the plastic strain magnitude in the vicinity of the punch approaches
100%.
1-123
The results differ very little between the linear and porous elastic models. This might be expected since
the problem involves loading well into the plastic range, so that elastic effects are not likely to be
significant (as long as the elasticity is sufficiently stiff to be realistic). However, in a case like this, the
linear elastic modeling is more efficient computationally. In addition, it is not possible to unload the
specimen for the rate-dependent case with porous elasticity. The crushable foam model with linear
elasticity is, therefore, recommended for most applications.
Figure 1.1.7-5 shows a contour plot of the position of the yield surface at the end of the indentation in
the rate-independent ABAQUS/Explicit case. Figure 1.1.7-6 shows the overall load-deflection
response of the foam specimen for both the rate-independent and the rate-dependent cases. Again, the
rate-dependent case predicts higher punch forces.
Input files

foamindent_ratedep_porous.inp
Rate-dependent case with porous elasticity, exponential hardening, and power law rate
dependence.
foamindent_piecerate_linear.inp
Rate-dependent case with linear elasticity, tabular hardening, and power law rate dependence
entered as a piecewise linear function.
foamindent_rateindep_porous.inp
Rate-independent case with porous elasticity and exponential hardening.
foamindent_rateindep_linear.inp
Rate-independent case with linear elasticity and tabular hardening.
foamindent_postoutput.inp
*POST OUTPUT analysis of foamindent_ratedep_porous.inp.

crushfoam_anl.inp
Rate-independent case.
crushfoam.inp
Rate-independent case using a faceted surface representation.
crushfoam_rate_anl.inp
Rate-dependent case with power law rate dependence.
crushfoam_rate.inp
Rate-dependent case with power law rate dependence using a faceted surface representation.
1-124
crushfoam_tabular_anl.inp
Rate-dependent case with power law rate dependence entered as a piecewise linear function.
crushfoam_tabular.inp
Rate-dependent case with power law rate dependence entered as a piecewise linear function for a
model using a faceted surface representation.
Figures
Figure 1.1.7-1 Model for foam indentation by spherical punch.
Figure 1.1.7-2 Punch force versus penetration response, ABAQUS/Standard.
1-125
Figure 1.1.7-3 Deformed configuration showing actual displacements, ABAQUS/Standard.
Figure 1.1.7-4 Contours of magnitude of plastic strain, ABAQUS/Standard.
Figure 1.1.7-5 Yield surface position contours, ABAQUS/Explicit.
1-126
Figure 1.1.7-6 Punch force versus penetration response, ABAQUS/Explicit.
Sample listings
1-127
Listing 1.1.7-1
*HEADING
RATE DEPENDENT FOAM INDENTATION WITH
ROUGH HEMISPHERICAL PUNCH
*NODE,NSET=ALLN
1,0.,.3
17,0.,.1
19,0.,0.
481,.3,.3
497,.3,.1
499,.3,0.
601,.6,.3
617,.6,.1
619,.6,0.
*NODE,NSET=N9999
9999,0.,.5
*NGEN,NSET=ALLN
1,17
17,19
481,497
497,499
601,617
617,619
*NSET,NSET=CL,GENERATE
1,19
*NSET,NSET=MID,GENERATE
481,499,1
*NSET,NSET=OUT,GENERATE
601,619,1
*NFILL,NSET=ALLN
CL,MID,24,20
*NFILL,NSET=ALLN
MID,OUT,6,20
19,619,20
1,3,43,41,1
*ELGEN,ELSET=FOAM
1,15,40,10,9,2,1
1,141,10
61,69
1-128
1,111,10
START,.14142,.64142
CIRCL,-.0001,.3,0.,.5
*SURFACE,NAME=ASURF
ETOP,S3
ASURF,BSURF
*FRICTION,ROUGH
*ELSET,ELSET=ELPRT
CENT,
*MATERIAL,NAME=FOAM
*POROUS ELASTIC
.09,0.,.02E6
*FOAM
.2E6,.02E6,.22E6,1.7,1.
*RATE DEPENDENT
80.,1.
*INITIAL CONDITIONS,TYPE=RATIO
ALLN,1.
*BOUNDARY
BOT,1,2
N9999,1
N9999,6
*STEP,INC=100,NLGEOM, UNSYMM=YES
*STATIC
20.E-5,15.E-3,,1.E-3
*BOUNDARY
9999,2,,-.15
*PRINT,CONTACT=YES
*ENERGY FILE,FREQUENCY=5
*CONTACT FILE,SLAVE=ASURF
S,
SINV,
1-129
E,
PE,
U,RF
*NODE FILE,NSET=N9999
U,RF
*END STEP
1-130
Listing 1.1.7-2
*HEADING
RATE DEPENDENT FOAM INDENTATION WITH
ROUGH HEMISPHERICAL PUNCH
*NODE,NSET=ALLN
1,0.,.3
17,0.,.1
19,0.,0.
481,.3,.3
497,.3,.1
499,.3,0.
601,.6,.3
617,.6,.1
619,.6,0.
*NODE,NSET=N9999
9999,0.,.5
*NGEN,NSET=ALLN
1,17
17,19
481,497
497,499
601,617
617,619
*NSET,NSET=CL,GENERATE
1,19
*NSET,NSET=MID,GENERATE
481,499,1
*NSET,NSET=OUT,GENERATE
601,619,1
*NFILL,NSET=ALLN
CL,MID,24,20
*NFILL,NSET=ALLN
MID,OUT,6,20
19,619,20
1,3,43,41,1
*ELGEN,ELSET=FOAM
1,15,40,10,9,2,1
1,141,10
61,69
1-131
*ELSET,ELSET=DBODY,GENERATE
1,111,10
*SURFACE,NAME=DSURF
DBODY,S3
*RIGID BODY,ANALYTICAL SURFACE=RSURF,REF NODE=9999
*SURFACE,TYPE=SEGMENTS,NAME=RSURF
START,.14142,.64142
CIRCL,-.0001,.3,0.,.5
*ELSET,ELSET=ELPRT
CENT,
*MATERIAL,NAME=FOAM
*ELASTIC
3.0E6 , 0.0
*FOAM,HARDENING=TABULAR
1.0,.02E6,.22E6,1.7,1.
*FOAM HARDENING
.20745E+05 , 0.000
.42916E+05 , 0.200
.75427E+05 , 0.400
.11738E+06 , 0.600
.16653E+06 , 0.800
.22000E+06 , 1.000
.24745E+06 , 1.100
.27494E+06 , 1.200
.30217E+06 , 1.300
.32890E+06 , 1.400
.35492E+06 , 1.500
.38006E+06 , 1.600
.40418E+06 , 1.700
.42720E+06 , 1.800
.44905E+06 , 1.900
.46969E+06 , 2.000
.50729E+06 , 2.200
.54008E+06 , 2.400
.56834E+06 , 2.600
.59247E+06 , 2.800
.61291E+06 , 3.000
.65083E+06 , 3.500
.67484E+06 , 4.000
.70810E+06 , 6.000
.71340E+06 , 11.000
*RATE DEPENDENT,TYPE=YIELD RATIO
1-132
1.00, 0.0
126.0,10000.0
*CONTACT PAIR,INTERACTION=RDINT1
DSURF,RSURF
*SURFACE INTERACTION,NAME=RDINT1
1.,
*FRICTION,ROUGH
*BOUNDARY
BOT,1,2
N9999,1
N9999,6
*STATIC
20.E-5,15.E-3,,1.E-3
*BOUNDARY
9999,2,,-.15
*PRINT,CONTACT=YES
*CONTACT PRINT,SLAVE=DSURF,FREQUENCY=999
*CONTACT FILE,SLAVE=DSURF,FREQUENCY=999
S,
SINV,
E,
PE,
U,RF
*NODE FILE,NSET=N9999
U,RF
*END STEP
*STATIC
20.E-5,15.E-3,,1.E-3
*BOUNDARY
9999,2,,0.0
*END STEP
1-133
Listing 1.1.7-3
*HEADING
FOAM INDENTATION WITH ROUGH HEMISPHERICAL PUNCH
*RESTART,TIMEMARKS=YES,WRITE,NUMBER=1
*NODE,NSET=ALLN
1,0.,0.3
19,0.,0.
481,0.3,0.3
499,0.3,0.
601,0.6,0.3
619,0.6,0.
*NSET,NSET=N1
1,
*NSET,NSET=N19
19,
*NSET,NSET=N481
481,
*NSET,NSET=N499
499,
*NFILL,NSET=TOP
N1,N481,12,40
*NGEN,NSET=TOP
481,601,40
*NFILL,NSET=BOT
N19,N499,12,40
*NGEN,NSET=BOT
499,619,40
*NFILL,NSET=ALLN
TOP,BOT,9,2
3,19
*ELEMENT,TYPE=CAX4R,ELSET=FOAM
1,3,43,41,1
*ELGEN,ELSET=FOAM
1,15,40,10,9,2,1
1,141,10
61,69
*NODE,NSET=PUNCH
1000,0.,0.51
*ELSET,ELSET=UPPER,GEN
1,141,10
1-134
*MATERIAL,NAME=FOAM
*ELASTIC
3.E6,0.0
*FOAM
1.0,0.02E6,0.22E6,
*FOAM HARDENING
0.20745E5,0.0
0.42916E5,0.2
0.75427E5,0.4
0.11738E6,0.6
0.16653E6,0.8
0.22E6,1.0
0.24745E6,1.1
0.27494E6,1.2
0.30217E6,1.3
0.3289E6,1.4
0.35492E6,1.5
0.38006E6,1.6
0.40418E6,1.7
0.4272E6,1.8
0.44905E6,1.9
0.46969E6,2.0
0.50729E6,2.2
0.54008E6,2.4
0.56834E6,2.6
0.59247E6,2.8
0.61291E6,3.0
0.65083E6,3.5
0.67484E6,4.0
0.70810E6,6.0
0.71340E6,11.0
*DENSITY
500.,
*BOUNDARY
BOT,1,2
CENTER,1
1,1
1000,1
1000,6
*AMPLITUDE,NAME=SMOOTH,TIME=TOTAL TIME,
DEFINITION=SMOOTH STEP
0.0,0.0,0.6,1.
1-135
UPPER,S3
*SURFACE, NAME=IMPACTOR, TYPE=SEGMENTS
START, 0.2,0.51
CIRCL, 0.,0.31, 0.,0.51
ANALYTICAL SURFACE =IMPACTOR
*STEP
*DYNAMIC,EXPLICIT
,0.6
*BOUNDARY,TYPE=DISPLACEMENT,AMPLITUDE=SMOOTH
1000,2,2,-0.2
*SURFACE INTERACTION,NAME=IMP_TO_FOAM
*FRICTION,TAUMAX=0.127E6
0.95,
*CONTACT PAIR,INTERACTION=IMP_TO_FOAM
IMPACTOR,TARGET
*FILE OUTPUT,TIMEMARKS=YES,NUMBER INTERVAL=1
*EL FILE
LE,ERV
*NODE FILE
U,
*END STEP
1.1.8 Notched beam under cyclic loading

This example illustrates the use of the nonlinear isotropic/kinematic hardening material model to
simulate the response of a notched beam under cyclic loading. The model has two features to simulate
plastic hardening in cyclic loading conditions: the center of the yield surface moves in stress space
(kinematic hardening behavior), and the size of the yield surface evolves with inelastic deformation
(isotropic hardening behavior). This combination of kinematic and isotropic hardening components is
introduced to model the Bauschinger effect and other phenomena such as plastic shakedown,
ratchetting, and relaxation of the mean stress.
The component investigated in this example is a notched beam subjected to a cyclic 4-point bending
load. The results are compared with the finite element results published by Benallal et al. (1988) and
Doghri (1993). No experimental data are available.
Geometry and model

The geometry and mesh are shown in Figure 1.1.8-1. Figure 1.1.8-2 shows the discretization in the
vicinity of the notch, which is the region of interest in this analysis. Only one-half of the beam is
modeled since the geometry and loading are symmetric with respect to the x= 0 plane. All dimensions
1-136
are given in millimeters. The beam is 1 mm thick and is modeled with plane strain, second-order,
reduced-integration elements (type CPE8R). The mesh is chosen to be similar to the mesh used by
Doghri (1993). No mesh convergence studies have been performed.
Material
The material properties reported by Doghri (1993) for a low-carbon ( AISI 1010), rolled steel are used
in this example.
A Young's modulus of E= 210 GPa and a Poisson's ratio of º = 0.3 define the elastic response of the
material. The initial yield stress is ¾j0 = 200 MPa.
The nonlinear evolution of the center of the yield surface is defined by the equation
1 pl pl
®_ = C (¾ ¡ ® )"¹_ ¡ ° ® "¹_ ;
¾0
where ® is the backstress, ¾ 0 is the size of the yield surface (size of the elastic range), "¹pl is the
equivalent plastic strain, and C = 25.5 GPa and ° = 81 are the material parameters that define the
initial hardening modulus and the rate at which the hardening modulus decreases with increasing
p
plastic strain, respectively. The quantity 2=3 C=° = 257 MPa defines the limiting value of the
p
equivalent backstress ® ¹ = ® dev : ® dev ; further hardening is possible only through the change in the
size of the yield surface (isotropic hardening).
The isotropic hardening behavior of this material is modeled with the exponential law
pl
¾ 0 = ¾j0 + Q1 (1 ¡ e¡b "¹ );
where ¾ 0 is the size of the yield surface (size of the elastic range), Q1 = 2000 MPa is the maximum
increase in the elastic range, and b = 0.26 defines the rate at which the maximum size is reached as
plastic straining develops.
The material used for this simulation is cold rolled. This work hardened state is represented by
specifying an initial equivalent plastic strain "¹pl j0 = 0.43 (so that ¾ 0 = 411 MPa) and an initial
backstress tensor
2 3
128 0 0
®j0 = 4 0 ¡181 0 5 MPa;
0 0 53
using the *INITIAL CONDITIONS, TYPE=HARDENING option.
Loading and boundary conditions

The beam is subjected to a 4-point bending load. Since only half of the beam is modeled, the model
contains one concentrated load at a distance of 26 mm from the symmetry plane (see Figure 1.1.8-1).
The pivot point is 42 mm from the symmetry plane. The simulation runs 3 1/2 cycles over 7 time units.
In each cycle the load is ramped from zero to 675 N and back to zero. An amplitude curve is used to
1-137
describe the loading and unloading. The increment size is restricted to a maximum of 0.125 to force
ABAQUS to follow the prescribed loading/unloading pattern closely.

Figure 1.1.8-3 shows the final deformed shape of the beam after the 3 1/2 cycles of load; the final load
on the beam is 675 N.
The deformation is most severe near the root of the notch. The results reported in Figure 1.1.8-4 and
Figure 1.1.8-5are measured in this area (element 166, integration point 3). Figure 1.1.8-4shows the
time evolution of stress versus strain. Several important effects are predicted using this material model.
First, the onset of yield occurs at a lower absolute stress level during the first unloading than during the
first loading, which is the Bauschinger effect. Second, the stress-strain cycles tend to shift and stabilize
so that the mean stress decreases from cycle to cycle, tending toward zero. This behavior is referred to
as the relaxation of the mean stress and is most pronounced in uniaxial cyclic tests in which the strain
is prescribed between unsymmetric strain values. Third, the yield surface shifts along the strain axis
with cycling, whereas the shape of the stress-strain curve tends to remain similar from one cycle to the
next. This behavior is known as ratchetting and is most pronounced in uniaxial cyclic tests in which
the stress is prescribed between unsymmetric stress values. Finally, the hardening behavior during the
first half-cycle is very flat relative to the hardening curves of the other cycles, which is typical of work
hardened metals whose initial hardened state is a result of a large monotonic plastic deformation
caused by a forming process such as rolling. The low hardening modulus is the result of the initial
conditions
q on backstress, which places the center of the yield surface at a distance of
dev dev
® j0 : ®j0 = 228 MPa away from the origin of stress space. Since this distance is close to the
maximum possible distance (257 MPa), most of the hardening during the first cycle is isotropic.
These phenomena are modeled in this example primarily by the nonlinear evolution of the backstress,
since the rate of isotropic hardening is very small. This behavior can be verified by conducting an
analysis in which the elastic domain remains fixed throughout the analysis.
Figure 1.1.8-5 shows the evolution of the direct components of the deviatoric part of the backstress
tensor. The backstress components evolve most during the first cycle as the Bauschinger effect
overcomes the initial hardening configuration. Only the deviatoric components of the backstress are
shown so that the results obtained using ABAQUS can be compared to those reported by Doghri
(1993). Since ABAQUS uses an extension of the Ziegler evolution law, a backstress tensor with
nonzero pressure is produced, whereas the backstress tensor produced with the law used by Doghri
(which is an extension of the linear Prager law) is deviatoric. Since the plasticity model considers only
the deviatoric part of the backstress, this difference in law does not affect the other solution
variables.
The results shown in Figure 1.1.8-4 and Figure 1.1.8-5 agree well with the results reported by Doghri
(1993).
Input files
cyclicnotchedbeam.inp
1-138
Input data.
cyclicnotchedbeam_mesh.inp
Element and node data.
References
· Benallal, A., R. Billardon, and I. Doghri, "An Integration Algorithm and the Corresponding
Consistent Tangent Operator for Fully Coupled Elastoplastic and Damage Equations, "
Communications in Applied Numerical Methods, vol. 4, pp. 731-740, 1988.
· Doghri, I., "Fully Implicit Integration and Consistent Tangent Modulus in Elasto-Plasticity, "
International Journal for Numerical Methods in Engineering, vol. 36, pp. 3915-3932, 1993.
Figures
Figure 1.1.8-1 Undeformed mesh (dimensions in mm).
Figure 1.1.8-2 Magnified view of the root of the notch.
1-139
Figure 1.1.8-3 Deformed mesh at the conclusion of the simulation. Displacement magnification factor
is 3.
Figure 1.1.8-4 Evolution of stress versus strain in the vicinity of the root of the notch.
1-140
Figure 1.1.8-5 Evolution of the diagonal components of the deviatoric part of the backstress tensor.
Sample listings
1-141
Listing 1.1.8-1
*HEADING
NOTCHED BEAM UNDER CYCLIC LOADING
*INCLUDE,INPUT=cyclicnotchedbeam_mesh.inp
*SOLID SECTION,ELSET=ELALL,MATERIAL=MAT
*MATERIAL,NAME=MAT
*ELASTIC
2.1E5,.3
*PLASTIC,HARDENING=COMBINED,DATA TYPE=PARAMETERS
200.,25500.,81.
*CYCLIC HARDENING,PARAMETERS
200.,2000.,.26
*INITIAL CONDITIONS,TYPE=HARDENING
ELALL,.43,128.,-181.,53.
*BOUNDARY
PIVOT,2
XSYM,1
*AMPLITUDE,NAME=AMP
0.,1.,1.,0.,2.,1.,3.,0.,
4.,1.,5.,0.,6.,1.
*RESTART,WRITE,FREQ=1
*STEP
first loading
*STATIC
.125,1.,,.125
*CLOAD, OP=NEW
2210,2,675.
*NODE FILE,FREQ=8
U,
*EL FILE,ELSET=REFINE,FREQ=8
S,E,PE,ALPHA
*NODE OUTPUT
U,
*OUTPUT,HISTORY,FREQ=1
*ELEMENT OUTPUT,ELSET=REFINE
S,E,PE,ALPHA
*NODE PRINT,FREQ=0
*EL PRINT,FREQ=0
*END STEP
*STEP,INC=100
additional unloadings and loadings
1-142
*STATIC
.125,6.,,.125
*CLOAD,OP=NEW,AMPLITUDE=AMP
2210,2,675.
*END STEP
1.1.9 Hydrostatic fluid elements: modeling an airspring

Airsprings are rubber or fabric actuators that support and contain a column of compressed air. They are
used as pneumatic actuators and vibration isolators. Unlike conventional pneumatic cylinders,
airsprings have no pistons, rods, or dynamic seals. This makes them better suited to handle off-center
loading and shock. In addition, airsprings are considerably more flexible than other types of isolators:
the airspring's inflation pressure can be changed to compensate for different loads or heights without
compromising isolation efficiency. Dils (1992) provides a brief discussion of various practical uses of
airsprings.
In this section two examples of the analysis of a cord-reinforced rubber airspring are discussed. Static
analyses are performed in ABAQUS/Standard, and quasi-static analyses are performed in
ABAQUS/Explicit. The first example is a three-dimensional, half-symmetry model that uses
finite-strain shell elements to model the rubber spring; three-dimensional, hydrostatic fluid elements to
model the air-filled cavity; and rebars to model the multi-ply steel reinforcements in the rubber
membrane. In addition, a three-dimensional, element-based rigid surface is used to define the contact
between the airspring and the lateral metal bead.
The second example is a two-dimensional, axisymmetric version of the first model that uses composite
axisymmetric, finite-strain shell elements to model the cord-reinforced rubber spring; axisymmetric,
hydrostatic fluid elements to model the air-filled cavity; and an axisymmetric, element-based rigid
surface in the contact definition.
The three-dimensional, 180° model uses a hyperelastic material model in conjunction with steel rebars
to model the cord-reinforced rubber membrane, and the axisymmetric shell model uses a composite
shell section consisting of a thin orthotropic elastic layer sandwiched between two hyperelastic layers.
The orthotropic layer captures the mechanical properties of the rebar definition used in the 180° model.
The orthotropic material constants have been obtained by performing simple tests on a typical element
of the three-dimensional model. The three-dimensional shell model uses rebars with material
properties that are initially identical to the properties of the composite shell section in the
axisymmetric shell model.
For comparison, ABAQUS/Standard input files for the two- and three-dimensional models that use
finite-strain membrane elements along with rebars to model the cord-reinforced rubber spring,
hydrostatic fluid elements to model the air-filled cavity, and an element-based rigid surface in the
contact definition are also included.
In all analyses the air inside the airspring cavity has been modeled as a compressible or "pneumatic"
fluid satisfying the ideal gas law.
1-143
Geometry and model

The dimensions of the airspring have been inferred from the paper by Fursdon (1990). This airspring,
shown in Figure 1.1.9-1, is fairly large and is used in secondary suspension systems on railway bogies.
However, the shape of the airspring is typical of airsprings used in other applications. The airspring's
cross-section is shown in Figure 1.1.9-2. The airspring is toroidal in shape, with an inner radius of 200
mm and an outer radius of 400 mm. The airspring has been idealized in the model as consisting of two
circular, metal disks connected to each other via a rubber component. The lower disk has a radius of
200 mm, and the upper disk has a radius of 362.11 mm. The disks are initially coaxial and are 100 mm
apart. The rubber component is doubly curved and toroidal in shape. The rubber is constrained in the
radial direction by a circular bead 55 mm in radius that goes around the circumference of the upper
disk.
The rubber "hose" in the half-symmetry, three-dimensional model is modeled with 550 S4R
finite-strain shell elements. The mesh in the upper hemisphere of the hose is more refined than that in
the lower hemisphere, because the rubber membrane undergoes a reversal in curvature in the upper
region as it contours the circular bead attached to the upper disk. The circular bead is modeled by an
axisymmetric, element-based rigid surface. Contact with the rubber is enforced by defining a contact
pair between this rigid surface and a surface defined on the (deformable) shell mesh in the contacting
region. The metal disks are assumed to be rigid relative to the rubber component of the airspring. In the
ABAQUS/Standard analysis both disks are modeled by means of boundary conditions and multi-point
constraints. In the ABAQUS/Explicit analysis the lower metal disk is modeled by means of boundary
conditions, while the upper disk is modeled as part of the rigid surface. The mesh of the rubber
membrane and the contact master surface is shown in Figure 1.1.9-3.
Three-dimensional F3D3 and F3D4 hydrostatic fluid elements are used to model the air-filled airspring
cavity. The F3D4 elements are used to cover the portion of the boundary of the cavity that is associated
with the rubber hose. These fluid elements share the same nodes as the S4R elements. To define the
cavity completely and to ensure proper calculation of its volume (see ``Modeling fluid-filled cavities,''
Section 7.9.1 of the ABAQUS/Standard User's Manual and Section 7.4.1 of the ABAQUS/Explicit
User's Manual), the F3D3 hydrostatic fluid elements are defined along the bottom and top rigid disk
boundaries of the cavity, even though no displacement elements exist along those surfaces. All the
hydrostatic fluid elements have been grouped into an element set named FLUID and share the cavity
reference node 50000. The cavity reference node has a single degree of freedom representing the
pressure inside the cavity. This node is specified on the property reference option for the hydrostatic
fluid elements, *FLUID PROPERTY. Because of symmetry only half of the cavity boundary has been
modeled. The cavity reference node 50000 has been placed on the model's symmetry plane, y = 0, to
assure proper calculation of the cavity volume. Figure 1.1.9-4 shows the mesh of the airspring's cavity.
To facilitate comparisons, the two-dimensional axisymmetric model uses the same cross-sectional
mesh refinement as the 180° model. For the shell model the rubber component is modeled with 25
SAX1 shell elements. For the membrane model the SAX1 elements are replaced with either MAX1
elements or MGAX1 elements. The circular bead is modeled in ABAQUS/Standard with a segmented
analytical rigid surface and in ABAQUS/Explicit with an element-based rigid surface constructed of
RAX2 rigid elements. Contact with the hose is enforced by defining a contact pair between this rigid
1-144
surface and a surface defined on the (deformable) shell mesh in the contacting region. The rigid metal
disks are again modeled by boundary conditions and multi-point constraints in ABAQUS/Standard; in
ABAQUS/Explicit the lower rigid metal disk is again modeled by boundary conditions, and the upper
rigid metal disk is modeled as part of the rigid body. Two-dimensional FAX2 hydrostatic fluid
elements are used to model the airspring cavity. The mesh of the rubber membrane and the contact
master surface is shown in Figure 1.1.9-5, and the mesh of the cavity is shown in Figure 1.1.9-6.
Symmetry boundary conditions and initial shell curvature

Symmetry has been exploited in the three-dimensional airspring model, and the plane y = 0 has been
made a plane of symmetry. Since S4R shell elements are true curved shell elements, accurate definition
of the initial curvature of the surface being modeled is required, especially on the plane of symmetry. If
the user does not provide this information by specifying the normal to the surface at the shell nodes,
ABAQUS will estimate the normal direction based on the coordinates of the surrounding nodes on the
shell. Normals computed in this fashion will be inaccurate on the symmetry plane: they will have
out-of-plane components, which will lead to convergence difficulties in ABAQUS/Standard and
inaccurate results. To avoid these difficulties, direction cosines have been specified for all shell nodes
in the model.
Material properties
The walls of an airspring's rubber component are made from plies of symmetrically placed, positively
and negatively oriented reinforcement cords. The walls of an actual component are made of several
such layers. However, for the purposes of the three-dimensional example problem being considered,
the airspring's wall is taken to be a rubber matrix with a single 6-mm-thick symmetric layer of
positively and negatively oriented cords. The cords are modeled by uniformly spaced skew rebars in
the shell elements. The rebars are assumed to be made of steel. The rubber is modeled as an
incompressible Mooney-Rivlin (hyperelastic) material with C10 = 3.2 MPa and C01 = 0.8 MPa, and
the steel is modeled as a linear elastic material with E = 210.0 GPa and º = 0.3.
Skew rebar orientations in shell elements are defined by giving the angle between the local 1-axis and
the rebars. The default local 1-direction is the projection of the global x-axis onto the shell surface (see
``Conventions,'' Section 1.2.2 of the ABAQUS/Standard User's Manual and the ABAQUS/Explicit
User's Manual). It is for this reason, and to make the rebar definition uniform for all elements, that the
axis of revolution of the airspring model has been chosen to be the global x-axis. Two rebar sets,
PLSBAR and MNSBAR, have been defined with orientation angles of 18° and -18°, respectively. The
cross-sectional area of the rebars is 1 mm 2, and they are spaced every 3.5 mm in the shell surface.
The above rebar specification is simplified and somewhat unrealistic. A more realistic simulation
would require different rebar definitions in each ring of elements in the airspring model. The
reinforced plies used in the manufacture of the airspring are located in an initially cylindrical tube with
uniform rebar angles. However, the transformation of these layers from a cylindrical geometry to a
toroidal one gives the airspring a variable rebar angle and rebar spacing that is dependent on the radius
from the axis of revolution of the torus and on the initial rebar angle (see Fursdon, 1990).
In the axisymmetric shell model the airspring walls are modeled by a three-layer composite shell
1-145
section. The two outer layers are each 2.5 mm thick and made up of the same Mooney-Rivlin material
that is used in the 180° model. The middle "rebar" layer is 1 mm thick and is made up of an orthotropic
elastic material that captures the mechanical behavior of the positively and negatively oriented rebar
definition used in the three-dimensional airspring model.
The plane stress orthotropic engineering constants are obtained by looking at the response of a typical
element in the three-dimensional model (element 14) subjected to uniaxial extensions along the local
1- and 2-directions. Using a shell thickness of 1 mm, the in-plane states of stress and strain resulting
from these two tests are
Test "1 "2 ¾1 (MPa) ¾2 (MPa)
1-direction 1.00 ´ 10 -2
-8.75 ´ 10 -2
2.48 ´ 10 1
-2.41 ´ 10-5
2-direction -1.05 ´ 10-3 1.00 ´ 10-2 -5.96 ´ 10-6 2.86 ´ 10-1
For a plane-stress orthotropic material the in-plane stress and strain components are related to each
other as follows:
µ ¶ µ ¶µ ¶
"1 1=E1 ¡º12 =E1 ¾1
= ;
"2 ¡º21 =E2 1=E2 ¾2
where E1 , E2 , º12 , and º21 are engineering constants. Solving for these constants using the above
stress-strain relation and the results of the two uniaxial tests yields
E1 = 2:48 £ 103 MPa

E2 = 2:86 £ 101 MPa
º12 = 9:1
º21 = 0:1
The remaining required engineering constants--G12 , G13 , and G23 --play no role in the rebar layer
definition. Consequently, they have been arbitrarily set to be equal to the shear modulus of the rubber,
which is given by 2(C10 + C01 ).
For the axisymmetric membrane model the bulk material is chosen to have the same material
properties (Mooney-Rivlin hyperelastic) as those used in the 180° model and the axisymmetric shell
model. The rebar parameters and material properties are chosen such that they capture the initial
material properties of the sandwiched steel layer in the axisymmetric shell model. The principal
material directions do not rotate in the axisymmetric shell model (they are the default element basis
directions--the meridional and the hoop directions, respectively). However, they do rotate with finite
strain in the axisymmetric membrane model as a result of the use of rebars. Initial stresses are applied
to the rebars in the axisymmetric membrane model.
In all analyses the air inside the airspring cavity has been modeled as an ideal gas with the following
properties: its reference density is 1.774 kg/m 3 at a room temperature of 27° C and an ambient pressure
of 101.36 kPa.
1-146
Loading
The airspring is first pressurized to 506.6 ´ 103 kPa (5 atms) while holding the upper disk fixed. This
pressure is applied by prescribing degree of freedom 8 at the cavity reference node using the
*BOUNDARY option. In this case the air volume is adjusted automatically to fill the cavity.
In the next step the *BOUNDARY, OP=NEW option is used to remove the boundary condition on the
pressure degree of freedom, thus sealing the cavity with the current air volume. In addition, during this
step in the ABAQUS/Standard analysis the boundary condition on the vertical displacement degree of
freedom of the rigid body reference node is removed, and in its place a downward load of 150 kN is
applied.
The next two steps in the ABAQUS/Standard axisymmetric model analysis are linear perturbation
steps to test the axial stiffness of the airspring with the cavity pressure allowed to vary (closed cavity
conditions) and with it fixed. The three-dimensional ABAQUS/Standard analysis contains three linear
perturbation steps, all under variable cavity pressure (closed cavity) conditions: the first to test the
axial stiffness of the airspring, the second to test its lateral stiffness, and the third to test the rotational
stiffness for rocking motion in the symmetry plane.
The ABAQUS/Explicit axisymmetric analysis concludes with a nonlinear step in which the airspring is
subjected to a downward displacement of 75 mm. The ABAQUS/Standard axisymmetric analysis
concludes with a nonlinear step in which the airspring is compressed by increasing the downward load
to 240.0 kN. The three-dimensional analyses conclude with a nonlinear step in which the airspring is
subjected to a lateral displacement of 20 mm.

Figure 1.1.9-7 and Figure 1.1.9-8 show displaced plots of the axisymmetric shell model at the end of
the pressurization step, Step 1. It is of interest to compare the results from this model with those from
the 180° model to validate the material model that was used for the rebar reinforcements in the
axisymmetric model. A close look at the nodal displacements reveals that the deformation is
practically identical for the two models. Moreover, the axial reaction force at the rigid body reference
node is 156 kN for the axisymmetric model and 155 kN for the 180° model (after multiplication by a
factor of 2). The cavity volume predicted by the axisymmetric model is 8.22 ´ 10-2 m3 versus 8.34 ´
10-2 m3 for the 180° model (again, after multiplication by a factor of 2).
ABAQUS/Standard results
Figure 1.1.9-9 shows a displaced plot of the axisymmetric model at the end of Step 2. The spring has
undergone an upward displacement along its axis because the downward load being applied to it is
slightly smaller than the axial reaction force at the end of the previous step. Once again, comparison
between the results from both models reveals almost identical solutions.
Linearized stiffnesses for the airspring are obtained from the linear perturbation steps. The stiffness is
computed by dividing the relevant reaction force at the rigid body reference node by the appropriate
displacement, which gives the airspring's axial stiffness under variable cavity pressure conditions as
826 kN/m for the axisymmetric model. From the results of Step 4 the axial stiffness under "fixed"
1-147
cavity pressure conditions is 134 kN/m. The difference in axial stiffness between these two cases (a
factor of 6) is the result of differences in cavity pressure experienced during axial compression. Under
variable cavity pressure conditions, a fixed mass of fluid (air) is contained in a cavity whose volume is
decreasing; thus, the cavity pressure increases. Under "fixed" cavity pressure conditions, the pressure
is prescribed as a constant value for the step. For the 180° model the predicted stiffnesses are as
follows: the axial stiffness is 821 kN/m, the lateral stiffness is 3.31 MN/m, and the rotational stiffness
is 273 kN/m.
Figure 1.1.9-10 shows a series of displaced plots associated with the compression of the axisymmetric
airspring model during Step 5. Figure 1.1.9-11 shows the load-deflection curve corresponding to this
deformation. The response of the airspring is slightly nonlinear; consequently, there is good agreement
between the axial stiffness obtained with the linear perturbation analysis (Step 3) and that obtained
from the slope of the load-displacement curve. Figure 1.1.9-12 shows a plot of cavity pressure versus
the downward displacement of the rigid body in Step 5, which shows that the gauge pressure in the
cavity increases by approximately 50% during this step. This pressure increase substantially affects the
deformation of the airspring structure and cannot be specified as an externally applied load during the
step since it is an unknown quantity. Figure 1.1.9-13 shows a plot of cavity volume versus the
downward displacement of the rigid body in Step 5. The corresponding results for the axisymmetric
membrane model are in good agreement with the above results.
Figure 1.1.9-14 shows the displaced plot of the 180° model at the end of Step 6, in which a lateral
displacement was applied to the airspring. Figure 1.1.9-15 shows the load-deflection curve
corresponding to this deformation. Once again, good agreement is found between the lateral stiffness
predicted from the linear perturbation analysis (Step 4) and that obtained from the slope of the
load-displacement curve.
ABAQUS/Explicit results
Figure 1.1.9-16 shows a series of displaced plots associated with the compression of the axisymmetric
model during the second step. Figure 1.1.9-17 shows the load-deflection curve corresponding to this
deformation. Although the displacement of the rigid body was applied over a short enough time period
to cause significant inertial effects in the model, there is still good agreement between the slope of the
load-displacement curve in this example and the slope of the load-displacement curve for the same
analysis performed statically in ABAQUS/Standard. Figure 1.1.9-18 shows a plot of cavity pressure
versus the downward displacement of the rigid body in Step 2, which shows that the gauge pressure in
the cavity increases by approximately 50 percent during this step. This pressure increase substantially
affects the deformation of the airspring structure and cannot be specified as an externally applied load
during the step since it is an unknown quantity. Figure 1.1.9-19 shows a plot of cavity volume versus
the downward displacement of the rigid body in Step 2.
Figure 1.1.9-20 shows the displaced plot of the 180° model at the end of Step 2, in which a lateral
displacement was applied to the airspring. Figure 1.1.9-21 shows the load-deflection curve
corresponding to this deformation. Although there is a significant amount of noise that results from the
contact conditions and the coarseness of the mesh, the load-deflection curve shows good agreement
between the analysis performed quasi-statically in ABAQUS/Explicit and the same analysis performed
statically in ABAQUS/Standard. The load versus displacement curve shown has been smoothed to
1-148
eliminate some of the noise.
Input files
hydrofluidairspring_3d_shell.inp
Three-dimensional ABAQUS/Standard model using shell elements.
hydrofluidairspring_axisymm.inp
Axisymmetric ABAQUS/Standard model.
airspring_s4r.inp
Three-dimensional ABAQUS/Explicit model using shell elements.
airspring_sax1.inp
Axisymmetric ABAQUS/Explicit model.
hydrofluidairspring_3d_mem.inp
Three-dimensional ABAQUS/Standard model using membrane elements.
hydrofluidairspring_max1.inp
ABAQUS/Standard analysis using MAX1 elements with rebars.
hydrofluidairspring_mgax1.inp
ABAQUS/Standard analysis using MGAX1 elements with rebars.
References
· Dils, M., `Àir Springs vs. Air Cylinders,'' Machine Design, May 7, 1992.
· Fursdon, P. M. T., ``Modelling a Cord Reinforced Component with ABAQUS,'' 6th UK

ABAQUS User Group Conference Proceedings, 1990.
Figures
Figure 1.1.9-1 A cord reinforced airspring.
1-149
Figure 1.1.9-2 The airspring model cross-section.
Figure 1.1.9-3 180° model: mesh of the rubber membrane and partial view of the axisymmetric
contact master surface.
Figure 1.1.9-4 180° model: mesh of the airspring cavity.
Figure 1.1.9-5 Axisymmetric model: mesh of the rubber membrane and the contact master surface.
1-150
Figure 1.1.9-6 Axisymmetric model: mesh of the airspring cavity.
Figure 1.1.9-7 Axisymmetric ABAQUS/Standard model: deformed configuration at the end of Step 1.
Figure 1.1.9-8 Axisymmetric ABAQUS/Explicit model: deformed configuration at the end of Step 1.
1-151
Figure 1.1.9-9 Axisymmetric ABAQUS/Standard model: deformed configuration at the end of Step 2.
Figure 1.1.9-10 Axisymmetric ABAQUS/Standard model: progressive deformed configurations

during Step 5.
Figure 1.1.9-11 Axisymmetric ABAQUS/Standard model: load-displacement curve for Step 5.
1-152
Figure 1.1.9-12 Axisymmetric ABAQUS/Standard model: cavity pressure versus downward

displacement in Step 5.
Figure 1.1.9-13 Axisymmetric ABAQUS/Standard model: cavity volume versus downward

1-153
Figure 1.1.9-14 180° ABAQUS/Standard model: deformed configuration at the end of Step 6.
Figure 1.1.9-15 180° ABAQUS/Standard model: load-displacement curve for Step 6.
1-154
Figure 1.1.9-16 Axisymmetric ABAQUS/Explicit model: progressive deformed configurations during

Step 2.
Figure 1.1.9-17 Axisymmetric ABAQUS/Explicit model: load-displacement curve for Step 2.
1-155
Figure 1.1.9-18 Axisymmetric ABAQUS/Explicit model: cavity pressure versus downward

Figure 1.1.9-19 Axisymmetric ABAQUS/Explicit model: cavity volume versus downward

1-156
Figure 1.1.9-20 180° ABAQUS/Explicit model: deformed configuration at the end of Step 2.
Figure 1.1.9-21 180° ABAQUS/Explicit model: load-displacement for Step 2.
Sample listings
1-157
Listing 1.1.9-1
*HEADING
AIRSPRING -- 180 DEG MODEL
*RESTART,WRITE,OVERLAY
*NODE, NSET = FPLANE
1,2.E-2, 0.,0.
4,2.E-2, 0.,20.E-2,0.,0.,1.0
6,0.,0.,20.E-2,0.,0.,1.0
26,0.,0.,40.E-2,0.,0.,-1.0
32,6.293E-2,0.,37.771E-2,-0.58431,0.,-0.81153
38,10.698E-2,0.,36.211E-2,0.,0.,-1.0
39,11.35E-2,0.,36.211E-2,0.,0., -1.0
40,12.E-2,0.,36.211E-2,0.,0.,-1.0
41,12.E-2,0.,0.
*NGEN, LINE = C, NSET = FPLANE
6, 26, 2, , 0. , 0. , 30.E-2 , 0., 1., 0
26, 32, 1, , 0. , 0. , 30.E-2
32, 38, 1, , 10.698E-2, 0. , 43.211E-2
*NCOPY, OLD SET = FPLANE, CHANGE NUMBER = 100,
SHIFT, MULTIPLE = 22
0., 0., 0.
0., 0., 0., 1., 0., 0., 8.18181818
*NODE, NSET = RIGID
10000, 12.E-2, 0. , 0.
*NODE, NSET = CAVITY
50000, 9.E-2 , 0. , 0.
**
** RUBBER SPRING
**
*ELEMENT, TYPE = S4R
2, 4, 6, 106, 104
26, 26, 27, 127, 126
*ELGEN, ELSET = SPRING1
2, 11, 2, 2, 22, 100, 100
26, 14, 1, 1, 22, 100, 100
*ELSET,ELSET=SPRING
SPRING1,SPRING2
*SHELL SECTION, MATERIAL = RUBBER, ELSET = SPRING
6.0E-3,
*HOURGLASS STIFFNESS
,,,100.
1-158
*MATERIAL, NAME = RUBBER

*HYPERELASTIC, POLYNOMIAL, N = 1
3.2E6, 0.8E6
*REBAR, ELEMENT = SHELL, MATERIAL = STEEL,
GEOMETRY = SKEW,
NAME = PLSREBAR
SPRING, 1.E-6, 3.5E-3, 0.0, 18.0
*REBAR, ELEMENT = SHELL, MATERIAL = STEEL,
GEOMETRY = SKEW,
NAME = MNSREBAR
SPRING, 1.E-6, 3.5E-3, 0.0, -18.0
*MATERIAL, NAME = STEEL
*ELASTIC
2.1E11, 0.3
**
** CONTACT DEFINITION
**
*SURFACE, NAME = ASURF
SPRING2,SNEG
*RIGID BODY,ANALYTICAL SURFACE=RSURF,
REF NODE=10000
*SURFACE,TYPE=REVOLUTION,NAME=RSURF,
FILLET RADIUS=0.5E-2
0., 0., 0., 1., 0., 0.
START, 47.E-2 , 5.198E-2
LINE, 41.711E-2, 5.198E-2
CIRCL, 36.211E-2, 10.698E-2, 41.711E-2, 10.698E-2
LINE, 36.211E-2, 11.9E-2
*CONTACT PAIR, INTERACTION = SMOOTH
ASURF, RSURF
*SURFACE INTERACTION, NAME = SMOOTH
**
** FLUID ELEMENTS
**
*ELEMENT, TYPE = F3D3, ELSET = FLUID
6001, 1, 4, 104
6040, 40, 41, 140
*ELGEN, ELSET = FLUID
6001, 22, 100, 100
6040, 22, 100, 100
6002, 4, 6, 106, 104
6026, 26, 27, 127, 126
1-159

6002, 11, 2, 2, 22, 100, 100
6026, 14, 1, 1, 22, 100, 100
*PHYSICAL CONSTANTS, ABSOLUTE ZERO=-273.16
*FLUID PROPERTY, REF NODE=50000, TYPE=PNEUMATIC,
ELSET=FLUID,
AMBIENT=101.36E+3
*FLUID DENSITY, PRESSURE=0.0, TEMPERATURE=27.0
1.774,
**
** KINEMATIC CONSTRAINTS
**
*NSET, NSET = N1M1, GENERATE
101, 2201, 100
*NSET, NSET = N4, GENERATE
4, 2204, 100
6, 2206, 100
40, 2240, 100
41, 2241, 100
*NSET, NSET = YPLN1, GENERATE
4, 26, 2
26, 40, 1
2204, 2226, 2
2226, 2240, 1
*MPC
TIE, N1M1, 1
TIE, N41, 10000
BEAM, N40, 10000
*BOUNDARY
1,1,3
N4,1,6
N6,2,3
YPLN1, 2
YPLN1, 4
YPLN1, 6
YPLN2, 2
YPLN2, 4
YPLN2, 6
RIGID,1,6
1-160
**
** STEP 1 - INFLATION STEP (5 ATMOSPHERES)
**
*STEP, NLGEOM, INC=20
*STATIC
0.02,1.0
*MONITOR, NODE=26, DOF=3
*PRINT, CONTACT = YES
*BOUNDARY
CAVITY,8,8, 5.066E5
*EL PRINT, FREQUENCY=99
*NODE PRINT, NSET=CAVITY, FREQUENCY=99
PCAV, CVOL
*NODE PRINT, NSET=RIGID, FREQUENCY=99
U,
RF,
*NODE FILE, NSET=CAVITY
PCAV, CVOL
*OUTPUT,FIELD
*NODE OUTPUT,NSET=CAVITY
PCAV,CVOL
*NODE FILE, NSET=RIGID
U,
RF,
CF,
*OUTPUT,FIELD
*NODE OUTPUT,NSET=RIGID
U,
RF,
CF,
*END STEP
**
** STEP 2 - SEAL CAVITY, RELEASE GRIP ON
** RIGID SURFACE, AND LOAD RIGID SURFACE
**
*STATIC
1.0,1.0
*BOUNDARY, OP=NEW
1,1,3
N4,1,6
N6,2,3
1-161
YPLN1, 2
YPLN1, 4
YPLN1, 6
YPLN2, 2
YPLN2, 4
YPLN2, 6
RIGID,2,6
*CLOAD
RIGID,1,-75.E3
*END STEP
**
** STEP 3 - PERTURBATION STEP 1
** UNIT AXIAL DISPLACEMENT
**
*STEP, PERTURBATION
*STATIC
*BOUNDARY
RIGID,1,1, 1.E-3
*NODE PRINT, NSET=CAVITY
PCAV, CVOL
*NODE PRINT, NSET=RIGID
U,
RF,
*END STEP
**
** UNIT LATERAL DISPLACEMENT
**
*STEP, PERTURBATION
*STATIC
*BOUNDARY
RIGID,3,3, 1.E-3
PCAV, CVOL
U,
RF,
*END STEP
**
** UNIT ROTATION
**
*STEP, PERTURBATION
1-162
*STATIC
*BOUNDARY
RIGID,5,5, 1.E-3
PCAV, CVOL
U,
RF,
*END STEP
**
** STEP 6 - NONLINEAR LATERAL
** LOAD DEFLECTION CURVE
**
*STATIC
0.1, 1.0
*CONTROLS,PARAMETERS=FIELD
,1.0
*BOUNDARY
RIGID,3,3, 2.E-2
*EL PRINT,REBAR
S,E,RBANG,RBROT
*END STEP
1-163
Listing 1.1.9-2
*HEADING
AIRSPRING -- AXISYMMETRIC MODEL
*RESTART,WRITE
*NODE
1, 0. , 2.E-2
4, 20.E-2 , 2.E-2
6, 20.E-2 , 0.
26, 40.E-2 , 0.
32, 37.771E-2, 6.293E-2
38, 36.211E-2, 10.698E-2
39, 36.211E-2, 11.35E-2
40, 36.211E-2, 12.E-2
500, 0. , 9.E-2
1000, 0. , 12.E-2
*NSET, NSET = RIGID
1000,
*NSET, NSET = CAVITY
500,
*NGEN, LINE = C
6, 26, 2, , 30.E-2 , 0. , 0., 0., 0., 1.
26, 32, 1, , 30.E-2 , 0. , 0.
32, 38, 1, , 43.211E-2, 10.698E-2, 0.
**
** RUBBER SPRING
**
*ELEMENT, TYPE = SAX1
2, 4, 6
26, 26, 27
*ELGEN, ELSET = SPRING
2, 11, 2, 2
26, 14, 1, 1
*SHELL SECTION, COMPOSITE, ELSET = SPRING
2.5E-3, 3, RUBBER
1.0E-3, 3, STEEL
2.5E-3, 3, RUBBER
3.2E6, 0.8E6
*ELASTIC, TYPE=LAMINA
2.48E9, 2.86E7, 9.105, 8.0E6, 8.0E6, 8.0E6
1-164
**
** CONTACT ELEMENTS
**
*ELSET,ELSET=ECON,GENERATE
26,39,1
*SURFACE,TYPE=SEGMENTS,FILLET RADIUS=0.5E-2,
NAME=BSURF
START, 41.711E-2, 5.198E-2
CIRCL, 36.211E-2, 10.698E-2, 41.711E-2, 10.698E-2
LINE, 36.211E-2, 12.E-2
*SURFACE,NAME=ASURF
ECON,SPOS
*CONTACT PAIR,INTERACTION=SMOOTH
ASURF,BSURF
**
** FLUID ELEMENTS
**
*ELEMENT, TYPE = FAX2, ELSET = FLUID
201, 1, 4
202, 4, 6
226, 26, 27
299, 40, 1000
202, 11, 2, 2
226, 14, 1, 1
ELSET=FLUID,
AMBIENT=101.36E+3
1.774,
**
*MPC
BEAM, 40, 1000
*BOUNDARY
1,1,3
4,1,6
6,1,1
RIGID,1,6
**
1-165
**
*STEP, NLGEOM
*STATIC
0.1,1.0
*PRINT, CONTACT = YES
*BOUNDARY
CAVITY,8,8, 5.066E5
*OUTPUT,FIELD
*CONTACT OUTPUT,VARIABLE=PRESELECT,SLAVE=ASURF
*NODE PRINT, NSET=CAVITY, FREQUENCY=99
PCAV, CVOL
*NODE PRINT, NSET=RIGID, FREQUENCY=99
U, RF
PCAV, CVOL
*OUTPUT,FIELD
*NODE OUTPUT,NSET=CAVITY
PCAV,CVOL
U, RF
CF,
*OUTPUT,FIELD
*NODE OUTPUT,NSET=RIGID
U,RF
CF,
*END STEP
**
** STEP 2 - SEAL CAVITY, RELEASE GRIP ON
** RIGID SURFACE,
** AND LOAD RIGID SURFACE
**
*STEP, NLGEOM
*STATIC
1.0, 1.0
*BOUNDARY, OP=NEW
1,1,3
4,1,6
6,1,1
1-166
RIGID,1
RIGID,6
*CLOAD
RIGID,2,-150.E3
*END STEP
**
** UNIT AXIAL DISPLACEMENT WITH CAVITY
** PRESS FREE
**
*STEP, PERTURBATION
*STATIC
*BOUNDARY
RIGID, 2, 2, 1.E-3
PCAV, CVOL
U, RF
*END STEP
**
** UNIT LOAD WITH CAVITY PRESS FIXED
**
*STEP, PERTURBATION
*STATIC
*BOUNDARY
CAVITY, 8, 8, 0
RIGID, 2, 2, 1.E-3
PCAV, CVOL
U, RF
*END STEP
**
** STEP 5 - NONLINEAR LOAD DEFLECTION CURVE
**
*STATIC
0.1,1.0
*CLOAD
RIGID,2,-240.E3
*END STEP
1-167
Listing 1.1.9-3
*HEADING
AIRSPRING -- 180 DEG MODEL
*RESTART,WRITE,NUMBER INTERVAL=10
*NODE, NSET = FPLANE
1,2.E-2,0.,0.
4,2.E-2,0.,20.E-2,0.,0.,1.0
6,0.,0.,20.E-2,0.,0.,1.0
26,0.,0.,40.E-2,0.,0.,-1.0
32,6.293E-2,0.,37.771E-2,-0.58431,0.,-0.81153
38,10.698E-2,0.,36.211E-2,0.,0.,-1.0
39,11.35E-2,0.,36.211E-2,0.,0.,-1.0
40,12.E-2,0.,36.211E-2,0.,0.,-1.0
41,12.E-2,0.,0.
*NGEN, LINE = C, NSET = FPLANE
6, 26, 2, , 0., 0., 30.E-2, 0., 1., 0.
26, 32, 1, , 0., 0., 30.E-2
32, 38, 1, , 10.698E-2, 0., 43.211E-2
*NCOPY,OLDSET=FPLANE,CHANGE NUMBER=100,SHIFT,
MULTIPLE=22,NEWSET=NALL
0., 0., 0.
0., 0., 0., 1., 0., 0., 8.18181818
*NODE, NSET = RIGID
11000, 12.E-2, 0. , 0.
*NODE, NSET = CAVITY
50000, 9.E-2 , 0. , 0.
**
** RUBBER SPRING
**
*ELEMENT, TYPE = S4R
2, 4, 6, 106, 104
26, 26, 27, 127, 126
2, 11, 2, 2, 22, 100, 100
26, 14, 1, 1, 22, 100, 100
*ELSET,ELSET=SPRING
SPRING1,SPRING2
*SHELL SECTION, MATERIAL = RUBBER, ELSET = SPRING
6.0E-3,
1-168
3.2E6, 0.8E6
*DENSITY
1000.,
*REBAR,ELEMENT=SHELL,MATERIAL=STEEL,GEOMETRY=SKEW,
NAME=PLSREBAR
SPRING, 1.E-6, 3.5E-3, 0., 18.0
*REBAR,ELEMENT=SHELL,MATERIAL=STEEL,GEOMETRY=SKEW,
NAME=MNSREBAR
SPRING, 1.E-6, 3.5E-3, 0., -18.0
*ELASTIC
2.1E11, 0.3
*DENSITY
7.8E3,
**
** FLUID ELEMENTS
**
6001, 1, 4, 104
6040, 40, 41, 140
6001, 22, 100, 100
6040, 22, 100, 100
6002, 4, 6, 106, 104
6026, 26, 27, 127, 126
6002, 11, 2, 2, 22, 100, 100
6026, 14, 1, 1, 22, 100, 100
ELSET=FLUID,
AMBIENT=101.36E+3
1.774,
**
** RIGID BODY/SURFACE
**
*NODE, NSET = RIGID
11000, 12.E-2, 0., 0.
*NODE,NSET=RIGNODE
11001, 12.E-2, 0., 36.511E-2
11021, 6.5E-2, 0., 42.011E-2
1-169
12000, 12.E-2, 0., 42.011E-2

*NGEN,NSET=RIGNODE,LINE=C
11001, 11021, 1, 12000
*NCOPY,OLDSET=RIGNODE,CHANGE NUMBER=100,SHIFT,
MULTIPLE=22,NEWSET=NALL
0., 0., 0.
0., 0., 0., 1., 0., 0., 8.18181818
*ELEMENT,TYPE=R3D4,ELSET=RIGELEM
11001, 11001,11101,11102,11002
*ELGEN,ELSET=RIGELEM
11001, 22,100,100, 20,1,1
*ELEMENT,TYPE=MASS,ELSET=MASS_RIGID
11000, 11000
*MASS,ELSET=MASS_RIGID
.5,
**
** KINEMATIC CONSTRAINTS
**
*NSET, NSET = N1M1, GENERATE
101, 2201, 100
4, 2204, 100
6, 2206, 100
40, 2240, 100
41, 2241, 100
4, 26, 2
26, 39, 1
2204, 2226, 2
2226, 2239, 1
*BOUNDARY
1,1,6
N4,1,6
N6,2,3
YPLN1, 2
YPLN1, 4
YPLN1, 6
YPLN2, 2
YPLN2, 4
1-170
YPLN2, 6
RIGID,1,6
**
**
*AMPLITUDE,NAME=RAMP,DEFINITION=SMOOTH STEP
0.,0., 1.E-2,1.0
*AMPLITUDE,NAME=RAMPSMOOTH,DEFINITION=SMOOTH STEP
0.,0., 2.E-2,1.0
*SURFACE,TYPE=ELEMENT, NAME = ASURF
ASURF,SNEG
*SURFACE,TYPE=ELEMENT, NAME = RSURF
RIGELEM,SPOS
*RIGID BODY,REFNODE=1,PINNSET=N1M1
*RIGID BODY,REFNODE=11000,ELSET=RIGELEM,
TIENSET=RIG
*STEP
*DYNAMIC,EXPLICIT
,1.E-2
*BOUNDARY,AMPLITUDE=RAMP
CAVITY,8,8, 5.066E5
**
**
*NSET,NSET=RIG
N40,N41
*ELSET,ELSET=ASURF,GEN
26,2126,100
27,2127,100
28,2128,100
29,2129,100
30,2130,100
31,2131,100
32,2132,100
33,2133,100
34,2134,100
35,2135,100
36,2136,100
37,2137,100
38,2138,100
*CONTACT PAIR, INTERACTION = SMOOTH
ASURF, RSURF
*SURFACE INTERACTION, NAME = SMOOTH
1-171
**
** OUTPUT REQUESTS
**
*FILE OUTPUT,NUMBER INTERVAL=2
PCAV, CVOL
U,
RF,
*HISTORY OUTPUT,TIME=1.E-5
*ENERGY HISTORY
ALLKE,ALLIE,ALLAE,ALLSE,ETOTAL,DT,DTFB,ALLVD
*ELSET,ELSET=TOP
39,2139,1039
*EL HISTORY,ELSET=TOP
S,E
*EL HISTORY,ELSET=TOP,REBAR=PLSREBAR
S,E,RBANG,RBROT
*NODE HISTORY,NSET=CAVITY
PCAV,CVOL
*NODE HISTORY,NSET=RIG
U,RF
*NODE HISTORY,NSET=RIGID
U,RF
*END STEP
**
** STEP 2 - SEAL CAVITY, LOAD FURTHER
**
*STEP
*DYNAMIC,EXPLICIT
,2.E-2
*BOUNDARY, OP=NEW,AMPLITUDE=RAMPSMOOTH
1,1,6
N4,1,6
N6,2,3
YPLN1, 2
YPLN1, 4
YPLN1, 6
YPLN2, 2
YPLN2, 4
YPLN2, 6
RIGID,1,2
RIGID,4,6
1-172
RIGID,3,3, 2.E-2
*END STEP
1-173
Listing 1.1.9-4
*HEADING
AIRSPRING -- AXISYMMETRIC MODEL
*NODE
1, 0. , 2.E-2
4, 20.E-2 , 2.E-2
6, 20.E-2 , 0.
26, 40.E-2 , 0.
32, 37.771E-2, 6.293E-2
38, 36.211E-2, 10.698E-2
39, 36.211E-2, 11.35E-2
40, 36.211E-2, 12.E-2
500, 0. , 9.E-2
1000, 0. , 12.E-2
*NSET, NSET = CAVITY
500,
*NGEN, LINE = C
6, 26, 2, , 30.E-2 , 0. , 0., 0., 0., 1.
26, 32, 1, , 30.E-2 , 0. , 0.
32, 38, 1, , 43.211E-2, 10.698E-2, 0.
**
** RUBBER SPRING
**
*ELEMENT, TYPE = SAX1
2, 4, 6
26, 26, 27
*ELGEN, ELSET = SPRING
2, 11, 2, 2
26, 14, 1, 1
*SHELL SECTION, COMPOSITE, ELSET = SPRING
2.5E-3, 3, RUBBER
1.0E-3, 3, STEEL
2.5E-3, 3, RUBBER
3.2E6, 0.8E6
*DENSITY
1000.,
*ELASTIC, TYPE=LAMINA
2.48E9, 2.86E7, 9.105, 8.0E6, 8.0E6, 8.0E6
*DENSITY
1-174
7.8E3,
**
** RIGID BODY/SURFACE
**
*NODE, NSET = RIGID
11000, 0. , 12.E-2
*NODE,NSET=RIGNODE
11001, 36.211E-2, 10.698E-2
11021, 41.711E-2, 5.198E-2
12000, 41.711E-2, 10.698E-2
*NGEN,NSET=RIGNODE,LINE=C
11001, 11021, 1, 12000
*ELEMENT,TYPE=RAX2,ELSET=RIGELEM
11001, 1000,40
11002, 40, 11002
11003, 11002, 11003
*ELGEN,ELSET=RIGELEM
11003, 19, 1, 1
*ELEMENT,TYPE=MASS,ELSET=MASS_RIGID
11000, 11000
*MASS,ELSET=MASS_RIGID
.5,
**
** FLUID ELEMENTS
**
*ELEMENT, TYPE = FAX2, ELSET = FLUID
201, 1, 4
202, 4, 6
226, 26, 27
299, 40, 1000
202, 11, 2, 2
226, 14, 1, 1
ELSET=FLUID,
AMBIENT=101.36E+3
1.774,
**
*BOUNDARY
1,1,3
4,1,6
1-175
6,1,1
RIGID,1,6
*NSET,NSET=RIG
40,1000
*AMPLITUDE,NAME=RAMP,DEFINITION=SMOOTHSTEP
0.,0., .05,1.0
**
**
*SURFACE,TYPE=ELEMENT,NAME=ASURF,NOTHICK
ECON,SPOS
*SURFACE,TYPE=ELEMENT,NAME=BSURF
RIGELEM,SNEG
*RIGID BODY,REFNODE=11000,ELSET=RIGELEM
*STEP
*DYNAMIC,EXPLICIT
,.05
*BOUNDARY,AMPLITUDE=RAMP
CAVITY,8,8, 5.066E5
**
**
26,38,1
ASURF,BSURF
**
** OUTPUT REQUESTS
**
*ELSET,ELSET=EOUT
39,
*FILE OUTPUT, NUMBER INTERVAL=1
PCAV, CVOL
U, RF
CF,
*ENERGY HISTORY
ALLSE,ALLKE,ALLAE,ALLIE,ETOTAL
*NODE HISTORY,NSET=CAVITY
1-176
PCAV,CVOL
*NODE HISTORY, NSET=RIGID
U, RF, CF
*EL HISTORY,ELSET=EOUT
S,E
*DLOAD
SPRING,VP,1E4
*END STEP
**
** STEP 2 - SEAL CAVITY (KEEP RIGID SURFACE FIXED)
** NONLINEAR LOAD DEFLECTION CURVE
**
*STEP
*DYNAMIC,EXPLICIT
,.05
*BOUNDARY, OP=NEW,AMPLITUDE=RAMP
1,1,3
4,1,6
6,1,1
RIGID,1
RIGID,3,6
RIGID,2,2,-.075
*END STEP
1.1.10 Shell-to-solid submodeling of a pipe joint

Submodeling is the technique of analyzing a local part of a model with a refined mesh, based on
interpolation of the solution from an initial, global model (usually with a coarser mesh) onto the nodes
on the appropriate parts of the boundary of the submodel. This local refinement procedure provides a
cost-effective approach to model enhancement. Shell-to-solid submodeling models a region with solid
elements, when the global model is made up of shell elements.
The purpose of this example is to demonstrate the shell-to-solid submodeling capability in ABAQUS.
Geometry and model

The joint between a pipe and a plate is analyzed. A pipe of radius 10 mm and thickness 0.75 mm is
attached to a plate that is 100 mm long, 50 mm wide, and 1 mm thick. The pipe-plate intersection has a
fillet radius of 1 mm. Taking advantage of the symmetry of the problem, only one-half of the assembly
is modeled. Both the pipe and the plate are assumed to be made up of aluminum with E =69 ´ 103
MPa and Poisson's ratio º = 0.3. The global model consists of S4R elements with the mesh layout as
shown in Figure 1.1.10-1. Since a shell model is used, the fillet radius is not taken into consideration.
1-177
In the submodel the joint and its vicinity are meshed using three-dimensional continuum elements
(C3D20R) with four layers through the thickness (see Figure 1.1.10-2). The solid model extends 10
mm along the pipe length and has a radius of 25 mm in the plane of the plate. The submodel accurately
models the fillet radius at the joint. Hence, the submodel capability makes it possible to calculate the
stress concentration in the fillet. The problem could be expanded by adding a ring of welded material
to simulate a welded joint (for this case the submodel would have to be meshed with new element
layers representing the welded material at the joint). The example could also be expanded by including
plastic material behavior in the submodel while using an elastic global model solution.
Loading
The pipe in the global model is subjected to concentrated loads acting in the 1-direction applied at the
nodes at the free end, representing a shear load on the pipe. The total value of all concentrated forces is
equal to 10 N.
Kinematic boundary conditions

In the global shell model the plate is clamped along all edges. In the solid submodel kinematic
conditions are interpolated from the global model at two surfaces of the submodel, one lying within the
pipe and the other within the plate. The default center zone size, equal to 10% of the maximum shell
thickness, is used. Thus, only one layer of driven nodes lies within the center zone, and only these
nodes have all three displacement components driven by the global solution. For the remaining driven
nodes only the displacement components parallel to the global model midsurface are driven from the
global model. Thus, a single row of nodes is transmitting the transverse shear forces from the shell
solution to the solid model.

The loading and boundary conditions are such that the pipe is subjected to bending. The end of the
pipe that is attached to the plate leads to deformation of the plate itself (see Figure 1.1.10-3). From a
design viewpoint the area of interest is the pipe-plate joint where the pipe is bending the plate. Hence,
this area is submodeled to gain better understanding of the deformation and stress state.
Figure 1.1.10-4 shows the contours of the out-of-plane displacement component in the plate and shows
the differences between the behavior in the models. As expected, the solid model is in good agreement
with the displacement of the shell model around the joint. The stress concentration in the fillet radius is
obtained with the solid model. The maximum Mises stress in this region is equal to 80.6 MPa, which is
51% more than the Mises stress in the shell model. Similarly, the maximum principal stress in the fillet
region in the solid model is 54% higher than the corresponding stress in the shell model.
Input files
shellsolidpipe_s4_global.inp
S4 global model.
shellsolidpipe_c3d20rsub_s4.inp
1-178
C3D20R submodel, which uses the S4 global model.

shellsolidpipe_s4r_global.inp
S4R global model.
shellsolidpipe_c3d20rsub_s4r.inp
Key input data for the C3D20R submodel, which uses the S4R global model.
shellsolidpipe_c3d20r_mesh.inp
Remainder of the input data for the C3D20R submodel.
shellsolidpipe_node.inp
Node definitions for the S4R and S4 global models.
shellsolidpipe_element.inp
Element definitions for the S4R and S4 global models.
Figures
Figure 1.1.10-1 Global shell model of pipe-plate structure.
Figure 1.1.10-2 Magnified solid submodel of the pipe-plate joint.
1-179
Figure 1.1.10-3 Solid submodel overlaid on the shell model in the deformed state, using
magnification factor of 20.
Figure 1.1.10-4 Comparison of out-of-plane displacement in plate for both models.
Sample listings
1-180
Listing 1.1.10-1
*HEADING
Global Shell Model
**
**RESTART,WRITE
*NODE,INPUT=shellsolidpipe_node.inp
*ELEMENT, TYPE=S4R,
INPUT=shellsolidpipe_element.inp
**
** plate
**
*ELSET,ELSET=PLATE,GENERATE
1,96,1
193,288,1
** pipe
**
*ELSET,ELSET=PIPE,GENERATE
97,192,1
289,384,1
*SHELL SECTION,ELSET=PLATE,MATERIAL=MAT1
0.001,
*SHELL SECTION,ELSET=PIPE,MATERIAL=MAT1
0.00075,
**
** Basic steel properties
*MATERIAL,NAME=MAT1
*ELASTIC
69.0E9,0.3
**
** built_in
**
*NSET, NSET=BUILT_IN
1, 10, 19, 28,
37, 46, 55, 64,
73, 82, 91, 100,
109, 222, 223, 224,
225, 226, 227, 228,
285, 286, 287, 288,
289,
**
** ysymm
**
1-181
*NSET, NSET=YSYMM
1, 2, 3, 4,
5, 6, 7, 8,
9, 118, 119, 120,
121, 122, 123, 124,
222, 229, 236, 243,
250, 257, 264, 271,
278, 330, 337, 344,
351, 358, 365, 372
**
** load
**
*NSET, NSET=LOAD
125, 133, 141, 149,
157, 165, 173, 181,
189, 197, 205, 213,
221, 379, 380, 381,
382, 383, 384, 385,
421, 422, 423, 424,
425,
*NSET, NSET=force
133, 141, 149, 157,
165, 173, 181, 189,
197, 205, 213, 221,
380, 381, 382, 383,
384, 385, 421, 422,
423, 424, 425
*BOUNDARY
BUILT_IN,ENCASTRE
YSYMM,YSYMM
**
** step 1,Default
**
*STEP
Total load of 10.0 N in the 1-direction
*STATIC
*CLOAD
force,1,0.416667
125,1,0.208333
379,1,0.208333
*NODE FILE
U,
*END STEP
1-182
Listing 1.1.10-2
*HEADING
Solid submodel for global model.
ABAQUS job create
**
**RESTART,WRITE,FREQ=999
***
**Node, element and node set definitions
**are written into the file
**shellsolidpipe_c3d20r_mesh.inp
***
*INCLUDE,INPUT=shellsolidpipe_c3d20r_mesh.inp
*NSET,NSET=CUT
NPIPE,NPLATE
*MATERIAL,NAME=ALUMINUM
*ELASTIC
69.E9,0.3
*SOLID SECTION,ELSET=EALL,MATERIAL=ALUMINUM
**
*SUBMODEL,SHELLTOSOLID,SHELLTHICKNESS=0.001
CUT,
*BOUNDARY
NYSYMM,YSYMM
*STEP
*STATIC
*BOUNDARY,SUBMODEL,STEP=1
CUT,
*NODE FILE,NSET=CUT
U,
*END STEP
1.1.11 Stress-free element reactivation

This example demonstrates element reactivation for problems where new elements are to be added in a
stress-free state. Typical examples include the construction of a gravity dam, in which unstressed
layers of material are added to a mesh that has already deformed under geostatic load, or a tunnel in
which a concrete or steel support liner is installed. The *MODEL CHANGE, ADD option (`Èlement
and contact pair removal and reactivation, '' Section 7.4.2 of the ABAQUS/Standard User's Manual)
1-183
provides for this type of application directly because the strain in newly added elements corresponds to
the deformation of the mesh since the reactivation.
Verification of the *MODEL CHANGE capability is provided in ``Model change,'' Section 3.8 of the
ABAQUS Verification Manual.
Problem description
The example considers the installation of a concrete liner to support a circular tunnel. Practical
geotechnical problems usually involve a complex sequence of construction steps. The construction
details determine the appropriate analysis method to represent these steps accurately. Such details have
been avoided here for the sake of simplifying the illustration.
The tunnel is assumed to be excavated in clay, with a Young's modulus of 200 MPa and a Poisson's
ratio of 0.2 (see Figure 1.1.11-1). The diameter of the tunnel is 8 m, and the tunnel is excavated 20 m
below ground surface. The material surrounding the excavation is discretized with first-order 4-node
plane strain elements (element type CPE4). The infinite extent of the soil is represented by a
30-m-wide mesh that extends from the surface to a depth of 50 m below the surface. The left-hand
boundary represents a vertical symmetry axis. Far-field conditions on the bottom and right-hand-side
boundaries are modeled by infinite elements (element type CINPE4). No mesh convergence studies
have been performed to establish if these boundary conditions are placed far enough away from the
excavation.
An initial stress field due to gravitational and tectonic forces exists through the depth of the soil. It is
assumed that this stress varies linearly with depth and that the ratio between the horizontal and vertical
stress components is 0.5. The self weight of the clay is 20.0 kN/m 3.
The excavation of the tunnel material is accomplished by applying the forces that are required to
maintain equilibrium with the initial stress state in the surrounding material as loads on the perimeter
of the tunnel. These loads are then reduced to zero to simulate the excavation. The three-dimensional
effect of face advancement during excavation is taken into account by relaxing the forces gradually
over several steps. The liner is installed after 40% relaxation of the loads. Further deformation
continues to occur as the face of the excavation advances. This ongoing deformation loads the liner.
In the first input file the 150-mm-thick liner is discretized with one layer of incompatible mode
elements (element type CPE4I). These elements are recommended in regions where bending response
must be modeled accurately. In the second input file beam elements are used to discretize the liner.
The liner is attached rigidly to the tunnel. The concrete is assumed to have cured to a strength
represented by the elastic properties shown in Figure 1.1.11-1 by the time the liner is loaded. The liner
is not shown in this diagram.
It is expected that an overburden load representing the weight of traffic and buildings exists after the
liner is installed.
Analysis method
The excavation and installation of the liner is modeled in four analysis steps. In the first step the initial
stress state is applied and the liner elements are removed using the *MODEL CHANGE, REMOVE
1-184
option. Concentrated loads that are in equilibrium with the initial stress field are applied on the
perimeter of the tunnel. These forces were obtained from an independent analysis where the
displacements on the tunnel perimeter were constrained. The reaction forces at the constrained nodes
are the loads applied here. The second step begins the tunnel excavation by reducing the concentrated
loads on the tunnel surface. The loads are reduced by 40% in this step before the liner is installed in
the third step using the *MODEL CHANGE, ADD option. No deformation takes place in the soil or
liner during the third step. In the fourth step the surface load is applied, and the excavation is
completed by removing the remainder of the load on the tunnel perimeter.
In problems involving geometric nonlinearities with finite deformation, it is important to recognize
that element reactivation occurs in the configuration at the start of the reactivation step. If the
NLGEOM parameter were used in this problem, the thickness of the liner, when modeled with the
continuum elements, would have a value at reactivation that would be different from its original value.
This result would happen because the outside nodes (the nodes on the tunnel/liner interface) displace
with the mesh, whereas the inside nodes remain at their current locations since liner elements are
inactive initially. This effect is not relevant in this problem because geometric nonlinearities are not
included. However, it may be significant for problems involving finite deformation, and it may lead to
convergence problems in cases where elements are severely distorted upon reactivation. This problem
would not occur in the model with beam elements because they have only one node through the
thickness. In the model where the liner is modeled with continuum elements, the problem can be
eliminated if the inner nodes are allowed to follow the outer nodes prior to reactivation, which can be
accomplished by applying displacement boundary conditions on the inner nodes. Alternatively, the
liner can be overlaid with (elastic) elements of very low stiffness. These elements use the same nodes
as the liner but are so compliant that their effect on the analysis is negligible when the liner is present.
They remain active throughout the analysis and ensure that the inner nodes follow the outer nodes,
thereby preserving the liner thickness.

Figure 1.1.11-2 shows the stress state at a material point in the liner. The figure clearly indicates that
the liner remains unstressed until reactivated.
Figure 1.1.11-3 compares the axial stress obtained from the CPE4I and beam elements at the top and
bottom of the liner section. A cylindrical *ORIENTATION (`Òrientations,'' Section 2.2.4 of the
ABAQUS/Standard User's Manual) is used to orient the liner stresses in the continuum element model
along the beam axis so that these stresses can be compared directly with the results of the beam
element model. The small difference between the results can be attributed to the element type used in
the discretization of the liner: the beam element model uses a plane stress condition, and the
continuum element model uses a plane strain condition.
Input files
modelchangedemo_continuum.inp
*MODEL CHANGE with continuum elements.
modelchangedemo_beam.inp
1-185
*MODEL CHANGE with beam elements.

modelchangedemo_node.inp
Nodal coordinates for the soil.
modelchangedemo_element.inp
Element definitions for the soil.
Figures
Figure 1.1.11-1 Geometry and finite element discretization.
Figure 1.1.11-2 Liner stress during analysis history.
1-186
Figure 1.1.11-3 Axial stress along beam inside and outside.
Sample listings
1-187
Listing 1.1.11-1
*HEADING
STRESS FREE INCLUSION OF TUNNEL LINER
(USING *MODEL CHANGE)
Units : N, m
*RESTART,WRITE
**
*NODE,NSET=SOIL,INPUT=modelchangedemo_node.inp
*NSET,NSET=TUNNEL,UNSORTED
100, 101, 102, 103, 104, 105, 106,
2105, 2104, 2103, 2102, 2101, 2100
*NODE
4000, 0.00, -16.00
4012, 0.00, -24.00
4100, 0.00, -16.15
4112, 0.00, -23.85
*NGEN,LINE=C
4000, 4012, 1, ,0.0, -20.0, 0.0, 0.0, 0.0, -1.0
4100, 4112, 1, ,0.0, -20.0, 0.0, 0.0, 0.0, -1.0
*NSET,NSET=LINER,GENERATE
4000, 4012
*NSET,NSET=RHS,GENERATE
8, 3108, 10
3100, 3108
*NSET,NSET=XSYMM,GENERATE
100, 150, 10
200, 1000, 100
2100, 2150, 10
2200, 3200, 100
*NSET,NSET=XSYMM
4000, 4012, 4100, 4112
*NODE,NSET=NINF
209, 60.00, -20.00
309, 60.00, -17.77
409, 60.00, -14.98
509, 60.00, -11.50
609, 60.00, -10.10
709, 60.00, -8.347
809, 60.00, -6.158
909, 60.00, -3.420
1009, 60.00, 0.000
1-188
2309, 60.00, -22.03

2409, 60.00, -25.02
2509, 60.00, -28.50
2609, 60.00, -29.90
2709, 60.00, -31.65
2809, 60.00, -33.84
2909, 60.00, -36.58
3009, 60.00, -40.00
3109, 60.00, -76.00
3200, 0.000, -76.00
3201, 2.245, -76.00
3202, 5.052, -76.00
3203, 8.560, -76.00
3204, 11.17, -76.00
3205, 14.44, -76.00
3206, 18.52, -76.00
3207, 23.62, -76.00
**
*ELEMENT,TYPE=CPE4,ELSET=SOIL,
INPUT=modelchangedemo_element.inp
*ELEMENT,ELSET=LINER,TYPE=CPE4I
4000, 4100, 4101, 4001, 4000
*ELGEN,ELSET=LINER
4000, 12
*ELSET,ELSET=SURFACE,GENERATE
900 , 907 , 1
*ELEMENT,TYPE=CINPE4,ELSET=ELINF
3100,3101,3100,3200,3201
3101,3102,3101,3201,3202
3102,3103,3102,3202,3203
3103,3104,3103,3203,3204
3104,3105,3104,3204,3205
3105,3106,3105,3205,3206
3106,3107,3106,3206,3207
3107,3108,3107,3207,3109
3108,3008,3108,3109,3009
2908,2908,3008,3009,2909
2808,2808,2908,2909,2809
2708,2708,2808,2809,2709
2608,2608,2708,2709,2609
2508,2508,2608,2609,2509
2408,2408,2508,2509,2409
2308,2308,2408,2409,2309
1-189
2208,208,2308,2309,209
208,308,208,209,309
308,408,308,309,409
408,508,408,409,509
508,608,508,509,609
608,708,608,609,709
708,808,708,709,809
808,908,808,809,909
908,1008,908,909,1009
*ELSET,ELSET=EALL
SOIL,ELINF
**
*ORIENTATION,NAME=OR,SYSTEM=CYLINDRICAL
0.0, -20.0, 0.0, 0.0, -20.0, 10.0
3, 0.0
*SOLID SECTION,MATERIAL=CONCRETE,ELSET=LINER,
ORIENTATION=OR
*SOLID SECTION,MATERIAL=CLAY,ELSET=SOIL
*SOLID SECTION,MATERIAL=CLAY,ELSET=ELINF
*MATERIAL,NAME=CLAY
*ELASTIC,TYPE=ISOTROPIC
0.2E9, 0.2
*MATERIAL,NAME=CONCRETE
19.0E9 , 0.2
**
*MPC
TIE, LINER, TUNNEL
EALL, 0.0, 0.0, -1.52E6, -76.00, 0.5
*AMPLITUDE,NAME=RELAX,TIME=TOTAL TIME
0.0, 1.0, 1.0, 1.0, 2.0, 0.6, 3.0, 0.6,
4.0, 0.0
** --------------------------------------
*STEP
step 1: add initial stress state & remove liner
*STATIC
*DLOAD
SOIL, BY, -20.0E3
*MODEL CHANGE,REMOVE
LINER,
*CLOAD,AMPLITUDE=RELAX
100, 1, 5.4086E+04
1-190
101, 1, 4.3918E+04
102, 1, 8.6901E+04
103, 1, 1.2732E+05
104, 1, 1.6185E+05
105, 1, 1.8949E+05
106, 1, 2.0701E+05
2100, 1, 8.2710E+04
2101, 1, 6.3287E+04
2102, 1, 1.2031E+05
2103, 1, 1.6549E+05
2104, 1, 1.9676E+05
2105, 1, 2.1052E+05
100, 2, 1.6652E+05
101, 2, 3.2459E+05
102, 2, 2.9838E+05
103, 2, 2.5160E+05
104, 2, 1.8487E+05
105, 2, 9.9587E+04
106, 2, -2142.
2100, 2, -2.4756E+05
2101, 2, -4.7534E+05
2102, 2, -4.1880E+05
2103, 2, -3.3399E+05
2104, 2, -2.2922E+05
2105, 2, -1.1476E+05
*BOUNDARY
XSYMM, 1
*NODE PRINT,FREQUENCY=100,NSET=TUNNEL
U, RF
*EL PRINT,FREQUENCY=100,ELSET=LINER,
POSITION=AVERAGED AT NODES
S,
*END STEP
** --------------------------------------
*STEP,INC=100
step 2: relax tunnel stress 40 %
*STATIC
*END STEP
** --------------------------------------
*STEP
step 3: add liner stress free
*STATIC
*MODEL CHANGE,ADD
1-191
LINER,
*END STEP
** --------------------------------------
*STEP,INC=100
step 4: relax tunnel stress to zero &
apply surface load
*STATIC
*DLOAD
SURFACE, P3, 50.0E3
*EL FILE,FREQUENCY=100,ELSET=LINER,
S,
*END STEP
1-192
Listing 1.1.11-2
*HEADING
STRESS FREE INCLUSION OF TUNNEL LINER
(USING *MODEL CHANGE)
Units : N, m
*RESTART,WRITE
**
*NODE,NSET=SOIL,INPUT=modelchangedemo_node.inp
*NSET,NSET=TUN_IX,UNSORTED
101, 102, 103, 104, 105, 106,
2105, 2104, 2103, 2102, 2101
*NSET,NSET=TUN_IY,UNSORTED
100, 101, 102, 103, 104, 105,
106, 2105, 2104, 2103, 2102, 2101, 2100
*NSET,NSET=TUNNEL,GENERATE
100, 106
2100, 2105
*NODE
4000, 0.00, -16.0
4012, 0.00, -24.0
*NGEN,NSET=LINER,LINE=C
4000, 4012, 1, ,0.0, -20.0, 0.0, 0.0, 0.0, -1.0
*NSET,NSET=LIN_JX,GENERATE
4001, 4011
*NSET,NSET=LIN_JY,GENERATE
4000, 4012
*NSET,NSET=RHS,GENERATE
8, 3108, 10
3100, 3108
*NSET,NSET=XSYMM,GENERATE
100, 150, 10
200, 1000, 100
2100, 2150, 10
2200, 3200, 100
*NSET,NSET=XSYMM
4000, 4012, 4100, 4112
*NODE,NSET=NINF
209, 60.00, -20.00
309, 60.00, -17.77
409, 60.00, -14.98
509, 60.00, -11.50
1-193
609, 60.00, -10.10

709, 60.00, -8.347
809, 60.00, -6.158
909, 60.00, -3.420
1009, 60.00, 0.000
2309, 60.00, -22.03
2409, 60.00, -25.02
2509, 60.00, -28.50
2609, 60.00, -29.90
2709, 60.00, -31.65
2809, 60.00, -33.84
2909, 60.00, -36.58
3009, 60.00, -40.00
3109, 60.00, -76.00
3200, 0.000, -76.00
3201, 2.245, -76.00
3202, 5.052, -76.00
3203, 8.560, -76.00
3204, 11.17, -76.00
3205, 14.44, -76.00
3206, 18.52, -76.00
3207, 23.62, -76.00
**
*ELEMENT,TYPE=CPE4,ELSET=SOIL,
INPUT=modelchangedemo_element.inp
*ELEMENT,TYPE=B21
4000, 4000, 4001
*ELGEN,ELSET=LINER
4000, 12
*ELSET,ELSET=SURFACE,GENERATE
900 , 907 , 1
*ELEMENT,TYPE=CINPE4,ELSET=ELINF
3100,3101,3100,3200,3201
3101,3102,3101,3201,3202
3102,3103,3102,3202,3203
3103,3104,3103,3203,3204
3104,3105,3104,3204,3205
3105,3106,3105,3205,3206
3106,3107,3106,3206,3207
3107,3108,3107,3207,3109
3108,3008,3108,3109,3009
2908,2908,3008,3009,2909
2808,2808,2908,2909,2809
1-194
2708,2708,2808,2809,2709
2608,2608,2708,2709,2609
2508,2508,2608,2609,2509
2408,2408,2508,2509,2409
2308,2308,2408,2409,2309
2208,208,2308,2309,209
208,308,208,209,309
308,408,308,309,409
408,508,408,409,509
508,608,508,509,609
608,708,608,609,709
708,808,708,709,809
808,908,808,809,909
908,1008,908,909,1009
*ELSET,ELSET=EALL
SOIL,ELINF
**
*BEAM SECTION,MATERIAL=CONCRETE,SECTION=RECT,
ELSET=LINER
1.0 , 0.15
0.0 , 0.0 , -1.0
*SOLID SECTION,MATERIAL=CLAY,ELSET=SOIL
*SOLID SECTION,MATERIAL=CLAY,ELSET=ELINF
*MATERIAL,NAME=CLAY
0.2E9, 0.2
*MATERIAL,NAME=CONCRETE
19.0E9 , 0.2
**
*EQUATION
2,
TUN_IX, 1, 1.0, LIN_JX, 1, -1.0
2,
TUN_IY, 2, 1.0, LIN_JY, 2, -1.0
EALL, 0.0, 0.0, -1.52E6, -76.00, 0.5
*AMPLITUDE,NAME=RELAX,TIME=TOTAL TIME
0.0, 1.0, 1.0, 1.0, 2.0, 0.6, 3.0, 0.6,
4.0, 0.0
** --------------------------------------
*STEP
step 1: add initial stress state
1-195
*STATIC
*DLOAD
SOIL, BY, -20.0E3
*MODEL CHANGE,REMOVE
LINER,
*CLOAD,AMPLITUDE=RELAX
100, 1, 5.4086E+04
101, 1, 4.3918E+04
102, 1, 8.6901E+04
103, 1, 1.2732E+05
104, 1, 1.6185E+05
105, 1, 1.8949E+05
106, 1, 2.0701E+05
2100, 1, 8.2710E+04
2101, 1, 6.3287E+04
2102, 1, 1.2031E+05
2103, 1, 1.6549E+05
2104, 1, 1.9676E+05
2105, 1, 2.1052E+05
100, 2, 1.6652E+05
101, 2, 3.2459E+05
102, 2, 2.9838E+05
103, 2, 2.5160E+05
104, 2, 1.8487E+05
105, 2, 9.9587E+04
106, 2, -2142.
2100, 2, -2.4756E+05
2101, 2, -4.7534E+05
2102, 2, -4.1880E+05
2103, 2, -3.3399E+05
2104, 2, -2.2922E+05
2105, 2, -1.1476E+05
*BOUNDARY
XSYMM, 1
XSYMM, 6
*NODE PRINT,FREQUENCY=100,NSET=TUNNEL
U, RF
*EL PRINT,FREQUENCY=100,ELSET=LINER,
S,
*END STEP
** --------------------------------------
*STEP,INC=100
1-196
step 2: relax tunnel stress 40 %

*STATIC
*END STEP
** --------------------------------------
*STEP
step 3: add liner stress free
*STATIC
*MODEL CHANGE,ADD
LINER,
*END STEP
** --------------------------------------
*STEP,IN=100
step 4: relax tunnel stress to zero &
apply surface load
*STATIC
*DLOAD
SURFACE, P3, 50.0E3
*EL FILE,FREQUENCY=100,ELSET=LINER,
S,
*END STEP
1.1.12 Symmetric results transfer for a rubber bushing

Products: ABAQUS/Standard ABAQUS/Design
The *SYMMETRIC RESULTS TRANSFER option (``Transferring results from a symmetric mesh to a
three-dimensional mesh,'' Section 7.7.2 of the ABAQUS/Standard User's Manual) allows the user to
transfer the solution obtained from an axisymmetric analysis onto a three-dimensional model with the
same geometry. It also allows the transfer of a symmetric three-dimensional solution to a full
three-dimensional model. This capability can reduce the analysis cost of structures that undergo
symmetric deformation, followed by nonsymmetric deformation later during the loading history.
The use of the results transfer capability is illustrated in this example by considering the response of an
axisymmetric bushing to loading imposed by a shaft that fits through the center of the bushing. Only
results transfer from an axisymmetric to a three-dimensional model is considered here.
The bushing consists of inner and outer steel tubes that are bonded to a central rubber cylinder ( Figure
1.1.12-1). It is assumed that the outer perimeter of the bushing is fully fixed. The loading
conditions--which include axial, twisting, and bending loads--are typical of the loading conditions that
would be applied to study the stiffness characteristics of the rubber bushing.
Two loading sequences are considered in two separate analyses. The first analysis considers axial
displacement of the shaft, followed by rotation of the rigid shaft about a transverse axis; the latter
loading is called bending in these examples. The second analysis considers relative twisting between
the shaft and the outer perimeter of the bushing, followed by the same bending load as in the first
1-197
example. In both examples the first loading step is fully axisymmetric so that only a two-dimensional
analysis is required. However, a full three-dimensional model is required to model the subsequent
bending load. The *SYMMETRIC RESULTS TRANSFER option is used to transfer the axisymmetric
solution to the full three-dimensional model. This solution becomes the base, or reference, state in the
subsequent bending analysis.
It is desirable to reduce the stress concentrations to increase the service life of the bushing. To this end
the sensitivity of stresses in the axisymmetric model is studied with respect to two shape design
parameters: the fillet radius, r, and the thickness, t, of the rubber bushing at the top and bottom ends
where it is bonded to the inner steel tube.
Geometry and model

The bushing is 457.2 mm (18.0 in) long, with an outside diameter of 508.0 mm (20.0 in) and an inside
diameter of 228.6 mm (9.0 in). The steel is elastic with Young's modulus = 206.0 GPa (3.0 ´ 107 psi)
and Poisson's ratio = 0.3. The rubber is modeled as a fully incompressible hyperelastic material that at
all strain levels is relatively soft compared to the steel. The nonlinear elastic behavior of the rubber is
described by a strain energy function that is a second-order polynomial in the strain invariants.
The first model is discretized with standard axisymmetric elements since the axial loading results in
pure axisymmetric deformation; CAX4 elements are used for the steel components, and CAX4H
elements are used for the rubber component. The second model is discretized with first-order,
reduced-integration axisymmetric elements with twist (CGAX4R type elements) to accommodate the
relative twisting between the shaft and outer perimeter of the bushing. Rigid elements (element type
RAX2) are attached to the inside of the bushing in both models to represent the relatively stiff shaft.
The use of these elements also simplifies the application of the loading conditions. The axisymmetric
finite element mesh is shown in Figure 1.1.12-1.
The corresponding three-dimensional models are generated using the *SYMMETRIC MODEL
GENERATION, REVOLVE option (``Symmetric model generation,'' Section 7.7.1 of the
ABAQUS/Standard User's Manual). The first model consists of a 180.0° revolution, and the second
consists of a 360.0° revolution of the axisymmetric cross-section about the symmetry axis. The model
generation capability converts the CAX4 and CGAX4R elements to C3D8 and C3D8R elements,
respectively. It also converts the CAX4H and CGAX4RH elements to C3D8H and C3D8RH,
respectively, and the RAX2 elements to R3D4 elements. Both models are discretized with eight
elements along the circumference in a 180.0° segment. No mesh convergence studies were performed.
Design sensitivity analysis

The objective is to modify the bushing geometry to lower the maximum axial stress. Sensitivity
analysis is used to provide an approximate assessment of the change needed in the design parameters
to achieve this objective. Only the axisymmetric model is considered. To carry out the design
sensitivity analysis with respect to a shape design parameter, the gradients of the nodal coordinates
with respect to the design parameter must be specified in the *PARAMETER SHAPE VARIATION
option. One simple approach to obtaining these gradients is to perturb the shape design parameters r
and t one at a time and to record the perturbed coordinates. The gradients are then found by
1-198
numerically differencing the initial and perturbed nodal coordinates. The perturbation of the thickness
is such that it causes the line of nodes connecting the thickness dimension to the fillet radius to rotate
about the point of tangency of this line to the fillet radius.

In the first example an axial force of magnitude 10675.0 N (2400.0 lbs) is applied to the rigid body
reference node in the first step, while the outer steel tube is fully fixed. This solution is transferred to a
180.0° segment of the three-dimensional model; the result forms the base state from which the
subsequent bending loading is applied. Symmetry boundary conditions are applied on the surfaces of
the three-dimensional model at µ = 0.0° and 180.0°. To ensure that the model is in equilibrium at the
beginning of the three-dimensional analysis, a *STEP definition is included using boundary conditions
and loading that are consistent with the state of the axisymmetric model. This step is completed in one
loading increment. The deformed mesh is shown in Figure 1.1.12-2. A bending moment of 294.0 N m
(2600.0 lbs in) is applied to the rigid body reference node in the second step. Figure 1.1.12-3 shows
the resulting deformation. A 22% saving in analysis time is obtained in this example by taking
advantage of symmetry conditions and the result transfer capability.
In the second example a rotation of 8.6° (0.15 rad) with respect to the axis of symmetry is applied to
the outer perimeter in the first step, while the shaft is held fixed. This solution is transferred to a full
360.0° three-dimensional model; the result forms the base state from which the subsequent bending
loading is applied. An initial *STEP definition is again included, using the boundary conditions and
loading that are consistent with the state of the axisymmetric model, to ensure that the
three-dimensional model is in equilibrium. Figure 1.1.12-4 and Figure 1.1.12-5 show the deformed
mesh that results from the twisting and bending loads, respectively. An 86% saving in analysis time is
obtained in this example by transferring the axisymmetric analysis onto the three-dimensional model.
The first example, the axial and bending problem, can also be solved using asymmetric-axisymmetric
CAXAnn elements rather than by using the axisymmetric model generation and results transfer option
(see ``Finite sliding between concentric cylinders--axisymmetric and CAXA models,'' Section 1.6.13 of
the ABAQUS Verification Manual). Depending on the number of Fourier modes used, the CAXAnn
elements should execute the analysis more quickly than the symmetric model generation followed by
the results transfer option. However, the twisting and bending problem cannot be solved using
CAXAnn elements; in this case, the axisymmetric model generation and the results transfer option are
the most efficient method available.
Figure 1.1.12-6 shows the contours of axial stress in the rubber part of the bushing at the end of the
axisymmetric analysis. The maximum stress occurs near the top fillet close to the axis. Figure 1.1.12-7
and Figure 1.1.12-8 show the contours of the sensitivities of the axial stress for the shape design
variables r and t, respectively. Table 1.1.12-1 shows the normalized sensitivities of the maximum axial
stress, ¾max = 0.18 MPa (26.01 psi), with respect to the shape design variables. The normalization has
been carried out by multiplying the sensitivities by a characteristic dimension (initial fillet radius r0 =
12.7 mm (0.5 in) and initial thickness t0 = 15.24 mm (0.6 in)) and dividing by the maximum stress. As
can be inferred from this table, a change in the fillet radius influences the maximum stress to a larger
extent than a change in the thickness of the rubber. Hence, it is desirable to change r to modify the
stresses. To obtain approximately a 10% reduction in the maximum stress in the axial direction, the
1-199
fillet radius is increased by
¢¾max
¢r = d¾max
:
dr
Substituting for ¢¾max = 0:1¾max and d¾max =dr = 0.008 MPa/mm (28.75 psi/in) (see Figure
1.1.12-7) gives ¢r = 2.25 mm (0.09049 in). A reanalysis of the problem with the radius changed to
r0 + ¢r = 14.99 mm (0.59049 in) yields a reduction of 8.8% in the maximum axial stress, which is
slightly less than the goal of 10%. This is expected because of the nonlinearity of the problem; to
achieve the 10% reduction, this process would have to be repeated, which is essentially an
optimization problem.
Input files
bushing_cax4_axi.inp
Axisymmetric model with CAX4 elements.
bushing_cax4_3d.inp
Three-dimensional model created from CAX4 elements.
bushing_cgax4r_axi.inp
Axisymmetric model with CGAX4R elements.
bushing_cgax4r_3d.inp
Three-dimensional model created from CGAX4R elements.
bushing_node.inp
Node definitions.
bushing_steel.inp
Element definitions for the steel.
bushing_rubber.inp
Element definitions for the rubber.
bushing_rigid.inp
Element definitions for the rigid body.
bushing_cax4_axi_dsa.inp
Design sensitivity analysis for the axisymmetric model.
Table
Table 1.1.12-1 Normalized sensitivities of the maximum stress.
1-200
r0 d¾max t0 d¾max
Parameter ¾max dr ¾max dt
r -0.55 --
t -- -0.11
Figures
Figure 1.1.12-1 Axisymmetric cross-section.
Figure 1.1.12-2 Deformed mesh after axial loading.
Figure 1.1.12-3 Deformed mesh after axial loading followed by nonaxisymmetric loading.
1-201
Figure 1.1.12-4 Deformed mesh after twist loading.
Figure 1.1.12-5 Deformed mesh after twist loading followed by nonaxisymmetric loading.
Figure 1.1.12-6 Variation of axial stress in the rubber after axial loading.
1-202
Figure 1.1.12-7 Variation of the sensitivity of the axial stress with respect to an increase in the radius
of the fillet, r.
1-203
Figure 1.1.12-8 Variation of the sensitivity of the axial stress with respect to a decrease in the
thickness of the rubber, t.
Sample listings
1-204
Listing 1.1.12-1
*HEADING
AXISYMMETRIC BUSHING
*RESTART,WRITE
*NODE,NSET=NALL,INPUT=bushing_node.inp
*NODE
229,0.,0.,0.
*ELEMENT, TYPE=CAX4,INPUT=bushing_steel.inp
*ELEMENT, TYPE=CAX4H,ELSET=RUBBER,INPUT=bushing_rubber.inp
*ELEMENT,TYPE=RAX2,ELSET=RIGID,INPUT=bushing_rigid.inp
*RIGID BODY,ELSET=RIGID,REF NODE=229
*ELSET,ELSET=INNER,GENERATE
1,5,1
76,79,1
163,171,1
*ELSET,ELSET=OUTER,GENERATE
178,184,1
86,92,1
*NSET, NSET=OUT_ST
110, 111, 112, 113, 114, 115, 116, 117,
222, 223, 224, 225, 226, 227, 228
*NSET,NSET=IN_ST
1, 2, 3, 4, 5, 6, 98, 99,
100, 101, 209, 210, 211, 212, 213, 214,
215, 216, 217,
*SOLID SECTION, ELSET=INNER, MATERIAL=STEEL-MATERIAL
1.,
*SOLID SECTION, ELSET=OUTER, MATERIAL=STEEL-MATERIAL
1.,
*SOLID SECTION, ELSET=RUBBER, MATERIAL=RUBBER-MATERIAL
1.,
*MATERIAL, NAME=STEEL-MATERIAL
*ELASTIC
3.E+7,0.3
*MATERIAL, NAME=RUBBER-MATERIAL
*HYPERELASTIC, N=2
11.5796, 3.47492, 2.269385E-1, -1.77868E-1, 8.5253E-3, 0.0, 0.0
*STEP, NLGEOM
*STATIC
1.0, 1.0
*BOUNDARY
OUT_ST,1,2,0.0
1-205
229,1,1
229,6,6
*CLOAD
229,2,2400.0
*NODE PRINT,FREQ=999
U,
RF,
*NODE FILE,FREQ=999
U,
RF,
*EL PRINT,FREQ=999
S,
E,
*EL FILE,FREQ=999
S,
E,
*END STEP
1-206
Listing 1.1.12-2
*HEADING
AXISYMMETRIC BUSHING
*SYMMETRIC MODEL GENERATION,REVOLVE
0,0,0,0,0,1
1,0,0
180.0,8,1.
*NSET,NSET=SYMM,GEN
1,228,1
1833,2060,1
*NSET,NSET=NOUT,GEN
917,1144,1
*ELSET,ELSET=EOUT,GEN
1555,1738,1
*SYMMETRIC RESULTS TRANSFER,STEP=1,INC=1
*TRANSFORM,NSET=NALL,TYPE=C
0.,0.,0.,0.,0.,10.
*FILE FORMAT,ZERO INCREMENT
*STEP,NLGEOM
*STATIC
1.0, 1.0
*BOUNDARY
OUT_ST,1,3
SYMM,2,2,0.0
229,1,2,0.0
229,4,6,0.0
*CLOAD
229,3,1200.
*NODE PRINT,NSET=NOUT,FREQ=999
U,
RF,
*NODE FILE,NSET=NOUT,FREQ=999
U,
RF,
*EL PRINT,ELSET=EOUT,FREQ=999
S,
E,
*EL FILE,ELSET=EOUT,FREQ=999
S,
E,
*END STEP
1-207
*STEP,NLGEOM,INC=20
*STATIC
.1, 1.0
*BOUNDARY,OP=NEW
OUT_ST,1,3
SYMM,2,2,0.0
229,2,2,0.0
*CLOAD
229,5,2600.
*NODE PRINT,NSET=NOUT,FREQ=999
U,
RF,
*NODE FILE,NSET=NOUT,FREQ=999
U,
RF,
*EL PRINT,ELSET=EOUT,FREQ=999
S,
E,
*EL FILE,ELSET=EOUT,FREQ=999
S,
E,
*END STEP
1.1.13 Transient loading of a viscoelastic bushing

This example demonstrates the automatic incrementation capability provided for integration of
time-dependent material models and the use of the viscoelastic material model in conjunction with
large-strain hyperelasticity in a typical design application. The structure is a bushing, modeled as a
hollow, viscoelastic cylinder. The bushing is glued to a rigid, fixed body on the outside and to a rigid
shaft on the inside, to which the loading is applied. A static preload is applied to the shaft, which
moves the inner shaft off center. This load is held for sufficient time for steady-state response to be
obtained. Then a torque is applied instantaneously and held for a long enough period of time to reach
steady-state response. We compute the bushing's transient response to these events.
Geometry and model

The viscoelastic bushing has an inner radius of 12.7 mm (0.5 in) and an outer radius of 25.4 mm (1.0
in). We assume that the bushing is long enough for plane strain deformation to occur. The problem is
modeled with first-order reduced-integration elements ( CPE4R). The mesh is regular, consisting of 6
elements radially, repeated 56 times to cover the 360° span in the hoop direction. The mesh is shown
in Figure 1.1.13-1. No mesh convergence studies have been performed.
The fixed outer body is modeled by fixing both displacement components at all the outside nodes. The
1-208
nodes in the inner boundary of the bushing are connected, using the *KINEMATIC COUPLING
option, to a node located in the center of the model. This node, thus, defines the inner shaft as a rigid
body.
Material
The material model is not defined from any particular physical material.
The instantaneous behavior of the viscoelastic material is defined by hyperelastic properties. A
polynomial model with N =1 (a Mooney-Rivlin model) is used for this, with the constants C10 =
27.56 MPa (4000 psi), C01 = 6.89 MPa (1000 psi), and D1 = 0.0029 MPa -1 (0.00002 psi -1).
The viscous behavior is modeled by a time-dependent shear modulus, GR (t), and a time-dependent
bulk modulus, KR (t), each of which is expanded in a Prony series in terms of the corresponding
instantaneous modulus,
2
X µ µ¶¶
t
GR (t)=G0 = 1 ¡ g¹iP
1 ¡ exp ¡
i=1
¿i
X2 µ µ ¶¶
¹ P t
KR (t)=K0 = 1 ¡ ki 1 ¡ exp ¡ :
i=1
¿ i
The relative moduli g¹iP and k¹iP and time constants ¿i are
i g¹iP k¹iP ¿i ; sec
1 0.2 0.5 0.1
2 0.1 0.2 0.2
This model results in an initial instantaneous Young's modulus of 206.7 MPa (30000 psi) and
Poisson's ratio of 0.45. It relaxes pressures faster than shear stresses.
Analysis
The analysis is done in four steps. The first step is a preload of 222.4 kN (50000 lbs) applied in the
x-direction to the node in the center of the model in 0.001 sec with a *STATIC procedure (``Static
stress analysis,'' Section 6.2.2 of the ABAQUS/Standard User's Manual). The *STATIC procedure
does not allow viscous material behavior, so this response is purely elastic. During the second step the
load stays constant and the material is allowed to creep for 1 sec by using the *VISCO procedure
(``Quasi-static analysis,'' Section 6.2.5 of the ABAQUS/Standard User's Manual). Since 1 sec is a long
time compared with the material time constants, the solution at that time should be close to steady
state. The CETOL parameter on the *VISCO option defines the accuracy of the automatic time
incrementation during creep response. CETOL is an upper bound on the allowable error in the creep
strain increment in each time increment. It is chosen as 5 ´ 10-4, which is small compared to the elastic
strains. The third step is another *STATIC step. Here the loading is a torque of 1129.8 N-m (10000
lb-in) applied in 0.001 sec. The fourth step is another *VISCO step with a time period of 1 sec.

Figure 1.1.13-2 through Figure 1.1.13-5depict the deformed shape of the bushing at the end of each
1-209
step. Each of the static loads produces finite amounts of deformation, which are considerably expanded
during the holding periods. Figure 1.1.13-6shows the displacement of the center of the bushing in the
x-direction and its rotation as functions of time.
Input file
viscobushing.inp
Input data for the analysis.
Figures
Figure 1.1.13-1 Finite element model of viscoelastic bushing.
Figure 1.1.13-2 Deformed model after horizontal static loading.
Figure 1.1.13-3 Deformed model after first holding period.
1-210
Figure 1.1.13-4 Deformed model after static moment loading.
Figure 1.1.13-5 Deformed model after second holding period.
Figure 1.1.13-6 Displacement and rotation of center of bushing.
1-211
Sample listings
1-212
Listing 1.1.13-1
*HEADING
TRANSIENT LOADING OF A VISCOELASTIC BUSHING --
CPE4R
*NODE
1,0.5,0.
7,1.0,0.
999,0.,0.
*NGEN,NSET=RADIAL
1,7
*NCOPY,CHANGE NUMBER=10,OLD SET=RADIAL,
NEW SET=ALL,SHIFT,MULTIPLE=55
0.,0.,0.,0.,0.,1.,6.4285714
*NSET,NSET=INSIDE,GENERATE
1,551,10
*NSET,NSET=OUTSIDE,GENERATE
7,557,10
*ELEMENT,TYPE=CPE4R,ELSET=ONE
1, 1,2,12,11
*ELGEN,ELSET=ALL
1,6,1,1,55,10,10
*ELSET,ELSET=TWO
1,2
*ELSET,ELSET=RIM,GENERATE
1,21,10
2,22,10
401,421,10
402,422,10
*ELEMENT,TYPE=CPE4R
551,551,552,2,1
*ELGEN,ELSET=ALL
551,6,1,1
*KINEMATIC COUPLING, REF NODE=999
INSIDE,1,2
*BOUNDARY
OUTSIDE,1,2
*MATERIAL,NAME=RUBBER
*HYPERELASTIC,N=1,MODULI=INSTANTANEOUS
4000.,1000.,0.00002
*VISCOELASTIC,TIME=PRONY
0.2,0.5,.1
1-213
0.1,0.2,.2
*SOLID SECTION,ELSET=ALL,MATERIAL=RUBBER
1.,
*NSET,NSET=CENTER
999,
*STEP,NLGEOM
*STATIC
0.001,0.001,0.0001
*CLOAD
999,1,50000.
*NODE PRINT,NSET=CENTER
*NODE FILE,FREQUENCY=150,NSET=RADIAL
U,RF
*EL FILE,FREQUENCY=150,ELSET=TWO
S,E,CE,CEP
*OUTPUT,VAR=PRESELECT,FIELD,FREQ=999
*NODE OUTPUT,NSET=CENTER
U,
*OUTPUT,FIELD,Frequency=150
*NODE OUTPUT,NSET=RADIAL
U,RF
*OUTPUT,HISTORY,Frequency=150
U,RF
*ELEMENT OUTPUT,ELSET=TWO
S,E,CE,CEP
*ELEMENT OUTPUT,ELSET=TWO
S,E,CE,CEP
*END STEP
*STEP,NLGEOM,INC=150
*VISCO,CETOL=5.E-4
0.04,1.
*END STEP
*STEP,NLGEOM
*STATIC
0.001,0.001
*CLOAD
999,6,10000.
*NODE FILE,FREQUENCY=150,NSET=RADIAL
1-214
U,RF
*EL FILE,FREQUENCY=150,ELSET=RIM
S,E,CE,CEP
U,RF
*NODE OUTPUT,NSET=CENTER
U,
U,RF
*OUTPUT,VAR=PRESELECT,FIELD,FREQ=999
*ELEMENT OUTPUT,ELSET=RIM
S,E,CE,CEP
*ELEMENT OUTPUT,ELSET=RIM
S,E,CE,CEP
*END STEP
*VISCO,CETOL=5.E-4
0.04,1.
*END STEP
1.1.14 Indentation of a thick plate

Product: ABAQUS/Explicit
This example illustrates the use of adaptive meshing in deep indentation problems with graded meshes.
Problem description
A deep indentation problem is solved for both axisymmetric and three-dimensional geometries, as
shown in Figure 1.1.14-1. Each model consists of a rigid punch and a deformable blank. The punch has
a semicircular nose section and a radius of 100 mm. The blank is modeled as a von Mises
elastic-plastic material with a Young's modulus of 3 ´ 106 MPa, an initial yield stress of 1.5 ´ 105
MPa, and a constant hardening slope of .45 ´ 105 MPa. Poisson's ratio is 0.3; the density is 1.0 ´ 10-5
kg/mm3.
In both cases the punch is fully constrained except in the vertical direction. A deep indentation is made
by moving the punch into the blank to a depth of 250 mm. The displacement of the punch is prescribed
using the SMOOTH STEP parameter on the *AMPLITUDE option so that a quasi-static response is
generated.
1-215
Case 1: Axisymmetric model

The blank is meshed with CAX4R elements and measures 300 ´ 300 mm. The punch is modeled as an
analytical rigid surface using the *SURFACE, TYPE=SEGMENTS option in conjunction with the
*RIGID BODY option. The bottom of the blank is constrained in the x- and z-directions, and symmetry
boundary conditions are prescribed at r=0.
Case 2: Three-dimensional models

Two models are analyzed. For one model the blank is meshed uniformly, while for the other a graded
mesh is used. For both models the blank is meshed with C3D8R elements and measures 600 ´ 300 ´
600 mm. The punch is modeled as an analytical rigid surface using the *SURFACE,
TYPE=REVOLUTION option in conjunction with the *RIGID BODY option. The bottom of the blank
is fully constrained.
Adaptive meshing
A single adaptive mesh domain that incorporates the entire blank is used for each model. A Lagrangian
boundary region type (the default) is used to define the constraints along the bottom of the plate for
both models and along the axis of symmetry in two dimensions. A sliding boundary region (the
default) is used to define the contact surface on the plate. To obtain a good mesh throughout the
simulation, the number of mesh sweeps is increased to 3 using the MESH SWEEPS parameter on the
*ADAPTIVE MESH option. For the graded three-dimensional model the SMOOTHING OBJECTIVE
parameter is set to GRADED on the *ADAPTIVE MESH CONTROLS option to preserve the
gradation of the mesh while adaptive meshing is performed.

Figure 1.1.14-2 to Figure 1.1.14-4show the initial configurations for the axisymmetric model, the
three-dimensional uniform mesh model, and the three-dimensional graded mesh model. Although the
punch is not shown in these figures, it is initially in contact with the plate. Figure 1.1.14-5 shows the
final deformed mesh for the axisymmetric indentation. The meshing algorithm attempts to minimize
element distortion both near and away from the contact surface with the punch. Figure 1.1.14-6 and
Figure 1.1.14-7show the deformed mesh of the entire blank and a quarter-symmetry, cutaway view,
respectively, for the three-dimensional model with an initially uniform mesh. Even under this depth of
indentation, elements appear to be nicely shaped both on the surface and throughout the cross-section
of the plate.
Figure 1.1.14-8 and Figure 1.1.14-9show the deformed mesh of the entire plate and a
quarter-symmetry, cutaway view, respectively, for the three-dimensional case with an initially graded
mesh. Adaptive meshing with the graded smoothing objective preserves the mesh gradation throughout
the indentation process while simultaneously minimizing element distortion. Preserving mesh
gradation in adaptivity problems is a powerful capability that allows mesh refinement to be
concentrated in the areas of highest strain gradients. A contour plot of equivalent plastic strain for the
graded mesh case is shown in Figure 1.1.14-10.
1-216
Input files
ale_indent_axi.inp
Case 1.
ale_indent_sph.inp
Case 2 with a uniform mesh.
ale_indent_gradedsph.inp
Case 2 with a graded mesh.
ale_indent_sphelset.inp
External file referenced by Case 2.
Figures
Figure 1.1.14-1 Axisymmetric and three-dimensional model geometries.
1-217
Figure 1.1.14-2 Initial configuration for the axisymmetric model.
Figure 1.1.14-3 Initial configuration for the three-dimensional model with a uniform mesh.
Figure 1.1.14-4 Initial configuration for the three-dimensional model with a graded mesh.
1-218
Figure 1.1.14-5 Deformed configuration for the axisymmetric model.
Figure 1.1.14-6 Deformed configuration for the three-dimensional model with an initially uniform
mesh.
1-219
Figure 1.1.14-7 Quarter-symmetry, cutaway view of the deformed configuration for the
three-dimensional model with an initially uniform mesh.
Figure 1.1.14-8 Deformed configuration for the three-dimensional model with an initially graded
mesh.
1-220
Figure 1.1.14-9 Quarter-symmetry, cutaway view of the deformed configuration for the
three-dimensional model with an initially graded mesh.
Figure 1.1.14-10 Contours of equivalent plastic strain for the three-dimensional model with an
initially graded mesh.
1-221
Sample listings
1-222
Listing 1.1.14-1
*HEADING
ADAPTIVE MESHING EXAMPLE
3D SPHERICAL INDENTATION
Units - N, mm, sec
*NODE,NSET=ALLN
1,0.,300.
10,0.,0.
401,600.,300.
410,600.,0.
*NSET,NSET=N1
1,
*NSET,NSET=N10
10,
*NSET,NSET=N401
401,
*NSET,NSET=N410
410,
*NFILL,NSET=TOP2D
N1,N401,40,10
*NFILL,NSET=BOT2D
N10,N410,40,10
*NFILL,NSET=HEAD
TOP2D,BOT2D,9,1
*NCOPY, SHIFT, OLD SET=HEAD, NEW SET=TAIL,
CHANGE NUMBER=16400
0., 0., 600.
0., 0., -1., 0., 0., 1., 0.
*NFILL, NSET=NALL
HEAD, TAIL, 40, 410
*ELEMENT,TYPE=C3D8R
1,2,12,11,1,412,422,421,411
*ELGEN,ELSET=BLANK
1,40,10,1,9,1,40,40,410,360
*NCOPY, SHIFT, OLD SET=TOP2D, NEW SET=TOPLAST,
CHANGE NUMBER=16400
0., 0., 600.
0., 0., -1., 0., 0., 1., 0.
*NFILL, NSET=TOP
TOP2D, TOPLAST, 40, 410
*NCOPY, SHIFT, OLD SET=BOT2D, NEW SET=BOTLAST,
1-223
CHANGE NUMBER=16400
0., 0., 600.
0., 0., -1., 0., 0., 1., 0.
*NFILL, NSET=BOT
BOT2D, BOTLAST, 40, 410
*INCLUDE,INPUT=ale_indent_sphelset.inp
*NODE,NSET=NOUT
100000,300.,410.,300.
*ELEMENT,TYPE=MASS,ELSET=PMASS
100000,100000
*MASS,ELSET=PMASS
0.2,
*SOLID SECTION,ELSET=BLANK,MATERIAL=MAT,
CONTROLS=SECT
*SECTION CONTROLS,HOURGLASS=STIFFNESS,
KINEMATICS=ORTHOGONAL,NAME=SECT
*MATERIAL,NAME=MAT
*ELASTIC
3.0E6,0.3
*PLASTIC
0.15E5, 0.0
0.6E5, 1.0
*DENSITY
1.E-5,
*BOUNDARY
BOT,1,3,
100000,1,
100000,3,
100000,4,6,
*AMPLITUDE,NAME=RAMPP,TIME=TOTAL TIME,
0.0,0.0,0.06,-250.,
*ELSET,ELSET=SMALL,GEN
6495,6505,1
6855,6865,1
7215,7225,1
*NSET,NSET=SMALL
7571,7581,7591,7981,7991,8001,8391,8401,
8411,
UPPER,S5
*SURFACE,TYPE=REVOLUTION,NAME=PUNCH
300., 400., 300.,300.,600., 300.
1-224
START,100.,0.
CIRCL,0.,-100.,0.,0.
*RIGID BODY,REF NODE=100000,
ANALYTICAL SURFACE =PUNCH
*STEP
*DYNAMIC,EXPLICIT
,0.06
*BOUNDARY,AMPLITUDE=RAMPP
100000,2,2,1.
*SURFACE INTERACTION,NAME=IMP_TARG
*CONTACT PAIR,INTERACTION=IMP_TARG
PUNCH,TARGET
*EL HISTORY,ELSET=SMALL
S,LE,LEP,NE,NEP,PEEQ
*NODE HISTORY,NSET=SMALL
RF,U
*ENERGY HISTORY
DT,
*EL FILE, ELSET=SMALL
S,LE,LEP,NE,NEP
*NODE FILE, NSET=NOUT
U,RF
*ENERGY FILE
*ADAPTIVE MESH, ELSET=BLANK, MESH SWEEPS=3
*END STEP
1.1.15 Damage and failure of a laminated composite plate

This example demonstrates how the *USER DEFINED FIELD option and user subroutine USDFLD
(`ÙSDFLD,'' Section 23.2.36 of the ABAQUS/Standard User's Manual) can be applied to model
nonlinear material behavior in a composite laminate. With this option it is possible to modify the
standard linear elastic material behavior (for instance, to include the effects of damage) or to change
the behavior of the nonlinear material models in ABAQUS. The material model in this example
includes damage, resulting in nonlinear behavior. It also includes various modes of failure, resulting in
abrupt loss of stress carrying capacity (Chang and Lessard, 1989). The analysis results are compared
with experimental results.
Problem description and material behavior

A composite plate with a hole in the center is subjected to in-plane compression. The plate is made of
1-225
24 plies of T300/976 graphite-epoxy in a [(-45/+45) 6 ]s layup. Each ply has a thickness of 0.1429 mm
(0.005625 in); thus, the total plate thickness is 3.429 mm (0.135 in). The plate has a length of 101.6
mm (4.0 in) and a width of 25.4 mm (1.0 in), and the diameter of the hole is 6.35 mm (0.25 in). The
plate is loaded in compression in the length direction. The thickness of the plate is sufficient that
out-of-plane displacements of the plate can be ignored. The compressive load is measured, as well as
the length change between two points, originally a distance of 25.4 mm (1.0 in) apart, above and below
the hole. The plate geometry and loading are shown in Figure 1.1.15-1.
The material behavior of each ply is described in detail by Chang and Lessard. The initial elastic ply
properties are longitudinal modulus Ex =156512 MPa (22700 ksi), transverse modulus Ey =12962
MPa (1880 ksi), shear modulus Gxy =6964 MPa (1010 ksi), and Poisson's ratio ºxy =0.23. The material
accumulates damage in shear, leading to a nonlinear stress-strain relation of the form
°xy = G¡1 3
xy ¾xy + ®¾xy ;
where Gxy is the (initial) ply shear modulus and the nonlinearity is characterized by the factor
®=2.44´10-8 MPa-3 (0.8´1-5 ksi-3).
Failure modes in laminated composites are strongly dependent on geometry, loading direction, and ply
orientation. Typically, one distinguishes in-plane failure modes and transverse failure modes
(associated with interlaminar shear or peel stress). Since this composite is loaded in-plane, only
in-plane failure modes need to be considered, which can be done for each ply individually. For a
unidirectional ply as used here, five failure modes can be considered: matrix tensile cracking, matrix
compression, fiber breakage, fiber matrix shearing, and fiber buckling. All the mechanisms, with the
exception of fiber breakage, can cause compression failure in laminated composites.
The failure strength in laminates also depends on the ply layup. The effective failure strength of the
layup is at a maximum if neighboring plies are orthogonal to each other. The effective strength
decreases as the angle between plies decreases and is at a minimum if plies have the same direction.
(This is called a ply cluster.) Chang and Lessard have obtained some empirical formulas for the
effective transverse tensile strength; however, in this model we ignore such effects. Instead, we use the
following strength properties for the T300/976 laminate: transverse tensile strength Yt =102.4 MPa
(14.86 ksi), ply shear strength Sc =106.9 MPa (15.5 ksi), matrix compressive strength Yc =253.0 MPa
(36.7 ksi), and fiber buckling strength Xc =2707.6 MPa (392.7 ksi).
The strength parameters can be combined into failure criteria for multiaxial loading. Four different
failure modes are considered in the model analyzed here.
· Matrix tensile cracking can result from a combination of transverse tensile stress, ¾y , and shear
stress, ¾xy . The failure index, em , can be defined in terms of these stresses and the strength
parameters, Yt and Sc . When the index exceeds 1.0, failure is assumed to occur. Without nonlinear
material behavior, the failure index has the simple form,
µ ¶2 µ ¶2
¾y ¾xy
e2m = + :
Yt Sc
1-226
With nonlinear shear behavior taken into consideration, the failure index takes the more complex
form,
µ ¶2 2 4
¾y 2¾xy =Gxy + 3®¾xy
e2m = + :
Yt 2Sc2 =Gxy + 3®Sc4
· Matrix compressive failure results from a combination of transverse compressive stress and shear
stress. The failure criterion has the same form as that for matrix tensile cracking:
µ ¶2 2 4
¾y 2¾xy =Gxy + 3®¾xy
e2m = + :
Yc 2Sc2 =Gxy + 3®Sc4
The same failure index is used since the previous two failure mechanisms cannot occur
simultaneously at the same point. After the failure index exceeds 1.0, both the transverse stiffness
and Poisson's ratio of the ply drop to zero.
· Fiber-matrix shearing failure results from a combination of fiber compression and matrix
shearing. The failure criterion has essentially the same form as the other two criteria:
µ ¶2 2 4
¾x 2¾xy =Gxy + 3®¾xy
e2f s = + :
Xc 2Sc2 =Gxy + 3®Sc4
This mechanism can occur simultaneously with the other two criteria; hence, a separate failure
index is used. Shear stresses are no longer supported after the failure index exceeds 1.0, but direct
stresses in the fiber and transverse directions continue to be supported.
· Fiber buckling failure occurs when the maximum compressive stress in the fiber direction ( ¡¾x )
exceeds the fiber buckling strength, Xc , independent of the other stress components:
¾x
eb = ¡ :
Xc
It is obvious that, unless the shear stress vanishes exactly, fiber-matrix shearing failure occurs
prior to fiber buckling. However, fiber buckling may follow subsequent to fiber shearing because
only the shear stiffness degrades after fiber-matrix shearing failure. Fiber buckling in a layer is a
catastrophic mode of failure. Hence, after this failure index exceeds 1.0, it is assumed that the
material at this point can no longer support any loads.
In this example the primary loading mode is shear. Therefore, failure of the plate occurs well before
the fiber stresses can develop to a level where fiber buckling takes place, and this failure mode need
not be taken into consideration.
Chang and Lessard assume that after failure occurs, the stresses in the failed directions drop to zero
immediately, which corresponds to brittle failure with no energy absorption. This kind of failure model
usually leads to immediate, unstable failure of the composite. This assumption is not very realistic: in
1-227
reality, the stress-carrying capacity degrades gradually with increasing strain after failure occurs.
Hence, the behavior of the composite after onset of failure is not likely to be captured well by this
model. Moreover, the instantaneous loss of stress-carrying capacity also makes the postfailure analysis
results strongly dependent on the refinement of the finite element mesh and the finite element type
used.
Material model implementation

To simulate the shear nonlinearity and the failure modes (matrix failure in tension or compression and
fiber-matrix shear failure), the elastic properties are made linearly dependent on three field variables.
The first field variable represents the matrix failure index, the second represents the fiber-matrix shear
failure index, and the third represents the shear nonlinearity (damage) prior to failure. The dependence
of the elastic material properties on the field variables is shown in Table 1.1.15-1.
To account for the nonlinearity, the nonlinear stress-strain relation must be expressed in a different
form: the stress at the end of the increment must be given as a linear function of the strain. The most
obvious way to do this is to linearize the nonlinear term, leading to the relation
³ ´
(i+1)
°xy = G¡1
xy + ®(¾ (i) 2
xy ) (i+1)
¾xy ;
where i represents the increment number. This relation can be written in inverted form as
(i+1) Gxy (i+1)

¾xy = (i)
°xy ;
1+ ®Gxy (¾xy )2
thus providing an algorithm to define the effective shear modulus.

However, this algorithm is not very suitable because it is unstable at higher strain levels, which is
readily demonstrated by stability analysis. Consider an increment where the strain does not change;
(i+1) (i) e (i)
i.e., °xy = °xy = °xy : Let the stress at increment i have a small perturbation from ¾xy , the exact
(i) e (i) (i)
solution at that increment: ¾xy = ¾xy + ¢¾xy . Similarly, at increment i+1,
(i+1) e (i+1) (i+1) (i+1)
¾xy = ¾xy + ¢¾xy . For the algorithm to be stable, ¢¾xy should not be larger than
(i) (i)
¢¾xy .The perturbation in increment i+1 is calculated by substituting ¾xy in the effective shear
e (i)
modulus equation and linearizing it about ¾xy :
(i+1)
¡2®G2xy ¾xy °xy (i)
¢¾xy = 2 2
¢¾xy ;
(1 + ®Gxy ¾xy )
e (i)
where ¾xy = ¾xy : The perturbation in increment i+1 is larger than the perturbation in increment i if
2®G2xy ¾xy °xy > (1 + ®Gxy ¾xy

2 2
) ;
which, after elimination of °xy , reduces to the expression
1-228
3 ¡1
®¾xy > Gxy ¾xy :
Hence, instability occurs when the "nonlinear" part of the shear strain is larger than the "linear" part of
the shear strain.
To obtain a more stable algorithm, we write the nonlinear stress-strain law in the form
3
°xy + ¯¾xy = G¡1 3
xy ¾xy + (® + ¯ )¾xy ;
where ¯ is an as yet unknown coefficient. In linearized form this leads to the update algorithm
³ ´
(1 + (i) 3
¯ (¾xy (i)
) =°xy (i+1)
)°xy = G¡1
xy + (® + (i) 2
¯ )(¾xy ) (i+1)
¾xy ;
or, in inverted form,
(i) (i)
(i+1) 1 + ¯ (¾xy )3 =°xy (i+1)
¾xy = (i)
Gxy °xy :
1 + (® + ¯ )Gxy (¾xy )2
Following the same procedure as that for the original update algorithm, it is readily derived that a small
(i)
perturbation, ¢¾xy , in increment i reduces to zero in increment i+1 if ¯ = 2®. Hence, the optimal
algorithm appears to be
(i) (i)
(i+1) 1 + 2®(¾xy )3 =°xy (i+1)
¾xy = (i)
Gxy °xy :
1+ 3®Gxy (¾xy )2
Finally, this relation is written in terms of the damage parameter d:
(i+1) (i+1)
¾xy = (1 ¡ d)Gxy °xy ;
where
(i) (i) (i)

3®Gxy (¾xy )2 ¡ 2®(¾xy )3 =°xy
d= (i)
:
1 + 3®Gxy (¾xy )2
This relation is implemented in user subroutine USDFLD, and the value of the damage parameter is
assigned directly to the third field variable used for definition of the elastic properties.
The failure indices are calculated with the expressions discussed earlier, based on the stresses at the
start of the increment:
1-229
Ã !2
(i) (i) (i)
¾y 2(¾xy )2 =Gxy + 3®(¾xy )4
e2m = + if ¾y(i) > 0;
Yt 2Sc2 =Gxy + 3®Sc4
Ã !2
(i) (i) (i)
¾y 2(¾xy )2 =Gxy + 3®(¾xy )4
e2m = + if ¾y(i) < 0;
Yc 2Sc2 =Gxy + 3®Sc4
Ã !2
(i) (i) (i)
¾x 2(¾xy )2 =Gxy + 3®(¾xy )4
e2f s = + :
Xc 2Sc2 =Gxy + 3®Sc4
The values of the failure indices are not assigned directly to the field variables: instead, they are stored
as solution-dependent state variables. Only if the value of a failure index exceeds 1.0 is the
corresponding user-defined field variable set equal to 1.0. After the failure index has exceeded 1.0, the
associated user-defined field variable continues to have the value 1.0 even though the stresses may
reduce significantly, which ensures that the material does not "heal" after it has become damaged.
Finite element model

The plate consists of 24 plies of T300/976 graphite-epoxy in a [(-45/+45) 6 ]s layup. Instead of
modeling each ply individually, we combine all plies in the -45° direction and all plies in the +45°
direction. Consequently, only two layers need to be modeled separately:
1. A layer in the -45° direction with a thickness of 1.715 mm (0.0675 in).
2. A layer in the +45° direction with a thickness of 1.715 mm (0.0675 in).
The corresponding finite element model consists of two layers of CPS4 plane stress elements, with
thicknesses and properties as previously discussed. The quarter-symmetry finite element model is
shown in Figure 1.1.15-1.
The implementation of nonlinear material behavior with user-defined field variables is explicit: the
nonlinearity is based on the state at the start of the increment. Hence, the user must ensure that the time
increments are sufficiently small, which is particularly important because the automatic time increment
control in ABAQUS is ineffective with the explicit nonlinearity implemented in USDFLD. If automatic
time incrementation is used, the maximum time increment can be controlled from within subroutine
USDFLD with the variable PNEWDT. This capability is useful if there are other nonlinearities that
require automatic time incrementation. In this example the only significant nonlinearity is the result of
the material behavior. Hence, fixed time incrementation can be used effectively.

For this problem experimental load-displacement results were obtained by Chang and Lessard. The
experimental results, together with the numerical results obtained by ABAQUS, are shown in Figure
1.1.15-2. The agreement between the experimental and numerical results is excellent up to the point
where the load maximum is reached. After that, the numerical load-displacement curve drops off
sharply, whereas the experimental data indicate that the load remains more or less constant. Chang and
Lessard also show numerical results: their results agree with the results obtained by ABAQUS but do
not extend to the region where the load drops off. The dominant failure mode in this plate is
1-230
fiber/matrix shear: failure occurs first at a load of approximately 12.15 kN (2700 lbs) and continues to
grow in a stable manner until a load of approximately 13.5 kN (3000 lbs) is reached. Figure 1.1.15-3
shows the extent of the damage in the finite element model at the point of maximum load. In this figure
an element is shaded if fiber/matrix shear failure has occurred at at least three integration points. These
results also show excellent agreement with the results obtained by Chang and Lessard.
As discussed earlier, the sharp load drop-off in the numerical results is the result of the lack of residual
stress carrying capacity after the failure criterion is exceeded. Better agreement could be reached only
if postfailure material data were available. Without postfailure data the results are very sensitive to the
mesh and element type, which is clearly demonstrated by changing the element type from CPS4 (full
integration) to CPS4R (reduced integration). The results are virtually identical up to the point where
first failure occurs. After that point the damage in the CPS4R model spreads more rapidly than in the
CPS4 model until a maximum load of about 12.6 kN (2800 lbs) is reached. The load then drops off
rapidly.
The problem is also analyzed with models consisting of S4R and S4 elements. The elements have a
composite section with two layers, with each layer thickness equal to the thickness of the plane stress
elements in the CPS4 and CPS4R models. The results that were obtained with the S4R and S4 element
models are indistinguishable from those obtained with the CPS4R element model.
Input files
damagefailcomplate_cps4.inp
CPS4 elements.
damagefailcomplate_cps4.f
User subroutine USDFLD used in damagefailcomplate_cps4.inp.
damagefailcomplate_node.inp
Node definitions.
damagefailcomplate_element.inp
Element definitions.
damagefailcomplate_cps4r.inp
CPS4R elements.
damagefailcomplate_cps4r.f
User subroutine USDFLD used in damagefailcomplate_cps4r.inp.
damagefailcomplate_s4.inp
S4 elements.
damagefailcomplate_s4.f
User subroutine USDFLD used in damagefailcomplate_s4.inp.
damagefailcomplate_s4r.inp
1-231
S4R elements.
damagefailcomplate_s4r.f
User subroutine USDFLD used in damagefailcomplate_s4r.inp.
Reference
· Chang, F-K., and L. B. Lessard, ``Damage Tolerance of Laminated Composites Containing an
Open Hole and Subjected to Compressive Loadings: Part I--Analysis,'' Journal of Composite
Materials, vol. 25, pp. 2-43, 1991.
Table
Table 1.1.15-1 Dependence of the elastic material properties on the field variables.
Material State Elastic Properties FV1 FV2 FV3
No failure Ex Ey ºxy Gxy 0 0 0
Matrix failure Ex 0 0 Gxy 1 0 0
Fiber/matrix shear Ex Ey 0 0 0 1 0
Shear damage Ex Ey ºxy 0 0 0 1
Matrix failure and fiber/matrix
Ex 0 0 0 1 1 0
shear
Matrix failure and shear
Ex 0 0 0 1 0 1
damage
Fiber/matrix shear and shear
Ex Ey 0 0 0 1 1
damage
All failure modes Ex 0 0 0 1 1 1
Figures
Figure 1.1.15-1 Plate geometry and loading.
1-232
Figure 1.1.15-2 Experimental and numerical load displacement curves.
Figure 1.1.15-3 Distribution of material damage at maximum load.
1-233
Sample listings
1-234
Listing 1.1.15-1
*HEADING
COMPOSITE PLATE WITH EXPLICIT DAMAGE AND
CPS4 FAILURE MODEL
** UNITS: IN, LBS
** -------------------------
** MODEL DEFINITION
** -------------------------
*NODE,INPUT=damagefailcomplate_node.inp,NSET=NALL
*ELEMENT,INPUT=damagefailcomplate_element.inp,
TYPE=CPS4,ELSET=L1
*ELCOPY,ELEMENT SHIFT=1000,OLD SET=L1,NEW SET=L2,
SHIFT NODES=0
*NSET,NSET=XSYMMTRY
1, 2, 3, 4, 5, 6, 7, 8,
9, 10, 11, 12, 13, 190, 197, 204,
211, 218, 225, 232, 239, 246, 253, 260,
267, 1000
*NSET,NSET=YSYMMTRY
170, 171, 172, 173, 174, 175, 176, 177,
178, 179, 180, 181, 182, 287, 294, 301,
308, 315, 322
*NSET,NSET=TOP
267, 268, 269, 270, 271, 272, 273, 408,
409, 410, 411, 412, 413
**
** MATERIAL: NONLINEAR SHEAR WITH BUILT-IN
** EXPLICIT FAILURE
**
** FV1: MATRIX COMPRESSIVE/TENSILE FAILURE
** FV2: FIBER-MATRIX SHEAR FAILURE
** FV3: SHEAR NONLINEARITY (DAMAGE) PRIOR TO
** FAILURE
** TOTAL OF 2^3 = 8 STATES
**
*MATERIAL,NAME=T300
*ELASTIC,TYPE=LAMINA,DEPENDENCIES=3
22.7E6,1.88E6,0.23,1.01E6,1.01E6,1.01E6,0.,0,
0,0
22.7E6,1.00E0,0.00,1.01E6,1.01E6,1.01E6,0.,1,
0,0
22.7E6,1.88E6,0.00,1.00E0,1.01E6,1.01E6,0.,0,
1-235
1,0
22.7E6,1.00E0,0.00,1.00E0,1.01E6,1.01E6,0.,1,
1,0
22.7E6,1.88E6,0.23,1.00E0,1.01E6,1.01E6,0.,0,
0,1
22.7E6,1.00E0,0.00,1.00E0,1.01E6,1.01E6,0.,1,
0,1
22.7E6,1.88E6,0.00,1.00E0,1.01E6,1.01E6,0.,0,
1,1
22.7E6,1.00E0,0.00,1.00E0,1.01E6,1.01E6,0.,1,
1,1
*DEPVAR
3,
*USER DEFINED FIELD
**
** LOCAL ORIENTATIONS: P45 AT +45, N45 AT -45
**
*ORIENTATION,NAME=P45
0.,1.,0.,-1.,0.,0.
3,+45.0
*ORIENTATION,NAME=N45
0.,1.,0.,-1.,0.,0.
3,-45.0
*SOLID SECTION,ELSET=L1,ORIENTATION=P45,
MATERIAL=T300
0.06750,
*SOLID SECTION,ELSET=L2,ORIENTATION=N45,
MATERIAL=T300
0.06750,
**
** CONSTRAIN TOP FOR THE PURPOSES OF LOAD
** APPLICATION
**
*EQUATION
2,
TOP,2,1.0,1000,2,-1.0
**
** CONSTRAIN OUTPUT NODE BETWEEN SURROUNDING NODES
*MPC
LINEAR,2000,197,204
*NSET,NSET=OUTPUT
1000,2000,204,197
*ELSET,ELSET=EOUT,GENERATE
1-236
133,144
11133,11144
** -------------------
** ANALYSIS HISTORY
** -------------------
*STEP,INC=200,NLGEOM
*STATIC,DIRECT
0.05,1.0
*BOUNDARY
XSYMMTRY,XSYMM
YSYMMTRY,YSYMM
1000,2,,-0.027
*NODE PRINT,NSET=OUTPUT
U2,RF2
*NODE FILE,NSET=OUTPUT
U,RF
*EL FILE,ELSET=EOUT
S,E
SDV,FV
NE,LE
*EL PRINT,FREQUENCY=5,ELSET=EOUT
S,E
*END STEP
1.1.16 Analysis of an automotive boot seal

Boot seals are used to protect constant velocity joints and steering mechanisms in automobiles. These
flexible components must accommodate the motions associated with angulation of the steering
mechanism. Some regions of the boot seal are always in contact with an internal metal shaft, while
other areas come into contact with the metal shaft during angulation. In addition, the boot seal may
also come into contact with itself, both internally and externally. The contacting regions affect the
performance and longevity of the boot seal.
In this example the deformation of the boot seal, caused by a typical angular movement of the shaft, is
studied. It provides a demonstration and verification of the finite-sliding capability in
three-dimensional deformable-to-deformable contact in ABAQUS.
Geometry and model

The boot seal with the internal shaft is shown in Figure 1.1.16-1. The boot seal and shaft are modeled
as separate parts, each instanced once. Symmetry is utilized to model one-half of the boot seal. The
corrugated shape of the boot seal tightly grips the steering shaft at one end, while the other end is
1-237
fixed. The rubber seal is modeled with 1728 first-order, hybrid brick elements with two elements
through the thickness. The seal has a nonuniform thickness varying from a minimum of 3.0 mm to a
maximum of 4.75 mm at the fixed end. The rubber is modeled as a slightly compressible
Mooney-Rivlin (hyperelastic) material with C10 = 0.752 MPa and D1 = 0.026 MPa -1. The internal
shaft is considered to be rigid and is modeled as an analytical rigid surface; the radius of the shaft is 14
mm. The rigid body reference node is located precisely in the center of the constant velocity joint.
Contact is specified between the rigid shaft and the regions on the inner surface of the seal that are
likely to come into contact with the shaft. In Figure 1.1.16-2 the slave surface on the interior of the seal
is shown. Contact is also specified between facing regions on the inner and outer surfaces of the seal.
Figure 1.1.16-3 shows the surfaces on the interior of the seal that are likely to come into contact with
each other, and Figure 1.1.16-4 shows the surfaces on the exterior that may come into contact with
each other. The interactions between the appropriate pairs are specified by using the *CONTACT
PAIR option.
If a three-dimensional deformable master surface is defined from a large number of underlying
elements and the equations would be ordered to account for any possible contact between the slave and
master surface, the resulting wavefront would be very large. Hence, by default, ABAQUS/Standard
employs an automated moving contact patch and equation reordering algorithm to reduce the
wavefront and and solution time (see ``Common difficulties associated with contact modeling,''
Section 21.10.1 of the ABAQUS/Standard User's Manual). Alternatively, the user can specify a fixed,
non-moving, contact patch size with the SLIDE DISTANCE parameter on the *CONTACT
CONTROLS option. In such a case the maximum slide distance and patch location remain fixed
throughout the analysis. This can be more efficient if the relative motion of the slave and master
surfaces is limited to a few elements; although the wavefront might not be optimal, the avoidance of
reordering may yield some cost savings.
As an illustration, an additional input file has been included in which the SLIDE DISTANCE
parameter is utilized. Due to the motion of the shaft, slave nodes will slide more in the longitudinal
direction than in the circumferential direction. We can assume the maximum sliding distance of each
slave node to be equal to the width of the master surface with which it comes into contact (see Figure
1.1.16-3 and Figure 1.1.16-4).
Loading
The mounting of the boot seal and the angulation of the shaft are carried out in a two-step analysis.
The inner radius at the neck of the boot seal is smaller than the radius of the shaft so as to provide a
tight fit between the seal and the shaft. In the first step the initial interference fit is resolved,
corresponding to the assembly process of mounting the boot seal onto the shaft. The automatic
"shrink" fit method is utilized by including the SHRINK parameter on the *CONTACT
INTERFERENCE option. The second step simulates the angulation of the inner shaft by specifying a
finite rotation of 20° at the rigid body reference node of the shaft. Symmetry boundary conditions are
specified for all nodes lying in the symmetry plane of the model. The wider end of the boot seal is
constrained completely.
1-238
Figure 1.1.16-5 shows the deformed configuration of the model. The rotation of the shaft causes the
stretching of one side and compression on the other side of the boot seal. The surfaces have come into
contact on the compressed side. Figure 1.1.16-6 shows the contours of maximum principal stresses in
the boot seal.
Comparison of the analysis times when using fixed and automated contact patches shows that both
analyses complete in approximately the same amount of time. This can be expected for this type of
problem since the fixed contact patches are limited in size to a few elements. For the case with fixed
contact patches the wavefront is somewhat larger, requiring more memory and solution time per
iteration. However, this is offset by the time required to form new contact patches and to reorder the
equations for the case with automatic contact patches.
Input files
bootseal.inp
Analysis with automatic contact patches.
bootseal_fixed_cpatch.inp
Analysis with fixed contact patches.
bootseal_mesh.inp
Node and element definitions.
Figures
Figure 1.1.16-1 Undeformed model.
Figure 1.1.16-2 The non-shaded surface is the slave surface on the seal that may come into contact
1-239
with the shaft.
Figure 1.1.16-3 Surfaces on the interior of the seal that may come into contact with each other. The
shaded surface is the master surface.
Figure 1.1.16-4 Surfaces on the exterior of the seal that may come into contact with each other. The
shaded surface is the master surface.
1-240
Figure 1.1.16-5 Deformed configuration of the model.
Figure 1.1.16-6 Contours of maximum principal stress in the seal.
1-241
Sample listings
1-242
Listing 1.1.16-1
*HEADING
Analysis of automotive boot seal.
**
** read nodes and elements from external file
** 1728 C3D8H elements
** 2849 nodes
** create element sets: OUT_SLAVE,OUT_MASTER,
** BOOT_IN,IN_SLAVE,
** and IN_MASTER
** create node sets: MGS,ZSYMM,and REFNODE
**
*PART,NAME=SEAL
*INCLUDE, INPUT=bootseal_mesh.inp
*SOLID SECTION,ELSET=BOOT,MATERIAL=RUBBER
*SURFACE, NAME=OUT_SLAV
OUT_SLAVE,S4
*SURFACE, NAME=OUT_MAST
OUT_MASTER,S4
*SURFACE, NAME=BOOT_IN
BOOT_IN,S6
*SURFACE, NAME=IN_SLAVE
IN_SLAVE,S6
*SURFACE, NAME=IN_MASTE
IN_MASTER,S1
*END PART
*PART,NAME=SHAFT
*NODE,NSET=REFNODE
9999, 0.0000E+00,-3.3000E+01, 0.0000E+00
*RIGID BODY,ANALYTICAL SURFACE=SHAFT,REF NODE=9999
*SURFACE,TYPE=REVOLUTION,NAME=SHAFT
START, 14.,150.
LINE, 14.,0.
*END PART
*ASSEMBLY,NAME=BOOT-SEAL
*INSTANCE,NAME=SEAL-1,PART=SEAL
*END INSTANCE
*INSTANCE,NAME=SHAFT-1,PART=SHAFT
*END INSTANCE
*NSET, NSET=NHIST
1-243
SHAFT-1.REFNODE, SEAL-1.1, SEAL-1.6001

*ELSET, ELSET=EHIST
SEAL-1.6052,
*END ASSEMBLY
*HYPERELASTIC,N=1
0.752,0., .026
*CONTACT PAIR,INTERACTION=OUT_SELF
BOOT-SEAL.SEAL-1.OUT_SLAV, BOOT-SEAL.SEAL-1.OUT_MAST
*SURFACE INTERACTION, NAME=OUT_SELF
*CONTACT PAIR,INTERACTION=IN_SELF
BOOT-SEAL.SEAL-1.IN_SLAVE, BOOT-SEAL.SEAL-1.IN_MASTE
*SURFACE INTERACTION, NAME=IN_SELF
*CONTACT PAIR,INTERACTION=SHAFT_C
BOOT-SEAL.SEAL-1.BOOT_IN,BOOT-SEAL.SHAFT-1.SHAFT
*SURFACE INTERACTION, NAME=SHAFT_C
*BOUNDARY
BOOT-SEAL.SEAL-1.MGS, 1, 3, 0.0
BOOT-SEAL.SEAL-1.ZSYMM, ZSYMM
** PERFORM INTERFERENCE FIT BETWEEN SHAFT AND BOOT
*STEP,NLGEOM,INC=20
*STATIC
.25,1.
*CONTACT CONTROLS, SLAVE=BOOT-SEAL.SEAL-1.OUT_SLAV,
MASTER=BOOT-SEAL.SEAL-1.OUT_MAST, SLIDEDISTANCE=15.0
*CONTACT CONTROLS, SLAVE=BOOT-SEAL.SEAL-1.IN_SLAVE,
MASTER=BOOT-SEAL.SEAL-1.IN_MASTE, SLIDEDISTANCE=15.0
*BOUNDARY
BOOT-SEAL.SHAFT-1.REFNODE, 1,6
*CONTACT INTERFERENCE, SHRINK
BOOT-SEAL.SEAL-1.BOOT_IN,BOOT-SEAL.SHAFT-1.SHAFT
*PRINT,CONTACT=YES
*EL PRINT,FREQ=0
*NODE PRINT,NSET=BOOT-SEAL.SHAFT-1.REFNODE
*NODE PRINT,FREQ=0
*NODE FILE, NSET=BOOT-SEAL.NHIST, FREQ=999
U, RF
*ENERGY FILE, ELSET=BOOT-SEAL.SEAL-1.BOOT, FREQ=999
ELSE,
*END STEP
** ROTATE SHAFT 20 DEGREES (.349 RADIANS)
*STEP,NLGEOM,INC=20
*STATIC
1-244
.1,1.
*BOUNDARY, TYPE=VELOCITY
BOOT-SEAL.SHAFT-1.REFNODE, 6,6, .349
*NODE FILE, NSET=BOOT-SEAL.NHIST, FREQ=1
U, RF
*EL FILE, ELSET=BOOT-SEAL.EHIST, FREQ=1
S,
*ENERGY FILE, ELSET=BOOT-SEAL.SEAL-1.BOOT, FREQ=1
ELSE,
*END STEP
1.1.17 Pressure penetration analysis of an air duct kiss seal

Seals are common structural components that often require design analyses. ABAQUS can be used to
perform nonlinear finite element analyses of seals and provide information needed to determine the
seal performance. Information such as a load-deflection curve, seal deformation and stresses, and
contact pressure distribution is readily obtained in these analyses. ABAQUS allows for pressure
penetration effects between the seal and the contacting surfaces to be considered in these analyses,
making routine analyses of seals more realistic and accurate. Analyses of clutch seals, threaded
connectors, car door seals, and air duct kiss seals are some applications where pressure penetration
effects are important.
The surface-based pressure penetration capability is used to simulate pressure penetration between
contacting surfaces. It is invoked by using the *PRESSURE PENETRATION option, which is
described in ``Pressure penetration loading,'' Section 21.3.5 of the ABAQUS/Standard User's Manual.
This capability is provided for simulating cases where a joint between two deforming bodies (for
example, between two components threaded onto each other) or between a deforming body and a rigid
surface (such as a soft gasket used in a joint) is exposed at one or multiple ends to a fluid or air
pressure. This air pressure will penetrate into the joint and load the surfaces forming the joint until
some area of the surfaces is reached where the contact pressure between the abutting surfaces exceeds
the critical value specified on the *PRESSURE PENETRATION option, cutting off further
penetration.
Geometry and model

The major consideration in an air duct kiss seal design is to provide sealing while avoiding excessive
closure force. A poorly designed air duct seal that minimizes the amount of effort to close the fan cowl
door may fail to prevent leakage and reduce wind noise. The model used in this example is a simplified
version of an air duct kiss seal. It illustrates how pressure penetration effects can be modeled using
ABAQUS.
The seal modeled is a rolled shape seal. An axisymmetric model of the seal is developed, as shown in
Figure 1.1.17-1. The top horizontal rigid surface represents the air fan cowl door, and the bottom
horizontal rigid surface represents the seal groove. The rolled seal is 2.54 mm (0.1 in) thick and 74.66
1-245
mm (2.9 in) high; and its inner diameters at the top and bottom surfaces are 508.5 mm (20 in) and
528.3 mm (20.8 in), respectively. A folded metal clip is partially bonded to the top surface of the seal.
The thickness of the metal clip is 0.48 mm (0.019 in).
The material of the seal is taken to be an incompressible rubberlike material. To obtain the material
constants, the Ogden form of the strain energy function with N = 4 is used to fit the uniaxial test data.
The metal clip is made of steel, with a Young's modulus of 206.8 GPa (3.0 ´ 107 lb/in2) and a Poisson's
ratio of 0.3. CAX4H elements are used to model the seal and the metal clip. The contact pair approach
is used to model the contact between the top surface of the metal clip and the top rigid surface
representing the fan cowl door, where the pressure penetration is likely to occur. The contact pair
approach is also used to model the contact between the seal and the bottom rigid surface, the contact
between the seal and the unbonded portion of the metal clip, and the self contact of the seal. The
mechanical interaction between the contact surfaces is assumed to be frictional contact. Therefore, the
*FRICTION option is used to specify friction coefficients. To increase computational efficiency, the
SLIP TOLERANCE parameter on the *FRICTION option is used for the contact surfaces between the
seal and the metal clip because the dimensions of these elements vary greatly. Fixed boundary
conditions are applied initially to the reference node of the top rigid surface, 5001, and the reference
node of the bottom rigid surface, 5002. The vertical edge at the bottom of the seal is constrained such
that it cannot be moved in the x-direction. The bottom node of the vertical edge, 1, touches the bottom
rigid surface and is held fixed in the y-direction. The top rigid surface is located initially 1.27 mm
(0.05 in) above the top surface of the metal clip.
The seal and the unbonded portion of the clip are loaded by air pressure on all of their inner surfaces
and by contact pressure generated by closing the air fan cowl door. Two nonlinear static steps, all of
which include large-displacement effects, are used to simulate these loading conditions.
In the first step the top rigid surface moves 35.56 mm (1.4 in) downward in the y-direction, simulating
the closing of the fan cowl door.
In the second step the inner surface of the seal is subjected to a uniform air pressure load of 206.8 KPa
(30.0 lb/in 2) since some gaps between the seal and the top rigid surface have been closed. The pressure
penetration is simulated between the top surface of the metal clip ( PPRES), which includes 31
elements, and the top rigid surface ( CFACE). Air pressure penetration does not need to be modeled
between the metal clip and the seal because they are well bonded.
The *PRESSURE PENETRATION option is invoked to define the node exposed to the air pressure,
the magnitude of the air pressure, and the critical contact pressure. The surface PPRES is exposed to
the air pressure at node 597, with a pressure magnitude of 206.8 KPa (30.0 lb/in 2). A default value of
zero for the critical contact pressure is used, indicating that the pressure penetration occurs only when
contact at a slave node is lost.

The deformed configuration and the contours of the contact pressures on the seal at the end of Step 1
are shown in Figure 1.1.17-2 and Figure 1.1.17-3. A nonuniform contact pressure is observed along the
surface of the seal. The contact pressure at the first five slave nodes is zero.
1-246
The penetrating pressure loads are applied during Step 2. The air pressure is applied immediately to
elements associated with the first five slave nodes since the contact pressure there is zero and the
pressure penetration criterion is satisfied. The spread of the penetration is captured in Figure 1.1.17-4
through Figure 1.1.17-12, which show the deformed seal, the contact pressure profile, and the air
pressure profile corresponding to load increments 2, 8, and 12 of Step 2. The pressures applied to the
surface corresponding to these three increments are 3.23 KPa (0.469 lb/in 2), 23.37 KPa (3.39 lb/in 2),
and 64.74 KPa (9.39 lb/in 2), respectively.
Increased penetrating pressure loads applied in Step 2 further reduce the contact pressure, eventually
causing complete air penetration through the seal. The seal was lifted off from the air fan cowl door
except at the last slave node, 663, where the contact pressure is well-maintained due to imposed
boundary conditions and the air pressures. The development of the weakening of the sealing is
captured in Figure 1.1.17-13 through Figure 1.1.17-16, which show the deformed seal and the contact
pressure profile corresponding to load increments 14 and at the end of Step 2. The pressures applied to
the surface corresponding to these two increments are 85.5 KPa (12.4 lb/in 2) and 206.8 KPa (30.0
lb/in2), respectively.
The behavior of the seal throughout the loading histories can be best described by plotting the air
penetration distance as a function of the air pressure, as shown in Figure 1.1.17-17. It is clear that air
penetration into the seal accelerates only when the pressure is on the order of 65.5 KPa (9.5 lb/in 2).
The air completely penetrates through the seal when the pressure is 82.7 KPa (12.0 lb/in 2), which is
approximately equal to 80% of the sea level atmospheric pressure.
Input files
presspenairductseal.inp
Pressure penetration simulation of an air duct kiss seal.
presspenairductseal_node.inp
Node definitions for the seal model.
presspenairductseal_elem_metal.inp
Element definitions for the metal part of the seal model.
presspenairductseal_elem_rub.inp
Element definitions for the rubber part of the seal model.
Figures
Figure 1.1.17-1 Air duct kiss seal model.
1-247
Figure 1.1.17-2 Deformed configuration of the seal at the end of Step 1.
Figure 1.1.17-3 Contact stress contours in the seal at the end of Step 1.
1-248
Figure 1.1.17-4 Deformed configuration of the seal at Step 2, increment 2.
Figure 1.1.17-5 Contact stress contours in the seal at Step 2, increment 2.
1-249
Figure 1.1.17-6 Air pressure contours in the seal at Step 2, increment 2.
1-250
1-251
1-252
Figure 1.1.17-15 Deformed configuration of the seal at the end of Step 2.
1-253
Figure 1.1.17-16 Contact stress contours in the seal at the end of Step 2.
Figure 1.1.17-17 Air penetration distance as a function of air pressure in the seal.
Sample listings
1-254
Listing 1.1.17-1
*HEADING
Surface-based Pressure Penetration Analysis of
Air Duct Kiss Seal
*NODE,NSET=NSEAL,
INPUT=presspenairductseal_node.inp
*ELEMENT,TYPE=CAX4H,ELSET=STEEL,
INPUT=presspenairductseal_elem_metal.inp
*ELEMENT,TYPE=CAX4H,ELSET=RUBFIB,
INPUT=presspenairductseal_elem_rub.inp
*ELSET,ELSET=FACE1,GENERATE
437,439,1
443,452,1
28,36,1
67,76,1
227,276,1
307,316,1
440,442,1
453,462,1
473,475,1
437,440,3
*ELSET,ELSET=PPRES
476,477,478,479,480,481,482,503,504,
505,506,507,508,509,510,511,512,513,
514,515,516,517,518,519,520,521,522,
533,534,535,536
*NSET,NSET=EDGE
1,2,3,4,5
*NODE,NSET=BCYL
5001,20.5,2.9,0.
5002,20.5,0.,0.
*RIGID BODY,ANALYTICAL SURFACE=CFACE,REFNODE=5001
*SURFACE,NAME=CFACE,TYPE=SEGMENTS
START,21.5,2.9,
LINE,19.5,2.9
*SURFACE,NAME=PPRES
PPRES,S3
*CONTACT PAIR,INTERACTION=PPEN
PPRES,CFACE
*SURFACE INTERACTION,NAME=PPEN
1-255
*FRICTION
.00001,
*SURFACE BEHAVIOR,
PRESSURE-OVERCLOSURE=EXPONENTIAL
.0002,1000.
*RIGID BODY,ANALYTICAL SURFACE=CFACE2,
REF NODE=5002
*SURFACE,NAME=CFACE2,TYPE=SEGMENTS
START,19.5,0.,
LINE,21.5,0.
*ELSET,ELSET=ELBOT,GENERATE
1,9,1
*SURFACE,NAME=BOTTOM
ELBOT,
*CONTACT PAIR, INTERACTION=FLANGE
BOTTOM,CFACE2
*SURFACE INTERACTION, NAME=FLANGE
*FRICTION
.1,
*SURFACE BEHAVIOR,
.0005,1000.
*ELSET,ELSET=MA1,GENERATE
568,582,1
*ELSET,ELSET=SL1,GENERATE
77,122,1
*SURFACE,NAME=MASTER1
MA1,
*SURFACE,NAME=SLAVE1
SL1,
*CONTACT PAIR, INTERACTION=SELF1
SLAVE1,MASTER1
*SURFACE INTERACTION, NAME=SELF1
*FRICTION,SLIPTOLERANCE=.2
.1,
*SURFACE BEHAVIOR,
.005,1000.
307,315,1
443,451,1
1-256
MA2,
SL2,
SLAVE2,MASTER2
.1,
*SURFACE BEHAVIOR,
.0002,1000.
28,36,1
227,270,1
MA3,
SL3,
SLAVE3,MASTER3
*FRICTION
.1,
*SURFACE BEHAVIOR,
.0002,1000.
613,622,1
550,562,1
MA4,
SL4,
SLAVE4,MASTER4
.1,
*SURFACE BEHAVIOR,
.001,1000.
1-257
603,610,1
116,124,1
MA5,
SL5,
SLAVE5,MASTER5
*FRICTION
.1,
*SURFACE BEHAVIOR,
.001,1000.
*ELSET,ELSET=EALL
STEEL,RUBFIB
*SOLID SECTION,ELSET=RUBFIB,MATERIAL=RUBFIB
*MATERIAL,NAME=RUBFIB
*HYPERELASTIC,OGDEN,N=4,TEST DATA INPUT
*UNIAXIAL TEST DATA
-211.,-.4
-172.,-.35
-145.,-.30
-122.,-.25
-106.,-.20
-89.,-.15
-66.,-.1
-36.,-.05
0.,0.
132.,.05
231.,.1
297.,.15
363.,.2
594.,.25
924.,.3
1188.,.4
*SOLID SECTION,ELSET=STEEL,MATERIAL=STEEL
*MATERIAL,NAME=STEEL
*ELASTIC
3.0e+7,.3
*BOUNDARY
1-258
EDGE,1,1
1,2,2
*STEP,UNSYMM=YES,NLGEOM,INC=100
*STATIC
.0125,1.,1.E-8,.05
*BOUNDARY,OP=MOD
5001,1,,0.
5001,2,,-1.4
5001,6,,0.
5002,1,,0.
5002,2,,0.
5002,6,,0.
S,
U,
*NODE PRINT,FREQUENCY=1,NSET=BCYL
U1,U2,RF1,RF2
*NODE FILE,FREQUENCY=100
U,
*CONTACT PRINT,SLAVE=PPRES,MASTER=CFACE,FREQ=100
CSTRESS,
*CONTACT FILE,SLAVE=PPRES,MASTER=CFACE,FREQ=100
CSTRESS,
*CONTACT OUTPUT,SLAVE=PPRES,MASTER=CFACE
CSTRESS,
*CONTACT CONTROLS, FRICTION ONSET=DELAYED
*END STEP
*STEP,UNSYMM=YES,NLGEOM,INC=100
*STATIC
.0125,1.,1.E-8,.05
*DLOAD,OP=NEW
FACE1,P1,30.
FACE3,P3,30.
FACE4,P4,30.
*PRESSURE PENETRATION,MASTER=CFACE,SLAVE=PPRES
597,,30.0
S,
1-259
U,
*NODE PRINT,FREQUENCY=1,NSET=BCYL
U1,U2,RF1,RF2
S,
U,
*CONTACT PRINT,SLAVE=PPRES,MASTER=CFACE,FREQ=2
CSTRESS,PPRESS
*CONTACT FILE,SLAVE=PPRES,MASTER=CFACE,FREQ=2
CSTRESS,PPRESS
*CONTACT OUTPUT,SLAVE=PPRES,MASTER=CFACE
CSTRESS,PPRESS
*END STEP
1.1.18 Self-contact in rubber/foam components: jounce bumper

The self-contact capability in ABAQUS is illustrated with two examples derived from the automotive
component industry: this problem and the following one, which discusses a rubber gasket. These
examples demonstrate the use of the single-surface contact capability available for two-dimensional
large-sliding analysis. Components that deform and change their shape substantially can fold and have
different parts of the surface come into contact with each other. In such cases it can be difficult to
predict at the outset of the analysis where such contact may occur and, therefore, it can be difficult to
define two independent surfaces to make up a contact pair.
A jounce bumper, sometimes referred to as a "helper spring," is a highly compressible component that
is used as part of the shock isolation system in a vehicle. It is typically located above the coil spring
that connects the wheels to the frame. Microcellular material is used because of its high
compressibility and low Poisson's ratio value at all but fully compressed configurations.
The bumper is mounted in a mandrel with a diameter larger than the bumper's inner diameter ( Figure
1.1.18-1). The first step of the analysis solves this interference fit problem. The bumper initially sits
against a fixed flat rigid surface on one end; on the other end, another flat rigid surface is used to
model the compression of the bumper. The geometry of the bumper is such that it folds in three
different locations. Three separate surfaces, one for each such location, are defined in this example.
Each surface is allowed to contact itself. This modeling technique produces a more economical
analysis because the scope of contact searches and the wavefront of the equation system are
minimized.
Geometry and model
1-260
The bumper is 76.0 mm (3.0 in) long and has an inside diameter of 20.0 mm (.8 in). The mandrel,
which is modeled as a rigid surface, has a diameter of 22.0 mm (.9 in). The bumper is modeled with
the hyperfoam material model. The compressible, nonlinear elastic behavior is described by a strain
energy function. The irregular shape of the bumper makes use of an automatic triangular mesher
convenient. The model is discretized with first-order triangular axisymmetric elements. In addition to
the portions of the bumper's surface used to define self-contact, two additional regions are defined: one
to model contact with the fixed surface and the other to model contact with the mandrel and the
moving rigid surface. A small amount of friction (a Coulomb coefficient of 0.05) is applied to all of the
surfaces.

The bumper analysis is a two-step process. In the first step the interference with the mandrel is
resolved using the *CONTACT INTERFERENCE, SHRINK option (see ``Common difficulties
associated with contact modeling,'' Section 21.10.1 of the ABAQUS/Standard User's Manual): the
calculated initial penetration is allowed at the beginning of the step and scaled linearly to zero at the
end of the step (Figure 1.1.18-2). In the second step the bottom surface compresses the bumper 42.0
mm (1.7 in) as a result of the application of displacement boundary conditions to the reference node of
the surface (Figure 1.1.18-3). The high compressibility of the material is apparent, as well as the
folding of the surface onto itself. Although a general knowledge of where the folding would occur was
used in the definition of the self-contacting surfaces, it is not necessary to know exactly where the
kinks in the surface will form. The energy absorbing capacity of the device is seen through the load
versus displacement curve of the bottom surface (see Figure 1.1.18-4).
Input files
selfcontact_bump.inp
Jounce bumper model.
selfcontact_bump_node.inp
Node definitions for the bumper model.
selfcontact_bump_element.inp
Element definitions for the bumper model.
Figures
Figure 1.1.18-1 Jounce bumper initial mesh.
1-261
Figure 1.1.18-2 Bumper mesh after interference is resolved.
1-262
Figure 1.1.18-3 Bumper mesh after crushing.
1-263
Figure 1.1.18-4 Bumper load-displacement curve.
1-264
Sample listings
1-265
Listing 1.1.18-1
*HEADING
SELF CONTACT: JOUNCE BUMPER
**
*NODE, INPUT=selfcontact_bump_node.inp
*NODE, NSET=DIES
1000, 50., 0.
1001, 50., 76.5
1002, 10., 80.
**
*NSET, NSET=BOTDIE
1000,
*NSET, NSET=TOPDIE
1001,
*NSET, NSET=SHAFTDIE
1002,
**
*ELEMENT, TYPE=CAX3, ELSET=ALLFEM,
INPUT=selfcontact_bump_element.inp
**
** allfem
**
*SOLID SECTION, ELSET=ALLFEM, MATERIAL=HYPRFOAM
1.,
**
** hyprfoam
**
*MATERIAL, NAME=HYPRFOAM
**
*HYPERFOAM,N=1
3.0,11.5,.1
*******************************
**
** Contact definition
**
*******************************
*RIGID BODY,ANALYTICAL SURFACE=BOTDIE,
REF NODE=1000
*SURFACE,NAME=BOTDIE, TYPE=SEGMENTS
START, 0.00, 0.00
LINE, 50.00, 0.00
*RIGID BODY,ANALYTICAL SURFACE=TOPDIE,
1-266
REF NODE=1001
*SURFACE,NAME=TOPDIE, TYPE=SEGMENTS
START, 50.00, 76.50
LINE, 0.00, 76.50
*RIGID BODY,ANALYTICAL SURFACE=SHAFTDIE,
REF NODE=1002
*SURFACE,NAME=SHAFTDIE, TYPE=SEGMENTS
START, 11.00, 80.00
LINE, 11.00, -5.00
**
*ELSET, ELSET=BOTSKIN3, GEN
19, 61, 1
63, 91, 1
*ELSET, ELSET=BOTSKIN3
BOTSKIN3, 210
*ELSET, ELSET=BOTSKIN1, GEN
1321, 1328, 1
**
*SURFACE,NAME=BOTSKIN
BOTSKIN3, S3
BOTSKIN1, S1
62, S2
62, S1
**
*ELSET, ELSET=TOPSKIN3, GEN
126, 140, 1
*ELSET, ELSET=TOPSKIN1
141, 1315, 1316, 1318, 1319
**
*SURFACE,NAME=TOPSKIN
TOPSKIN3, S3
TOPSKIN1, S1
**
*ELSET, ELSET=SC1,GEN
1, 17, 1
**
*SURFACE,NAME=SC1
SC1, S3
**
114, 124, 1
**
*SURFACE,NAME=SC3
1-267
SC3, S3
**
93, 104, 1
**
*SURFACE,NAME=SC5
SC5, S3
**
*CONTACT PAIR,INTERACTION=INT1
TOPSKIN,SHAFTDIE
BOTSKIN,SHAFTDIE
TOPSKIN,TOPDIE
BOTSKIN,BOTDIE
SC1,SC1
SC3,SC3
SC5,SC5
*SURFACE INTERACTION,NAME=INT1
*FRICTION
.05,
*FRICTION
.05,
*FRICTION
.05,
*FRICTION
.05,
**
********************
**
**
********************
*BOUNDARY
BOTDIE, 1,6,0.
TOPDIE, 1,6,0.
1-268
SHAFTDIE, 1,6,0.
****************************************
**
** Step definition: Shrink fit
**
****************************************
*STEP, INC=200, NLGEOM
*STATIC
1.E-1, 1.0,
*CONTACT INTERFERENCE,SHRINK
TOPSKIN,SHAFTDIE
BOTSKIN,SHAFTDIE
*PRINT, CONTACT=YES
*CONTACT FILE,FREQ=999
*EL PRINT, ELSET=ALLFEM, FREQ=999
S,
E,
*NODE PRINT, FREQ=999
*NODE OUTPUT
U,RF
*ELEMENT OUTPUT,ELSET=ALLFEM
S,E
*CONTACT OUTPUT
CSTRESS,CDISP
*NODE OUTPUT,NSET=DIES
U,RF
*CONTACT CONTROLS, FRICTION ONSET=DELAYED
*END STEP
****************************************
**
** Step definition
**
****************************************
*STATIC
1.E-1, 1.0,1.e-5 , 0.1
*BOUNDARY,OP=MOD
BOTDIE,2, ,42.
*END STEP
1.1.19 Self-contact in rubber/foam components: rubber gasket
1-269
The self-contact capability in ABAQUS is illustrated with two examples derived from the automotive
component industry: this problem and the preceding one, which discusses a jounce bumper. These
examples demonstrate the use of the single-surface contact capability available for two-dimensional
large-sliding analysis. Components that deform and change their shape substantially can fold and have
different parts of the surface come into contact with each other. In such cases it can be difficult to
predict at the outset of the analysis where such contact may occur and, therefore, it can be difficult to
define two independent surfaces to make up a contact pair.
This model is used to analyze an oil pan gasket, which enhances the sealing of the oil pan against the
engine block. A primary objective of gasket designers is to reach or exceed a threshold value of contact
pressure at the gasket bead/cover/engine block interfaces. Experience shows that, above such a
threshold, oil will not leak. Another item of interest is the load-deflection curve obtained when
compressing the gasket cross-section since it is indicative of the bolt load required to attain a certain
gap between the oil pan and the engine block. Finally, the analysis provides details to ensure that
stresses and strains are within acceptable bounds.
The gasket is embedded in a plastic backbone. It has two planes of symmetry and a bead that, when
compressed, provides the sealing effect (Figure 1.1.19-1). A flat rigid surface, parallel to one of the
symmetry planes, pushes the gasket into the backbone. The geometry of the gasket is such that it folds
in two different locations. In this model the entire free surface of the gasket and of the backbone is
declared as a single surface allowed to contact itself. This modeling technique, although very simple, is
more expensive because of the extensive contact searches required, as well as a larger wavefront of the
equation system.
Geometry and model

The gasket is modeled as a quarter of a plane strain section, initially in contact with a flat rigid surface.
The clearance between the plastic backbone and the surface is 0.612 mm (.02 in). The height of the
rubber bead in the gasket is 1.097 mm (.04 in). The backbone is modeled with a linear elastic material
with a Young's modulus of 8000.0 GPa (1160 ksi) and a Poisson's ratio of 0.4. The gasket is modeled
as a fully incompressible hyperelastic material, which is much softer than the backbone material at all
strain levels. The nonlinear elastic behavior of the rubber is described by a strain energy function that
is a first-order polynomial in the strain invariants. The model is discretized with lower-order
quadrilaterals. Standard elements are used for the backbone, and hybrid elements are used for the
rubber. The interface between the rubber and the backbone is assumed to be glued with no special
treatment required. A single surface definition covers all of the free surface of the gasket and the
backbone. Through the definition of contact pairs, this surface is allowed to contact both the rigid
surface and itself. A small amount of friction (Coulomb coefficient of 0.05) is applied to the interface
with the rigid surface, which is assumed to be lubricated. Sticking friction, through the option
*FRICTION, ROUGH (`Ìnteraction tangential to the surface,'' Section 21.3.4 of the
ABAQUS/Standard User's Manual), is applied when the rubber contacts itself, denoting a clean
surface.
1-270
The gasket analysis is a single-step procedure in which the rigid surface moves down almost all of the
backbone clearance (0.61 mm or .02 in). The relative rigidity of the backbone forces the rubber gasket
to fit inside the cavity provided by the backbone, folding in two regions ( Figure 1.1.19-2). Although
the general vicinity of the location of the folds can be estimated from the initial configuration, their
exact locations are difficult to predict.
Acknowledgements
HKS would like to thank Mr. DeHerrera of Freudenberg-NOK General Partnership for providing these
examples.
Input files
selfcontact_gask.inp
Gasket model.
selfcontact_gask_node.inp
Node definitions for the gasket model.
selfcontact_gask_element1.inp
Element definitions for the rubber part of the gasket model.
selfcontact_gask_element2.inp
Element definitions for the backbone part of the gasket model.
Figures
Figure 1.1.19-1 Gasket initial mesh.
Figure 1.1.19-2 Gasket mesh after loading.
1-271
Sample listings
1-272
Listing 1.1.19-1
*HEADING
SELF CONTACT: GASKET
**
*NODE, INPUT=selfcontact_gask_node.inp, NSET=NSET1
*NODE, NSET=DIE
1001, 8.08006, 2.062
*NSET,NSET=NALL
NSET1, DIE
**
*NSET, NSET=BOTTOM
370, 404, 432, 460, 486, 508, 532, 561, 587, 609,
633, 657, 681, 694, 702, 703, 706, 707, 708, 709,
710, 711
**
*NSET, NSET=LEFT
126, 153, 186, 222, 256, 292, 330, 368, 369, 370
**
**
*ELEMENT, TYPE=CPE4H, ELSET=GASKET,
INPUT=selfcontact_gask_element1.inp
*ELEMENT, TYPE=CPE4, ELSET=BACKBONE,
INPUT=selfcontact_gask_element2.inp
**
*ELSET,ELSET=ALLFEM
GASKET,BACKBONE
**
** gasket
**
*SOLID SECTION, ELSET=GASKET, MATERIAL=ELSTMR01
1.,
**
** backbone
**
*SOLID SECTION, ELSET=BACKBONE, MATERIAL=PLASTIC
1.,
**
** symm
**
*BOUNDARY
LEFT , 1,, 0.
BOTTOM, 2,, 0.
1-273
**
** plastic
**
*MATERIAL, NAME=PLASTIC
**
*ELASTIC, TYPE=ISO
8000., 0.4
**
** Elastomer
**
*MATERIAL, NAME=ELSTMR01
**
*HYPERELASTIC, N=1
0.35, 0.25
*****************************
**
** Contact definition
**
*****************************
*RIGID BODY,ANALYTICAL SURFACE=DIE,REF NODE=1001
*SURFACE,NAME=DIE, TYPE=SEGMENTS
START, 8.080, 2.062
LINE, 0.000, 2.062
**
*ELSET, ELSET=SKIN4, GEN
1, 64, 1
626, 638, 1
**
*SURFACE,NAME=SKIN
SKIN4, S4
65, S3
**
SKIN,DIE
SKIN,SKIN
*FRICTION
.05,
*FRICTION, ROUGH
*CONTACT DAMPING, DEFINITION=DAMPING COEFFICIENT
1.0,0.02,
1-274
**
********************
**
**
********************
*BOUNDARY, OP=NEW
**
** topdie
**
DIE, 1,6, 0.
****************************************
**
** Step definition
**
****************************************
*STATIC
0.05 , 1.0, , 0.05
*CONTACT CONTROLS,FRICTION ONSET=IMMEDIATE
*BOUNDARY,OP=MOD
DIE,2, ,-0.61
*PRINT, CONTACT=YES
*CONTACT FILE,FREQ=999
*CONTACT PRINT,FREQ=999
*EL PRINT, ELSET=GASKET, FREQ=999
S,
E,
*NODE PRINT, FREQ=999
U,
*NODE OUTPUT,NSET=NALL
U,
*ELEMENT OUTPUT,ELSET=ALLFEM
S,
*CONTACT OUTPUT
CSTRESS,CDISP
*NODE OUTPUT,NSET=DIE
U,RF
*END STEP
1.1.20 Submodeling of a stacked sheet metal assembly
1-275
Sheet metal stampings stacked and fitted on top of each other and secured together via mechanical
fasteners such as bolts or rivets are commonly used in the automotive industry. Examples include seat
belt anchors and seating track assemblies. The submodeling capability in ABAQUS facilitates
economical, yet detailed, prediction of the ultimate strength and integrity of such jointed assemblies. A
global model analysis of an assembly is first performed to capture the overall deformation of the
system. Subsequently, the displacement results of this global analysis are used to drive the boundaries
of a submodeled region of critical concern. The submodeling methodology provides accurate modeling
that is more economical than using a globally refined mesh in a single analysis.
In a finite element analysis of such a structure, shell elements are commonly used to represent the sheet
metal stampings. The nodes of each shell typically lie along the mid-plane of the shell thickness. The
thickness of the shells is used in the structural calculations but is not taken into account in the contact
calculations. Hence, a structure composed of a stack-up of several sheet stampings may have the nodes
of each sheet all lying in the same spatial plane. This close proximity creates uncertainty in a submodel
analysis since ABAQUS will not be able to determine the correct correspondence between the sheets
in the submodel and the global model. Therefore, ABAQUS provides a capability that allows the user
to specify particular elements of the global model that are used to drive a particular set of nodes in a
submodel, which eliminates the uncertainty. This capability is demonstrated in this example problem.
Geometry and model

The global model consists of five separate metal stampings meshed with S4R and S3R shell elements.
An exploded view of the global finite element model is shown in Figure 1.1.20-1. The stampings are
stacked one upon the other by collapsing the configuration in the 3-direction. All the shell elements are
0.5 mm thick, with all nodes positioned at the mid-surface of each shell. The separate meshes are
connected together with BEAM-type MPCs between corresponding perimeter nodes on the large bolt
holes through each layer. The nodes on the edges of the two small holes at the bottom of Layer 1 are
constrained in all six degrees of freedom, representing the attachment point to ground. The
translational degrees of freedom of the nodes around the perimeter of Layer 2 are also constrained,
representing the far-field boundary condition in that plate.
Several surface definitions are used to model the contact between the various adjacent layers. The
contact definitions prevent unwanted penetration between shell element layers. The small-sliding
contact formulation is employed. Most of the contact in this problem is between adjacent layers, but
there is also direct contact between Layer 2 and Layer 4. To avoid overconstraints, it is important that
no point on Layer 4 simultaneously contact Layer 3 and Layer 2; therefore, node-based surfaces are
used for the slave surfaces. This precludes accurate calculation of contact stresses, but that is not
important in this case since more accurate contact stresses are obtained in the submodel.
All five stampings are made of steel and are modeled as an elastic-plastic material. The elastic
modulus is 207,000 MPa, Poisson's ratio is 0.3, and the yield stress is 250 MPa. The *PLASTIC
definition includes moderate strain hardening.
The submodel stampings are truncated versions of the global model, located in the same physical
position as the global model. In this case these are the regions of concern for high stresses and
1-276
potential failure of the joint. The submodel is discretized with a finer mesh than the global model to
provide a higher level of accuracy. Figure 1.1.20-2 shows an exploded view of the submodel. Because
the stampings in the submodel contain the large bolt holes, the submodel contains BEAM-type MPCs
in a manner analogous to that in the global model.
The submodel has several surface definitions and contact pairs to avoid penetration of one stamping
into another. The submodel contains no node-based surfaces, however. The contact is modeled as
element-based surface-to-surface in each layer.
The material definition and shell thicknesses in the submodel are the same as those in the global
model.

The global model is loaded by enforcing prescribed boundary conditions on the protruding edge of
Layer 3. This edge is displaced -5.0 mm in the 1-direction and -12.5 mm in the 3-direction. Figure
1.1.20-3 shows the deformed shape of the global model. The displacements at the nodes are saved to
the results file for later use by the submodel analysis.
The submodel driven nodes are loaded using the *BOUNDARY option with the SUBMODEL
parameter. The perimeter nodes of each layer of the submodel that correspond to a "cut" out of the
global geometry are driven by the interpolated nodal displacement results in the global results file.
Each driven node set is in a separate shell layer. Therefore, the submodel contains multiple
*SUBMODEL options, which designate the global model element sets to be searched for the responses
that drive the submodel driven node sets. For example, the driven nodes in submodel Layer 1 (node set
L1BC) are driven by the results for the global element set which contains the elements of (global)
Layer 1 (element set LAYER1). Similar commands exist for Layers 2-4. Because submodel Layer 5 has
no driven nodes, only four *SUBMODEL options are required.
*SUBMODEL, GLOBAL ELSET=LAYER1
L1BC,
L2BC,
L3BC,
L4BC ,
Figure 1.1.20-4 shows the deformed shape of the submodel. Figure 1.1.20-5 and Figure 1.1.20-6 show
contour plots of the out-of-plane displacements in Layer 2 for the global model and submodel,
respectively. In both cases the displacement patterns are similar; however, the maximum displacement
predicted by the global model is about 7.8% larger than that predicted by the submodel.
Input files
stackedassembly_s4r_global.inp
S4R global model.
1-277
stackedassembly_s4r_global_mesh.inp
Key input data for the S4R global model.
stackedassembly_s4r_sub.inp
S4R submodel.
stackedassembly_s4r_sub_mesh.inp
Key input data for the S4R submodel.
Figures
Figure 1.1.20-1 Exploded view of global model.
1-278
Figure 1.1.20-2 Exploded view of submodel.
Figure 1.1.20-3 Deformed shape of global model.
1-279
Figure 1.1.20-4 Deformed shape of submodel.
Figure 1.1.20-5 Out-of-plane displacement in Layer 2, global model.
1-280
Figure 1.1.20-6 Out-of-plane displacement in Layer 2, submodel.
Sample listings
1-281
Listing 1.1.20-1
*HEADING
GLOBAL MODEL FOR SUBMODELING EXAMPLE
STACKEDASSEMBLY_S4R_GLOBAL
MULTIPLE *SUBMODEL OPTIONS
*INCLUDE,INPUT=stackedassembly_s4r_global_mesh.inp
*SHELLSECTION,ELSET=LAYER1,MATERIAL=STEEL
0.5,
0.5,
0.5,
0.5,
0.5,
*DENSITY
7.8E-09,
*ELASTIC
207000.,0.28
*PLASTIC
250.0,0.0
420.0,0.2
*SURFACE, NAME=BLAYER2
LA2SPOS,SPOS
LA2SNEG,SNEG
*SURFACE, NAME=TLAYER2
LA2SPOS,SNEG
LA2SNEG,SPOS
LAYER3,SNEG
TLAYER4,SPOS
*SURFACE, TYPE=NODE, NAME=CNODE1
CNODE1,
CNODE3,
*SURFACE, TYPE=NODE, NAME=CNODE4A
CNODE4A,
*SURFACE, TYPE=NODE, NAME=CNODE4B
CNODE4B,
1-282

CNODE5,
*SURFACE INTERACTION,NAME=INTER1
*CONTACT PAIR,INTERACTION=INTER1,
SMALL SLIDING
CNODE1,BLAYER2
CNODE3,TLAYER2
CNODE4A,TLAYER3
CNODE4B,TLAYER2
CNODE5,TLAYER4
*BOUNDARY
BC1,1,6
BC3,1,6
*STATIC
0.025,1.0
*BOUNDARY
BC2,3,3,-12.5
BC2,1,1,-5.0
*ELPRINT,FREQ=0
*NODEPRINT,FREQ=0
*NODEFILE,NSET=ALLNO,FREQ=100
U,
*END STEP
1-283
Listing 1.1.20-2
*HEADING
SUBMODEL FOR SUBMODELING EXAMPLE
STACKEDASSEMBLY_S4R_SUB
MULTIPLE *SUBMODEL OPTIONS
*INCLUDE,INPUT=stackedassembly_s4r_sub_mesh.inp
0.5,
0.5,
0.5,
0.5,
0.5,
*ELASTIC
207000.0,0.3
*PLASTIC
250.0,0.0
420.0,0.2
OLAYER1,SNEG
LAYER2,SNEG
LAYER2,SPOS
LAYER3,SNEG
OLAYER3,SPOS
OLAYER4,SNEG
OLAYER4,SPOS
LAYER5,SPOS
*CONTACT PAIR,INTERACTION=ROUGH,
SMALL SLIDING
TLAYER1,BLAYER2
1-284
BLAYER3,TLAYER2
BLAYER4,TLAYER3
BLAYER5,TLAYER4
BLAYER4,TLAYER2
*SUBMODEL,ABSOLUTE EXTERIOR TOLERANCE=2.75,
GLOBAL ELSET=LAYER1
L1BC,
GLOBAL ELSET=LAYER2
L2BC,
GLOBAL ELSET=LAYER3
L3BC,
GLOBAL ELSET=LAYER4
L4BC,
*NSET,NSET=NLAYER2,ELSET=LAYER2
**
*STATIC
0.0025,1.0
*BOUNDARY,SUBMODEL,OP=NEW,STEP=1
L1BC,1,6,
L2BC,1,6,
L3BC,1,6,
L4BC,1,6,
*PRINT,CONTACT=YES
*ELPRINT,FREQ=0
*ELFILE, FREQ=99999, ELSET=LAYER2
S,
*NODEFILE, FREQ=99999, NSET=NLAYER2
U,
*NODEPRINT,FREQ=0
*END STEP
1.2 Buckling and collapse analyses

1.2.1 Snap-through buckling analysis of circular arches
It is often necessary to study the postbuckling behavior of a structure whose response is unstable
during part of its loading history. Two of the models in this example illustrate the use of the modified
Riks method, which is provided to handle such cases. The method is based on moving with fixed
1-285
increments along the static equilibrium path in a space defined by the displacements and a proportional
loading parameter. The actual load value may increase or decrease as the solution progresses. The
modified Riks method implemented in ABAQUS is described in ``Modified Riks algorithm,'' Section
2.3.2 of the ABAQUS Theory Manual.
The other two models illustrate the use of viscous damping. One example applies viscous damping as a
feature of surface contact, which allows for the definition of a "viscous" pressure that is proportional to
the relative velocity between the surfaces. The implementation of this option in ABAQUS is described
in ``Contact pressure definition,'' Section 5.2.1 of the ABAQUS Theory Manual. The other example
applies volume proportional damping to the model. The implementation of this option is described in
the automatic stabilization section of ``Solving nonlinear problems,'' Section 8.2.1 of the
ABAQUS/Standard User's Manual.
Three separate cases are considered here. The first is a clamped shallow arch subjected to a pressure
load. Reference solutions for this case are given by Ramm (1981) and Sharafi and Popov (1971). The
second case is the instability analysis of a clamped-hinged circular arch subjected to a point load. The
exact analytical solution for this problem is given by DaDeppo and Schmidt (1975). The third case is a
modification of the shallow arch problem in which the ends are pinned rather than clamped and the
arch is depressed with a rigid punch.
Model and solution control

The shallow circular arch is shown in Figure 1.2.1-1. Since the deformation is symmetric, one-half of
the arch is modeled. Ten elements of type B21 (linear interpolation beams) are used. A uniform
pressure is first applied to snap the arch through. The loading is then reversed so that the behavior is
also found as the pressure is removed.
The deep circular arch is shown in Figure 1.2.1-2. One end of the arch is clamped, and the other is
hinged. A concentrated load is applied at the apex of the arch. The arch undergoes extremely large
deflections but small strains. Because of the asymmetric boundary conditions, the arch will sway
toward the hinged end and then collapse. The arch is almost inextensible for most of the response
history. Sixty elements of type B31H are used. Hybrid elements are used because they are most
suitable for problems such as this.
The *CONTROLS option is used to set a very tight convergence tolerance because the problem
contains more than one equilibrium path. If tight tolerances are not used, the response might follow a
path that is different from the one shown.
In the Riks procedure actual values of load magnitudes cannot be specified. Instead, they are computed
as part of the solution, as the "load proportionality factor" multiplying the load magnitudes given on
the loading data lines. User-prescribed load magnitudes serve only to define the direction and to
estimate the magnitude of the initial increment of the load for a step. This initial load increment is the
product of the ratio of the initial time increment to the time period and the load magnitudes given in
the loading options. The user can terminate a Riks analysis by specifying either a maximum load
proportionality factor or a maximum displacement at a node, or both. When a solution point is
computed at which either of these limits is crossed, the analysis will stop. In any event, or if neither
option is used, the analysis ends when the maximum number of increments for the step is exceeded.
1-286
In snap-through studies such as these, the structure can carry increasing load after a complete snap.
Therefore, the analysis is terminated conveniently by specifying a maximum load proportionality
factor.
For the clamped shallow arch the initial snap occurs at a pressure of about -1000 (force/length 2 units).
Thus, -250 (force/length 2 units) seems to be a reasonable estimate for the first increment of load to be
applied. Accordingly, an initial time increment of 0.05 is specified for a time period of 1.0 and a
pressure load of -5000 (force/length 2 units). The solution will have been sufficiently developed at a
pressure of about -2000 (force/length 2 units). Therefore, the analysis is terminated when the load
proportionality factor exceeds 0.4.
To illustrate the use of Riks in several steps, a second step is included in which the pressure is taken
off the arch so that it will snap back toward its initial configuration. At any point in a Riks analysis, the
actual load is given by P = P0 + ¸(Pref ¡ P0 ) , where P0 is the load at the end of the previous step,
Pref is the load magnitude prescribed in the current step, and ¸ is the load proportionality factor. The
arch is unloaded so that in the initial time increment, a pressure of approximately 0.15 P0 is removed.
Using an initial time increment of 0.05 in a time period of 1.0, a load of Pref = ¡2P0 is prescribed for
this restarted step. Furthermore, we want the analysis to end when all the load is removed and the arch
has returned to its initial configuration. Therefore, a displacement threshold of 0.0 is set for the center
of the arch. The analysis terminates when this limit is crossed. Because ABAQUS must pick up the
load magnitude at the end of the initial Riks step to start the next step, any step following a Riks step
can be done only as a restart job, using the *RESTART option with the END STEP parameter.
For the deep clamped-hinged arch, the initial snap occurs at a load of about 900 (force units). The load
magnitude specified is 100 (force units), and the maximum load proportionality factor is specified as
9.5.
The shallow arch depressed with a rigid punch is shown in Figure 1.2.1-3. The analysis uses the same
model of the arch as the first problem. However, the end is pinned rather than clamped, and load is
applied through the displacement of the punch. The pinned boundary condition makes the problem
more unstable than the clamped-end case. A preliminary analysis in which the arch is depressed with a
prescribed displacement of the midpoint of the arch shows that the force will become negative during
snap-through. Thus, if the arch is depressed with a rigid punch, the Riks method will not help
convergence because, at the moment of snap-through, the arch separates from the punch, and the
movement of the punch no longer controls the displacement of the arch. Therefore, damping is
introduced to aid in convergence. Viscous damping with surface contact adds a pressure that is
proportional to the relative velocity to slow down the separation of the arch from the punch.
The viscous damping clearance is set to 10.0, and the fraction of the clearance interval is set to 0.9; the
damping is constant for a clearance of up to 9.0. Since the arch is 4.0 units high, the distance traveled
by the top of the arch from the initial position to the final snap-through position is 8.0 units. This
distance is clearly larger than the clearance between the middle of the arch and the tip of the punch at
any time during the analysis. Thus, the viscous damping is in effect for the whole period when the arch
has separated from the punch.
To choose the viscous damping coefficient, note that it is given as pressure per relative velocity. The
1-287
relevant pressure is obtained by dividing the approximate peak force (10000.0) by the contact area
(1.0). The relevant velocity is obtained by dividing the distance over which the top of the arch travels
(8.0 from initial to snapped position, which can be rounded to 10.0) by the time (approximately 1.0,
the total time of the step). A small percentage (0.1%) of this value is used for the viscous damping
coefficient:
Ft (10000:0)(1:0)
¹ = 0:001 = 0:001 = 1:0:
Al (1:0)(10:)
With ¹ = 1.0, the analysis runs to completion. Another analysis was run with a smaller value of ¹ =
0.1, but the viscous damping was not sufficient to enable the analysis to pass the point of
snap-through. Thus, a damping coefficient of 1.0 was determined to be an appropriate value.
Alternatively, including the STABILIZE parameter on the *STATIC analysis procedure option applies
volume proportional damping to the model. The default damping intensity is used in this case.

The results for the clamped shallow arch are shown in Figure 1.2.1-4, where the downward
displacement of the top of the arch is plotted as a function of the pressure. The algorithm obtains this
solution in 12 increments, with a maximum of three iterations in an increment. At the end of 12
increments the displacement of the top of the arch is about 7.5 length units. This represents a complete
snap through, as the original rise of the arch was 4 length units. Figure 1.2.1-5 and Figure 1.2.1-6show
a series of deformed configuration plots for this problem. Several other authors have examined this
same case and have obtained essentially the same solution (see Ramm, 1981, and Sharafi and Popov,
1971).
The results for the deep clamped-hinged arch are shown in Figure 1.2.1-7, where the displacement of
the top of the arch is plotted as a function of the applied load. Figure 1.2.1-8 shows a series of
deformed configuration plots for this problem. The arch collapses unstably at the peak load. Following
this, the beam stiffens rapidly as the load increases. The ability of the Riks method to handle unstable
response is well-illustrated by this example.
The results of the preliminary analysis of the prescribed displacement of a pinned shallow arch are
shown in Figure 1.2.1-9, with the displacement of the top of the arch plotted as a function of the
reaction force at that point. This plot shows the negative force that develops during snap-through. A
series of deformed configuration plots for the pinned shallow arch depressed with a punch and with
viscous damping introduced is shown in Figure 1.2.1-10, with one plot showing the arch separated
from the punch. Figure 1.2.1-11is a plot of the force between the punch and the top of the arch. The
force is positive until snap-through, when the arch separates from the punch and a negative viscous
force develops. Once the snap-through is complete, the force drops to zero as the punch continues to
move down while separated from the arch. When the punch contacts the arch, a positive force develops
again.
Similar results are produced when the contact viscous damping is replaced by volume proportional
damping. A sequence of configurations like Figure 1.2.1-10is obtained, in which separation of the arch
1-288
from the punch occurs during snap-through. At the end of the analysis the amount of energy dissipated
is similar to the amount dissipated with the viscous damping option.
Input files
snapbuckling_shallow_step1.inp
Initial analysis step for the shallow arch.
snapbuckling_shallow_unload.inp
Restart run to obtain the unloading response of the shallow arch.
snapbuckling_deep.inp
Deep arch.
snapbuckling_shallow_midpoint.inp
Shallow arch loaded by a fixed displacement of the midpoint.
snapbuckling_shallow_punch.inp
Shallow arch loaded by the displacement of a rigid punch.
snapbuckling_b21h_deep.inp
60 elements of type B21H used for the deep clamped-hinged arch analysis.
snapbuckling_b32h_deep.inp
30 elements of type B32H used for the deep clamped-hinged arch analysis.
snapbuckling_restart1.inp
Restart analysis of snapbuckling_shallow_step1.inp during the RIKS step.
snapbuckling_restart2.inp
Restart analysis of snapbuckling_restart1.inp during the RIKS step. This illustrates restarting an
existing RIKS restart analysis.
snapbuckling_shallow_stabilize.inp
Same as snapbuckling_shallow_punch.inp with the surface contact viscous damping replaced by
the volume proportional damping of *STATIC, STABILIZE.
References
· DaDeppo, D. A., and R. Schmidt, "Instability of Clamped-Hinged Circular Arches Subjected to a
Point Load," Transactions of the American Society of Mechanical Engineers , Journal of Applied
Mechanics, pp. 894-896, Dec. 1975.
· Ramm, E., "Strategies for Tracing the Nonlinear Response Near Limit Points," in Nonlinear
Finite Element Analysis in Structural Mechanics , edited by W. Wunderlich, E. Stein and K. J.
Bathe, Springer Verlag, Berlin, 1981.
1-289
· Sharifi, P., and E. P. Popov, "Nonlinear Buckling Analysis of Sandwich Arches," Proc. ASCE,
Journal of the Engineering Mechanics Division, vol. 97, pp. 1397-1412, 1971.
Figures
Figure 1.2.1-1 Clamped shallow circular arch.
Figure 1.2.1-2 Deep clamped-hinged arch.
Figure 1.2.1-3 Pinned shallow arch with rigid punch.
1-290
Figure 1.2.1-4 Load versus displacement curve for clamped shallow arch.
Figure 1.2.1-5 Deformed configuration plots for clamped shallow arch-Step 1.
1-291
Figure 1.2.1-6 Deformed configuration plots for clamped shallow arch-Step 2.
Figure 1.2.1-7 Load versus displacement curves for deep clamped-hinged arch.
Figure 1.2.1-8 Deformed configuration plots for deep clamped-hinged arch.
1-292
Figure 1.2.1-9 Force versus displacement curve for fixed displacement of pinned shallow arch.
Figure 1.2.1-10 Deformed configuration plots for pinned arch depressed with rigid punch.
1-293
Figure 1.2.1-11 Force between the punch and the top of the pinned arch.
Sample listings
1-294
Listing 1.2.1-1
*HEADING
SHALLOW ARCH
*NODE
1,0.,4.,0.,0.,1.
11,28.,0.,0.,.28,.96
*NGEN,LINE=C
1,11,1,0,0.,-96.
*ELEMENT,TYPE=B21
1,1,2
*ELGEN,ELSET=ARCH
1,10
*SURFACE, NAME=SURFACE
ARCH, SPOS
*BEAM SECTION,SECTION=RECT,ELSET=ARCH,MATERIAL=A1
1.,2.
*MATERIAL,NAME=A1
*ELASTIC
1.E7,.25
*BOUNDARY
1,1
1,6
11,1,6
*ELSET,ELSET=ONE
1,
*NSET,NSET=ONE
1,
*STEP,NLGEOM,INC=50
LOADING
*STATIC,RIKS
.05,1.,0.,.2,.4
*DSLOAD
SURFACE,P,5000.
*PRINT,RESIDUAL=NO
*EL PRINT,ELSET=ONE,SUMMARY=NO
E,
LOADS,
*NODE PRINT,SUMMARY=NO
U,RF
*EL FILE,ELSET=ONE
LOADS,
1-295
S,
E,
*NODE FILE,NSET=ONE
U,
*END STEP
1-296
Listing 1.2.1-2
*HEADING
UNLOADING OF SHALLOW ARCH
*RESTART, READ, STEP=1, INC=12, WRITE, FREQUENCY=0,
END STEP
*STEP,NLGEOM,INC=25
*STATIC,RIKS
.05,1.,,,,1,2,0.
*DSLOAD
SURFACE,P,-4685.6
*END STEP
1-297
Listing 1.2.1-3
*HEADING
DADEPPO'S ARCH
*NODE
1, 95.37169508,-30.07057995
61,-95.37169508,-30.07057995
*NGEN,LINE=C
1,61,1,,0.0,0.0,0.0, 0.0,0.0,1.0
*NSET,NSET=APEX
31 ,
*ELEMENT,TYPE=B31H
1,1,2
*ELGEN,ELSET=EALL
1,60,1,1
*ELSET,ELSET=CENT
30,31
*BEAM SECTION,SECTION=RECT,ELSET=EALL,
MATERIAL=MAT
1.0,2.289428295
*MATERIAL,NAME=MAT
*ELASTIC
1.E6,.0
*BOUNDARY
1,1,6
61,1,3
*STATIC,RIKS
0.40,1.0,,,9.5
*CONTROLS, PARAMETERS=FIELD
5.0E-6,
*CLOAD
31,2,-100.0
*EL PRINT,ELSET=CENT,FREQUENCY=5
SF,
SE,
*EL FILE,ELSET=CENT,FREQUENCY=10
SF,
SE,
*NODE FILE,NSET=APEX
U,
1-298
CF,
*NODE FILE
RF,
*END STEP
1-299
Listing 1.2.1-4
*HEADING
SHALLOW CLAMPED ARCH WITH DISPLACED MIDPOINT
*NODE
1,0.,4.,0.,0.,1.
11,28.,0.,0.,.28,.96
*NGEN,LINE=C
1,11,1,0,0.,-96.
*ELEMENT,TYPE=B21
1,1,2
*ELGEN,ELSET=ARCH
1,10
1.,2.
*MATERIAL,NAME=A1
*ELASTIC
1.E7,.25
*BOUNDARY
1,1
1,6
11,PINNED
*NSET,NSET=NENDS
1,11
*STEP,NLGEOM,INC=50
LOADING
*STATIC
.05,1., ,.05
*BOUNDARY,OP=MOD
1, 2,2, -9.0
*PRINT,RESIDUAL=NO
*NODE PRINT,SUMMARY=NO,FREQUENCY=2
U,RF
*NODE FILE,NSET=NENDS
U,RF
*END STEP
1-300
Listing 1.2.1-5
*HEADING
SHALLOW CLAMPED ARCH LOADED WITH PUNCH,
*NODE
1,0.,4.,0.,0.,1.
11,28.,0.,0.,.28,.96
100, 0., 9.
*NGEN,LINE=C
1,11,1,0,0.,-96.
*NSET,NSET=NENDS
1,11
*ELEMENT,TYPE=B21
1,1,2
*ELGEN,ELSET=ARCH
1,10
1.,2.
*MATERIAL,NAME=A1
*ELASTIC
1.E7,.25
*ELSET,ELSET=ET1
1,
*ELSET,ELSET=ET2,GENERATE
2,4,1
START, 5.,9.
CIRCL, 0.,4., 0.,9.
*SURFACE,NAME=ASURF
ET1,SPOS
*CONTACT PAIR,INTERACTION=DAMP
ASURF,BSURF
*SURFACE INTERACTION,NAME=DAMP
1.,
*CONTACT DAMPING, DEFINITION=DAMPING COEFFICIENT
1.0,10.0,0.9
*SURFACE,NAME=CSURF
ET2,SPOS
*CONTACT PAIR,INTERACTION=DAMP
CSURF,BSURF
*BOUNDARY
1,1
1-301
1,6
11,PINNED
100,1
100,6
LOADING
*STATIC
.05,1.,,.05
*BOUNDARY,OP=MOD
100,2,2,-9.0
*PRINT,RESIDUAL=NO
*CONTACT PRINT,SLAVE=ASURF,FREQUENCY=5
*CONTACT PRINT,SLAVE=CSURF,FREQUENCY=5
*CONTACT FILE,SLAVE=CSURF
S,E
*NODE PRINT,SUMMARY=NO,FREQUENCY=5
U,RF
*END STEP
1.2.2 Laminated composite shells: buckling of a cylindrical panel

with a circular hole
This example illustrates a type of analysis that is of interest in the aerospace industry. The objective is
to determine the strength of a thin, laminated composite shell, typical of shells used to form the outer
surfaces of aircraft fuselages and rocket motors. Such analyses are complicated by the fact that these
shells typically include local discontinuities--stiffeners and cutouts--which can induce substantial
stress concentrations that can delaminate the composite material. In the presence of buckling this
delamination can propagate through the structure to cause failure. In this example we study only the
geometrically nonlinear behavior of the shell: delamination or other section failures are not considered.
Some estimate of the possibility of material failure could presumably be made from the stresses
predicted in the analyses reported here, but no such assessment is included in this example.
The example makes extensive use of material orientation as part of the *SHELL GENERAL SECTION
option to define the multilayered, anisotropic, laminated section. The various orientation options for
shells are discussed in `Ànalysis of an anisotropic layered plate,'' Section 1.1.2 of the ABAQUS
Benchmarks Manual.
The *SHELL GENERAL SECTION option has two methods of defining laminated sections: defining
the thickness, material, and orientation of each layer or defining the equivalent section properties
1-302
directly. The last method is particularly useful if the laminate properties are obtained directly from
experiments or a separate preprocessor. This example uses both methods with the *SHELL GENERAL
SECTION option. Alternatively, the *SHELL SECTION option could be used to analyze the model;
however, because the material behavior is linear, no difference in solution would be obtained and the
computational costs would be greater.
Geometry and model

The structure analyzed is shown in Figure 1.2.2-1and was originally studied experimentally by Knight
and Starnes (1984). The test specimen is a cylindrical panel with a 355.6 mm (14 in) square platform
and a 381 mm (15 in) radius of curvature, so that the panel covers a 55.6° arc of the cylinder. The
panel contains a centrally located hole of 50.8 mm (2 in) diameter. The shell consists of 16 layers of
unidirectional graphite fibers in an epoxy resin. Each layer is 0.142 mm (.0056 in) thick. The layers are
arranged in the symmetric stacking sequence {§45/90/0/0/90/¨45} degrees repeated twice. The
nominal orthotropic elastic material properties as defined by Stanley (1985) are
E11 = 135 kN/mm2 (19.6 ´ 106 lb/in2),
2
E22 = 13 kN/mm (1.89 ´ 106 lb/in2),
G12 = G13 = 6.4 kN/mm2 (.93 ´ 106 lb/in2),
G23 = 4.3 kN/mm2 (0.63 ´ 106 lb/in2),
º12 = 0.38,
where the 1-direction is along the fibers, the 2-direction is transverse to the fibers in the surface of the
lamina, and the 3-direction is normal to the lamina.
The panel is fully clamped on the bottom edge, clamped except for axial motion on the top edge and
simply supported along its vertical edges. Three analyses are considered. The first is a linear
(prebuckling) analysis in which the panel is subjected to a uniform end shortening of 0.8 mm (.0316
in). The total axial force and the distribution of axial force along the midsection are used to compare
the results with those obtained by Stanley (1985). The second analysis consists of an eigenvalue
extraction of the first five buckling modes. The buckling loads and mode shapes are also compared
with those presented by Stanley (1985). Finally, a nonlinear load-deflection analysis is done to predict
the postbuckling behavior, using the modified Riks algorithm. For this analysis an initial imperfection
is introduced. The imperfection is based on the fourth buckling mode extracted during the second
analysis. These results are compared with those of Stanley (1985) and with the experimental
measurements of Knight and Starnes (1984).
The mesh used in ABAQUS is shown in Figure 1.2.2-2. The anisotropic material behavior precludes
any symmetry assumptions, hence the entire panel is modeled. The same mesh is used with the 4-node
shell element (type S4R5) and also with the 9-node shell element (type S9R5); the 9-node element
mesh, thus, has about four times the number of degrees of freedom as the 4-node element mesh. The
6-node triangular shell element STRI65 is also used; it employs two triangles for each quadrilateral
element of the second-order mesh. Mesh generation is facilitated by using the *NFILL and *NMAP
options, as shown in the input data. In this model specification of the relative angle of orientation to
define the material orientation within each layer, along with the *ELASTIC, TYPE=LAMINA option,
makes the definition of the laminae properties straightforward.
The shell elements used in this example use an approximation to thin shell theory, based on a
1-303
numerical penalty applied to the transverse shear strain along the element edges. These elements are
not universally applicable to the analysis of composites since transverse shear effects can be significant
in such cases and these elements are not designed to model them accurately. Here, however, the
geometry of the panel is that of a thin shell; and the symmetrical lay-up, along with the relatively large
number of laminae, tends to diminish the importance of transverse shear deformation on the response.
Relation between stress resultants and generalized strains

The shell section is most easily defined by giving the layer thickness, material, and orientation, in
which case ABAQUS preintegrates to obtain the section stiffness properties. However, the user can
choose to input the section stiffness properties directly instead, as follows.
In ABAQUS a lamina is considered as an orthotropic sheet in plane stress. The principal material axes
of the lamina (see Figure 1.2.2-3) are longitudinal, denoted by L; transverse to the fiber direction in the
surface of the lamina, denoted by T ; and normal to the lamina surface, denoted by N: The constitutive
relations for a general orthotropic material in the principal directions ( L; T; N ) are
8 9 2 38 9
> ¾L > C11 C12 0 C14 0 0 > ²L >
>
> >
> >
> >
>
>
> ¾ >
> 6 C12 C22 0 C24 0 0 7> ² >
< T = 6 7>< T >
=
¿LT 6 0 0 C33 0 0 0 7 °LT
=6 7 :
>
> ¾N >
> 6 C14 C24 0 C44 0 0 7> ²N >
>
> >
> 4 5>>
>
>
>
>
>
: ¿LN >
; 0 0 0 0 C55 0 >
: °LN >
;
¿T N 0 0 0 0 0 C66 °T N
In terms of the data required by the *ELASTIC, TYPE=ORTHO option in ABAQUS these are
C11 = D1111 ; C12 = D1122 ; C14 = D1133 ; C22 = D2222
C24 = D2233 ; C33 = D1212 ; C44 = D3333 ; C55 = D1313

C66 = D2323
This matrix is symmetric and has nine independent constants. If we assume a state of plane stress, then
¾N is taken to be zero. This yields
8 9 2 38 9
> ¾L > Q11 Q12 0 0 0 > ²L >
>
> >
> >
> >
>
< ¾T = 6 Q12 Q22 0 0 0 7 < ²T =
6 7
¿LT =6 0 0 Q33 0 0 7 °LT ;
>
> >
> 4 5>> >
>
>
: ¿LN >
; 0 0 0 Q55 0 >
: °LN >
;
¿T N 0 0 0 0 Q66 °T N
where
1-304
C14 C14
Q11 = C11 ¡ ;
C44
C24 C14
Q12 = C12 ¡ ;
C44
C24 C24
Q22 = C22 ¡ ;
C44
Q33 = C33 ;
Q55 = C55 ;
Q66 = C66 :
The correspondence between these terms and the usual engineering constants that might be given for a
simple orthotropic layer in a laminate is
E1
Q11 = ;
1 ¡ º12 º21
º12 E2 º21 E1
Q12 = = ;
1 ¡ º12 º21 1 ¡ º12 º21
E2
Q22 = ;
1 ¡ º12 º21
Q33 = G12 ;
Q55 = G13 ;
Q66 = G23 :
The parameters used on the right-hand side of the above equation are those that must be provided as
data on the *ELASTIC, TYPE=LAMINA option.
If the (1; 2; N ) system denotes the standard shell basis directions that ABAQUS chooses by default,
the local stiffness components must be rotated to this system to construct the lamina's contribution to
the *SHELL GENERAL SECTION stiffness. Since Qij represent fourth-order tensors, in the case of a
lamina they are oriented at an angle µ to the standard shell basis directions used in ABAQUS. Hence,
the transformation is
¹ 11 = Q11 cos4 µ + 2(Q12 + 2Q33 ) sin2 µ cos2 µ + Q22 sin4 µ;

Q
¹ 12 = (Q11 + Q22 ¡ 4Q33 ) sin2 µ cos2 µ + Q12 (sin4 µ + cos4 µ);
Q
¹ 22 = Q11 sin4 µ + 2(Q12 + 2Q33 ) sin2 µ cos2 µ + Q22 cos4 µ;
Q
¹ 13 = (Q11 ¡ Q12 ¡ 2Q33 ) sin µ cos3 µ + (Q12 ¡ Q22 + 2Q33 ) sin3 µ cos µ;
Q
¹ 23 = (Q11 ¡ Q12 ¡ 2Q33 ) sin3 µ cos µ + (Q12 ¡ Q22 + 2Q33 ) sin µ cos3 µ;
Q
¹ 33 = (Q11 + Q22 ¡ 2Q12 ¡ 2Q33 ) sin2 µ cos2 µ + Q33 (sin4 µ + cos4 µ);
Q
¹ 55 = Q55 cos2 µ ¡ Q66 sin2 µ;
Q
¹ 56 = Q55 sin µ cos µ ¡ Q66 sin µ cos µ;
Q
¹ 66 = Q55 sin2 µ ¡ Q66 cos2 µ;
Q
1-305
¹ ij are the stiffness coefficients in the standard shell basis directions used by ABAQUS.
where Q
ABAQUS assumes that a laminate is a stack of laminae arranged with the principal directions of each
layer in different orientations. The various layers are assumed to be rigidly bonded together. The
section force and moment resultants per unit length in the normal basis directions in a given layer can
be defined on this basis as
Z h=2
(N1 ; N2 ; N12 ) = (¾1 ; ¾2 ; ¿12 ) dz;
¡h=2
Z h=2
(M1 ; M2 ; M12 ) = (¾1 ; ¾2 ; ¿12 )z dz;
¡h=2
Z h=2
(V1 ; V2 ) = (¿13 ; ¿23 ) dz;
¡h=2
where h is the thickness of the layer.

This leads to the relations
8 9 2 38 9
> N1 > A11 A12 A13 B11 B12 B13 0 0 > ²1 >
>
> >
> > >
>
> N >
2 > 6 A A22 A23 B12 B22 B23 0 0 7> > ²2 > >
7> >
12
>
> >
> 6 >
> >
>
>
> N12 >
> 6 A A23 A33 B13 B23 B33 0 0 7> > ° >
>
< = 6 13 7< 12
=
M1 6B B12 B13 D11 D12 D13 0 0 7 ∙1
= 6 11 7 ;
>
> M2 >> 6 B12 B22 B23 D12 D22 D23 0 0 7> > ∙2 > >
>
> >
> 6 7> >
>
> >
M12 > 6 B13 B23 B33 D13 D23 D33 0 0 7> >
> ∙12 >>
>
>
> >
> 4 0 5>> >
>
: V1 >
> ; 0 0 0 0 0 E11 E12 > : 13 >
° ;
V2 0 0 0 0 0 0 E12 E22 °23
where the components of this section stiffness matrix are given by

Z h=2
(Aij ; Bij ; Dij ) = ¹m
Q 2
ij (1; z; z ) dz; (i; j = 1; 2; 3)
¡h=2
Z h=2
Eij = ¹m
Q®¯ ki kj dz; (i; j = 1; 2; and ®; ¯ = i + 4; j + 4):
¡h=2
Here m indicates a particular layer. Thus, the Q¹ m depend on the material properties and fiber
ij
orientation of the mth layer. The ki ; i = 1,2 parameters are the shear correction coefficients as defined
by Whitney (1973). If there are n layers in the lay-up, we can rewrite the above equations as a
summation of integrals over the n laminae. The material coefficients will then take the form
1-306
n
X
Aij = ¹m
Qij (hm ¡ hm¡1 );
m=1
Xn
1 ¹ m (h2 ¡ h2m¡1 );
Bij = Q
2 m=1 ij m
n
1 X ¹m 3
Dij = Q (h ¡ h3m¡1 );
3 m=1 ij m
n
X
Eij = ¹m
Q®¯ (hm ¡ hm¡1 )ki kj ;
m=1
where the hm and hm¡1 in these equations indicate that the mth lamina is bounded by surfaces
z = hm and z = hm¡1 : See Figure 1.2.2-4for the nomenclature.
These equations define the coefficients required for the direct input of the section stiffness matrix
method with the *SHELL GENERAL SECTION option. Only the [A], [B ], and [D ] submatrices are
needed for that option. The three terms in [E ], if required, are defined using the *TRANSVERSE
SHEAR STIFFNESS option. The section forces as defined above are in the normal shell basis
directions.
Applying these equations to the laminate defined for this example leads to the following overall
section stiffness:
2 3 2 3
138:385 44:0189 0 0 0 0
4
[A] = 44:0189 138:385 0 5 kN/mm; 4
[B] = 0 0 05;
0 0 47:1831 0 0 0
2 3
55:670 21:638 2:138 ∙ ¸
12:2387 0
[D ] = 4 21:638 58:521 2:138 5 kN-mm; [E ] = kN/mm,
0 12:2387
2:138 2:138 23:004
or
2 3 2 3
790:239 251:367 0 0 0 0
4
[A] = 251:367 790:239 0 5 £ 103 lb/in; 4
[B] = 0 0 05;
0 0 269:436 0 0 0
2 3
492:719 191:513 18:9245 ∙ ¸
4 49:573 0:002
[D ] = 191:513 517:951 18:9245 5 lb-in; [E ] = £ 103 lb/in :
0:002 52:967
18:9245 18:9245 203:602
1-307
The total axial force necessary to compress the panel 0.803 mm (0.0316 in) is 100.2 kN (22529 lb) for
the mesh of S9R5 elements, 99.5 kN (22359 lb) for the mesh of S4R5 elements, and 100.3 kN (22547
lb) for the mesh of STRI65 elements. These values match closely with the result of 100 kN (22480 lb)
reported by Stanley (1985). Figure 1.2.2-5 shows the displaced configuration and a profile of axial
force along the midsection of the panel (at z = L=2). It is interesting to note that the axial load is
distributed almost evenly across the entire panel, with only a very localized area near the hole
subjected to an amplified stress level. This suggests that adequate results for this linear analysis could
also be obtained with a coarser mesh that has a bias toward the hole.
The second stage of the analysis is the eigenvalue buckling prediction. To obtain the buckling
predictions with ABAQUS, a *BUCKLE step is run. In this step nominal values of load are applied.
The magnitude that is used is not of any significance, since eigenvalue buckling is a linear perturbation
procedure: the stiffness matrix and the stress stiffening matrix are evaluated at the beginning of the
step without any of this load applied. The *BUCKLE step calculates the eigenvalues that, multiplied
with the applied load and added to any "base state" loading, are the predicted buckling loads. The
eigenvectors associated with the eigenvalues are also obtained. This procedure is described in more
detail in `Èigenvalue buckling prediction,'' Section 6.2.3 of the ABAQUS/Standard User's Manual.
The buckling predictions are summarized in Table 1.2.2-1and Figure 1.2.2-6. The buckling load
predictions from ABAQUS are higher than those reported by Stanley. The eigenmode predictions
given by the mesh using element types S4R5, S9R5, and STRI65 are all the same and agree well with
those reported by Stanley. Stanley makes several important observations that remain valid for the
ABAQUS results: (1) the eigenvalues are closely spaced; (2) nevertheless, the mode shapes vary
significantly in character; (3) the first buckling mode bears the most similarity to the linear prebuckling
solution; (4) there is no symmetry available that can be utilized for computational efficiency.
Following the eigenvalue buckling analyses, nonlinear postbuckling analysis is carried out by imposing
an imperfection based on the fourth buckling mode. The maximum initial perturbation is 10% of the
thickness of the shell. The load versus normalized displacement plots for the S9R5 mesh, the S4R5
mesh, and the STRI65 mesh are compared with the experimental results and those given by Stanley in
Figure 1.2.2-7. The overall response prediction is quite similar for the ABAQUS elements, although
the general behavior predicted by Stanley is somewhat different. The ABAQUS results show a peak
load slightly above the buckling load predicted by the eigenvalue extraction, while Stanley's results
show a significantly lower peak load. In addition, the ABAQUS results show rather less loss of
strength after the initial peak, followed quite soon by positive stiffness again. Neither the ABAQUS
results nor Stanley's results agree closely with the experimentally observed dramatic loss of strength
after peak load. Stanley ascribes this to material failure (presumably delamination), which is not
modeled in his analyses or in these.
Figure 1.2.2-8 shows the deformed configurations for the panel during its postbuckling response. The
plots show the results for S4R5, but the pattern is similar for S9R5 and STRI65. The response is quite
symmetric initially; but, as the critical load is approached, a nonsymmetric dimple develops and grows,
presumably accounting for the panel's loss of strength. Later in the postbuckling response another
wrinkle can be seen to be developing.
Input files
1-308
laminpanel_s9r5_prebuckle.inp
Prebuckling analysis for the 9-node (element type S9R5) mesh.
laminpanel_s9r5_buckle.inp
Corresponding eigenvalue buckling prediction.
laminpanel_s9r5_postbuckle.inp
Corresponding nonlinear postbuckling analysis.
Eigenvalue buckling prediction with direct input of shell section stiffness properties using the
*SHELL GENERAL SECTION option.
Nonlinear postbuckling analysis with direct input of shell section stiffness properties using the
*SHELL GENERAL SECTION option.
laminpanel_s4r5_prebuckle.inp
Prebuckling analysis using element type S4R5.
Eigenvalue buckling prediction using element type S4R5.
Nonlinear postbuckling analysis using element type S4R5.
laminpanel_s4r5_node.inp
Nodal coordinate data for the imperfection imposed for the postbuckling analysis using element
type S4R5.
laminpanel_s9r5_stri65_node.inp
Nodal coordinate data for the imperfection imposed for the postbuckling analysis using element
types S9R5 and STRI65.
laminpanel_stri65_prebuckle.inp
Prebuckling analysis using element type STRI65.
laminpanel_stri65_buckle.inp
Eigenvalue buckling prediction using element type STRI65.
laminpanel_stri65_postbuckle.inp
Nonlinear postbuckling analysis using element type STRI65.
laminpanel_s4_prebuckle.inp
Prebuckling analysis using element type S4.
1-309
laminpanel_s4_buckle.inp
Eigenvalue buckling prediction using element type S4.
laminpanel_s4_postbuckle.inp
Nonlinear postbuckling analysis using element type S4.
References
· Knight, N. F., and J. H. Starnes Jr., "Postbuckling Behavior of Axially Compressed
Graphite-Epoxy Cylindrical Panels with Circular Holes," presented at the 1984 ASME Joint
Pressure Vessels and Piping/Applied Mechanics Conference, San Antonio, Texas, 1984.
· Stanley, G.M., Continuum-Based Shell Elements, Ph.D. Dissertation, Department of Mechanical

Engineering, Stanford University, 1985.
· Whitney, J.M., "Shear Correction Factors for Orthotropic Laminates Under Static Loads,"
Journal of Applied Mechanics, Transactions of the ASME, vol. 40, pp. 302-304, 1973.
Table
Table 1.2.2-1 Summary of buckling load predictions.

Stanley 107.0 kN (24054 lb)
Mode 1 S9R5 113.4 kN (25503 lb)
S4R5 115.5 kN (25964 lb)
S4 118.5 kN (26651 lb)
STRI65 113.8 kN (25579 lb)
Mode 2 S9R5 117.6 kN (26427 lb)
S4R5 121.2 kN (27244 lb)
S4 122.6 kN (27560 lb)
STRI65 117.8 kN (26490 lb)
Mode 3 S9R5 120.3 kN (27051 lb)
S4R5 124.7 kN (28042 lb)
S4 127.7 kN (28713 lb)
STRI65 121.1 kN (27218 lb)
Mode 4 S9R5 147.5 kN (33161 lb)
S4R5 156.1 kN (35092 lb)
S4 157.8 kN (35478 lb)
STRI65 146.9 kN (33015 lb)
Mode 5 S9R5 171.3 kN (38510 lb)
S4R5 181.5 kN (40800 lb)
S4 186.8 kN (41992 lb)
STRI65 172.8 kN (38842 lb)
1-310
Figures
Figure 1.2.2-1 Geometry for cylindrical panel with hole.
Figure 1.2.2-2 Mesh for cylindrical panel with hole.
1-311
Figure 1.2.2-3 Typical lamina.
1-312
Figure 1.2.2-4 Typical laminate.
1-313
Figure 1.2.2-5 Displaced shape and axial force distribution.
1-314
Figure 1.2.2-6 Buckling modes, element types S4R5, S9R5, and STRI65.
1-315
Figure 1.2.2-7 Load-displacement response.
1-316
Figure 1.2.2-8 Postbuckling deformations: 10% h imperfection with S4R5.
1-317
Sample listings
1-318
Listing 1.2.2-1
*HEADING
COMPOSITE CYLINDRICAL PANEL WITH CIRCULAR HOLE-
S9R5
** FROM STUDY IN THESIS OF G.M. STANLEY
*NODE
2001,15., 0., 0.
2017,15., 14.556, 0.
2033,15., 14.556, 14.
2049,15., 0. , 14.
2065,15., 0., 0.
*NGEN,NSET=OUTSIDE
2001,2017
2017,2033
2033,2049
2049,2065
** DEFINE CIRCLE
*SYSTEM
15.,7.27802298,7.0
*NODE
9999, 0.,0.
1,0., -.70710678, -.70710678
17,0., .70710678,-.70710678
33,0., .70710678, .70710678
49,0., -.70710678, .70710678
65,0., -.70710678, -.70710678
*NGEN,LINE=C,NSET=HOLE
1,17,1,9999
17,33,1,9999
33,49,1,9999
49,65,1,9999
*SYSTEM
0.,0.,0.
*NFILL,NSET=ALL
HOLE,OUTSIDE,20,100
*NSET,NSET=XPARB,GENERATE
2001,2017
*NSET,NSET=XPART,GENERATE
2033,2049
*NSET,NSET=YPAR,GENERATE
2049,2065
1-319
2017,2033
*NSET,NSET=PROFILE,GENERATE
57,2057,100
25,2025,100
*NMAP,TYPE=CYLINDRICAL,NSET=ALL
0.,0.,0., 0.,0.,1.
15.,0.,0.
1.,3.8197307 ,1.
*ELEMENT,TYPE=S9R5
1,1,201,203,3,101,202,103,2,102
*ELGEN,ELSET=ALL
1,31,2,1,10,200,100
*ELEMENT,TYPE=S9R5
32,63,263,201,1,163,264,101,64,164
*ELGEN,ELSET=ALL
32,10,200,100
*ELSET,ELSET=PROFILE,GENERATE
12, 912,100
13, 913,100
28, 928,100
29, 929,100
*SHELL GENERAL SECTION,ELSET=ALL,COMPOSITE,
ORIENTATION=SECORI
.0056,,LAMINA, 45.
.0056,,LAMINA,-45.
.0056,,LAMINA, 90.
.0056,,LAMINA, 0.
.0056,,LAMINA, 0.
.0056,,LAMINA, 90.
.0056,,LAMINA,-45.
.0056,,LAMINA, 45.
*** CENTER LINE
.0056,,LAMINA, 45.
.0056,,LAMINA,-45.
.0056,,LAMINA, 90.
.0056,,LAMINA, 0.
.0056,,LAMINA, 0.
.0056,,LAMINA, 90.
.0056,,LAMINA,-45.
.0056,,LAMINA, 45.
*MATERIAL,NAME=LAMINA
*ELASTIC,TYPE=LAMINA
19.6E6, 1.89E6, .38, .93E6, .93E6, .63E6
1-320
*ORIENTATION,SYSTEM=CYLINDRICAL,NAME=SECORI
0.,0.,0., 0.,0., 1.
1, 0.
*STEP
*STATIC
*BOUNDARY
XPARB,1,6
XPART,1,2
XPART,4,6
YPAR,1,2
XPART,3,,-.0316
*EL PRINT,ELSET=PROFILE,POSITION=AVERAGED AT NODES
SF,
*EL FILE,ELSET=PROFILE,POSITION=AVERAGED AT NODES
SF,
*NODE PRINT,NSET=XPART,TOTALS=YES
U,
RF,
*NODE PRINT,NSET=XPARB,TOTALS=YES
U,
RF,
*NODE FILE,NSET=XPART
U,RF
*END STEP
1-321
Listing 1.2.2-2
*HEADING
COMPOSITE CYLINDRICAL PANEL WITH CIRCULAR HOLE-- S9R5
**
**
*PREPRINT,ECHO=YES,MODEL=NO,HISTORY=NO
*NODE
2001,15., 0., 0.
2017,15., 14.556, 0.
2033,15., 14.556, 14.
2049,15., 0. , 14.
2065,15., 0., 0.
*NGEN,NSET=OUTSIDE
2001,2017
2017,2033
2033,2049
2049,2065
**
** DEFINE CIRCLE
**
*SYSTEM
15.,7.27802298,7.0
*NODE
9999, 0.,0.
1,0., -.70710678, -.70710678
17,0., .70710678,-.70710678
33,0., .70710678, .70710678
49,0., -.70710678, .70710678
65,0., -.70710678, -.70710678
*NGEN,LINE=C,NSET=HOLE
1,17,1,9999
17,33,1,9999
33,49,1,9999
49,65,1,9999
*SYSTEM
0.,0.,0.
*NFILL,NSET=ALL
HOLE,OUTSIDE,20,100
2001,2017
1-322
2033,2049
2049,2065
2017,2033
57,2057,100
25,2025,100
*NMAP,TYPE=CYLINDRICAL,NSET=ALL
0.,0.,0., 0.,0.,1.
15.,0.,0.
1.,3.8197307 ,1.
*EQUATION
2,
2033,3,1.,2041,3,-1.
2,
2034,3,1.,2041,3,-1.
2,
2035,3,1.,2041,3,-1.
2,
2036,3,1.,2041,3,-1.
2,
2037,3,1.,2041,3,-1.
2,
2038,3,1.,2041,3,-1.
2,
2039,3,1.,2041,3,-1.
2,
2040,3,1.,2041,3,-1.
2,
2042,3,1.,2041,3,-1.
2,
2043,3,1.,2041,3,-1.
2,
2044,3,1.,2041,3,-1.
2,
2045,3,1.,2041,3,-1.
2,
2046,3,1.,2041,3,-1.
2,
2047,3,1.,2041,3,-1.
2,
2048,3,1.,2041,3,-1.
1-323
2,
2049,3,1.,2041,3,-1.
*ELEMENT,TYPE=S9R5
1,1,201,203,3,101,202,103,2,102
*ELGEN,ELSET=ALL
1,31,2,1,10,200,100
*ELEMENT,TYPE=S9R5
32,63,263,201,1,163,264,101,64,164
*ELGEN,ELSET=ALL
32,10,200,100
12, 912,100
13, 913,100
28, 928,100
29, 929,100
*SHELL GENERAL SECTION,ELSET=ALL
7.90239D5,2.51367D5,7.90239D5,-3.08578D-6,-7.94285D-5,2.69436D5,0.,0.
0., 4.92719D+02,0.,0.,0.,1.91513D+02,5.17951D+02,0.
0.,0., 1.89245D+01 , 1.89245D+01 ,2.03602D+02
*BOUNDARY
XPARB,1,6
XPART,1,2
XPART,4,6
YPAR,1,2
*STEP
*BUCKLE
5,,60,20
*CLOAD
2041,3,-1000.
*MODAL FILE
*MODAL OUTPUT
*END STEP
1-324
Listing 1.2.2-3
*HEADING
NONLINEAR POSTBUCKLING ANALYSIS--10% IMPERFECTION-
WITH CONTRIBUTIONS FROM FIRST FOUR MODES
COMPOSITE CYLINDRICAL PANEL WITH CIRCULAR HOLE
**
**
*NODE,INPUT=PANEL9N.NOD
*NSET,NSET=OUTSIDE,GENERATE
2001,2017
2017,2033
2033,2049
2049,2065
*NSET,NSET=HOLE,GENERATE
1,17,1
17,33,1
33,49,1
49,65,1
2001,2017
2033,2049
2049,2065
2017,2033
57,2057,100
25,2025,100
*EQUATION
2,
2033,3,1.,2041,3,-1.
2,
2034,3,1.,2041,3,-1.
2,
2035,3,1.,2041,3,-1.
2,
2036,3,1.,2041,3,-1.
2,
2037,3,1.,2041,3,-1.
2,
2038,3,1.,2041,3,-1.
1-325
2,
2039,3,1.,2041,3,-1.
2,
2040,3,1.,2041,3,-1.
2,
2042,3,1.,2041,3,-1.
2,
2043,3,1.,2041,3,-1.
2,
2044,3,1.,2041,3,-1.
2,
2045,3,1.,2041,3,-1.
2,
2046,3,1.,2041,3,-1.
2,
2047,3,1.,2041,3,-1.
2,
2048,3,1.,2041,3,-1.
2,
2049,3,1.,2041,3,-1.
*NSET,NSET=MASTER
2041,
*ELEMENT,TYPE=S9R5
1,1,201,203,3,101,202,103,2,102
*ELGEN,ELSET=ALL
1,31,2,1,10,200,100
*ELEMENT,TYPE=S9R5
32,63,263,201,1,163,264,101,64,164
*ELGEN,ELSET=ALL
32,10,200,100
12, 912,100
13, 913,100
28, 928,100
29, 929,100
*SHELL GENERAL SECTION,ELSET=ALL
7.90239D5,2.51367D5,7.90239D5,-3.08578D-6,-7.94285D-5,2.69436D5,0.,0.
0., 4.92719D+02,0.,0.,0.,1.91513D+02,5.17951D+02,0.
0.,0., 1.89245D+01 , 1.89245D+01 ,2.03602D+02
*BOUNDARY
XPARB,1,6
XPART,1,2
XPART,4,6
1-326
YPAR,1,2
*STEP,NLGEOM,INC=14
*STATIC,RIKS
1.,1.,,,,2041,3,-0.08
*CLOAD
2041,3,-10000.
*EL PRINT,ELSET=PROFILE,POSITION=AVERAGED AT NODES,FREQUENCY=0
SF,
*NODE PRINT,NSET=XPART,TOTALS=YES,FREQUENCY=0
U,
RF,
*NODE PRINT,NSET=XPARB,TOTALS=YES,FREQUENCY=0
U,
RF,
*NODE FILE,FREQUENCY=1,NSET=MASTER
U,CF
*END STEP
1.2.3 Buckling of a column with spot welds

This example illustrates the dynamic collapse of a steel column constructed by spot welding two
channel sections. It is intended to illustrate the modeling of spot welds. ``Spot welds,'' Section 20.3.5
of the ABAQUS/Explicit User's Manual, discusses the spot weld modeling capabilities provided in
ABAQUS/Explicit.
Problem description
The pillar is composed of two columns of different cross-sections, one box-shaped and the other
W-shaped,welded together with spot welds (Figure 1.2.3-1). The top end of the pillar is connected to a
rigid body, which makes the deformation of the pillar easy to control by manipulating the rigid body
reference node. The column with the box-shaped cross-section is defined to be the slave surface in
contact with the column with the W-shaped cross-section. The box-shaped column is welded to the
W-shaped column with five spot welds on either side of the box-shaped column.
The columns are both composed of aluminum-killed steel, which is assumed to satisfy the
Ramberg-Osgood relation between true stress and logarithmic strain:
" = ¾=E + (¾=K )n ;
where Young's modulus (E) is 206.8 GPa, the reference stress value (K) is 0.510 GPa, and the
work-hardening exponent (n) is 4.76. The material is assumed to be linear elastic below a 0.5% offset
yield stress of 170.0 MPa. (The 0.5% offset yield stress is defined from the Ramberg-Osgood fit by
1-327
taking (" ¡ ¾=E ) to be 0.5% and solving for the stress.) Poisson's ratio is 0.3.
The spot welds on the two sides of the box-shaped column are modeled with different yield forces and
post-yield behavior to illustrate the two failure models.
Spot welded nodes 5203, 15203, 25203, 35203, and 45203 are all located on the positive z-side of the
box-shaped column, with node 5203 at the bottom end of the column and node 45203 at the top end of
the column. The force to cause failure for the spot welds is 3000 N in pure tension and 1800 N in pure
shear. Once the spot welds start to fail, the maximum force that they can bear is assumed to decay
linearly with time over the course of 2.0 msec, which illustrates modeling of complete loss of strength
over a given time period. These spot welds are shown in Figure 1.2.3-2.
Spot welded nodes 5211, 15211, 25211, 35211, and 45211 are all located on the negative z-side of the
box-shaped column, with node 5211 at the bottom end of the column and node 45211 at the top end of
the column. The force to cause failure for these spot welds is 4000 N in pure tension and 2300 N in
pure shear. The spot welds fail according to the damaged failure model, which assumes that the
maximum forces that the spot welds can carry decay linearly with relative displacement between the
welded node and the master surface. The welds are defined to fail completely once their total relative
displacement reaches 0.3 mm, which illustrates modeling of loss of strength in the spot welds based on
energy absorption.
Loading
The bottom of the pillar is fully built-in. The reference node for the rigid body at the top of the pillar
moves at a constant velocity of 25 m/sec in the y-direction, thus loading it in compression, together
with a velocity of 2 m/sec in the z-direction that shears it slightly. At the same time the end of the pillar
is rotated about the negative z-axis at 78.5 rad/sec and rotated about the negative x-axis at 7 rad/sec.
This loading is applied by prescribing the velocities of the reference node of the rigid body that is
attached to the top end of the compound pillar.
The analysis is carried out over 10 milliseconds.

Figure 1.2.3-3 shows the deformed shape of the pillar after 5.0 msec. Figure 1.2.3-4shows the
deformed shape of the pillar after 10.0 msec. Figure 1.2.3-5 and Figure 1.2.3-6show the status of the
spot welds on the positive z-side of the column and the negative z-side of the column, respectively. In
these figures a status of 1.0 means that the weld is fully intact, and 0.0 means that the weld has failed
completely. Figure 1.2.3-7shows the load on spot weld node 25203 relative to the failure load. This
relative value is called the bond load and is defined to be 1.0 when the spot weld starts to fail and 0.0
when the spot weld is broken. Figure 1.2.3-8 shows the time history of the total kinetic energy, the
total work done on the model, the total energy dissipated by friction, the total internal energy, and the
total energy balance.
Input files
pillar.inp
1-328
Input data for this analysis.

pillar_rest.inp
Used to test the restart capability with spot welds.
pillar_ds.inp
Analysis using the double-sided surface capability.
Figures
Figure 1.2.3-1 Initial configuration of the compound pillar.
Figure 1.2.3-2 Initial configuration of the box-shaped column showing spot welds.
Figure 1.2.3-3 Deformed shape at 5.0 msec.
1-329
Figure 1.2.3-4 Deformed shape at 10.0 msec.
Figure 1.2.3-5 Time histories of the status of all spot welds on positive z-side of column.
Figure 1.2.3-6 Time histories of the status of all spot welds on negative z-side of column.
1-330
Figure 1.2.3-7 Time histories of the load on spot weld node 25203 relative to the failure load.
Figure 1.2.3-8 Time histories of the total kinetic energy, energy dissipated by friction, work done on
the model, internal energy, and total energy.
1-331
Sample listings
1-332
Listing 1.2.3-1
*HEADING
Buckling of car with spot welds
************************************************
**
*NODE,NSET=W0
101, 0., 0., 0.065
105, 0.02, 0., 0.065
109, 0.02, 0., 0.045
113, 0., 0., 0.045
117, 0., 0., 0.02
121, 0.02, 0., 0.02
125, 0.02, 0., 0.
129, 0., 0., 0.
*NCOPY,CHANGE=50000,OLD=W0,NEW=W50,SHIFT
0.12,0.45,0.
0.,0.45,0., 0.,0.45,1., -20.
*NGEN,LINE=P,NSET=WPILA
101,50101,1000,, 0.03,0.2,0.065
*NGEN,LINE=P,NSET=WPILB
105,50105,1000,, 0.05,0.2,0.065
*NGEN,LINE=P,NSET=WPILC
109,50109,1000,, 0.05,0.2,0.045
*NGEN,LINE=P,NSET=WPILD
113,50113,1000,, 0.03,0.2,0.045
*NGEN,LINE=P,NSET=WPILE
117,50117,1000,, 0.03,0.2,0.02
*NGEN,LINE=P,NSET=WPILF
121,50121,1000,, 0.05,0.2,0.02
*NGEN,LINE=P,NSET=WPILG
125,50125,1000,, 0.05,0.2,0.0
*NGEN,LINE=P,NSET=WPILH
129,50129,1000,, 0.03,0.2,0.0
*NFILL,NSET=WPIL
WPILA,WPILB,4,1
WPILB,WPILC,4,1
WPILC,WPILD,4,1
WPILD,WPILE,4,1
WPILE,WPILF,4,1
WPILF,WPILG,4,1
WPILG,WPILH,4,1
************************************************
1-333
*ELEMENT,TYPE=S4R,ELSET=WPIL
101, 101,102,1102,1101
*ELGEN,ELSET=WPIL
101, 28,1,1, 50,1000,1000
************************************************
** box
*NODE,NSET=B0
200, 0.003,0.,0.03725
201, 0.003,0.,0.042
205, 0.025,0.,0.042
209, 0.025,0.,0.023
213, 0.003,0.,0.023
214, 0.003,0.,0.02775
*NCOPY,CHANGE=50000,OLD=B0,NEW=B50,SHIFT
0.12,0.45,0.
0.,0.45,0., 0.,0.45,1., -20.
*NGEN,LINE=P,NSET=BPILA
200,50200,1000,, 0.033,0.2,0.03725
*NGEN,LINE=P,NSET=BPILB
201,50201,1000,, 0.033,0.2,0.042
*NGEN,LINE=P,NSET=BPILC
205,50205,1000,, 0.055,0.2,0.042
*NGEN,LINE=P,NSET=BPILD
209,50209,1000,, 0.055,0.2,0.023
*NGEN,LINE=P,NSET=BPILE
213,50213,1000,, 0.033,0.2,0.023
*NGEN,LINE=P,NSET=BPILF
214,50214,1000,, 0.033,0.2,0.02775
*NFILL,NSET=BPIL
BPILA,BPILB,1,1
BPILB,BPILC,4,1
BPILC,BPILD,4,1
BPILD,BPILE,4,1
BPILE,BPILF,1,1
************************************************
*ELEMENT,TYPE=S4R,ELSET=BPIL
200, 200,201,1201,1200
*ELGEN,ELSET=BPIL
200, 14,1,1, 50,1000,1000
************************************************
** roof
*NODE,NSET=ROOF
60000, 0.0125, 0., 0.0325
1-334
*NCOPY,CHANGE=0,OLD=ROOF,NEW=ROOF,SHIFT
0.12,0.45,0.
0.,0.45,0., 0.,0.45,1., -20.
*ELEMENT,ELSET=ROOF,TYPE=R3D4
60101, 49101,49102,50102,50101
60201, 49201,49202,50202,50201
*ELGEN,ELSET=ROOF
60101, 28,1,1
60201, 12,1,1
*ELEMENT,TYPE=MASS,ELSET=REF
60000, 60000
*MASS,ELSET=REF
1.E-3,
************************************************
**
*SHELL SECTION,ELSET=WPIL,MATERIAL=STEEL,
SECTION INTEGRATION=GAUSS
0.002,
*SHELL SECTION,ELSET=BPIL,MATERIAL=STEEL,
0.003,
************************************************
**
*DENSITY
7850.,
*ELASTIC
206.8E9,0.3
*PLASTIC
170.0E6, 0.0000000E+00
180.0E6, 1.7205942E-03
190.0E6, 3.8296832E-03
200.0E6, 6.3897874E-03
210.0E6, 9.4694765E-03
220.0E6, 1.3143660E-02
230.0E6, 1.7493792E-02
240.0E6, 2.2608092E-02
250.0E6, 2.8581845E-02
260.0E6, 3.5517555E-02
270.0E6, 4.3525275E-02
280.0E6, 5.2722659E-02
290.0E6, 6.3235357E-02
300.0E6, 7.5197279E-02
1-335
310.0E6, 8.8750519E-02
320.0E6, 0.1040458
330.0E6, 0.1212430
340.0E6, 0.1405106
350.0E6, 0.1620263
360.0E6, 0.1859779
370.0E6, 0.2125620
380.0E6, 0.2419857
390.0E6, 0.2744660
400.0E6, 0.3102303
410.0E6, 0.3495160
420.0E6, 0.3925720
430.0E6, 0.4396578
440.0E6, 0.4910434
450.0E6, 0.5470111
460.0E6, 0.6078544
470.0E6, 0.6738777
480.0E6, 0.7453985
490.0E6, 0.8227461
500.0E6, 0.9062610
510.0E6, 0.9962980
************************************************
*NSET,NSET=BASE,GEN
101,129,1
200,214,1
*BOUNDARY
BASE,ENCASTRE
************************************************
**
*SURFACE,TYPE=ELEMENT,NAME=BPIL
BPIL,SPOS
*SURFACE,TYPE=ELEMENT,NAME=WPIL
WPIL,SPOS
*RIGID BODY,ELSET=ROOF,REF=60000
*STEP
*DYNAMIC,EXPLICIT
,10.E-3
*RESTART,WRITE,NUM=2, TIMEMARKS=NO
********************
*CONTACT PAIR
BPIL,
********************
*NSET,NSET=WELDA,GEN
1-336
5203,45203,10000
*NSET,NSET=WELDB,GEN
5211,45211,10000
*NSET,NSET=WELDS
WELDA,WELDB
*CONTACT PAIR,INTER=WELDS,WEIGHT=0.
BPIL, WPIL
*SURFACE INTERACTION,NAME=WELDS
*FRICTION,TAUMAX=1.4E8
.3,
*BOND
WELDA, 3000., 1800., 0.5E-3, 2.E-3, 0., 0.
WELDB, 4000., 2300., 0.5E-3, 0., 0.3E-3, 0.3E-3
********************
*BOUNDARY,TYPE=VELOCITY
60000, 1,1, 0.
60000, 2,2, -25.
60000, 3,3, -2.
60000, 4,4, -7.
60000, 5,5, 0.
60000, 6,6, -78.5
********************
*NODE HISTORY,NSET=WELDS
RF,U,BONDSTAT,BONDLOAD
*NODE HISTORY,NSET=ROOF
RF,U
*ENERGY HISTORY
ALLIE,ALLKE,ALLWK,ALLVD,ETOTAL,ALLFD,ALLSE,
ALLAE,ALLCD,ALLPD,DT
*FILE OUTPUT,TIMEMARKS=YES,NUM=1
*NODE FILE
U,
*ENERGY FILE
*END STEP
1.2.4 Elastic-plastic K-frame structure

This example illustrates the use of the frame element FRAME2D. Frame elements (``Frame elements,''
Section 15.4.1 of the ABAQUS/Standard User's Manual) can be used to model elastic, elastic-plastic,
and buckling strut responses of frame-like structures. The elastic response is defined by
Euler-Bernoulli beam theory. The elastic-plastic response is modeled with nonlinear kinematic
1-337
hardening plasticity concentrated at the element's ends, simulating the development of plastic hinges.
The buckling strut response is a simplified, phenomenological representation of the highly nonlinear
cross-section collapse and material yielding that takes place when slender members are loaded in
compression; therefore, frame elements can be elastic, elastic-plastic, behave as struts (with or without
buckling), or switch during the analysis to strut behavior followed by postbuckling behavior. Both the
elastic-plastic and buckling strut responses are simplifications of highly nonlinear responses. They are
designed to approximate these complex responses with a single finite element representing a structural
member between connections. For parts of the model where higher solution resolution is required, such
as stress prediction, the model should be refined with beam elements.
The geometry in this example is a typical K-frame construction used in applications such as offshore
structures (see Figure 1.2.4-1). A push-over analysis is performed to determine the maximum
horizontal load that the structure can support before collapse results from the development of plastic
hinges or buckling failure. During a push-over test, many structural members are loaded in
compression. Slender members loaded in compression often fail due to geometric buckling,
cross-section collapse, and/or material yielding. The buckling strut response, which models such
compressive behavior, is added in separate simulations to investigate the effect of the compressive
failure of critical members in the structure. Push-over analyses are either load or displacement control
tests. A dead load is applied to the top of the structure representing the weight supported by the
K-frame.
Geometry and model

The structure consists of 19 members between structural connections. Hence, 19 frame elements are
used: 17 elements with PIPE cross-sections of varying properties and 2 elements (the top platform)
with I cross-sections. The plastic response of the elements is calculated from the yield stress of the
material, using the plastic default values provided by ABAQUS. (The default values for the plastic
response are based on experiments with slender steel members. For details on the default values, see
``Frame section behavior,'' Section 15.4.2 of the ABAQUS/Standard User's Manual.) The default
plastic response includes mild hardening for axial forces and strong hardening for bending moments.
The default hardening responses for a typical element in the model are shown in Figure 1.2.4-2 and
Figure 1.2.4-3.
A dead load of 444.8 kN (1.0 ´ 105 lb) is applied to the top of the K-frame, representing the part of the
structure above the K-frame. Subsequently, the top platform is loaded or displaced horizontally. The
load level or applied displacement is chosen to be large enough so that the entire structure fails by the
formation of plastic hinges and, consequently, loses load carrying capacity.
Three different models are investigated. A limit load is expected, since the goal of the analysis is to
determine when the structure loses overall stiffness. Large- and small-displacement analyses are
performed for all three models for comparison. (Frame elements assume that the strains are small.
Large-displacement analyses using frame elements are valid for large overall rotations but small
strains.) In the first model all elements use elastic-plastic material response. In the second model
buckling is checked for all elements with PIPE cross-sections. The ISO equation is used as a criteria
for buckling, and the default Marshall strut envelope is followed for the postbuckling behavior. The
buckling strut envelope is calculated from the yield stress of the material and the default Marshall Strut
1-338
theory. (For details on the default buckling strut envelope, see ``Frame section behavior,'' Section
15.4.2 of the ABAQUS/Standard User's Manual.) All frame members use the BUCKLING parameter
on the *FRAME SECTION option to check the ISO criteria for the switching-to-strut algorithm. The
displacement-control analyses are performed for both small and large displacements. In the third model
the two elements that switched to strut behavior in the second model (elements 7 and 9) are replaced
by frame elements with buckling strut response from the beginning of the analysis. To proceed beyond
the unstable phase of the response, the Riks static solution procedure is used in the elastic-plastic
problems. In large-displacement analysis with the switching algorithm and frame elements with
buckling, the STABILIZE parameter is used on the *STATIC option to stabilize the results for the
loads close to the limit load point. To decrease the number of solution iterations, the *CONTROLS
option is used in some cases with the value of the ratio of the largest solution correction to the largest
incremental solution set to 1.0, since displacement increments are very small for increments where
switching occurs.

The structure is loaded or displaced to the point at which all load carrying capacity is lost. In the first
model with elastic-plastic frame elements, the results for the linear and nonlinear geometries compare
as expected. The limit load for the large-displacement analysis is reached at a load of 1141 kN (2.56 ´
105 lb) as compared to a higher load of 1290 kN (2.91 ´ 105 lb) in the small-displacement analysis.
The plastic hinge pattern is the same in both cases.
The second model uses the switching algorithm. It shows that element 7 first violates the ISO equation
(buckles) at a prescribed displacement equal to 1.85 ´ 10--2, before any elements form plastic hinges.
The critical compressive force in this element is -303 kN (-68.12 ´ 103 lb) for the large-displacement
analysis. Next, element 9 buckles after several elements develop plasticity. The frame elements with
the switching algorithm predict the structural behavior in the most accurate way, checking the buckling
criteria for all elements in the model and switching automatically to postbuckling behavior for highly
compressed members (see the plastic and buckled frame elements in Figure 1.2.4-4). When the
structure can no longer support horizontal loading, the patterns of plastic hinges for linear and
nonlinear geometry are very similar. They only show small differences for loads close to the limit load.
To compare the results of two different frame behaviors, the first and the third models are investigated
(kframe_loadcntrl_nlgeom.inp and kframe_dispcntrl_buckle_nlgeom.inp). Load versus horizontal
deflection curves for the large-displacement analyses are shown in Figure 1.2.4-5. The model with two
elements using the buckling strut response becomes unstable as soon as the first element buckles. As
the other elements deform and absorb the load no longer carried by the buckled element, the structure
regains stiffness until the second element buckles. At this point the structure can no longer support
additional horizontal loading because of the presence of buckled elements and the formation of plastic
hinges. The limit load in the third model reaches only about 22% of the limit load in the model without
buckling. The load displacement curves for the switching algorithm and for the example with elements
7 and 9 using the buckling strut response compare well.
Input files
1-339
kframe_loadcntrl_nlgeom.inp
Elastic-plastic analysis with load control; large-displacement analysis.
kframe_loadcntrl.inp
Elastic-plastic analysis with load control; small-displacement analysis.
kframe_dispcntrl_switch_nlgeom.inp
Elastic-plastic frame element with the switching algorithm and displacement control;
large-displacement analysis.
kframe_dispcntrl_switch.inp
Elastic-plastic frame element with the switching algorithm and displacement control;
small-displacement analysis.
kframe_dispcntrl_buckle_nlgeom.inp
Elastic-plastic and buckling strut response with load control; large-displacement analysis.
kframe_dispcntrl_buckle.inp
Elastic-plastic and buckling strut response with displacement control; small-displacement analysis.
Figures
Figure 1.2.4-1 Two-dimensional K-frame structure.
Figure 1.2.4-2 Default hardening response for axial force in a typical element with PIPE cross-section
1-340
(element 7 in the model).
Figure 1.2.4-3 Default hardening response for bending moments in a typical element with PIPE
cross-section (element 7 in the model).
Figure 1.2.4-4 Results of analysis with switching algorithm: K-frame model with plastic and two
buckled elements.
1-341
Figure 1.2.4-5 Applied force versus horizontal displacement of the load point for the elastic-plastic
model and the model including buckling strut response.
Sample listings
1-342
Listing 1.2.4-1
*HEADING
ELASTIC-PLASTIC K-FRAME:
*NODE,NSET=ALL
1,-139.98,0.0
2,
3,139.98,0.0
4,-131.4996,72.0
5,0.0,72.0
6,131.4996,72.0
7,-110.9664,256.8
9, 110.9664,256.8
10,-104.3004,316.8
11,0.0,316.8
12,104.3004,316.8
*NSET,NSET=TEN
10
*ELEMENT,TYPE=FRAME2D,ELSET=BOTTOM
1,1,2
2,2,3
3,2,4
4,2,6
*ELEMENT,TYPE=FRAME2D,ELSET=MID
5,4,5
6,5,6
9,7,9
*ELEMENT,TYPE=FRAME2D,ELSET=PINNED1
7,5,7
8,5,9
*ELEMENT,TYPE=FRAME2D,ELSET=TOP
10,7,11
11,9,11
*ELEMENT,TYPE=FRAME2D,ELSET=W
12,10,11
13,11,12
*ELEMENT,TYPE=FRAME2D,ELSET=BOTTOM_C
14,1,4
15,3,6
*ELEMENT,TYPE=FRAME2D,ELSET=MID_COL
16,4,7
17,6,9
1-343
18,7,10
19,9,12
*FRAME SECTION,SECTION=PIPE,ELSET=BOTTOM,
YIELDSTRESS=51.9E3,PLASTIC DEFAULTS
4.3125,0.25
0.,0.,-1.
3.0E7,1.5E7
*FRAME SECTION,SECTION=PIPE,ELSET=MID,
3.125,0.156
0.,0.,-1.
3.0E7,1.5E7
*FRAME SECTION,SECTION=PIPE,ELSET=PINNED1,
3.125,0.156
0.,0.,-1.
3.0E7,1.5E7
*FRAME SECTION,SECTION=PIPE,ELSET=TOP,
4.3125,0.25
0.,0.,-1.
3.0E7,1.5E7
*FRAME SECTION,SECTION=I,ELSET=W,
7.01,14.02,14.52,14.52,0.71,0.71,0.44
0.,0.,-1.
3.0E7,1.5E7
*FRAME SECTION,SECTION=PIPE,ELSET=BOTTOM_C,
6.375,0.5
0.,0.,-1.
3.0E7,1.5E7
*FRAME SECTION,SECTION=PIPE,ELSET=MID_COL,
6.375,0.33
0.,0.,-1.
3.0E7,1.5E7
*BOUNDARY
1,1,2
3,1,2
*STEP,NLGEOM
*STATIC
*CLOAD
1-344
10,2,-2.5E4
11,2,-5.0E4
12,2,-2.5E4
*EL PRINT,ELSET=PINNED1
*****
** TO FIND ALL PLASTIC HINGES
** PRINT SEP FOR ALL ELEMENTS
*****
SF,
SEE,
SEP,
SALPHA,
**NODE FILE,FREQ=100
*NODE FILE,FREQ=1
U,
CF,
*EL FILE,FREQ=100
SF,
SEE,
SEP,
SALPHA,
*ENERGY FILE,FREQ=100
*MONITOR,DOF=1,NODE=10
*END STEP
*STEP,INC=78,NLGEOM
******
** FOR THE LIMIT LOAD ANALYSIS INCREASE
** THE NUMBER OF INCREMENTS TO INC=100
******
*STATIC,RIKS
0.001,1.0,,0.03
*controls,parameter=field
,1.0
*CLOAD
10,1, 2.75E5
*NODE FILE,FREQ=1,NSET=TEN
U,
CF,
*END STEP
1-345
Listing 1.2.4-2
*HEADING
ELASTIC-PLASTIC K-FRAME WITH SWITCHING ALGORITHM
*NODE,NSET=ALL
1,-139.98,0.0
2,
3,139.98,0.0
4,-131.4996,72.0
5,0.0,72.0
6,131.4996,72.0
7,-110.9664,256.8
9, 110.9664,256.8
10,-104.3004,316.8
11,0.0,316.8
12,104.3004,316.8
1,1,2
2,2,3
3,2,4
4,2,6
5,4,5
6,5,6
9,7,9
7,5,7
8,5,9
10,7,11
11,9,11
12,10,11
13,11,12
14,1,4
15,3,6
16,4,7
17,6,9
18,7,10
19,9,12
1-346
YIELDSTRESS=51.9E3,PLASTIC DEFAULTS,BUCKLING
4.3125,0.25
0.,0.,-1.
3.0E7,1.5E7
3.125,0.156
0.,0.,-1.
3.0E7,1.5E7
3.125,0.156
0.,0.,-1.
3.0E7,1.5E7
4.3125,0.25
0.,0.,-1.
3.0E7,1.5E7
7.01,14.02,14.52,14.52,0.71,0.71,0.44
0.,0.,-1.
3.0E7,1.5E7
6.375,0.5
0.,0.,-1.
3.0E7,1.5E7
6.375,0.33
0.,0.,-1.
3.0E7,1.5E7
*ELSET,ELSET=OUTPUT
7,8,9
*NSET,NSET=TEN
10
*BOUNDARY
1,1,2
3,1,2
1-347
*STATIC,stabilize
0.01,1.0,,0.03
,1.0
*BOUNDARY
*****
** TO MAKE A PLOT AND COMPARE WITH THE LINEAR
** SOLUTION USE THE COMMENTED
** PRESCRIBED DISPLACEMENT
**10,1,1,26.
**10,2,2,10.
**11,2,2,10.
**12,2,2,10.
*****
10,1,1,5.72
10,2,2,2.2
11,2,2,2.2
12,2,2,2.2
*EL PRINT,FREQ=100,ELSET=OUTPUT
SF,
SEE,
*EL PRINT,FREQ=100,ELSET=OUTPUT
SEP,
SALPHA,
*NODE PRINT,NSET=TEN,FREQ=100
RF,
*NODE FILE,NSET=TEN,FREQ=100
****
** TO PLOT USE FREQ=1
****
U,
RF,
*EL FILE,FREQ=100,ELSET=OUTPUT
SF,
SEE,
*EL FILE,FREQ=100
SEP,
SALPHA,
*END STEP
1-348
Listing 1.2.4-3
*HEADING
ELASTIC-PLASTIC, BUCKLING STRUT, DISPLACEMENT
CONTROL, K-FRAME:
*NODE,NSET=ALL
1,-139.98,0.0
2,
3,139.98,0.0
4,-131.4996,72.0
5,0.0,72.0
6,131.4996,72.0
7,-110.9664,256.8
9, 110.9664,256.8
10,-104.3004,316.8
11,0.0,316.8
12,104.3004,316.8
*NSET,NSET=TEN
10
1,1,2
2,2,3
3,2,4
4,2,6
5,4,5
6,5,6
9,7,9
7,5,7
*ELEMENT,TYPE=FRAME2D,ELSET=BRACER
8,5,9
10,7,11
11,9,11
12,10,11
13,11,12
14,1,4
15,3,6
1-349
16,4,7
17,6,9
18,7,10
19,9,12
4.3125,0.25
0.,0.,-1.
3.0E7,1.5E7
YIELDSTRESS=51.9E3,PINNED,BUCKLING
3.125,0.156
0.,0.,-1.
3.0E7,1.5E7
3.125,0.156
0.,0.,-1.
3.0E7,1.5E7
YIELDSTRESS=51.9E3,PINNED,BUCKLING
3.125,0.156
0.,0.,-1.
3.0E7,1.5E7
*FRAME SECTION,SECTION=PIPE,ELSET=BRACER,
3.125,0.156
0.,0.,-1.
3.0E7,1.5E7
4.3125,0.25
0.,0.,-1.
3.0E7,1.5E7
7.01,14.02,14.52,14.52,0.71,0.71,0.44
0.,0.,-1.
3.0E7,1.5E7
6.375,0.5
1-350
0.,0.,-1.
3.0E7,1.5E7
6.375,0.33
0.,0.,-1.
3.0E7,1.5E7
*ELSET,ELSET=OUTPUT
7,8,9
*BOUNDARY
1,1,2
3,1,2
*STEP,NLGEOM
*STATIC,stabilize
*CLOAD
10,2,-2.5E4
11,2,-5.0E4
12,2,-2.5E4
*EL PRINT,FREQ=50
SF,
SEE,
SEP,
SALPHA,
U,
RF,
*EL FILE,FREQ=50,ELSET=OUTPUT
SF,
SEE,
SEP,
SALPHA,
*END STEP
*STATIC,stabilize
0.001,1.0,,0.02
,1.0
*BOUNDARY
****
** TO PLOT USE THE COMMENTED
1-351
** PRESCRIBED DISPLACEMENT
**10,1,1,26.
****
10,1,1,20.8
****
** TO PLOT USE FREQ=1
****
U,
RF,
*END STEP
1.2.5 Unstable static problem: reinforced plate under

compressive loads
This example demonstrates the use of automatic techniques to stabilize unstable static problems.
Geometrically nonlinear static problems can become unstable for a variety of reasons. Instability may
occur in contact problems, either because of chattering or because contact intended to prevent rigid
body motions is not established initially. Localized instabilities can also occur; they can be either
geometrical, such as local buckling, or material, such as material softening.
This problem models a reinforced plate structure subjected to in-plane compressive loading that
produces localized buckling. Structures are usually designed for service loads properly augmented by
safety factors. However, it is quite often of interest to explore their behavior under extreme accident
loads. This example looks into a submodel of a naval construction structure. It is a rectangular plate
reinforced with beams in its two principal directions ( Figure 1.2.5-1). The plate has symmetry
boundary conditions along the longer edges and is pinned rigidly along the shorter sides. An in-plane
load is applied to one of the pinned sides, compressing the plate. Gravity loads are also applied. The
plate buckles under the load. The buckling is initially localized within each of the sections bounded by
the reinforcements. At higher load levels the plate experiences global buckling in a row of sections
closest to the applied load.
Standard analysis procedures typically provide the load at which the structure starts to buckle. The user
may be interested in knowing the structure's additional load carrying capacity. This information could
translate, for instance, into knowing when the onset of global buckling takes place or how far into the
structure damage propagates. In such situations more sophisticated analysis techniques are necessary.
Arc length methods such as the Riks method available in ABAQUS are global load-control methods
that are suitable for global buckling and postbuckling analyses; they do not function well when
buckling is localized. Alternatives are to analyze the problem dynamically or to introduce damping. In
the dynamic case the strain energy released locally from buckling is transformed into kinetic energy; in
the damping case this strain energy is dissipated. To solve a quasi-static problem dynamically is
typically an expensive proposition. In this example the automatic stabilization capability in ABAQUS,
which applies volume proportional damping to the structure, is used.
1-352
Geometry and model

The model consists of a rectangular plate 10.8 m (425.0 in) long, 6.75 m (265.75 in) wide, and 5.0 mm
(0.2 in) thick. This plate has several reinforcements in both the longitudinal and transverse directions
(Figure 1.2.5-1). The plate represents part of a larger structure: the two longitudinal sides have
symmetry boundary conditions, and the two transverse sides have pinned boundary conditions. In
addition, springs at two major reinforcement intersections represent flexible connections to the rest of
the structure. The mesh consists of S4 shell elements for both the plate and larger reinforcements and
additional S3 shell and B31 beam elements for the remaining reinforcements. The entire structure is
made of the same construction steel, with an initial flow stress of 235.0 MPa (34.0 ksi).

The analysis consists of two steps. In the first step a gravity load perpendicular to the plane of the plate
is applied. In the second step a longitudinal compressive load of 6.46 ´ 106 N (1.45 ´ 106 lbf) is
applied to one of the pinned sides of the plate. All the nodes on that side are forced to move equally by
means of multi-point constraints. The analysis is quasi-static, but buckling is expected. The volume
proportional damping stabilizing capability in ABAQUS is invoked with the *STATIC, STABILIZE
option, with the default damping intensity. This option applies a damping coefficient such that the
viscous dissipated energy extrapolated from the first increment to the total step is a small fraction (2.0
´ 10-4) of the strain energy also extrapolated from the first increment to the total step. The algorithm
works quite well in situations such as this problem, in which the first increment of a step is stable but
instabilities develop later in the analysis. Initially local out-of-plane buckling develops throughout the
plate in an almost checkerboard pattern inside each one of the sections delimited by the reinforcements
(Figure 1.2.5-2). Later, global buckling develops along a front of sections closer to the applied load
(Figure 1.2.5-3). The evolution of the displacements produced by the applied load is very smooth
(Figure 1.2.5-4) and does not reflect the early local instabilities in the structure. However, when the
global instability develops, the curve becomes almost flat, indicating the complete loss of load carrying
capacity. An inspection of the model's energy content (Figure 1.2.5-5 and Figure 1.2.5-6) reveals that
while the load is increasing, the amount of dissipated energy is negligible. As soon as the load flattens
out, the strain energy also flattens out (indicating a more or less constant load carrying capacity), while
the dissipated energy increases dramatically to absorb the work done by the applied loads.
Acknowledgements
HKS would like to thank IRCN (France) for providing this example.
Input files
unstablestatic_plate.inp
Plate model.
unstablestatic_plate_node.inp
Node definitions for the plate model.
1-353
unstablestatic_plate_elem.inp
Element definitions for the plate model.
Figures
Figure 1.2.5-1 Reinforced plate initial mesh.
Figure 1.2.5-2 Plate localized buckling.
Figure 1.2.5-3 Plate global buckling.
1-354
Figure 1.2.5-4 Plate load-displacement curve.
Figure 1.2.5-5 Dissipated and strain energies as functions of load.
1-355
Figure 1.2.5-6 Dissipated and strain energies as functions of displacement.
Sample listings
1-356
Listing 1.2.5-1
*HEADING
STABILIZED COMPRESSION OF REINFORCED PLATE
*PREPRINT,ECHO=NO,HISTORY=NO,MODEL=YES
**
************************************************
** FILE WITH NODE DEFINITIONS AND NSETS:
** COTE1
** COTE2
** COTE3
** COTE4
************************************************
**
*INCLUDE,INPUT=unstablestatic_plate_node.inp
**
************************************************
** FILE WITH ELEMENT DEFINITIONS AND ELSETS:
** E0000001 144 B31
** E0000002 144 B31
** E0000003 6 S3
** E0000004 8 S3
** TOLE 1008 S4
** E0000006 288 S4
** E0000007 252 S4
** E0000008 178 S4
** E0000009 108 S4
** E0000010 78 S4
************************************************
**
*INCLUDE,INPUT=unstablestatic_plate_elem.inp
**
*ELEMENT,TYPE=SPRING1,ELSET=EPONT
50000,340
50001,2881
*SPRING,ELSET=EPONT
3,
5.3E+08,
*BEAM SECTION,MATERIAL=ACIER,SECTION=RECT,
ELSET=E0000001
1.500000E-02, 1.500000E-02
0.000000E+00, 6.557236E-03,-9.999785E-01
*MATERIAL,NAME=ACIER
1-357
*ELASTIC
2.100E+11, 3.000E-01
*PLASTIC
235.259E+06,0.0
471.9E+06,0.18837
*DENSITY
7.850E+03,
*BEAM SECTION,MATERIAL=ACIER,SECTION=RECT,
ELSET=E0000002
1.500000E-02, 1.500000E-02
0.000000E+00, 8.695324E-03,-9.999622E-01
*SHELL SECTION,ELSET=E0000003,MATERIAL=ACIER
7.000E-03, 3
1.000E-02, 3
*SHELL SECTION,ELSET=TOLE,MATERIAL=ACIER
5.000E-03, 3
6.000E-03, 3
7.000E-03, 3
1.000E-02, 3
1.000E-02, 3
1.000E-02, 3
*NSET,NSET=NM1
329,
*ELSET,ELSET=HILOIRE
E0000004,E0000009,E0000010
*ELSET,ELSET=BARROTS
E0000003,E0000007,E0000008
*ELSET,ELSET=TOT
TOLE,HILOIRE,BARROTS,E0000001,E0000002,E0000006
**
*MPC
BEAM,COTE1,NM1
**
*BOUNDARY
NM1,3,3
NM1,4,4
NM1,6,6
1-358
COTE2,1,1
COTE2,3,3
COTE2,4,4
COTE2,6,6
8896,2,2
COTE3,2,2
COTE3,4,4
COTE3,6,6
COTE4,2,2
COTE4,4,4
COTE4,6,6
**
*RESTART,WRITE,F=10
**
************************************************
**
*STEP,NLGEOM,INC=1
GRAVITY LOAD
*STATIC
1.,1.
*DLOAD
TOT,GRAV,9.81,0,0,-1.
*NODE FILE,NSET=NM1,FREQ=1
U,CF
*EL FILE,ELSET=HILOIRE,FREQ=200
S,E,PE
*ENERGY FILE
*OUTPUT,FIELD
*NODE OUTPUT
U,VF
*ELEMENT OUTPUT
S,E,PE
*OUTPUT,HISTORY
*ENERGY OUTPUT,VARIABLE=ALL
*NODE OUTPUT,NSET=NM1
U,CF
*EL PRINT,FREQ=0
*NODE PRINT,FREQ=0
*END STEP
**
BUCKLING LOAD
1-359
*STATIC,STABILIZE
0.1,1.,,
*CLOAD
NM1,1,646.E+04
*END STEP
1.2.6 Buckling of an imperfection sensitive cylindrical shell

This example serves as a guide to performing a postbuckling analysis using ABAQUS for an
imperfection sensitive structure. A structure is imperfection sensitive if small changes in an
imperfection change the buckling load significantly. Qualitatively, this behavior is characteristic of
structures with closely spaced eigenvalues. For such structures the first eigenmode may not
characterize the deformation that leads to the lowest buckling load. A cylindrical shell is chosen as an
example of an imperfection sensitive structure.
Geometry and model

The cylinder being analyzed is depicted in Figure 1.2.6-1. The cylinder is simply supported at its ends
and is loaded by a uniform, compressive axial load. A uniform internal pressure is also applied to the
cylinder. The material in the cylinder is assumed to be linear elastic. The thickness of the cylinder is
1/500 of its radius, so the structure can be considered to be a thin shell.
The finite element mesh uses the fully integrated S4 shell element. This element is based on a finite
membrane strain formulation and is chosen to avoid hourglassing. A full-length model is used to
account for both symmetric and antisymmetric buckling modes. A fine mesh, based on the results of a
refinement study of the linear eigenvalue problem, is used. The convergence of the mesh density is
based on the relative change of the eigenvalues as the mesh is refined. The mesh must have several
elements along each spatial deformation wave; therefore, the level of mesh refinement depends on the
modes with the highest wave number in the circumferential and axial directions.
Solution procedure
The solution strategy is based on introducing a geometric imperfection in the cylinder. In this study the
imperfections are linear combinations of the eigenvectors of the linear buckling problem. If details of
imperfections caused in a manufacturing process are known, it is normally more useful to use this
information as the imperfection. However, in many instances only the maximum magnitude of an
imperfection is known. In such cases assuming the imperfections are linear combinations of the
eigenmodes is a reasonable way to estimate the imperfect geometry (Arbocz, 1987).
Determining the most critical imperfection shape that leads to the lowest collapse load of an axially
compressed cylindrical shell is an open research issue. The procedure discussed in this example does
not, therefore, claim to compute the lowest collapse load. Rather, this example discusses one approach
that can be used to study the postbuckling response of an imperfection sensitive structure.
The first stage in the simulation is a linear eigenvalue buckling analysis. To prevent rigid body motion,
1-360
a single node is fixed in the axial direction. This constraint is in addition to the simply supported
boundary conditions noted earlier and will not introduce an overconstraint into the problem since the
axial load is equilibrated on opposing edges. The reaction force in the axial direction should be zero at
this node.
The second stage involves introducing the imperfection into the structure using the *IMPERFECTION
option. A single mode or a combination of modes is used to construct the imperfection. To compare
the results obtained with different imperfections, the imperfection size must be fixed. The measure of
the imperfection size used in this problem is the out-of-roundness of the cylinder, which is computed
as the radial distance from the axis of the cylinder to the perturbed node minus the radius of the perfect
structure. The scale factor associated with each eigenmode used to seed the imperfection is computed
with a FORTRAN program. The program reads the results file produced by the linear analysis and
determines the scale factors so that the out-of-roundness of the cylinder is equal to a specified value.
This value is taken as a fraction of the cylinder thickness.
The final stage of the analysis simulates the postbuckling response of the cylinder for a given
imperfection. The primary objective of the simulation is to determine the static buckling load. The
modified Riks method is used to obtain a solution since the problem under consideration is unstable.
The Riks method can also be used to trace the unstable and stable solution branches of a buckled
structure. However, with imperfection sensitive structures the first buckling mode is usually
catastrophic, so further continuation of the analysis is usually not undertaken. When using the
*STATIC, RIKS option, the tolerance used for the force residual convergence criteria may need to be
tightened to ensure that the solution algorithm does not retrace its original loading path once the limit
point is reached. Simply restricting the maximum arc length allowed in an increment is normally not
sufficient.
Parametric study
There are two factors that significantly alter the buckling behavior: the shape of the imperfection and
the size of the imperfection. A convenient way to investigate the effects of these factors on the
buckling response is to use the parametric study capabilities of ABAQUS. A Python script file is used
to perform the study. The script executes the linear analysis, runs the FORTRAN routine to create an
input file with a specified imperfection size, and finally executes the postbuckling analysis.
Before executing the script, copy the FORTRAN routine cylsh_maximp.f to your work directory using
the ABAQUS fetch command,
abaqus fetch job=cylsh_maximp.f
and compile it using the ABAQUS make command,
abaqus make job=cylsh_maximp.f
Parametrized template input data are used to generate variations of the parametric study. The script
allows the analyst to vary the eigenmodes used to construct the imperfection, out-of-roundness
measure, cylindrical shell geometry (radius, length, thickness), mesh density, material properties
(Young's modulus and Poisson's ratio), etc. The results presented in the following section, however,
are based on an analysis performed with a single set of parameters.
1-361

The results for both the linear eigenvalue buckling and postbuckling analyses are discussed below.
Linear eigenvalue buckling

The Lanczos eigensolver is used to extract the linear buckling modes. This solver is chosen because of
its superior accuracy and convergence rate relative to wavefront solvers for problems with closely
spaced eigenvalues. Table 1.2.6-1 lists the first 19 eigenvalues of the cylindrical shell. The eigenvalues
are closely spaced with a maximum percentage difference of 1.3%.
The geometry, loading, and material properties of the cylindrical shell analyzed in this example are
characterized by their axisymmetry. As a consequence of this axisymmetry the eigenmodes associated
with the linear buckling problem will be either (1) axisymmetric modes associated with a single
eigenvalue, including the possibility of eigenmodes that are axially symmetric but are twisted about the
symmetry axis or (2) nonaxisymmetric modes associated with repeated eigenvalues (Wohlever, 1999).
The nonaxisymmetric modes are characterized by sinusoidal variations (n-fold symmetry) about the
circumference of the cylinder. For most practical engineering problems and as illustrated in Table
1.2.6-1, it is usually found that a majority of the buckling modes of the cylindrical shell are
nonaxisymmetric.
The two orthogonal eigenmodes associated with each repeated eigenvalue span a two-dimensional
space, and as a result any linear combination of these eigenmodes is also an eigenmode; i.e., there is no
preferred direction. Therefore, while the shapes of the orthogonal eigenmodes extracted by the
eigensolver will always be the same and span the same two-dimensional space, the phase of the modes
is not fixed and might vary from one analysis to another. The lack of preferred directions has
consequences with regard to any imperfection study based upon a linear combination of
nonaxisymmetric eigenmodes from two or more distinct eigenvalues. As the relative phases of
eigenmodes change, the shape of the resulting imperfection and, therefore, the postbuckling response,
also changes. To avoid this situation, postprocessing is performed after the linear buckling analysis on
each of the nonaxisymmetric eigenmode pairs to fix the phase of the eigenmodes before the
imperfection studies are performed. The basic procedure involves calculating a scaling factor for each
of the eigenvectors corresponding to a repeated eigenvalue so that their linear combination generates a
maximum displacement of 1.0 along the global X-axis. This procedure is completely arbitrary but
ensures that the postbuckling response calculations are repeatable.
For the sake of consistency the maximum radial displacement associated with a unique eigenmode is
also scaled to 1.0. These factors are further scaled to satisfy the out-of-roundness criterion mentioned
earlier.
Postbuckling response
The modes used to seed the imperfection are taken from the first 19 eigenmodes obtained in the linear
eigenvalue buckling analysis. Different combinations are considered: all modes, unique eigenmodes,
and pairs of repeated eigenmodes. An imperfection size (i.e., out-of-roundness) of 0.5 times the shell
thickness is used in all cases. The results indicate that the cylinder buckles at a much lower load than
the value predicted by the linear analysis (i.e., the value predicted using only the lowest eigenmode of
1-362
the system). An imperfection based on mode 1 (a unique eigenmode) results in a buckling load of
about 90% of the predicted value. When the imperfection was seeded with a combination of all modes
(1-19), a buckling load of 35% of the predicted value was obtained. Table 1.2.6-2 lists the buckling
loads predicted by ABAQUS (as a fraction of linear eigenvalue buckling load) when different modes
are used to seed the imperfection.
The smallest predicted buckling load in this study occurs when using modes 12 and 13 to seed the
imperfection, yet the results obtained when the imperfection is seeded using all 19 modes indicate that
a larger buckling load can be sustained. One possible explanation for this is that the solution strategy
used in this study (discussed earlier) involves using a fixed value for the out-of-roundness of the
cylinder as a measure of the imperfection size. Thus, when multiple modes are used to seed the
imperfection, the overall effect of any given mode is less than it would be if only that mode were used
to seed the imperfection. The large number of closely spaced eigenvalues and innumerable
combinations of eigenmodes clearly demonstrates the difficulty of determining the collapse load of
structures such as the cylindrical shell. In practice, designing imperfection sensitive structures against
catastrophic failure usually requires a combination of numerical and experimental results as well as
practical building experience.
The deformed configuration shown in Figure 1.2.6-2uses a displacement magnification factor of 5 and
corresponds to using all the modes to seed the imperfection. Even though the cylinder appears to be
very short, it can in fact be classified as a moderately long cylinder using the parameters presented in
Chajes (1985). The cylinder exhibits thin wall wrinkling; the initial buckling shape can be
characterized as dimples appearing on the side of the cylinder. The compression of the cylinder causes
a radial expansion due to Poisson's effect; the radial constraint at the ends of the cylinder causes
localized bending to occur at the ends. This would cause the shell to fold into an accordion shape.
(Presumably this would be seen if self-contact was specified and the analysis was allowed to run
further. This is not a trivial task, however, and modifications to the solution controls would probably
be required. Such a simulation would be easier to perform with ABAQUS/Explicit.) This deformed
configuration is in accordance with the perturbed reference geometry, shown in Figure 1.2.6-3. To
visualize the imperfect geometry, an imperfection size of 5.0 times the shell thickness (i.e., 10 times
the value actually used in the analysis) was used to generate the perturbed mesh shown in this figure.
The deformed configuration in the postbuckling analysis depends on the shape of the imperfection
introduced into the structure. Seeding the structure with different combinations of modes and
imperfection sizes produces different deformed configurations and buckling loads. As the results vary
with the size and shape of the imperfection introduced into the structure, there is no solution to which
the results from ABAQUS can be compared.
The load-displacement curve for the case when the first 19 modes are used to seed the imperfection is
shown in Figure 1.2.6-4. The figure shows the variation of the applied load (normalized with respect to
the linear eigenvalue buckling load) versus the axial displacement of an end node. The peak load that
the cylinder can sustain is clearly visible.
Input files
cylsh_buck.inp
1-363
Linear eigenvalue buckling problem.

cylsh_postbuck.inp
Postbuckling problem.
cylsh_maximp.f
FORTRAN program to compute the scaling factors for the imperfection size.
cylsh_script.psf
Python script to generate the parametrized input files.
References
· Arbocz, J., "Post-Buckling Behaviour of Structures: Numerical Techniques for More
Complicated Structures," in Lecture Notes in Physics, Ed. H. Araki et al., Springer-Verlag,
Berlin, 1987, pp. 84-142.
· Chajes, A., "Stability and Collapse Analysis of Axially Compressed Cylindrical Shells," in Shell
Structures: Stability and Strength , Ed. R. Narayanan, Elsevier, New York, 1985, pp. 1-17.
· Wohlever, J. C., "Some Computational Aspects of a Group Theoretic Finite Element Approach to
the Buckling and Postbuckling Analyses of Plates and Shells-of-Revolution ," in Computer
Methods in Applied Mechanics and Engineering , vol. 170, pp. 373-406, 1999.
Tables
Table 1.2.6-1 Eigenvalue estimates for the first 19 modes.

Mode Eigenvalue
number
1 11721
2, 3 11722
4, 5 11726
6, 7 11733
8, 9 11744
10, 11 11758
12, 13 11777
14, 15 11802
16, 17 11833
18, 19 11872
Table 1.2.6-2 Summary of predicted buckling loads.

Mode used to Normalized
seed the buckling
imperfection load
1 0.902
1-364
2, 3 0.707
4, 5 0.480
6, 7 0.355
8, 9 0.351
10, 11 0.340
12, 13 0.306
14, 15 0.323
16, 17 0.411
18, 19 0.422
All modes 0.352
(1-19)
Figures
Figure 1.2.6-1 Cylindrical shell with uniform axial loading.
Figure 1.2.6-2 Deformed configuration of the cylindrical shell (first 19 eigenmodes used to seed the
imperfection; displacement magnification factor of 5.0).
1-365
Figure 1.2.6-3 Perturbed geometry of the cylindrical shell (imperfection factor = 5 ´ thickness for
illustration only; actual imperfection factor used = .5 ´ thickness).
Figure 1.2.6-4 Normalized applied load versus axial displacement at an end node (first 19 modes used
to seed the imperfection).
1-366
Sample listings
1-367
Listing 1.2.6-1
import string
import os
#
########################################################################
#
# THIS SCRIPT RUNS A SEQUENCE OF PARAMETERIZED INPUT FILES TO STUDY THE
# POSTBUCKLING BEHAVIOR OF A LINEAR ELASTIC, AXIALLY LOADED,
# CYLINDRICAL SHELL.
#
# PARAMETERS USED IN STUDY:
#
# SHELL THICKNESS: thickness (set by xthickne
# LENGTH OF THE CYLINDER: length (set by xlength)
# MEAN RADIUS OF THE CYLINDER: radius (set by xradius)
# NUMBER OF NODES AROUND CIRCUMF: node_circum (set by nel_c)
# NUMBER OF NODES ALONG LENGTH: node_length (set by nel_l)
# POISSON'S RATIO: poisson (set by xpoisson
# YOUNG'S MODULUS: young (set by xyoung)
# APPLIED LOAD (BUCKLING ANALYSIS): tot_load (set by xload)
# INTERNAL PRESSURE: int_press (set by press)
# NUMBER OF BUCKLING MODES: num_modes (set by nmodes)
#
# ADDITIONAL PARAMETERS FOR POST-BUCKLING ANALYSIS
#
# APPLIED AXIAL LOAD (BASED ON
# LOWEST BUCKLING MODE): eig1_load (set by xload1)
# BUCKLING ANALYSIS RESULTS FILE NAME: buckle file (set by bklfname
# *INCLUDE FILE WITH *IMPERFECTION DATA : imperf file (set by impfname
#
# ADDITIONAL VARIABLES
#
# IMPERFECTION SCALE FACTOR: RadialImperfFactor
# PRECRIBED RADIAL IMPERFECTION: rad_imp
# EIGENMODES TO SEED THE IMPERFECTION: eigmodes
# SCALE FACTOR ASSOC. WITH EACH MODE: modefctr
# -THIS IS A GUESS; THE TRUE FACTORS
# WILL BE COMPUTED LATER.
#
########################################################################
#
# 1. DEFINE THE SHELL PARAMETERS AND MAX-OUT-OF ROUND IMPERFECTION FACTO
1-368
xthickness = 0.01
xlength = 2.0
xradius = 5.0
nel_c = 240
nel_l = 21
xpoisson = 0.3
xyoung = 30.0e6
xload = 1.0
press = 1.0
nmodes = 19
radialImperfFactor = 0.5
# 2. SPECIFY THE EIGENMODES WHICH WILL BE USED TO SEED THE IMPERFECTION

# (TAKE ALL MODES, 1 THROUGH NMODES)
eigmodes = [ ]
for i in range(nmodes): eigmodes.append(i+1)
# 3. MISC. DEFINITIONS
zero = 0
########################################################################
#
# PERFORM EIGENVALUE BUCKLING ANALYSIS
#
# CREATE THE STUDY
buckle = ParStudy(par=('thickness','length','radius',
'node_circum','node_length','poisson','young',
'tot_load','int_press','num_modes'))
# DEFINE THE NAMES OF THE INPUT DECKS
names = ['cylsh_buck']
# DEFINE THE PARAMETERS
buckle.define(DISCRETE,par='thickness')
buckle.define(DISCRETE,par='length')
buckle.define(DISCRETE,par='radius')
1-369
buckle.define(DISCRETE,par='node_circum')
buckle.define(DISCRETE,par='node_length')
buckle.define(DISCRETE,par='poisson')
buckle.define(DISCRETE,par='young')
buckle.define(DISCRETE,par='tot_load')
buckle.define(DISCRETE,par='num_modes')
buckle.define(DISCRETE,par='int_press')
# SAMPLE THE PARAMETERS - INPUT THE APPROPRIATE VALUES
buckle.sample(VALUES,par='thickness',values=(xthickness))
buckle.sample(VALUES,par='length',values=(xlength))
buckle.sample(VALUES,par='radius',values=(xradius))
buckle.sample(VALUES,par='node_circum',values=(nel_c))
buckle.sample(VALUES,par='node_length',values=(nel_l))
buckle.sample(VALUES,par='poisson',values=(xpoisson))
buckle.sample(VALUES,par='young',values=(xyoung))
buckle.sample(VALUES,par='tot_load',values=(xload))
buckle.sample(VALUES,par='num_modes',values=(nmodes))
buckle.sample(VALUES,par='int_press',values=(press))
# COMBINE THE SAMPLES INTO ANALYSES
buckle.combine(MESH,name='short')
for temp in names:
# GENERATE INPUT DECKS AND EXECUTION SCRIPT

# FOR VARIOUS TEMPLATES
buckle.generate(template=temp)
# EXECUTE RUNS SEQUENTIALLY
buckle.execute()
# GATHER RESULTS FOR FIRST BUCKLING MODE AND WRITE TO OUTPUT FILES
buckle.gather(results='mode',variable='MODAL',step=2)
buckle.report(PRINT,par=('length','radius','thickness',
'poisson','young'),results=('mode.2'))
1-370
# 3. STORE LOWEST EIGENVALUE IN EIG_VALS; COMPUTE RAD_IMP.

#
# STORE THE RESULTS (.FIL) FILE NAMES OF EACH ANALYSIS (IN BKLFNAME
# THE PARAMETER SETTINGS FOR EACH ANALYSIS (IN PARAMS), AND A
# DESCRIPTIVE STRING OF CHARACTERS FOR EACH ANALYSIS (IN XNAMES).
# STORE THE NUMBER OF NODES ALONG THE CIRCUMFERENCE (IN N CIRC) AND
# LENGTH (N_LGTH)
eig_vals=[ ]
rad_imp =[ ]
n_circ = [ ]
n_lgth = [ ]
bklfnames=[ ]
xnames=[ ]
params = []
i = -1
res = buckle.table.results
for jname in buckle.job.keys():
i = i+1
des = buckle.job[jname].design
thickness = des[0]
r_imp = radialImperfFactor*thickness
length = des[1]
radius = des[2]
ncrc = des[3]
nlen = des[4]
value = res[i][0]
xname = buckle.job[jname].designName
root = buckle.job[jname].root
bklfname = root + '_' + xname
bklfnames.append(bklfname)
xnames.append(xname)
param = des
params.append(param)
eig_vals.append(value)
rad_imp.append(r_imp)
n_circ.append(ncrc)
n_lgth.append(nlen)
# INTERMEDIATE STEP: RUN FORTRAN PROGRAM TO

# DETERMINE TRUE SCALE FOR IMPERFECTIONS
1-371
#
# NOW CREATE THE INPUT FILE REQUIRED TO RUN THE FORTRAN PROGRAM;
# INCLUDES THE FOLLOWING:
#
# 1. NAME ASSIGNED TO THE OUTPUT FILE CREATED BY THE PROGRAM
# THIS FILE CONTAINS THE SCALE FACTOR FOR EACH EIGENMODE
# USED TO SEED THE IMPERFECTION AND WILL BE INCLUDED IN THE
# POST BUCKLING ANALYSIS FILE.)
# 2. RESULTS FILE NAME FOR EIGENVALUE BUCKLING ANALYSIS
# 3. NUMBER OF CIRCUMF. AND LONG. NODES
# 4. PRESCRIBED RADIAL IMPERFECTION (FROM RAD_IMP)
# 5. LIST OF EIGENMODES (FROM EIGMODES) FOLLOWED BY ZERO
# 6. LIST OF "GUESS" SCALE FACTORS (FROM MODEFCTR)
#
file = 'max_round_input.dat'
impfnames = [ ]
names = [ ]
i = -1
for bklfname in bklfnames:
i = i+1
impfname = bklfname + str('_imp')
impfnames.append(impfname)
names.append(bklfname)
modefctr = [ ]
for m in eigmodes: modefctr.append(rad_imp[i])
f1 = open(file,'w')
f1.write(impfnames[i] + '\n')
f1.write(names[i] + '\n')
f1.write(str(n_circ[i]) + '\n')
f1.write(str(n_lgth[i]) + '\n')
f1.write(str(rad_imp[i]) + '\n')
for imode in eigmodes: f1.write(str(imode) + '\n')

f1.write(str(zero) + '\n')
for ifctr in modefctr: f1.write(str(ifctr) + '\n')
f1.close()
# RUN THE FORTRAN PROGRAM
os.system('./cylsh_maximp.x')
1-372
# PERFORM THE POST-BUCKLING ANALYSIS

#
# CREATE THE STUDY
riks = ParStudy(par=('thickness','length','radius',
'node_circum','node_length','poisson','young',
'int press','eig1 load','buckle file','imperf file'
# DETERMINE THE APPROPRIATE LOAD LEVEL FOR THE ANALYSIS
xload1 = [ ]
for l in eig_vals:
xl = float(xload)*float(l)
xload1.append(xl)
# DEFINE THE NAMES OF THE INPUT DECKS
names2 = ['cylsh_postbuck']
# DEFINE THE PARAMETERS
riks.define(DISCRETE,par='thickness')
riks.define(DISCRETE,par='length')
riks.define(DISCRETE,par='radius')
riks.define(DISCRETE,par='node_circum')
riks.define(DISCRETE,par='node_length')
riks.define(DISCRETE,par='poisson')
riks.define(DISCRETE,par='young')
riks.define(DISCRETE,par='eig1_load')
riks.define(DISCRETE,par='buckle_file')
riks.define(DISCRETE,par='imperf_file')
riks.define(DISCRETE,par='int_press')
# SAMPLE THE PARAMETERS - INPUT THE APPROPRIATE VALUES
for i in range(len(xload1)):
riks.sample(VALUES,par='thickness',values=(params[i][0]))
riks.sample(VALUES,par='length',values=(params[i][1]))
riks.sample(VALUES,par='radius',values=(params[i][2]))
riks.sample(VALUES,par='node_circum',values=(params[i][3]))
riks.sample(VALUES,par='node_length',values=(params[i][4]))
riks.sample(VALUES,par='poisson',values=(params[i][5]))
1-373
riks.sample(VALUES,par='young',values=(params[i][6]))
riks.sample(VALUES,par='int_press',values=(params[i][8]))
riks.sample(VALUES,par='eig1_load',values=(xload1[i]))
riks.sample(VALUES,par='buckle_file',values=(bklfnames[i]))
riks.sample(VALUES,par='imperf_file',values=(impfnames[i]))
# COMBINE THE SAMPLES INTO ANALYSES

#
# (APPEND XNAME TO THE JOBNAMES OTHERWISE THEY WILL NOT BE UNIQUE;
# ALSO PROVIDES A MEANS OF ASSOCIATING THE POSTBUCKLING ANALYSIS
# WITH ITS CORRESPONDING EIGENVALUE BUCKLING ANALYSIS.)
riks.combine(MESH,name=xnames[i])
for temp in names2:
# GENERATE INPUT DECKS AND EXECUTION SCRIPT

# FOR VARIOUS TEMPLATES
riks.generate(template=temp)
# EXECUTE RUNS SEQUENTIALLY
riks.execute()
riks.gather(results='lpf',variable='LPF',step=2,inc=LAST)
# GATHER RESULTS FOR HKS QA PURPOSES ONLY
riks.report(PRINT,par=('length','radius',
'thickness','poisson','young'),
results=('lpf'))
buck_res = buckle.table.results
riks_res = riks.table.results
resfile = open('cylsh_script.psr','w')
for x in buck_res: resfile.write('%14.6g \n' % x[0])
for y in riks_res: resfile.write('%14.3g \n' % y[0])
1-374
Listing 1.2.6-2
*heading
Input file for the linear buckling analysis.
*parameter
#
#geometric/load parameters (can be modified)
#
radius = 5.0
length = 2.0
thickness = 0.01
tot_load = 1.0
#
#elastic material properties (can be modified)
#
young = 30e+06
poisson=0.3
#
#number of buckling modes to be extractend (can be modified)
#
num_modes=20
#
#internal pressure (can be modified)
#
int_press = 0.0
#
#mesh parameters (can be modified)
#
node_circum = 240
node_length = 21
##
##dependent parameters (do not modify)
##
chn = node_circum*node_length-node_circum
node_ang = -360.0/float(node_circum)
node_tot = node_circum*node_length
node_tmp = node_tot-node_circum+1
node_int = node_length-1
node_circum1 = node_circum+1
node_circum0 = node_circum-1
e1 = node_circum*2
p = -tot_load/float(node_circum)
1-375
pn = tot_load/float(node_circum)
#
#end of parameter list
#
*node,system=c
1,<radius>,0.0,0.0
<node_circum>,<radius>,<node_ang>,0.0
<node_tmp>,<radius>,0.0,<length>
<node_tot>,<radius>,<node_ang>,<length>
*ngen,line=c,nset=bottom
1,<node_circum>,1,,0.0,0.0,0.0,0.0,0.0,1.0
*ncopy,new set=top,old set=bottom,shift,change number=<chn>
0.0,0.0,<length>
0.0,0.0,0.0,0.0,0.0,1.0,0.0
*nfill
bottom,top,<node_int>,<node_circum>
*element,type=s4
1,1,2,<node_circum2>,<node_circum1>
<node_circum>,<node_circum>,1,<node_circum1>,<e1>
*elgen,elset=cylinder
1,<node_circum0>,1,1,<node_int>,<node_circum>,<node_circum>
<node_circum>,1,,,<node_int>,<node_circum>,<node_circum>
*shell section, elset=cylinder, material=mat_1
<thickness>,
*material,name=mat_1
*elastic
<young>,<poisson>
*nset,nset=ends
bottom,top
*transform,type=c,nset=ends
0.0,0.0,0.0,0.0,0.0,1.0
*boundary
ends,1,2
ends,4,4
ends,6,6
1,3
**
*step,nlgeom,inc=10
static preload for internal pressure
*static
1.0,1.0
*monitor,node=<node_tot>, dof=3
*dload
1-376
cylinder,p,<int_press>
*end step
*step,nlgeom
*buckle,eigensolver=lanczos
<num_modes>,
*cload
top,3,<p>
bottom,3,<pn>
*output,field
*node output
u,
*node file,global=yes
u,
*end step
1-377
Listing 1.2.6-3
*heading
Input file for the postbuckling analysis.
*parameter
#
# filenames
#
buckle_file = 'file1.dat'
imperf_file = 'file2.dat'
#
# workaround to allow parametrization of a
# filename read with *INCLUDE
#
line1 = '*include, input='+imperf_file
#
# geometric/load parameters
#
radius = 5.0
length = 2.0
thickness = 0.01
#
# this is the pcritical for the 1st value from
# the linear eigenvalue analysis
#
eig1_load = 1.18305e+4
#
# elastic material properties
#
young = 30e+06
poisson=0.3
#
# internal pressure
#
int_press = 0.0
#
# mesh parameters
#
node_circum = 240
node_length = 21
##
## dependent parameters (do not modify)
##
1-378
chn = node_circum*node_length-node_circum
node_ang = -360.0/float(node_circum)
node_tot = node_circum*node_length
node_tmp = node_tot-node_circum+1
node_int = node_length-1
node_circum0 = node_circum-1
e1 = node_circum*2
p = -eig1_load/float(node_circum)
pn = eig1_load/float(node_circum)
#
# end of parameter list
#
*node,system=c
1,<radius>,0.0,0.0
<node_circum>,<radius>,<node_ang>,0.0
<node_tmp>,<radius>,0.0,<length>
<node_tot>,<radius>,<node_ang>,<length>
*ngen,line=c,nset=bottom
1,<node_circum>,1,,0.0,0.0,0.0,0.0,0.0,1.0
*ncopy,new set=top,old set=bottom,shift,change number=<chn>
0.0,0.0,<length>
0.0,0.0,0.0,0.0,0.0,1.0,0.0
*nfill
bottom,top,<node_int>,<node_circum>
**specify the imperfection as a function of modeshape amplitude
*imperfection,file=<buckle_file>,step=2
<line1>
*element,type=s4
1,1,2,<node_circum2>,<node_circum1>
<node_circum>,<node_circum>,1,<node_circum1>,<e1>
*elgen,elset=cylinder
1,<node_circum0>,1,1,<node_int>,<node_circum>,<node_circum>
<node_circum>,1,,,<node_int>,<node_circum>,<node_circum>
*shell section, elset=cylinder, material=mat_1
<thickness>,
*material,name=mat_1
*elastic
<young>,<poisson>
*nset,nset=ends
bottom,top
*transform,type=c,nset=ends
1-379
0.0,0.0,0.0,0.0,0.0,1.0
*boundary
ends,1,2
ends,4,4
ends,6,6
1,3
**
*step,nlgeom,inc=10
static preload for internal pressure
*static
1.0,1.0
*monitor,node=<node_tot>, dof=3
*dload
cylinder,p,<int_press>
*end step
**
*step,nlgeom,inc=60
postbuckling (riks) analysis
*static,riks
0.05,1.0,,0.05,,<node_tot>,3,-0.1
*monitor,node=<node_tot>,dof=3
*controls,parameter=field,field=global
1.e-5,
*cload
top,3,<p>
bottom,3,<pn>
*node file,freq=1,nset=top
u,cf
*output,field,variable=preselect,freq=10
*output,history,freq=1
*node output,nset=top
u,cf
*end step
1.3 Forming analyses

1.3.1 Upsetting of a cylindrical billet in ABAQUS/Standard:
quasi-static analysis with rezoning
This example illustrates the use of the rezoning capabilities of ABAQUS/Standard in a metal forming
application. The same problem is analyzed using the coupled temperature-displacement elements in
1-380
`Ùpsetting of a cylindrical billet: coupled temperature-displacement and adiabatic analysis, '' Section
1.3.17. Coupled temperature-displacement elements are included in this example only for rezoning
verification purposes; no heat generation occurs in these elements for this example. The same test case
is done with ABAQUS/Explicit in `Ùpsetting of a cylindrical billet in ABAQUS/Explicit,'' Section
1.3.2.
When the strains become large in geometrically nonlinear analysis, the elements often become so
severely distorted that they no longer provide a good discretization of the problem. When this occurs,
it is necessary to "rezone": to map the solution onto a new mesh that is better designed to continue the
analysis. The procedure is to monitor the distortion of the mesh--for example, by observing deformed
configuration plots--and decide when the mesh needs to be rezoned. At that point a new mesh must be
generated using the mesh generation options in ABAQUS or some external mesh generator. The results
file output is useful in this context since the current geometry of the model can be extracted from the
data in the results file. Once a new mesh is defined, the analysis is continued by beginning a new
problem using the solution from the old mesh at the point of rezoning as initial conditions. This is
done by including the *MAP SOLUTION option and specifying the step number and increment
number at which the solution should be read from the previous analysis. ABAQUS interpolates the
solution from the old mesh onto the new mesh to begin the new problem. This technique provides
considerable generality. For example, the new mesh might be more dense in regions of high-strain
gradients and have fewer elements in regions that are distorting rigidly--there is no restriction that the
number of elements be the same or that element types agree between the old and new meshes. In a
typical practical analysis of a manufacturing process, rezoning may have to be done several times
because of the large shape changes associated with such a process.
The interpolation technique used in rezoning is a two-step process. First, values of all solution
variables are obtained at the nodes of the old mesh. This is done by extrapolation of the values from
the integration points to the nodes of each element and averaging those values over all elements
abutting each node. The second step is to locate each integration point in the new mesh with respect to
the old mesh (this assumes all integration points in the new mesh lie within the bounds of the old
mesh: warning messages are issued if this is not so, and new model solution variables at the integration
point are set to zero). The variables are then interpolated from the nodes of the element in the old mesh
to the location in the new mesh. All solution variables are interpolated automatically in this way so
that the solution can proceed on the new mesh.
Whenever a model is rezoned, it can be expected that there will be some discontinuity in the solution
because of the change in the mesh. If the discontinuity is significant, it is an indication that the meshes
are too coarse or that the rezoning should have been done at an earlier stage before too much distortion
occurred.
Geometry and model

The geometry is the standard test case of Lippmann (1979) and is defined in `Ùpsetting of a
cylindrical billet: coupled temperature-displacement and adiabatic analysis, '' Section 1.3.17. It is a
circular billet, 30 mm long, with a radius of 10 mm, compressed between two flat, rigid dies that are
defined to be perfectly rough.
1-381
The mesh used to begin the analysis is shown in Figure 1.3.1-1. The finite element model is
axisymmetric and includes the top half of the billet only since the middle surface of the billet is a plane
of symmetry. Element type CAX4R is used: this is a 4-node quadrilateral with a single integration
point and "hourglass control" to control spurious mechanisms caused by the fully reduced integration.
The element is chosen here because it is relatively inexpensive for problems involving nonlinear
constitutive behavior since the material calculations are only done at one point in each element.
The contact between the top and lateral exterior surfaces of the billet and the rigid die is modeled with
the *CONTACT PAIR option. The billet surface is defined by means of the *SURFACE option. The
rigid die is modeled as an analytical rigid surface with the *SURFACE option in conjunction with the
*RIGID BODY option. The mechanical interaction between the contact surfaces is assumed to be
nonintermittent, rough frictional contact. Therefore, two suboptions are used with the *SURFACE
INTERACTION property option: the *FRICTION, ROUGH suboption to enforce a no slip constraint
between the two surfaces, and the *SURFACE BEHAVIOR, NO SEPARATION suboption to ensure
that separation does not occur once contact has been established.
No mesh convergence studies have been done, but the agreement with the results given in Lippmann
(1979) suggests that the meshes used here are good enough to provide reasonable predictions of the
overall force on the dies.
Material
The material behavior is similar to that used in `Ùpsetting of a cylindrical billet: coupled
temperature-displacement and adiabatic analysis,'' Section 1.3.17, except that rate dependence of the
yield stress is not included. Thermal properties are not needed in this case since the analysis is
mechanical only (we assume the loading is applied so slowly that the response is isothermal).
Boundary conditions and loading

Kinematic boundary conditions are symmetry on the axis (nodes at r =0, in node set AXIS, have
ur =0 prescribed), symmetry about z =0 (all nodes at z =0, in node set MIDDLE, have uz =0
prescribed). The node on the top surface of the billet that lies on the symmetry axis is not part of the
node set AXIS to avoid overconstraint: the radial motion of this node is already constrained by a no
slip frictional constraint (see ``Common difficulties associated with contact modeling,'' Section 21.10.1
of the ABAQUS/Standard User's Manual). The uz -displacement of the rigid body reference node for
the die is prescribed as having a constant velocity in the axial direction so that the total displacement
of the die is -9 mm over the history of the upsetting.
The first analysis is done in two steps so that the first step can be stopped at a die displacement
corresponding to 44% upsetting. The second step carries the first analysis to 60% upsetting. The
second analysis restarts from the first step of the first analysis with a new mesh. A FORTRAN routine
is used to extract the coordinates of the nodes along the outer boundary of the original mesh at 44%
upsetting. These coordinates are then used to define the outer boundary of the new mesh.

The results from the first mesh are illustrated in Figure 1.3.1-2. This figure shows the configuration
1-382
and plastic strain magnitude that are predicted at 44% upsetting (73.3% of the total die displacement).
The folding of the top outside surface of the billet onto the die is clearly visible, as well as the severe
straining of the middle of the specimen. At this point the mesh is rezoned. The new mesh for the
rezoned model is shown in Figure 1.3.1-3. It is based on placing nodes on straight lines between the
outer surface of the billet and the axis of the billet. The final configuration and plastic strain
magnitudes predicted with rezoning are shown in Figure 1.3.1-4. Figure 1.3.1-5 shows the predictions
of the total upsetting force versus displacement of the die. The results shown on the plot include the
results for the analysis that includes the rezoning and the data obtained when the original mesh is used
for the entire analysis. The results show that the rezoning of the mesh does not have a significant
effect, in this case, on the overall die force. The results compare well with the rate independent results
obtained by Taylor (1981).
Input files
rezonebillet_cax4r.inp
Original CAX4R mesh.
rezonebillet_cax4r_rezone.inp
Rezoned CAX4R mesh; requires the external file generated by rezonebillet_fortran_cax4r.f.
rezonebillet_fortran_cax4r.f
FORTRAN routine used to access the results file of rezonebillet_cax4r.inp and generate a file
containing the nodal coordinates of the outer boundary at 44% upsetting.
rezonebillet_cax4i.inp
Original CAX4I mesh.
rezonebillet_cax4i_rezone.inp
Rezoned CAX4I mesh. The FORTRAN routine given in file rezonebillet_fortran_cax4r.f is not
used as part of this analysis sequence.
rezonebillet_cgax4r.inp
Original CGAX4R mesh.
rezonebillet_cgax4r_rezone.inp
Rezoned CGAX4R mesh; requires the external file generated by rezonebillet_fortran_cax4r.f.
rezonebillet_cgax4t.inp
Original CGAX4T mesh.
rezonebillet_cgax4t_rezone.inp
Rezoned CGAX4T mesh; requires the external file generated by rezonebillet_fortran_cgax4t.f.
rezonebillet_fortran_cgax4t.f
FORTRAN routine used to access the results file from file rezonebillet_cgax4t.inp and generate a
1-383
file containing the nodal coordinates of the outer boundary at 44% upsetting.
rezonebillet_deftorigid.inp
Rigid die simulated by declaring deformable elements (SAX1) as rigid. The billet is meshed with
CAX4R elements.
rezonebillet_deftorigid_rezone.inp
Rezoned CAX4R mesh. The rigid die is simulated by declaring deformable elements (SAX1) as
rigid.
References
· Lippmann, H., Metal Forming Plasticity, Springer-Verlag, Berlin, 1979.
· Taylor, L. M., "A Finite Element Analysis for Large Deformation Metal Forming Problems
Involving Contact and Friction," Ph.D. Thesis, U. of Texas at Austin, 1981.
Figures
Figure 1.3.1-1 Axisymmetric upsetting example: initial mesh.
Figure 1.3.1-2 Deformed configuration and plastic strain at 44% upset.
1-384
Figure 1.3.1-3 New mesh at 44% upset.
1-385
Figure 1.3.1-4 Deformed configuration and plastic strain of original mesh at 60% upset.
Figure 1.3.1-5 Force-deflection response for cylinder upsetting. (Results from the rezoned mesh start
at 73.6% of applied displacement.)
1-386
Sample listings
1-387
Listing 1.3.1-1
*HEADING
AXISYMMETRIC UPSETTING PROBLEM
REZONING
*NODE,NSET=RSNODE
9999,0.,.015
*NODE
1,
13,.01
1201,0.,.015
1213,.01,.015
*NGEN,NSET=MIDDLE
1,13
*NGEN,NSET=TOP
1201,1213
*NFILL
MIDDLE,TOP,12,100
*NSET,NSET=AXIS,GENERATE
1,1201,100
*NSET,NSET=OUTER,GENERATE
13,1013,100
*NSET,NSET=NAXIS,GENERATE
1,1101,100
*ELEMENT,TYPE=CAX4R,ELSET=METAL
1,1,2,102,101
*ELGEN,ELSET=METAL
1,12,1,1,12,100,100
*ELSET,ELSET=ECON1,GENERATE
1101,1112,1
12,1112,100
START,.020,.015
LINE,-.001,.015
*SURFACE,NAME=ASURF
ECON1,S3
ECON2,S2
ASURF,BSURF
1-388
*FRICTION,ROUGH
*SURFACE BEHAVIOR,NO SEPARATION
*SOLID SECTION,ELSET=METAL,MATERIAL=EL
*MATERIAL,NAME=EL
*ELASTIC
200.E9,.3
*PLASTIC
7.E8,0.00
3.7E9,10.0
*BOUNDARY
MIDDLE,2
NAXIS,1
*STEP,INC=200,AMPLITUDE=RAMP,NLGEOM
73.3 PERCENT OF DIE DISPLACEMENT
*STATIC
0.015,1.
*BOUNDARY
9999,1
9999,6
9999,2,,-.0066
*EL PRINT, ELSET=METAL,FREQUENCY=40
S,MISES
E,
PEEQ,
*NODE PRINT, FREQUENCY=10
*NODE FILE,NSET=RSNODE
U,RF
COORD,
*OUTPUT, FIELD, OP=NEW, FREQUENCY=9999
*NODE OUTPUT
U,
*END STEP
100 PERCENT OF DIE DISPLACEMENT
*STATIC
0.015,1.
*BOUNDARY
9999,1
9999,6
1-389
9999,2,,-.009
S,MISES
E,
PEEQ,
U,RF
*END STEP
1-390
Listing 1.3.1-2
*HEADING
AXISYMMETRIC UPSETTING PROBLEM
REZONED
*NODE,NSET=RSNODE
9999,0.,.0084
*NODE
1,
1001,0.,.0084
*NODE,NSET=OUTER,INPUT=BOUNDARY.OUT
*NGEN,NSET=AXIS
1,1001,100
*NSET,NSET=NAXIS,GENERATE
1,901,100
*NSET,NSET=MIDDLE,GENERATE
1,13
*NSET,NSET=TOP,GENERATE
1001,1013
*NFILL
AXIS,OUTER,12,1
*ELEMENT,TYPE=CAX4R,ELSET=METAL
1,1,2,102,101
*ELGEN,ELSET=METAL
1,12,1,1,10,100,100
901,912,1
12,912,100
START,.020,.0084
LINE,-.001,.0084
*SURFACE,NAME=ASURF
ECON1,S3
ECON2,S2
ASURF,BSURF
*FRICTION,ROUGH
1-391
*MATERIAL,NAME=EL
*ELASTIC
200.E9,.3
*PLASTIC
7.E8,0.00
3.7E9,10.0
*BOUNDARY
MIDDLE,2
NAXIS,1
*MAP SOLUTION,STEP=1,INC=16
*STATIC
.03,1.
*BOUNDARY
9999,1
9999,6
9999,2,,-.0024
S,MISES
E,
PEEQ,
U,RF
*END STEP
1-392
Listing 1.3.1-3
SUBROUTINE HKSMAIN
C
C PROGRAM READSETS
C
INCLUDE 'aba_param.inc'
C
PARAMETER (MAXNODES = 500)
DIMENSION ARRAY(513), JRRAY(NPRECD,513), NMEMS(MAXNODES)
EQUIVALENCE (ARRAY(1), JRRAY(1,1))
C
INTEGER LRUNIT(2,1)
LOGICAL READNODES
CHARACTER FNAME*80, ASET*8
C
C THIS PROGRAM WILL EXTRACT THE CURRENT COORDINATES OF THE NODES OF THE
C "OUTER" AT THE STEP/INC DEFINED BY THE PARAMETERS K STEP AND K INC. TH
C NUMBERS AND COORDINATES ARE WRITTEN TO THE OUTPUT FILE BOUNDARY.OUT IN
C SUITABLE FOR INPUT INTO AN ABAQUS INPUT FILE.
C
C FOR EXAMPLE rezonebillet_cax4r:
PARAMETER (K_STEP = 1, K_INC = 16, FNAME = 'rezonebillet_cax4r')
C
C FOR EXAMPLE rezonebillet_cgax4t:
C PARAMETER (K_STEP = 1, K_INC = 17, FNAME = 'rezonebillet_cgax4t')
C
C THIS MAY BE USED TO EXTRACT COORDINATES FROM EXA rezonebillet cax4i U
C PARAMETER (K_STEP = 1, K_INC = 40, FNAME = 'rezonebillet_cax4i')
C
LRUNIT(1,1)=8
LRUNIT(2,1)=2
LOUTF=0
NRU = 1
C
CALL INITPF (FNAME, NRU, LRUNIT, LOUTF)
C
JOUT = 6
KEYPRV= 0
KSTEP = 0
KINC = 0
READNODES = .FALSE.
NUMMEM= 0
1-393
C
JUNIT = LRUNIT(1,1)
CALL DBRNU (JUNIT)
C
OPEN (UNIT=6,STATUS='UNKNOWN',FILE='BOUNDARY.OUT')
C
C READ RECORDS FROM RESULTS FILE, UP TO 100000 RECORDS:
C
DO 50 IXX = 1, 99999
CALL DBFILE(0,ARRAY,JRCD)
IF (JRCD .NE. 0 .AND. KEYPRV .EQ. 2001) THEN
WRITE(0,*) 'END OF FILE'
CLOSE (JUNIT)
GOTO 100
ELSE IF (JRCD .NE. 0) THEN
WRITE(0,*) 'ERROR READING FILE'
CLOSE (JUNIT)
GOTO 100
ENDIF
C
KEY=JRRAY(1,2)
C
C RECORD 2000: INCREMENT START RECORD
C
IF(KEY.EQ.2000) THEN
KSTEP = JRRAY(1,8)
KINC = JRRAY(1,9)
END IF
C
C RECORD 1931: NODE SET DEFINITION RECORD
C
IF(KEY.EQ.1931) THEN
C
WRITE(ASET,110) array(3)
110 FORMAT(a8)
IF (ASET(1:5).EQ.'OUTER') THEN
NUMMEM = JRRAY(1,1) - 3
IF (NUMMEM.GT.MAXNODES) THEN
WRITE(0,*)'ERROR: TOO MANY NODES ON RECORD'
WRITE(0,*)' INCREASE MAXNODES TO ',NUMMEM
CLOSE (JUNIT)
GOTO 100
END IF
1-394
DO KMEM = 1,NUMMEM
NMEMS(KMEM)=JRRAY(1,3+KMEM)
END DO
READNODES = .TRUE.
END IF
END IF
C
C RECORD 107: NODAL COORDINATES RECORD
C
IF (KEY.EQ.107 .AND. READNODES) THEN
IF (KSTEP.EQ.K_STEP .AND. KINC.EQ.K_INC) THEN
KNODE = JRRAY(1,3)
DO KMEM = 1,NUMMEM
IF (KNODE.EQ.NMEMS(KMEM)) THEN
WRITE(JOUT,70)KNODE,ARRAY(4),ARRAY(5)
70 FORMAT(I5,2(',',D18.8))
END IF
END DO
END IF
END IF
C
KEYPRV = KEY
C
50 CONTINUE
C
100 CONTINUE
C
STOP
END
1.3.2 Upsetting of a cylindrical billet in ABAQUS/Explicit

The example illustrates the forming of a small, circular billet of metal that is reduced in length by 60%.
This is the standard test case that is defined in Lippmann et al. (1979), so some verification of the
result is available by comparing the results with the numerical results presented in that reference. The
same test case is done with ABAQUS/Standard in `Ùpsetting of a cylindrical billet in
ABAQUS/Standard: quasi-static analysis with rezoning,'' Section 1.3.1. The same problem is analyzed
using the coupled temperature-displacement elements in `Ùpsetting of a cylindrical billet: coupled
temperature-displacement and adiabatic analysis,'' Section 1.3.17.
Problem description
The specimen is a circular billet, 30 mm long with a radius of 10 mm, compressed between flat, rough,
1-395
rigid dies. The finite element model is axisymmetric and includes the top half of the billet only, since
the middle surface of the billet is a plane of symmetry. Axisymmetric continuum elements of types
CAX4R and CAX6M are used to model the billet. The rigid die is modeled in several different ways,
as described below.
1. The die is modeled as an analytical rigid surface using the *SURFACE, TYPE=SEGMENTS and
the *RIGID BODY options. The rigid surface is associated with a rigid body by its specified
reference node.
2. Axisymmetric rigid elements of type RAX2 are used to model the rigid die. Three input files that
use progressively finer meshes to model the billet exist for this case.
3. The die is modeled with RAX2 elements, as in Case 2. However, the die is assigned a mass by
specifying point masses at the nodes of the RAX2 elements. The reference node of the rigid die is
repositioned at its center of mass by specifying POSITION=CENTER OF MASS on the *RIGID
BODY option.
4. The die is modeled with RAX2 elements, as in Case 2. The rigid elements are assigned thickness
and density values such that the mass of the die is the same as in Case 3.
5. The die is modeled with RAX2 elements, as in Case 2. The NODAL THICKNESS parameter is
used on the *RIGID BODY option to specify the thickness of the die at its nodes. The same
thickness value is prescribed as in Case 4.
6. Axisymmetric shell elements of type SAX1 are used to model the die, and they are included in the
rigid body by referring to them on the *RIGID BODY option. The thickness and the material
density of the SAX1 elements is the same as that of the rigid elements in Case 4.
7. The die is modeled with axisymmetric shell elements of type SAX1 and with axisymmetric rigid
elements of type RAX2. The deformable elements are included in the rigid body by referring to
them on the *RIGID BODY option. Both element types have the same thickness and density as in
Case 4.
A coefficient of friction of 1.0 is used between the rigid surface and the billet. This value is large
enough to ensure a no-slip condition so that, when the billet comes in contact with the rigid surface,
there is virtually no sliding between the two.
The material model assumed for the billet is that given in Lippmann et al. (1979). Young's modulus is
200 GPa, Poisson's ratio is 0.3, and the density is 7833 kg/m 3. A rate-independent von Mises
elastic-plastic material model is used, with a yield stress of 700 MPa and a hardening slope of 0.3
GPa.
Kinematic boundary conditions are symmetry on the axis (all nodes at r = 0 have ur = 0 prescribed)
and symmetry about z = 0 (all nodes at z = 0 have uz = 0 prescribed). The rigid body reference node
for the rigid body is constrained to have no rotation or ur -displacement. In Case 1 and Case 2 the
uz -displacement is prescribed using a velocity boundary condition whose value is ramped up to a
velocity of 20 m/s and then held constant until the die has moved a total of 9 mm. In the remaining
cases a concentrated force of magnitude 410 kN is applied in the z-direction at the reference node. The
1-396
magnitude of the concentrated force is such as to ensure that the resulting displacement of the die at
the end of the time period is the same as in Case 1 and Case 2. The total time of the analysis is 0.55
millisec and is slow enough to be considered quasi-static.
For Case 1 several different analyses are performed to compare section control options and mesh
refinement for the billet modeled with CAX4R elements. Table 1.3.2-1 lists the analysis options used.
A coarse mesh (analysis COARSE_SS) and a fine (analysis FINE_SS) mesh are analyzed with the pure
stiffness form of hourglass control (HOURGLASS=STIFFNESS). A coarse mesh (analysis
COARSE_CS) is analyzed with the combined hourglass control. The default section controls, using the
integral viscoelastic form of hourglass control (HOURGLASS=RELAX STIFFNESS), are tested on a
coarse mesh (analysis COARSE). Since this is a quasi-static analysis, the viscous hourglass control
option (HOURGLASS=VISCOUS) should not be used. All other cases use the default section
controls.
In addition, Case 1 of the above problem has been analyzed using CAX6M elements with both coarse
and fine meshes.

Figure 1.3.2-1 through Figure 1.3.2-4 show results for Case 1, where the billet is modeled with
CAX4R elements. The rigid die is modeled using an analytical rigid surface, and the pure stiffness
hourglass control is used. Figure 1.3.2-1 shows the original and deformed shape at the end of the
analysis for the coarse mesh. Figure 1.3.2-2 shows the original and deformed shape at the end of the
analysis for the fine mesh. Figure 1.3.2-3 and Figure 1.3.2-4 show contours of equivalent plastic strain
for both meshes.
Corresponding coarse and fine mesh results for the billet modeled with CAX6M elements are shown in
Figure 1.3.2-5 to Figure 1.3.2-8. The number of nodes used for the coarse and fine CAX6M meshes are
the same as those used for the coarse and fine CAX4R meshes. However, the analysis using CAX4R
elements performs approximately 55% faster than the analysis using CAX6M elements. The equivalent
plastic strain distributions obtained from each analysis compare closely for both the coarse and fine
meshes; the CAX6M elements predict a slightly higher peak value. In both analyses the folding of the
top outside surface of the billet onto the die is clearly visible, as well as the severe straining of the
middle of the specimen.
Figure 1.3.2-9 is a plot of vertical displacement versus reaction force at the rigid surface reference
node for Case 1 with the section control options identified in Table 1.3.2-1. Results for an analysis run
with ABAQUS/Standard (labeled COARSE_STD) are included for comparison. The curves obtained
using CAX4R and CAX6M elements are very close and agree well with independent results obtained
by Taylor (1981). The results from analysis COARSE_SS are virtually the same as the results from
analysis COARSE, but at a much reduced cost; therefore, such analysis options are recommended for
this problem. The results for all the other cases (which use the default section controls but different
rigid surface models) are the same as the results for Case 1 using the default section controls.
Input files
upset_anl_ss.inp
1-397
Coarse mesh (CAX4R elements) of Case 1 using STIFFNESS hourglass control.

upset_fine_anl_ss.inp
Fine mesh of Case 1 (CAX4R elements) usingSTIFFNESS hourglass control.
upset_anl_cs.inp
Coarse mesh of Case 1 (CAX4R elements) using COMBINED hourglass control.
upset_anl.inp
Coarse mesh of Case 1 (CAX4R elements) using the default section control options.
upset_fine_anl.inp
Fine mesh of Case 1 (CAX4R elements) using the default section control options.
upset_case2.inp
Coarse mesh of Case 2 (CAX4R elements).
upset_fine_case2.inp
Fine mesh of Case 2 (CAX4R elements).
upset_vfine_case2.inp
An even finer mesh of Case 2 (CAX4R elements) included to test the performance of the code.
upset_case3.inp
Case 3 using CAX4R elements.
upset_case4.inp
upset_case5.inp
upset_case6.inp
upset_case7.inp
upset_anl_cax6m.inp
Case 1 using the coarse mesh of CAX6M elements.
upset_fine_anl_cax6m.inp
Case 1 using the fine mesh of CAX6M elements.
References
1-398
· H. Lippmann, editor, Metal Forming Plasticity, Springer-Verlag, Berlin, 1979.
Involving Contact and Friction," Ph. D. Dissertation, U. of Texas at Austin, 1981.
Table
Table 1.3.2-1 Analysis options for Case 1 using CAX4R elements.

Analysis Mesh Hourglass
Label Type Control
COARSE_SS coarse STIFFNESS
FINE_SS fine STIFFNESS
COARSE_CS coarse COMBINED
COARSE coarse RELAX
Figures
Figure 1.3.2-1 Undeformed and deformed shape for the coarse mesh (CAX4R) of Case 1 (using the
STIFFNESS hourglass control).
Figure 1.3.2-2 Undeformed and deformed shape for the fine mesh (CAX4R) of Case 1 (using the
1-399
Figure 1.3.2-3 Contours of equivalent plastic strain for the coarse mesh ( CAX4R) of Case 1 (using
the STIFFNESS hourglass control).
Figure 1.3.2-4 Contours of equivalent plastic strain for the fine mesh ( CAX4R) of Case 1 (using the
1-400
Figure 1.3.2-5 Undeformed and deformed shape for the coarse mesh (CAX6M) of Case 1.
Figure 1.3.2-6 Contours of equivalent plastic strain for the coarse mesh ( CAX6M) of Case 1.
Figure 1.3.2-7 Undeformed and deformed shape for the fine mesh (CAX6M) of Case 1.
1-401
Figure 1.3.2-8 Contours of equivalent plastic strain for the fine mesh ( CAX6M) of Case 1.
Figure 1.3.2-9 Comparison of reaction force versus vertical displacement for the different analyses
tested for Case 1.
Sample listings
1-402
Listing 1.3.2-1
*HEADING
AXISYMMETRIC UPSETTING PROBLEM -- COARSE MESH
(WITH RIGID SURFACE)
SECTION CONTROLS USED (HOURGLASS=STIFFNESS)
*NODE
1,
13,.01
1201,0.,.015
1213,.01,.015
*NGEN,NSET=MIDDLE
1,13
*NGEN,NSET=TOP
1201,1213
*NFILL
MIDDLE,TOP,12,100
*NSET,NSET=AXIS,GEN
1,1201,100
*ELEMENT,TYPE=CAX4R,ELSET=BILLET
1,1,2,102,101
*ELGEN,ELSET=BILLET
1,12,1,1,12,100,100
*NODE, NSET=NRIGID
2003,0.01,.02
*SOLID SECTION,ELSET=BILLET,MATERIAL=METAL,
CONTROL=B
*SECTION CONTROLS, HOURGLASS=STIFFNESS, NAME=B
*MATERIAL,NAME=METAL
*ELASTIC
200.E9,.3
*PLASTIC
7.E8,0.00
3.7E9,10.0
*DENSITY
7833.,
*BOUNDARY
MIDDLE,2
AXIS,1
2003,1
2003,3,6
*AMPLITUDE,NAME=RAMP
1-403
0.,0., 2.E-4,1., 5.5E-4,1.

*SURFACE,TYPE=ELEMENT,NAME=BILLET
TOP,S3
SIDE,S2
*SURFACE,NAME=RIGID,TYPE=SEGMENTS
START, 0.02,.015
LINE, 0.00,.015
ANALYTICAL SURFACE =RIGID
*STEP
*DYNAMIC,EXPLICIT
,5.5E-4
*BOUNDARY,TYPE=VELOCITY,AMPLITUDE=RAMP
2003,2,,-20.
*ELSET,ELSET=TOP,GEN
1101,1112,1
*ELSET,ELSET=SIDE,GEN
12,1112,100
*SURFACE INTERACTION,NAME=RIG_BILL
*FRICTION
1.0,
*CONTACT PAIR,INTERACTION=RIG_BILL
RIGID,BILLET
*FILE OUTPUT,NUM=2,TIMEMARKS=YES
*EL FILE
PEEQ,MISES
*HISTORY OUTPUT, TIME=4.E-7
*NODE HISTORY, NSET=NRIGID
U2,RF2
*NODE FILE, NSET=NRIGID
U,RF
*END STEP
1.3.3 Superplastic forming of a rectangular box

In this example we consider the superplastic forming of a rectangular box. The example illustrates the
use of rigid elements to create a smooth three-dimensional rigid surface.
Superplastic metals exhibit high ductility and very low resistance to deformation and are, thus, suitable
for forming processes that require very large deformations. Superplastic forming has a number of
advantages over conventional forming methods. Forming is usually accomplished in one step rather
than several, and intermediate annealing steps are usually unnecessary. This process allows the
1-404
production of relatively complex, deep-shaped parts with quite uniform thickness. Moreover, tooling
costs are lower since only a single die is usually required. Drawbacks associated with this method
include the need for tight control of temperature and deformation rate. Very long forming times make
this method impractical for high volume production of parts.
A superplastic forming process usually consists of clamping a sheet against a die whose surface forms
a cavity of the shape required. Gas pressure is then applied to the opposite surface of the sheet, forcing
it to acquire the die shape.
Rigid surface
The *SURFACE option allows the creation of a rigid faceted surface created from an arbitrary mesh of
three-dimensional rigid elements (either triangular R3D3 or quadrilateral R3D4 elements). See
``Defining analytical rigid surfaces,'' Section 2.3.4 of the ABAQUS/Standard User's Manual, for a
discussion of smoothing of master surfaces. ABAQUS automatically smoothes any discontinuous
surface normal transitions between the surface facets.
Solution-dependent amplitude
One of the main difficulties in superplastically forming a part is the control of the processing
parameters. The temperature and the strain rates that the material experiences must remain within a
certain range for superplasticity to be maintained. The former is relatively easy to achieve. The latter is
more difficult because of the unknown distribution of strain rates in the part. The manufacturing
process must be designed to be as rapid as possible without exceeding a maximum allowable strain
rate at any material point. For this purpose ABAQUS has a feature that allows the loading (usually the
gas pressure) to be controlled by means of a solution-dependent amplitude. The options invoked are
*AMPLITUDE, DEFINITION=SOLUTION DEPENDENT and a target maximum *CREEP STRAIN
RATE CONTROL. In the loading options the user specifies a reference value. The amplitude
definition requires an initial, a minimum, and a maximum load multiplier. During a *VISCO procedure
ABAQUS will then monitor the maximum creep strain rate and compare it with the target value. The
load amplitude is adjusted based on this comparison. This controlling algorithm is simple and
relatively crude. The purpose is not to follow the target values exactly but to obtain a practical loading
schedule.
Geometry and model

The example treated here corresponds to superplastic forming of a rectangular box whose final
dimensions are 1524 mm (60 in) long by 1016 mm (40 in) wide by 508 mm (20 in) deep with a 50.8
mm (2 in) flange around it. All fillet radii are 101.6 mm (4 in). The box is formed by means of a
uniform fluid pressure.
A quarter of the blank is modeled using 704 membrane elements of type M3D4R. These are bilinear
membrane elements with fully reduced integration and hourglass control. The initial dimensions of the
blank are 1625.6 mm (64 in) by 1117.6 mm (44 in), and the thickness is 3.175 mm (0.125 in). The
blank is clamped at all its edges. The flat initial configuration of the membrane model is entirely
singular in the normal direction unless it is stressed in biaxial tension. This difficulty is prevented by
applying a small biaxial initial stress of 6.89 kPa (1 lb/in 2) by means of the *INITIAL CONDITIONS,
1-405
TYPE=STRESS option.
The female die is modeled as a rigid body and is meshed with rigid R3D3 elements. The rigid surface
is defined with the *SURFACE option by grouping together those faces of the 231 R3D3 elements
used to model the die that face the contact direction. See Figure 1.3.3-1 for an illustration of the rigid
element mesh.
To avoid having points "fall off" the rigid surface during the course of the analysis, more than a quarter
of the die has been modeled, as shown in Figure 1.3.3-2. It is always a good idea to extend the rigid
surface far enough so that contacting nodes will not slide off the master surface.
By default, ABAQUS generates a unique normal to the rigid surface at each node point, based on the
average of the normals to the elements sharing each node. There are times, however, when the normal
to the surface should be specified directly. This is discussed in ``Node definition,'' Section 2.1.1 of the
ABAQUS/Standard User's Manual. In this example the flange around the box must be flat to ensure
compatibility between the originally flat blank and the die. Therefore, an outer normal (0, 1, 0) has
been specified at the 10 nodes that make up the inner edge of the flange. This is done by entering the
direction cosines after the node coordinates. The labels of these 10 vertices are 9043, 9046, 9049,
9052, 9089, 9090, 9091, 9121, 9124, and 9127; and their definitions can be found in
superplasticbox_node.inp.
Material
The material in the blank is assumed to be elastic-viscoplastic, and the properties roughly represent the
2004 (Al-6Cu-0.4Zr)-based commercial superplastic aluminum alloy Supral 100 at 470°C. It has a
Young's modulus of 71 GPa (10.3 ´ 106 lb/in2) and a Poisson's ratio of 0.34. The flow stress is
assumed to depend on the plastic strain rate according to
¾ f = A("_pl )1=2 ;
where A is 179.2 MPa (26. ´ 103 lb/in2) and the time is in seconds.

We perform two analyses to compare constant pressure loading and a pressure schedule automatically
adjusted to achieve a maximum strain rate of 0.02/sec. In the constant load case the prestressed blank
is subjected to a rapidly applied external pressure of 68.8 kPa (10 lb/in 2), which is then held constant
for 3000 sec until the box has been formed. In the second case the prestressed blank is subjected to a
rapidly applied external pressure of 1.38 kPa (0.2 lb/in 2). The pressure schedule is then chosen by
ABAQUS.
The initial application of the pressures is assumed to occur so quickly that it involves purely elastic
response. This is achieved by using the *STATIC procedure. The creep response is developed in a
second step using the *VISCO procedure.
During the *VISCO step the parameter CETOL controls the time increment and, hence, the accuracy of
the transient creep solution. ABAQUS compares the equivalent creep strain rate at the beginning and
1-406
the end of an increment. The difference should be less than CETOL divided by the time increment.
Otherwise, the increment is reattempted with a smaller time increment. The usual guideline for setting
CETOL is to decide on an acceptable error in stress and convert it to an error in strain by dividing by
the elastic modulus. For this problem we assume that moderate accuracy is required and choose
CETOL as 0.5%. In general, larger values of CETOL allow ABAQUS to use larger time increments,
resulting in a less accurate and less expensive analysis.
In the automatic scheduling analysis the pressure is referred to an amplitude that allows for a
maximum pressure of 1.38 MPa (200 lb/in 2) and a minimum pressure of 0.138 kPa (0.02 lb/in 2). The
target creep strain rate is a constant entered using the *CREEP STRAIN RATE CONTROL option.

Figure 1.3.3-3 through Figure 1.3.3-5show a sequence of deformed configurations during the
automatically controlled forming process. The stages of deformation are very similar in the constant
load process. However, the time necessary to obtain the deformation is much shorter with automatic
loading--the maximum allowable pressure is reached after 83.3 seconds. The initial stages of the
deformation correspond to inflation of the blank because there is no contact except at the edges of the
box. Contact then occurs at the box's bottom, with the bottom corners finally filling. Although there is
some localized thinning at the bottom corners, with strains on the order of 100%, these strains are not
too much larger than the 80% strains seen on the midsides.
Figure 1.3.3-6 shows the equivalent plastic strain at the end of the process. The constant load case
provides similar results.
Figure 1.3.3-7 shows the evolution in time of the ratio between the maximum creep strain rate found in
the model and the target creep strain rate. The load applied initially produces a low maximum creep
strain rate at the beginning of the analysis. At the end the maximum creep strain rate falls substantially
as the die cavity fills up. Although the curve appears very jagged, it indicates that the maximum peak
strain rate is always relatively close to the target value. This is quite acceptable in practice. Figure
1.3.3-8 shows the pressure schedule that ABAQUS calculates for this problem. For most of the time,
while the sheet does not contact the bottom of the die, the pressure is low. Once the die starts
restraining the deformation, the pressure can be increased substantially without producing high strain
rates.
Input files
superplasticbox_constpress.inp
Constant pressure main analysis.
superplasticbox_autopress.inp
Automatic pressurization main analysis.
superplasticbox_node.inp
Node definitions for the rigid elements.
superplasticbox_element.inp
1-407
Element definitions for the rigid R3D3 elements.
Figures
Figure 1.3.3-1 Rigid surface for die.
Figure 1.3.3-2 Initial position of blank with respect to die.
Figure 1.3.3-3 Automatic loading: deformed configuration after 34 sec in Step 2.
1-408
1-409
Figure 1.3.3-6 Automatic loading: inelastic strain in the formed box.
Figure 1.3.3-7 History of ratio between maximum creep strain rate and target creep strain rate.
Figure 1.3.3-8 History of pressure amplitude.
1-410
Sample listings
1-411
Listing 1.3.3-1
*HEADING
SUPERPLASTIC FORMING OF BOX WITH
MEMBRANES - CONSTANT LOADING
**
** PLATE DEFINITION
**
*NODE
1, 4.0, 20.0, -4.0
23, 26.0, 20.0, -4.0
3201, 4.0, 20.0, -36.0
3223, 26.0, 20.0, -36.0
*NGEN, NSET=EDGE3
1,3201,100
*NGEN, NSET=EDGE5
23,3223,100
*NFILL, NSET=PLATE
EDGE3,EDGE5,22,1
*NSET, NSET=EDGE1, GENERATE
1,23,1
3201,3223,1
*ELEMENT, TYPE=M3D4, ELSET=PLATE1
1,1,2,102,101
*ELGEN, ELSET=PLATE1
1,21,1,1,31,100,100
*NSET,ELSET=PLATE1, NSET=NCONT
22,22,23,123,122
3101,3101,3102,3202,3201
22,32,100,100
3101,21,1,1
*ELSET,ELSET=PLATE
PLATE1,PLATE2
*MEMBRANE SECTION, ELSET=PLATE, MATERIAL=SUPRAL
.125,
**
** MATERIAL IS CLOSE TO SUPRAL100 AT 470C
**
*MATERIAL, NAME=SUPRAL
1-412
*ELASTIC
10.3E6, .34
*CREEP, LAW=TIME
1.48E-9, 2., 0.
**
** CONTACT surface
**
*NODE, NSET=DIE
10000, 0.0, 0.0, 0.0
*RIGID BODY, ELSET=ERIGID, REFNODE=10000
*NODE, INPUT=superplasticbox_node.inp
*ELEMENT,TYPE=R3D3,ELSET=ERIGID,
INPUT=superplasticbox_element.inp
*SURFACE,NAME=DIE
ERIGID,SPOS
*SURFACE,TYPE=NODE,NAME=SLAVES
NCONT,
*CONTACT PAIR, INTERACTION=DIE_NODE, smooth=0.2
SLAVES,DIE
*SURFACE INTERACTION,NAME=DIE_NODE
**
** BOUNDARY AND INITIAL CONDITIONS
**
*BOUNDARY
EDGE1,3
EDGE7,2,3
EDGE3,1
EDGE5,1,2
10000,1,6
*INITIAL CONDITIONS, TYPE=STRESS
PLATE, 1.0, 1.0
*NSET,NSET=NSELECT
101,121,1001,1101,1507,2509
1,113,1211,1308,1408,1508,1509,1705
**
** STEP 1
**
*STEP, INC=50, NLGEOM, unsymm=yes
*STATIC
1.E-4,1.0,
*DLOAD
PLATE,P,-10.
*CONTACT PRINT, FREQUENCY=100
1-413
*CONTACT FILE, FREQUENCY=100, NSET=NSELECT

*EL PRINT, ELSET=PLATE , FREQUENCY=100
S, E
CE,
SINV,
*PRINT, CONTACT=YES
*NODE PRINT, NSET=EDGE3, FREQUENCY=100
U,
*NODE FILE, NSET=EDGE3, FREQUENCY=100
U,
*END STEP
**
** STEP 2
**
*VISCO, CETOL=0.005
0.005, 3000.0
*DLOAD
PLATE,P,-10.
*END STEP
1-414
Listing 1.3.3-2
*HEADING
SUPERPLASTIC FORMING OF BOX WITH
MEMBRANES - AUTOMATIC LOADING
**
** PLATE DEFINITION
**
*NODE
1, 4.0, 20.0, -4.0
23, 26.0, 20.0, -4.0
3201, 4.0, 20.0, -36.0
3223, 26.0, 20.0, -36.0
*NGEN, NSET=EDGE3
1,3201,100
*NGEN, NSET=EDGE5
23,3223,100
*NFILL, NSET=PLATE
EDGE3,EDGE5,22,1
1,23,1
3201,3223,1
*NSET, NSET=CENTER
1,
1,1,2,102,101
1,21,1,1,31,100,100
*NSET,ELSET=PLATE1, NSET=NCONT
22,22,23,123,122
3101,3101,3102,3202,3201
22,32,100,100
3101,21,1,1
*ELSET,ELSET=PLATE
PLATE1,PLATE2
*MEMBRANE SECTION, ELSET=PLATE, MATERIAL=SUPRAL
.125,
**
** MATERIAL IS CLOSE TO SUPRAL100 AT 470C
1-415
**
*MATERIAL, NAME=SUPRAL
*ELASTIC
10.3E6, .34
*CREEP, LAW=TIME
1.48E-9, 2., 0.
**
** CONTACT surface
**
*NODE, NSET=DIE
10000, 0.0, 0.0, 0.0
*RIGID BODY, ELSET=ERIGID, REFNODE=10000
*NODE, INPUT=superplasticbox_node.inp
*ELEMENT,TYPE=R3D3,ELSET=ERIGID,
INPUT=superplasticbox_element.inp
*SURFACE,NAME=DIE
ERIGID,SPOS
*SURFACE,type=node,NAME=SLAVES
NCONT,
*CONTACT PAIR, INTERACTION=DIE_NODE, smooth=0.2
SLAVES,DIE
*SURFACE INTERACTION,NAME=DIE_NODE
**
** BOUNDARY AND INITIAL CONDITIONS
**
*BOUNDARY
EDGE1,3
EDGE7,2,3
EDGE3,1
EDGE5,1,2
10000,1,6
*INITIAL CONDITIONS, TYPE=STRESS
PLATE, 1.0, 1.0
*AMPLITUDE,DEFINITION=SOLUTION DEPENDENT,NAME=AUTO
1.,0.1,1000.
*NSET,NSET=NSELECT
101,121,1001,1101,1507,2509
1,113,1211,1308,1408,1508,1509,1705
**
** STEP 1
**
*STATIC
1-416
2.E-3,1.0,
*DLOAD
PLATE,P,-0.2
*CONTACT FILE, FREQUENCY=100, NSET=NSELECT
*EL PRINT, ELSET=PLATE , FREQUENCY=100
S, E
CE,
SINV,
*PRINT, CONTACT=YES
*NODE PRINT, NSET=EDGE3, FREQUENCY=100
U,
*NODE FILE, NSET=EDGE3, FREQUENCY=100
U,
*END STEP
**
** STEP 2
**
*VISCO, CETOL=0.005
0.2, 2000.0
*DLOAD,AMPLITUDE=AUTO
PLATE,P,-0.2
*CREEP STRAIN RATE CONTROL, ELSET=PLATE,
AMPLITUDE=AUTO
0.02 ,
** TO WRITE THE AUTOMATIC SOLUTION CONTROL
** VARIABLES AMPCU AND RATIO TO THE RESULTS
** FILE EVERY INCREMENT SUCH FILE HAS TO BE
** ACTIVATED
*NODE FILE, NSET=CENTER, FREQUENCY=1
U,
*END STEP
1.3.4 Stretching of a thin sheet with a hemispherical punch

Stamping of sheet metals by means of rigid punches and dies is a standard manufacturing process. In
most bulk forming processes the loads required for the forming operation are often the primary
concern. However, in sheet forming the prediction of strain distributions and limit strains (which
define the onset of local necking) are most important. Such analysis is complicated in that it requires
consideration of large plastic strains during deformation, an accurate description of material response
including strain hardening, the treatment of a moving boundary that separates the region in contact
1-417
with the punch head from the unsupported one, and the inclusion of friction between the sheet and the
punch head.
The stretching of a thin circular sheet with a hemispherical punch is considered in this example.
Geometry and model

The geometry of this problem is shown in Figure 1.3.4-1. The sheet being stretched has a clamping
radius, r0 , of 59.18 mm. The radius of the punch, rp , is 50.8 mm; the die radius, rd , is 6.35 mm; and
the initial thickness of the sheet, t0 , is 0.85 mm. Such a sheet has been tested experimentally by Ghosh
and Hecker (1975) and has been analyzed by Wang and Budiansky (1978) using an axisymmetric
membrane shell finite element formulation. The analysis is conducted statically in ABAQUS/Standard
and dynamically in ABAQUS/Explicit such that inertial forces are relatively small. The initial
configuration for the ABAQUS/Explicit analysis is shown in Figure 1.3.4-2.
The sheet, the punch, and the die are modeled as separate parts, each instanced once. As an
axisymmetric problem in ABAQUS/Standard the sheet is modeled using 50 elements of type SAX1 (or
MAX1) or 25 elements of type SAX2 (or MAX2). The ABAQUS/Explicit model uses 50 elements of
type SAX1. Mesh convergence studies (not reported here) have been done and indicate that these
meshes give acceptably accurate results for most of the values of interest. To test the three-dimensional
membrane and shell elements in ABAQUS/Standard, a 10° sector is modeled using 100 elements of
type S4R, S4, or M3D4R or 25 elements of type M3D9R. All these meshes are reasonably fine; they
are used to obtain good resolution of the moving contact between the sheet and the dies. In the
ABAQUS/Standard shell models nine integration points are used through the thickness of the sheet to
ensure the development of yielding and elastic-plastic bending response; in ABAQUS/Explicit five
integration points are used through the thickness of the sheet.
The rigid punch and die are modeled in ABAQUS/Standard as analytical rigid surfaces with the
*SURFACE option in conjunction with the *RIGID BODY option. The top and bottom surfaces of the
sheet are defined with the *SURFACE option. In ABAQUS/Explicit the punch and die are modeled as
rigid bodies using the *RIGID BODY option; the surface of the punch and die are modeled either by
analytical rigid surfaces or RAX2 elements. The rigid surfaces are offset from the blank by half of the
thickness of the blank because the contact algorithm in ABAQUS/Explicit takes the shell thickness
into account.
Material properties
The material (aluminum-killed steel) is assumed to satisfy the Ramberg-Osgood relation between true
stress and logarithmic strain:
" = ¾=E + (¾=K )n ;
where Young's modulus, E, is 206.8 GPa; the reference stress value, K, is 0.510 GPa; and the
work-hardening exponent, n, is 4.76. The material is assumed to be linear elastic below a 0.5% offset
yield stress of 170.0 MPa and the stress-strain curve beyond that value is defined in piecewise linear
segments using the *PLASTIC option. (The 0.5% offset yield stress is defined from the
Ramberg-Osgood fit by taking (" ¡ ¾=E ) to be 0.5% and solving for the stress.) Poisson's ratio is
1-418
0.3.
The membrane element models in ABAQUS/Standard are inherently unstable in a static analysis
unless some prestress is present in the elements prior to the application of external loading. Therefore,
an equibiaxial initial stress condition equal to 5% of the initial yield stress is prescribed for the
membrane elements in ABAQUS/Standard.
Contact interactions
The contact between the sheet and the rigid punch and the rigid die is modeled with the *CONTACT
PAIR option. The mechanical interaction between the contact surfaces is assumed to be frictional
contact, with a coefficient of friction of 0.275 in ABAQUS/Standard and 0.265 in ABAQUS/Explicit.
Loading
The ABAQUS/Standard analysis is carried out in six steps; the ABAQUS/Explicit analysis is carried
out in four steps. In each of the first four steps of the ABAQUS/Standard analysis either the die or the
punch head is moved using the *BOUNDARY option. In ABAQUS/Explicit the velocity of the punch
head is prescribed using the *BOUNDARY option; the magnitude of the velocity is specified with the
*AMPLITUDE option. It is ramped up to 30 m/s at 1.24 milliseconds during the first step and then
kept constant until time reaches 1.57 milliseconds at the end of the second step. It is then ramped down
to zero at a time of 1.97 milliseconds at the end of the third step. In the first step of the
ABAQUS/Standard analysis the die is moved so that it just touches the sheet. In the next three steps
(the first three steps of the ABAQUS/Explicit analysis) the punch head is moved toward the sheet
through total distances of 18.6 mm, 28.5 mm, and 34.5 mm, respectively. The purpose of these three
steps is to compare the results with those provided experimentally by Ghosh and Hecker for these
punch displacements. More typically the punch would be moved through its entire travel in one step.
Two final steps are included in the ABAQUS/Standard analysis. In the first step the metal sheet is held
in place and the contact pairs are removed from the model with the *MODEL CHANGE,
TYPE=CONTACT PAIR, REMOVE option. In the second step the original boundary conditions for
the metal sheet are reintroduced for springback analysis. However, this springback step is not included
for the analyses using membrane elements, since these elements do not have any bending stiffness and
residual bending stress is often a key determinant of springback.
In the final step of the ABAQUS/Explicit analysis the punch head is moved away from the sheet for
springback analysis. A viscous pressure load is applied to the surface of the shell during this step to
damp out transient wave effects so that quasi-static equilibrium can be reached quickly. This effect
happens within approximately 2 milliseconds from the start of unloading. The coefficient of viscous
pressure is chosen to be 0.35 MPa sec/m, approximately 1% of the value of ½cd , where ½ is the
material density of the sheet and cd is the dilatational wave speed. A value of viscous pressure of ½cd
would absorb all the energy in a pressure wave. For typical structural problems choosing a small
percentage of this value provides an effective way of minimizing ongoing dynamic effects. Static
equilibrium is reached when residual stresses in the sheet are reasonably constant over time.
1-419
Figure 1.3.4-2 shows the initial, undeformed profile of the blank, the die, and the punch. Figure 1.3.4-3
illustrates the deformed sheet and the punch and the die. Figure 1.3.4-4 shows a plot of the same
system after the punch is lifted back, showing the springback of the sheet.
Figure 1.3.4-5 and Figure 1.3.4-6 show the distribution of nominal values of radial and circumferential
membrane strain in the sheet for 18.6 mm punch head displacement. Figure 1.3.4-7 and Figure 1.3.4-8
show the strain distributions at a punch head displacement of 28.5 mm, and Figure 1.3.4-9 and Figure
1.3.4-10 show the strain distributions at a punch head displacement of 34.5 mm. The strain
distributions for the SAX1 models compare well with those obtained experimentally by Ghosh and
Hecker (1975) and those obtained numerically by Wang and Budiansky (1978), who used a membrane
shell finite element formulation. The important phenomenon of necking during stretching is
reproduced at nearly the same location, although slightly different strain values are obtained. Draw
beads are used to hold the edge of the sheet in the experiment, but in this analysis the sheet is simply
clamped at its edge. Incorporation of the draw bead boundary conditions may further improve the
correlation with the experimental data.
A spike can be observed in the radial strain distribution toward the edge of the sheet in some of the
ABAQUS/Standard shell models. This strain spike is the result of the sheet bending around the die.
The spike is not present in the membrane element models since they possess no bending stiffness.
The results obtained with the axisymmetric membrane models are compared with those obtained from
the axisymmetric shell models and were found to be in good agreement.
These analyses assume values of 0.265 or 0.275 for the coefficient of friction. Ghosh and Hecker do
not give a value for their experiments, but Wang and Budiansky assume a value of 0.17. The
coefficient of friction has a marked effect on the peak strain during necking and may be a factor
contributing to the discrepancy of peak strain results during necking. The values used in these analyses
have been chosen to provide good correlation with the experimental data.
The distributions of the residual stresses on springback of the sheet in ABAQUS/Explicit are shown in
Figure 1.3.4-11 and Figure 1.3.4-12.
Input files

thinsheetstretching_m3d4r.inp
Element type M3D4R.
thinsheetstretching_m3d9r.inp
Element type M3D9R.
thinsheetstretching_max1.inp
Element type MAX1.
thinsheetstretching_max2.inp
Element type MAX2.
1-420
thinsheetstretching_s4.inp
Element type S4.
thinsheetstretching_s4r.inp
Element type S4R.
thinsheetstretching_sax1.inp
Element type SAX1.
thinsheetstretching_sax2.inp
Element type SAX2.
thinsheetstretching_restart.inp
Restart of thinsheetstretching_sax2.inp.

hemipunch_anl.inp
Model using analytical rigid surfaces to describe the rigid surface.
hemipunch.inp
Model using rigid elements to describe the rigid surface.
References
· Ghosh, A. K., and S. S. Hecker, ``Failure in Thin Sheets Stretched Over Rigid Punches ,''
Metallurgical Transactions, vol. 6A, pp. 1065-1074, 1975.
· Wang, N. M., and B. Budiansky, `Ànalysis of Sheet Metal Stamping by a Finite Element
Method,'' Journal of Applied Mechanics, vol. 45, pp. 73-82, 1978.
Figures
Figure 1.3.4-1 Configuration and dimensions for hemispherical punch stretching.
1-421
Figure 1.3.4-2 Initial ABAQUS/Explicit configuration.
Figure 1.3.4-3 Configuration for punch head displacement of 34.5 mm, ABAQUS/Explicit.
Figure 1.3.4-4 Final configuration after springback, ABAQUS/Explicit.
1-422
Figure 1.3.4-5 Strain distribution for punch head displacement of 18.6 mm, ABAQUS/Standard.
Figure 1.3.4-6 Strain distribution for punch head displacement of 18.6 mm, ABAQUS/Explicit.
1-423
1-424
Figure 1.3.4-11 Residual stress on top surface after springback, ABAQUS/Explicit.
1-425
Figure 1.3.4-12 Residual stress on bottom surface after springback, ABAQUS/Explicit.
Sample listings
1-426
Listing 1.3.4-1
*HEADING
WANG AND BUDIANSKY'S SPHERICAL PUNCH WITH SAX1
50 ELEMENTS, 9 LAYERS
*PREPRINT,ECHO=YES
*PART,NAME=BLANK
*NODE,NSET=MID
1,0.0,0.0
*NODE,NSET=REFD
401,50.59,0.
*NFILL,BIAS=1.0,NSET=METND
MID,REFD,40,10
*NODE,NSET=END
501,59.18,0.
*NFILL,BIAS=1.0,NSET=METND
REFD,END,10,10
*NSET,NSET=NODWR,GENERATE
1,501,10
*ELEMENT,TYPE=SAX1
1,1,11
*ELGEN,ELSET=METAL
1,50,10
*ELSET,ELSET=EDIE,GENERATE
42,49,1
1,41,1
*SHELL SECTION,ELSET=METAL,MATERIAL=SAMP
0.85,9
*END PART
*MATERIAL,NAME=SAMP
*ELASTIC
206.8,0.3
*PLASTIC
0.1700000, 0.0000000E+00
0.1800000 , 1.7205942E-03
0.1900000 , 3.8296832E-03
0.2000000 , 6.3897874E-03
0.2100000, 9.4694765E-03
0.2200000, 1.3143660E-02
0.2300000, 1.7493792E-02
0.2400000, 2.2608092E-02
1-427
0.2500000, 2.8581845E-02
0.2600000, 3.5517555E-02
0.2700000, 4.3525275E-02
0.2800000, 5.2722659E-02
0.2900000, 6.3235357E-02
0.3000000, 7.5197279E-02
0.3100000, 8.8750519E-02
0.3200000, 0.1040458
0.3300000, 0.1212430
0.3400000, 0.1405106
0.3500000, 0.1620263
0.3600000, 0.1859779
0.3700000, 0.2125620
0.3800000, 0.2419857
0.3900000, 0.2744660
0.4000000, 0.3102303
0.4100000, 0.3495160
0.4200000, 0.3925720
0.4300000, 0.4396578
0.4400000, 0.4910434
0.4500000, 0.5470111
0.4600000, 0.6078544
0.4700000, 0.6738777
0.4800000, 0.7453985
0.4900000, 0.8227461
0.5000000, 0.9062610
0.5100000 , 0.9962980
*PART,NAME=PUNCH
*NODE,NSET=PUNCH
1000,0.,0.
*END PART
*PART,NAME=DIE
*NODE,NSET=DIE
2000,59.18,0.05
*RIGID BODY,ANALYTICAL SURFACE=DSURF,REF NODE=2000
*END PART
*ASSEMBLY,NAME=FORM
*INSTANCE,NAME=BLANK-1,PART=BLANK
*SURFACE,NAME=ASURF
ECON,SNEG
*SURFACE,NAME=CSURF
EDIE,SPOS
1-428
*END INSTANCE
*INSTANCE,NAME=PUNCH-1,PART=PUNCH
*SURFACE,NAME=BSURF,TYPE=SEGMENTS
START,0.0,0.0
CIRCL,50.8,-50.80,0.0,-50.80
*END INSTANCE
*INSTANCE,NAME=DIE-1,PART=DIE
*SURFACE,NAME=DSURF,TYPE=SEGMENTS
START,61.00,0.05
LINE,59.18,0.05
CIRCL,52.83,6.4,59.18,6.4
LINE,52.83,8.
*END INSTANCE
*NSET,NSET=PUNK
PUNCH-1.PUNCH,DIE-1.DIE
*END ASSEMBLY
FORM.BLANK-1.ASURF,FORM.PUNCH-1.BSURF
FORM.BLANK-1.CSURF,FORM.DIE-1.DSURF
*FRICTION
0.275,
*BOUNDARY
FORM.BLANK-1.1,1,1
FORM.BLANK-1.1,6,6
FORM.BLANK-1.501,1,1
FORM.PUNCH-1.1000,6,6
FORM.DIE-1.2000,1,1
FORM.DIE-1.2000,6,6
*STATIC
1.,1.
*BOUNDARY
FORM.DIE-1.2000,2,2,-0.05
*PRINT,RESIDUAL=NO,FREQUENCY=10
U,RF
COORD,
*CONTACT FILE,SLAVE=FORM.BLANK-1.ASURF,FREQUENCY=1000
1-429
*CONTACT FILE,SLAVE=FORM.BLANK-1.CSURF,FREQUENCY=1000
*END STEP
*STATIC
0.05,100.,1.E-5
*BOUNDARY
FORM.PUNCH-1.1000,2,2,18.6
FORM.DIE-1.2000,2,2,-0.05
*MONITOR,NODE=FORM.PUNCH-1.1000,DOF=2
*NODE PRINT,NSET=FORM.PUNK,FREQUENCY=100
U,RF
COORD,
*END STEP
*STATIC
0.05,100.,1.E-5
*EL FILE,ELSET=FORM.BLANK-1.METAL,FREQUENCY=1000
5,
S,E
*BOUNDARY
FORM.PUNCH-1.1000,2,2,28.5
FORM.DIE-1.2000,2,2,-0.05
*END STEP
*STATIC
0.05,100.,1.E-5
*BOUNDARY
FORM.PUNCH-1.1000,2,2,34.5
FORM.DIE-1.2000,2,2,-0.05
*END STEP
*STATIC
100.,100.
*BOUNDARY,FIXED
FORM.BLANK-1.METND,1,2
FORM.BLANK-1.METND,6
*MODEL CHANGE,TYPE=CONTACT PAIR,REMOVE
FORM.BLANK-1.ASURF,FORM.PUNCH-1.BSURF
FORM.BLANK-1.CSURF,FORM.DIE-1.DSURF
5,
S,
*END STEP
1-430
*STATIC
1.,100.
*BOUNDARY,OP=NEW
FORM.BLANK-1.1,1,1
FORM.BLANK-1.1,6,6
FORM.DIE-1.2000,1,2
FORM.DIE-1.2000,6,6
*MONITOR,NODE=FORM.BLANK-1.1,DOF=2
5,
S,
*END STEP
1-431
Listing 1.3.4-2
*HEADING
WANG AND BUDIANSKY'S SPHERICAL PUNCH WITH SAX1
ELEMENTS
PUNCH AND DIE ARE ANALYTICAL RIGID SEGMENTS
*PREPRINT,ECHO=YES
*PART,NAME=BLANK
*NODE
1,0.0,0.0
401,.05059,0.
501,.05918,0.
*NGEN
1,401,10
401,501,10
*ELEMENT,TYPE=SAX1,ELSET=BLANK
1,1,11
*ELGEN,ELSET=BLANK
1,50,10
*ELSET,ELSET=CENTER,GEN
1,10,1
*SHELL SECTION,ELSET=BLANK,MATERIAL=SAMP
0.00085
*SURFACE, NAME=TOP
BLANK,SPOS
*SURFACE, NAME=BOTTOM
BLANK,SNEG
*END PART
*MATERIAL,NAME=SAMP
*DENSITY
7850.
*ELASTIC
206.8E9,0.3
*PLASTIC
170.0E6, 0.0000000E+00
180.0E6, 1.7205942E-03
190.0E6, 3.8296832E-03
200.0E6, 6.3897874E-03
210.0E6, 9.4694765E-03
220.0E6, 1.3143660E-02
230.0E6, 1.7493792E-02
240.0E6, 2.2608092E-02
250.0E6, 2.8581845E-02
1-432
260.0E6, 3.5517555E-02
270.0E6, 4.3525275E-02
280.0E6, 5.2722659E-02
290.0E6, 6.3235357E-02
300.0E6, 7.5197279E-02
310.0E6, 8.8750519E-02
320.0E6, 0.1040458
330.0E6, 0.1212430
340.0E6, 0.1405106
350.0E6, 0.1620263
360.0E6, 0.1859779
370.0E6, 0.2125620
380.0E6, 0.2419857
390.0E6, 0.2744660
400.0E6, 0.3102303
410.0E6, 0.3495160
420.0E6, 0.3925720
430.0E6, 0.4396578
440.0E6, 0.4910434
450.0E6, 0.5470111
460.0E6, 0.6078544
470.0E6, 0.6738777
480.0E6, 0.7453985
490.0E6, 0.8227461
500.0E6, 0.9062610
510.0E6, 0.9962980
**
*PART,NAME=PUNCH
*NODE,NSET=REF_NODE
1000,0.,.051225
*RIGID BODY, REF NODE=REF_NODE,
ANALYTICAL SURFACE=PUNCH_BOT
*SURFACE, NAME=PUNCH_BOT, TYPE=SEGMENTS
START, .0508,.051225
CIRCL, 0.,0.000425, 0.,.051225
*END PART
*PART,NAME=DIE
*NODE,NSET=REF_NODE
2000, .05918,-.006775
*RIGID BODY, REF NODE=REF_NODE,
ANALYTICAL SURFACE=DIE_TOP
*SURFACE, NAME=DIE_TOP, TYPE=SEGMENTS
START, .05283,-.030425
1-433
LINE, .05283,-.006775
CIRCL, .05918,-0.000425, .05918,-.006775
LINE, .05930,-0.000425
*END PART
**
*ASSEMBLY,NAME=ASSEMBLY-1
*INSTANCE,NAME=BLANK-1,PART=BLANK
*NSET,NSET=NOUT
1,
*ELSET,ELSET=EOUT
22,23,24,25
*END INSTANCE
*INSTANCE,NAME=PUNCH-1,PART=PUNCH
*NSET,NSET=PUNCH
1000,
*END INSTANCE
*INSTANCE,NAME=DIE-1,PART=DIE
*END INSTANCE
*END ASSEMBLY
*BOUNDARY
ASSEMBLY-1.BLANK-1.1,1,1
ASSEMBLY-1.PUNCH-1.1000,1,1
ASSEMBLY-1.PUNCH-1.1000,1,6
ASSEMBLY-1.DIE-1.2000,1,2
ASSEMBLY-1.DIE-1.2000,6,6
*AMPLITUDE,NAME=LOAD,TIME=TOTAL
0.,0.,1.24E-3,1.,1.57E-3,1.,1.97E-3,0.,
3.97E-3,-.25
*RESTART,WRITE,NUM=2,TIMEMARKS=NO
*STEP
*DYNAMIC,EXPLICIT
,1.24E-3
*BOUNDARY,TYPE=VELOCITY,AMP=LOAD
ASSEMBLY-1.PUNCH-1.1000,2,2,-30.
*SURFACE INTERACTION,NAME=PUNCH_TOP
*FRICTION
0.265,
*CONTACT PAIR, INTERACTION=PUNCH_TOP,
CPSET=PUNCH_DIE
ASSEMBLY-1.PUNCH-1.PUNCH_BOT,ASSEMBLY-1.BLANK-1.TOP
1-434
ASSEMBLY-1.DIE-1.DIE_TOP,ASSEMBLY-1.BLANK-1.BOTTOM
*FILE OUTPUT,NUMBER INTERVAL = 1
*EL FILE
S,LE
*ENERGY FILE
*HISTORY OUTPUT,TIME=0.
*NODE HISTORY,NSET=ASSEMBLY-1.PUNCH-1.PUNCH
U2,RF2
*NODE HISTORY,NSET=ASSEMBLY-1.BLANK-1.NOUT
U,V
*EL HISTORY,ELSET=ASSEMBLY-1.BLANK-1.EOUT
STH
MISES,S,LE
*ENERGY HISTORY
ALLKE,ALLSE,ALLWK,ALLPD,ALLIE,ALLVD,ETOTAL,ALLAE,
ALLCD,ALLFD,DT
*OUTPUT, FIELD, NUMBER INTERVAL=4, TIMEMARKS=NO
*CONTACT OUTPUT, CPSET=PUNCH_DIE,
VARIABLE=PRESELECT
*END STEP
*STEP
*DYNAMIC,EXPLICIT
,.33E-3
*END STEP
*STEP
*DYNAMIC,EXPLICIT
,.40E-3
*END STEP
*STEP
**
** Unloading step
**
*DYNAMIC,EXPLICIT
,2.E-3
*DLOAD,OP=NEW
ASSEMBLY-1.BLANK-1.BLANK,VP,0.35E6
*END STEP
1.3.5 Deep drawing of a cylindrical cup

Deep drawing of sheet metal is an important manufacturing technique. In the deep drawing process a
"blank" of sheet metal is clamped by a blank holder against a die. A punch is then moved against the
1-435
blank, which is drawn into the die. Unlike the operation described in the hemispherical punch
stretching example (``Stretching of a thin sheet with a hemispherical punch, '' Section 1.3.4), the blank
is not assumed to be fixed between the die and the blank holder; rather, the blank is drawn from
between these two tools. The ratio of drawing versus stretching is controlled by the force on the blank
holder and the friction conditions at the interface between the blank and the blank holder and the die.
Higher force or friction at the blank/die/blank holder interface limits the slip at the interface and
increases the radial stretching of the blank. In certain cases drawbeads, shown in Figure 1.3.5-1, are
used to restrain the slip at this interface even further.
To obtain a successful deep drawing process, it is essential to control the slip between the blank and its
holder and die. If the slip is restrained too much, the material will undergo severe stretching, thus
potentially causing necking and rupture. If the blank can slide too easily, the material will be drawn in
completely and high compressive circumferential stresses will develop, causing wrinkling in the
product. For simple shapes like the cylindrical cup here, a wide range of interface conditions will give
satisfactory results. But for more complex, three-dimensional shapes, the interface conditions need to
be controlled within a narrow range to obtain a good product.
During the drawing process the response is determined primarily by the membrane behavior of the
sheet. For axisymmetric problems in particular, the bending stiffness of the metal yields only a small
correction to the pure membrane solution, as discussed by Wang and Tang (1988). In contrast, the
interaction between the die, the blank, and the blank holder is critical. Thus, thickness changes in the
sheet material must be modeled accurately in a finite element simulation, since they will have a
significant influence on the contact and friction stresses at the interface. In these circumstances the
most suitable elements in ABAQUS are the 4-node reduced-integration axisymmetric quadrilateral,
CAX4R; the first-order axisymmetric shell element, SAX1; the first-order axisymmetric membrane
element, MAX1; the first-order finite-strain quadrilateral shell element, S4R; and the fully integrated
general-purpose finite-membrane-strain shell element, S4.
Membrane effects and thickness changes are modeled properly with CAX4R. However, the bending
stiffness of the element is low. The element does not exhibit "locking" due to incompressibility or
parasitic shear. It is also very cost-effective. In the shells and membranes the thickness change is
calculated from the assumption of incompressible deformation of the material. This simplifying
assumption does not allow for the development of stress in the thickness direction of the shell, thus
making it difficult to model the contact pressure between the blank and the die and the blank holder.
This situation is resolved in the shell and membrane models by using the *SURFACE BEHAVIOR,
PRESSURE-OVERCLOSURE=EXPONENTIAL option (`Ìnteraction normal to the surface,'' Section
21.3.3 of the ABAQUS/Standard User's Manual) to impose the proper clamping pressure in the
thickness direction of the shell or membrane between the blank and the die and the blank holder.
Geometry and model

The geometry of the problem is shown in Figure 1.3.5-2. The circular blank being drawn has an initial
radius of 100 mm and an initial thickness of 0.82 mm. The punch has a radius of 50 mm and is
rounded off at the corner with a radius of 13 mm. The die has an internal radius of 51.25 mm and is
rounded off at the corner with a radius of 5 mm. The blank holder has an internal radius of 56.25 mm.
1-436
The blank is modeled using 40 elements of type CAX4R or 31 elements of either type SAX1, MAX1,
S4R, or S4. An 11.25° wedge of the circular blank is used in the three-dimensional S4R and S4
models. These meshes are rather coarse for this analysis. However, since the primary interest in this
problem is to study the membrane effects, the analysis will still give a fair indication of the stresses
and strains occurring in the process.
The contact between the blank and the rigid punch, the rigid die, and the rigid blank holder is modeled
with the *CONTACT PAIR option. The top and bottom surfaces of the blank are defined by means of
the *SURFACE option. The rigid punch, the die, and the blank holder are modeled as analytical rigid
surfaces with the *RIGID BODY option in conjunction with the *SURFACE option. The mechanical
interaction between the contact surfaces is assumed to be frictional contact. Therefore, the *FRICTION
option is used in conjunction with the various *SURFACE INTERACTION property options to
specify coefficients of friction. For the shell models the interaction between the blank and the blank
holder is also assumed to be "softened" contact, as discussed previously.
At the start of the analysis for the CAX4R model, the blank is positioned precisely on top of the die
and the blank holder is precisely in touch with the top surface of the blank. The punch is positioned
0.18 mm above the top surface of the blank.
The shell and membrane models begin with the same state except that the blank holder is positioned a
fixed distance above the blank. This fixed distance is the distance at which the contact pressure is set
to zero by means of the *SURFACE BEHAVIOR, PRESSURE-OVERCLOSURE=EXPONENTIAL
option.
Material properties
The material (aluminum-killed steel) is assumed to satisfy the Ramberg-Osgood relation between true
stress and logarithmic strain:
² = (¾=K )1=n :
The reference stress value, K, is 513 MPa; and the work-hardening exponent, n, is 0.223. The Young's
modulus is 211 GPa, and the Poisson's ratio is 0.3. An initial yield stress of 91.3 MPa is obtained with
these data. The stress-strain curve is defined in piecewise linear segments in the *PLASTIC option, up
to a total (logarithmic) strain level of 107%.
The coefficient of friction between the interface and the punch is taken to be 0.25; and that between
the die and the blank holder is taken as 0.1, the latter value simulating a certain degree of lubrication
between the surfaces. The stiffness method of sticking friction is used in these analyses. The numerics
of this method make it necessary to choose an acceptable measure of relative elastic slip between
mating surfaces when sticking should actually be occurring. The basis for the choice is as follows.
Small values of elastic slip best simulate the actual behavior but also result in a slower convergence of
the solution. Permission of large relative elastic displacements between the contacting surfaces can
cause higher strains at the center of the blank. In these runs we let ABAQUS choose the allowable
elastic slip, which is done by determining a characteristic interface element length over the entire mesh
and multiplying by a small fraction to get an allowable elastic slip measure. This method typically
1-437
gives a fairly small amount of elastic slip.

Although the material in this process is fully isotropic, the *ORIENTATION option is used with the
CAX4R elements to define a local orientation that is coincident initially with the global directions. The
reason for using this option is to obtain the stress and strain output in more natural coordinates: if the
*ORIENTATION option is used in a geometrically nonlinear analysis, stress and strain components
are given in a corotational framework. Hence, in our case throughout the motion, S11 will be the stress
in the r-z plane in the direction of the middle surface of the cup. S22 will be the stress in the thickness
direction, S33 will be the hoop stress, and S12 will be the transverse shear stress, which makes
interpreting the results considerably easier. This orientation definition is not necessary with the SAX1
or MAX1 elements since the output for shell and membrane elements is already given in the local shell
system. For the SAX1 and MAX1 model, S11 is the stress in the meridional direction and S22 is the
circumferential (hoop) stress. An orientation definition would normally be needed for the S4R and S4
models but can be avoided by defining the wedge in such a manner that the single integration point of
each element lies along the global x-axis. Such a model definition, along with appropriate kinematic
boundary conditions, keeps the local stress output definitions for the shells as S11 being the stress in
the meridional plane and S22 the hoop stress. There should be no in-plane shear, S12, in this problem.
A transformation is used in the S4R and S4 models to impose boundary constraints in a cylindrical
system.
Loading
The entire analysis is carried out in five steps. In the first step the blank holder is pushed onto the
blank with a prescribed displacement of -17.5 ´ 10-6 mm. This value is chosen to obtain a reaction
force that is approximately equal to the applied force. In the shell models this displacement
corresponds to zero clearance across the interface, thus resulting in the application of a predetermined
clamping pressure across the shell thickness via the *SURFACE BEHAVIOR,
PRESSURE-OVERCLOSURE=EXPONENTIAL option.
In the second step the boundary condition is removed and replaced by the applied force of 100 kN on
the blank holder. This force is kept constant during Steps 2 and 3. This technique of simulating the
clamping process is used to avoid potential problems with rigid body modes of the blank holder, since
there is no firm contact between the blank holder, the blank, and the die at the start of the process. The
two-step procedure creates contact before the blank holder is allowed to move freely.
In the third step the punch is moved toward the blank through a total distance of 60 mm. This step
models the actual drawing process. During this step the option *CONTROLS,
ANALYSIS=DISCONTINUOUS is included since contact with friction tends to create a severely
discontinuous nonlinearity and we wish to avoid premature cutbacks of the automatic time
incrementation scheme.
The last two steps are used to simulate springback. In the fourth step all the nodes in the model are
fixed in their current positions and the contact pairs are removed from the model with the *MODEL
CHANGE, TYPE=CONTACT PAIR, REMOVE option. This is the most reliable method for releasing
contact conditions. In the fifth, and final, step the regular set of boundary conditions is reinstated and
the springback is allowed to take place. This part of the analysis with the CAX4R elements is included
1-438
to demonstrate the feasibility of the unloading procedure only and is not expected to produce realistic
results, since the reduced-integration elements have a purely elastic bending behavior. The springback
is modeled with more accuracy in the shell element models.

Figure 1.3.5-3 shows deformed shapes that are predicted at various stages of the drawing process for
the CAX4R model. The profiles show that the metal initially bends and stretches and is then drawn in
over the surface of the die. The distributions of radial and circumferential strain for all three models
and thickness strain for the CAX4R model are shown in Figure 1.3.5-4. The thickness for the shell or
membrane models can be monitored with output variable STH (current shell or membrane thickness).
The thickness does not change very much: the change ranges from approximately -12% in the
cylindrical part to approximately +16% at the edge of the formed cup. Relatively small thickness
changes are usually desired in deep drawing processes and are achieved because the radial tensile
strain and the circumferential compressive strain balance each other. The drawing force as a function
of punch displacement is shown in Figure 1.3.5-5. The curves for the three models compare closely.
The oscillations in the force are a result of the rather coarse mesh--each oscillation represents an
element being drawn over the corner of the die. Compared to the shell models, the membrane model
predicts a smaller punch force for a given punch displacement. Thus, toward the end of the analysis the
results for punch force versus displacement for the MAX1 model are closer to those for the CAX4R
model.
The deformed shape after complete unloading is shown in Figure 1.3.5-6, superimposed on the
deformed shape under complete loading. The analysis shows the lip of the cup springing back strongly
after the blank holder is removed for the CAX4R model. No springback is evident in the shell models.
As was noted before, this springback in the CAX4R model is not physically realistic: in the first-order
reduced-integration elements an elastic "hourglass control" stiffness is associated with the "bending"
mode, since this mode is identical to the "hourglass" mode exhibited by this element in continuum
situations. In reality the bending of the element is an elastic-plastic process, so that the springback is
likely to be much less. A better simulation of this aspect would be achieved by using several elements
through the thickness of the blank, which would also increase the cost of the analysis. The springback
results for the shell models do not exhibit this problem and are clearly more representative of the
actual elastic-plastic process.
Input files
deepdrawcup_cax4r.inp
CAX4R model.
deepdrawcup_cax4i.inp
Model using the incompatible mode element, CAX4I, as an alternative to the CAX4R element. In
contrast to the reduced-integration, linear isoparametric elements such as the CAX4R element, the
incompatible mode elements have excellent bending properties even with one layer of elements
through the thickness (see ``Geometrically nonlinear analysis of a cantilever beam,'' Section 2.1.2
of the ABAQUS Benchmarks Manual) and have no hourglassing problems. However, they are
1-439
computationally more expensive.

deepdrawcup_s4.inp
S4 model.
deepdrawcup_s4r.inp
S4R model.
deepdrawcup_sax1.inp
SAX1 model.
deepdrawcup_postoutput.inp
*POST OUTPUT analysis of deepdrawcup_sax1.inp.
deepdrawcup_max1.inp
MAX1 model.
deepdrawcup_mgax1.inp
MGAX1 model.
Reference
· Wang, N. M., and S. C. Tang, "Analysis of Bending Effects in Sheet Forming Operations,"
International Journal for Numerical Methods in Engineering, vol. 25, pp. 253-267, January 1988.
Figures
Figure 1.3.5-1 A typical drawbead used to limit slip between the blank and die.
Figure 1.3.5-2 Geometry and mesh for the deep drawing problem.
1-440
Figure 1.3.5-3 Deformed shapes at various stages of the analysis.
1-441
Figure 1.3.5-4 Strain distribution at the end of the deep drawing step.
1-442
Figure 1.3.5-5 Punch force versus punch displacement.
1-443
Figure 1.3.5-6 Deformed shape after unloading.
1-444
Sample listings
1-445
Listing 1.3.5-1
*HEADING
DEEP DRAWING OF CYLINDRICAL CUP WITH CAX4R
*NODE
101,
181,0.1
301,0.0,0.00082
381,0.1,0.00082
*NGEN,NSET=BOT
101,181,2
*NGEN,NSET=TOP
301,381,2
*NSET,NSET=WRKPC
BOT,TOP
*NODE,NSET=DIE
100,0.1,-0.05
*NODE,NSET=PUNCH
200,0.,.05
*NODE,NSET=HOLDER
300,0.1,0.05
*NSET,NSET=TOOLS
PUNCH,DIE,HOLDER
*NSET,NSET=CENTER
101,301
*ELEMENT,TYPE=CAX4R,ELSET=BLANK
201,101,103,303,301
*ELGEN,ELSET=BLANK
201,40,2,2
*ELSET,ELSET=ALL
BLANK,
*SOLID SECTION,MATERIAL=STEEL,ORIENTATION=LOCAL,
ELSET=BLANK
*ORIENTATION,NAME=LOCAL
1.,0.,0.,0.,1.,0.
0,0.,
*ELASTIC
2.1E11,0.3
*PLASTIC,HARDENING=ISOTROPIC
0.91294E+08, 0.00000E+00
0.10129E+09, 0.21052E-03
1-446
0.11129E+09, 0.52686E-03
0.12129E+09, 0.97685E-03
0.13129E+09, 0.15923E-02
0.14129E+09, 0.24090E-02
0.15129E+09, 0.34674E-02
0.16129E+09, 0.48120E-02
0.17129E+09, 0.64921E-02
0.18129E+09, 0.85618E-02
0.19129E+09, 0.11080E-01
0.20129E+09, 0.14110E-01
0.21129E+09, 0.17723E-01
0.22129E+09, 0.21991E-01
0.23129E+09, 0.26994E-01
0.24129E+09, 0.32819E-01
0.25129E+09, 0.39556E-01
0.26129E+09, 0.47301E-01
0.27129E+09, 0.56159E-01
0.28129E+09, 0.66236E-01
0.29129E+09, 0.77648E-01
0.30129E+09, 0.90516E-01
0.31129E+09, 0.10497E+00
0.32129E+09, 0.12114E+00
0.33129E+09, 0.13916E+00
0.34129E+09, 0.15919E+00
0.35129E+09, 0.18138E+00
0.36129E+09, 0.20588E+00
0.37129E+09, 0.23287E+00
0.38129E+09, 0.26252E+00
0.39129E+09, 0.29502E+00
0.40129E+09, 0.33054E+00
0.41129E+09, 0.36929E+00
0.42129E+09, 0.41147E+00
0.43129E+09, 0.45729E+00
0.44129E+09, 0.50696E+00
0.45129E+09, 0.56073E+00
0.46129E+09, 0.61881E+00
0.47129E+09, 0.68145E+00
0.48129E+09, 0.74890E+00
0.49129E+09, 0.82142E+00
0.50129E+09, 0.89928E+00
0.51129E+09, 0.98274E+00
0.52129E+09, 0.10721E+01
*ELSET,ELSET=EDIE,GENERATE
1-447
231,279,2
*ELSET,ELSET=EHOLDER,GENERATE
241,279,2
*ELSET,ELSET=EPUNCH,GENERATE
201,249,2
START,0.05125,-0.060
LINE,0.05125,-0.005
CIRCL,0.05625,0.0,0.05625,-0.005
LINE,0.1,0.0
*RIGID BODY,ANALYTICAL SURFACE=DSURF,REF NODE=300
*SURFACE,TYPE=SEGMENTS,NAME=DSURF
START,0.1,0.00082
LINE,0.05630,0.00082
CIRCL,0.05625,.00087,.05630,.00087
LINE,0.05625,.06
*RIGID BODY,ANALYTICAL SURFACE=FSURF,REF NODE=200
*SURFACE,TYPE=SEGMENTS,FILLET RADIUS=.013,
NAME=FSURF
START,0.05,0.060
LINE,0.05,2.250782E-3
CIRCL,0.0,0.001,0.0,1.001
LINE,-0.001,0.001
*SURFACE,NAME=ASURF
EDIE,S1
*SURFACE,NAME=CSURF
EHOLDER,S3
*SURFACE,NAME=ESURF
EPUNCH,S3
*CONTACT PAIR,INTERACTION=ROUGH1
ASURF,BSURF
CSURF,DSURF
ESURF,FSURF
*SURFACE INTERACTION,NAME=ROUGH1
*FRICTION
0.1,
*FRICTION
0.1,
1-448
*FRICTION
0.25,
PUSH THE BLANKHOLDER DOWN BY A PRESCRIBED
DISPLACEMENT
*STATIC
1.,1.
*BOUNDARY
CENTER,1,1
DIE,1,1
DIE,2,2
DIE,6,6
PUNCH,1,1
PUNCH,2,2
PUNCH,6,6
HOLDER,1,1
HOLDER,2,2,-1.75E-8
HOLDER,6,6
*PRINT,CONTACT=YES
*NODE PRINT,NSET=TOOLS,FREQUENCY=100
COORD,U,RF
*EL PRINT,ELSET=ALL,FREQUENCY=500
S,E
*NODE FILE,NSET=TOOLS,FREQUENCY=10
U,RF
*CONTACT FILE,SLAVE=CSURF,FREQUENCY=10
*CONTACT FILE,SLAVE=ESURF,FREQUENCY=10
*END STEP
*STEP,INC=10,NLGEOM
APPLY PRESCRIBED FORCE ON BLANKHOLDER AND
RELEASE DISPLACEMENT
*STATIC
1.,1.
*BOUNDARY,OP=NEW
CENTER,1,1
DIE,1,1
DIE,2,2
DIE,6,6
PUNCH,1,1
PUNCH,2,2
1-449
PUNCH,6,6
HOLDER,1,1
HOLDER,6,6
*CLOAD
HOLDER,2,-100000.
*END STEP
MOVE THE PUNCH DOWN
*STATIC
.01,1.,1.E-6
*BOUNDARY,OP=NEW
CENTER,1,1
DIE,1,1
DIE,2,2
DIE,6,6
PUNCH,1,1
PUNCH,2,2,-.06
PUNCH,6,6
HOLDER,1,1
HOLDER,6,6
*CLOAD
HOLDER,2,-100000.
*END STEP
FIX ALL NODES AND REMOVE THE CONTACT SURFACES
*STATIC
1.,1.,1.,1.
*BOUNDARY,FIXED
WRKPC,1,2
ASURF,BSURF
CSURF,DSURF
ESURF,FSURF
*CLOAD,OP=NEW
HOLDER,2,0.
*END STEP
*STEP,INC=50,NLGEOM, UNSYMM=NO
REPLACE THE BOUNDARY CONDITIONS BY THE REGULAR SET
*STATIC
.1,1.,1.E-6
*BOUNDARY,OP=NEW
181,2
1-450
CENTER,1,1
*END STEP
1-451
Listing 1.3.5-2
*HEADING
DEEP DRAWING OF CYLINDRICAL CUP WITH SAX1
*NODE
101,
181,0.1
*NGEN,NSET=BOT
101,181,2
*NSET,NSET=WRKPC,GENERATE
121,181,2
*NODE,NSET=DIE
100,0.,-0.05
*NODE,NSET=PUNCH
200,0.,.05
*NODE,NSET=HOLDER
300,0.,0.05
*NSET,NSET=TOOLS
PUNCH,DIE,HOLDER
*NSET,NSET=CENTER
101,
201,101,121
202,121,123
*ELGEN,ELSET=BLANK
202,30,2,2
*SHELL SECTION,MATERIAL=STEEL,ELSET=BLANK
.00082,5
*ELASTIC
2.1E11,0.3
0.91294E+08, 0.00000E+00
0.10129E+09, 0.21052E-03
0.11129E+09, 0.52686E-03
0.12129E+09, 0.97685E-03
0.13129E+09, 0.15923E-02
0.14129E+09, 0.24090E-02
0.15129E+09, 0.34674E-02
0.16129E+09, 0.48120E-02
0.17129E+09, 0.64921E-02
0.18129E+09, 0.85618E-02
1-452
0.19129E+09, 0.11080E-01
0.20129E+09, 0.14110E-01
0.21129E+09, 0.17723E-01
0.22129E+09, 0.21991E-01
0.23129E+09, 0.26994E-01
0.24129E+09, 0.32819E-01
0.25129E+09, 0.39556E-01
0.26129E+09, 0.47301E-01
0.27129E+09, 0.56159E-01
0.28129E+09, 0.66236E-01
0.29129E+09, 0.77648E-01
0.30129E+09, 0.90516E-01
0.31129E+09, 0.10497E+00
0.32129E+09, 0.12114E+00
0.33129E+09, 0.13916E+00
0.34129E+09, 0.15919E+00
0.35129E+09, 0.18138E+00
0.36129E+09, 0.20588E+00
0.37129E+09, 0.23287E+00
0.38129E+09, 0.26252E+00
0.39129E+09, 0.29502E+00
0.40129E+09, 0.33054E+00
0.41129E+09, 0.36929E+00
0.42129E+09, 0.41147E+00
0.43129E+09, 0.45729E+00
0.44129E+09, 0.50696E+00
0.45129E+09, 0.56073E+00
0.46129E+09, 0.61881E+00
0.47129E+09, 0.68145E+00
0.48129E+09, 0.74890E+00
0.49129E+09, 0.82142E+00
0.50129E+09, 0.89928E+00
0.51129E+09, 0.98274E+00
0.52129E+09, 0.10721E+01
*RIGID BODY,ANALYTICAL SURFACE=HOLDER,REF NODE=300
*SURFACE,TYPE=SEGMENTS,NAME=HOLDER,
FILLET RADIUS=0.001
START,0.12,1.75E-8
LINE,0.05625,1.75E-8
LINE,0.05625,.06
*RIGID BODY,ANALYTICAL SURFACE=DIE,REF NODE=100
*SURFACE,TYPE=SEGMENTS,NAME=DIE,
FILLET RADIUS=0.00541
1-453
START,0.05125,-0.060
LINE,0.05125,0.0
LINE,0.12,0.0
*RIGID BODY,ANALYTICAL SURFACE=PUNCH,REF NODE=200
*SURFACE,TYPE=SEGMENTS,NAME=PUNCH,
FILLET RADIUS=.01341
START,0.05,0.060
LINE,0.05,2.250782E-3
CIRCL,0.0,.0001,0.0,1.0001
LINE,-0.001,0.0001
**
** Contact with holder
**
*ELSET,ELSET=HOLD_CON,GENERATE
222,260,2
*SURFACE,NAME=HLD_SURF
HOLD_CON,SPOS
*CONTACT PAIR,INTERACTION=FRIC1
HLD_SURF,HOLDER
*SURFACE INTERACTION,NAME=FRIC1
*SURFACE BEHAVIOR,
1.75E-8,4.5E6
*FRICTION
0.0,
**
** Contact with die
**
*ELSET,ELSET=DIE_CON,GENERATE
212,260,2
*SURFACE,NAME=DIE_SURF
DIE_CON,SNEG
DIE_SURF,DIE
*FRICTION
0.0,
**
** Contact with punch
**
*ELSET,ELSET=PUN_CON2,GENERATE
202,230,2
*SURFACE,NAME=PCH_SURF
1-454
PUN_CON2,SPOS
PCH_SURF,PUNCH
*FRICTION
0.0,
*SURFACE,TYPE=NODE,NAME=PCH_PNT
101,
PCH_PNT,PUNCH
*FRICTION
0.0,
**
**
*STEP,INC=10, UNSYMM=YES
PUSH THE BLANKHOLDER DOWN BY A PRESCRIBED
DISPLACEMENT
*STATIC
1.,1.
*BOUNDARY
CENTER,1,1
CENTER,6,6
DIE,1,1
DIE,2,2
DIE,6,6
PUNCH,1,1
PUNCH,2,2
PUNCH,6,6
HOLDER,1,1
HOLDER,2,2,-1.75E-8
HOLDER,6,6
*PRINT,CONTACT=YES
*NODE PRINT,NSET=TOOLS,FREQUENCY=100
COORD,U,RF
*EL PRINT,ELSET=BLANK,FREQUENCY=500
S,E
STH,
*CONTACT PRINT,FREQUENCY=500
*CONTACT FILE,FREQUENCY=500
*NODE FILE,NSET=TOOLS,FREQUENCY=10
U,RF
1-455
*END STEP
APPLY PRESCRIBED FORCE ON BLANKHOLDER AND
RELEASE DISPLACEMENT
*STATIC
1.,1.
*BOUNDARY,OP=NEW
CENTER,1,1
CENTER,6,6
DIE,1,1
DIE,2,2
DIE,6,6
PUNCH,1,1
PUNCH,2,2
PUNCH,6,6
HOLDER,1,1
HOLDER,6,6
*CLOAD
HOLDER,2,-100000.
*CHANGE FRICTION,INTERACTION=FRIC2
*FRICTION
0.1,
*FRICTION
0.1,
*FRICTION
0.25,
*END STEP
*STEP,INC=500,NLGEOM,UNSYMM=YES
MOVE THE PUNCH DOWN
*STATIC
.01,1.,1.E-6
*BOUNDARY,OP=NEW
CENTER,1,1
CENTER,6,6
DIE,1,1
DIE,2,2
DIE,6,6
PUNCH,1,1
PUNCH,2,2,-.06
PUNCH,6,6
1-456
HOLDER,1,1
HOLDER,6,6
*CLOAD
HOLDER,2,-100000.
*END STEP
FIX ALL NODES AND REMOVE THE CONTACT SURFACES
*STATIC
1.,1.,1.,1.
*BOUNDARY,FIXED
WRKPC,1,6
*BOUNDARY,FIXED
CENTER,1,6
HLD_SURF,HOLDER
PCH_PNT,PUNCH
PCH_SURF,PUNCH
DIE_SURF,DIE
*CLOAD,OP=NEW
HOLDER,2,0.
*END STEP
REPLACE THE BOUNDARY CONDITIONS BY THE REGULAR SET
*STATIC
.1,1.,1.E-6
*BOUNDARY,OP=NEW
CENTER,1,1
CENTER,6,6
181,2,2
*END STEP
1.3.6 Extrusion of a cylindrical metal bar with frictional heat

generation
This analysis illustrates how extrusion problems can be simulated with ABAQUS. In this particular
problem the radius of an aluminum cylindrical bar is reduced 33% by an extrusion process. The
generation of heat due to plastic dissipation inside the bar and the frictional heat generation at the
workpiece/die interface are considered.
Geometry and model

The bar has an initial radius of 100 mm and is 300 mm long. Figure 1.3.6-1 shows half of the
cross-section of the bar, modeled with first-order axisymmetric elements ( CAX4T elements in
1-457
ABAQUS/Standard and CAX4RT elements in ABAQUS/Explicit).

The die is assumed to be rigid. In ABAQUS/Standard the die is modeled with CAX4T elements, which
are made into an isothermal rigid body with the *RIGID BODY, ISOTHERMAL option. The
*SURFACE option is used to define the slave surface on the outside of the bar and the master surface
on the inside of the die. To model a die that has no sharp corners and is smooth in the transition region,
the SMOOTH parameter on the *CONTACT PAIR option is set to 0.48.
In ABAQUS/Explicit the die is modeled with either an analytical rigid surface or discrete rigid
elements (RAX2). The analytical rigid surface is defined using the *RIGID BODY option in
conjunction with the *SURFACE option. The FILLET RADIUS parameter on the *SURFACE option
is set to 0.075 to remove sharp corners in the transition region of the die.
For simplicity we do not model any heat transfer in the die--we simply fix the temperature of the rigid
body reference node and assume that no heat is transmitted between the bar and the die. Half the heat
dissipated as a result of friction is assumed to be conducted into the workpiece; the other half is
conducted into the die. 90% of the nonrecoverable work because of plasticity is assumed to heat the
work material. More realistic analysis would include thermal modeling of the die.
The ABAQUS/Explicit simulations are performed both with and without adaptive meshing.
Material model and interface behavior

The material model is chosen to reflect the response of a typical commercial purity aluminum alloy.
The material is assumed to harden isotropically. The dependence of the flow stress on the temperature
is included, but strain rate dependence is ignored. Instead, representative material data at a strain rate
of 0.1 sec-1 are selected to characterize the flow strength.
The interface is assumed to have no conductive properties. Coulomb friction is assumed for the
mechanical behavior, with a friction coefficient of 0.1. The *GAP HEAT GENERATION option is
used to specify the fraction, fg , of total heat generated by frictional dissipation that is transferred to the
two bodies in contact. Half of this heat is conducted into the workpiece, and the other half is conducted
into the die.
Boundary conditions, loading, and solution control

In the first step the bar is moved to a position where contact is established and slipping of the
workpiece against the die begins. In the second step the bar is extruded through the die to realize the
extrusion process. This is accomplished by prescribing displacements to the nodes at the top of the bar.
In the third step the contact elements are removed in preparation for the cool down portion of the
simulation. In ABAQUS/Standard this is performed in a single step: the bar is allowed to cool down
using film conditions, and deformation is driven by thermal contraction during the fourth step. In
ABAQUS/Explicit the cool down simulation is broken into two steps: the first introduces viscous
pressure to damp out dynamic effects and, thus, allow the bar to reach static equilibrium quickly; the
balance of the cool down simulation is performed in a fifth step. The relief of residual stresses through
creep is not analyzed in this example.
In ABAQUS/Explicit mass scaling is used to reduce the computational cost of the analysis; nondefault
1-458
hourglass control is used to control the hourglassing in the model. The default integral viscoelastic
approach to hourglass control generally works best for problems where sudden dynamic loading
occurs; a stiffness-based hourglass control is recommended for problems where the response is
quasi-static. A combination of stiffness and viscous hourglass control is used in this problem.
For purposes of comparison a second problem is also analyzed, in which the first two steps of the
previous analysis are repeated in a static analysis with the adiabatic heat generation capability. The
adiabatic analysis neglects heat conduction in the bar. Frictional heat generation must also be ignored
in this case. This problem is analyzed only in ABAQUS/Standard.

The following discussion centers around the results obtained with ABAQUS/Standard. The results of
the ABAQUS/Explicit simulation are in close agreement with those obtained with ABAQUS/Standard.
Figure 1.3.6-2 shows the deformed configuration after Step 2 of the analysis. Figure 1.3.6-3 and Figure
1.3.6-4 show contour plots of plastic strain and temperature at the end of Step 2 for the fully coupled
analysis. The plastic deformation is most severe near the surface of the workpiece, where plastic
strains exceed 100%. The peak temperature also occurs at the surface of the workpiece because of
plastic deformation and frictional heating. The peak temperature occurs immediately after the radial
reduction zone of the die. This is expected for two reasons. First, the material that is heated by
dissipative processes in the reduction zone will cool by conduction as the material progresses through
the postreduction zone. Second, frictional heating is largest in the reduction zone because of the larger
values of shear stress in that zone. The peak surface temperature is approximately 106°C (i.e., ¢T ¼
86°C).
If we ignore the zone of extreme distortion at the end of the bar, the temperature increase on the
surface is not as large for the adiabatic analysis (Figure 1.3.6-5) because of the absence of frictional
heating. The surface temperatures in this analysis are approximately 80°C. As expected, the
temperature field contours for the adiabatic heating analysis, Figure 1.3.6-5, are very similar to the
contours of plastic strain, Figure 1.3.6-3, from the thermally coupled analysis.
As noted earlier, excellent agreement is observed between the results obtained with ABAQUS/Explicit
and ABAQUS/Standard. Figure 1.3.6-6 compares the effects of adaptive meshing on the element
quality. The results obtained with adaptive meshing show significantly reduced mesh distortion. The
material point in the bar that experiences the largest temperature rise during the course of the
simulation is indicated (node 2029 in the model without adaptivity). Figure 1.3.6-7 compares the
results obtained with ABAQUS/Explicit for the temperature history of this material point against the
same results obtained with ABAQUS/Standard.
Input files

metalbarextrusion_coupled_fric.inp
Thermally coupled extrusion with frictional heat generation.
1-459
metalbarextrusion_adiab.inp
Extrusion with adiabatic heat generation and without frictional heat generation.
metalbarextrusion_stabil.inp
Thermally coupled extrusion with frictional heat generation and automatic stabilization.

metalbarextrusion_x_cax4rt.inp
Thermally coupled extrusion with frictional heat generation and without adaptive meshing; die
modeled with an analytical rigid surface; kinematic mechanical contact.
metalbarextrusion_xad_cax4rt.inp
Thermally coupled extrusion with frictional heat generation and adaptive meshing; die modeled
with an analytical rigid surface; kinematic mechanical contact.
metalbarextrusion_xd_cax4rt.inp
modeled with RAX2 elements; kinematic mechanical contact.
metalbarextrusion_xp_cax4rt.inp
modeled with an analytical rigid surface; penalty mechanical contact.
Figures
Figure 1.3.6-1 Mesh and geometry: axisymmetric extrusion, ABAQUS/Standard.
Figure 1.3.6-2 Deformed configuration, Step 2, ABAQUS/Standard.
1-460
Figure 1.3.6-3 Plastic strain contours, Step 2, thermally coupled analysis (frictional heat generation),
ABAQUS/Standard.
Figure 1.3.6-4 Temperature contours, Step 2, thermally coupled analysis (frictional heat generation),
1-461
ABAQUS/Standard.
Figure 1.3.6-5 Temperature contours, Step 2, adiabatic heat generation (without heat generation due
to friction), ABAQUS/Standard.
1-462
Figure 1.3.6-6 Deformed shape of the workpiece: without adaptive remeshing, left; with adaptive
remeshing, right; ABAQUS/Explicit.
1-463
Figure 1.3.6-7 Temperature history of node 2029 (nonadaptive result), ABAQUS/Explicit.
Sample listings
1-464
Listing 1.3.6-1
*HEADING
Extrusion
*NODE
1,0.,0.
61,0.,.3
2001,.1,0.
2061,.1,.3
*NODE, NSET=NREF
99999, 0.5 , 0.0
*NGEN,NSET=AXIS
1,61,1
*NGEN,NSET=OUTSIDE
2001,2061,1
*NFILL,NSET=ALL
AXIS,OUTSIDE,20,100
*ELEMENT,TYPE=CAX4T,ELSET=WORK
1,1,201,203,3
*ELGEN,ELSET=WORK
1,30,2,1,10,200,100
*ELSET,ELSET=BOT,GENERATE
1,901,100
*ELSET,ELSET=SIDE,GENERATE
901,930,1
*ELSET,ELSET=TOP,GENERATE
30,930,100
**
*** Node & element definitions for die
**
*NODE, NSET=CONTACT
10001, 0.250000000, -0.180000000
10002, 0.250000000, -0.114444000
10003, 0.250000000, -0.048888900
10004, 0.250000000, 0.016666700
10005, 0.250000000, 0.082222200
10006, 0.250000000, 0.147778000
10007, 0.250000000, 0.213333000
10008, 0.250000000, 0.278889000
10009, 0.250000000, 0.344444000
10010, 0.250000000, 0.410000000
10011, 0.188867000, -0.180000000
1-465
10012, 0.189030000, -0.106122000

10013, 0.199098000, -0.041751700
10014, 0.199227000, 0.022784300
10015, 0.199356000, 0.087320200
10016, 0.199484000, 0.151856000
10017, 0.199613000, 0.216392000
10018, 0.199742000, 0.280928000
10019, 0.199871000, 0.345464000
10020, 0.200000000, 0.410000000
10021, 0.127733000, -0.180000000
10022, 0.127815000, -0.103061000
10023, 0.149548000, -0.045875800
10024, 0.149613000, 0.012106400
10025, 0.149677000, 0.070088700
10026, 0.149742000, 0.128071000
10027, 0.149806000, 0.186053000
10028, 0.149871000, 0.244035000
10029, 0.149935000, 0.302018000
10030, 0.149999000, 0.360000000
10031, 0.066600000, -0.180000000
10032, 0.066600000, -0.100000000
10033, 0.099999000, -0.050000000
10034, 0.099999000, 0.001428570
10035, 0.099999000, 0.052857100
10036, 0.099999000, 0.104286000
10037, 0.099999000, 0.155714000
10038, 0.099999000, 0.207143000
10039, 0.099999000, 0.258571000
10040, 0.099999000, 0.310000000
*ELEMENT, TYPE=CAX4T, ELSET=CONTACT
10001, 10001, 10002, 10012, 10011
10002, 10002, 10003, 10013, 10012
10003, 10003, 10004, 10014, 10013
10004, 10004, 10005, 10015, 10014
10005, 10005, 10006, 10016, 10015
10006, 10006, 10007, 10017, 10016
10007, 10007, 10008, 10018, 10017
10008, 10008, 10009, 10019, 10018
10009, 10009, 10010, 10020, 10019
10010, 10011, 10012, 10022, 10021
10011, 10012, 10013, 10023, 10022
10012, 10013, 10014, 10024, 10023
10013, 10014, 10015, 10025, 10024
1-466
10014, 10015, 10016, 10026, 10025

10015, 10016, 10017, 10027, 10026
10016, 10017, 10018, 10028, 10027
10017, 10018, 10019, 10029, 10028
10018, 10019, 10020, 10030, 10029
10019, 10021, 10022, 10032, 10031
10020, 10022, 10023, 10033, 10032
10021, 10023, 10024, 10034, 10033
10022, 10024, 10025, 10035, 10034
10023, 10025, 10026, 10036, 10035
10024, 10026, 10027, 10037, 10036
10025, 10027, 10028, 10038, 10037
10026, 10028, 10029, 10039, 10038
10027, 10029, 10030, 10040, 10039
*SOLID SECTION, ELSET=CONTACT, MATERIAL=RIG
*RIGID BODY, ELSET=CONTACT, ISOTHERMAL=YES,
REF NODE=99999
*SOLID SECTION,ELSET=WORK,MATERIAL=METAL
*ELASTIC
6.9E10,.33
*PLASTIC
** STRAIN RATE APPX .1 **
60.E6,0.0 ,20.
90.E6,.125 ,20.
113.E6,.25 ,20.
124.E6,.375 ,20.
133.E6,0.5 ,20.
165.E6,1.0 ,20.
166.E6,2.0 ,20.
60.E6,0. ,50.
80.E6,.125 ,50.
97.E6,.25 ,50.
110.E6,.375 ,50.
120.E6,0.5 ,50.
150.E6,1.0 ,50.
151.E6,2.0 ,50.
50.E6,0.0 ,100.
65.E6,.125 ,100,
81.5E6,.25 ,100.
91.E6,.375 ,100.
100.E6,0.5 ,100.
125.E6,1.0 ,100.
1-467
126.E6,2.0 ,100.
45.E6,0.0 ,150.
63.E6,.125 ,150.
75.E6,.25 ,150.
89.E6,.5 ,150.
110.E6,1. ,150.
111.E6,2. ,150.
*SPECIFIC HEAT
880.,
*DENSITY
2700.,
*CONDUCTIVITY
204.,0.
225.,300.
*EXPANSION,ZERO=20.0
8.42E-5,
*INELASTIC HEAT FRACTION
.9,
**
*** material properties are inconsequential
*** for rigid elements
***
*MATERIAL, NAME=RIG
*ELASTIC
1.0E10,0.3
*SPECIFIC HEAT
880.,
*DENSITY
2700.,
*CONDUCTIVITY
204.,0.
225.,300.
8.42E-5,
**
**
**
61,2061,100
*NSET,NSET=ALL
1,2061,1
*INITIAL CONDITIONS,TYPE=TEMPERATURE
ALL,20.
1-468
**
*** surface definitions ***
**
*ELSET, ELSET=INDIE, GEN
10019,10027,1
*ELSET, ELSET=BOTDIE, GEN
10001,10019,9
*SURFACE, NAME=RIGID, TYPE=ELEMENT
INDIE, S3
BOTDIE,S4
*SURFACE, NAME=DEF1, TYPE=ELEMENT
SIDE, S2
BOT, S1
**
*** Interaction definitions
**
*SURFACE INTERACTION, NAME=INTER
*FRICTION
0.1
*GAP HEAT GENERATION
1.0,
**
*** Contact pair definitions
**
*CONTACT PAIR, INTERACTION=INTER, SMOOTH=0.48
DEF1, RIGID
DEF2, RIGID
**
*** elset for output purposes
**
*ELSET,ELSET=EFILEOUT
BOT,SIDE,TOP
**
*** step 1
**
*STEP,INC=100,AMPLITUDE=RAMP,NLGEOM, UNSYMM=YES
STABILIZE WORKPIECE INSIDE DIE
*COUPLED TEMPERATURE-DISPLACEMENT,DELTMX=100.
.1,1.
*BOUNDARY
NREF, 1, 2, 0.0
NREF, 6, 6, 0.0
1-469
NREF,11,11,20.0
AXIS,1,1,0.0
TOP,2,2,-.000125
ALL,11,11,20.0
**2061,1,1,0.0
*PRINT,CONTACT=YES
U
RF
NT
RFL
*END STEP
**
*** step 2
**
EXTRUSION
.1,10.
*BOUNDARY,OP=NEW
NREF, 1, 2, 0.0
NREF, 6, 6, 0.0
NREF, 11,11,20.0
AXIS,1,1,0.0
TOP,2,2,-.25
**2061,1,1,0.0
*PRINT,CONTACT=YES
*EL PRINT,ELSET=CONTACT,FREQUENCY=10
S
E
*EL FILE,FREQUENCY=999,ELSET=EFILEOUT
S,E
PE,
*NSET,NSET=NFILEOUT
AXIS,OUTSIDE,TOP
*NODE FILE,FREQUENCY=999,NSET=NFILEOUT
NT,
U
RF
NT
RFL
1-470
*END STEP
**
*** step 3
**
REMOVE CONTACT PAIRS
.1,.1,
*MODEL CHANGE,REMOVE, TYPE=CONTACT PAIR
DEF1, RIGID
DEF2, RIGID
*PRINT,CONTACT=YES
U
RF
NT
RFL
*END STEP
**
*** step 4
**
LET WORKPIECE COOL DOWN
100.,10000.,
*FILM
BOT,F1,20.,10.
TOP,F3,20.,10.
SIDE,F2,20.,10.
*PRINT,CONTACT=YES
U
RF
NT
RFL
*EL FILE,FREQUENCY=999,ELSET=EFILEOUT
S,E
PE,
1-471
*NODE FILE,FREQUENCY=999,NSET=NFILEOUT
NT,
*END STEP
1-472
Listing 1.3.6-2
*HEADING
Extrusion - adiabatic analysis
*NODE
1,0.,0.
61,0.,.3
2001,.1,0.
2061,.1,.3
*NODE, NSET=NREF
99999, 0.5, 0.0
**
*NGEN,NSET=AXIS
1,61,1
*NGEN,NSET=OUTSIDE
2001,2061,1
*NFILL,NSET=ALL
AXIS,OUTSIDE,20,100
*ELEMENT,TYPE=CAX4,ELSET=WORK
1,1,201,203,3
*ELGEN,ELSET=WORK
1,30,2,1,10,200,100
1,901,100
901,930,1
30,930,100
**
*** Node & element definitions for die
**
*NODE, NSET=CONTACT
10001, 0.250000000, -0.180000000
10002, 0.250000000, -0.114444000
10003, 0.250000000, -0.048888900
10004, 0.250000000, 0.016666700
10005, 0.250000000, 0.082222200
10006, 0.250000000, 0.147778000
10007, 0.250000000, 0.213333000
10008, 0.250000000, 0.278889000
10009, 0.250000000, 0.344444000
10010, 0.250000000, 0.410000000
1-473
10011, 0.188867000, -0.180000000

10012, 0.189030000, -0.106122000
10013, 0.199098000, -0.041751700
10014, 0.199227000, 0.022784300
10015, 0.199356000, 0.087320200
10016, 0.199484000, 0.151856000
10017, 0.199613000, 0.216392000
10018, 0.199742000, 0.280928000
10019, 0.199871000, 0.345464000
10020, 0.200000000, 0.410000000
10021, 0.127733000, -0.180000000
10022, 0.127815000, -0.103061000
10023, 0.149548000, -0.045875800
10024, 0.149613000, 0.012106400
10025, 0.149677000, 0.070088700
10026, 0.149742000, 0.128071000
10027, 0.149806000, 0.186053000
10028, 0.149871000, 0.244035000
10029, 0.149935000, 0.302018000
10030, 0.149999000, 0.360000000
10031, 0.066600000, -0.180000000
10032, 0.066600000, -0.100000000
10033, 0.099999000, -0.050000000
10034, 0.099999000, 0.001428570
10035, 0.099999000, 0.052857100
10036, 0.099999000, 0.104286000
10037, 0.099999000, 0.155714000
10038, 0.099999000, 0.207143000
10039, 0.099999000, 0.258571000
10040, 0.099999000, 0.310000000
*ELEMENT, TYPE=CAX4, ELSET=CONTACT
10001, 10001, 10002, 10012, 10011
10002, 10002, 10003, 10013, 10012
10003, 10003, 10004, 10014, 10013
10004, 10004, 10005, 10015, 10014
10005, 10005, 10006, 10016, 10015
10006, 10006, 10007, 10017, 10016
10007, 10007, 10008, 10018, 10017
10008, 10008, 10009, 10019, 10018
10009, 10009, 10010, 10020, 10019
10010, 10011, 10012, 10022, 10021
10011, 10012, 10013, 10023, 10022
10012, 10013, 10014, 10024, 10023
1-474
10013, 10014, 10015, 10025, 10024

10014, 10015, 10016, 10026, 10025
10015, 10016, 10017, 10027, 10026
10016, 10017, 10018, 10028, 10027
10017, 10018, 10019, 10029, 10028
10018, 10019, 10020, 10030, 10029
10019, 10021, 10022, 10032, 10031
10020, 10022, 10023, 10033, 10032
10021, 10023, 10024, 10034, 10033
10022, 10024, 10025, 10035, 10034
10023, 10025, 10026, 10036, 10035
10024, 10026, 10027, 10037, 10036
10025, 10027, 10028, 10038, 10037
10026, 10028, 10029, 10039, 10038
10027, 10029, 10030, 10040, 10039
*SOLID SECTION, ELSET=CONTACT, MATERIAL=RIG
*RIGID BODY, ELSET=CONTACT, REF NODE=99999
*SOLID SECTION,ELSET=WORK,MATERIAL=METAL
*ELASTIC
6.9E10,.33
*PLASTIC
60.E6,0.0 ,20.
90.E6,.125 ,20.
113.E6,.25 ,20.
124.E6,.375 ,20.
133.E6,0.5 ,20.
165.E6,1.0 ,20.
166.E6,2.0 ,20.
60.E6,0. ,50.
80.E6,.125 ,50.
97.E6,.25 ,50.
110.E6,.375 ,50.
120.E6,0.5 ,50.
150.E6,1.0 ,50.
151.E6,2.0 ,50.
50.E6,0.0 ,100.
65.E6,.125 ,100,
81.5E6,.25 ,100.
91.E6,.375 ,100.
100.E6,0.5 ,100.
125.E6,1.0 ,100.
1-475
126.E6,2.0 ,100.
45.E6,0.0 ,150.
63.E6,.125 ,150.
75.E6,.25 ,150.
89.E6,.5 ,150.
110.E6,1. ,150.
111.E6,2. ,150.
*SPECIFIC HEAT
880.,
*DENSITY
2700.,
*CONDUCTIVITY
204.,0.
225.,300.
8.42E-5,
.9,
61,2061,100
*NSET,NSET=ALL
1,2061,1
**
*** material properties are inconsequential
*** for rigid elements
***
*MATERIAL, NAME=RIG
*ELASTIC
1.0E10,0.3
*PLASTIC
60.E6,0.0 ,20.
90.E6,.125 ,20.
113.E6,.25 ,20.
124.E6,.375 ,20.
133.E6,0.5 ,20.
165.E6,1.0 ,20.
166.E6,2.0 ,20.
60.E6,0. ,50.
80.E6,.125 ,50.
97.E6,.25 ,50.
110.E6,.375 ,50.
120.E6,0.5 ,50.
1-476
150.E6,1.0 ,50.
151.E6,2.0 ,50.
50.E6,0.0 ,100.
65.E6,.125 ,100,
81.5E6,.25 ,100.
91.E6,.375 ,100.
100.E6,0.5 ,100.
125.E6,1.0 ,100.
126.E6,2.0 ,100.
45.E6,0.0 ,150.
63.E6,.125 ,150.
75.E6,.25 ,150.
89.E6,.5 ,150.
110.E6,1. ,150.
111.E6,2. ,150.
*SPECIFIC HEAT
880.,
*DENSITY
2700.,
*CONDUCTIVITY
204.,0.
225.,300.
8.42E-5,
**
*** surface definitions ***
**
*ELSET, ELSET=INDIE, GEN
10019,10027,1
*ELSET, ELSET=BOTDIE, GEN
10001,10019,9
*SURFACE, NAME=RIGID, TYPE=ELEMENT
INDIE, S3
BOTDIE,S4
SIDE, S2
BOT, S1
**
*** Interaction definitions
**
*FRICTION
1-477
0.1
1.0,
**
*** Contact pair definitions
**
*CONTACT PAIR, INTERACTION=INTER, SMOOTH=0.48
DEF1, RIGID
DEF2, RIGID
ALL,20.
*NSET, NSET=NALL
CONTACT,ALL
**
** step 1
**
*STATIC,ADIABATIC
.1,1.
*BOUNDARY
NREF,1,2,0.0
NREF,6,6,0.0
AXIS,1,1,0.0
TOP,2,2,-.000125
*PRINT,CONTACT=YES
S
E
MISES
PE
U
RF
*OUTPUT, FIELD, FREQ=10
*ELEMENT OUTPUT, ELSET=WORK
TEMP
S
E
PE
PEEQ
*NODE OUTPUT, NSET=ALL
1-478
U
*END STEP
**
*** step 2
**
EXTRUSION
*STATIC,ADIABATIC
.1,10.
*BOUNDARY,OP=NEW
NREF,1,2,0.0
NREF,6,6,0.0
AXIS,1,1,0.0
TOP,2,2,-.25
*PRINT,CONTACT=YES
S,E
U
RF
S
E
PE
TEMP
*OUTPUT, FIELD, FREQ=10
*ELEMENT OUTPUT, ELSET=WORK
TEMP
S
E
PE
PEEQ
*NODE OUTPUT, NSET=ALL
U
*END STEP
1-479
Listing 1.3.6-3
*HEADING
EXTRUSION OF A BAR WITH FRICTIONAL HEAT GENERATION
EXPLICIT [CAX4RT]
*NODE
1,0.,0.
61,0.,.3
2001,.1,0.
2061,.1,.3
*NGEN,NSET=AXIS
1,61,1
*NGEN,NSET=OUTSIDE
2001,2061,1
*NFILL,NSET=ALL
AXIS,OUTSIDE,20,100
*NSET, NSET=TEMP
2025, 2027, 2029, 2031
*ELEMENT, TYPE=CAX4RT, ELSET=WORK
1,1,201,203,3
*ELGEN,ELSET=WORK
1,30,2,1,10,200,100
1,901,100
901,930,1
30,930,100
*SOLID SECTION,ELSET=WORK,MATERIAL=METAL,
CONTROLS=HGLASS
*SECTION CONTROLS,NAME=HGLASS,HOURGLASS=COMBINED
*ELASTIC
6.9E10,.33
*PLASTIC
60.E6,0.0 ,20.
90.E6,.125 ,20.
113.E6,.25 ,20.
124.E6,.375 ,20.
133.E6,0.5 ,20.
165.E6,1.0 ,20.
166.E6,2.0 ,20.
1-480
60.E6,0. ,50.
80.E6,.125 ,50.
97.E6,.25 ,50.
110.E6,.375 ,50.
120.E6,0.5 ,50.
150.E6,1.0 ,50.
151.E6,2.0 ,50.
50.E6,0.0 ,100.
65.E6,.125 ,100,
81.5E6,.25 ,100.
91.E6,.375 ,100.
100.E6,0.5 ,100.
125.E6,1.0 ,100.
126.E6,2.0 ,100.
45.E6,0.0 ,150.
63.E6,.125 ,150.
75.E6,.25 ,150.
89.E6,.5 ,150.
110.E6,1. ,150.
111.E6,2. ,150.
*SPECIFIC HEAT
880.,
*DENSITY
2700.,
*CONDUCTIVITY
204.,0.
225.,300.
8.42E-5,
.9,
*NODE, NSET=REFNODE
9999, 0.2, 0.0, 0.0
*ELEMENT, TYPE=HEATCAP, ELSET=CAP
99001,9999
*HEATCAP, ELSET=CAP
1.,
**
61,2061,100
*NSET,NSET=ALL
1,2061,1
1-481
ALL,20.
*NSET,NSET=NFILEOUT
AXIS,OUTSIDE,TOP
*ELSET,ELSET=EFILEOUT
BOT,SIDE,TOP
*SURFACE, TYPE=S, NAME=DIE, FILLET=0.075
START,.25,-.18
LINE,.0866,-.18
LINE,.0666,-.18
LINE,.0666,-.17
LINE,.0666,-.15
LINE,.0666,-.1
LINE,.099999,-.05
LINE,.099999,0.0
LINE,.099999,.3
LINE,.099999,.31
LINE,.2,.41
*SURFACE,TYPE=ELEMENT, NAME=BAR
BOT, S1
SIDE, S2
*RIGID BODY, REFNODE=9999, ISOTHERMAL=YES,
ANALYTICAL SURFACE =DIE
*STEP
*DYNAMIC TEMPERATURE-DISPLACEMENT, EXPLICIT
,1.
*FIXED MASS SCALING, ELSET=WORK, FACTOR=1.E5
*CONTACT PAIR, INTERACTION=CONTACT
BAR, DIE
*SURFACE INTERACTION, NAME=CONTACT
*FRICTION
0.1 ,
1.0,0.5
REFNODE,1,6,0.0
AXIS,1,1,0.0
TOP,2,2,-.000125
2061,1,1,0.0
*BOUNDARY
ALL,11,11,20.0
REFNODE,11,11,20.0
*OUTPUT,FIELD,NUM=1,VAR=PRESELECT
1-482
*NODE OUTPUT, NSET=TEMP
NT,
*END STEP
**
*STEP
EXTRUSION
,10.
*BOUNDARY,OP=NEW, TYPE=VELOCITY
REFNODE,1,6,0.0
AXIS,1,1,0.0
TOP,2,2,-.0249875
2061,1,1,0.0
*BOUNDARY,OP=NEW
REFNODE,11,11,20.0
*FILE OUTPUT, NUM=1
*EL FILE,,ELSET=EFILEOUT
S,E
PE,
*NODE FILE,NSET=NFILEOUT
NT,
*EL FILE,ELSET=EFILEOUT
S,E
PE,
NT,
*OUTPUT, HISTORY, TIME INTERVAL=35.0
NT,
*OUTPUT, FIELD,NUM=4
*ELEMENT OUTPUT,ELSET=WORK
S,PEEQ, TEMP, HFL
*NODE OUTPUT
NT, U
*END STEP
**
*STEP
REMOVE CONTACT
,.1
1-483
*CONTACT PAIR, OP=DELETE

BAR,DIE
*BOUNDARY,OP=NEW, TYPE=VELOCITY
REFNODE,1,6,0.0
AXIS,1,1,0.0
TOP,2,2,0.0
*END STEP
**
*STEP
LET WORKPIECE COOL DOWN--I (ADD VISCOUS PRESSURE)
,10.
*FIXED MASS SCALING,ELSET=WORK,FACTOR=1.E6
*AMPLITUDE, NAME=RAMP,TIME=TOTAL TIME
0.0,0.0,11.1,0.0,10011.1,1.0
*FILM, AMP=RAMP
BOT,F1,20.,10.
TOP,F3,20.,10.
SIDE,F2,20.,10.
**
*DLOAD
WORK,VP1,1.E8
WORK,VP2,1.E8
WORK,VP3,1.E8
WORK,VP4,1.E8
**
*FILE OUTPUT, NUM=2
*EL FILE,ELSET=EFILEOUT
S,E
PE,
NT,
*OUTPUT, HISTORY, TIME INTERVAL=35.0
NT,
*OUTPUT, FIELD,NUM=2
*ELEMENT OUTPUT,ELSET=WORK
S,PEEQ, TEMP, HFL
*NODE OUTPUT
NT, U
*END STEP
*STEP
LET WORKPIECE COOL DOWN--II
1-484

,9990.
*FIXED MASS SCALING,ELSET=WORK,FACTOR=4.E9
**
*DLOAD,OP=NEW
**
*END STEP
1.3.7 Rolling of thick plates

Hot rolling is a basic manufacturing technique used to transform preformed shapes into a form suitable
for further processing. Rolling processes can be divided into different categories, depending on the
complexity of metal flow and on the geometry of the rolled product. Finite element computations are
used increasingly to analyze the elongation and spread of the material during rolling (Kobayashi,
1989). Although the forming process is often carried out at low roll speed, this example shows that a
considerable amount of engineering information can be obtained by using the explicit dynamics
procedure in ABAQUS/Explicit to model the process.
The rolling process is first investigated using plane strain computations. These results are used to
choose the modeling parameters associated with the more computationally expensive
three-dimensional analysis.
Since rolling is normally performed at relatively low speeds, it is natural to assume that static analysis
is the proper modeling approach. Typical rolling speeds (surface speed of the roller) are on the order of
1 m/sec. At these speeds inertia effects are not significant, so the response--except for rate effects in
the material behavior--is quasi-static. Representative rolling geometries generally require
three-dimensional modeling, resulting in very large models, and include nonlinear material behavior
and discontinuous effects--contact and friction. Because the problem size is large and the
discontinuous effects dominate the solution, the explicit dynamics approach is often less expensive
computationally and more reliable than an implicit quasi-static solution technique.
The computer time involved in running a simulation using explicit time integration with a given mesh
is directly proportional to the time period of the event. This is because numerical stability
considerations restrict the time increment to
µ r ¶
el ½
¢t ∙ min L ;
¸ + 2¹
where the minimum is taken over all elements in the mesh, Lel is a characteristic length associated
with an element, ½ is the density of the material in the element, and ¸ and ¹ are the effective Lamé's
constants for the material in the element. Since this condition effectively means that the time increment
can be no larger than the time required to propagate a stress wave across an element, the computer time
involved in running a quasi-static analysis can be very large. The cost of the simulation is directly
proportional to the number of time increments required, n = T =¢t if ¢t remains constant, where T is
1-485
the time period of the event being simulated. ( ¢t will not remain constant in general, since element
distortion will change Lel and nonlinear material response will change the effective Lamé constants
and density. But the assumption is acceptable for the purposes of this discussion.) Thus,
Ã s !
1 ¸ + 2¹
n = T max :
Lel ½
To reduce n, we can speed up the simulation compared to the time of the actual process; that is, we can
artificially reduce the time period of the event, T . This will introduce two possible errors. If the
simulation speed is increased too much, the inertia forces will be larger and will change the predicted
response (in an extreme case the problem will exhibit wave propagation response). The only way to
avoid this error is to find a speed-up that is not too large. The other error is that some aspects of the
problem other than inertia forces--for example, material behavior--may also be rate dependent. This
implies that we cannot change the actual time period of the event being modeled. But we can see a
simple equivalent--artificially increasing the material density, ½, by a factor f 2 reduces n to n=f , just
as decreasing T to T =f . This concept, which is called "mass scaling," reduces the ratio of the event
time to the time for wave propagation across an element while leaving the event time fixed, thus
allowing treatment of rate-dependent material and other behaviors, while having exactly the same
effect on inertia forces as speeding up the time of simulation. Mass scaling is attractive because it
allows us to treat rate-dependent quasi-static problems efficiently. But we cannot take it too far or we
allow the inertia forces to dominate and, thus, change the solution. This example illustrates the use of
mass scaling and shows how far we can take it for a practical case.
Problem description
A steel plate of an original square cross-section of 40 mm by 40 mm and a length of 92 mm is reduced
to a 30 mm height by rolling through one roll stand. The radius of the rollers is 170 mm. The single
roller in the model (taking advantage of symmetry) is assumed to be rigid and is modeled as an
analytical rigid surface. The isotropic hardening yield curve of the steel is taken from Kopp and
Dohmen (1990). Isotropic elasticity is assumed, with Young's modulus of 150 GPa and Poisson's ratio
of 0.3. The strain hardening is described using 11 points on the yield stress versus plastic strain curve,
with an initial yield stress of 168.2 MPa and a maximum yield stress of 448.45 MPa. No rate
dependence or temperature dependence is taken into account.
Coulomb friction is assumed between the roller and the plate, with a friction coefficient of 0.3. Friction
plays an important role in this process, as it is the only mechanism by which the plate is pulled through
the roll stand. If the friction coefficient is too low, the plate cannot be drawn through the roll stand.
Initially, when a point on the surface of the plate has just made contact with the roller, the roller
surface is moving faster than the point on the surface of the plate and there is a relative slip between
the two surfaces. As the point on the plate is drawn into the process zone under the roller, it moves
faster and, after a certain distance, sticks to the roller. As the point on the surface of the plate is pushed
out of the process zone, it picks up speed and begins to move faster than the roller. This causes slip in
the opposite direction before the point on the surface of the sheet finally loses contact with the
roller.
1-486
For plane strain computations a half-symmetry model with CPE4R elements is used. For the
three-dimensional computations a one-quarter symmetry model with C3D8R elements is used. The
roller is modeled with analytical rigid surfaces for both the two-dimensional and three-dimensional
cases. For quasi-static rolling problems perfectly round analytical surfaces can provide a more accurate
representation of the revolved roller geometry, improve computational efficiency, and reduce noise
when compared to element-based rigid surfaces.
The roller is rotated through 32° at a constant angular velocity of 1 revolution per second (6.28
rad/sec), which corresponds to a roller surface speed of 1.07 m/sec. The plate is given an initial
velocity in the global x-direction. The initial velocity is chosen to match the x-component of velocity
of the roller at the point of first contact. This choice of initial velocity results in a net acceleration of
zero in the x-direction at the point of contact and minimizes the initial impact between the plate and
the roller. This minimizes the initial transient disturbance.
In each analysis performed in this example, the *FIXED MASS SCALING option is used to scale the
masses of all the elements in the model by factors of either 110, 2758, or 68962. These scaling factors
translate into effective roller surface speeds of 11.2 m/sec, 56.1 m/sec, and 280.5 m/sec. An
alternative, but equivalent, means of mass scaling could be achieved by scaling the actual density
(entered on the *DENSITY option) by the aforementioned factors.
The element formulation for the two-dimensional (using CPE4R elements) and three-dimensional
(using C3D8R elements) analyses uses the pure stiffness form of hourglass control
(HOURGLASS=STIFFNESS). The element formulation is selected using the *SECTION CONTROLS
option. In addition, the three-dimensional model (using C3D8R elements) uses the centroidal
(KINEMATIC SPLIT=CENTROID) kinematic formulation. These options are economical yet provide
the necessary level of accuracy for this class of problems. Other cases using more computationally
intensive element formulations are included for comparison: analyses that use the default section
control options and a mass scaling factor of 2758 and two- and three-dimensional analyses that use an
element formulation intermediate in computational cost between the two previous formulations.
For the sole purpose of testing the performances of the modified triangular and tetrahedral elements,
the problem is also analyzed in two dimensions using CPE6M elements and in three dimensions using
C3D10M elements.

Table 1.3.7-1 shows the effective rolling speeds and the relative CPU cost of the cases using the
element formulations recommended for this problem. The relative costs are normalized with respect to
the CPU time for the two-dimensional model (using CPE4R elements) with the intermediate mass
scaling value. In addition, Table 1.3.7-2 compares the relative CPU cost and accuracy between the
different element formulations of the solid elements using the intermediate mass scaling value.
Plane strain rolling (CPE4R elements)

A plane strain calculation allows the user to resolve a number of modeling questions in two
dimensions before attempting a more expensive three-dimensional calculation. In particular, an
acceptable effective mass scaling factor for running the transient dynamics procedure can be
1-487
determined.
Figure 1.3.7-1 through Figure 1.3.7-3 show contours of equivalent plastic strain for the three mass
scaling factors using the STIFFNESS hourglass control. Figure 1.3.7-4 through Figure 1.3.7-6 show
contours of shear stress for the same cases. These results show that there is very little difference
between the lowest and the intermediate mass scaling cases. All the results are in good agreement with
the quasi-static analysis results obtained with ABAQUS. The results of the largest mass scaling case
show pronounced dynamic effects. Table 1.3.7-1 shows the relative run time of the quasi-static
calculation, and Table 1.3.7-2 compares the different element formulations at the same level of mass
scaling. The intermediate mass scaling case gives essentially the same results as the quasi-static
calculation, using about one-seventh of the CPU time. In addition to the savings provided by the mass
scaling option, more computational savings are achieved using the chosen element formulation of
STIFFNESS hourglass control; the results for this formulation compare well to the results for the
computationally more expensive element formulations.
Three-dimensional rolling (C3D8R elements)

We have ascertained with the two-dimensional calculations that mass scaling by a factor of 2758 gives
results that are essentially the same as a quasi-static solution. Figure 1.3.7-7 shows the distribution of
the equivalent plastic strain of the deformed sheet for the three-dimensional case using the
CENTROID kinematic and STIFFNESS hourglass section control options. Figure 1.3.7-8 shows the
distribution of the equivalent plastic strain of the deformed sheet for the three-dimensional case using
the default section control options ( AVERAGE STRAIN kinematic and RELAX STIFFNESS
hourglass). Table 1.3.7-1 compares this three-dimensional case with the plane strain and quasi-static
cases, and Table 1.3.7-2 compares the three different three-dimensional element formulations included
here with the two-dimensional cases at the same level of mass scaling. The accuracy for all three
element formulations tested is very similar for this problem, but significant savings are realized in the
three-dimensional analyses when using more economical element formulations.
Analyses using CPE6M and C3D10M elements

The total number of nodes in the CPE6M model is identical to the number in the CPE4R model. The
number of nodes in the C3D10M model is 3440 (compared to 3808 in the C3D8R model). The
analyses using the CPE6M and C3D10M elements use a mass scaling factor of 2758. Figure 1.3.7-9
and Figure 1.3.7-10 show the distribution of the equivalent plastic strain of the plate for the
two-dimensional and three-dimensional cases, respectively. The results are in reasonably good
agreement with other element formulations. However, the CPU costs are higher since the modified
triangular and tetrahedral elements use more than one integration point in each element and the stable
time increment size is somewhat smaller than in analyses that use reduced integration elements with
the same node count. For the mesh refinements used in this problem, the CPE6M model takes about
twice the CPU time as the CPE4R model, while the C3D10M model takes about 5.75 times the CPU
time as the C3D8R model.
Input files
roll2d330_anl_ss.inp
1-488
Two-dimensional case (using CPE4R elements) with a mass scaling factor of 2758 and the
STIFFNESS hourglass control.
roll3d330_rev_anl_css.inp
Three-dimensional case (using C3D8R elements) with a mass scaling factor of 2758; an analytical
rigid surface of TYPE=REVOLUTION; and the CENTROID kinematic and STIFFNESS hourglass
section control options.
roll2d66_anl_ss.inp
Two-dimensional case (using CPE4R elements) with a mass scaling factor of 110 using the
roll2d330_anl_cs.inp
COMBINED hourglass control.
roll2d330_cs.inp
COMBINED hourglass control and rigid elements.
roll3d330_css.inp
Three-dimensional case (using C3D8R elements) with a mass scaling factor of 2758, rigid
elements, and the CENTROID kinematic and STIFFNESS hourglass section control options.
roll3d330_ocs.inp
Three-dimensional case (using C3D8R elements) with a mass scaling factor of 2758, rigid
elements, and the ORTHOGONAL kinematic and COMBINED hourglass section control options.
roll2d1650_anl_ss.inp
roll3d330_rev_anl_ocs.inp
Three-dimensional model (using C3D8R elements) with a mass scaling factor of 2758; an
analytical rigid surface of TYPE=REVOLUTION; and the ORTHOGONAL kinematic and
COMBINED hourglass section control options.
roll3d330_rev_anl.inp
Three-dimensional model (using C3D8R elements) with a mass scaling factor of 2758, an
analytical rigid surface of TYPE=REVOLUTION, and the default section control options.
roll3d330_cyl_anl.inp
Three-dimensional model (using C3D8R elements) with a mass scaling factor of 2758, an
analytical rigid surface of TYPE=CYLINDER, and the default section control options.
roll2d66.inp
1-489
Two-dimensional model (using CPE4R elements) with a mass scaling factor of 110 and default
section controls.
roll2d330.inp
section controls.
roll2d1650.inp
section controls.
roll3d330.inp
Three-dimensional model using rigid elements and default section controls.
roll2d66_anl.inp
Two-dimensional model (using CPE4R elements) with a mass scaling factor of 110 using
analytical rigid surfaces and default section controls.
roll2d330_anl.inp
roll2d1650_anl.inp
roll2d330_anl_cpe6m.inp
Two-dimensional case (using CPE6M elements) with a mass scaling factor of 2758.
roll3d330_anl_c3d10m.inp
Three-dimensional case (using C3D10M elements) with a mass scaling factor of 2758.
roll3d_medium.inp
Additional mesh refinement case (using C3D8R elements) included for the sole purpose of testing
the performance of the code.
References
· Kobayashi, S., S. I. Oh, and T. Altan, Metal Forming and the Finite Element Method , Oxford
University Press, 1989.
· Kopp, R., and P. M. Dohmen, "Simulation und Planung von Walzprozessen mit Hilfe der
Finite-Elemente-Methode (FEM)," Stahl U. Eisen, no. 7, pp. 131-136, 1990.
Tables
1-490
Table 1.3.7-1 Analysis cases and relative CPU costs. (The two-dimensional explicit analyses all use
CPE4R elements and the STIFFNESS hourglass control. The three-dimensional explicit analysis uses
C3D8R elements and the CENTROID kinematic, STIFFNESS hourglass section control options.)
Analysis Type Mass Scaling Effective Roll Surface Relative
CPU
Factor Speed (m/sec) Time
Explicit, plane strain 110.3 11.2 4.99
Implicit, plane strain quasi-static 6.90
Explicit, 3-D 2758.5 56.1 16.0
Table 1.3.7-2 Explicit section control options tested (mass scaling factor=2758.5). CPE4R and
C3D8R elements are employed for the two-dimensional and three-dimensional cases, respectively.
Spread values are reported for the half-model at node 24015.
Analysis Type Section Controls Relative Sprea
Kinematic Hourglass CPU d
Time (mm)
Explicit, plane strain n/a STIFFNES 1.00 n/a
S
Explicit, plane strain n/a RELAX 1.11 n/a
Explicit, plane strain n/a COMBINE 1.04 n/a
D
Explicit, 3-D AVERAGE RELAX 30.7 2.06
STRAIN STIFFNES
S
Explicit, 3-D ORTHOGON COMBINE 21.2 2.07
AL D
Explicit, 3-D CENTROID STIFFNES 16.0 2.10
S
Figures
Figure 1.3.7-1 Equivalent plastic strain for the plane strain case ( CPE4R) with STIFFNESS hourglass
control (mass scaling factor=110.3).
1-491
Figure 1.3.7-4 Shear stress for the plane strain case (CPE4R) with STIFFNESS hourglass control
(mass scaling factor=110.3).
1-492
Figure 1.3.7-7 Equivalent plastic strain for the three-dimensional case ( C3D8R) using the
CENTROID kinematic and STIFFNESS hourglass section control options (mass scaling
1-493
factor=2758.5).
Figure 1.3.7-8 Equivalent plastic strain for the three-dimensional case ( C3D8R) using the AVERAGE
STRAIN kinematic and RELAX STIFFNESS hourglass section control options (mass scaling
factor=2758.5).
Figure 1.3.7-9 Equivalent plastic strain for the plane strain case ( CPE6M) (mass scaling
factor=2758.5).
1-494
Figure 1.3.7-10 Equivalent plastic strain for the three-dimensional case ( C3D10M) (mass scaling
factor=2758.5).
Sample listings
1-495
Listing 1.3.7-1
*HEADING
Thick plate rolling: Plane Strain,
ABAQUS/Explicit
(Analytical rigid surfaces)
SECTION CONTROLS USED (HOURGLASS=STIFFNESS)
**
*NODE
** Bar
1, 0., 0.
401, 0., 0.020
47, -00.092, 0.
447, -00.092, 0.020
**
*NGEN,NSET=BOTTOM
1,47,1
*NGEN,NSET=TOP
401,447,1
**
*NFILL,NSET=BAR
BOTTOM,TOP,8,50
**
***** Bar
**
*ELEMENT,TYPE=CPE4R,ELSET=METAL
1,1,51,52,2
*ELGEN,ELSET=METAL
1,8,50,50,46,1,1
**
351,396,1
*ELSET,ELSET=BACK,GEN
46,396,50
**
*SOLID SECTION,ELSET=METAL,MAT=C15,CONTROL=B
1.,
**
*MATERIAL,NAME=C15
*ELASTIC
1.5E11,.3
1-496
*PLASTIC
168.72E06,0
219.33E06,0.1
272.02E06,0.2
308.53E06,0.3
337.37E06,0.4
361.58E06,0.5
382.65E06,0.6
401.42E06,0.7
418.42E06,0.8
434.01E06,0.9
448.45E06,1.0
*DENSITY
7.85E3,
**
** Node for rigid surface
*NODE
10000, 0.0409 , 0.185
*INITIAL CONDITIONS,TYPE=VELOCITY
BAR,1,1.0367
**
**
*SURFACE, TYPE=SEGMENTS,NAME=ROLLER
START, 0.040900, 0.015000
CIRCL, -.129100, 0.185000 , 0.0409 , 0.185
*SURFACE,TYPE=ELEMENT, NAME=SURF1
TOP,S2
ANALYTICAL SURFACE=ROLLER
*STEP
*DYNAMIC,EXPLICIT
,0.089286
*FIXED MASS SCALING,FACTOR=2758.5,ELSET=METAL
**
** Roller, Radius = .170 m
**
*BOUNDARY
BOTTOM,2,2
10000,1,2
**
10000,6,6,6.2832
**
1-497
*SURFACE INTERACTION,NAME=FRICT
*FRICTION
0.3,
*CONTACT PAIR,INTERACTION=FRICT
SURF1,ROLLER
**
**
*EL FILE
PEEQ,MISES,PE,LE
*NODE FILE
U,
*ENERGY FILE
*END STEP
**
1-498
Listing 1.3.7-2
*HEADING
Thick plate rolling: 3-Dimensional,
ABAQUS/Explicit
SECTION CONTROLS USED
(KINEMATIC=CENTROID, HOURGLASS=STIFFNESS)
*NODE
** Bar
1, 0., 0.
801, 0., 0.020
47, -00.092, 0.
847, -00.092, 0.020
24001, 0., 0. , -0.020
24801, 0., 0.020, -0.020
24047, -00.092, 0. , -0.020
24847, -00.092, 0.020, -0.020
*NGEN,NSET=BOT1
1,47,1
*NGEN,NSET=TOP1
801,847,1
*NGEN,NSET=BOT2
24001,24047,1
*NGEN,NSET=TOP2
24801,24847,1
*NFILL,NSET=ZSYMM
BOT1,TOP1,8,100
*NFILL,NSET=SIDE
BOT2,TOP2,8,100
*NFILL,NSET=BAR
ZSYMM,SIDE,8,3000
*NSET,NSET=BOTTOM,GEN
1,47,1
3001,3047,1
6001,6047,1
9001,9047,1
12001,12047,1
15001,15047,1
18001,18047,1
21001,21047,1
24001,24047,1
**
1-499
***** Bar
**
*ELEMENT,TYPE=C3D8R,ELSET=METAL
1,2,1,3001,3002,102,101,3101,3102
*ELGEN,ELSET=METAL
1,8,100,100,46,1,1,8,3000,1000
**
701,746,1
1701,1746,1
2701,2746,1
3701,3746,1
4701,4746,1
5701,5746,1
6701,6746,1
7701,7746,1
*ELSET,ELSET=BACK,GEN
46,746,100
1046,1746,100
2046,2746,100
3046,3746,100
4046,4746,100
5046,5746,100
6046,6746,100
7046,7746,100
7001,7046,1
7101,7146,1
7201,7246,1
7301,7346,1
7401,7446,1
7501,7546,1
7601,7646,1
7701,7746,1
**
*SOLID SECTION,ELSET=METAL,MAT=C15,CONTROL=C
1.,
*SECTION CONTROLS,KINEMATIC=CENTROID,
HOURGLASS=STIFFNESS, NAME=C
**
*MATERIAL,NAME=C15
*ELASTIC
1.5E11,.3
1-500
*PLASTIC
168.72E06,0
219.33E06,0.1
272.02E06,0.2
308.53E06,0.3
337.37E06,0.4
361.58E06,0.5
382.65E06,0.6
401.42E06,0.7
418.42E06,0.8
434.01E06,0.9
448.45E06,1.0
*DENSITY
7.85E3,
**
*NODE
** Reference node
30000, 0.0409 , 0.185 , -0.010
BAR,1,1.0367
**
***************** Step 1
*SURFACE, NAME=ROLLER, TYPE=REVOL
0.0409,0.185,-0.025,0.0409,0.185,0.005
START, 0.170,0.03
LINE, 0.170,0.0
*SURFACE,TYPE=ELEMENT, NAME=SURF1
TOP,S2
SIDE,S5
ANALYTICAL SURFACE=ROLLER
*STEP
*DYNAMIC,EXPLICIT
,0.089286
*FIXED MASS SCALING,FACTOR=2758.5,ELSET=METAL
** Roller, Radius = .170 m
*BOUNDARY
BOTTOM,2,2
ZSYMM,ZSYMM
30000,1,5
**
**30000,6,6,330.0
1-501
30000,6,6,6.2832
**
*FRICTION
0.3,
SURF1,ROLLER
**
*EL FILE
PEEQ,MISES,PE,LE
*NODE FILE
U,
*ENERGY FILE
*END STEP
1.3.8 Axisymmetric forming of a circular cup

This example illustrates the hydroforming of a circular cup using an axisymmetric model. In this case a
two-stage forming sequence is used, with annealing between the stages. Two analysis methods are
used: in one the entire process is analyzed using ABAQUS/Explicit; in the other the forming sequences
are analyzed with ABAQUS/Explicit, while the springback analyses are run in ABAQUS/Standard.
Here, the import capability is used to transfer results between ABAQUS/Explicit and
ABAQUS/Standard and vice versa.
Problem description
The model consists of a deformable blank and three rigid dies. The blank has a radius of 150.0 mm, is
1.0 mm thick, and is modeled using axisymmetric shell elements, SAX1. The coefficient of friction
between the blank and the dies is taken to be 0.1. Dies 1 and 2 are offset from the blank by half of the
thickness of the blank, because the contact algorithm takes into account the shell thickness. To avoid
pinching of the blank while die 3 is put into position for the second forming stage, the radial gap
between dies 2 and 3 is set to be 20% bigger than the initial shell thickness. Figure 1.3.8-1 and Figure
1.3.8-2 show the initial geometry of the model.
The three dies are modeled with either two-dimensional analytical rigid surfaces or RAX2 rigid
elements. An analytical rigid surface can yield a more accurate representation of two-dimensional
curved punch geometries and result in computational savings. Contact pressure can be viewed on the
specimen surface, and the reaction force is available at the rigid body reference node. In addition, both
the kinematic (default) and penalty contact formulations are tested. Results for the kinematic contact
formulation using rigid elements are presented here.
The blank is made of aluminum-killed steel, which is assumed to satisfy the Ramberg-Osgood relation
between true stress and logarithmic strain,
1-502
² = (¾=K )1=n ;
with a reference stress value (K) of 513 MPa and work-hardening exponent (n) of 0.223. Isotropic
elasticity is assumed, with Young's modulus of 211 GPa and Poisson's ratio of 0.3. With these data an
initial yield stress of 91.3 MPa is obtained. The stress-strain behavior is defined by piecewise linear
segments matching the Ramberg-Osgood curve up to a total (logarithmic) strain level of 107%, with
Mises yield, isotropic hardening, and no rate dependence.
The analysis that is performed entirely within ABAQUS/Explicit consists of six steps. In the first step
contact is defined between the blank and dies 1 and 2. Both dies remain fixed while a distributed load
of 10 MPa in the negative z-direction is ramped onto the blank. This load is then ramped off in the
second step, allowing the blank to spring back to an equilibrium state.
The third step is an annealing step. The annealing procedure in ABAQUS/Explicit sets all appropriate
state variables to zero. These variables include stresses, strains (excluding the thinning strain for
shells, membranes, and plane stress elements), plastic strains, and velocities. There is no time
associated with an annealing step. The process occurs instantaneously.
In the fourth step contact is defined between the blank and die 3 and contact is removed between the
blank and die 1. Die 3 moves down vertically in preparation for the next pressure loading.
In the fifth step another distributed load is applied to the blank in the positive z-direction, forcing the
blank into die 3. This load is then ramped off in the sixth step to monitor the springback of the
blank.
To obtain a quasi-static response, an investigation was conducted to determine the optimum rate for
applying the pressure loads and removing them. The optimum rate balances the computational time
against the accuracy of the results; increasing the loading rate will reduce the computer time but lead to
less accurate quasi-static results.
The analysis that uses the import capability consists of four runs. The first run is identical to Step 1 of
the ABAQUS/Explicit analysis described earlier. In the second run the ABAQUS/Explicit results for
the first forming stage are imported into ABAQUS/Standard (using UPDATE=NO and STATE=YES
on the *IMPORT option) for the first springback analysis. The third run imports the results of the first
springback analysis into ABAQUS/Explicit for the subsequent annealing process and the second
forming stage. By setting UPDATE=YES and STATE=NO on the *IMPORT option, this run begins
with no initial stresses or strains, effectively simulating the annealing process. The final run imports
the results of the second forming stage into ABAQUS/Standard for the second springback analysis.

Figure 1.3.8-3 to Figure 1.3.8-5show the results of the analysis conducted entirely within
ABAQUS/Explicit using the rigid element approach and the kinematic contact formulation. Figure
1.3.8-3shows the deformed shape at the end of Step 2, after the elastic springback. Figure 1.3.8-4
shows the deformed shape at the end of the analysis, after the second elastic springback. Although it is
not shown here, the amount of springback observed during the unloading steps is negligible. Figure
1.3.8-5 shows a contour plot of the shell thickness ( STH) at the end of the analysis. The thickness of
1-503
the material at the center of the cup has been reduced by about 20%, while the thickness at the edges of
the cup has been increased by about 10%.
The results obtained using the import capability to perform the springback analyses in
ABAQUS/Standard are nearly identical, as are those obtained using analytical rigid surfaces and/or
penalty contact formulations.
Input files
axiform.inp
ABAQUS/Explicit analysis that uses rigid elements and kinematic contact. This file is also used
for the first step of the analysis that uses the import capability.
axiform_anl.inp
Model using analytical rigid surfaces and kinematic contact.
axiform_pen.inp
Model using rigid elements and penalty contact.
axiform_anl_pen.inp
Model using analytical rigid surfaces and penalty contact.
axiform_sprbk1.inp
First springback analysis using the import capability.
axiform_form2.inp
Second forming analysis using the import capability.
axiform_sprbk2.inp
Second springback analysis using the import capability.
axiform_restart.inp
Restart of axiform.inp included for the purpose of testing the restart capability.
axiform_rest_anl.inp
Restart of axiform_anl.inp included for the purpose of testing the restart capability.
Figures
Figure 1.3.8-1 Configuration at the beginning of stage 1.
1-504
Figure 1.3.8-2 Configuration of dies in forming stage 2. (The dotted line shows the initial position of
die 3.)
Figure 1.3.8-3 Deformed configuration after the first forming stage.
1-505
Figure 1.3.8-4 Final configuration.
Figure 1.3.8-5 Contour plot of shell thickness.
Sample listings
1-506
Listing 1.3.8-1
*HEADING
SHEET METAL FORMING WITH ANNEALING
DIE3 SLIDES DOWN FROM ABOVE TO AVOID
INITIAL OVERCLOSURE
*NODE,NSET=BLANK
1,0.,0.0005
41,.150,0.0005
*NGEN,NSET=BLANK
1,41,1
1, 1,2
*ELGEN,ELSET=BLANK
1, 40,1,1
*SHELL SECTION,ELSET=BLANK,MATERIAL=STEEL,
.001,5
*DENSITY
7800.,
*ELASTIC
2.1E11,0.3
*PLASTIC
0.91294E+08, 0.00000E+00
0.10129E+09, 0.21052E-03
0.11129E+09, 0.52686E-03
0.12129E+09, 0.97685E-03
0.13129E+09, 0.15923E-02
0.14129E+09, 0.24090E-02
0.15129E+09, 0.34674E-02
0.16129E+09, 0.48120E-02
0.17129E+09, 0.64921E-02
0.18129E+09, 0.85618E-02
0.19129E+09, 0.11080E-01
0.20129E+09, 0.14110E-01
0.21129E+09, 0.17723E-01
0.22129E+09, 0.21991E-01
0.23129E+09, 0.26994E-01
0.24129E+09, 0.32819E-01
0.25129E+09, 0.39556E-01
0.26129E+09, 0.47301E-01
0.27129E+09, 0.56159E-01
1-507
0.28129E+09, 0.66236E-01
0.29129E+09, 0.77648E-01
0.30129E+09, 0.90516E-01
0.31129E+09, 0.10497E+00
0.32129E+09, 0.12114E+00
0.33129E+09, 0.13916E+00
0.34129E+09, 0.15919E+00
0.35129E+09, 0.18138E+00
0.36129E+09, 0.20588E+00
0.37129E+09, 0.23287E+00
0.38129E+09, 0.26252E+00
0.39129E+09, 0.29502E+00
0.40129E+09, 0.33054E+00
0.41129E+09, 0.36929E+00
0.42129E+09, 0.41147E+00
0.43129E+09, 0.45729E+00
0.44129E+09, 0.50696E+00
0.45129E+09, 0.56073E+00
0.46129E+09, 0.61881E+00
0.47129E+09, 0.68145E+00
0.48129E+09, 0.74890E+00
0.49129E+09, 0.82142E+00
0.50129E+09, 0.89928E+00
0.51129E+09, 0.98274E+00
0.52129E+09, 0.10721E+01
*NODE,NSET=DIE1
1001, 0.,-0.05
1002,.090,-0.05
1011,.100,-.040
199991, 0., -0.05
*NGEN,NSET=DIE1,LINE=C
1002,1011,1,,.090,-.040,0.
*ELEMENT,TYPE=RAX2,ELSET=DIE1
1001, 1001,1002
*ELGEN,ELSET=DIE1
1001, 10,1,1
*NODE,NSET=DIE2
2001,.100,-.060
2002,.100,-.010
2011,.110, 0.
2012,.160, 0.
299991, .100, -.060
1-508
2002,2011,1,,.110,-.010,0.,0.,0.,-1.
2001, 2001,2002
*ELGEN,ELSET=DIE2
2001, 11,1,1
*NODE,NSET=DIE3
** raised by 0.05 from original shift outer
** surface inward by then lower by 0.005
** (half the radius of curvature)
** further lower by 0.005
3001,.0 ,0.0044
3014,.050,-0.009
3015,.0888,-0.009
3024,.0988,0.001
3025,.0988,0.0405
399991,.0,0.0044
3001,3014,1,, 0.0,-0.0956,0.,0.,0.,-1.
3015,3024,1,,.0888, 0.001, 0.
3001, 3001,3002
*ELGEN,ELSET=DIE3
3001, 24,1,1
*BOUNDARY
1,XSYMM
199991,1,6
299991,1,6
399991,1,2
399991,6,6
*AMPLITUDE,NAME=R1,TIME=STEP TIME
0.,0., .8E-3,1.5E6, 1.7E-3,1.5E6, 3.E-3,1.E7,
3.5E-3,1.E7
0.,1.E7, 1.E-3,0.
*AMPLITUDE,NAME=R3A,TIME=STEP TIME,
0.,0., 1.E-3, 1.0
0.,0., 1.E-3,.6E7
0.,.6E7, 1.E-3,0.
**
1-509
** First downward pressure loading

*SURFACE,TYPE=ELEMENT,NAME=TOP
BLANK,SPOS
*SURFACE,TYPE=ELEMENT,NAME=DIE1
DIE1,SPOS
DIE2,SPOS
DIE3,SNEG
*SURFACE,TYPE=ELEMENT,NAME=BOTTOM
BLANK,SNEG
*RIGID BODY,ELSET=DIE3,REF NODE=399991
*STEP
*DYNAMIC,EXPLICIT
,3.5E-3
*DLOAD,AMP=R1
BLANK,P,-1.0
*FRICTION
0.1,
DIE2,BOTTOM
DIE1,BOTTOM
*FILE OUTPUT,NUM=2, TIMEMARKS=YES
*EL FILE
STH,
*OUTPUT,FIELD
*ELEMENT OUTPUT
STH,
*NODE FILE
U,
*OUTPUT,FIELD
*NODE OUTPUT
U,
*HISTORY OUTPUT,TIME=3.5E-6
*NSET,NSET=NOUT
1,
*NODE HISTORY,NSET=NOUT
U,V
*ELSET,ELSET=EOUT
24,25,26
1-510
*EL HISTORY,ELSET=EOUT
STH,
PEEQ,MISES,S,LE,PE
*ENERGY HISTORY
ALLKE,ALLSE,ALLWK,ALLPD,ALLAE,ALLCD,ALLFD,ALLIE,
ALLVD,ETOTAL,DT
*END STEP
**
** First springback
*STEP
*DYNAMIC,EXPLICIT
,1.E-3
*DLOAD,AMP=R2,OP=NEW
BLANK,P,-1.0
*END STEP
**
** anneal
*STEP
*ANNEAL
*END STEP
**
** gradually slide die 3 into position
*STEP
*DYNAMIC,EXPLICIT
,1.0E-3
*BOUNDARY, AMP=R3A
399991,2,2,-0.04
***DLOAD,OP=NEW
**BLANK, VP, 100.0
*CONTACT PAIR,INTERACTION=FRICT,OP=ADD
DIE3,TOP
*CONTACT PAIR,OP=DELETE
DIE1,BOTTOM
*END STEP
**
** Second upward pressure loading
*STEP
*DYNAMIC,EXPLICIT
,1.0E-3
*ELSET,ELSET=LOAD,GEN
1,26,1
LOAD,P,1.0
1-511
*END STEP
**
** Second springback
*STEP
*DYNAMIC,EXPLICIT
,1.E-3
LOAD,P,1.0
*END STEP
1.3.9 Cup/trough forming

This example illustrates the use of adaptive meshing in forging problems that include large amounts of
shearing at the tool-blank interface; a cup and a trough are formed.
Problem description
Three different geometric models are considered, as shown in Figure 1.3.9-1. Each model consists of a
rigid punch, a rigid die, and a deformable blank. The outer top and bottom edges of the blank are
cambered, which facilitates the flow of material against the tools. The punch and die have semicircular
cross-sections; the punch has a radius of 68.4 mm, and the die has a radius of 67.9 mm. The blank is
modeled as a von Mises elastic, perfectly plastic material with a Young's modulus of 4000 MPa and a
yield stress of 5 MPa. The Poisson's ratio is 0.21; the density is 1.E-4 kg/mm 3.
In each case the punch is moved 61 mm, while the die is fully constrained. The SMOOTH STEP
parameter on the *AMPLITUDE option is used to ramp the punch velocity to a maximum, at which it
remains constant. The SMOOTH STEP specification of the velocity promotes a quasi-static response
to the loading.
Case 1: Axisymmetric model for cup forming

The blank is meshed with CAX4R elements and measures 50 ´ 64.77 mm. The punch and the die are
modeled as TYPE=SEGMENTS analytical rigid surfaces. Symmetry boundary conditions are
prescribed at r=0. The finite element model is shown in Figure 1.3.9-2.
Case 2: Three-dimensional model for trough forming

The blank is meshed with C3D8R elements and measures 50 ´ 64.7 ´ 64.7 mm. The punch and the die
are modeled as TYPE=CYLINDER analytical rigid surfaces. Symmetry boundary conditions are
applied at the x=0 and z=0 planes. The finite element model of the blank is shown in Figure 1.3.9-3.
Case 3: Three-dimensional model for cup forming

The blank is meshed with C3D8R elements. A 90° wedge of the blank with a radius of 50 mm and a
height of 64.7 mm is analyzed. The punch and the die are modeled as TYPE=REVOLUTION
analytical rigid surfaces. Symmetry boundary conditions are applied at the x=0 and y=0 planes. The
1-512
finite element model of the blank is shown in Figure 1.3.9-4.
Adaptive meshing
A single adaptive mesh domain that incorporates the entire blank is used for each model. Symmetry
planes are defined as Lagrangian boundary regions (the default), and contact surfaces are defined as
sliding boundary regions (the default). Since this problem is quasi-static with relatively small amounts
of deformation per increment, the default values for frequency, mesh sweeps, and other adaptive mesh
parameters and controls are sufficient.

Figure 1.3.9-5 through Figure 1.3.9-7show the mesh configuration at the end of the forging simulation
for Cases 1-3. In each case a quality mesh is maintained throughout the simulation. As the blank
flattens out, geometric edges and corners that exist at the beginning of the analysis are broken and
adaptive meshing is allowed across them. The eventual breaking of geometric edges and corners is
essential for this type of problem to minimize element distortion and optimize element aspect ratios.
For comparison purposes Figure 1.3.9-8 shows the deformed mesh for a pure Lagrangian simulation of
Case 1 (the axisymmetric model). The mesh is clearly better when continuous adaptive meshing is
used. Several diamond-shaped elements with extremely poor aspect ratios are formed in the pure
Lagrangian simulation. Adaptive meshing improves the element quality significantly, especially along
the top surface of the cup where solution gradients are highest. Figure 1.3.9-9 and Figure 1.3.9-10show
contours of equivalent plastic strain at the completion of the forging for the adaptive meshing and pure
Lagrangian analyses of Case 1, respectively. Overall plastic strains compare quite closely. Slight
differences exist only along the upper surface, where the pure Lagrangian mesh becomes very distorted
at the end of the simulation. The time histories of the vertical punch force for the adaptive and pure
Lagrangian analyses agree closely for the duration of the forging, as shown in Figure 1.3.9-11.
Input files
ale_cupforming_axi.inp
Case 1.
ale_cupforming_axinodes.inp
ale_cupforming_axielements.inp
ale_cupforming_cyl.inp
Case 2.
ale_cupforming_sph.inp
Case 3.
lag_cupforming_axi.inp
1-513
Lagrangian solution of Case 1.
Figures
Figure 1.3.9-1 Model geometries for each case.
Figure 1.3.9-2 Undeformed mesh for Case 1.
1-514
Figure 1.3.9-5 Deformed mesh for Case 1.
1-515
1-516
Figure 1.3.9-8 Deformed mesh for Case 1 using a pure Lagrangian formulation.
Figure 1.3.9-9 Contours of equivalent plastic strain for Case 1 using adaptive meshing.
1-517
Figure 1.3.9-10 Contours of equivalent plastic strain for Case 1 using a pure Lagrangian fomulation.
Figure 1.3.9-11 Comparison of time histories for the vertical punch force for Case 1.
1-518
Sample listings
1-519
Listing 1.3.9-1
*HEADING
BULK FORMING OF A CUP.
Units - N, mm, sec
*NODE, INPUT=ale_cupforming_axinodes.inp
**
*ELEMENT, TYPE=CAX4R, ELSET=BLANK,
INPUT=ale_cupforming_axielements.inp
**
*SOLID SECTION, ELSET=BLANK, MATERIAL=AL1
**
*MATERIAL,NAME=AL1
4000,0.21
5,0
5,0.22
*DENSITY
1.E-4,
*ELSET, ELSET=BLANK_T, GEN
441, 450, 1
461, 470, 1
480, 480, 1
*ELSET, ELSET=BLANK_B, GEN
350, 440, 10
150, 240, 10
30, 40, 10
31, 39, 1
11, 20, 1
*NODE
900001,5,90,0
900003,0,-25,0
*NSET, NSET=XSYM, GEN
309,408, 11
67, 166, 11
1, 23, 11
551, 562, 11
*BOUNDARY
900003, 1, 6
900001, 1, 6
XSYM, XSYMM
1-520
*NSET, NSET=REFN
900001, 900003
*AMPLITUDE,NAME=AMP,DEFINITION=SMOOTH STEP
0.0, 0.0, .5, 81.333, 1.,81.333
*RESTART, WRITE, NUMBER=30
*SURFACE,TYPE=ELEMENT,NAME=BLANK_B, REGION TYPE=SLIDING
BLANK_B,
*SURFACE, TYPE=SEGMENTS, NAME=DIE
START, 0.0000000E+00, -0.1300000E+02
CIRCL, 0.7500000E+02 , 0.5200001E+02, 7.06897, 54.612727
*SURFACE,TYPE=ELEMENT,NAME=BLANK_T, REGION TYPE=SLIDING
BLANK_T,
*SURFACE, TYPE=SEGMENTS, NAME=PUNCH
START, 80,144
LINE, 64,144
LINE, 64,134
CIRCL, 0,65.7,-4.44445,134
LINE, -1, 65.7
*RIGID BODY, REF NODE=900001, ANALYTICAL SURFACE =PUNCH
*RIGID BODY, REF NODE=900003, ANALYTICAL SURFACE =DIE
*STEP
*DYNAMIC,EXPLICIT
, 1.
*BOUNDARY,TYPE=VELOCITY,AMPLITUDE=AMP
900001, 2,2, -1.00
*CONTACT PAIR, INTERACTION=TOP
BLANK_T, PUNCH
*SURFACE INTERACTION, NAME=TOP
*FRICTION, TAUMAX=4.
0.1,
*CONTACT PAIR, INTERACTION=BOTTOM
BLANK_B, DIE
*SURFACE INTERACTION, NAME=BOTTOM
*FRICTION, TAUMAX=4.
0.1,
*HISTORY OUTPUT, TIME INTERVAL=2.E-3
*NODE HISTORY, NSET=REFN
U,RF
*EL FILE, ELSET=BLANK_T
MISES,PEEQ,
*NODE FILE,NSET=REFN
U,RF
1-521
*ENERGY FILE
*ADAPTIVE MESH, ELSET=BLANK
*END STEP
1.3.10 Forging with sinusoidal dies

This example illustrates the use of adaptive meshing in forging problems that incorporate
geometrically complex dies and involve substantial material flow.
Problem description
Three different geometric models are considered, as shown in Figure 1.3.10-1. Each model consists of
a rigid die and a deformable blank. The cross-sectional shape of the die is sinusoidal with an amplitude
and a period of 5 and 10 mm, respectively. The blank is steel and is modeled as a von Mises
elastic-plastic material with a Young's modulus of 200 GPa, an initial yield stress of 100 MPa, and a
constant hardening slope of 300 MPa. Poisson's ratio is 0.3; the density is 7800 kg/m 3.
In all cases the die is moved downward vertically at a velocity of 2000 mm/sec and is constrained in all
other degrees of freedom. The total die displacement is 7.6 mm for Cases 1 and 2 and 5.6 mm for Case
3. These displacements represent the maximum possible given the refinement and topology of the
initial mesh (if the quality of the mesh is retained for the duration of the analysis). Although each
analysis uses a sinusoidal die, the geometries and flow characteristics of the blank material are quite
different for each problem.

The blank is meshed with CAX4R elements and measures 20 ´ 10 mm. The dies are modeled as
TYPE=SEGMENTS analytical rigid surfaces. The bottom of the blank is constrained in the z-direction,
and symmetry boundary conditions are prescribed at r=0. The initial configuration of the blank and the
die is shown in Figure 1.3.10-2.
Case 2: Three-dimensional model

The blank is meshed with C3D8R elements and measures 20 ´ 10 ´ 10 mm.The dies are modeled as
TYPE=CYLINDER analytical rigid surfaces. The bottom of the blank is constrained in the y-direction,
and symmetry boundary conditions are applied at the x=0 and z=10 planes. The finite element model of
the blank and the die is shown in Figure 1.3.10-3.

The blank is meshed with C3D8R elements and measures 20 ´ 10 ´ 20 mm.The dies are modeled as
TYPE=REVOLUTION analytical rigid surfaces. The bottom of the blank is constrained in the
y-direction, and symmetry boundary conditions are applied at the x=0 and z=10 planes. The finite
element model of the blank and the die is shown in Figure 1.3.10-4. The revolved die is displaced
1-522
upward in the figure from its initial position for clarity.
Adaptive meshing
sliding boundary regions (the default). Because the material flow for each of the geometries is
substantial, the frequency and the intensity of adaptive meshing must be increased to provide an
accurate solution. The value of the FREQUENCY parameter on the *ADAPTIVE MESH option is
reduced from the default of 10 to 5 for all cases. The value of the MESH SWEEPS parameter is
increased from the default of 1 to 3 for all cases.

Figure 1.3.10-5 and Figure 1.3.10-6show the deformed mesh and contours of equivalent plastic strain
at the completion of the forming step for Case 1. Adaptive meshing maintains reasonable element
shapes and aspect ratios. This type of forging problem cannot typically be solved using a pure
Lagrangian formulation. Figure 1.3.10-7shows the deformed mesh for Case 2. A complex, doubly
curved deformation pattern is formed on the free surface as the material spreads under the die. Element
distortion appears to be reasonable. Figure 1.3.10-8and Figure 1.3.10-9 show the deformed mesh and
contours of equivalent plastic strain for Case 3. Although the die is a revolved geometry, the
three-dimensional nature of the blank gives rise to fairly complex strain patterns that are symmetric
with respect to the planes of quarter symmetry.
Input files
ale_sinusoid_forgingaxi.inp
Case 1.
ale_sinusoid_forgingaxisurf.inp
ale_sinusoid_forgingcyl.inp
Case 2.
ale_sinusoid_forgingrev.inp
Case 3.
Figures
Figure 1.3.10-1 Model geometries for each of the three cases.
1-523
Figure 1.3.10-2 Initial configuration for Case 1.
1-524
1-525
Figure 1.3.10-6 Contours of equivalent plastic strain for Case 1.
1-526
Sample listings
1-527
Listing 1.3.10-1
*HEADING
2D AXISYMMETRIC FORGING EXAMPLE
Units - N, m, sec
*NODE
1, 0.00, 0.00
97, 0.02, 0.00
1165, 0.00, 0.01
1261, 0.02, 0.01
10000, 0.01, 0.02
*NGEN, NSET=BOT
1,97,1
*NGEN, NSET=TOP
1165,1261,1
*NFILL,NSET=NALL
BOT, TOP, 12, 97
*NGEN, NSET=CENTER
1,1165,97
*ELEMENT, TYPE=CAX4R
1,1,2,99,98
*ELGEN, ELSET=METAL0
1,95,1,1,12,97,96,
96, 96,97,194,193
*ELGEN, ELSET=METAL1
96,12,97,96
*ELSET,ELSET=METAL
METAL0,METAL1
*ELEMENT, TYPE=MASS, ELSET=PMASS
10000, 10000
*MASS, ELSET=PMASS
0.2,
*ELSET, ELSET=UPPER, GEN
1057,1152,1
*ELSET, ELSET=SIDE, GEN
96,1152,96
*SOLID SECTION, ELSET=METAL, MATERIAL=STEEL
*MATERIAL, NAME=STEEL
*ELASTIC
200.E+9, .3
1-528
*PLASTIC
1.E+8, 0.0
3.1E+9, 10.0
*DENSITY
7800.E+1,
*BOUNDARY
BOT, 2,2
CENTER, 1,1
10000,1,1
10000,6,6
*SURFACE, TYPE=SEGMENTS, NAME=RSURF,
FILLET RADIUS=.001
*INCLUDE, INPUT=ale_sinusoid_forgingaxisurf.inp
*SURFACE,TYPE=ELEMENT, NAME=TARGET,
REGION TYPE=SLIDING
UPPER, S3
SIDE, S2
ANALYTICAL SURFACE =RSURF
*STEP
*DYNAMIC, EXPLICIT
,.00038
*CONTACT PAIR, INTERACTION=INTER
RSURF, TARGET
10000, 2, 2, -20.
*EL HISTORY,ELSET=UPPER
S,LE,LEP,NE,NEP,PEEQ
*ENERGY HISTORY
DT,
*EL FILE
S,LE,LEP,NE,NEP
*NODE FILE
U,RF
*ENERGY FILE
*ADAPTIVE MESH,ELSET=METAL,FREQUENCY=5,
MESH SWEEPS=3
*END STEP
1-529
1.3.11 Forging with multiple complex dies

This example illustrates the use of adaptive meshing in forging problems that use multiple
geometrically complex dies. The problem is based on a benchmark presented at the " FEM-Material
Flow Simulation in the Forging Industry" workshop.
Problem description
The benchmark problem is an axisymmetric forging, but in this example both axisymmetric and
three-dimensional geometric models are considered. Each model is shown in Figure 1.3.11-1. Both
models consist of two rigid dies and a deformable blank. The blank's maximum radial dimension is
895.2 mm, and its thickness is 211.4 mm. The outer edge of the blank is rounded to facilitate the flow
of material through the dies. The blank is modeled as a von Mises elastic-plastic material with a
Young's modulus of 200 GPa, an initial yield stress of 360 MPa, and a constant hardening slope of 30
MPa. The Poisson's ratio is 0.3; the density is 7340 kg/m 3.
Both dies are fully constrained, with the exception of the top die, which is moved 183.4 mm downward
at a constant velocity of 166.65 mm/s.

The blank is meshed with CAX4R elements. A fine discretization is required in the radial direction
because of the geometric complexity of the dies and the large amount of material flow that occurs in
that direction. Symmetry boundary conditions are prescribed at r=0. The dies are modeled as
TYPE=SEGMENTS analytical rigid surfaces. The initial configuration is shown in Figure 1.3.11-2.

The blank is meshed with C3D8R elements. A 90° wedge of the blank is analyzed. The level of mesh
refinement is the same as that used in the axisymmetric model. Symmetry boundary conditions are
applied at the x=0 and z=0 planes. The dies are modeled as TYPE=REVOLUTION analytical rigid
surfaces. The initial configuration of the blank only is shown in Figure 1.3.11-3. Although the tools are
not shown in the figure, they are originally in contact with the blank.
Adaptive meshing
sliding boundary regions (the default). Since this problem is quasi-static with relatively small amounts
of deformation per increment, the defaults for frequency, mesh sweeps, and other adaptive mesh
parameters and controls are sufficient.

Figure 1.3.11-4 and Figure 1.3.11-5show the deformed mesh for the axisymmetric case at an
intermediate stage (t = 0.209 s) and in the final configuration ( t = 0.35 s), respectively. The elements
1-530
remain well shaped throughout the entire simulation, with the exception of the elements at the extreme
radius of the blank, which become very coarse as material flows radially during the last 5% of the top
die's travel. Figure 1.3.11-6 shows contours of equivalent plastic strain at the completion of forming.
Figure 1.3.11-7and Figure 1.3.11-8 show the deformed mesh for the three-dimensional case at t =
0.209 and t = 0.35, respectively. Although the axisymmetric and three-dimensional mesh smoothing
algorithms are not identical, the elements in the three-dimensional model also remain well shaped until
the end of the analysis, when the same behavior that is seen in the two-dimensional model occurs.
Contours of equivalent plastic strain for the three-dimensional model (not shown) are virtually
identical to those shown in Figure 1.3.11-6.
Input files
ale_duckshape_forgingaxi.inp
Case 1.
ale_duckshape_forg_axind.inp
External file referenced by the Case 1 analysis.
ale_duckshape_forg_axiel.inp
ale_duckshape_forg_axiset.inp
ale_duckshape_forg_axirs.inp
ale_duckshape_forgingrev.inp
Case 2.
Reference
· Industrieverband Deutscher Schmieden e.V.(IDS), "Forging of an Axisymmetric Disk,"
FEM-Material Flow Simulation in the Forging Industry, Hagen, Germany, October 1997.
Figures
Figure 1.3.11-1 Axisymmetric and three-dimensional model geometries.
1-531
Figure 1.3.11-2 Initial configuration for the axisymmetric model.
Figure 1.3.11-3 Initial configuration mesh for the three-dimensional model.
1-532
Figure 1.3.11-4 The deformed mesh for the axisymmetric model at an intermediate stage.
Figure 1.3.11-5 The deformed mesh for the axisymmetric model at the end of forming.
Figure 1.3.11-6 Contours of equivalent plastic strain for the axisymmetric model at the end of
forming.
1-533
Figure 1.3.11-7 The deformed mesh for the three-dimensional model at an intermediate stage.
Figure 1.3.11-8 The deformed mesh for the three-dimensional model at the end of forming.
Sample listings
1-534
Listing 1.3.11-1
*HEADING
FORGING WITH DUCK-SHAPED DIE (AXISYMMETRIC)
Units - N, mm, sec
*NODE,INPUT=ale_duckshape_forg_axind.inp
*ELEMENT, TYPE=CAX4R , ELSET=BLANK,
INPUT=ale_duckshape_forg_axiel.inp
*INCLUDE, INPUT=ale_duckshape_forg_axiset.inp
*NSET,NSET=SIDE
1,83,164,245,326,407,488,569,650,731,812
*MATERIAL,NAME=BLANK
*DENSITY
7340.e-9,
*ELASTIC
2.E5, 0.3
*PLASTIC
360., 0.
390., 1.
*SOLID SECTION,ELSET=BLANK,MATERIAL=BLANK
*ELSET,ELSET=OUT,GEN
1,10,1
*NSET,NSET=REF
2000,2001
*SURFACE,TYPE=ELEMENT,NAME=BLANK,
REGION TYPE=SLIDING
BLANK,
*SURFACE, NAME=TOP, TYPE=SEGMENTS
*INCLUDE,INPUT=ale_duckshape_forg_axirs.inp
ANALYTICAL SURFACE =BOT_1
ANALYTICAL SURFACE =TOP
*STEP
*DYNAMIC,EXPLICIT
,0.105
*BOUNDARY
2000,1
2000,3,6
2001,1,6
SIDE,1,1
1-535
2000,2,2,-166.652
BLANK,TOP
BLANK,BOT_1
*EL FILE,ELSET=OUT
PEEQ,MISES
*NODE FILE,NSET=REF
U,
*ADAPTIVE MESH,ELSET=BLANK
*END STEP
1.3.12 Flat rolling: transient and steady-state

This example illustrates the use of adaptive meshing to simulate a rolling process using both transient
and steady-state approaches, as shown in Figure 1.3.12-1. A transient flat rolling simulation is
performed using three different methods: a "pure" Lagrangian approach, an adaptive meshing approach
using a Lagrangian domain, and a mixed Eulerian-Lagrangian adaptive meshing approach in which
material upstream from the roller is drawn from an Eulerian inflow boundary but the downstream end
of the blank is handled in a Lagrangian manner. In addition, a steady-state flat rolling simulation is
performed using an Eulerian adaptive mesh domain as a control volume and defining inflow and
outflow Eulerian boundaries. Solutions using each approach are compared.
Problem description
For each analysis case quarter symmetry is assumed; the model consists of a rigid roller and a
deformable blank. The blank is meshed with C3D8R elements. The roller is modeled as an analytical
rigid surface using the *SURFACE, TYPE=CYLINDER and *RIGID BODY options. The radius of the
cylinder is 175 mm. Symmetry boundary conditions are prescribed on the right (z=0 plane) and bottom
(y=0 plane) faces of the blank.
Coulomb friction with a friction coefficient of 0.3 is assumed between the roller and the plate. All
degrees of freedom are constrained on the roller except rotation about the z-axis, where a constant
angular velocity of 6.28 rad/sec is defined. For each analysis case the blank is given an initial velocity
of 0.3 m/s in the x-direction to initiate contact.
The blank is steel and is modeled as a von Mises elastic-plastic material with isotropic hardening. The
Young's modulus is 150 GPa, and the initial yield stress is 168.2 MPa. The Poisson's ratio is 0.3; the
density is 7800 kg/m3. The *FIXED MASS SCALING option is used to scale the masses of all the
blank elements by a factor of 2750 so that the analysis can be performed more economically. This
scaling factor represents an approximate upper bound on the mass scaling possible for this problem,
above which significant inertial effects would be generated.
1-536
The *STEADY STATE DETECTION option is used to define the criteria for stopping the rolling
analyses based on the achievement of a steady-state condition. The criteria used require the satisfaction
of the steady-state detection norms of equivalent plastic strain, spread, force, and torque within the
default tolerances. The exit plane for each norm is defined as the plane passing through the center of
the roller with the normal to the plane coincident with the rolling direction. The SAMPLING
parameter is set to PLANE BY PLANE for Case 1 through Case 3 for the steady-state detection norms
to be evaluated as each plane of elements passes the exit plane. Case 4 requires that the SAMPLING
parameter is set to UNIFORM since the initial mesh is roughly stationary due to the initial geometry
and the inflow and outflow Eulerian boundaries.
The finite element models used for each analysis case are shown in Figure 1.3.12-2. A description of
each model and the adaptive meshing techniques used follows:
Case 1: Transient simulation--pure Lagrangian approach

The blank is initially rectangular and measures 224 ´ 20 ´ 50 mm. No adaptive meshing is performed.
The analysis is run until steady-state conditions are achieved.
Case 2: Transient simulation--Lagrangian adaptive mesh domain

The finite element model for this case is identical to that used for Case 1, with the exception that a
single adaptive mesh domain that incorporates the entire blank is defined to allow continuous adaptive
meshing. Symmetry planes are defined as Lagrangian surfaces (the default), and the contact surface on
the blank is defined as a sliding surface (the default). The analysis is run until steady-state conditions
are achieved.
Case 3: Transient simulation--mixed Eulerian-Lagrangian

approach
This analysis is performed on a relatively short initial blank measuring 65 ´ 20 ´ 50 mm. Material is
continuously drawn by the action of the roller on the blank through an inflow Eulerian boundary
defined on the upstream end. The blank is meshed with the same number of elements as in Cases 1 and
2 so that similar aspect ratios are obtained as the blank lengthens and steady-state conditions are
achieved.
An adaptive mesh domain is defined that incorporates the entire blank. Because it contains at least one
Eulerian surface, this domain is considered Eulerian for the purpose of setting parameter defaults.
However, the analysis model has both Lagrangian and Eulerian aspects. The amount of material flow
with respect to the mesh will be large at the inflow end and small at the downstream end of the
domain. To account for the Lagrangian motion of the downstream end, the MESHING PREDICTOR
option on the *ADAPTIVE MESH CONTROLS option is changed from the default of PREVIOUS to
CURRENT for this problem. To mesh the inflow end accurately and to perform the analysis
economically, the FREQUENCY parameter is set to 5 and the MESH SWEEPS parameter is set to 5.
As in Case 2, symmetry planes are defined as Lagrangian boundary regions (the default), and the
contact surface on the blank is defined as a sliding boundary region (the default). In addition, an
Eulerian boundary region is defined on the upstream end using the *SURFACE, REGION
1-537
TYPE=EULERIAN option. Adaptive mesh constraints are defined on the Eulerian surface using the
*ADAPTIVE MESH CONSTRAINT option to hold the inflow surface mesh completely fixed while
material is allowed to enter the domain normal to the surface. The *EQUATION option is used to
ensure that the velocity normal to the inflow boundary is uniform across the surface. The velocity of
nodes in the direction tangential to the inflow boundary surface is constrained.
Case 4: Steady-state simulation--Eulerian adaptive mesh domain

This analysis employs a control volume approach in which material is drawn from an inflow Eulerian
boundary and is pushed out through an outflow boundary by the action of the roller. The blank
geometry for this analysis case is defined such that it approximates the shape corresponding to the
steady-state solution: this geometry can be thought of as an "initial guess" to the solution. The blank
initially measures 224 mm in length and 50 mm in width and has a variable thickness such that it
conforms to the shape of the roller. The surface of the blank transverse to the rolling direction is not
adjusted to account for the eventual spreading that will occur in the steady-state solution. Actually, any
reasonable initial geometry will reach a steady state, but geometries that are closer to the steady-state
geometry often allow a solution to be obtained in a shorter period of time.
As in the previous two cases an adaptive mesh domain is defined on the blank, symmetry planes are
defined as Lagrangian surfaces (the default), and the contact surface is defined as a sliding surface (the
default). Inflow and outflow Eulerian surfaces are defined on the ends of the blank using the same
techniques as in Case 3, except that for the outflow boundary adaptive mesh constraints are applied
only normal to the boundary surface and no material constraints are applied tangential to the boundary
surface.
To improve the computational efficiency of the analysis, the frequency of adaptive meshing is
increased to every fifth increment because the Eulerian domain undergoes very little overall
deformation and the material flow speed is much less than the material wave speed. This frequency
will cause the mesh at Eulerian boundaries to drift slightly. However, the amount of drift is extremely
small and does not accumulate. There is no need to increase the mesh sweeps because this domain is
relatively stationary and the default MESHING PREDICTOR setting for Eulerian domains is
PREVIOUS. Very little mesh smoothing is required.

The final deformed configurations of the blank for each of the three transient cases are shown in Figure
1.3.12-3. The transient cases have reached a steady-state solution and have been terminated based on
the criteria defined using the *STEADY STATE DETECTION option. Steady-state conditions are
determined to have been reached when the reaction forces and moments on the roller have stabilized
and the cross-sectional shape and distribution of equivalent plastic strain under the roller become
constant over time. When using the *STEADY STATE DETECTION option, these conditions imply
that the force, moment, spread, and equivalent plastic strain norms have stabilized such that the
changes in the norms over three consecutive sampling intervals have fallen below the user-prescribed
tolerances. See ``Steady-state detection,'' Section 7.7.1 of the ABAQUS/Explicit User's Manual, for a
detailed discussion on the definition of the norms. Contours of equivalent plastic strain for each of the
three transient cases are in good agreement and are shown in the final configuration of each blank in
1-538
Figure 1.3.12-4. Figure 1.3.12-5 shows the initial and final mesh configurations at steady state. With
the exception of Case 3 all analyses were terminated using the default steady-state norm tolerances.
Case 3 required that the force and torque norm tolerances be increased from .005 to .01 due to the
force and torque at the roller being rather noisy.
To compare the results from the transient and steady-state approaches, the steady-state detection norms
are summarized for each case in Table 1.3.12-1. The table shows a comparison of the values of the
steady-state detection norms after the analyses have been terminated. The only significant difference is
in the value of the spread norm for Case 4, which is higher than the others. The spread norm is defined
as the largest of the second principle moments of inertia of the workpiece's cross-section. Since the
spread norm is a cubic function of the lateral deformation of the workpiece, rather small differences in
displacements between the test cases can lead to significant differences in the spread norms.
Time history plots of the steady-state detection norms are also shown. Figure 1.3.12-9 and Figure
1.3.12-10 show time history plots of the steady-state force and torque norms, respectively, for all cases.
The force and torque norms are essentially running averages of the force and moment on the roller and
show good agreement for all four test cases. Figure 1.3.12-7 and Figure 1.3.12-8 show time history
plots of the steady-state equivalent plastic strain and spread norms, respectively, for all cases. The
equivalent plastic strains norms are in good agreement for all cases.
Input files
lag_flatrolling.inp
Case 1.
ale_flatrolling_noeuler.inp
Case 2.
ale_flatrolling_inlet.inp
Case 3.
ale_flatrolling_inletoutlet.inp
Case 4.
Table
Table 1.3.12-1 Comparison of steady-state detection norms.

Formulatio Spread norm Effective Force norm Torque norm
n plastic strain
norm
Case 1 1.349 E-7 .8037 -1.43 E6 3.59 E4
Case 2 1.369 E-7 .8034 -1.43 E6 3.55 E4
Case 3 1.365 E-7 .8018 -1.43 E6 3.61 E4
Case 4 1.485 E-7 .8086 -1.40 E6 3.65 E4
1-539
Figures
Figure 1.3.12-1 Diagram illustrating the four analysis approaches used in this problem.
Figure 1.3.12-2 Initial configurations for each case.
1-540
Figure 1.3.12-3 Deformed mesh for Cases 1-3.
1-541
Figure 1.3.12-4 Contours of equivalent plastic strain for Cases 1-3.
1-542
Figure 1.3.12-5 Deformed mesh for Case 4 (shown with initial mesh for comparison).
1-543
Figure 1.3.12-7 Comparison of equivalent plastic strain norm versus time for all cases.
Figure 1.3.12-8 Comparison of spread norm versus time for all cases.
1-544
Figure 1.3.12-9 Comparison of force norm versus time for all cases.
Figure 1.3.12-10 Comparison of torque norm versus time for all cases.
1-545
Sample listings
1-546
Listing 1.3.12-1
*HEADING
FLAT ROLLING - ADAPTIVE MESH,
EULERIAN INLET NODES
Units - N, m, second
**
*RESTART,W,N=20
*NODE
1, -.0851, 0.00000, 0.00000
253, -.0851, 0.02000, 0.00000
42, -.0200, 0.00000, 0.00000
294, -.0200, 0.02000, 0.00000
2059,-.0851, 0.00000, 0.05000
2100,-.0200, 0.00000, 0.05000
2311,-.0851, 0.02000, 0.05000
2352,-.0200, 0.02000, 0.05000
*NGEN, NSET=BOT1
1,42,1
*NGEN, NSET=TOP1
253,294,1
*NFILL,NSET=FRONT
BOT1, TOP1, 6, 42
*NGEN, NSET=BOT2
2059,2100,1
*NGEN, NSET=TOP2
2311,2352, 1
*NFILL,NSET=BACK
BOT2,TOP2, 6, 42
*NFILL, NSET=BAR
FRONT,BACK, 7, 294
*ELEMENT, TYPE=C3D8R
1, 1, 2, 44, 43,
295, 296, 338, 337
*ELGEN, ELSET=BAR
1,41,1,1,6,42,41,7,294,246
*NSET,NSET=BOT,GEN
1,42,1
295,336,1
589,630,1
883,924,1
1177,1218,1
1-547
1471,1512,1
1765,1806,1
2059,2100,1
*SOLID SECTION,ELSET=BAR,MAT=C15,CONTROLS=SECT
1.,
*SECTION CONTROLS, NAME=SECT,HOURGLASS=STIFFNESS
**
*MATERIAL,NAME=C15
*ELASTIC
1.5E11,.3
*PLASTIC
168.72E06,0
219.33E06,0.1
272.02E06,0.2
308.53E06,0.3
337.37E06,0.4
361.58E06,0.5
382.65E06,0.6
401.42E06,0.7
418.42E06,0.8
434.01E06,0.9
448.45E06,1.0
*DENSITY
7.85E3,
***********************************************
**** ROLL
*NODE,NSET=REF
10000, 0.0409 , 0.185
BAR,1,.30
*NSET,NSET=LEFT,GEN
1,294,1
*BOUNDARY
LEFT,ZSYMM
BOT,YSYMM
1477,1722,1
206,246,1
452,492,1
698,738,1
944,984,1
1190,1230,1
1-548
1436,1476,1
1682,1722,1
*NSET,NSET=EULER,GEN
1,253,42
295,547,42
589,841,42
883,1135,42
1177,1429,42
1471,1723,42
1765,2017,42
2059,2311,42
*NSET,NSET=EULERINT,GEN
337,505,42
631,799,42
925,1093,42
1219,1387,42
1513,1681,42
1807,1975,42
*NSET, NSET=EULERSMALL1, GEN
295,547,42
589,841,42
883,1135,42
1177,1429,42
1471,1723,42
1765,2017,42
2059,2311,42
*NSET, NSET=EULERSMALL2, GEN
43,253,42
337,547,42
631,841,42
925,1135,42
1219,1429,42
1513,1723,42
1807,2017,42
2101,2311,42
*NSET,NSET=EQN,GEN
43,253,42
295,547,42
589,841,42
883,1135,42
1177,1429,42
1471,1723,42
1765,2017,42
1-549
2059,2311,42
*EQUATION
2,
EQN,1,1.,1,1,-1.
*ELSET, ELSET=EULER, GEN
1,1477,246
42,1518,246
83,1559,246
124,1600,246
165,1641,246
206,1682,246
*SURFACE, NAME=SURF1, REGION TYPE=SLIDING
TOP,S5
SIDE,S2
*SURFACE, REGION TYPE=EULERIAN, NAME=EULER1
EULER,S6
*SURFACE,TYPE=CYLINDER,NAME=RIGID,
FILLET RADIUS=.001
0.0409 , 0.185, 0.0, 0.05, 0.185,0.0
0.0409 , 0.185, -0.05
START,0.0,-0.175
CIRCL,-0.175,0.0,0.0,0.0
CIRCL,0.0,0.175,0.0,0.0
*RIGID BODY, REFNODE=10000,
ANALYTICAL SURFACE = RIGID
*STEP
*DYNAMIC,EXPLICIT
,0.50
*STEADY STATE DETECTION,ELSET=BAR,
SAMPLING=PLANE BY PLANE
1.0, 0., 0., .1, 0.0, 0.0
*STEADY STATE CRITERIA
SSPEEQ, , .0409, 0., 0.
SSSPRD, , .0409, 0., 0.
SSTORQ, .01, .0409, 0., 0., 10000, 0., 0., 1.
SSFORC, .01, .0409, 0., 0., 10000, 0., 1., 0.
*BOUNDARY
10000,1,5
10000,6,6,6.2832
*BOUNDARY,TYPE=VELOCITY,REGION TYPE=EULERIAN
EULERSMALL1,3,3,0.0
EULERSMALL2,2,2,0.0
1-550
*FIXED MASS SCALING, FACTOR=2750.

*FRICTION
0.3,
SURF1,RIGID
*FILE OUTPUT, NUMBER INTERVAL=10, TIMEMARKS=YES
*EL FILE, ELSET=TOP
MISES,PEEQ,
*NODE FILE,NSET=REF
U,RF
*OUTPUT, FIELD,NUMBER INTERVAL=10
*ELEMENT OUTPUT
MISES, PEEQ
*NODE OUTPUT
U,
*OUTPUT,HISTORY,TIME INTERVAL=1.E-4
*NODE OUTPUT,NSET=REF
RF2,RM3
*INCREMENTATION OUTPUT
SSPEEQ,SSSPRD,SSTORQ,SSFORC
*ADAPTIVE MESH, ELSET=BAR,FREQUENCY=5,
MESH SWEEPS=5,CONTROLS=ALE
*ADAPTIVE MESH CONSTRAINT
EULER,1,3,0.0
*ADAPTIVE MESH CONTROLS,NAME=ALE,MESHING =CURRENT
*ENDSTEP
1.3.13 Section rolling

This example illustrates the use of adaptive meshing in a transient simulation of section rolling.
Results are compared to a pure Lagrangian simulation.
Problem description
This analysis shows a stage in the rolling of a symmetric I-section. Because of the cross-sectional
shape of the I-section, two planes of symmetry exist and only a quarter of the section needs to be
modeled. The quarter-symmetry model, shown in Figure 1.3.13-1, consists of two rigid rollers and a
blank. Roller 1 has a radius of 747 mm, and roller 2 has a radius of 452 mm. The blank has a length of
775 mm, a web half-width of 176.5 mm, a web half-thickness of 24 mm, and a variable flange
thickness.
The finite element model is shown in Figure 1.3.13-2. The blank is meshed with C3D8R elements.
1-551
Symmetry boundary conditions are applied on the y and z symmetry planes of the blank. The rollers are
modeled as TYPE=REVOLUTION analytical rigid surfaces. Roller 1 has all degrees of freedom
constrained except rotation about the z-axis, where a constant angular velocity of 5 rad/sec is specified.
Roller 2 has all degrees of freedom constrained except rotation about the y-axis. An initial velocity of
5602.5 mm/sec in the negative x-direction is applied to the blank to initiate contact between the blank
and the rollers. This velocity corresponds to the velocity of the rollers at the point of initial
contact.
The *VARIABLE MASS SCALING, TYPE=BELOW MIN option is used to scale the masses of all
the blank elements so that a desired minimum stable time increment is achieved initially and the stable
time increment does not fall below this minimum throughout the analysis. The loading rates and mass
scaling definitions are such that a quasi-static solution is generated.
The blank is steel and is modeled as a von Mises elastic-plastic material with a Young's modulus of
212 GPa, an initial yield stress of 80 MPa, and a constant hardening slope of 258 MPa. Poisson's ratio
is 0.3; the density is 7833 kg/m 3. Coulomb friction with a friction coefficient of 0.3 is assumed
between the rollers and the blank.
Adaptive meshing
Adaptive meshing can improve the solution and mesh quality for section rolling problems that involve
large deformations. A single adaptive mesh domain that incorporates the entire blank is defined.
Symmetry planes are defined as Lagrangian boundary regions (the default), and the contact surface on
the blank is defined as a sliding boundary region (the default). The default values are used for all
adaptive mesh parameters and controls.

Figure 1.3.13-3 shows the deformed configuration of the blank when continuous adaptive meshing is
used. For comparison purposes Figure 1.3.13-4 shows the deformed configuration for a pure
Lagrangian simulation. The mesh at the flange-web interface is distorted in the Lagrangian simulation,
but the mesh remains nicely proportioned in the adaptive mesh analysis. A close-up view of the
deformed configuration of the blank is shown for each analysis in Figure 1.3.13-5 and Figure 1.3.13-6
to highlight the differences in mesh quality. Contours of equivalent plastic strain for each analysis are
shown in Figure 1.3.13-5 and Figure 1.3.13-6. The plastic strain distributions are very similar.
Figure 1.3.13-7 and Figure 1.3.13-8show time history plots for the y-component of reaction force and
the reaction moment about the z-axis, respectively, for roller 1. The results for the adaptive mesh
simulation compare closely to those for the pure Lagrangian simulation.
Input files
ale_rolling_section.inp
Analysis that uses adaptive meshing.
ale_rolling_sectionnode.inp
External file referenced by the adaptive mesh analysis.
1-552
ale_rolling_sectionelem.inp
ale_rolling_sectionnelset.inp
ale_rolling_sectionsurf.inp
lag_rolling_section.inp
Lagrangian analysis.
Figures
Figure 1.3.13-1 Geometry of the quarter-symmetry blank and the rollers.
1-553
Figure 1.3.13-2 Quarter-symmetry finite element model.
Figure 1.3.13-3 Deformed blank for the adaptive mesh simulation.
Figure 1.3.13-4 Deformed blank for the pure Lagrangian simulation.
1-554
Figure 1.3.13-5 Close-up of the deformed blank for the adaptive mesh simulation.
Figure 1.3.13-6 Close-up of the deformed blank for the pure Lagrangian simulation.
1-555
Figure 1.3.13-7 Contours of equivalent plastic strain for the adaptive mesh simulation.
Figure 1.3.13-8 Contours of equivalent plastic strain for the pure Lagrangian simulation.
1-556
Figure 1.3.13-9 Time history of the reaction force in the y-direction at the reference node of Roller 1.
Figure 1.3.13-10 Time history of the reaction moment about the z-axis at the reference node of Roller
1.
1-557
Sample listings
1-558
Listing 1.3.13-1
*HEADING
SECTION ROLLING
Units - N, m, seconds
*RESTART, WRITE,NUMBER=10
*SYSTEM
-0.0949794,0.,0., -0.0949794,1.,0.
-0.0949794,0.,1.
*INCLUDE, INPUT=ale_rolling_sectionnode.inp
*INCLUDE, INPUT=ale_rolling_sectionelem.inp
*INCLUDE, INPUT=ale_rolling_sectionnelset.inp
** SECTION: METAL
*SOLID SECTION, ELSET=METAL, MATERIAL=STEEL
1.,
*DENSITY
7833.,
*ELASTIC
2.12E+11, 0.281
*PLASTIC
8e+07, 0.,
2.35e+08, 0.6
*ELEMENT, TYPE=ROTARYI, ELSET=ROTI
50000,10117
*ROTARY INERTIA, ELSET=ROTI
1.E-4,11.05,1.E-4
** INITIAL CONDITION: VELOCITY
*INITIAL CONDITIONS, TYPE=VELOCITY
ROLNODES, 1, -4.187
ROLNODES, 2, 0.
ROLNODES, 3, 0.
** STEP: STEP-1
**
*SURFACE, TYPE=REVOLUTION, NAME=VROLSURF
-0.1800, 0., 0.6221, -0.1800, 100., 0.6221
START, 0.2, 0.2
LINE, 0.4354, 0.2
LINE, 0.45182, 0.035821
CIRCL, 0.45182,-0.035821, 0.093604, 6.9389E-18
LINE, 0.4354, -0.2
LINE, 0.2, -0.2
1-559
*SURFACE, TYPE=REVOLUTION, NAME=HROLSURF

-0.1800, 0.760, 0., -0.180, 0.760, 100.
START, 0.547, 0.137
LINE, 0.71051, 0.12261
CIRCL, 0.747, 0.082765, 0.707, 0.082765
LINE, 0.747, -0.082765
CIRCL, 0.71051, -0.12261, 0.707, -0.082765
LINE, 0.547, -0.137
ANALYTICAL SURFACE =HROLSURF
ANALYTICAL SURFACE =VROLSURF
*INCLUDE, INPUT=ale_rolling_sectionsurf.inp
*STEP
*DYNAMIC, EXPLICIT
, 0.227
*BOUNDARY, OP=NEW
HROLREF, 1, 1
HROLREF, 2, 2
HROLREF, 3, 3
HROLREF, 4, 4
HROLREF, 5, 5
*BOUNDARY, OP=NEW, TYPE=VELOCITY
HROLREF, 6, 6,-5.
*BOUNDARY, OP=NEW
VROLREF, 1, 1
VROLREF, 2, 2
VROLREF, 3, 3
VROLREF, 4, 4
VROLREF, 6, 6
*BOUNDARY, OP=NEW
YSYM, YSYMM
*BOUNDARY, OP=NEW
ZSYM, ZSYMM
*SURFACE INTERACTION, NAME=PRO-1
*FRICTION
0.3,
** INTERACTION: HOR_ROLL
*CONTACT PAIR, INTERACTION=PRO-1
FLANGSURF, HROLSURF
** INTERACTION: VER_ROLL
*CONTACT PAIR, INTERACTION=PRO-1
WEBSURF, VROLSURF
1-560
*FILE OUTPUT, NUMBER INTERVAL=10, TIME MARKS=NO

*NODE FILE, NSET=HROLREF
RF,
*HISTORY OUTPUT, TIME INTERVAL=0.0026
*ENERGY HISTORY
ALLAE, ALLCD, ALLFD, ALLIE, ALLKE, ALLPD, ALLSE
ALLVD, ALLWK, ETOTAL
*NODE HISTORY, NSET=HROLREF
RF2, RF3, RM2, RM3
*NODE HISTORY, NSET=VROLREF
RF2, RF3, RM3, RM2
** MASS SCALING: WHOLEMODEL 1
** ROLLING
*VARIABLE MASS SCALING,TYPE=BELOW MIN,
FREQUENCY =50,
ELSET=METAL,DT=2.E-5
*ADAPTIVE MESH,ELSET=METAL
*END STEP
1.3.14 Ring rolling

This example illustrates the use of adaptive meshing in a two-dimensional rolling simulation. Results
are compared to those obtained using a pure Lagrangian approach.
Problem description
Ring rolling is a specialized process typically used to manufacture parts with revolved geometries such
as bearings. The three-dimensional rolling setup usually includes a freely mounted, idle roll; a
continuously rotating driver roll; and guide rolls in the rolling plane. Transverse to the rolling plane,
conical rolls are used to stabilize the ring and provide a forming surface in the out-of-plane direction.
In this example a two-dimensional, plane stress idealization is used that ignores the effect of the
conical rolls. A schematic diagram of the ring and the surrounding tools is shown in Figure 1.3.14-1.
The driver roll has a diameter of 680 mm, and the idle and guide rolls have diameters of 102 mm. The
ring has an initial inner diameter of 127.5 mm and a thickness of 178.5 mm. The idle and driver rolls
are arranged vertically and are in contact with the inner and outer surfaces of the ring, respectively.
The driver roll is rotated around its stationary axis, while the idle roll is moved vertically downward at
a specified feed rate. For this simulation the x-y motion of the guide rolls is determined a priori and is
prescribed so that the rolls remain in contact with the ring throughout the analysis but do not exert
appreciable force on it. In practice the guide rolls are usually connected through linkage systems, and
their motion is a function of both force and displacement.
The ring is meshed with CPS4R elements, as shown in Figure 1.3.14-2. The ring is steel and is
1-561
modeled as a von Mises elastic-plastic material with a Young's modulus of 150 GPa, an initial yield
stress of 168.7 MPa, and a constant hardening slope of 884 MPa. The Poisson's ratio is 0.3; the density
is 7800 kg/m3.
The analysis is run so that the ring completes approximately 20 revolutions (16.5 seconds). The rigid
rolls are modeled as TYPE=SEGMENTS analytical rigid surfaces. The driver roll is rotated at a
constant angular velocity of 3.7888 rad/sec about the z-axis, while the idle roll has a constant feed rate
of 4.9334 mm/sec and is free to rotate about the z-axis. All other degrees of freedom for the driver and
idle rolls are constrained. A friction coefficient of 0.5 is defined at the blank-idle roll and blank-drive
roll interfaces. Frictionless contact is used between the ring and guide rolls, and the rotation of the
guide rolls is constrained since the actual guide rolls are free to rotate and exert negligible torque on
the ring.
To obtain an economical solution, the *FIXED MASS SCALING option is used to scale the masses of
all elements in the ring by a factor of 2500. This scaling factor represents a reasonable upper limit on
the mass scaling possible for this problem, above which significant inertial effects would be generated.
Furthermore, since the two-dimensional model does not contain the conical rolls, the ring oscillates
from side to side even under the action of the guide rolls. An artificial viscous pressure of 300 MPa
sec/m is applied on the inner and outer surfaces of the ring to assist the guide rolls in preserving the
circular shape of the ring. The pressure value was chosen by trial and error.
Adaptive meshing
A single adaptive mesh domain that incorporates the ring is defined. Contact surfaces on the ring are
defined as sliding boundary regions (the default). Because of the large number of increments required
to simulate 20 revolutions, the deformation per increment is very small. Therefore, the frequency of
adaptive meshing is changed from the default of 10 to every 50 increments. The cost of adaptive
meshing at this frequency is negligible compared to the underlying analysis cost.

Figure 1.3.14-3 shows the deformed configuration of the ring after completing 20 revolutions with
continuous adaptive meshing. High-quality element shapes and aspect ratios are maintained throughout
the simulation. Figure 1.3.14-4 shows the deformed configuration of the ring when a pure Lagrangian
simulation is performed. The pure Lagrangian mesh is distorted, especially at the inner radius where
elements become skewed and very small in the radial direction.
Figure 1.3.14-5 and Figure 1.3.14-6show time history plots for the y-component of reaction force on
the idle roll and the reaction moment about the z-axis for the driver roll, respectively, for both the
adaptive mesh and pure Lagrangian approaches. Although the final meshes are substantially different,
the roll force and torque match reasonably well.
For both the adaptive and pure Lagrangian solutions the plane stress idealization used here results in
very localized through-thickness straining at the inner and outer radii of the ring. This specific type of
localized straining is unique to plane stress modeling and does not occur in ring rolling processes. It is
also not predicted by a three-dimensional finite element model. If adaptivity is used and refined
meshing is desired to capture strong gradients at the inner and outer extremities, the initially uniform
1-562
mesh can be replaced with a graded mesh. Although not shown here, a graded mesh concentrates
element refinement in areas of strong gradients. The adaptive meshing technique preserves the initial
grading when the SMOOTHING OBJECTIVE=GRADED parameter is used on the *ADAPTIVE
MESH CONTROLS option.
Input files
ale_ringroll_2d.inp
Analysis that uses adaptive meshing.
ale_ringroll_2dnode.inp
ale_ringroll_2delem.inp
guideamp.inp
lag_ringroll_2d.inp
Lagrangian analysis.
Figures
Figure 1.3.14-1 Model geometry for the two-dimensional ring rolling analysis.
1-563
Figure 1.3.14-2 Initial mesh configuration.
Figure 1.3.14-3 Deformed configuration after 20 revolutions using adaptive meshing.
1-564
Figure 1.3.14-4 Deformed configuration after 20 revolutions using a pure Lagrangian approach.
Figure 1.3.14-5 Time history of the reaction force in the y-direction for the idle roll.
1-565
Figure 1.3.14-6 Time history of the reaction moment about the z-axis for the driver roll.
Sample listings
1-566
Listing 1.3.14-1
*HEADING
ROLLING OF A RING IN 2D (PLANE STRESS) WITH
SMALL VISCOUS PRESSURE TO STABILIZE THE RING.
Units - N, m, sec
**
*NODE, INPUT=ale_ringroll_2dnode.inp
*ELEMENT, TYPE=CPS4R, ELSET=RING,
INPUT=ale_ringroll_2delem.inp
*SOLID SECTION, ELSET=RING, MAT=C15, CONTROLS=SECT
0.119,
*SECTION CONTROLS, NAME=SECT,HOURGLASS=STIFFNESS
*MATERIAL,NAME=C15
*ELASTIC
1.5E11,.3
*PLASTIC
168.72E6,0.0
1053.00E6,1.0
*DENSITY
7.85E3,
***************** rigid bodies
*** Driver roll
*BOUNDARY, OP=NEW
1660, 1,, 0.
1660, 2,, 0.
1660, 3,, 0.
**
*BOUNDARY, OP=NEW
1660, 4,, 0.
1660, 5,, 0.
**
*** Idle roll
*BOUNDARY, OP=NEW
1648, 1,, 0.
1648, 3,, 0.
**
*BOUNDARY, OP=NEW
1648, 4,, 0.
1648, 5,, 0.
*ELEMENT, TYPE=ROTARYI, ELSET=IDLEI
9000, 1648
1-567
*ROTARY INERTIA, ELSET=IDLEI

**0., 0., 0.3176634
0., 0., 0.3176634E4
***
*** Left guide roll
*BOUNDARY, OP=NEW
1651, 3,, 0.
1651, 4,, 0.
1651, 6,, 0.
*ELEMENT, TYPE=MASS, ELSET=MASSGUIDE
8001, 1651
*** Right guide roll
*BOUNDARY, OP=NEW
1649, 3,, 0.
1649, 4,, 0.
1649, 6,, 0.
**
*ELEMENT, TYPE=MASS, ELSET=MASSGUIDE
8002, 1649
***
*MASS, ELSET=MASSGUIDE
**61.06563
61.06563E4,
**************
*AMPLITUDE, NAME=IDLEVEL, DEFINITION=SMOOTH STEP
0., 1., 16.3, 1.0, 16.5, 0.0
*AMPLITUDE, NAME=DRVRVEL, DEFINITION=SMOOTH STEP
0., 0., 0.2, 1.0, 16.5, 1.0
*INCLUDE, INPUT=guideamp.inp
*SURFACE,TYPE=ELEMENT, NAME=EXTERIOR,
REGION TYPE=SLIDING
EXTERIOR,
*SURFACE, TYPE=SEGMENTS, NAME=LFTGUIDE
START,-.39192431,.77060147,-.39192431,.87260147
CIRCL,-.49392431,.87260147,-.39192431,.87260147
CIRCL,-.39192431,.97460147,-.39192431,.87260147
CIRCL,-.28992431,.87260147,-.39192431,.87260147
CIRCL,-.39192431,.77060147,-.39192431,.87260147
*SURFACE,TYPE=ELEMENT, NAME=INTERIOR,
REGION TYPE=SLIDING
INTERIOR,
*SURFACE, TYPE=SEGMENTS, NAME=RHTGUIDE
START,.39192431,.77060147,.39192431,.87260147
1-568
CIRCL,.28992431,.87260147,.39192431,.87260147
CIRCL,.39192431,.97460147,.39192431,.87260147
CIRCL,.49392431,.87260147,.39192431,.87260147
CIRCL,.39192431,.77060147,.39192431,.87260147
*SURFACE, TYPE=SEGMENTS, NAME=DRIVER
START, 0., -.680
CIRCL,-0.680, 0., 0., 0.
CIRCL, 0., 0.680, 0., 0.
CIRCL, 0.680, 0., 0., 0.
CIRCL, 0., -0.680, 0.,0.
*SURFACE, TYPE=SEGMENTS, NAME=IDLE
START, 0., 0.8585
CIRCL,-0.102, 0.9605, 0., 0.9605
CIRCL, 0., 1.0625, 0., 0.9605
CIRCL, 0.102, 0.9605, 0., 0.9605
CIRCL, 0., 0.8585, 0., 0.9605
ANALYTICAL SURFACE =IDLE
ANALYTICAL SURFACE =RHTGUIDE
ANALYTICAL SURFACE =LFTGUIDE
ANALYTICAL SURFACE =DRIVER
*STEP
Ring Rolling process in Plane Strain.
Reduce the thickness of the ring by 55.0%
*DYNAMIC, EXPLICIT
, 16.5,
*FIXED MASS SCALING, ELSET=RING,FACTOR=2500.
*********************** rigid surfaces:
********* driver roll
*BOUNDARY, TYPE=VELOCITY, AMP=DRVRVEL
1660, 6, 6, -3.78884
********* idle roll
*BOUNDARY, TYPE=VELOCITY, AMP=IDLEVEL
1648, 2, 2, -4.933417E-3
********* left guide roll
*BOUNDARY,TYPE=DISPLACEMENT,AMP=X1651
1651,1,1,-1.0
*BOUNDARY,TYPE=DISPLACEMENT,AMP=Y1651
1651,2,2,1.0
********* right guide roll
1-569
*BOUNDARY,TYPE=DISPLACEMENT,AMP=X1649
1649,1,1,1.0
*BOUNDARY,TYPE=DISPLACEMENT,AMP=Y1649
1649,2,2,1.0
******** define the contact pairs
*ELSET, ELSET=INTERIOR, GEN
1, 1433, 8
*ELSET, ELSET=EXTERIOR, GEN
8, 1440, 8
*CONTACT PAIR, INTERACTION=FRIC
EXTERIOR, DRIVER
INTERIOR, IDLE
*SURFACE INTERACTION, NAME=FRIC
*FRICTION
0.5,
*CONTACT PAIR
EXTERIOR, LFTGUIDE
EXTERIOR, RHTGUIDE
*DLOAD
INTERIOR, VP4, 3.0e8
EXTERIOR, VP2, 3.0e8
******
*RESTART, WRITE, NUM=10
*NSET, NSET=QRTRPNTS
1, 406, 820, 1234, 9, 414, 828, 1242
*ELSET, ELSET=THRURING, GEN
1, 8, 1
*NSET, NSET=REFNODES
1660, 1648, 1651, 1649
*NODE HISTORY, NSET=REFNODES
U, UR3, RF, RM3
*NODE HISTORY, NSET=QRTRPNTS
U,
*EL HISTORY, ELSET=THRURING
MISES, PEEQ, PRESS, ERV
*ENERGY HISTORY
*FILE OUTPUT, NUMBER INTER=5, TIME MARKS=YES
*NODE FILE, NSET=REFNODES
U, RF
*NODE FILE, NSET=QRTRPNTS
U,
1-570
*EL FILE, ELSET=THRURING

MISES, PEEQ, PRESS, ERV
*ADAPTIVE MESH, ELSET=RING,FREQUENCY=50
*END STEP
1.3.15 Axisymmetric extrusion: transient and steady-state

This example illustrates the use of adaptive meshing in simulations of extrusion processes with three
axisymmetric analysis cases. First, a transient simulation is performed for a backward, flat-nosed die,
extrusion geometry using adaptivity on a Lagrangian mesh domain. Second, a transient simulation is
performed on the analogous forward, square die, extrusion geometry, also using adaptivity on a
Lagrangian mesh domain. Finally, a steady-state simulation is performed for the forward extrusion
geometry using adaptivity on an Eulerian mesh domain.
Problem description
The model configurations for the three analysis cases are shown in Figure 1.3.15-1. Each of the models
is axisymmetric and consists of one or more rigid tools and a deformable blank. The rigid tools are
modeled as TYPE=SEGMENTS analytical rigid surfaces. All contact surfaces are assumed to be
well-lubricated and, thus, are treated as frictionless. The blank is made of aluminum and is modeled as
a von Mises elastic-plastic material with isotropic hardening. The Young's modulus is 38 GPa, and the
initial yield stress is 27 MPa. The Poisson's ratio is 0.33; the density is 2672 kg/m 3.
Case 1: Transient analysis of a backward extrusion

The model geometry consists of a rigid die, a rigid punch, and a blank. The blank is meshed with
CAX4R elements and measures 28 ´ 89 mm. The blank is constrained along its base in the z-direction
and at the axis of symmetry in the r-direction. Radial expansion is prevented by contact between the
blank and the die. The punch and the die are fully constrained, with the exception of the prescribed
vertical motion of the punch. The punch is moved downward 82 mm to form a tube with wall and
endcap thicknesses of 7 mm each. The punch velocity is specified using the SMOOTH STEP
parameter on the *AMPLITUDE option so that the response is essentially quasi-static.
The deformation that occurs in extrusion problems, especially in those that involve flat-nosed die
geometries, is extreme and requires adaptive meshing. Since adaptive meshing in ABAQUS/Explicit
works with the same mesh topology throughout the step, the initial mesh must be chosen such that the
mesh topology will be suitable for the duration of the simulation. A simple meshing technique has
been developed for extrusion problems such as this. In two dimensions it uses a four-sided, mapped
mesh domain that can be created with nearly all finite element mesh preprocessors. The vertices for the
four-sided, mapped mesh are shown in Figure 1.3.15-1 and are denoted A, B, C, and D. Two vertices
are located on either side of the extrusion opening, the third is in the corner of the dead material zone
(the upper right corner of the blank), and the fourth vertex is located in the diagonally opposite corner.
A 10 ´ 60 element mesh using this meshing technique is created for this analysis case and is shown in
Figure 1.3.15-2. The mesh refinement is oriented such that the fine mesh along sides BC and DA will
1-571
move up along the extruded walls as the punch is moved downward.

An adaptive mesh domain is defined that incorporates the entire blank. Because of the extremely large
distortions expected in the backward extrusion simulation, three mesh sweeps, instead of the default
value of one, are specified using the MESH SWEEPS parameter on the *ADAPTIVE MESH option.
The default adaptive meshing frequency of 10 is used. Alternatively, a higher frequency could be
specified to perform one mesh sweep per adaptive mesh increment. However, this method would result
in a higher computational cost because of the increased number of advection sweeps it would require.
A substantial amount of initial mesh smoothing is performed by increasing the value of the INITIAL
MESH SWEEPS parameter on the *ADAPTIVE MESH option to 100. The initially smoothed mesh is
shown in Figure 1.3.15-2. Initial smoothing reduces the distortion of the mapped mesh by rounding out
corners and easing sharp transitions before the analysis is performed; therefore, it allows the best mesh
to be used throughout the analysis.
Case 2: Transient analysis of a forward extrusion

The model geometry consists of a rigid die and a blank. The blank geometry and the mesh are identical
to those described for Case 1, except that the mapped mesh is reversed with respect to the vertical
plane so that the mesh lines are oriented toward the forward extrusion opening. The blank is
constrained at the axis of symmetry in the r-direction. Radial expansion is prevented by contact
between the blank and the die. The die is fully constrained. The blank is pushed up 19 mm by
prescribing a constant velocity of 5 m/sec for the nodes along the bottom of the blank. As the blank is
pushed up, material flows through the die opening to form a solid rod with a 7 mm radius.
Adaptive meshing for Case 2 is defined in a similar manner as for Case 1. The undeformed mesh
configurations, before and after initial mesh smoothing, are shown in Figure 1.3.15-3.
Case 3: Steady-state analysis of a forward extrusion

The model geometry consists of a rigid die, identical to the die used for Case 2, and a blank. The blank
geometry is defined such that it closely approximates the shape corresponding to the steady-state
solution: this geometry can be thought of as an "initial guess" to the solution. As shown in Figure
1.3.15-4, the blank is discretized with a simple graded pattern that is most refined near the die fillet.
No special mesh is required for the steady-state case since minimal mesh motion is expected during the
simulation. The blank is constrained at the axis of symmetry in the r-direction. Radial expansion of the
blank is prevented by contact between it and the die.
An adaptive mesh domain is defined that incorporates the entire blank. Because the Eulerian domain
undergoes very little overall deformation and the material flow speed is much less than the material
wave speed, the frequency of adaptive meshing is changed to 5 from the default value of 1 to improve
the computational efficiency of the analysis.
The outflow boundary is assumed to be traction-free and is located far enough downstream to ensure
that a steady-state solution can be obtained. This boundary is defined using the *SURFACE, REGION
TYPE=EULERIAN option. A multi-point constraint is defined on the outflow boundary to keep the
velocity normal to the boundary uniform. The inflow boundary is defined using the *BOUNDARY,
REGION TYPE=EULERIAN option to prescribe a velocity of 5 m/sec in the vertical direction.
1-572
Adaptive mesh constraints are defined on both the inflow and outflow boundaries to fix the mesh in
the vertical direction using the *ADAPTIVE MESH CONSTRAINT option. This effectively creates a
stationary control volume with respect to the inflow and outflow boundaries through which material
can pass.

The results for each analysis case are described below.
Case 1
The use of the mapped meshing technique along with adaptive meshing allows the backward extrusion
analysis to run to completion, creating the long tube with an endcap. Three plots of the deformed mesh
at various times are shown in Figure 1.3.15-5. These plots clearly show how the quality of the mesh is
preserved for the majority of the simulation. Despite the large amount of deformation involved, the
mesh remains smooth and concentrated in the areas of high strain gradients. Extreme deformation and
thinning at the punch fillet occurs near the end of the analysis. This thinning can be reduced by
increasing the fillet radius of the punch. Corresponding contours of equivalent plastic strain are plotted
in Figure 1.3.15-6. The plastic strains are highest along the inner surface of the tube.
Case 2
Adaptive meshing enables the transient forward extrusion simulation to proceed much further than
would be possible using a pure Lagrangian approach. After pushing the billet 19 mm through the die,
the analysis cannot be continued because the elements become too distorted. Since the billet material is
essentially incompressible and the cross-sectional area of the die opening at the top is 1/16 of the
original cross-sectional area of the billet, a rod measuring approximately 304 mm (three times the
length of the original billet) is formed.
Three plots of the deformed mesh at various times in the transient forward extrusion are shown in
Figure 1.3.15-7. As in the backward extrusion case, the plots show that the quality of the mesh is
preserved for a majority of the simulation. The last deformed shape has been truncated for clarity
because the extruded column becomes very long and thin. Contours of equivalent plastic strain at
similar times are shown in Figure 1.3.15-8. The plastic strain distribution developing in the vertical
column does not reach a steady-state value, even at a height of 304 mm. The steady-state results
reported in the discussion for Case 3 show that a steady-state solution based on the equivalent plastic
strain distribution is not reached until much later. An absolute steady-state solution cannot be reached
until the material on the upstream side of the dead material zone first passes along that zone and
through the die opening. The dead material zone is roughly the shape of a triangle and is located in the
upper right-hand corner of the die.
Case 3
The steady-state solution to the forward extrusion analysis is obtained at an extruded column height of
800 mm, which corresponds to pushing the billet 50 mm through the die. Thus, this analysis runs 2.5
times longer than Case 2.
1-573
Contours of equivalent plastic strain in the middle and at the end of the simulation are shown in Figure
1.3.15-9. Time histories of the equivalent plastic strains on the outer edge of the extruded column at
the outflow boundary and 27.5 mm below the outflow boundary are shown in Figure 1.3.15-10. The
plastic strains at both locations converge to the same value by the end of the simulation, which
indicates that the solution has reached a steady state. The final mesh configuration is shown in Figure
1.3.15-11. The mesh undergoes very little change from the beginning to the end of the analysis because
of the accurate initial guess made for the steady-state domain shape and the ability of the adaptive
meshing capability in ABAQUS/Explicit to retain the original mesh gradation.
As a further check on the accuracy of the steady-state simulation and the conservation properties of
adaptive meshing, a time history of the velocity at the outflow boundary is shown in Figure 1.3.15-12.
The velocity reaches a steady value of approximately 80 m/s, which is consistent with the
incompressible material assumption and the 1/16 ratio of the die opening to the billet size.
Input files
ale_extrusion_back.inp
Case 1.
ale_extrusion_backnode.inp
Node data for Case 1.
ale_extrusion_backelem.inp
Element data for Case 1.
ale_extrusion_forward.inp
Case 2.
ale_extrusion_forwardnode.inp
ale_extrusion_forwardelem.inp
ale_extrusion_eulerian.inp
Case 3.
ale_extrusion_euleriannode.inp
ale_extrusion_eulerianelem.inp
Figures
1-574
Figure 1.3.15-1 Axisymmetric model geometries used in the extrusion analysis.
Figure 1.3.15-2 Undeformed configuration for Case 1, before and after initial smoothing.
1-575
Figure 1.3.15-3 Undeformed configuration for Case 2, before and after initial smoothing.
1-576
Figure 1.3.15-4 Undeformed configuration for Case 3.
1-577
Figure 1.3.15-5 Deformed mesh at various times for Case 1.
1-578
Figure 1.3.15-6 Contours of equivalent plastic strain at various times for Case 1.
1-579
Figure 1.3.15-7 Deformed mesh at various times for Case 2.
1-580
Figure 1.3.15-8 Contours of equivalent plastic strain at various times for Case 2.
1-581
Figure 1.3.15-9 Contours of equivalent plastic strain at an intermediate stage and at the end of the
analysis for Case 3.
1-582
Figure 1.3.15-10 Time history of equivalent plastic strain along the outer edge of the extruded column
for Case 3.
Figure 1.3.15-11 Final deformed mesh for Case 3.
1-583
Figure 1.3.15-12 Time history of material velocity at the outflow boundary for Case 3.
Sample listings
1-584
Listing 1.3.15-1
*HEADING
BACKWARD EXTRUSION,MATERIAL: ALUMINIUM
Units: N, m, second
*RESTART,W,N=10
*NODE, INPUT=ale_extrusion_backnode.inp
**
*ELEMENT, TYPE=CAX4R,ELSET=BLANK,
INPUT=ale_extrusion_backelem.inp
*NSET, NSET=BOT, GENERATE
979, 1001, 1
*NSET, NSET=RIGHT,GENERATE
911, 979, 1
*NSET, NSET=TOP, GENERATE
1, 92, 1
183, 183, 1
274, 274, 1
365, 365, 1
456, 456, 1
547, 547, 1
638, 638, 1
729, 729, 1
820, 820, 1
911, 911, 1
*NSET, NSET=LEFT
91, 182, 273, 364,
455, 546, 637, 728,
819, 910, 1001
**
*SOLID SECTION, ELSET=BLANK, MATERIAL=ALUMINIUM
*MATERIAL, NAME=ALUMINIUM
*ELASTIC
38E9,0.33
*PLASTIC
27E6,0
31E6,0.25
32.5E6,0.5
*DENSITY
2672,
*BOUNDARY
9999,1,1
1-585
9999,6,6
LEFT,XSYMM
BOT,YSYMM
9998,1,6
*NSET,NSET=REF
9999,9998
*ELSET, ELSET=OUT
1,2,3
*SURFACE,NAME=RIGHTS,TYPE=SEGMENTS
START,0.028,-0.001
LINE,0.028,0.2
*SURFACE,TYPE=NODE,NAME=NSURF
NCON,
*SURFACE,NAME=TOPS,TYPE=SEGMENTS,FILLET =0.002
START,0.021,0.2
LINE,0.021,0.089
LINE,-0.0001,0.089
*RIGID BODY,REF NODE= 9998,
ANALYTICAL SURFACE =RIGHTS
*RIGID BODY,REF NODE= 9999,
ANALYTICAL SURFACE =TOPS
*STEP
*DYNAMIC, EXPLICIT
,1.50337E-3
*BOUNDARY,TYPE=VELOCITY, AMP=STEP
9999,2,2,-60.0
*AMPLITUDE, DEFINITION=SMOOTH STEP, NAME=STEP
0.,0.,1.36667e-4,1.,1.36667e-3,1.,1.50337E-3, 0.
*NSET,NSET=NCON
TOP,RIGHT
*CONTACT PAIR,INTERACTION=I1
NSURF,TOPS
NSURF,RIGHTS
*SURFACE INTERACTION,NAME=I1
*FILE OUTPUT, NUMBER=4, TIME MARKS=YES
*NODE FILE, NSET=REF
U,RF
*EL FILE,ELSET=OUT
MISES,PEEQ
*ADAPTIVE MESH,ELSET=BLANK,FREQUENCY=10,
MESH SWEEPS=3, INITIAL MESH SWEEPS=100
*ENDSTEP
1-586
1.3.16 Two-step forming simulation

This example illustrates the use of adaptive meshing in simulations of a two-step, bulk metal forming
process. The problem is based on a benchmark problem presented at the Metal Forming Process
Simulation in Industry conference.
Problem description
The model consists of two sets of rigid forming tools (one set for each forming step) and a deformable
blank. The blank and forming die geometries used in the simulation are shown in Figure 1.3.16-1. The
initial configurations of the blank and the tools for each step are shown in Figure 1.3.16-2 and Figure
1.3.16-4. All forming tools are modeled as discrete rigid bodies and meshed with R3D4 and R3D3
elements. The blank, which is meshed with C3D8R elements, is cylindrical and measures 14.5 ´ 21
mm. A half model is constructed, so symmetry boundary conditions are prescribed at the y=0 plane.
The blank is made of a steel alloy that is assumed to satisfy the Ramberg-Osgood relation for true
stress and logarithmic strain,
² = (¾=K )1=n ;
with a reference stress value (K) of 763 MPa and a work-hardening exponent (n) of 0.245. Isotropic
elasticity is assumed, with a Young's modulus of 211 GPa and a Poisson's ratio of 0.3. An initial yield
stress of 200 MPa is obtained with these data. The stress-strain behavior is defined by piecewise linear
segments matching the Ramberg-Osgood curve up to a total (logarithmic) strain level of 140%, with
von Mises yield and isotropic hardening.
The analysis is conducted in two steps. For the first step the rigid tools consist of a planar punch, a
planar base, and a forming die. The initial configuration for this step is shown in Figure 1.3.16-2. The
base, which is not shown, is placed at the opening of the forming die to prevent material from passing
through the die. The motion of the tools is fully constrained, with the exception of the prescribed
displacement in the z-direction for the punch, which is moved 12.69 mm toward the blank at a constant
velocity of 30 m/sec consistent with a quasi-static response. The deformed configuration of the blank
at the completion of the first step is shown in Figure 1.3.16-3.
In the second step the original punch and die are removed from the model and replaced with a new
punch and die, as shown in Figure 1.3.16-4. The removal of the tools is accomplished by deleting the
contact pairs between them and the blank with the *CONTACT PAIR, OP=DELETE option. Although
not shown in the figure, the base is retained; both it and the new die are fully constrained. The punch is
moved 10.5 mm toward the blank at a constant velocity of 30 m/sec consistent with a quasi-static
response. The deformed configuration of the blank at the completion of the second step is shown in
Figure 1.3.16-5.
Adaptive meshing
1-587
A single adaptive mesh domain that incorporates the entire blank is used for both steps. A Lagrangian
boundary region type (the default) is used to define the constraints on the symmetry plane, and a
sliding boundary region type (the default) is used to define all contact surfaces. The frequency of
adaptive meshing is increased to 5 for this problem since material flows quickly near the end of the
step.

Figure 1.3.16-6 shows the deformed mesh at the completion of forming for an analysis in which a pure
Lagrangian mesh is used. Comparing Figure 1.3.16-5 and Figure 1.3.16-6, the resultant mesh for the
simulation in which adaptive meshing is used is clearly better than that obtained with a pure
Lagrangian mesh.
In Figure 1.3.16-7 through Figure 1.3.16-9path plots of equivalent plastic strain in the blank are shown
using the pure Lagrangian and adaptive mesh domains for locations in the y=0 symmetry plane at an
elevation of z=10 mm. The paths are defined in the positive x-direction (from left to right in Figure
1.3.16-4 to Figure 1.3.16-6). As shown in Figure 1.3.16-7, the results are in good agreement at the end
of the first step. At the end of the second step the path is discontinuous. Two paths are considered: one
that spans the left-hand side and another that spans the right-hand side of the U-shaped cross-section
along the symmetry plane. The left- and right-hand paths are shown in Figure 1.3.16-8and Figure
1.3.16-9, respectively. The solutions from the second step compare qualitatively. Small differences can
be attributed to the increased mesh resolution and reduced mesh distortion for the adaptive mesh
domain.
Input files
ale_forging_steelpart.inp
Analysis with adaptive meshing.
ale_forging_steelpartnode1.inp
ale_forging_steelpartelem1.inp
1-588

ale_forging_steelpartsets.inp
lag_forging_steelpart.inp
Pure Lagrangian analysis.
Reference
· Hermann, M. and A. Ruf, "Forming of a Steel Part," Metal Forming Process Simulation in
Industry, Stuttgart, Germany, September 1994.
Figures
Figure 1.3.16-1 Two-step forging process.
1-589
Figure 1.3.16-2 Initial configuration for the first step.
1-590
Figure 1.3.16-3 Deformed blank at the end of the first step.
Figure 1.3.16-4 Configuration at the beginning of the second step.
Figure 1.3.16-5 Deformed blank at the end of the second step for the adaptive mesh analysis.
1-591
Figure 1.3.16-6 Deformed blank at the end of the second step for the pure Lagrangian analysis.
Figure 1.3.16-7 Path plot of equivalent plastic strain at the end of the first step.
Figure 1.3.16-8 Path plot of equivalent plastic strain along the left side at the end of the second
step.
1-592
Figure 1.3.16-9 Path plot of equivalent plastic strain along the right side at the end of the second
step.
Sample listings
1-593
Listing 1.3.16-1
*HEADING
TWO-STAGE FORGING OF A STEEL PART
Units: N, mm, seconds
*NODE,NSET=MOLD,
INPUT=ale_forging_steelpartnode1.inp
*NODE,NSET=DIE1,
*ELEMENT, TYPE=C3D8R, ELSET=MOLD,
INPUT=ale_forging_steelpartelem1.inp
*NODE,NSET=PUNCH1
2000001,-1.,-1.,29.
200000,-1.,-1.,29.
200001,44.,-1.,29.
200002,44.,20.,29.
200003,-1.,20.,29.
*ELEMENT, TYPE=R3D4, ELSET=PUNCH1
200000,200000,200003,200002,200001
*NODE,NSET=BASE
3000001,-1.,-1.,.000001
300000,-1.,-1.,.000001
300001,44.,-1.,.000001
300002,44.,20.,.000001
300003,-1.,20.,.000001
*ELEMENT, TYPE=R3D4, ELSET=BASE
300000,300000,300001,300002,300003
*ELEMENT, TYPE=R3D4, ELSET=DIE1,
*NODE,NSET=PUNCH2,
*ELEMENT, TYPE=R3D4, ELSET=PUNCH2,
*ELEMENT, TYPE=R3D3, ELSET=PUNCH2,
*NODE,NSET=DIE2,
*ELEMENT, TYPE=R3D4, ELSET=DIE2,
*INCLUDE,INPUT=ale_forging_steelpartsets.inp
*BOUNDARY
1000001,1,6
1-594
2000001,1,2
2000001,4,6
3000001,1,6
4000001,1,6
5000001,1,6
YSYMM,YSYMM
RIG1,YSYMM
RIG2,YSYMM
RIG3,YSYMM
RIG4,YSYMM
*SOLID SECTION,ELSET=MOLD,MATERIAL=C15,
CONTROLS=SECT
*SECTION CONTROLS,NAME=SECT,HOURGLASS=STIFFNESS,
KINEMATICS=ORTHOGONAL
**
*MATERIAL,NAME=C15
*ELASTIC
210000.,0.3
*PLASTIC
200.000, 0.000
246.934, 0.010
434.092, 0.100
514.439, 0.200
568.167, 0.300
609.658, 0.400
643.916, 0.500
673.331, 0.600
699.247, 0.700
722.501, 0.800
743.654, 0.900
763.100, 1.000
781.129, 1.100
797.960, 1.200
813.762, 1.300
828.672, 1.400
*DENSITY
7.85E-9,
*NSET,NSET=P1
2000001,
*NSET,NSET=P2
4000001,
*SURFACE,TYPE=ELEMENT,NAME=PUNCH2
1-595
PUNCH2,SPOS
*SURFACE,TYPE=ELEMENT,NAME=TOP,REGION TYPE=SLIDING
TOP,S2
DIE1,SPOS
DIE2,SPOS
*SURFACE,TYPE=ELEMENT,NAME=SIDE,
REGION TYPE=SLIDING
SIDE2,S5
SIDE4,S4
*SURFACE,TYPE=ELEMENT,NAME=BOTTOM,
REGION TYPE=SLIDING
BOTTOM,S1
*SURFACE,TYPE=ELEMENT,NAME=BASE
BASE,SPOS
*SURFACE,TYPE=ELEMENT,NAME=PUNCH1
PUNCH1,SPOS
*RIGID BODY,ELSET=PUNCH1,REF NODE=2000001
*RIGID BODY,ELSET=BASE,REF NODE=3000001
*RIGID BODY,ELSET=PUNCH2,REF NODE=4000001
*STEP
*DYNAMIC,EXPLICIT
,.000423333
2000001,3,3,-30000.
*CONTACT PAIR,INTERACTION=FRIC
BOTTOM,BASE
BOTTOM,DIE1
SIDE,DIE1
TOP,PUNCH1
**
*FRICTION
0.1,
**
*HISTORY OUTPUT,TIMEINTERVAL=0.0000042
*NODE HISTORY,NSET=P1
U,RF
*ADAPTIVE MESH,ELSET=MOLD,FREQUENCY=5,
CONTROLS=TEST
1-596
*ADAPTIVE MESH CONTROLS, NAME=TEST,

TRANSITION FEATURE ANGLE=0.0
*OUTPUT,FIELD,number=2
*ELEMENT OUTPUT
peeq,mises
*node output
u,
*END STEP
*STEP
*DYNAMIC,EXPLICIT
,.00035
4000001,3,3,-30000.
*CONTACT PAIR,INTERACTION=FRIC,OP=DELETE
BOTTOM,DIE1
SIDE,DIE1
TOP,PUNCH1
**
*CONTACT PAIR,INTERACTION=FRIC
BOTTOM,DIE2
SIDE,DIE2
TOP,PUNCH2
**
*FRICTION
0.1,
**
*HISTORY OUTPUT,TIMEINTERVAL=0.0000035
*NODE HISTORY,NSET=P2
U,RF
*EL HISTORY,ELSET=TOP
PEEQ,
EDT,
*ENERGY HISTORY
DT,
*ADAPTIVE MESH,ELSET=MOLD,FREQUENCY=5
*OUTPUT,FIELD,number=2
*ELEMENT OUTPUT
peeq,mises
*node output
u,
*END STEP
1-597
1.3.17 Upsetting of a cylindrical billet: coupled

temperature-displacement and adiabatic analysis
This example illustrates coupled temperature-displacement analysis in a metal forming application.
The case studied is an extension of the standard test case that is defined in Lippmann (1979); thus,
some verification of the results is available by comparison with the numerical results presented in that
reference. The example is that of a small, circular billet of metal that is reduced in length by 60%. Here
the problem is analyzed as a viscoplastic case, including heating of the billet by plastic work. Such
analysis is often important in manufacturing processes, especially when significant temperature rises
degrade the material. The problem is also analyzed in ABAQUS/Standard using a porous metal
material model. The same problem is used to illustrate mesh rezoning in ABAQUS/Standard in
`Ùpsetting of a cylindrical billet in ABAQUS/Standard: quasi-static analysis with rezoning,'' Section
1.3.1; the same test case is also used in `Ùpsetting of a cylindrical billet in ABAQUS/Explicit,''
Section 1.3.2.
Geometry and model

The specimen is shown in Figure 1.3.17-1: a circular billet, 30 mm long, with a radius of 10 mm,
compressed between flat, rough, rigid dies. All surfaces of the billet are assumed to be fully insulated:
this thermal boundary condition is chosen to maximize the temperature rise.
The finite element model is axisymmetric and includes the top half of the billet only since the middle
surface of the billet is a plane of symmetry. In ABAQUS/Standard elements of type CAX8RT, 8-node
quadrilaterals with reduced integration that allow for fully coupled temperature-displacement analysis,
are used. A regular mesh with six elements in each direction is used, as shown in Figure 1.3.17-1. In
ABAQUS/Explicit the billet is modeled with CAX4RT elements in a 12 ´ 12 mesh.
The contact between the top and the lateral exterior surfaces of the billet and the rigid die is modeled
with the *CONTACT PAIR option. The billet surface is defined by means of the *SURFACE option.
The rigid die is modeled as an analytical rigid surface or as an element-based rigid surface, using the
*RIGID BODY option in conjunction with the *SURFACE option. The mechanical interaction
between the contact surfaces is assumed to be nonintermittent, rough frictional contact in
ABAQUS/Standard. Therefore, two options are used in conjunction with the *SURFACE
INTERACTION property option: the *FRICTION, ROUGH option to enforce a no-slip constraint
between the two surfaces and the *SURFACE BEHAVIOR, NO SEPARATION option to ensure that
separation does not occur once contact has been established. In ABAQUS/Explicit the friction
coefficient between the billet and the rigid die is 1.0.
The problem is also solved in ABAQUS/Standard with the first-order fully coupled
temperature-displacement CAX4T elements in a 12 ´ 12 mesh. Similarly, the problem is solved using
CAX8RT elements and user subroutines UMAT and UMATHT to illustrate the use of these
subroutines.
No mesh convergence studies have been performed, but the comparison with results given in
1-598
Lippmann (1979) suggests that these meshes provide accuracy similar to the best of those analyses.
The ABAQUS/Explicit simulations are performed both with and without adaptive meshing.
Material
The material definition is basically that given in Lippmann (1979), except that the metal is assumed to
be rate dependent. The thermal properties are added, with values that correspond to a typical steel, as
well as the data for the porous metal plasticity model. The material properties are then
Young's modulus: 200 GPa
Poisson's ratio: 0.3
Thermal expansion coefficient: 1.2´10-5 per °C
Initial static yield stress: 700 MPa
Work hardening rate: 300 MPa¡ ¢p
Strain rate dependence: "_pl = D (¾=¾ ± ) ¡ 1 ; D = 40/s, p = 5
Specific heat: 586 J/(kg°C)
Density: 7833 kg/m3
Conductivity: 52 J/(m-s-°C)
Porous material parameters: q 1 = q2 = q3 = 1:0
Initial relative density: 0.95 (f0 = 0.05)
Since the problem definition in ABAQUS/Standard assumes that the dies are completely rough, no
tangential slipping is allowed wherever the metal contacts the die.

The kinematic boundary conditions are symmetry on the axis (nodes at r =0, in node set AXIS, have
ur =0 prescribed), symmetry about z =0 (all nodes at z =0, in node set MIDDLE, have uz =0
prescribed). To avoid overconstraint, the node on the top surface of the billet that lies on the symmetry
axis is not part of the node set AXIS: the radial motion of this node is already constrained by a no-slip
frictional constraint (see ``Common difficulties associated with contact modeling,'' Section 21.10.1 of
the ABAQUS/Standard User's Manual). The rigid body reference node for the rigid surface that defines
the die is constrained to have no rotation or ur -displacement, and its uz -displacement is prescribed to
move 9 mm down the axis at constant velocity. The reaction force at the rigid reference node
corresponds to the total force applied by the die.
The thermal boundary conditions are that all external surfaces are insulated (no heat flux allowed).
This condition is chosen because it is the most extreme case: it must provide the largest temperature
rises possible, since no heat can be removed from the specimen.
DELTMX is the limit on the maximum temperature change allowed to occur in any increment and is
one of the controls for the automatic time incrementation scheme in ABAQUS/Standard. It is set to
100°C, which is a large value and indicates that we are not restricting the time increments because of
accuracy considerations in integrating the heat transfer equations. In fact, the automatic time
incrementation scheme will choose fairly small increments because of the severe nonlinearity present
in the problem and the resultant need for several iterations per increment even with a relatively large
number of increments. The large value is used for DELTMX to obtain a reasonable solution at low
cost.
1-599
In ABAQUS/Explicit the automatic time incrementation scheme is used to ensure numerical stability
and to advance the solution in time. Mass scaling is used to reduce the computational cost of the
analysis.
The AMPLITUDE=RAMP parameter is included because the default amplitude variation for a
transient, coupled temperature-displacement analysis is a step function, but here we want the die to
move down at a constant velocity.
Two versions of the analysis are run: a slow upsetting, where the upsetting occurs in 100 seconds, and
a fast upsetting, where the event takes 0.1 second. Both versions are analyzed with the coupled
temperature-displacement procedure. The fast upsetting is also run in ABAQUS/Standard as an
adiabatic static stress analysis. The time period values are specified on the data line associated with the
*COUPLED TEMPERATURE-DISPLACEMENT procedure, *DYNAMIC
TEMPERATURE-DISPLACEMENT procedure, and the *STATIC procedure options. The adiabatic
stress analysis is performed in the same time frame as the fast upsetting case. In all cases analyzed with
ABAQUS/Standard an initial time increment of 1.5% of the time period is used; that is, 1.5 seconds in
the slow case and 0.0015 second in the fast case. This value is chosen because it will result in a
nominal axial strain of about 1% per increment, and experience suggests that such increment sizes are
generally suitable for cases like this.

The results of the ABAQUS/Standard simulations are discussed first, beginning with the results for the
viscoplastic fully dense material. The results of the slow upsetting are illustrated in Figure 1.3.17-2 to
Figure 1.3.17-4. The results for the fast upsetting coupled temperature-displacement analysis are
illustrated in Figure 1.3.17-5 to Figure 1.3.17-7; those for the adiabatic static stress analysis are shown
in Figure 1.3.17-8 and Figure 1.3.17-9. Figure 1.3.17-2 and Figure 1.3.17-5 show the configuration
that is predicted at 60% upsetting. The configuration for the adiabatic analysis is not shown since it is
almost identical to the fast upsetting coupled case. Both the slow and the fast upsetting cases show the
folding of the top outside surface of the billet onto the die, as well as the severe straining of the middle
of the specimen. The second figure in each series (Figure 1.3.17-3 for the slow case, Figure 1.3.17-6
for the fast case, and Figure 1.3.17-8 for the adiabatic case) shows the equivalent plastic strain in the
billet. Peak strains of around 180% occur in the center of the specimen. The third figure in each series
(Figure 1.3.17-4 for the slow case, Figure 1.3.17-7 for the fast case, and Figure 1.3.17-9 for the
adiabatic case) shows the temperature distributions, which are noticeably different between the slow
and fast upsetting cases. In the slow case there is time for the heat to diffuse (the 60% upsetting takes
place in 100 sec, on a specimen where a typical length is 10 mm), so the temperature distribution at
100 sec is quite uniform, varying only between 180°C and 185°C through the billet. In contrast, the
fast upsetting occurs too quickly for the heat to diffuse. In this case the middle of the top surface of the
specimen remains at 0°C at the end of the event, while the center of the specimen heats up to almost
600°C. There is no significant difference in temperatures between the fast coupled case and the
adiabatic case. In the outer top section of the billet there are differences that are a result of the severe
distortion of the elements in that region and the lack of dissipation of generated heat. The temperature
in the rest of the billet compares well. This example illustrates the advantage of an adiabatic analysis,
since a good representation of the results is obtained in about 60% of the computer time required for
1-600
the fully coupled analysis.

The results of the slow and fast upsetting of the billet modeled with the porous metal plasticity model
are shown in Figure 1.3.17-10 to Figure 1.3.17-15. The deformed configuration is identical to that of
Figure 1.3.17-2 and Figure 1.3.17-5. The extent of growth/closure of the voids in the specimen at the
end of the analysis is shown in Figure 1.3.17-10 and Figure 1.3.17-13. The porous material is almost
fully compacted near the center of the billet because of the compressive nature of the stress field in that
region; on the other hand, the corner element is folded up and stretched out near the outer top portion
of the billet, increasing the void volume fraction to almost 0.1 (or 10%) and indicating that tearing of
the material is likely. The equivalent plastic strain is shown in Figure 1.3.17-11 (slow upsetting) and
Figure 1.3.17-14 (fast upsetting) for the porous material; Figure 1.3.17-12 and Figure 1.3.17-15 show
the temperature distribution for the slow and the fast upsetting of the porous metal. The porous metal
needs less external work to achieve the same deformation compared to a fully dense metal.
Consequently, there is less plastic work being dissipated as heat; hence, the temperature increase is not
as much as that of fully dense metal. This effect is more pronounced in the fast upsetting problem,
where the specimen heats up to only 510°C, compared to about 600°C for fully dense metal.
Figure 1.3.17-16 to Figure 1.3.17-18 show predictions of total upsetting force versus displacement of
the die. In Figure 1.3.17-16 the slow upsetting viscoplastic and porous plasticity results are compared
with several elastic-plastic and rigid-plastic results that were collected by Lippmann (1979) and slow
viscoplastic results obtained by Taylor (1981). There is general agreement between all the rate
independent results, and these correspond to the slow viscoplastic results of the present example and of
those found by Taylor (1981). In Figure 1.3.17-17 rate dependence of the yield stress is investigated.
The fast viscoplastic and porous plasticity results show significantly higher force values throughout the
event than the slow results. This effect can be estimated easily. A nominal strain rate of 6 sec is
maintained throughout the event. With the viscoplastic model that is used, this effect increases the
yield stress by 68%. This factor is very close to the load amplification factor that appears in Figure
1.3.17-17. Figure 1.3.17-18 shows that the force versus displacement prediction of the fast viscoplastic
adiabatic analysis agrees well with the fully coupled results.
Two cases using an element-based rigid surface to model the die are also considered in
ABAQUS/Standard. To define the element-based rigid surface, the elements are assigned to rigid
bodies using *RIGID BODY, ISOTHERMAL=YES. The results agree very well with the case when
the analytical rigid surface is used.
The automatic load incrementation results suggest that overall nominal strain increments of about 2%
per increment were obtained, which is slightly better than what was anticipated in the initial time
increment suggestion. These values are typical for problems of this class and are useful guidelines for
estimating the computational effort required for such cases.
The results obtained with ABAQUS/Explicit compare well with those obtained with
ABAQUS/Standard, as illustrated in Figure 1.3.17-19, which compares the results obtained with
ABAQUS/Explicit (without adaptive meshing) for the total upsetting force versus the displacement of
the die against the same result obtained with ABAQUS/Standard. The agreement between the two
solutions is excellent. Similar agreement is obtained with the results obtained from the
ABAQUS/Explicit simulation using adaptive meshing. The mesh distortion is significantly reduced in
1-601
this case, as illustrated in Figure 1.3.17-20.
Input files

cylbillet_cax4t_slow_dense.inp
Slow upsetting case with 144 CAX4T elements, using the fully dense material.
cylbillet_cax4t_fast_dense.inp
Fast upsetting case with 144 CAX4T elements, using the fully dense material.
cylbillet_cax8rt_slow_dense.inp
Slow upsetting case with CAX8RT elements, using the fully dense material.
cylbillet_cax8rt_rb_s_dense.inp
Slow upsetting case with CAX8RT elements, using the fully dense material and an element-based
rigid surface for the die.
cylbillet_cax8rt_fast_dense.inp
Fast upsetting case with CAX8RT elements, using the fully dense material.
cylbillet_cax8rt_slow_por.inp
Slow upsetting case with CAX8RT elements, using the porous material.
cylbillet_cax8rt_fast_por.inp
Fast upsetting case with CAX8RT elements, using the porous material.
cylbillet_cgax4t_slow_dense.inp
Slow upsetting case with 144 CGAX4T elements, using the fully dense material.
cylbillet_cgax4t_fast_dense.inp
Fast upsetting case with 144 CGAX4T elements, using the fully dense material.
cylbillet_cgax4t_rb_f_dense.inp
Fast upsetting case with 144 CGAX4T elements, using the fully dense material and an
element-based rigid surface for the die.
cylbillet_cgax8rt_slow_dense.inp
Slow upsetting case with CGAX8RT elements, using the fully dense material.
cylbillet_cgax8rt_fast_dense.inp
Fast upsetting case with CGAX8RT elements, using the fully dense material.
cylbillet_c3d10m_adiab_dense.inp
Adiabatic static analysis with fully dense material modeled with C3D10M elements.
1-602
cylbillet_cax6m_adiab_dense.inp
Adiabatic static analysis with fully dense material modeled with CAX6M elements.
cylbillet_cax8r_adiab_dense.inp
Adiabatic static analysis with fully dense material modeled with CAX8R elements.
cylbillet_postoutput.inp
*POST OUTPUT analysis, using the fully dense material.
cylbillet_slow_usr_umat_umatht.inp
Slow upsetting case with the material behavior defined in user subroutines UMAT and UMATHT.
cylbillet_slow_usr_umat_umatht.f
User subroutines UMAT and UMATHT used in cylbillet_slow_usr_umat_umatht.inp.

cylbillet_x_cax4rt_slow.inp
Slow upsetting case with fully dense material modeled with CAX4RT elements and without
adaptive meshing; kinematic mechanical contact.
cylbillet_x_cax4rt_fast.inp
Fast upsetting case with fully dense material modeled with CAX4RT elements and without
adaptive meshing; kinematic mechanical contact.
cylbillet_x_cax4rt_slow_adap.inp
Slow upsetting case with fully dense material modeled with CAX4RT elements and with adaptive
meshing; kinematic mechanical contact.
cylbillet_x_cax4rt_fast_adap.inp
Fast upsetting case with fully dense material modeled with CAX4RT elements and with adaptive
meshing; kinematic mechanical contact.
cylbillet_xp_cax4rt_fast.inp
Fast upsetting case with fully dense material modeled with CAX4RT elements and without
adaptive meshing; penalty mechanical contact.
References
· Lippmann, H., Metal Forming Plasticity, Springer-Verlag, Berlin, 1979.
Involving Contact and Friction," Ph.D. Thesis, U. of Texas at Austin, 1981.
Figures
1-603
Figure 1.3.17-1 Axisymmetric upsetting example: geometry and mesh (element type CAX8RT).
Figure 1.3.17-2 Deformed configuration at 60% upsetting: slow case, coupled

temperature-displacement analysis, ABAQUS/Standard.
Figure 1.3.17-3 Plastic strain at 60% upsetting: slow case, coupled temperature-displacement
analysis, ABAQUS/Standard.
1-604
Figure 1.3.17-4 Temperature at 60% upsetting: slow case, coupled temperature-displacement

analysis, ABAQUS/Standard.
Figure 1.3.17-5 Deformed configuration at 60% upsetting: fast case, coupled

Figure 1.3.17-6 Plastic strain at 60% upsetting: fast case, coupled temperature-displacement analysis,
ABAQUS/Standard.
1-605
Figure 1.3.17-7 Temperature at 60% upsetting: fast case, coupled temperature-displacement analysis,
ABAQUS/Standard.
Figure 1.3.17-8 Plastic strain at 60% upsetting: fast case, adiabatic stress analysis,
ABAQUS/Standard.
Figure 1.3.17-9 Temperature at 60% upsetting: fast case, adiabatic stress analysis,
ABAQUS/Standard.
1-606
Figure 1.3.17-10 Void volume fraction at 60% upsetting: porous material, slow coupled
Figure 1.3.17-11 Plastic strain at 60% upsetting: porous material, slow coupled
Figure 1.3.17-12 Temperature at 60% upsetting: porous material, slow coupled

1-607
Figure 1.3.17-13 Void volume fraction at 60% upsetting: porous material, fast coupled
Figure 1.3.17-14 Plastic strain at 60% upsetting: porous material, fast coupled
Figure 1.3.17-15 Temperature at 60% upsetting: porous material, fast coupled

Figure 1.3.17-16 Force-deflection response for slow cylinder upsetting, ABAQUS/Standard.
1-608
Figure 1.3.17-17 Rate dependence of the force-deflection response, ABAQUS/Standard.
1-609
Figure 1.3.17-18 Force-deflection response: adiabatic versus fully coupled analysis,

ABAQUS/Standard.
1-610
Figure 1.3.17-19 Force-deflection response: ABAQUS/Explicit versus ABAQUS/Standard.
1-611
Figure 1.3.17-20 Deformed configuration of the at 60% upsetting: slow case; without adaptive
meshing, left; with adaptive meshing, right (ABAQUS/Explicit).
Sample listings
1-612
Listing 1.3.17-1
*HEADING
- AXISYMMETRIC UPSETTING PROBLEM
- SLOW CASE
- COUPLED TEMPERATURE-DISPLACEMENT ANALYSIS
- (SLOW UPSETTING)
*NODE,NSET=RSNODE
9999,0.,.015
*NODE
1,
13,.01
1201,0.,.015
1213,.01,.015
*NGEN,NSET=MIDDLE
1,13
*NGEN,NSET=TOP
1201,1213
*NSET,NSET=TOPBND,GENERATE
1201,1212,1
*NFILL
MIDDLE,TOP,12,100
1,1101,100
*ELEMENT,TYPE=CAX8RT,ELSET=METAL
1,1,3,203,201,2,103,202,101
*ELGEN,ELSET=METAL
1,6,2,1,6,200,100
*ELSET,ELSET=ESID,GENERATE
6,506,100
501,506,1
*SURFACE,NAME=ASURF
ESID,S2
ETOP,S3
START,.020,.015
LINE,-.001,.015
ASURF,BSURF
1-613
*MATERIAL,NAME=EL
*ELASTIC
200.E9,.3
*PLASTIC
7.E8,0.00
3.7E9,10.0
*RATE DEPENDENT
40.,5.
*SPECIFIC HEAT
586.,
*DENSITY
7833.,
*CONDUCTIVITY
52.,
*EXPANSION
1.2E-5,
0.9,
*FRICTION,ROUGH
*BOUNDARY
MIDDLE,2
AXIS,1
1.5,100.,5.E-8,5.0
*BOUNDARY
RSNODE,1
RSNODE,6
RSNODE,2,,-.009
*PRINT,CONTACT=YES
*CONTACT OUTPUT,VARIABLE=PRESELECT,SLAVE=ASURF
S,MISES
E,PEEQ
U,RF,NT,RFL
1-614
RF,
*OUTPUT,FIELD
*NODE OUTPUT,NSET=RSNODE
RF,
NT,
*NODE OUTPUT
NT,
*END STEP
1-615
Listing 1.3.17-2
*HEADING
- AXISYMMETRIC UPSETTING PROBLEM
- WITH RIGID SURFACE
- SECTION CONTROLS USED (HOURGLASS=STIFFNESS)
- SLOW UPSETTING
*NODE
1,
13,.01
1201,0.,.015
1213,.01,.015
*NGEN,NSET=MIDDLE
1,13
*NGEN,NSET=TOP
1201,1213
*NFILL
MIDDLE,TOP,12,100
*NSET,NSET=AXIS,GEN
1,1201,100
*ELEMENT,TYPE=cax4rt,ELSET=BILLET
1,1,2,102,101
*ELGEN,ELSET=BILLET
1,12,1,1,12,100,100
*NODE, NSET=NRIGID
2003,0.01,.02
*SOLID SECTION,ELSET=BILLET,
MATERIAL=METAL,CONTROL=B
*ELASTIC
200.E9,.3
*PLASTIC
7.E8,0.00
3.7E9,10.0
*RATE DEPENDENT
40.,5.
*SPECIFIC HEAT
586.,
*DENSITY
7833.,
*CONDUCTIVITY
52.,
1-616
*EXPANSION
1.2E-5,
0.9,
*BOUNDARY
MIDDLE,2
AXIS,1
2003,1
2003,3,6
*SURFACE,TYPE=ELEMENT,NAME=BILLET
TOP,S3
SIDE,S2
*SURFACE,NAME=RIGID,TYPE=SEGMENTS
START, 0.02,.015
LINE, 0.00,.015
ANALYTICAL SURFACE =RIGID
*STEP
,100.0
*FIXED MASS SCALING, ELSET=BILLET,FACTOR=1.E+10
2003,2,,-9.e-5
1101,1112,1
12,1112,100
*SURFACE INTERACTION,NAME=RIG_BILL
*FRICTION
1.0,
*CONTACT PAIR,INTERACTION=RIG_BILL
RIGID,BILLET
*FILE OUTPUT,NUM=2
*EL FILE
PEEQ,MISES
*NODE FILE, NSET=NRIGID
U,RF
*NODE FILE
NT,
*OUTPUT,FIELD,VAR=PRESELECT,NUM=5
*OUTPUT,HISTORY
*NODE OUTPUT,NSET=NRIGID
1-617
U2,RF2
*ENERGY OUTPUT
ALLIE,ALLKE,ALLAE,ALLPD
*END STEP
1.3.18 Unstable static problem: thermal forming of a metal sheet

This example demonstrates the use of automatic techniques to stabilize unstable static problems.
Geometrically nonlinear static problems can become unstable for a variety of reasons. Instability may
occur in contact problems, either because of chattering or because contact intended to prevent rigid
body motions is not established initially. Localized instabilities can also occur; they can be either
geometrical, such as local buckling, or material, such as material softening.
This problem models the thermal forming of a metal sheet; the shape of the die may make it difficult to
place the undeformed sheet exactly in initial contact, in which case the initial rigid body motion
prevention algorithm is useful. Metal forming problems are characterized by relatively simply shaped
parts being deformed by relatively complex-shaped dies. The initial placement of the workpiece on a
die or the initial placement of a second die may not be a trivial geometrical exercise for an engineer
modeling the forming process. ABAQUS accepts initial penetrations in contact pairs and
instantaneously tries to resolve them; as long as the geometry allows for this to happen without
excessive deformation, the misplacement of the workpiece usually does not cause problems. On the
other hand, if the workpiece is initially placed away from the dies, serious problems may arise. Unless
there are enough boundary conditions applied, singular finite element systems of equations result
because one or more of the bodies has free rigid body motions. This typically arises when the
deformation is applied through loads instead of boundary conditions. It is possible to eliminate this
problem by modifying the model, which can be cumbersome for the analyst. Alternatively, the
*CONTACT CONTROLS, APPROACH option allows initial placement of a body apart from others
with smooth load-controlled motion until contact gets established.
This example looks at the thermal forming of an aluminum sheet. The deformation is produced by
applying pressure and gravity loads to push the sheet against a sculptured die. The deformation is
initially elastic. Through heating, the yield stress of the material is lowered until permanent plastic
deformations are produced. Subsequently, the assembly is cooled and the pressure loads are removed,
leaving a formed part with some springback. Although the sheet is initially flat, the geometrical nature
of the die makes it difficult to determine the exact location of the sheet when it is placed on the die.
Therefore, an initial gap between the two bodies is modeled, as shown in Figure 1.3.18-1.
Geometry and model

The model consists of a trapezoidal sheet 10.0 m (394.0 in) long, tapering from 2.0 m (78.75 in) to 3.0
m (118.0 in) wide, and 10.0 mm (0.4 in) thick. The die is a ruled surface controlled by two circles of
radii 13.0 m (517.0 in) and 6.0 m (242.0 in) and dimensions slightly larger than the sheet. The sheet is
initially placed over 0.2 m (7.9 in) apart from the die. The sheet has a longitudinal symmetry boundary
condition, and one node prevents the remaining nodes from experiencing in-plane rigid body motion.
1-618
The die is fixed throughout the analysis. The sheet mesh consists of 640 S4R shell elements, while the
die is represented by 640 R3D4 rigid elements. The material is an aluminum alloy with a flow stress of
1.0 ´ 108 Pa (14.5 ksi) at room temperature. A flow stress of 1.0 ´ 103 Pa (0.15 psi) at 400°C is also
provided, essentially declaring that at the higher temperature the material will flow plastically at any
stress. A Coulomb friction coefficient of 0.1 is used to model the interaction between the sheet and
die.

The analysis consists of three steps. In the first step a gravity load and a pressure load of 1.0 ´ 105 Pa
(14.5 psi) are applied, both pushing the sheet against the die. This step is aided by the automatic
contact approach procedure to prevent unrestrained motion of the sheet. This procedure consists of the
application of viscous pressure along the contact direction (normal to the die), opposing the relative
motion between the sheet and the die. During the first increment of the step ABAQUS applies a very
high amount of damping so that it can judge the magnitude of the external loads being applied, as well
as determine the initial distances between the slave and master surfaces. Based on the results of this
initial attempt a suitable damping coefficient is calculated and the increment is repeated, such that a
smooth approach is produced during the initial part of the step. In this particular case five increments
take place before the first contact point closes; from then on the sheet is pressed against the die. As
soon as contact is established, the relative velocities between the sheet and the die decrease and
become almost zero at the end of the step, which essentially eliminates the damping forces. In addition,
ABAQUS ramps down the damping coefficient to zero from the middle of the step on. This guarantees
that the viscous forces decrease to zero, thus avoiding any discontinuity in the forces at the start of the
next step. The shape and relatively low curvatures of the die are such that the deformation at the end of
the step is elastic (Figure 1.3.18-2). In the second step a two-hour heating (from room temperature to
360°C) and cooling (back to 50°C) cycle is applied to the loaded assembly. As a result of the decrease
in flow stress permanent (plastic) deformation develops, as shown in Figure 1.3.18-3. Finally, in the
third step the pressure load is removed and the springback of the deformed sheet is calculated, as
depicted in Figure 1.3.18-4.
Acknowledgements
HKS would like to thank British Aerospace Airbus, Ltd. for providing the basic data from which this
example was derived.
Input file
unstablestatic_forming.inp
Thermal forming model.
Figures
Figure 1.3.18-1 Initial placement of the sheet apart from the die.
1-619
Figure 1.3.18-2 Elastic deformation after gravity and pressure loading.
Figure 1.3.18-3 Permanent deformation produced by heating.
1-620
Figure 1.3.18-4 Springback.
Sample listings
1-621
Listing 1.3.18-1
*HEADING
THERMAL FORMING OF AN ALUMINUM SHEET
S.I Units, Kg, M, s, N
**
*RESTART, WRITE, FREQUENCY=20, OVERLAY
**
** Rigid body node definitions
*NODE,NSET=RIGID
1, -1.800000, 0.000000, 0.000000
17, 1.800000, 0.000000, 0.000000
901, 0.000000, 0.000000, 13.000000
681, -1.300000, 10.300000, 0.000000
697, 1.300000, 10.300000, 0.000000
902, 0.000000, 10.300000, 6.000000
*NGEN,NSET=END1,LINE=C
1,17,1,901
*NGEN,NSET=END2,LINE=C
681,697,1,902
*NFILL,NSET=RIGID
END1,END2,40,17
**
** Plate node definitions
*NODE,NSET=DEFORM
10001, -1.000000, 10.150000, 0.1725
10017, 1.000000, 10.150000, 0.1725
10681, -1.500000, 0.150000, 0.1725
10697, 1.500000, 0.150000, 0.1725
*NGEN,NSET=SHORT
10001,10017
*NGEN,NSET=LONG
10681,10697
*NFILL,NSET=DEFORM
SHORT,LONG,40,17
**
** Rigid body reference node
*NODE,NSET=FIX
99999, 0., 0., 0.
**
** Plate element definitions
*ELEMENT, TYPE=S4R, ELSET=SHEET
10001, 10001, 10002, 10019, 10018
1-622
*ELGEN,ELSET=SHEET
10001, 16, 1, 1, 40, 17, 16
*ELSET,ELSET=SHORT,GENERATE
10001,10016,1
**
** Rigid surface element definitions
*ELEMENT, TYPE=R3D4, ELSET=CONTACT
1, 1, 2, 19, 18
*ELGEN,ELSET=CONTACT
1, 16, 1, 1, 40, 17, 16
**
** Plate 10mm in thickness
*SHELL SECTION, ELSET=SHEET, MATERIAL=ALUMINUM
0.01, 5
**
*MATERIAL, NAME=ALUMINUM
**
*ELASTIC
7.919714E10, 0.3, 292.
*DENSITY
2.9e3,
*EXPANSION
8.367e-6,
**
*PLASTIC
1.e8,0,293.
1.e3,0,673.
**
*RIGID BODY, ELSET=CONTACT, REF NODE=99999
**
*SURFACE, NAME=M1
CONTACT, SPOS
*SURFACE, NAME=S1
SHEET, SPOS
**
*CONTACT PAIR, INTERACTION=I1
S1, M1
*SURFACE INTERACTION, NAME=I1
*FRICTION, SLIP TOLERANCE=0.02
0.1,
**
** Nodes down the middle of the model on which
** boundary conditions are placed.
1-623
*NSET,NSET=MIDDLE
10009, 10026, 10043, 10060,
10077, 10094, 10111, 10128,
10145, 10162, 10179, 10196,
10213, 10230, 10247, 10264,
10281, 10298, 10315, 10332,
10349, 10366, 10383, 10400,
10417, 10434, 10451, 10468,
10485, 10502, 10519, 10536,
10553, 10570, 10587, 10604,
10621, 10638, 10655, 10672,
10689,
**
**
*BOUNDARY
10009,2,2,0.0
MIDDLE,1,1,0.0
99999, 1,6,0.0
**
** Temperature amplitude card for use
** during forming step
**
*AMPLITUDE, NAME=TEMP_PROF, VALUE=ABSOLUTE
0,293., 3600,633. , 7200,323.
**
** Initial ambient temperature for the plate
**
*INITIAL CONDITIONS, TYPE=TEMPERATURE
DEFORM,293.,293.,293.,293.,293.
**
** ---------------------------------------------
**
** Step 1 - Apply gravity loading together with
** 1 Bar pressure loading on plate to force it
** into the die.
*STATIC
0.1, 1., 1.E-10
**
** pressure + gravity
**
*DLOAD, OP=NEW
SHEET,GRAV,9.81, 0., 0., -1.
1-624
SHEET, P, 1.e5
**
*CONTACT CONTROLS,MASTER=M1, SLAVE=S1,APPROACH
**
*NODE PRINT, FREQ=0
*EL PRINT, FREQ=0
*CONTACT PRINT, FREQ=999
*NODE FILE,NSET=SHORT,FREQ=999
U,
*EL FILE, ELSET=SHORT, FREQ=999
S,E,PE
*NODE OUTPUT
U,
*ELEMENT OUTPUT
S,E,PEEQ
*END STEP
**
** ---------------------------------------------
**
** Step 2 - Now perform the forming step.
** The temperature follows the profile on the
** *AMPLITUDE card
*STATIC
100.,7200.,1e-1,5.e2
**
*TEMPERATURE, OP=NEW, AMPLITUDE=TEMP_PROF
DEFORM,1.,1.,1.,1.,1.
**
*END STEP
**
** --------------------------------------------
**
** Step 3 - Now release 1 Bar pressure
** (but maintain gravity load)
*STATIC
0.1, 1.
**
*DLOAD, OP=NEW
SHEET,GRAV,9.81, 0., 0., -1.
**
1-625
*END STEP
1.4 Fracture mechanics

1.4.1 A plate with a part-through crack: elastic line spring
modeling
The line spring elements in ABAQUS allow inexpensive evaluation of the effects of surface flaws in
shell structures, with sufficient accuracy for use in design studies. The basic concept of these elements
is that they introduce the local solution, dominated by the singularity at the crack tip, into a shell model
of the uncracked geometry. The relative displacements and rotations across the cracked section,
calculated in the line spring elements, are then used to determine the magnitude of the local strain field
and hence the J -integral and stress intensity factor values, as functions of position along the crack
front. This example illustrates the use of these elements and provides some verification of the results
they provide by comparison with a published solution and also by making use of the shell-to-solid
submodeling technique.
Problem description
A large plate with a symmetric, centrally located, semi-elliptic, part-through crack is subjected to edge
tension and bending. The objective is to estimate the Mode I stress intensity factor, KI , as a function
of position along the crack front. Symmetry allows one quarter of the plate to be modeled, as shown in
Figure 1.4.1-1. The 8-node shell element, S8R, and the corresponding 3-node (symmetry plane) line
spring element LS3S are used in the model.
A mesh using LS6 elements is also included. Only half-symmetry is used in this case. When LS6
elements are used, the shell elements on either side of an LS6 element must be numbered such that the
normals to these shell elements point in approximately the same direction.
Geometry and model

For each load case (tension and bending) two plate thicknesses are studied: a "thick" case, for which
the plate thickness is 76.2 mm (3.0 in); and a "thin" case, for which the plate thickness is 19.05 mm
(0.75 in). For both thicknesses the semi-elliptic crack has a maximum depth ( a0 in Figure 1.4.1-2) of
15.24 mm (0.6 in) and a half-length, c, of 76.2 mm (3.0 in). The plate is assumed to be square, with
dimensions 609.6 ´ 609.6 mm (24 ´ 24 in).
The material is assumed to be linear elastic, with Young's modulus 207 GPa (30 ´ 106 lb/in2) and
Poisson's ratio 0.3.
A quarter of the plate is modeled, with symmetry along the edges of the quarter-model at x =0 and
y =0. On the edge containing the flaw (y =0), the symmetry boundary conditions are imposed only on
the unflawed segment of the edge, since they are built into the symmetry plane of the line spring
element being used (LS3S).
The loading consists of a uniform edge tension (per unit length) of 52.44 kN/m (300 lb/in) or a uniform
1-626
edge moment (per unit length) of 1335 N-m/m (300 lb-in/in).

The stress intensity factors for the thick and thin plates are compared with the detailed solutions of
Raju and Newman (1979) and Newman and Raju (1979) in Figure 1.4.1-3 (tension load) and Figure
1.4.1-4 (bending load). These plots show that the present results agree reasonably well with those of
Raju and Newman over the middle portion of the flaw (Á >30°), with better correlation being provided
for the thick case, possibly because the crack is shallower in that geometry. The accuracy is probably
adequate for basic assessment of the criticality of the flaw for design purposes. For values of Á less
than about 30° (that is, at the ends of the flaw), the stress intensity values predicted by the line spring
model lose accuracy. This accuracy loss arises from a combination of the relative coarseness of the
mesh, (especially in this end region where the crack depth varies rapidly), as well as from theoretical
considerations regarding the appropriateness of line spring modeling at the ends of the crack. These
points are discussed in detail by Parks (1981) and Parks et al. (1981).
Shell-to-solid submodeling around the crack tip

An input file for the case a0 =t= 0.2, which uses the shell-to-solid submodeling capability, is included.
This C3D20R element mesh allows the user to study the local crack area using the energy domain
integral formulation for the J -integral. The submodel uses a focused mesh with four rows of elements
p
around the crack tip. A 1= r singularity is utilized at the crack tip, the correct singularity for a linear
elastic solution. Symmetry boundary conditions are imposed on two edges of the submodel mesh,
while results from the global shell analysis are interpolated to two edges by using the submodeling
technique. The global shell mesh gives satisfactory J -integral results; hence, we assume that the
displacements at the submodel boundary are sufficiently accurate to drive the deformation in the
submodel. No attempt has been made to study the effect of making the submodel region larger or
smaller. The submodel is shown superimposed on the global shell model in Figure 1.4.1-5.
The variations of the J -integral values along the crack in the submodeled analysis are compared to the
line spring element analysis in Figure 1.4.1-3 (tension load) and Figure 1.4.1-4 (bending load).
Excellent correlation is seen between the three solutions. A more refined mesh in the shell-to-solid
submodel near the plate surface would be required to obtain J -integral values that more closely match
the reference solution.
Input files
crackplate_ls3s.inp
LS3S elements.
crackplate_surfaceflaw.f
A small program used to create a data file containing the surface flaw depths.
crackplate_ls6_nosym.inp
LS6 elements without symmetry about y = 0.
1-627
crackplate_postoutput.inp
crackplate_submodel.inp
Shell-to-solid submodel.
References
· Newman, J. C., Jr., and I. S. Raju, "Analysis of Surface Cracks in Finite Plates Under Tension or
Bending Loads," NASA Technical Paper 1578, National Aeronautics and Space Administration,
December 1979.
· Parks, D. M., "The Inelastic Line Spring: Estimates of Elastic-Plastic Fracture Mechanics
Parameters for Surface-Cracked Plates and Shells," Journal of Pressure Vessel Technology, vol.
13, pp. 246-254, 1981.
· Parks, D. M., R. R. Lockett, and J. R. Brockenbrough, "Stress Intensity Factors for

Surface-Cracked Plates and Cylindrical Shells Using Line Spring Finite Elements," Advances in
Aerospace Structures and Materials , Edited by S. S. Wang and W. J. Renton, ASME, AD-01, pp.
279-286, 1981.
· Raju, I. S. and J. C. Newman Jr., "Stress Intensity Factors for a Wide Range of Semi-Elliptic
Surface Cracks in Finite Thickness Plates," Journal of Engineering Fracture Mechanics, vol. 11,
pp. 817-829, 1979.
Figures
Figure 1.4.1-1 Quarter model of large plate with center surface crack.
1-628
Figure 1.4.1-2 Schematic surface crack geometry for a semi-elliptical crack.
Figure 1.4.1-3 Stress intensity factor dependence on crack front position: tension loading.
1-629
Figure 1.4.1-4 Stress intensity factor dependence on crack front position: moment loading.
Figure 1.4.1-5 Solid submodel superimposed on shell global model.
1-630
Sample listings
1-631
Listing 1.4.1-1
*HEADING
QUARTER MODEL OF LARGE PLATE WITH CENTER SURFACE
CRACK [S8R]
*NODE
1,0.,0.
97,12.,0.
4801,0.,12.
4897,12.,12.
*NGEN,NSET=ONY
1,4801,100
*NGEN,NSET=FREE
97,4897,100
*NGEN
1,97
101,197
201,297
401,497
601,697
1001,1097
1401,1497,2
2201,2297,2
3001,3097,2
3901,3997,2
4801,4897,2
*NSET,NSET=ONX
27,29,33,37,41,45,49,53,57
61,69,77,87,97
*ELEMENT,TYPE=LS3S,ELSET=LS
1,25,24,23
5,17,15,13
*ELGEN,ELSET=LS
1,4,-2,1
5,4,-4,1
*ELEMENT,TYPE=S8R,ELSET=SHELL
9,1,5,205,201,3,105,203,101
13,17,19,219,217,18,119,218,117
17,25,29,229,225,27,129,227,125
18,29,37,237,229,33,137,233,129
22,61,77,277,261,69,177,269,161
23,77,97,297,277,87,197,287,177
24,201,209,609,601,205,409,605,401
1-632
26,601,609,1409,1401,605,1009,1405,1001
28,217,221,621,617,219,421,619,417
31,617,621,1421,1417,619,1021,1419,1017
34,229,237,637,629,233,437,633,429
38,629,637,1437,1429,633,1037,1433,1029
42,261,277,677,661,269,477,669,461
43,661,677,1477,1461,669,1077,1469,1061
44,277,297,697,677,287,497,687,477
45,677,697,1497,1477,687,1097,1487,1077
46,1401,1417,3017,3001,1409,2217,3009,2201
48,1417,1421,3021,3017,1419,2221,3019,2217
50,1421,1429,3029,3021,1425,2229,3025,2221
58,1477,1497,3097,3077,1487,2297,3087,2277
47,3001,3017,4817,4801,3009,3917,4809,3901
49,3017,3021,4821,4817,3019,3921,4819,3917
51,3021,3029,4829,4821,3025,3929,4825,3921
59,3077,3097,4897,4877,3087,3997,4887,3977
*ELGEN,ELSET=SHELL
9,4,4,1
13,4,2,1
18,4,8,1
24,2,8,1
26,2,8,1
28,3,4,1
31,3,4,1
34,4,8,1
38,4,8,1
46,2,28,6
52,3,16,2
47,2,28,6
53,3,16,2
*ELSET,ELSET=PRINT
9,59,15,16,17
*MATERIAL,NAME=A1
*ELASTIC
30.E6,.3
*SHELL SECTION,ELSET=SHELL,MATERIAL=A1
3.0,3
*SHELL SECTION,ELSET=LS,MATERIAL=A1
3.0,
*SURFACE FLAW,SIDE=POSITIVE,INPUT=CRACK.FLW
** DATA GENERATED FROM PROGRAM 7-1-1-2
*ELSET,ELSET=TOPL
1-633
47,49,51,53,55,57,59
*MPC
QUADRATIC,203,201,205,209
QUADRATIC,207,201,205,209
QUADRATIC,211,209,213,217
QUADRATIC,215,209,213,217
QUADRATIC,218,217,219,221
QUADRATIC,220,217,219,221
QUADRATIC,222,221,223,225
QUADRATIC,224,221,223,225
QUADRATIC,1405,1401,1409,1417
QUADRATIC,1413,1401,1409,1417
QUADRATIC,1423,1421,1425,1429
QUADRATIC,1427,1421,1425,1429
QUADRATIC,1433,1429,1437,1445
QUADRATIC,1441,1429,1437,1445
QUADRATIC,1449,1445,1453,1461
QUADRATIC,1457,1445,1453,1461
*BOUNDARY
ONY,1
ONY,5,6
ONX,2
ONX,4
ONX,6
1,3
*STEP
*STATIC
0.1,1.0
*CLOAD
4801,2,100.
4809,2,400.
4817,2,125.
4819,2,100.
4821,2,75.
4825,2,200.
4829,2,150.
4837,2,400.
4845,2,200.
4853,2,400.
4861,2,200.
4869,2,400.
4877,2,225.
1-634
4887,2,500.
4897,2,125.
*EL PRINT,ELSET=LS
JK,
S,
*EL PRINT,ELSET=PRINT
COORD,
S,
E,
*EL FILE,ELSET=LS
JK,
S,
*EL FILE,ELSET=PRINT
COORD,
S,
E,
*NODE PRINT
U,
*NODE FILE
U,
*END STEP
1-635
Listing 1.4.1-2
CC--- PROGRAM TO GENERATE CRACK DEPTH DATA
C PROGRAM CRACK
C
SUBROUTINE HKSMAIN
C
IMPLICIT REAL*8(A-H,O-Z)
OPEN(UNIT=16,STATUS='NEW',ACCESS='SEQUENTIAL',
1 FORM='FORMATTED',FILE='CRACK.FLW')
C=3.
CC=C*C
N=24
NNODE=N+1
X0=C/DBLE(N)
X=0.
DO 100 I=1,NNODE
IF(I.GE.17) GO TO 1
IF((I/2)*2.EQ.I) GO TO 10
1 CONTINUE
XX=X*X
TMP=.2
Z=TMP*SQRT(CC-XX)
WRITE(6,99) I,Z
WRITE(16,99)I,Z
99 FORMAT(I5,', ',F10.7)
10 CONTINUE
X=X+X0
100 CONTINUE
REWIND 16
STOP
END
1-636
Listing 1.4.1-3
*HEADING
PLATE WITH A PART THROUGH CRACK: LS6 ELEMENTS
*RESTART,WRITE
*NODE
1,0.,0.,0.,0.,0.,1.
97,12.,0.,0.,0.,0.,1.
4801,0.,12.,0.,0.,0.,1.
4897,12.,12.,0.,0.,0.,1.
*NGEN,NSET=ONY
1,4801,100
*NGEN,NSET=FREE
97,4897,100
*NGEN,NSET=TOP
1,97
101,197
201,297
401,497
601,697
1001,1097
1401,1497,2
2201,2297,2
3001,3097,2
3901,3997,2
4801,4897,2
*NSET,NSET=ONX
27,29,33,37,41,45,49,53,57
61,69,77,87,97
*NCOPY, OLD SET=TOP, NEW SET=BOTTOM,
CHANGE NUMBER=10000, REFLECT=LINE
0.,0.,0.,12.,0.,0.
*NSET,NSET=NALL
TOP,BOTTOM
*NSET,NSET=ONY,GENERATE
10001,14801,100
*NSET,NSET=ONX1
10027,10029,10033,10037,10041,10045,10049,10053,
10057,10061,10069,10077,10087,10097
*MPC
TIE,ONX,ONX1
*ELEMENT,TYPE=S8R,ELSET=SHELLT
9,1,5,205,201,3,105,203,101
1-637
13,17,19,219,217,18,119,218,117
17,25,29,229,225,27,129,227,125
18,29,37,237,229,33,137,233,129
22,61,77,277,261,69,177,269,161
23,77,97,297,277,87,197,287,177
24,201,209,609,601,205,409,605,401
26,601,609,1409,1401,605,1009,1405,1001
28,217,221,621,617,219,421,619,417
31,617,621,1421,1417,619,1021,1419,1017
34,229,237,637,629,233,437,633,429
38,629,637,1437,1429,633,1037,1433,1029
42,261,277,677,661,269,477,669,461
43,661,677,1477,1461,669,1077,1469,1061
44,277,297,697,677,287,497,687,477
45,677,697,1497,1477,687,1097,1487,1077
46,1401,1417,3017,3001,1409,2217,3009,2201
48,1417,1421,3021,3017,1419,2221,3019,2217
50,1421,1429,3029,3021,1425,2229,3025,2221
58,1477,1497,3097,3077,1487,2297,3087,2277
47,3001,3017,4817,4801,3009,3917,4809,3901
49,3017,3021,4821,4817,3019,3921,4819,3917
51,3021,3029,4829,4821,3025,3929,4825,3921
59,3077,3097,4897,4877,3087,3997,4887,3977
*ELGEN,ELSET=SHELLT
9,4,4,1
13,4,2,1
18,4,8,1
24,2,8,1
26,2,8,1
28,3,4,1
31,3,4,1
34,4,8,1
38,4,8,1
46,2,28,6
52,3,16,2
47,2,28,6
53,3,16,2
*ELCOPY, ELEMENT SHIFT=10000, OLD SET=SHELLT,
NEW SET=SHELLB,SHIFT NODES=10000
*ELEMENT,TYPE=LS6,ELSET=LS
1,25,24,23,10025,10024,10023
5,17,15,13,10017,10015,10013
*ELGEN,ELSET=LS
1-638
1,4,-2,1
5,4,-4,1
*ELSET,ELSET=SALL
SHELLT,SHELLB,LS
*ELSET,ELSET=SHELLS
SHELLT,SHELLB
*NORMAL
SALL,NALL,0.,0.,1.
*ELSET,ELSET=PRINT
9,59,15,16,17
*MATERIAL,NAME=A1
*ELASTIC
30.E6,.3
*SHELL SECTION,ELSET=SHELLS,MATERIAL=A1
.75,3
*SHELL SECTION,ELSET=LS,MATERIAL=A1
.75,
*SURFACE FLAW,SIDE=POSITIVE
10001, 0.6000000
10003, 0.5979130
10005, 0.5916080
10007, 0.5809475
10009, 0.5656854
10011, 0.5454356
10013, 0.5196152
10015, 0.4873397
10017, 0.4472136
10018, 0.4235269
10019, 0.3968627
10020, 0.3665720
10021, 0.3316625
10022, 0.2904738
10023, 0.2397916
10024, 0.1713914
10025, 0.00
*ELSET,ELSET=TOPL
47,49,51,53,55,57,59
*ELSET,ELSET=BOTL
10047,10049,10051,10053,10055,10057,10059
*MPC
QUADRATIC,203,201,205,209
QUADRATIC,207,201,205,209
QUADRATIC,211,209,213,217
1-639
QUADRATIC,215,209,213,217
QUADRATIC,218,217,219,221
QUADRATIC,220,217,219,221
QUADRATIC,222,221,223,225
QUADRATIC,224,221,223,225
QUADRATIC,1405,1401,1409,1417
QUADRATIC,1413,1401,1409,1417
QUADRATIC,1423,1421,1425,1429
QUADRATIC,1427,1421,1425,1429
QUADRATIC,1433,1429,1437,1445
QUADRATIC,1441,1429,1437,1445
QUADRATIC,1449,1445,1453,1461
QUADRATIC,1457,1445,1453,1461
QUADRATIC,10203,10201,10205,10209
QUADRATIC,10207,10201,10205,10209
QUADRATIC,10211,10209,10213,10217
QUADRATIC,10215,10209,10213,10217
QUADRATIC,10218,10217,10219,10221
QUADRATIC,10220,10217,10219,10221
QUADRATIC,10222,10221,10223,10225
QUADRATIC,10224,10221,10223,10225
QUADRATIC,11405,11401,11409,11417
QUADRATIC,11413,11401,11409,11417
QUADRATIC,11423,11421,11425,11429
QUADRATIC,11427,11421,11425,11429
QUADRATIC,11433,11429,11437,11445
QUADRATIC,11441,11429,11437,11445
QUADRATIC,11449,11445,11453,11461
QUADRATIC,11457,11445,11453,11461
*BOUNDARY
ONY,1
ONY,5,6
4897,3
14897,3
10097,2
10097,4
*STEP
*STATIC
*CLOAD
10025,1,0.,
10025,2,0.
10025,3,0.
10025,4,0.
1-640
10025,5,0.
10025,6,0.
4801,4,-100.
4809,4,-400.
4817,4,-125.
4819,4,-100.
4821,4,-75.
4825,4,-200.
4829,4,-150.
4837,4,-400.
4845,4,-200.
4853,4,-400.
4861,4,-200.
4869,4,-400.
4877,4,-225.
4887,4,-500.
4897,4,-125.
14801,4,100.
14809,4,400.
14817,4,125.
14819,4,100.
14821,4,75.
14825,4,200.
14829,4,150.
14837,4,400.
14845,4,200.
14853,4,400.
14861,4,200.
14869,4,400.
14877,4,225.
14887,4,500.
14897,4,125.
*EL FILE
JK,
*END STEP
1.4.2 Conical crack in a half-space with and without

submodeling
The purpose of this example is to verify that ABAQUS correctly evaluates contour integrals when the
crack extension direction varies along the crack front. For the conical-shaped crack shown in Figure
1.4.2-1, the crack extension direction changes as the crack front is swept around the circle. The
1-641
problem is axisymmetric and can, therefore, also be modeled using axisymmetric elements. The
contour integrals for the three-dimensional model are verified by comparing them to results using
axisymmetric elements.
Full modeling vs. submodeling

The full three-dimensional crack model has an RMS wavefront of 2800. The submodeling capability is
used to obtain accurate results by running two, much smaller, models: first, a global model to get the
displacement solution with moderate accuracy away from the crack tip; and then a submodel to obtain
a more accurate solution, and, hence, more accurate J -integrals, along the crack front. These models
have RMS wavefronts of less than 1700, allowing them to be run on a smaller computer than is
required to run the full three-dimensional model.
Geometry and model

The geometry analyzed is a conical crack in a half-space, as shown in Figure 1.4.2-1. The crack
intersects the free surface at 45° and extends 15 units into the half-space. Pressure loading is applied
on the region of the half-space surface circumscribed by the crack. The full three-dimensional and
axisymmetric meshes are shown in Figure 1.4.2-2 and Figure 1.4.2-3, respectively. The full
three-dimensional model represents one-quarter of the problem, using symmetry about the x-y and y-z
planes, and is composed of 10 sectors parallel to the y-axis. In the region up to a distance of
approximately 10 times the crack length away from the crack, reduced-integration elements (C3D20R
and CAX8R) are used. Beyond this region infinite elements (CIN3D12R and CINAX5R) are used. The
focused mesh surrounding the crack tip in a plane parallel to the y-axis consists of 8 rings of 16
elements. It encompasses half of the crack length and extends the same distance ahead of the crack tip.
p
To obtain the desired 1/ r strain singularity, all the nodes in each crack front node set are tied
together using multi-point constraints; and on element edges radial to the crack front, the midside
nodes are moved to the 1/4 point position. This improves the modeling of the strain field near the crack
tip, which results in more accurate contour integral values.
There are three regions of degenerate elements. At the crack tip collapsed elements are necessary to
provide the desired singularity. The elements at the crack opening and the elements along the y-axis
are collapsed to simplify the meshing.
Figure 1.4.2-4 shows the displaced shape of the mesh near the crack for the three-dimensional case.
Submodel
Submodeling is also used to solve the problem with smaller meshes. The global model represents the
same problem as the full model, but with a coarser mesh. Only one ring of elements is used in the
focused part of the global model mesh compared to eight rings in the full model. For the
three-dimensional global model only five sectors of elements parallel to the 2-axis are used, as
opposed to 10 sectors in the full model. The axisymmetric global model is shown in Figure 1.4.2-5.
The submodels consist of only the focused region of the mesh around the crack tip and contain eight
rings and, in the three-dimensional case, 10 sectors. Quarter-symmetry boundary conditions are applied
to the three-dimensional submodel as well as to the global model. The axisymmetric and
three-dimensional submodels are shown in Figure 1.4.2-6 and Figure 1.4.2-7, respectively. It is
1-642
assumed that the global model's coarse mesh is sufficiently accurate to drive the submodel: if the
global model's displacement field far from the crack tip is accurate, the submodel can obtain accurate
contour integral results.
If two "driven" nodes on opposite faces of the crack share exactly the same location in the submodel,
the submodeling capability is unable to assign the driven nodes to an element uniquely. The two nodes
will behave as if they are tied together across the crack. In this example problem nodes 1033 and
65033 are defined to be approximately 0.01% of their typical element length away from their intended
location along the crack. Moving the nodes this small amount does not affect the results but alleviates
the assignment problem.

J -integral results are shown in Table 1.4.2-1. Neither the axisymmetric nor the three-dimensional
global models used to drive submodels provide useful J values. The submodels driven by these global
models, however, give J -integral results that differ from their related full models by only 2%.
In the three-dimensional models the J -integrals on planes that include element corner nodes differ
slightly from the J -integrals on planes that include only element midside nodes. The difference occurs
because the strain singularity at the crack tip is reproduced on planes of nodes that include corner
nodes, whereas on planes of nodes passing through midside nodes there is no singularity since we use
20-node elements. The use of 27-node elements adjacent to the crack line should eliminate this
problem.
In addition, the stress intensity factors and the T -stresses are calculated. The interaction integral
method, in which the auxiliary plane strain crack-tip fields are employed, is used for their calculations.
Since the crack front is very close to the symmetry axis, more refined meshes should be used to make
the plane strain condition prevail locally around the crack front so that contour-independent results can
be obtained. The calculated values of the stress intensity factors KI , KII , and KIII ; J -integral
(estimated from both stress intensity factors and ABAQUS); and the T -stresses are shown in Table
1.4.2-2, Table 1.4.2-3, Table 1.4.2-4, Table 1.4.2-5, and Table 1.4.2-6, respectively. ABAQUS
automatically outputs the J -integrals based on the stress intensity factors when the latter are evaluated.
These J values are compared with the J values calculated directly by ABAQUS in Table 1.4.2-5, and
good agreement is observed between them.
The sign of KII is different in the three-dimensional model and in the axisymmetric model. This is not
a problem since the sign of KII will depend on the order of the crack front node sets arranged for the
contour integral computation.
Input files
conicalcrack_3dglobal.inp
Three-dimensional global model.
conicalcrack_3dsubmodel.inp
Three-dimensional submodel.
1-643
conicalcrack_full3d.inp
Full three-dimensional model.
conicalcrack_node.inp
Node definitions for conicalcrack_full3d.inp.
conicalcrack_element.inp
Element definitions for conicalcrack_full3d.inp.
conicalcrack_axiglobal.inp
Axisymmetric global model.
conicalcrack_axisubmodel.inp
Axisymmetric submodel.
conicalcrack_fullaxi.inp
Full axisymmetric model.
conicalcrack_3dsubmodel_rms.inp
Three-dimensional submodel with refined meshes.
conicalcrack_full3d_rms.inp
Full three-dimensional model with refined meshes.
conicalcrack_node_rms.inp
Node definitions for conicalcrack_full3d_rms.inp.
conicalcrack_element_rms.inp
Element definitions for conicalcrack_full3d_rms.inp.
conicalcrack_axisubmodel_rms.inp
Axisymmetric submodel with refined meshes.
conicalcrack_fullaxi_rms.inp
Full axisymmetric model with refined meshes.
Tables
Table 1.4.2-1 J -integral estimates (´ 10-7) for conical crack. Contour 1 is omitted from the average
value calculations.
Solution Crack Front Contour Average
Location 1 2 3 4 Value,
5 6 7 8 Contours 2-8
Full Crack tip 1.360 1.331 1.336 1.337
1-644
Axisymmetric 5 6 3 2
1.337 1.337 1.337 1.337 1.3364
5 7 9 9
Corner nodes 1.361 1.331 1.335 1.335
5 4 5 6
1.331 1.334 1.333 1.332 1.3340
Full
5 5 6 5
Three-dimensiona
Midside 1.360 1.331 1.336 1.338
l
nodes 3 6 7 0
1.338 1.339 1.340 1.340 1.3379
7 4 0 6
Submodel Crack tip 1.390 1.361 1.366 1.367
Axisymmetric 3 0 0 1
1.367 1.368 1.368 1.367 1.3665
6 0 2 3
Corner nodes 1.392 1.361 1.365 1.366
1 4 8 1
1.365 1.365 1.363 1.363 1.3646
Submodel
7 3 9 8
Three-dimensiona
Midside 1.390 1.361 1.367 1.368
l
nodes 5 6 0 4
1.369 1.370 1.371 1.369 1.3682
3 1 1 9
Table 1.4.2-2 Stress intensity factor KI estimates for conical crack using refined meshes. Contour 1 is
omitted from the average value calculations.
Solution Crack Contour Average
Front 1 2 3 4 5 Value,
Location Contours
2-5
Full Crack tip 0.487 0.478 0.479 0.479 0.480 0.4794
Axisymmetric 7 3 3 7 1
Corner 0.472 0.477 0.478 0.479 0.479 0.4786
Full
nodes 4 2 4 1 8
Three-dimensiona
Midside 0.492 0.476 0.477 0.477 0.477 0.4770
l
nodes 8 4 1 2 2
Submodel Crack tip 0.537 0.526 0.528 0.528 0.528 0.5280
Axisymmetric 3 9 0 4 8
Corner 0.521 0.526 0.527 0.528 0.528 0.5277
Submodel
nodes 0 2 5 2 9
Three-dimensiona
Midside 0.543 0.525 0.526 0.526 0.526 0.5260
l
nodes 5 3 2 3 3
Table 1.4.2-3 Stress intensity factor KII estimates for conical crack using refined meshes. Contour 1
1-645
is omitted from the average value calculations.

Location Contours
2-5
Full Crack tip -2.078 -2.039 -2.04 -2.041 -2.039 -2.040
Axisymmetric 1
Corner 2.015 2.041 2.045 2.045 2.045 2.044
Full
nodes
Three-dimensiona
Midside 2.108 2.037 2.041 2.041 2.041 2.040
l
nodes
Submodel Crack tip -2.090 -2.050 -2.05 -2.052 -2.051 -2.051
Axisymmetric 3
Corner 2.027 2.053 2.057 2.057 2.057 2.056
Submodel
nodes
Three-dimensiona
Midside 2.121 2.049 2.052 2.053 2.053 2.052
l
nodes
Table 1.4.2-4 Stress intensity factor KIII estimates for conical crack using refined meshes. Contour 1
is omitted from the average value calculations.
Location Contours
2-5
Corner 0.000 0.000 0.000 0.000 0.000 0.0000
Full
nodes 0 0 0 0 0
Three-dimensiona
Midside 0.015 0.014 0.014 0.013 0.013 0.0140
l
nodes 0 0 0 0 0
Corner 0.000 0.000 0.000 0.000 0.000 0.0000
Submodel
nodes 0 0 0 0 0
Three-dimensiona
Midside 0.016 0.016 0.015 0.015 0.014 0.015
l
nodes
Table 1.4.2-5 J -integral estimates (´ 10-7) for conical crack using refined meshes. JK denotes the J
estimated from stress intensity factors; JA denotes the J estimated directly by ABAQUS.
Location (JK) (JK) (JK) (JK) (JK) Contours
1 2 3 4 5 2-5
(JA) (JA) (JA) (JA) (JA)
Full Crack tip 1.382 1.330 1.334 1.333 1.332 1.332
Axisymmetric 1.377 1.331 1.335 1.336 1.336 1.334
Corner 1.300 1.333 1.337 1.338 1.338 1.337
1-646
Full
nodes 1.306 1.331 1.336 1.336 1.336 1.335
Three-dimensiona
Midside 1.422 1.327 1.332 1.332 1.332 1.331
l 1.412 1.330 1.335 1.335 1.336 1.334
nodes
Submodel Crack tip 1.413 1.359 1.363 1.363 1.361 1.362
Axisymmetric 1.407 1.360 1.365 1.365 1.365 1.364
Corner 1.329 1.363 1.367 1.368 1.368 1.367
Submodel
nodes 1.336 1.361 1.366 1.366 1.366 1.365
Three-dimensiona
Midside 1.454 1.357 1.362 1.362 1.362 1.361
l 1.443 1.360 1.365 1.365 1.366 1.364
nodes
Table 1.4.2-6 T -stress estimates for conical crack using refined meshes. Contour 1-2 is omitted from
the average value calculations.
Location Contours 3-5
Full Crack tip -1.161 -0.981 -0.98 -0.981 -0.981 -0.981
Axisymmetric 2
Corner -0.640 -0.971 -0.97 -0.976 -0.976 -0.976
Full
nodes 6
Three-dimensiona
Midside -1.315 -0.973 -0.97 -0.978 -0.979 -0.978
l
nodes 7
Submodel Crack tip -1.182 -0.983 -0.98 -0.984 -0.984 -0.984
Axisymmetric 5
Corner -0.598 -0.974 -0.97 -0.979 -0.979 -0.979
Submodel
nodes 9
Three-dimensiona
Midside -1.366 -0.976 -0.98 -0.981 -0.982 -0.981
l
nodes 0
Figures
Figure 1.4.2-1 Conical crack in a half-space.
1-647
Figure 1.4.2-2 Full three-dimensional mesh.
Figure 1.4.2-3 Full axisymmetric mesh.
1-648
Figure 1.4.2-4 Three-dimensional displaced shape.
Figure 1.4.2-5 Axisymmetric global model for use with submodeling.
1-649
Figure 1.4.2-6 Three-dimensional global model with submodel overlaid.
Figure 1.4.2-7 Axisymmetric global model with submodel overlaid.
1-650
Sample listings
1-651
Listing 1.4.2-1
*HEADING
FULL THREE-DIMENSIONAL CONICAL
CRACK MODEL
*WAVEFRONT MINIMIZATION,SUPPRESS
** INCLUDE NODE DATA
*INCLUDE,INPUT=conicalcrack_node.inp
** INCLUDE ELEMENT DATA
*INCLUDE,INPUT=conicalcrack_element.inp
*ELASTIC
30.E6,0.3
*SOLID SECTION, MATERIAL=STEEL, ELSET=EALL
*MPC
TIE,TIP01,TIP02
TIE,TIP11,TIP12
TIE,TIP21,TIP22
TIE,TIP31,TIP32
TIE,TIP41,TIP42
TIE,TIP51,TIP52
TIE,TIP61,TIP62
TIE,TIP71,TIP72
TIE,TIP81,TIP82
TIE,TIP91,TIP92
TIE,TIP101,TIP102
TIE,TIP111,TIP112
TIE,TIP121,TIP122
TIE,TIP131,TIP132
TIE,TIP141,TIP142
TIE,TIP151,TIP152
TIE,TIP161,TIP162
TIE,TIP171,TIP172
TIE,TIP181,TIP182
TIE,TIP191,TIP192
TIE,TIP201,TIP202
*NSET,NSET=N571,GENERATE
57833,2057833,100000
57865,2057865,100000
65833,2065833,100000
*MPC
1-652
TIE,N572,N571
TIE,N65,N571
1833,2001833,100000
9833,2009833,100000
*MPC
TIE,N18,N98
*MPC
TIE,Y1,Y0
TIE,Y2,Y0
TIE,Y3,Y0
TIE,Y4,Y0
TIE,Y5,Y0
TIE,Y6,Y0
TIE,Y7,Y0
TIE,Y8,Y0
TIE,Y9,Y0
TIE,Y10,Y0
TIE,Y11,Y0
TIE,Y12,Y0
TIE,Y13,Y0
TIE,Y14,Y0
TIE,Y15,Y0
TIE,Y16,Y0
TIE,Y17,Y0
TIE,Y18,Y0
TIE,Y19,Y0
TIE,Y20,Y0
*BOUNDARY
N0,3,3
N20,1,1
Y0,1,1
Y0,3,3
*STEP
APPLY PRESSURE LOAD
*STATIC
1.0,1.0
*DLOAD
TOP9,P4,10.
*CONTOUR INTEGRAL,CONTOURS=8,OUTPUT=BOTH
T0,0.70710678,-0.70710678,0.
T1,0.70492701,-0.70710678,0.055478959
1-653
T2,0.69840112,-0.70710678,0.11061587
T3,0.68756936,-0.70710678,0.1650708
T4,0.67249851,-0.70710678,0.21850801
T5,0.65328148,-0.70710678,0.27059805
T6,0.63003676,-0.70710678,0.32101976
T7,0.60290764,-0.70710678,0.36946228
T8,0.5720614,-0.70710678,0.41562694
T9,0.53768821,-0.70710678,0.45922912
T10,0.5,-0.70710678,0.5
T11,0.45922912,-0.70710678,0.53768821
T12,0.41562694,-0.70710678,0.5720614
T13,0.36946228,-0.70710678,0.60290764
T14,0.32101976,-0.70710678,0.63003676
T15,0.27059805,-0.70710678,0.65328148
T16,0.21850801,-0.70710678,0.67249851
T17,0.1650708,-0.70710678,0.68756936
T18,0.11061587,-0.70710678,0.69840112
T19,0.055478959,-0.70710678,0.70492701
T20,0.,-0.70710678,0.70710678
*CONTOUR INTEGRAL,CONTOURS=8,OUTPUT=BOTH,
TYPE=K FACTORS
T0,0.70710678,-0.70710678,0.
T1,0.70492701,-0.70710678,0.055478959
T2,0.69840112,-0.70710678,0.11061587
T3,0.68756936,-0.70710678,0.1650708
T4,0.67249851,-0.70710678,0.21850801
T5,0.65328148,-0.70710678,0.27059805
T6,0.63003676,-0.70710678,0.32101976
T7,0.60290764,-0.70710678,0.36946228
T8,0.5720614,-0.70710678,0.41562694
T9,0.53768821,-0.70710678,0.45922912
T10,0.5,-0.70710678,0.5
T11,0.45922912,-0.70710678,0.53768821
T12,0.41562694,-0.70710678,0.5720614
T13,0.36946228,-0.70710678,0.60290764
T14,0.32101976,-0.70710678,0.63003676
T15,0.27059805,-0.70710678,0.65328148
T16,0.21850801,-0.70710678,0.67249851
T17,0.1650708,-0.70710678,0.68756936
T18,0.11061587,-0.70710678,0.69840112
T19,0.055478959,-0.70710678,0.70492701
T20,0.,-0.70710678,0.70710678
1-654
TYPE=T-STRESS
T0,0.70710678,-0.70710678,0.
T1,0.70492701,-0.70710678,0.055478959
T2,0.69840112,-0.70710678,0.11061587
T3,0.68756936,-0.70710678,0.1650708
T4,0.67249851,-0.70710678,0.21850801
T5,0.65328148,-0.70710678,0.27059805
T6,0.63003676,-0.70710678,0.32101976
T7,0.60290764,-0.70710678,0.36946228
T8,0.5720614,-0.70710678,0.41562694
T9,0.53768821,-0.70710678,0.45922912
T10,0.5,-0.70710678,0.5
T11,0.45922912,-0.70710678,0.53768821
T12,0.41562694,-0.70710678,0.5720614
T13,0.36946228,-0.70710678,0.60290764
T14,0.32101976,-0.70710678,0.63003676
T15,0.27059805,-0.70710678,0.65328148
T16,0.21850801,-0.70710678,0.67249851
T17,0.1650708,-0.70710678,0.68756936
T18,0.11061587,-0.70710678,0.69840112
T19,0.055478959,-0.70710678,0.70492701
T20,0.,-0.70710678,0.70710678
*EL PRINT
S,
*NODE PRINT
U,RF
*ENDSTEP
1-655
Listing 1.4.2-2
*HEADING
AXISYMMETRIC CONICAL CRACK MODEL--GLOBAL WITH 1 RING
*SYSTEM
10.,0.,0.
*NODE,SYSTEM=C
1833,0.,0.,0.
9833,0.,0.,0.
57833,0.,0.,0.
57865,0.,0.,0.
65833,0.,0.,0.
1001,15.,-45.,0.
65001,15.,-45.,0.
1033,8.,-45.,0.
65033,8.,-45.,0.
33033,25.,-45.,0.
41033,25.,-18.,0.
57033,10.,-18.,0.
9033,10.,-72.,0.
25033,25.,-72.,0.
41065,25.,0.,0.
57065,10.,0.,0.
*SYSTEM
0.,0.,0.
*NODE,SYSTEM=C
9865,0.,0.,0.
9065,15.,-90.,0.
25065,30.,-90.,0.
25865,170.,-90.,0.
25965,340.,-90.,0.
25833,170.,-60.,0.
25933,340.,-60.,0.
33833,170.,-45.,0.
33933,340.,-45.,0.
41833,170.,-30.,0.
41933,340.,-30.,0.
41865,170.,0.,0.
41965,340.,0.,0.
**
**CRACK TIP REGION
*NGEN,NSET=TIP
1-656
1001,65001,1000
*NGEN,NSET=OUTER1
1033,9033,1000
*NGEN,NSET=OUTER2
9033,25033,1000
*NGEN,NSET=OUTER3
25033,33033,1000
*NGEN,NSET=OUTER4
33033,41033,1000
*NGEN,NSET=OUTER5
41033,57033,1000
*NGEN,NSET=OUTER6
57033,65033,1000
*NSET,NSET=OUTER
OUTER1,OUTER2,OUTER3,OUTER4,OUTER5,OUTER6
*NFILL,NSET=JREGION,SINGULAR=1
TIP,OUTER,2,16
**
**SECTION 9
*NGEN,NSET=BOT9
9033,9065,1
*NGEN,NSET=TOP9
9833,9865,1
*NFILL,NSET=ALL9
BOT9,TOP9,16,50
**
**SECTION 25
*NGEN,NSET=BOT25
25033,25065,1
*NGEN,NSET=TOP25,LINE=C
25833,25865,1,9865
*NFILL,NSET=ALL25,BIAS=0.8
BOT25,TOP25,16,50
**
**SECTION 41
*NGEN,NSET=BOT41
41033,41065,1
41833,41865,1,9865
BOT41,TOP41,16,50
**
**SECTION 925
1-657
*NFILL,NSET=ALL925
BOT9,BOT25,16,1000
**
**SECTION 2541
*NSET,NSET=BOT2533,GENERATE
25033,33033,1000
25833,33833,1000,9865
33033,41033,1000
33833,41833,1000,9865
*NSET,NSET=BOT2541
BOT2533,BOT3341
*NSET,NSET=TOP2541
TOP2533,TOP3341
BOT2541,TOP2541,16,50
**
**SECTION 4157
*NGEN,NSET=BOT57
57033,57065,1
*NFILL,NSET=ALL4157
BOT41,BOT57,16,1000
**
**SECTION 57
*NGEN,NSET=TOP57
57833,57865,1
BOT57,TOP57,16,50
**
**SECTION 5765
*NGEN,NSET=TOP5765
57833,65833,1000
OUTER6,TOP5765,16,50
**
**SECTION 19
*NGEN,NSET=TOP19
1833,9833,1000
*NFILL,NSET=ALL19
OUTER1,TOP19,16,50
**
1-658
**INFINITE ELEMENT REGION

*NGEN,NSET=INF25,LINE=C
25933,25965,8,9865
25933,33933,4000,9865
33933,41933,4000,9865
41933,41965,8,9865
*NSET,NSET=INF
INF25,INF2533,INF3341,INF41
**
**CRACK TIP REGION ELEMENTS
*ELEMENT,TYPE=CAX8R
1001,1001,1033,5033,5001,1017,3033,5017,3001
*ELGEN,ELSET=RING
1001,16,4000,4000
**
**ELEMENTS SECTION 9
*ELEMENT,TYPE=CAX8R
9033,9033,9233,9241,9041,9133,9237,9141,9037
*ELGEN,ELSET=SECT9
9033,4,8,8,4,200,200
*ELSET,ELSET=TOP9,GENERATE
9633,9657,8
**
*ELEMENT,TYPE=CAX8R
25233,25233,25033,25041,25241,25133,25037,
25141,25237
*ELGEN,ELSET=SECT25
25233,4,8,8,4,200,200
**
*ELEMENT,TYPE=CAX8R
1033,1033,1233,5233,5033,1133,3233,5133,3033
*ELGEN,ELSET=SECT19
1033,2,4000,4000,4,200,200
**
**ELEMENT SECTION 925
*ELEMENT,TYPE=CAX8R
13033,13033,9033,9041,13041,11033,9037,
11041,13037
1-659
*ELGEN,ELSET=SECT925
13033,4,8,8,4,4000,4000
**
*ELEMENT,TYPE=CAX8R
29233,29233,29033,25033,25233,29133,27033,
25133,27233
29233,4,4000,4000,4,200,200
**
41241,41241,41041,41033,41233,41141,41037,
41133,41237
*ELGEN,ELSET=SECT41
41241,4,8,8,4,200,200
**
*ELEMENT,TYPE=CAX8R
41041,41041,45041,45033,41033,43041,45037,
43033,41037
41041,4,8,8,4,4000,4000
**
57041,57041,57241,57233,57033,57141,57237,
57133,57037
*ELGEN,ELSET=SECT57
57041,4,8,8,4,200,200
**
*ELEMENT,TYPE=CAX8R
57033,57033,57233,61233,61033,57133,59233,
61133,59033
57033,2,4000,4000,4,200,200
**
**INFINITE ELEMENTS
*ELEMENT,TYPE=CINAX5R
41941,41841,41833,41933,41941,41837
*ELGEN,ELSET=INF41
41941,4,8,8
1-660
29933,29833,25833,25933,29933,27833
*ELGEN,ELSET=INF2541
29933,4,4000,4000
25933,25833,25841,25941,25933,25837
*ELGEN,ELSET=INF25
25933,4,8,8
*ELSET,ELSET=INF
INF41,INF2541,INF25
**
*ELSET,ELSET=E1
RING,SECT9,SECT25,SECT41,SECT57,SECT19,SECT925,
SECT2541,SECT4157,SECT5765,INF
*ELASTIC
30.E6,0.3
*SOLID SECTION, MATERIAL=STEEL, ELSET=E1
**
**ADD BOUNDARY CONDITIONS
*NSET,NSET=R9,GENERATE
9065,9865,50
9065,25065,1000
25065,25865,50
*NSET,NSET=YAXIS
R9,R925,R25
*NSET,NSET=L57,GENERATE
57065,57865,50
41065,57065,1000
41065,41865,50
*NSET,NSET=XAXIS
TOP9,L57,L4157,L41
*BOUNDARY
YAXIS,XSYMM
**
**MPC'S TO TIE REDUNDANT NODES
*NSET,NSET=TIP1,GENERATE
1001,64001,1000
1-661
2001,65001,1000
*MPC
TIE,TIP1,TIP2
*MPC
TIE,57865,57833
*MPC
TIE,57833,65833
*MPC
TIE,1833,9833
**
*STEP
APPLY PRESSURE LOAD
*STATIC
1.0,1.0
*DLOAD
TOP9,P2,10.
TIP,0.707107,-0.707107
TYPE=K FACTORS
TIP,0.707107,-0.707107
TYPE=T-STRESS
TIP,0.707107,-0.707107
*EL PRINT
S,
*NODE PRINT
U,RF
*NODE FILE
U,
*OUTPUT,FIELD
*NODE OUTPUT
U,
*ENDSTEP
1-662
Listing 1.4.2-3
*HEADING
AXISYMMETRIC CONICAL CRACK MODEL
SUBMODEL WITH 8 RINGS
*SYSTEM
10.,0.,0.
*NODE,SYSTEM=C
1833,0.,0.,0.
9833,0.,0.,0.
57833,0.,0.,0.
57865,0.,0.,0.
65833,0.,0.,0.
1001,15.,-45.,0.
65001,15.,-45.,0.
1033,8.,-45.,0.
65033,8.,-45.,0.
33033,25.,-45.,0.
41033,25.,-18.,0.
57033,10.,-18.,0.
9033,10.,-72.,0.
25033,25.,-72.,0.
41065,25.,0.,0.
57065,10.,0.,0.
*SYSTEM
0.,0.,0.
*NODE,SYSTEM=C
9865,0.,0.,0.
9065,15.,-90.,0.
25065,30.,-90.,0.
25865,170.,-90.,0.
25965,340.,-90.,0.
25833,170.,-60.,0.
25933,340.,-60.,0.
33833,170.,-45.,0.
33933,340.,-45.,0.
41833,170.,-30.,0.
41933,340.,-30.,0.
41865,170.,0.,0.
41965,340.,0.,0.
*NODE,SYSTEM=R
1033,15.656,-5.6579
1-663
65033,15.658,-5.6559
**
**CRACK TIP REGION
*NGEN,NSET=TIP
1001,65001,2000
*NGEN,NSET=OUTER1
1033,9033,2000
*NGEN,NSET=OUTER2
9033,25033,2000
*NGEN,NSET=OUTER3
25033,33033,2000
*NGEN,NSET=OUTER4
33033,41033,2000
*NGEN,NSET=OUTER5
41033,57033,2000
*NGEN,NSET=OUTER6
57033,65033,2000
*NSET,NSET=OUTER
OUTER1,OUTER2,OUTER3,OUTER4,OUTER5,OUTER6
*NFILL,NSET=JREGION,SINGULAR=1
TIP,OUTER,16,2
**
**CRACK TIP REGION ELEMENTS
*ELEMENT,TYPE=CAX8R
1001,1001,1005,5005,5001,1003,3005,5003,3001
*ELGEN,ELSET=RINGS
1001,16,4000,4000,8,4,4
**
*ELASTIC
30.E6,0.3
*SOLID SECTION, MATERIAL=STEEL, ELSET=RINGS
**
**MPC'S TO TIE REDUNDANT NODES
1001,63001,2000
3001,65001,2000
*MPC
TIE,TIP1,TIP2
*SUBMODEL
OUTER,
**
1-664
*STEP
*STATIC
1.,1.
*BOUNDARY,SUBMODEL,STEP=1
OUTER,1,2
TIP,0.707107,-0.707107
TYPE=K FACTORS
TIP,0.707107,-0.707107
TYPE=T-STRESS
TIP,0.707107,-0.707107
*EL PRINT
S,
*NODE PRINT
U,RF
*ENDSTEP
1.4.3 Elastic-plastic line spring modeling of a finite length

cylinder with a part-through axial flaw
The elastic-plastic line spring elements in ABAQUS are intended to provide inexpensive solutions for
problems involving part-through surface cracks in shell structures loaded predominantly in Mode I by
combined membrane and bending action in cases where it is important to include the effects of
inelastic deformation. This example illustrates the use of these elements. The case considered is a long
cylinder with an axial flaw in its inside surface, subjected to internal pressure. It is taken from the
paper by Parks and White (1982).
When the line spring element model reaches theoretical limitations, the shell-to-solid submodeling
technique is utilized to provide accurate J -integral results. The energy domain integral is used to
evaluate the J -integral for this case.
Geometry and model

The cylinder has an inside radius of 254 mm (10 in), wall thickness of 25.4 mm (1 in), and is assumed
to be very long. The mesh is shown in Figure 1.4.3-1. It is refined around the crack by using
multi-point constraints (MPCs). There are 70 shell elements of type S8R in the symmetric
quarter-model and eight symmetric line spring elements (type LS3S) along the crack. The mesh is
taken from Parks and White, who suggest that this mesh is adequately convergent with respect to the
fracture parameters (J -integral values) that are the primary objective of the analysis. No independent
mesh studies have been done. The use of MPCs to refine a mesh of reduced integration shell elements
(such as S8R) is generally satisfactory in relatively thick shells as in this case. However, it is not
1-665
recommended for thin shells because it introduces constraints that "lock" the response in the finer
mesh regions. In a thin shell case the finer mesh would have to be carried out well away from the
region of high strain gradients.
Three different flaws are studied. All have the semi-elliptic geometry shown in Figure 1.4.3-2, with, in
all cases, c = 3a0 : The three flaws have a0 =t ratios of 0.25 (a shallow crack), 0.5, and 0.8 (a deep
crack). In all cases the axial length of the cylinder is taken as 14 times the crack half-length, c: this is
assumed to be sufficient to approximate the infinite length.
An input data file for the case a0 =t =.5 without making the symmetry assumption about z =0 is also
included. This mesh uses the LS6 line spring elements and serves to check the elastic-plastic capability
of the LS6 elements. The results are the same as for the corresponding mesh using LS3S elements and
symmetry about z =0. The formulation of the LS6 elements assumes that the plasticity is
predominately due to Mode I deformation around the flaw and neglects the effect of the Mode II and
Mode III deformation around the flaw. In the global mesh the displacement in the z-direction is
constrained to be zero at the node at the end of the flaw where the flaw depth goes to zero. To
duplicate this constraint in the mesh using LS6 elements, the two nodes at the end of the flaw (flaw
depth = 0) are constrained to have the same displacements.
Material
The cylinder is assumed to be made of an elastic-plastic metal, with a Young's modulus of 206.8 GPa
(30 ´ 106 lb/in2), a Poisson's ratio of 0.3, an initial yield stress of 482.5 MPa (70000 lb/in 2, and
constant work hardening to an ultimate stress of 689.4 MPa (10 5 lb/in2) at 10% plastic strain, with
perfectly plastic behavior at higher strains.
Loading
The loading consists of uniform internal pressure applied to all of the shell elements, with edge loads
applied to the far end of the cylinder to provide the axial stress corresponding to a closed-end
condition. Even though the flaw is on the inside surface of the cylinder, the pressure is not applied on
the exposed crack face. Since pressure loads on the flaw surface of line spring elements are
implemented using linear superposition in ABAQUS, there is no theoretical basis for applying these
loads when nonlinearities are present. We assume that this is not a large effect in this problem. For
consistency with the line spring element models, pressure loading of the crack face is not applied to the
shell-to-solid submodel.

The line spring elements provide J -integral values directly. Figure 1.4.3-3 shows the J -integral values
at the center of the crack as functions of applied pressure for the three flaws. In the input data the
maximum time increment size has been limited so that adequately smooth graphs can be obtained.
Figure 1.4.3-4 shows the variations of the J -integral values along the crack for the half-thickness crack
¹
¹ y t, is used, where R
(a0 =t =0.5), at several different pressure levels (a normalized pressure, p^ = pR=¾
is the mean radius of the cylinder). These results all agree closely with those reported by Parks and
White (1982), where the authors state that these results are also confirmed by other work. In the region
1-666
Á <30° the results are inaccurate for two reasons. First, the depth of the flaw is changing very rapidly
in this region, which makes the line spring approximation quite inaccurate. Second, J el is of the same
order of magnitude as J pl , but the line spring plasticity model is only valid when J el << J pl : The
results toward the center of the crack (Á >30°) are more accurate than those at the ends of the crack
since the flaw depth changes less rapidly with position in this region and J pl is much larger than J el :
For this reason only J values for Á >30° are shown in Figure 1.4.3-4.
Shell-to-solid submodeling around the crack tip

An input file for the case a0 =t =0.25, which uses the shell-to-solid submodeling capability, is
included. This C3D20R element mesh allows the user to study the local crack area using the energy
domain integral formulation for the J -integral. The submodel uses a focused mesh with four rows of
elements around the crack tip. A 1/r singularity is utilized at the crack tip, the correct singularity for a
fully developed perfectly plastic solution. Symmetry boundary conditions are imposed on two edges of
the submodel mesh, while results from the global shell analysis are interpolated to two surfaces via the
submodeling technique. The global shell mesh gives satisfactory J -integral results; hence, we assume
that the displacements at the submodel boundary are sufficiently accurate to drive the deformation in
the submodel. No attempt has been made to study the effect of making the submodel region larger or
smaller. The submodel is shown superimposed on the global shell model in Figure 1.4.3-5.
In addition, an input file for the case a0 =t =0.25, which consists of a full three-dimensional C3D20R
solid element model, is included for use as a reference solution. This model has the same general
characteristics as the submodel mesh. See inelasticlinespring_c3d20r_ful.inp for further details about
this mesh. One important difference exists in performing this analysis with shell elements as opposed
to continuum elements. The pressure loading is applied to the midsurface of the shell elements as
opposed to the continuum elements, where the pressure is accurately applied along the inside surface
of the cylinder. For this analysis this discrepancy results in about 10% higher J -integral values for the
line spring shell element analysis as compared to the full three-dimensional solid element model.
Results from the submodeled analyses are compared to the LS3S line spring element analysis and full
solid element mesh for variations of the J -integral values along the crack at the a normalized pressure
loading of pR=¹ (¾y t) = 0.898, where R ¹ is the mean radius of the cylinder. As seen in Figure 1.4.3-6,
the line spring elements underestimate the J -integral values for Á <50° for reasons described
previously. Note that at Á =0° the J -integral should be zero due to the lack of crack-tip constraint at
the cylinder surface. A more refined mesh would be required to model this phenomenon properly. It is
quite obvious that the use of shell-to-solid submodeling is required to augment a line spring element
model analysis to obtain accurate J -integral values near the surface of the cylinder.
Input files
inelasticlinespring_05.inp
a0 =t = 0.5.
inelasticlinespring_05_nosym.inp
a0 =t = 0.5 without the symmetry assumption across z =0, using line spring element type LS6.
1-667
inelasticlinespring_progcrack.f
A program used to create a data file giving the flaw depths as a function of position along the
crack.
Shallow crack case, a0 =t =0.25.
Deep crack case, a0 =t =0.8.
inelasticlinespring_c3d20r_sub.inp
C3D20R (a0 =t =0.25) submodel.
inelasticlinespring_c3d20r_ful.inp
C3D20R (a0 =t =0.25) full model.
Reference
· Parks, D. M., and C. S. White, "Elastic-Plastic Line-Spring Finite Elements for Surface-Cracked
Plates and Shells," Transactions of the ASME, Journal of Pressure Vessel Technology, vol. 104,
pp. 287-292, November 1982.
Figures
Figure 1.4.3-1 Finite element model for an axial flaw in a pressurized cylinder.
1-668
Figure 1.4.3-2 Schematic of a semi-elliptical surface crack.
1-669
Figure 1.4.3-3 Normalized J -integral values EJ =(¾y2 t) versus normalized applied pressure
¹ is the mean radius of the cylinder.
¹ (¾y t), where R
pR=
Figure 1.4.3-4 Normalized J -integral values EJ =(¾y2 t) versus position along the flaw surface given
¹ is the
¹ (¾y t)= .574, 1.097, and 1.172. R
by 2Á=¼, for a0 =t =0.5, and normalized applied pressures pR=
mean radius of the cylinder.
Figure 1.4.3-5 Solid submodel superimposed on shell global model.
1-670
Figure 1.4.3-6 Normalized J -integral values EJ =(¾y2 t) versus position along the flaw surface given
¹ is the mean radius of the
¹ (¾y t) =0.898. R
by 2Á=¼ for a0 =t =0.25 and at the normalized pressure. pR=
cylinder.
1-671
Sample listings
1-672
Listing 1.4.3-1
*HEADING
FINITE LENGTH AXIAL FLAW IN PRESSURIZED CYLINDER
*NODE,SYSTEM=C
1, 10.5,-90.,0.
97, 10.5,-90.,6.
99, 10.5,-90.,9.
103, 10.5,-90.,21.
9601, 10.5,234.,0.
9697, 10.5,234.,6.
9699, 10.5,234.,9.
9703, 10.5,234.,21.
11201, 10.5,90.,0.
11297, 10.5,90.,6.
11299, 10.5,90.,9.
11303, 10.5,90.,21.
99991, 0.,0.,0.
99992, 0.,0.,6.
99993, 0.,0.,9.
99994, 0.,0.,21.
*NGEN,NSET=BOTTOM,LINE=C
1,9601,200,99991,,,,,,-1.
9601,11201,200,99991,,,,,,-1.
*NGEN,NSET=MID1,LINE=C
97,9697,200,99992,,,,,,-1.
9697,11297,200,99992,,,,,,-1.
*NGEN,NSET=MID2,LINE=C
99,9699,200,99993,,,,,,-1.
9699,11299,200,99993,,,,,,-1.
*NGEN,NSET=TOP,LINE=C
103,9703,200,99994,,,,,,-1.
9703,11303,200,99994,,,,,,-1.
*NGEN
1,97
97,99
99,103
*NSET,NSET=EDGE1,GENERATE
25,103
*NGEN,NSET=EDGE2
11201,11297
11297,11299
11299,11303
1-673
*NFILL
BOTTOM,MID1,96
MID1,MID2,2
MID2,TOP,4
*NSET,NSET=N1
103,11303
*NSET,NSET=N2
4903,9903,10303,10703,11103
*NSET,NSET=N3
9703,10103,10503,10903
*NSET,NSET=ND1
1,
**
*ELEMENT,TYPE=S8R,ELSET=SHELL
9, 1,5,405,401,3,205,403,201
13, 17,19,419,417,18,219,418,217
17, 25,29,429,425,27,229,427,225
18, 29,37,437,429,33,237,433,229
22, 61,77,477,461,69,277,469,261
23, 77,97,497,477,87,297,487,277
24, 401,409,1209,1201,405,809,1205,801
26, 1201,1209,2809,2801,1205,2009,2805,2001
28, 417,421,1221,1217,419,821,1219,817
31, 1217,1221,2821,2817,1219,2021,2819,2017
34, 429,437,1237,1229,433,837,1233,829
38, 1229,1237,2837,2829,1233,2037,2833,2029
42, 461,477,1277,1261,469,877,1269,861
43, 1261,1277,2877,2861,1269,2077,2869,2061
44, 477,497,1297,1277,487,897,1287,877
45, 1277,1297,2897,2877,1287,2097,2887,2077
46, 2801,2817,6017,6001,2809,4417,6009,4401
48, 2817,2821,6021,6017,2819,4421,6019,4417
50, 2821,2829,6029,6021,2825,4429,6025,4421
58, 2877,2897,6097,6077,2887,4497,6087,4477
47, 6001,6017,9617,9601,6009,7817,9609,7801
49, 6017,6021,9621,9617,6019,7821,9619,7817
51, 6021,6029,9629,9621,6025,7829,9625,7821
59, 6077,6097,9697,9677,6087,7897,9687,7877
1002, 97,99,9699,9697,98,4899,9698,4897
1101, 9601,9697,10097,10001,9649,9897,10049,9801
1102, 9697,9699,10099,10097,9698,9899,10098,9897
*ELGEN,ELSET=SHELL
9,4,4,1
1-674
13,4,2,1
18,4,8,1
24,2,8,1
26,2,8,1
28,3,4,1
31,3,4,1
34,4,8,1
38,4,8,1
46,2,28,6
52,3,16,2
47,2,28,6
53,3,16,2
1002,3,2,1
1102,3,2,1
1101,4,400,100
1102,4,400,100
1103,4,400,100
1104,4,400,100
*ELSET,ELSET=SHELL1,GENERATE
1,59
*MATERIAL,NAME=MSHELL
*ELASTIC
30.E6,.3
*PLASTIC
7.E4,0.
1.E5,.0976
*SHELL SECTION,ELSET=SHELL,MATERIAL=MSHELL
1.,
*ELEMENT,TYPE=LS3S
1, 25,24,23
5, 17,15,13
*ELGEN,ELSET=CRACK
1, 4,-2,1
5, 4,-4,1
*SURFACE FLAW,SIDE=NEGATIVE
1, 0.5000000
3, 0.4982609
5, 0.4930066
7, 0.4841229
9, 0.4714045
11, 0.4545297
13, 0.4330127
15, 0.4061164
1-675
17, 0.3726780
18, 0.3529390
19, 0.3307189
20, 0.3054766
21, 0.2763854
22, 0.2420615
23, 0.1998263
24, 0.1428261
25, 0.0000000
*SHELL SECTION,ELSET=CRACK,MATERIAL=MSHELL
1.,
*MPC
QUADRATIC, 403,401,405,409
QUADRATIC, 407,401,405,409
QUADRATIC, 411,409,413,417
QUADRATIC, 415,409,413,417
QUADRATIC, 418,417,419,421
QUADRATIC, 420,417,419,421
QUADRATIC, 422,421,423,425
QUADRATIC, 424,421,423,425
QUADRATIC, 2805,2801,2809,2817
QUADRATIC, 2813,2801,2809,2817
QUADRATIC, 2823,2821,2825,2829
QUADRATIC, 2827,2821,2825,2829
QUADRATIC, 2833,2829,2837,2845
QUADRATIC, 2841,2829,2837,2845
QUADRATIC, 2849,2845,2853,2861
QUADRATIC, 2857,2845,2853,2861
QUADRATIC, 9609,9601,9649,9697
QUADRATIC, 9617,9601,9649,9697
QUADRATIC, 9619,9601,9649,9697
QUADRATIC, 9621,9601,9649,9697
QUADRATIC, 9625,9601,9649,9697
QUADRATIC, 9629,9601,9649,9697
QUADRATIC, 9637,9601,9649,9697
QUADRATIC, 9645,9601,9649,9697
QUADRATIC, 9653,9601,9649,9697
QUADRATIC, 9661,9601,9649,9697
QUADRATIC, 9669,9601,9649,9697
QUADRATIC, 9677,9601,9649,9697
QUADRATIC, 9687,9601,9649,9697
QUADRATIC, 7897,9697,4897,97
QUADRATIC, 6097,9697,4897,97
1-676
QUADRATIC, 4497,9697,4897,97
QUADRATIC, 2897,9697,4897,97
QUADRATIC, 2097,9697,4897,97
QUADRATIC, 1297,9697,4897,97
QUADRATIC, 897,9697,4897,97
QUADRATIC, 497,9697,4897,97
QUADRATIC, 297,9697,4897,97
**
*STEP,INC=10
*STATIC
1.,1.
*CLOAD
N1,3,1.15E4
N3,3,2.30E4
N2,3,4.60E4
*DLOAD
SHELL,P,3325.
*BOUNDARY
EDGE1,XSYMM
EDGE2,XSYMM
BOTTOM,ZSYMM
11201,2
*NODE PRINT,NSET=ND1,SUMMARY=NO
U,
RF,
*EL PRINT,ELSET=CRACK
JK,
*ENERGY PRINT
*NODE FILE,NSET=ND1
U,RF
*OUTPUT,FIELD
*NODE OUTPUT,NSET=ND1
U,RF
*EL FILE,ELSET=CRACK
JK,
*OUTPUT,FIELD
*ELEMENT OUTPUT,ELSET=CRACK
JK,
*EL FILE,ELSET=SHELL
LOADS,
*OUTPUT,FIELD
*ELEMENT OUTPUT,ELSET=SHELL
1-677
LOADS,
*ENERGY FILE
*ENERGY OUTPUT,VARIABLE=PRESELECT
*END STEP
**
*STEP,INC=20
*STATIC
.1,1.,,.1
*CLOAD
N1,3,.2875E5
N3,3,.575E5
N2,3,1.15E5
*DLOAD
SHELL,P,8312.5
*END STEP
1-678
Listing 1.4.3-2
CC--- PROGRAM TO GENERATE CRACK DEPTH DATA
C PROGRAM CRACK
C
SUBROUTINE HKSMAIN
C
IMPLICIT REAL*8(A-H,O-Z)
C=1.5
CC=C*C
N=24
NNODE=N+1
X0=C/DBLE(N)
X=0.
DO 100 I=1,NNODE
IF(I.GE.17) GO TO 1
IF((I/2)*2.EQ.I) GO TO 10
1 CONTINUE
XX=X*X
TMP=3.
Z=SQRT(CC-XX)/TMP
WRITE(6,99) I,Z
WRITE(16,99)I,Z
99 FORMAT(I5,F10.7)
10 CONTINUE
X=X+X0
100 CONTINUE
REWIND 16
STOP
END
1.4.4 Crack growth in a three-point bend specimen

This example illustrates the modeling of crack length versus time to simulate crack propagation and
the use of crack opening displacement as a crack propagation criterion. For stable crack growth in
ductile materials, experimental evidence indicates that the value of the crack opening displacement
(COD) at a specified distance behind the crack tip associated with ongoing crack extension is usually a
constant. ABAQUS provides the critical crack opening displacement, at a specified distance behind the
crack tip, as a crack propagation criterion. The other crack propagation model used in this
example--prescribed crack length versus time--is usually used to verify the results obtained from
experiments. ABAQUS also provides the critical stress criterion for crack propagation in brittle
1-679
materials.
In this example an edge crack in a three-point bend specimen is allowed to grow based on the crack
opening displacement criterion. Crack propagation is first modeled by giving the crack length as a
function of time. The data for the crack length are taken from Kunecke, Klingbeil, and Schicker
(1993). The data for the crack propagation analysis using the COD criterion are taken from the first
analysis. This example demonstrates how the COD criterion can be used in stable crack growth
analysis.
Problem description
An edge crack in a three-point bend specimen in plane strain, subjected to Mode I loading, is
considered (see Figure 1.4.4-1). The crack length to specimen width ratio is 0.2. The length of the
specimen is 55 mm, and its width is 10 mm. The specimen is subjected to bending loads such that
initially a well-contained plastic zone develops for the stationary crack. Subsequently, the crack is
allowed to grow.
Geometry and model

Due to symmetry only one-half of the specimen is analyzed. The crack tip is modeled as initially
blunted so that finite deformation effects near the crack tip can be taken into account (the NLGEOM
parameter is used on the *STEP option). The mesh is composed of 1737 CPE4 elements (Figure
1.4.4-2). A reasonably fine mesh, necessary to obtain a smooth load versus crack length relation, is
used to model the area in which the plastic zone grows and crack propagation occurs. The loading
point and the support points for the specimen are simulated by analytical rigid surfaces, as shown in
Figure 1.4.4-2.
Material
The material is assumed to be elastic-plastic, with a Young's modulus of E = 200 GPa and Poisson's
ratio of 0.3. The plastic work hardening data are given in Table 1.4.4-1.
Loading and solution control

The analysis is carried out in two stages. The first stage consists of pushing the rigid surface 1.0 mm
into the specimen. No crack growth is specified during this stage.
In the second stage the crack is allowed to propagate while the rigid surface is moved an additional
1.951 mm.
Once a crack-tip node debonds, the traction at the tip is initially carried as a reaction force at that node.
This force is ramped down to zero according to the amplitude curve specified under the *DEBOND
option. The manner in which the forces at the debonded nodes are ramped down greatly influences the
convergence of the solution. The convergence of the solution is also affected by reversals in plastic
flow due to crack propagation. In such circumstances, very small time increments are required to
continue the analysis. In the present analysis the *CONTROLS, PARAMETER=FIELD,
FIELD=DISPLACEMENT option is used to relax the tolerances so that more rapid convergence is
achieved. Because of the localized nature of the nonlinearity in this problem, the resulting loss of
1-680
accuracy is not significant. The *CONTROLS option is generally not recommended.
Crack length versus time

In the case when the crack length is given as a function of time, the second step in the analysis consists
of letting the crack grow according to a prescribed crack length versus time relationship, using the data
taken from Kunecke, Klingbeil, and Schicker.
COD criterion
The loading of the specimen and the specification of the COD criterion for crack growth demonstrates
the flexibility of the COD criterion on the *DEBOND option. Frequently, the crack opening
displacement is measured at the mouth of the crack tip: this is called the crack mouth opening
displacement (CMOD). The crack opening displacement can also be measured at the position where
the initial crack tip was located. Alternatively, the crack-tip opening angle ( CTOA), defined as the
angle between the two surfaces at the tip of the crack, is measured. The crack-tip opening angle can be
easily reinterpreted as the crack opening at a distance behind the crack tip. In this example the COD
specification required to use both the CMOD and the CTOA criteria is demonstrated.
For the purposes of demonstration the crack opening displacement at the mouth of the crack is used as
the initial debond criterion. The first three nodes along the crack propagation surface are allowed to
debond when the crack opening displacement at the mouth of the crack reaches a critical value. To
achieve this, the following loading sequence is adopted: in Step 1, the specimen is loaded to a
particular value (*DEBOND is not used), and in Step 2 the first crack-tip node is allowed to debond
(*DEBOND is used). Steps 3 and 4 and Steps 5 and 6 follow the same sequence as Steps 1 and 2 so
that the two successive nodes can debond. Since, the crack opening displacement is measured at the
mouth of the crack, the value of the DISTANCE parameter on the *FRACTURE CRITERION option
is different in Steps 2, 4, and 6.
The loading sequence adopted above outlines a way in which the CMOD measurements can be
simulated without encountering the situation in which the COD is measured beyond the bound of the
specimen, which would lead to an error message. In this example, the loads at which the crack-tip
nodes debonded were known a priori. In general, such information may not be available, and the
restart capabilities in ABAQUS can be used to determine the load at which the fracture criterion is
satisfied.
The remaining bonded nodes along the crack propagation surface are allowed to debond based on
averaged values of the crack-tip opening angles for different accumulated crack lengths. The data
prescribed under the *FRACTURE CRITERION option in Step 7 are the crack opening displacement
values that were computed from the crack-tip opening angles observed in the analysis that uses the
prescribed crack length versus time criterion. These crack-tip opening angles are converted to critical
crack opening displacements at a fixed distance of 0.04 mm behind the crack tip. Hence, the crack
opening displacement is measured very close to the current crack tip.

Figure 1.4.4-3 shows a plot of the accumulated incremental crack length versus time. The
1-681
user-specified data, as well as the results obtained from the finite element analysis based on the two
criteria, are plotted. Good agreement is observed between the user input values and the results from the
analysis. The curve based on the COD criterion does not correspond with the user-specified data
toward the end of the analysis because an average crack opening displacement was assumed.
Figure 1.4.4-4 shows the reaction force at the node where the displacements are applied as a function
of the accumulated incremental crack length, obtained from the analysis in which the crack length was
specified as a function of time. The curve obtained when the COD criterion is used is almost identical
and is not shown in this figure.
Figure 1.4.4-5 depicts the variation of the reaction force as a function of the displacement at the rigid
body reference node.
The contours of equivalent plastic strain in the near crack-tip region for two different crack advance
positions are shown in Figure 1.4.4-6and Figure 1.4.4-7. Contours of the Mises equivalent stress at the
final stage of the analysis are shown in Figure 1.4.4-8.
Input files
crackgrowth_lengthvtime.inp
Analysis with the crack length versus time criterion.
crackgrowth_cod.inp
Analysis with the COD criterion.
crackgrowth_model.inp
Model data for the two analysis files.
Reference
· G. Kunecke, D. Klingbeil, and J. Schicker, "Rißfortschrittssimulation mit der ABAQUS-option
DEBOND am Beispiel einer statisch belasteten Kerbschlagbiegeprobe," presented at the
ABAQUS German Fracture Mechanics group meeting in Stuttgart, November 1993.
Table
Table 1.4.4-1 Stress-strain data for isotropic plastic behavior.

True Stress True Strain
(MPa)
461.000 0.0
472.810 0.0187
521.390 0.0280
628.960 0.0590
736.306 0.1245
837.413 0.2970
905.831 0.5756
1208.000 1.9942
1-682
Figures
Figure 1.4.4-1 Schematic of the three-point bend specimen.
Figure 1.4.4-2 Finite element mesh for the three-point bend specimen.
Figure 1.4.4-3 Accumulated incremental crack length versus time.
1-683
Figure 1.4.4-4 Variation of the reaction force as a function of the cumulative crack length.
Figure 1.4.4-5 Variation of the reaction force as a function of displacement.
Figure 1.4.4-6 Plastic zone for an accumulated crack length of 1.03 mm.
1-684
Figure 1.4.4-7 Plastic zone for an accumulated crack length of 2.18 mm.
Figure 1.4.4-8 Contours of Mises stress for an accumulated crack length of 2.18 mm.
Sample listings
1-685
Listing 1.4.4-1
*HEADING
CRACK GROWTH IN A THREE POINT BEND SPECIMEN:
CRACK LENGTH CRITERION.
**
** Input file defining the model data.
**
*INCLUDE, INPUT=crackgrowth_model.inp
**
** step 1: Load the specimen
**
**
*STATIC
0.001, 1.0
**
*** load application
**
*BOUNDARY
9997, 1, 1, 1.0
*CONTACT PRINT ,SLAVE=DBDSL,MASTER=DBDMS
DBT, DBSF, DBS
CSTRESS, CDISP
*CONTACT FILE,FREQUENCY=10,SLAVE=DBDSL,MASTER=DBDMS
DBT, DBSF, DBS
*PRINT, CONTACT=YES
*END STEP
**
*** Step 2: Debond the bonded nodes
**
** Use the following lines if the analysis
** is to be run through to completion.
**
**STEP,NLGEOM,INC=500
**STATIC
**0.005, 1.951, 0.00001, 0.05
1-686
**BOUNDARY
**9997, 1, 1, 2.951
**
*STATIC
0.005, 0.4, 0.00001, 0.05
*BOUNDARY
9997, 1, 1, 1.4
**
*DEBOND,SLAVE=DBDSL,MASTER=DBDMS,
TIME INCREMENT=.001,OUTPUT=BOTH,FREQUENCY=1
0.0 , 1.0
0.005, 0.0
*FRACTURE CRITERION, TYPE=CRACK LENGTH, NSET=REF,
TOLERANCE=0.1
1.000001, 0.000000001,
1.256, 0.097,
1.416, 0.181,
1.577, 0.284,
1.737, 0.404,
1.898, 0.541,
2.058, 0.696,
2.114, 0.751,
2.190, 0.830,
2.263, 0.909,
2.332, 0.988,
2.398, 1.067,
2.461, 1.146,
2.521, 1.225,
2.577, 1.304,
2.631, 1.383,
2.681, 1.462,
2.727, 1.541,
2.771, 1.620,
2.811, 1.699,
2.848, 1.778,
2.881, 1.857,
2.912, 1.936,
2.939, 2.015,
2.951, 2.054
*CONTROLS, PARAMETERS=FIELD, FIELD=DISPLACEMENT
, 0.1
1-687
CSTRESS, CDISP
*CONTACT FILE,FREQUENCY=10,SLAVE=DBDSL,
MASTER=DBDMS
DBT,DBSF, DBS
*CONTACT PRINT
CSTRESS, CDISP
*CONTACT PRINT,SLAVE=DBDSL,MASTER=DBDMS
DBT,DBSF,DBS
*PRINT, CONTACT=YES
*END STEP
1-688
Listing 1.4.4-2
*HEADING
CRACK GROWTH IN A THREE POINT BEND SPECIMEN:
COD CRITERION.
**
** Input file defining the model data.
**
*INCLUDE, INPUT=crackgrowth_model.inp
**
** step 1: load the specimen
**
*STATIC
0.001, 1.0
**
*** load application
**
*BOUNDARY
9997, 1, 1, 1.0
DBT, DBSF, DBS
MASTER=DBDMS
DBT, DBSF, DBS
*PRINT, CONTACT=YES
*END STEP
**
*** step 2: Debond the first node.
**
*STEP,NLGEOM,INC=1
*STATIC
0.005, 0.005, 0.00001
*BOUNDARY
9997, 1, 1, 1.005
TIME INCREMENT=.005,FREQUENCY=1,OUTPUT=BOTH
0.0 , 1.0
0.005, 0.0
*FRACTURE CRITERION,TYPE=COD,TOLERANCE=0.01,
DISTANCE=2.36,SYMMETRY
1-689
2.446, 0.028
*CONTROLS, PARAMETERS=FIELD, FIELD=DISPLACEMENT
, 0.1
*CONTACT FILE,SLAVE=DBDSL,MASTER=DBDMS
DBT,DBSF,DBS
DBT,DBSF,DBS
*PRINT, CONTACT=YES
*END STEP
**
** step 3: Load to the point where the second
** node is about to debond.
**
*STEP,NLGEOM,INC=20
*STATIC
0.001, 0.0644, 0.00001
*BOUNDARY
9997, 1, 1, 1.0694
MASTER=DBDMS
DBT,DBSF,DBS
DBT,DBSF,DBS
*PRINT, CONTACT=YES
*END STEP
***
***** step 4: Debond the second node.
***
*STEP,NLGEOM,INC=20
*STATIC
1.0E-06, 1.E-06
*BOUNDARY
9997, 1, 1, 1.069401
0.0 , 1.0
0.005, 0.0
1-690
2.475, 0.06
DBT,DBSF,DBS
DBT,DBSF,DBS
*PRINT, CONTACT=YES
*END STEP
**
** step 5: Load to the point where the third
** node is about to debond
**
*STEP,NLGEOM,INC=50
*STATIC
0.001, 0.085, 0.00001
*BOUNDARY
9997, 1, 1, 1.1544
MASTER=DBDMS
DBT,DBSF,DBS
DBT,DBSF,DBS
*PRINT, CONTACT=YES
*END STEP
***
***** step 6: Debond the third node
***
*STEP,NLGEOM,INC=20
*STATIC
1.0E-06, 1.E-06
*BOUNDARY
9997, 1, 1, 1.154401
0.0 , 1.0
0.005, 0.0
2.518, 0.0934
1-691
DBT,DBSF,DBS
DBT,DBSF,DBS
*PRINT, CONTACT=YES
*END STEP
**
*** step 7: Debond the rest of the nodes.
**
**
** Use the following lines if the analysis is
** to run through to completion
**STEP,NLGEOM,INC=500
**STATIC
**0.001, 1.752, 0.00001
**BOUNDARY
**9997, 1, 1, 2.951
**
*STEP,NLGEOM,INC=75
*STATIC
0.001, 0.40, 0.00001
*BOUNDARY
9997, 1, 1, 1.40
0.0 , 1.0
0.005, 0.0
0.023, 0.262
0.021, 1.037
0.019, 1.66
0.012, 2.18
MASTER=DBDMS
DBT,DBSF,DBS
DBT,DBSF,DBS
*PRINT, CONTACT=YES
*END STEP
1-692
1.5 Import analyses

1.5.1 Springback of two-dimensional draw bending
This example illustrates the forming and springback analysis of a two-dimensional draw bending
process. The forming analysis is performed using ABAQUS/Explicit, and the springback analysis is
run with ABAQUS/Standard using the *IMPORT option.
Problem description
The example described here is one of the benchmark tests reported at the Numisheet '93 Conference.
The benchmark contains a series of six problems performed with three different materials and two
different blank holder forces. One of the six problems is described here. The simulations for all the
problems are described in the paper by Taylor et al. (1993).
The blank initially measures 350 mm by 35 mm and is 0.78 mm thick. The problem is essentially a
plane strain problem (the out-of-plane dimension for the blank is 35 mm). A cross-section of the
geometry of the die, the punch, the blank holder, and the blank is shown in Figure 1.5.1-1. The total
blank holder force is 2.45 kN, and a mass of 5 kg is attached to the blank holder. A coefficient of
friction of 0.144 is used for all interacting surfaces.
The blank is made of mild steel. The material is modeled as an elastic-plastic material with isotropic
elasticity, using the Hill anisotropic yield criterion for the plasticity. The following material properties
are used:
Young's modulus = 206.0 GPa
Poisson's ratio = 0.3
Density = 7800.
Yield stress ¾0 = 167.0 MPa
Anisotropic yield criterion: R11 =1.0, R22 =1.0402, R33 =1.24897, R12 =1.07895,
R13 =1.0, R23 =1.0
The problem is symmetric about a plane through the center of the punch, and only half of the problem
is modeled. The blank is modeled with a single row of 175 first-order shell elements. Symmetry
boundary conditions are applied on the plane of symmetry, and boundary conditions are applied on all
the nodes of the blank to simulate the plane strain conditions.
The forming process is simulated in two steps with ABAQUS/Explicit. The blank holder force is
applied in the first step of the analysis. The force is ramped on to minimize inertia effects. In the
second step of the analysis the punch is moved down 70 mm by prescribing the velocity of the rigid
body reference node for the punch. The velocity is applied with a triangular amplitude function,
starting and ending with zero velocity, and with a peak velocity occurring at the middle of the time
period.
A significant amount of springback occurs in this case. Because the blank is very flexible and the
fundamental mode of vibration is low, it would take a long simulation to obtain a quasi-static solution
of the springback analysis in ABAQUS/Explicit.
1-693
The springback analysis is performed with ABAQUS/Standard using the *IMPORT option. The results
from the forming simulation in ABAQUS/Explicit are imported into ABAQUS/Standard, and a static
analysis calculates the springback. During this step an artificial stress state that equilibrates the
imported stress state is applied automatically by ABAQUS/Standard and gradually removed during the
step. The displacement obtained at the end of the step is the springback, and the stresses give the
residual stress state.
The UPDATE parameter on the *IMPORT option determines the reference configuration. When the
UPDATE parameter is set equal to YES on the *IMPORT option, the deformed sheet with its material
state at the end of the ABAQUS/Explicit analysis is imported into ABAQUS/Standard and the
deformed configuration becomes the reference configuration. This procedure is most convenient if,
during postprocessing, the displacements due to springback need to be displayed. When the UPDATE
parameter is set equal to NO on the *IMPORT option, the material state, displacements, and strains of
the deformed sheet at the end of the ABAQUS/Explicit analysis are imported into ABAQUS/Standard,
and the original configuration remains as the reference configuration. This procedure should be used if
it is desirable to obtain a continuous displacement solution.
In this two-dimensional draw bending problem significant springback occurs, and large-displacement
effects are included in the calculations by including the NLGEOM parameter on the *STEP option.
Further details of the import capability are discussed in ``Transferring results between
ABAQUS/Explicit and ABAQUS/Standard,'' Section 7.6.1 of the ABAQUS/Standard User's Manual
and Section 7.3.1 of the ABAQUS/Explicit User's Manual.

The optimum peak velocity for the punch (the value that gives quasi-static results at least cost) is
determined by running the explicit analysis with peak velocities of 30 m/s, 15 m/s, and 5 m/s. The
energy histories are shown in Figure 1.5.1-2, Figure 1.5.1-3, and Figure 1.5.1-4, respectively. From
these results it is evident that the amount of kinetic energy in the model is too large at a peak velocity
of 30 m/s for the analysis to simulate the quasi-static forming process, while at a peak velocity of 5 m/s
the kinetic energy is virtually zero. A peak velocity for the punch of 15 m/s is chosen for the forming
analysis, as the kinetic energy for this case is considered low enough not to affect the results
significantly. For accurate springback analysis it is important that stresses are not influenced by inertia
effects.
The blank at the end of the ABAQUS/Explicit forming analysis is shown in Figure 1.5.1-5. The shape
after springback is shown in Figure 1.5.1-6. The results compare well with the reported experimental
data. In the numerical results the angle between the outside flange and the horizontal axis is 20.4°. The
average angle measured in the experiments is 17.1°, with a range from 9° to 23° in the experimental
results. The results of the springback analysis when the reference configuration is updated are nearly
identical to that when the reference configuration is not updated.
Input files
springback_exp_form.inp
1-694
Forming analysis in ABAQUS/Explicit with a punch velocity of 15 m/s.

springback_std_importyes.inp
Springback analysis in ABAQUS/Standard with the *IMPORT, UPDATE=YES option.
springback_std_importno.inp
Springback analysis in ABAQUS/Standard with the *IMPORT, UPDATE=NO option.
springback_std_both.inp
Input data used with ABAQUS/Standard for both the forming and the springback analyses.
springback_exp_punchv30.inp
springback_exp_punchv5.inp
Reference
· Taylor, L. M., J. Cao, A. P. Karafillis, and M. C. Boyce, "Numerical Simulations of Sheet Metal
Forming," Proceedings of 2nd International Conference, NUMISHEET 93, Isehara, Japan, Ed. A.
Makinovchi, et al.
Figures
Figure 1.5.1-1 Cross-section showing the geometry of the die, the punch, the blank holder, and the
blank.
1-695
Figure 1.5.1-2 Energy history for forming analysis: 30 m/s peak velocity of punch.
1-696
Figure 1.5.1-5 Blank at the end of the forming analysis in ABAQUS/Explicit.
Figure 1.5.1-6 Blank after springback in ABAQUS/Standard.
Sample listings
1-697
Listing 1.5.1-1
*HEADING
2D Draw Bending problem for Numisheet'93
Mild steel, 2.45kN blankholder force
Shell elements ; Explicit analysis
**
**------------------------ blank
*NODE
1,
351, 0.175
*NGEN, NSET=BLANK1
1,351,1
*NCOPY, OLD=BLANK1,NEW=BLANK2,CHANGE=1000,SHIFT
0., 0.005
*NSET, NSET=BLANK
BLANK1,BLANK2
*ELEMENT, TYPE=S4R
1,1,3,1003,1001
*ELGEN, ELSET=BLANK
1,175,2,2
*SHELL SECTION, ELSET=BLANK, MATERIAL=STEEL
0.00078,5
*DENSITY
7800.,
*ELASTIC
206.E9,0.3
*PLASTIC
.15403E+09,.00000E+00
.19410E+09,.10000E-01
.21913E+09,.20000E-01
.23803E+09,.30000E-01
.25348E+09,.40000E-01
.26668E+09,.50000E-01
.27826E+09,.60000E-01
.28864E+09,.70000E-01
.29807E+09,.80000E-01
.30674E+09,.90000E-01
.31477E+09,.10000E+00
.32226E+09,.11000E+00
.32930E+09,.12000E+00
1-698
.33594E+09,.13000E+00
.34224E+09,.14000E+00
.34822E+09,.15000E+00
.35393E+09,.16000E+00
.35940E+09,.17000E+00
.36464E+09,.18000E+00
.36968E+09,.19000E+00
.37454E+09,.20000E+00
.37922E+09,.21000E+00
.38375E+09,.22000E+00
.38814E+09,.23000E+00
.39239E+09,.24000E+00
.39652E+09,.25000E+00
.40053E+09,.26000E+00
.40443E+09,.27000E+00
.40822E+09,.28000E+00
.41193E+09,.29000E+00
.41554E+09,.30000E+00
.41906E+09,.31000E+00
.42251E+09,.32000E+00
.42588E+09,.33000E+00
.42917E+09,.34000E+00
.43240E+09,.35000E+00
.43556E+09,.36000E+00
.43866E+09,.37000E+00
.44169E+09,.38000E+00
.44468E+09,.39000E+00
.44760E+09,.40000E+00
.45047E+09,.41000E+00
.45330E+09,.42000E+00
.45607E+09,.43000E+00
.45880E+09,.44000E+00
.46148E+09,.45000E+00
.46413E+09,.46000E+00
.46672E+09,.47000E+00
.46928E+09,.48000E+00
.47181E+09,.49000E+00
.47429E+09,.50000E+00
*POTENTIAL
1., 1.0402,1.24897,1.07895,1.,1.
**
**---------------------------- die
*NODE
1-699
2000, 0.026, -0.001, -0.080

*NODE, NSET=DIE
2001, 0.026, -0.001, -0.080
2002, 0.026, -0.001, -0.00539
2008, 0.031, -0.001, -0.00039
2009, 0.180, -0.001, -0.00039
*NGEN, NSET=DIE, LINE=C
2002, 2008, 1,, 0.031, -0.001, -0.00539
*NCOPY, CHANGE=100, OLD=DIE, NEW=DIE, SHIFT
0., 0.007
*ELEMENT, TYPE=R3D4
2001, 2001, 2002, 2102, 2101
*ELGEN, ELSET=DIE
2001, 8, 1, 1
*RIGID BODY, ELSET=DIE, REF NODE=2000
**
**---------------------------- punch
*NODE
3000, 0.000, -0.001, 0.001
*NODE, NSET=PUNCH
3001, 0.000, -0.001, 0.001
3002, 0.020, -0.001, 0.001
3008, 0.025, -0.001, 0.006
3009, 0.025, -0.001, 0.081
*NGEN, NSET=PUNCH, LINE=C
3002, 3008, 1, , 0.02, -0.001, 0.006
*NCOPY, CHANGE=100, OLD=PUNCH, NEW=PUNCH, SHIFT
0., 0.007
*ELEMENT, TYPE=R3D4
3001, 3001, 3002, 3102, 3101
*ELGEN, ELSET= PUNCH
3001, 8,1
*RIGID BODY, ELSET=PUNCH, REF NODE=3000
**
**---------------------------- holder
*NODE, NSET=HOLDER
4001, 0.031, -0.001, 0.010
4002, 0.031, -0.001, 0.00039
4003, 0.081, -0.001, 0.00039
4004, 0.086, -0.001, 0.00539
*NODE
1-700
4005, 0.0585, 0.0025, 0.010

*NCOPY, OLD=HOLDER, NEW=HOLDER, CHANGE=100, SHIFT
0, 0.007
*ELEMENT, TYPE=R3D4
4001, 4001, 4002, 4102, 4101
*ELGEN, ELSET=HOLDER
4001, 3, 1,
*RIGID BODY, ELSET=HOLDER, REF NODE=4005
*ELEMENT, TYPE=MASS, ELSET=EM1
4100, 4005
*MASS, ELSET=EM1
2.5,
**
**----------------------------- surfaces
*SURFACE, NAME=PUNCH
PUNCH,S2
*SURFACE, NAME=HOLDER
HOLDER,S2
*SURFACE, NAME=DIE
DIE,S1
*SURFACE, NAME=TOP
BLANK,S1
*SURFACE, NAME=BOTTOM
BLANK,S2
**
**----------------------- boundary conditions
*BOUNDARY
BLANK, 2,2
BLANK, 4,4
BLANK, 6,6
1, 1,1
1001, 1,1
2000, 1,6
3000, 1,3
3000, 4,6
4005, 1,2
4005, 4,6
**
*RESTART, WRITE, NUMBER INT=1
**
*AMPLITUDE,NAME=APUNCH
0.,0., 0.00466666667, 1., 0.00933333333, 0.
1-701
0.0, 0.0, 0.00001, 1.0
*ELSET,ELSET=CHECK
57,151
*NSET,NSET=CHECKN
1,20,57
**
**---------- apply blankholder force
*STEP
*DYNAMIC, EXPLICIT
,0.00001
*CLOAD,AMP=RAMP
4005, 3, -175.0
**
*SURFACE INTERACTION, NAME=ALLCONT
*FRICTION
0.144,
*CONTACT PAIR, INTERACTION=ALLCONT
PUNCH,TOP
HOLDER,TOP
DIE,BOTTOM
**
*FILE OUTPUT,NUM=1
*NODE FILE
U,V
*EL FILE
S, PEEQ, MISES
STH,
*END STEP
**
** ------------------- move punch down
*STEP
*DYNAMIC,EXPLICIT
, 0.00933333333
*BOUNDARY,AMP=APUNCH,TYPE=VELOCITY
3000,3,3,-15.0
**
*RESTART,WRITE, NUMBER INT=1
*EL HISTORY, ELSET=CHECK, SECTION=1
S,
1-702
*NODE HISTORY, NSET=CHECKN

U3, V3, A3
*ENERGY HISTORY
ALLKE,ALLSE,ALLPD,ALLIE,ALLWK,ETOTAL,ALLFD
*END STEP
1-703
Listing 1.5.1-2
*HEADING
2D Draw bending problem for Numisheet'93
Mild steel, 2.45kN blankholder force
*IMPORT,STEP=2,INT=1,STATE=YES, UPDATE=YES
BLANK,
*IMPORT NSET
BLANK,
*BOUNDARY
BLANK,2,2
BLANK, 4,4
BLANK, 6,6
1,1,6
1001,1,6
**
*STEP,NLGEOM,INC=50
*STATIC
0.1,1.
*NODE FILE,FREQ=1
U,
*PRINT,FREQ=1
*NODE PRINT,FREQ=50
U,
RF,
*EL PRINT,FREQ=0
*END STEP
1.5.2 Deep drawing of a square box

This example illustrates the forming of a three-dimensional shape by a deep drawing process. The most
efficient way to analyze this type of problem is to analyze the forming step with ABAQUS/Explicit and
to import the results in ABAQUS/Standard to analyze the springback that occurs after the blank is
removed from the tool with a static procedure. Since the forming process is essentially a quasi-static
problem, the computations with ABAQUS/Explicit are performed over a sufficiently long time period
to render inertial effects negligible. For verification purposes the complete analysis is also carried out
with ABAQUS/Standard. However, this is computationally more expensive and will be prohibitively
expensive for simulation of the forming of realistic, complex components.
Problem description
1-704
The blank is initially square, 200 mm by 200 mm, and is 0.82 mm thick. The rigid die is a flat surface
with a square hole 102.5 mm by 102.5 mm, rounded at the edges with a radius of 10 mm. The rigid
square punch measures 100 mm by 100 mm and is rounded at the edges with the same 10 mm radius.
The blank holder can be considered a flat plate, since the blank never comes close to its edges. The
geometry of these parts is illustrated in Figure 1.5.2-1and Figure 1.5.2-2. The rigid surfaces are offset
from the blank by half of the thickness of the blank. The contact algorithm in ABAQUS/Explicit takes
into account the shell thickness. When the forming step is modeled with ABAQUS/Standard, the
thickness is accounted for indirectly by shifting the pressure penetration curve defined with the
*SURFACE BEHAVIOR option. A mass of 0.6396 kg is attached to the blank holder, and a
concentrated load of 2.287 ´ 104 N is applied to the control node of the blank holder. The blank holder
is then allowed to move only in the vertical direction to accommodate changes in blank thickness
during deformation. The coefficient of friction between the sheet and the punch is taken to be 0.25, and
that between the sheet and the die is 0.125. It is assumed that there is no friction between the blank and
the blank holder.
The blank is made of aluminum-killed steel, which is assumed to satisfy the Ramberg-Osgood relation
between true stress and logarithmic strain:
² = (¾=K )1=n ;
with a reference stress value (K) of 513 MPa and a work-hardening exponent (n) of 0.223. Isotropic
elasticity is assumed, with a Young's modulus of 211 GPa and a Poisson's ratio of 0.3. An initial yield
stress of 91.3 MPa is obtained from these data. The stress-strain behavior is defined by piecewise
linear segments matching the Ramberg-Osgood curve up to a total (logarithmic) strain level of 107%,
with Mises yield, isotropic hardening, and no rate dependence.
Given the symmetry of the problem, it is sufficient to model only a one-eighth sector of the box.
However, we have employed a full one-quarter model to make it easier to visualize. We use 4-node,
three-dimensional rigid surface elements (type R3D4) to model the die, the punch, and the blank
holder. The blank is modeled with 4-node, bilinear finite-strain shell elements (type S4R).
This problem was used by Nagtegaal and Taylor (1991) to compare implicit and explicit finite element
techniques for the analysis of sheet metal forming problems. The computer time involved in running
the simulation using explicit time integration with a given mesh is directly proportional to the time
period of the event, since the stable time increment size is a function of the mesh size (length) and the
material stiffness. Thus, it is usually desirable to run the simulation at an artificially high speed
compared to the physical process. If the speed in the simulation is increased too much, the solution
does not correspond to the low-speed physical problem; i.e., inertial effects begin to dominate. In a
typical forming process the punch may move at speeds on the order of 1 m/sec, which is extremely
slow compared to typical wave speeds in the materials to be formed. (The wave speed in steel is
approximately 5000 m/sec.) In general, inertia forces will not play a dominant role for forming rates
that are considerably higher than the nominal 1 m/sec rates found in the physical problem. The explicit
solutions obtained with punch speeds of 10, 30, and 100 m/sec are compared with the static solution
obtained with ABAQUS/Standard. The results at 10 m/sec were virtually indistinguishable from the
static results. Minor differences could be observed at the intermediate speed of 30 m/sec. The results at
1-705
100 m/sec were considerably different from the static results. In the results presented here, the drawing
process is simulated by moving the reference node for the punch downward through a total distance of
36 mm at a constant velocity of 30 m/sec. Comparison of analyses of various metal forming problems
using explicit dynamic and static procedures is discussed in the paper by Nagtegaal and Taylor.
Although this example does not contain rate-dependent material properties, it is common in sheet
metal forming applications for this to be a consideration. If the material is rate-dependent, the
velocities cannot be artificially increased without affecting the material response. Instead, the analyst
can use the technique of mass scaling to adjust the effective punch velocity without altering the
material properties. ``Rolling of thick plates,'' Section 1.3.7, contains an explanation and an example of
the mass scaling technique.
The results from the forming simulation obtained using ABAQUS/Explicit are made available to
ABAQUS/Standard by using the *IMPORT option with the parameter UPDATE=YES. The springback
that occurs and the residual stress state are then determined by performing a static analysis in
ABAQUS/Standard. During this step an artificial stress state that equilibrates the imported stress state
is applied automatically by ABAQUS/Standard and gradually removed during the step. The
displacement obtained at the end of the step is the springback, and the stresses give the residual stress
state. Only the deformed sheet with its material state at the end of the ABAQUS/Explicit analysis is
imported into ABAQUS/Standard. Boundary conditions are imposed in the ABAQUS/Standard
analysis to prevent rigid body motion and for symmetry. The node at the center of the box is fixed in
the z-direction.
The springback of the formed sheet is also analyzed in ABAQUS/Standard by setting UPDATE=NO
on the *IMPORT option. In this case the displacements are the total values relative to the original
reference configuration. This makes it easy to compare the results with the analysis in which both the
forming and springback are analyzed with ABAQUS/Standard.
Further details of the import capability are discussed in ``Transferring results between
ABAQUS/Explicit and ABAQUS/Standard,'' Section 7.6.1 of the ABAQUS/Standard User's Manual
and Section 7.3.1 of the ABAQUS/Explicit User's Manual.
Contact modeling in the forming step

In ABAQUS/Explicit the contact constraints, which take the current shell thickness into account, are
enforced with the default, kinematic contact method in the primary input file, although a model that
uses penalty contact is also included. (Penalty contact is invoked if MECHANICAL
CONSTRAINT=PENALTY is included on the *CONTACT PAIR option.) The results with these two
methods are very similar. It would also be valid to use a combination of both methods in one analysis;
for example, penalty contact between the blank holder and the blank and kinematic contact for the
other contact pairs. If the penalty method is used to model contact between the punch and the blank,
the blank will tend to bounce off the punch after an impact at the beginning of the analysis (the blank
is initially at rest). This phenomenon does not occur with kinematic contact because impacts are
perfectly plastic for kinematic contact. The significance of this phenomenon has been greatly reduced
in the analysis that uses penalty contact by ramping up the velocity of the punch smoothly over the
initial 1% of the step duration.
1-706
Because contact occurs on both sides of the blank, a double-sided surface is used to model the blank in
ABAQUS/Explicit. In ABAQUS/Explicit when a shell is pinched between two contact surfaces, such
as between the blank holder and die in this problem, at least one of the constraints will not be enforced
exactly, even if kinematic enforcement is used for both contact pairs. This aspect is discussed in
``Common difficulties associated with contact modeling,'' Section 20.5.1 of the ABAQUS/Explicit
User's Manual. The slight noncompliance does not affect the predicted properties of the formed part
significantly.
Double-sided surfaces are not available in ABAQUS/Standard, so two single-sided surfaces are used to
model the blank when the forming step is modeled in ABAQUS/Standard: one surface to model the top
of the blank and one to model the bottom of the blank. When a shell in ABAQUS/Standard is pinched
between two surfaces, at least one of the constraints must use "softened" contact. In the analysis in
which the forming and the springback steps are carried out with ABAQUS/Standard, softened contact
is used for all contact constraints. The contact stiffness is chosen sufficiently high so that the results
are not affected significantly.

Figure 1.5.2-3 shows contours of shell thickness in the blank after forming. Figure 1.5.2-4 shows
contours of equivalent plastic strain in the blank in the final deformed shape.
Closer inspection of the results reveals that the corners of the box are formed by stretching, whereas
the sides are formed by drawing action. This effect leads to the formation of shear bands that run
diagonally across the sides of the box, resulting in a nonhomogeneous wall thickness. Note also the
uneven draw of the material from the originally straight sides of the blank. Applying a more localized
restraint near the midedges of the box (for example, by applying drawbeads) and relaxing the restraint
near the corners of the box is expected to increase the quality of the formed product.
Figure 1.5.2-5 shows the reaction force on the punch. Figure 1.5.2-6 shows the vertical displacement
of the blank holder over time. The rise of the blank holder is attributable to the increased thickness of
the blank, as well as to the tendency of the blank to lift up off of the die. Figure 1.5.2-7 shows the
thinning of an element at the corner of the box.
The springback analysis runs in 6 increments in ABAQUS/Standard. Most of the springback occurs in
the z-direction, and the springback is not significant. The corner of the outside edge of the formed box
drops approximately 0.45 mm, while the vertical side of the box rises by approximately 0.2 mm. Figure
1.5.2-8 shows a contour plot of the displacements in the z-direction obtained from the springback
analysis.
The analysis with UPDATE=NO on the *IMPORT option yields similar results. However, in this case
the displacements are interpreted as total values relative to the original configuration.
Input files
deepdrawbox_exp_form.inp
Forming analysis with ABAQUS/Explicit.
deepdrawbox_std_importyes.inp
1-707
ABAQUS/Standard springback analysis with the UPDATE=YES parameter on the *IMPORT

option.
deepdrawbox_std_importno.inp
ABAQUS/Standard springback analysis with the UPDATE=NO parameter on the *IMPORT
option.
deepdrawbox_std_both.inp
Forming and springback analyses done in ABAQUS/Standard.
deepdrawbox_exp_form_penalty.inp
Original mesh using penalty contact in ABAQUS/Explicit.
deepdrawbox_exp_finemesh.inp
Forming analysis of a fine mesh case (included for the sole purpose of testing the performance of
the ABAQUS/Explicit code).
deepdrawbox_std_finesprngback.inp
Springback analysis of a fine mesh case (included for the sole purpose of testing the performance
of the ABAQUS/Explicit code).
Reference
· Nagtegaal, J. C., and L. M. Taylor, "Comparison of Implicit and Explicit Finite Element Methods
for Analysis of Sheet Forming Problems," VDI Berichte No. 894, 1991.
Figures
Figure 1.5.2-1 Meshes for the die, punch, and blank holder.
1-708
Figure 1.5.2-2 Undeformed mesh for the blank.
Figure 1.5.2-3 Contours of shell thickness.
Figure 1.5.2-4 Contours of equivalent plastic strain.
1-709
Figure 1.5.2-5 Reaction force on the punch versus punch displacement.
Figure 1.5.2-6 Vertical holder displacement versus time.
Figure 1.5.2-7 Shell thickness of the thinnest part of the blank versus time.
1-710
Figure 1.5.2-8 Contour plot showing the springback in the z-direction.
Sample listings
1-711
Listing 1.5.2-1
*HEADING
DEEP DRAWING OF A SQUARE BOX
(1/4 SYMMETRY MODEL)
**
** Generate the mesh for the PUNCH
**
*NODE
101,.0 ,.0,0.00041
107,.04,.0,0.00041
113,.05,.0 ,.01
119,.05,.0 ,.07
199,.04,.0 ,.01
701,.0 ,.04,0.00041
707,.04,.04,0.00041
713,.05,.04,.01
719,.05,.04,.07
799,.04,.04,.01
1307,.04,.04,.00041
1313,.04,.05,.01
1319,.04,.05,.07
1399,.04,.04,.01
1907,.0 ,.04,.00041
1913,.0 ,.05,.01
1919,.0 ,.05,.07
1999,.0 ,.04,.01
199991,0.,0.,0.00041
*NGEN,NSET=P1
101,107,1
113,119,1
*NGEN,NSET=P1,LINE=C
107,113,1,199
*NGEN,NSET=P2
701,707,1
713,719,1
707,713,1,799
*NSET,NSET=P2C,GEN
707,719,1
*NGEN,NSET=P3
1313,1319,1
1-712
1307,1313,1,1399
*NGEN,NSET=P4
1913,1919,1
1907,1913,1,1999
*NFILL
P1,P2,6,100
P3,P4,6,100
*NCOPY,CHANGE NUMBER=100,OLD SET=P2C,SHIFT,
MULTIPLE=6
0.,0.,0.
.04,.04,0., .04,.04,1.,15.
*NSET,NSET=P4,GEN
101,701,100
*ELEMENT,TYPE=R3D4,ELSET=PUNCH
1, 101,102,202,201
109, 708,709,809,808
241, 707, 708, 808,707
242, 707, 808, 908,707
243, 707, 908,1008,707
244, 707,1008,1108,707
245, 707,1108,1208,707
246, 707,1208,1308,707
247, 707,1308,1408,706
248, 706,1408,1508,705
249, 705,1508,1608,704
250, 704,1608,1708,703
251, 703,1708,1808,702
252, 702,1808,1908,701
*ELGEN,ELSET=PUNCH
1, 18,1,1, 6,100,18
109, 11,1,1, 12,100,11
*RIGID BODY,ELSET=PUNCH,REF NODE=199991
**
** Generate the mesh for the HOLDER
**
*NODE
20101,.06,.0,.01
20102,.06,.0,.00041
20108,.13,.0,.00041
20109,.13,.0,.01
20701,.06,.05,.01
20702,.06,.05,.00041
1-713
20708,.13,.05,.00041
20709,.13,.05,.01
21301,.05,.06,.01
21302,.05,.06,.00041
21308,.05,.13,.00041
21309,.05,.13,.01
21901,.0,.06,.01
21902,.0,.06,.00041
21908,.0,.13,.00041
21909,.0,.13,.01
299991,0.06,0.,0.01
*NGEN,NSET=H1
20101,20102,1
20102,20108,1
20108,20109,1
*NGEN,NSET=H2
20701,20702,1
20702,20708,1
20708,20709,1
*NGEN,NSET=H3
21301,21302,1
21302,21308,1
21308,21309,1
*NGEN,NSET=H4
21901,21902,1
21902,21908,1
21908,21909,1
*NFILL
H1,H2,6,100
H3,H4,6,100
*NCOPY,CHANGE NUMBER=100,OLD SET=H2,SHIFT,
MULTIPLE=6
0.,0.,0.
.05,.05,0., .05,.05,1.,15.
*ELEMENT,TYPE=R3D4,ELSET=HOLDER
20001, 20101,20201,20202,20102
*ELGEN,ELSET=HOLDER
20001, 8,1,1, 18,100,8
*RIGID BODY,ELSET=HOLDER,REF NODE=299991
*ELEMENT,TYPE=MASS,ELSET=EMASS
20000,299991
*MASS,ELSET=EMASS
6.396E-1,
1-714
**
** Generate the mesh for the DIE
**
*NODE
599991, .05125, 0., -0.06
*NODE,NSET=DIE
50101, .05125,0.,-.06
50107, .05125,0.,-.01
50113, .06125,0.,-0.00041
50119, .13000,0.,-0.00041
50199, .06125,0.,-.01
**
50701, .05125,.04125,-.06
50707, .05125,.04125,-.01
50713, .06125,.04125,-0.00041
50719, .13000,.04125,-0.00041
50799, .06125,.04125,-.01
**
51301, .04125,.05125,-.06
51307, .04125,.05125,-.01
51313, .04125,.06125,-0.00041
51319, .04125,.13000,-0.00041
51399, .04125,.06125,-.01
**
51901, 0.,.05125,-.06
51907, 0.,.05125,-.01
51913, 0.,.06125,-0.00041
51919, 0.,.13000,-0.00041
51999, 0.,.06125,-.01
*NGEN,NSET=D1
50101,50107,1
50113,50119,1
*NGEN,NSET=D1,LINE=C
50107,50113,1,50199
*NGEN,NSET=D2
50701,50707,1
50713,50719,1
50707,50713,1,50799
*NGEN,NSET=D3
51301,51307,1
51313,51319,1
1-715
51307,51313,1,51399
*NGEN,NSET=D4
51901,51907,1
51913,51919,1
51907,51913,1,51999
*NFILL
D1,D2,6,100
D3,D4,6,100
*NCOPY,CHANGE NUMBER=100,OLD SET=D2,SHIFT,
MULTIPLE=6
0.,0.,0.
.04125,.04125,0., .04125,.04125,1., 15.
*ELEMENT,TYPE=R3D4
50001, 50101,50102,50202,50201
*ELGEN,ELSET=DIE
50001, 18,1,1, 18, 100,18
*RIGID BODY,ELSET=DIE,REF NODE=599991
**
** Generate the mesh for the BLANK
**
*NODE
90001,.0,.0,0.
90036,.1,.0,0.
91261,.0,.1,0.
91296,.1,.1,0.
*NGEN,NSET=B1
90001,90036,1
*NGEN,NSET=B2
91261,91296,1
*NFILL
B1,B2,35,36
*NSET,NSET=B4,GEN
90001,91261,36
*ELEMENT,TYPE=S4R,ELSET=BLANK
3001, 90001,90002,90038,90037
*ELGEN,ELSET=BLANK
3001, 35,1,1, 35,36,35
*SHELL SECTION,MATERIAL=STEEL,ELSET=BLANK,
.00082,5
**
** Define material properties for STEEL
1-716
**
*DENSITY
7800.,
*ELASTIC
2.1E11,0.3
*PLASTIC
0.91294E+08, 0.00000E+00
0.10129E+09, 0.21052E-03
0.11129E+09, 0.52686E-03
0.12129E+09, 0.97685E-03
0.13129E+09, 0.15923E-02
0.14129E+09, 0.24090E-02
0.15129E+09, 0.34674E-02
0.16129E+09, 0.48120E-02
0.17129E+09, 0.64921E-02
0.18129E+09, 0.85618E-02
0.19129E+09, 0.11080E-01
0.20129E+09, 0.14110E-01
0.21129E+09, 0.17723E-01
0.22129E+09, 0.21991E-01
0.23129E+09, 0.26994E-01
0.24129E+09, 0.32819E-01
0.25129E+09, 0.39556E-01
0.26129E+09, 0.47301E-01
0.27129E+09, 0.56159E-01
0.28129E+09, 0.66236E-01
0.29129E+09, 0.77648E-01
0.30129E+09, 0.90516E-01
0.31129E+09, 0.10497E+00
0.32129E+09, 0.12114E+00
0.33129E+09, 0.13916E+00
0.34129E+09, 0.15919E+00
0.35129E+09, 0.18138E+00
0.36129E+09, 0.20588E+00
0.37129E+09, 0.23287E+00
0.38129E+09, 0.26252E+00
0.39129E+09, 0.29502E+00
0.40129E+09, 0.33054E+00
0.41129E+09, 0.36929E+00
0.42129E+09, 0.41147E+00
0.43129E+09, 0.45729E+00
0.44129E+09, 0.50696E+00
1-717
0.45129E+09, 0.56073E+00
0.46129E+09, 0.61881E+00
0.47129E+09, 0.68145E+00
0.48129E+09, 0.74890E+00
0.49129E+09, 0.82142E+00
0.50129E+09, 0.89928E+00
0.51129E+09, 0.98274E+00
0.52129E+09, 0.10721E+01
**
** Define surfaces
**
*SURFACE, NAME=PUNCH
PUNCH,SNEG
*SURFACE, NAME=HOLDER
HOLDER,SPOS
*SURFACE, NAME=DIE
DIE,SPOS
*SURFACE, NAME=BLANK
BLANK,
**
** Apply symmetry boundary conditions
**
*BOUNDARY
D1,2,2
P1,2,2
H1,2,2
D2,1,1
P2,1,1
H2,1,1
B1,YSYMM
B4,XSYMM
199991, 1,2
199991, 4,6
299991,1,2
299991,4,6
599991,1,3
599991,4,6
**
** Simulate the deep drawing operation
**
*RESTART,WRITE,NUMBER INTERVAL=2, TIMEMARKS=NO
**
*STEP
1-718
*DYNAMIC,EXPLICIT
,.0012
199991,3,3,-30.
*CLOAD
299991, 3, -2.287E4
**
*SURFACE INTERACTION,NAME=PUNCH_TOP
*FRICTION
0.25,
*CONTACT PAIR,INTERACTION=PUNCH_TOP
PUNCH,BLANK
*CONTACT PAIR
HOLDER,BLANK
*SURFACE INTERACTION,NAME=DIE_BOTT
*FRICTION
0.125,
*CONTACT PAIR,INTERACTION=DIE_BOTT
DIE,BLANK
**
*EL FILE
PEEQ,MISES
*NODE FILE
U,V
*ENERGY FILE
**
** history output
**
*NSET,NSET=NPUNCH
199991,
*NSET,NSET=HOLDER
299991,
*ELSET,ELSET=ELHIST
3647,
*HISTORY OUTPUT,TIME INTERVAL=.6E-4
*NODE HISTORY,NSET=NPUNCH
U3,RF3
*NODE HISTORY,NSET=HOLDER
U3,V3,A3,
*EL HISTORY,ELSET=ELHIST
STH,
PEEQ,
1-719
*ENERGY HISTORY
ALLKE,ALLSE,ALLWK,ALLPD,ALLIE,ALLVD,ETOTAL,
ALLFD,ALLCD,DT,ALLAE
*OUTPUT,FIELD,OP=NEW,NUMBER INTERVAL=5,
TIMEMARKS=NO
*ELEMENT OUTPUT
S, PEEQ
*NODE OUTPUT
U,
*OUTPUT,HISTORY,OP=NEW,TIME INTERVAL=2.4E-6
*ENERGY OUTPUT
ALLAE,ALLIE,ALLKE,ALLPD,ALLSE,ALLVD,ALLWK,ETOTAL
*END STEP
1-720
Listing 1.5.2-2
*HEADING
SHEET METAL FORMING: import model
spring back analysis
*IMPORT,STEP=1,INT=2,STATE=YES, UPDATE=YES
BLANK,
*IMPORT NSET
B1,B4
*BOUNDARY
B1,YSYMM
B4,XSYMM
90001,1,6
**
**
*STEP,NLGEOM,INC=50
*STATIC
0.1,1.
*NODE FILE,FREQ=1
U,
*PRINT,FREQ=50
*NODE PRINT,FREQ=50
U,
RF,
*EL PRINT,FREQ=0
*END STEP
1-721
Dynamic Stress/Displacement Analyses
2. Dynamic Stress/Displacement Analyses

2.1 Dynamic stress analyses
2.1.1 Nonlinear dynamic analysis of a structure with local
inelastic collapse
This example illustrates an inexpensive approach to the prediction of the overall response of a
structure that exhibits complex local behavior. The case studied is an unrestrained pipe whip example,
where an initially straight pipe undergoes so much motion that the pipe section collapses. A two-stage
technique is used to predict the response. First, the collapse of the section is studied under static
conditions using a generalized plane strain model. This analysis defines the moment-curvature
relationship for the section under conditions of pure bending. It also shows how the section deforms as
it collapses. This information can be used to judge whether the deformation is reasonable with respect
to possible failure (fracture) of the section. In addition, this first stage analysis can be used to calculate
the change in the cross-sectional area enclosed by the pipe as a function of the curvature of the pipe. In
a pipe whip case the driving force is caused by fluid jetting from a break in the pipe; and, if the pipe
does undergo such large motion, a section may be deformed sufficiently to choke the flow. The second
stage of the analysis is to predict the overall dynamic response of the pipe, using the moment-curvature
response of the section that has been obtained in the first analysis to define the inelastic bending
behavior of the beam. This two-stage approach provides a straightforward, inexpensive method of
evaluating the event. The method is approximate and may give rise to significant errors. That aspect of
the approach is discussed in the last section below.
Modeling
The problem is shown in Figure 2.1.1-1. To investigate the static collapse of the section, we consider a
unit length of an initially straight pipe subjected to a pure bending moment and assume that plane
sections remain plane. We can think of this unit length of pipe as being bounded at its ends by rigid
walls and imagine the bending to be achieved by rotation of the walls relative to each other, the end
sections being allowed to distort only in the plane of the walls (see Figure 2.1.1-2). With this
idealization the pipe section can be modeled and discretized using generalized plane strain elements, as
shown in Figure 2.1.1-2. Bending occurs about the x-axis, and symmetry conditions are prescribed
along the y-axis. There will not be symmetry about the x-axis because of the Poisson's effect. To
remove rigid body motion in the y-direction, point A is fixed in that direction. Symmetry implies no
x-displacement at x =0 and that there is no rotation of the section about the y-axis. 8-node and 4-node
generalized plane strain elements with reduced integration are used. In addition to the eight or four
regular nodes used for interpolation, these elements require two extra nodes that are common to all
elements in the model. Degree of freedom 1 at the first of these extra nodes is the relative displacement
between the boundary planes, while the first two degrees of freedom at the second extra node are the
relative rotation of these planes.
Since the problem involves bending the pipe cross-section, fully integrated 4-node elements will not
2-722
provide accurate results, especially when the pipe is fairly thin, because they will suffer from "shear
locking"--they will not provide the bending deformation because to do so requires that they shear at
their integration points and this shearing requires an unrealistically large amount of strain energy. This
problem is avoided by integrating the elements only at their centroids but the elements then exhibit
singular modes--modes of deformation that do not cause strain. ABAQUS uses orthogonal hourglass
generalized strains and associated stiffness to avoid such spurious singular mode behavior. Although
these techniques are not always reliable, they can work well and do so in this example. A superior
approach would be to use the fully integrated incompatiable mode element CGPE6I element. For
additional discussion of these points see ``Performance of continuum and shell elements for linear
analysis of bending problems,'' Section 2.3.5 of the ABAQUS Benchmarks Manual.
For the dynamic analysis of the pipe whip event the pipeline is modeled with 10 beam elements of type
B21. These are planar beam elements that use linear interpolation of displacement and rotation. The
moment-curvature relation obtained from the static analysis (shown in Figure 2.1.1-7) is used in the
*BEAM GENERAL SECTION, SECTION=NONLINEAR GENERAL option input data to define the
bending behavior of the beams. A definition for the axial force versus strain behavior of the beams is
also required and is provided by conversion of the uniaxial stress-strain relation given in Figure 2.1.1-1
into force versus strain by multiplying the stress by the current area, A, of the cross section. This
current area is computed from the original cross-sectional area A0 by assuming that the material is
incompressible, so A = A0 l0 =l, where l is the current length and l0 is the original length.
This definition of the beam section behavior provides for no interaction between the bending and axial
stretching, although in most real cases there will actually be some interaction. However, this
approximation is probably reasonable in this particular problem since the response is predominantly
bending and no appreciable error is introduced by the little stretching that does occur.
Loading and solution control

In the large-displacement static analysis of the inelastic collapse of the section, rotation of the
boundary planes about the x-axis is prescribed at the second extra node of the generalized plane strain
model. The Riks procedure is used: this method usually provides rapid convergence in such cases,
especially when unstable response occurs.
In the large-displacement dynamic analysis the blowdown force is treated as a follower force. During
the first 0.06 seconds of the event it has a constant magnitude of 30 kN (this is about three times the
load required to produce maximum moment in the static response of the section). After that time the
load is zero. The response is computed for a time period of 0.4 seconds, using automatic time
incrementation. A half-step residual tolerance ( HAFTOL) of 30 kN (which is the magnitude of the
applied load) is used. Since we expect considerable plastic deformation, high frequency response
should be damped quickly in the actual event, so that this value of HAFTOL should be adequate to
give reasonably accurate results.

Figure 2.1.1-3 shows a series of deformed configuration plots from the static analysis using element
type CGPE10R, and Figure 2.1.1-4 shows the same plots for the analysis using element type CGPE6R.
2-723
Figure 2.1.1-5 and Figure 2.1.1-6are corresponding plots of equivalent plastic strain. Figure 2.1.1-3and
Figure 2.1.1-5 clearly show that the discretization is too coarse or should be rezoned later in the
deformation, but it is judged that this is not critical to the overall moment-rotation response prediction.
Figure 2.1.1-7 shows the moment-curvature responses predicted by the analyses. The unstable nature
of the response is clearly illustrated.
Figure 2.1.1-8 shows a series of deformed configuration plots from the dynamic analysis. After the
shutdown of the force at 0.06 seconds, the momentum of the pipe is enough to cause localization of the
deformation at the root of the cantilever as the section collapses there: the pipe whips around this
hinge in a full circle and beyond its initial configuration. As well as this major hinge at the root,
permanent plastic deformation develops throughout most of the pipe, leaving it bent into an arc. Time
history plots of the tip displacement are shown in Figure 2.1.1-9and of the curvature strain at the
localized hinge in Figure 2.1.1-10. Figure 2.1.1-11 shows the moment-curvature response for the
element at the support and shows the elastic unloading and reloading that takes place during and at the
end of the event. Figure 2.1.1-12shows the history of the energy content during the dynamic analysis
and clearly shows the initial build-up of kinetic energy, which is then converted almost entirely to
plastic dissipation.
This two-stage approach to the problem has the advantages of being simple and computationally
inexpensive. It contains some obvious approximations. One is that interaction effects between bending,
axial, and torsional behavior are neglected. This lack of interaction between the various modes of
cross-sectional response is a basic approximation of the nonlinear beam general section option. In
reality, axial or torsional strain will have the effect of reducing the strength of the section in bending.
This effect is unlikely to be significant in a case that is dominated by bending, but it can be important
if large axial or torsional loadings occur. The approach also neglects the effect of the axial gradient of
the cross-sectional behavior on the response. This may be a significant error, but its evaluation would
require a detailed, three-dimensional analysis for comparison; and that exercise is beyond the scope of
this example. Another possibly significant error is the neglect of rate effects on the response. The
cross-sectional collapse involves large strains, which occur in a very short time in the dynamic
loadings, so high strain rates arise. It is likely that the material will exhibit strain rate dependence in its
yield behavior and will, therefore, be rather stiffer than the static analysis predicts it to be. This should
have the effect of spreading the hinge along the pipe and reducing the localization (because the strain
rates increase at the section where most deformation is occurring, and that increased strain rate
increases the resistance of the section). The magnitude of this effect can be estimated from the solution
we have obtained. From Figure 2.1.1-5we see that typical strains in the section are about 10-20% when
the section is far into collapse; and Figure 2.1.1-10shows that, in the dynamic event, it takes about 0.2
seconds for this to occur. This implies average gross strain rates of about 1.0 per second in that period
of the response. In typical piping steels such a strain rate might raise the yield stress 5-10% above its
static value. This is not a large effect, so the mitigation of localization by rate effects is probably not a
major aspect of this event. Again, a more precise assessment of this error would require a fully
three-dimensional analysis. Overall it seems likely that this simple and computationally inexpensive
two-stage approach to the problem is providing results that are sufficiently realistic to be used in
design, although it would be most desirable to compare these results with physical experimental data
or data from a full, detailed, three-dimensional analysis to support that statement. Finally, it should be
2-724
noted that the section considered here is relatively thick ( R=t =3.5). In pipes with thin walls
(R=t >20) it is to be expected that the behavior will be affected strongly by internal fluid pressure in
the pipe and by the interaction between axial and bending forces. Such thin-walled pipes could be
modeled at relatively low cost by using ELBOW elements directly in the dynamic analysis instead of
this two-stage approach. An additional concern with very thin pipes is that they are more likely to tear
and leak, rather than choke the flow.
Input files
nonlindyncollapse_cge10r.inp
Static analysis of the elastic-plastic collapse of the pipe section using CGPE10R elements.
nonlindyncollapse_nonlingsect.inp
Dynamic analysis of the inelastic pipe whip response using nonlinear beam general section
definitions for the axial and bending behaviors of the pipe.
nonlindyncollapse_cgpe6i.inp
Static analysis using element type CGPE6I.
nonlindyncollapse_cgpe6r.inp
Static analysis using element type CGPE6R.
nonlindyncollapse_postoutput1.inp
nonlindyncollapse_postoutput2.inp
Figures
Figure 2.1.1-1 Elastic-plastic pipe subjected to rupture force.
2-725
Figure 2.1.1-2 Initially straight pipe collapsing under pure bending; generalized plane strain model.
2-726
Figure 2.1.1-3 Deformed shapes of collapsing pipe section under pure bending, element type
CGPE10R.
2-727
Figure 2.1.1-4 Deformed shapes of collapsing pipe section under pure bending, element type
CGPE6R.
2-728
Figure 2.1.1-5 Equivalent plastic strain contours in collapsing pipe section, element type CGPE10R.
2-729
Figure 2.1.1-6 Equivalent plastic strain contours in collapsing pipe section, element type CGPE6R.
2-730
Figure 2.1.1-7 Moment-curvature response predicted for collapsing section under pure bending.
2-731
Figure 2.1.1-8 Displaced positions of pipe, every 20 increments. Initial increments in top figure, final
increments in bottom figure.
2-732
Figure 2.1.1-9 Tip displacement history.
2-733
Figure 2.1.1-10 Curvature-time history for the element at the support.
Figure 2.1.1-11 Moment versus curvature in the element at the support.
2-734
Figure 2.1.1-12 Energy history for the beam.
Sample listings
2-735
Listing 2.1.1-1
*HEADING
COLLAPSING PIPE SECTION
*NODE
1,
2,
*NODE,NSET=SYM
101,0.,-28.55
201,0.,-30.925
301,0.,-33.3
401,0.,-35.675
501,0.,-38.05
117,0.,28.55
217,0.,30.925
317,0.,33.3
417,0.,35.675
517,0.,38.05
*NGEN,LINE=C
101,117,1,1,,,,0.,0.,1.
201,217,2,1,,,,0.,0.,1.
301,317,1,1,,,,0.,0.,1.
401,417,2,1,,,,0.,0.,1.
501,517,1,1,,,,0.,0.,1.
*NSET,NSET=MON
2,
*ELEMENT,TYPE=CGPE10R
11, 101,301,303,103,201,302,203,102,1,2
*ELGEN,ELSET=ALL
11,2,200,1
11,8,2,10
12,8,2,10
*ELSET,ELSET=PRINT
11,12,41,42,51,52,81,82
*SOLID SECTION,MATERIAL=MAT,ELSET=ALL
*MATERIAL,NAME=MAT
*ELASTIC
2.08E5,.3
*PLASTIC
316.,0.
324.,.02
388.,.04
431.,.06
2-736
464.,.08
492.,.1
547.,.15
589.,.2
624.,.25
654.,.3
705.,.4
747.,.5
784.,.6
816.,.7
845.,.8
871.,.9
895.,1.
*BOUNDARY
SYM,XSYMM
2,2
509,2
*STATIC,RIKS
.01,1.,.01, , ,2,1,.03
*BOUNDARY
2,1,1,.05
U,RF
*NODE FILE,NSET=MON
U,RF
*EL PRINT,ELSET=PRINT,FREQUENCY=5
E,
PE,
*EL FILE,ELSET=PRINT,FREQUENCY=5
E,PE
*END STEP
2-737
Listing 2.1.1-2
*HEADING
INELASTIC NONLINEAR BEAM GENERAL SECTION DYNAMICS
*NODE
1,
*NODE,NSET=N11
11,3.
*NGEN
1,11
*ELEMENT,TYPE=B21
1,1,2
*ELGEN,ELSET=EALL
1,10,1,1
*BEAM GENERAL SECTION,DENSITY=7827.,
SECTION=NONLINEAR GENERAL,ELSET=EALL
1.9877E-3,1.1245E-6,0.,1.1245E-6,2.2490E-6
0.,0.,-1.
*AXIAL
0.,0.
6.48E5,.02
7.46E5,.04
8.12E5,.06
8.57E5,.08
8.90E5,.1
*M1
0.,0.
1.35E4,.4
1.43E4,1.
1.60E4,1.6
1.73E4,2.2
1.80E4,3.
1.90E4,4.
1.80E4,5.
1.15E4,10.2
.90E4,13.
.75E4,16.2
.65E4,21.
.58E4,30.
*BOUNDARY
1,1,6
*AMPLITUDE,NAME=P,TIME=STEP TIME
0.,1.,.06,1.,.0601,0.
2-738
*DYNAMIC,HAFTOL=30.E3
.001,.4
*CLOAD,FOLLOWER,AMPLITUDE=P
11,2,-30000.
*PRINT,RESIDUAL=NO
SF,
SPE,SEPE
SF,
SPE,SEPE
*NODE FILE,NSET=N11
U,CF
*ENERGY FILE
*END STEP
2.1.2 Detroit Edison pipe whip experiment

This example is a model of a simple, small-displacement pipe whip experiment conducted by the
Detroit Edison Company and reported by Esswein et al. (1978). The problem involves rather small
displacements but provides an interesting case because some (limited) experimental results are
available. It is a typical pipe whip restraint design case. It is a rather straightforward analysis because
the restraint limits the motion and the geometry is so simple.
Geometry and model

The geometry and loading are shown in Figure 2.1.2-1. The pipe has a straight run of length 2.286 m
(90 in), a very stiff elbow, and a cantilever "stick" 482.6 mm (19 in) long. A bursting diaphragm is
installed at the end of the "stick" to initiate the blowdown. The restraint is a set of three U-bolts
coupled together. The blowdown force history measured in the experiment is also shown in Figure
2.1.2-1. All dimensions, material properties, and this force history are taken from Esswein et al.
(1978).
The horizontal pipe run is modeled with eight elements of type B23 (cubic interpolation beam with
planar motion), and the stick is modeled with two elements of the same type. The elbow is treated as a
fully rigid junction, so the node at the elbow is shared between the two branches. The bursting
diaphragm structure is modeled as a lumped mass of 106.8 kg (0.61 lb s 2/in). The restraint is modeled
as a single truss element. For the pipe the Young's modulus is 207 GPa (30 ´ 106 lb/in2), the initial
yield stress is 214 MPa (31020 lb/in 2), and the work hardening modulus is 846 "0.2 MPa (122700"0.2
lb/in2) after yield. The restraint has an elastic stiffness of 131.35 MN/m (750000 lb/in), a yield force of
2-739
16681 N (3750 lb), and--when yielding--a force-displacement response F = 2.2716 ± 0.235 MN/m
(12971± 0.235 lb/in). These values are taken from Esswein et al. (1978), where it is stated that they are
based on measurements of static values with the stresses and forces increased by 50% in the plastic
range to account for strain-rate effects. When it is known that strain-rate effects are important to the
response it is preferable to model them directly, using the *RATE DEPENDENT suboption of the
*PLASTIC option. This has not been done in this case because the actual material is not specified.
Isotropic hardening is assumed for both the pipe and the restraint since the plastic flows are presumed
to be in the large flow regime and not just incipient plasticity (where the Bauschinger effect can be
important). The cross-section of the pipe is integrated with a seven-point Simpson rule: this should be
of sufficient accuracy for this problem. Generally, in beam-like problems without repeated large
magnitude excitation, a higher-order integration scheme would show only significantly different results
at late times in the response, and then the differences are not too important in models of this rather
unrefined level.
Esswein et al. (1978) provide the blowdown force-time history shown in Figure 2.1.2-1. This is
applied as a point load at the end of the stick. In reality the fluid force during blowdown occurs at the
piping elbows; but, since the displacements remain small, this detail is not important.
Solution control
Automatic time stepping is used, with an initial time increment of 100 ¹sec and the value of the
half-step residual, HAFTOL on the *DYNAMIC option, set to 4448 N (1000 lb). This value is based
on actual force values expected (in this case, the blowdown force): HAFTOL is chosen to be about
10% of peak real forces. This should give good accuracy in the dynamic integration.

The displacement of the node that hits the restraint is shown in Figure 2.1.2-2, and the force between
the pipe and restraint is shown in Figure 2.1.2-3. Some experimental results from Esswein et al. (1978)
are shown in Figure 2.1.2-3.
The analysis appears to predict the closure time and the peak force between the pipe and restraint quite
well. However, the numerical solution (like the numerical solution given by Esswein et al., 1978)
shows a slower force rise time than the experiment. A possible explanation may be the material model,
where viscoplastic (strain-rate-dependent yield) effects have been modeled as enhanced yield values,
as discussed above: this means that, at the high strain rate that occurs just after impact, the actual
material can carry higher stresses than the model, and so will respond more stiffly. The oscillation in
the gap force in Figure 2.1.2-3after the initial loading of the restraint is presumably caused by the
difference in the basic natural frequencies of the restraint and the pipe: this oscillation is sufficiently
severe to cause two slight separations.
Input files
detroitedison.inp
2-740
detroitedison_postoutput.inp
Reference
· Esswein, G., S. Levy, M. Triplet, G. Chan, and N. Varadavajan, Pipe Whip Dynamics, ASME
Special Publication, 1978.
Figures
Figure 2.1.2-1 Detroit Edison experiment.
Figure 2.1.2-2 Displacement history at constrained end.
2-741
Figure 2.1.2-3 Gap force history.
2-742
Sample listings
2-743
Listing 2.1.2-1
*HEADING
DECO EXAMPLE 10
*WAVEFRONT MINIMIZATION, SUPPRESS
*NODE
1, 97.,19.
2,97.,9.
3,97.
4,90.
5,78.
6,66.
7,54.
8,42.
9,30.
10,15.
11,0.
12,90.,-3.18
13,90.,997.
*NSET,NSET=NFIL
3,4
*ELEMENT,TYPE=B23
1,1,2
*ELGEN,ELSET=BEAM
1,10
*ELEMENT,TYPE=T3D2,ELSET=TRUS
11,12,13
*SOLID SECTION, ELSET=TRUS, MATERIAL=A2
1.,
*MATERIAL,NAME=A2
*DENSITY
1.174E-6,
*ELASTIC
7.5E8,
3741.,0.
1.1046E4,4.85E-4
1.3E4,9.83E-4
1.53E4,1.98E-3
1.8E4,3.98E-3
*BEAM SECTION,SECTION=PIPE,ELSET=BEAM,MATERIAL=A1
2.25,.337
*MATERIAL,NAME=A1
2-744
*DENSITY
7.338E-4,
*ELASTIC
30.E6,
3.102E4,0.
4.252E4,3.58E-3
6.74E4,4.775E-2
8.893E4,.197
1.227E5,.9959
*ELEMENT,TYPE=MASS,ELSET=MASS
21,1
*MASS,ELSET=MASS
.61,
*ELEMENT,TYPE=GAPUNI,ELSET=GAP1
31,4,12
*GAP,ELSET=GAP1
3.181,0.,-1.,0.
*AMPLITUDE,NAME=P
0.,7200.,.00045,7200.,.00046,7272.,.0291,7272.
.0292,7920.,.1,7920.
*BOUNDARY
11,1,2
11,6
12,1
12,3
13,1,3
APPLY CONCENTRATED FORCE
*DYNAMIC,HAFTOL=1000.
100.E-6,.1
*CLOAD,AMPLITUDE=P
1,2,-1.
*NODE PRINT,NSET=NFIL,FREQUENCY=10
U,
*EL PRINT,FREQUENCY=10,ELSET=MASS
ELEN,
*PRINT,RESIDUAL=NO,FREQUENCY=5
*EL PRINT,ELSET=TRUS,FREQUENCY=20
S,
2-745
E,
PE,PEEQ
U,
V,
A,
CF,
RF,
*EL PRINT,ELSET=BEAM,FREQUENCY=80
S,
SF,
E,
PE,PEEQ
*NODE FILE,NSET=NFIL
U,
*EL FILE,ELSET=GAP1
S,E
*EL FILE,ELSET=MASS
ELEN,
*END STEP
2.1.3 Plate impact simulation

This example simulates the oblique impact of a rigid spherical projectile onto a flat armor plate at a
velocity of 1000 m/sec. A failure model is used for the plate, thus allowing the projectile to perforate
the plate. The example illustrates impact, shear failure, and the use of infinite elements.
Problem description
The armor plate has a thickness of 10 mm and is assumed to be semi-infinite in size compared to the
projectile. This is accomplished by using CIN3D8 infinite elements around the perimeter of the plate.
The plate is modeled using 4480 C3D8R elements. The armor plate material has Young's modulus of
206.8 GPa, Poisson's ratio of 0.3, density of 7800 kg/m 3, yield stress of 1220 MPa, and a constant
hardening slope of 1220 MPa. The material definition also includes a shear failure model, which
causes ABAQUS/Explicit to remove elements from the mesh as they fail. Failure is assumed to occur
at an equivalent plastic strain of 100%, at which point the element is removed from the model
instantaneously. (The value of the failure strain is chosen somewhat arbitrarily; it is not intended to
model any particular material.)
The sphere has a diameter of 20 mm and is assumed to be rigid, with a mass corresponding to a
uniform material with a density of 37240 kg/m 3. The rotary inertia of the sphere is not needed in the
model because we assume there is no friction between the sphere and the plate. Boundary conditions
are applied to constrain the motion of the sphere in the y-direction. Two approaches for modeling the
2-746
surface of the sphere are tested: using an analytical rigid surface and using R3D4 rigid elements.
Analytical rigid surfaces are the preferred means for representing simple rigid geometries such as this
in terms of both accuracy and computational performance. However, more complex three-dimensional
surface geometries that occur in practice must be modeled with surfaces formed by element faces.
Results for the faceted representations are presented here. The element formulation for the C3D8R
elements is modified with the *SECTION CONTROLS option. The advocated formulation for this
problem uses the CENTROID kinematic formulation and the COMBINED hourglass control. Two
other formulations are included for comparison: the default formulation using the AVERAGE STRAIN
kinematic formulation and the RELAX STIFFNESS hourglass control and the formulation using the
ORTHOGONAL kinematic formulation and the COMBINED hourglass control.
Only half of the plate is modeled, using appropriate symmetry boundary conditions in the x-z plane.
The model is shown in Figure 2.1.3-1. The complete sphere is modeled for visualization purposes.
There are 17094 degrees of freedom in the model.
Since elements in the plate will fail and be removed from the model, nodes in the interior of the plate
will be exposed to contact with the surface of the rigid sphere. Thus, contact must be modeled between
the surface of the sphere, defined as an element-based surface using the *SURFACE,
TYPE=ELEMENT option, and a node-based surface that contains all of the nodes in the plate within a
radius of 20 mm of the point of impact, defined with the *SURFACE, TYPE=NODE option. The
*CONTACT PAIR option is used to define contact between the surface of the sphere and any of the
nodes contained in the node set.

The spherical projectile impacts the plate at 1000 m/sec at an angle of 30° to the normal to the plate.
Deformed shapes at different stages of the analysis are shown in Figure 2.1.3-2through Figure 2.1.3-4
for the CENTROID kinematic and COMBINED hourglass section control options (analysis case
pl3d_erode_ccs). Early in the analysis, shown in Figure 2.1.3-2, a relatively small amount of material
has been eroded from the surface of the plate and the plate is still deforming under the sphere. In
Figure 2.1.3-3the plate has been perforated and the projectile is still in contact with the edge of the
hole. In Figure 2.1.3-4 the projectile has exited the plate and is moving as a rigid body. Figure
2.1.3-5and Figure 2.1.3-6 show the history of the projectile's velocity (Table 2.1.3-1 shows the analysis
options used to obtain these results for analyses using different section controls). The analysis
pl3d_erode_ccs shows comparable accuracy to the other analyses at a much reduced computational
cost; however, strain and stress values vary more than the primary variables (velocity, displacement,
etc.).
In Figure 2.1.3-2 through Figure 2.1.3-4the failed elements have been eliminated by creating a display
group in ABAQUS/Viewer that contains only the active elements.
Input files
pl3d_erode_ccs.inp
Model using the CENTROID kinematic and COMBINED hourglass section control options.
2-747
sphere_n.inp
External file referenced in this input.
sphere_e.inp
pl3d_erode.inp
Model using the default section controls.
pl3d_erode_ale.inp
Model using the default section controls and the *ADAPTIVE MESH option.
pl3d_erode_ocs.inp
Model using the ORTHOGONAL kinematic and COMBINED hourglass section control options.
pl3d_erode_anl.inp
Model using an analytical rigid surface and the default section controls.
Table
Table 2.1.3-1 Analysis options tested.

Analysis File Relative Section Controls
CPU Time Kinematic Hourglass
pl3d_erode 1.0 average relax
pl3d_erode_oc 0.70 orthogonal combined
s
pl3d_erode_cc 0.57 centroid combined
s
Figures
Figure 2.1.3-1 Undeformed mesh.
2-748
Figure 2.1.3-2 Deformed shape at 10 microseconds (analysis using the following section control
options: CENTROID kinematics and COMBINED hourglass control).
2-749
Figure 2.1.3-5 Vertical component of the projectile velocity.
Figure 2.1.3-6 Horizontal component of the projectile velocity.
2-750
Sample listings
2-751
Listing 2.1.3-1
*HEADING
RIGID SPHERE IMPACTING ON THICK PLATE WITH
PENETRATION AND ELEMENT EROSION
SECTION CONTROLS USED (KINEMA=CENTROID,
HOURGLASS=COMBINED)
*PREPRINT,ECHO=NO,MODEL=NO,HISTORY=NO
*NODE
1,-.0040, 0.0,0.0
9, .0040, 0.0,0.0
37,-.0031,.0031,0.0
41, 0.0,.0040,0.0
45, .0031,.0031,0.0
46, .0040, 0.0,0.0
50, .0031,.0031,0.0
54, 0.0,.0040,0.0
58,-.0031,.0031,0.0
62,-.0040, 0.0,0.0
318, .02,0.0,0.0
334,-.02,0.0,0.0
488, .04,0.0,0.0
504,-.04,0.0,0.0
505, .05,0.0,0.0
521,-.05,0.0,0.0
*NGEN,NSET=A
1,9,1
*NGEN,NSET=B
37,41,1
41,45,1
*NGEN,NSET=C
46,50,1
50,54,1
54,58,1
58,62,1
*NGEN,NSET=D,LINE=C
318,334,1,,0.,0.,0.,0.,0.,1.
*NGEN,NSET=E,LINE=C
488,504,1,,0.,0.,0.,0.,0.,1.
*NGEN,NSET=TOP,LINE=C
505,521,1,,0.,0.,0.,0.,0.,1.
*NFILL,NSET=TOP
A,B,4,9
2-752
C,D,16,17
*NFILL,BIAS=.9,NSET=TOP
D,E,10,17
*NCOPY,OLD SET=TOP,SHIFT,CHANGE NUMBER=1000
0.,0.,-.001
0.,0.,0.,0.,0.,1.,0.0
0.,0.,-.002
0.,0.,0.,0.,0.,1.,0.0
0.,0.,-.003
0.,0.,0.,0.,0.,1.,0.0
0.,0.,-.004
0.,0.,0.,0.,0.,1.,0.0
0.,0.,-.005
0.,0.,0.,0.,0.,1.,0.0
0.,0.,-.006
0.,0.,0.,0.,0.,1.,0.0
0.,0.,-.007
0.,0.,0.,0.,0.,1.,0.0
0.,0.,-.008
0.,0.,0.,0.,0.,1.,0.0
0.,0.,-.009
0.,0.,0.,0.,0.,1.,0.0
0.,0.,-.010
0.,0.,0.,0.,0.,1.,0.0
**
*NSET,NSET=YSYM,GEN
1, 9, 1
46, 488,17
62, 504,17
1001,1009,1
1046,1488,17
1062,1504,17
2001,2009,1
2046,2488,17
2-753
2062,2504,17
3001,3009,1
3046,3488,17
3062,3504,17
4001,4009,1
4046,4488,17
4062,4504,17
5001,5009,1
5046,5488,17
5062,5504,17
6001,6009,1
6046,6488,17
6062,6504,17
7001,7009,1
7046,7488,17
7062,7504,17
8001,8009,1
8046,8488,17
8062,8504,17
9001,9009,1
9046,9488,17
9062,9504,17
10001,10009,1
10046,10488,17
10062,10504,17
*NSET,NSET=FRONT,GEN
1, 385,1
**
*ELEMENT,TYPE=C3D8R,ELSET=TOP
1, 1002,1003,1012,1011, 2, 3,12,11
19, 1008,1046,1047,1017, 8,46,47,17
20, 1017,1047,1048,1026, 17,47,48,26
21, 1026,1048,1049,1035, 26,48,49,35
22, 1035,1049,1050,1051, 35,49,50,51
23, 1034,1035,1051,1052, 34,35,51,52
24, 1033,1034,1052,1053, 33,34,52,53
25, 1032,1033,1053,1054, 32,33,53,54
26, 1031,1032,1054,1055, 31,32,54,55
27, 1030,1031,1055,1056, 30,31,55,56
28, 1029,1030,1056,1057, 29,30,56,57
29, 1059,1029,1057,1058, 59,29,57,58
30, 1060,1020,1029,1059, 60,20,29,59
31, 1061,1011,1020,1060, 61,11,20,60
2-754
32, 1062,1002,1011,1061, 62, 2,11,61

33, 1046,1063,1064,1047, 46,63,64,47
*ELGEN,ELSET=TOP
1, 6,1,1, 3, 9, 6
33, 16,1,1, 26,17,16
*ELEMENT,TYPE=CIN3D8,ELSET=TINF
449, 489,488,1488,1489, 506,505,1505,1506
*ELGEN,ELSET=TINF
449, 16,1,1
**
*ELCOPY,OLD SET=TOP,NEW SET=PLATE,SHIFT NODES=1000,
ELEMENT SHIFT=1000
ELEMENT SHIFT=2000
ELEMENT SHIFT=3000
ELEMENT SHIFT=4000
ELEMENT SHIFT=5000
ELEMENT SHIFT=6000
ELEMENT SHIFT=7000
ELEMENT SHIFT=8000
ELEMENT SHIFT=9000
*ELSET,ELSET=PLATE
TOP,
**
*ELCOPY,OLD SET=TINF,NEW SET=INF,SHIFT NODES=1000,
ELEMENT SHIFT=1000
ELEMENT SHIFT=2000
ELEMENT SHIFT=3000
ELEMENT SHIFT=4000
ELEMENT SHIFT=5000
ELEMENT SHIFT=6000
2-755

ELEMENT SHIFT=7000
ELEMENT SHIFT=8000
ELEMENT SHIFT=9000
*ELSET,ELSET=INF
TINF,
**
*SOLID SECTION,ELSET=PLATE,MATERIAL=RHA,CONTROL=B
*SOLID SECTION,ELSET=INF, MATERIAL=INF,CONTROL=B
*SECTION CONTROLS, KINEMA=CENTROID,
HOURGLASS=COMBINED, NAME=B
**
** Sphere with radius of .01 m.
**
*NODE,NSET=SPHERE,INPUT=sphere_n.inp
*ELEMENT,TYPE=R3D4,ELSET=SPHERE,
INPUT=sphere_e.inp
*ELEMENT,TYPE=MASS,ELSET=MASS
50000,599991
*MASS,ELSET=MASS
** Mass based on a density of 37240 Kg/m^3
** and radius of .01 m.
.078,
**
*MATERIAL,NAME=RHA
*DENSITY
7800.,
*ELASTIC
207.8E9,.3
*PLASTIC
1220.E6,0.
2440.E6,1.
*SHEAR FAILURE
1.0,
*MATERIAL,NAME=INF
*DENSITY
7800.,
*ELASTIC
206.8E9,.3
599991,1, 500.
2-756
599991,3,-866.6
*BOUNDARY
YSYM,2,2
599991, 2
599991, 4,6
*SURFACE,TYPE=ELEMENT,NAME=SPHERE
SPHERE,SPOS
*NSET,NSET=ERODE_NSET,GEN
1, 385,1
1001,1385,1
2001,2385,1
3001,3385,1
4001,4385,1
5001,5385,1
6001,6385,1
7001,7385,1
8001,8385,1
9001,9385,1
10001,10385,1
*SURFACE,TYPE=NODE,NAME=ERODE
ERODE_NSET,
*RIGID BODY,ELSET=SPHERE,REF NODE=599991
*STEP
*DYNAMIC,EXPLICIT
,40.E-6
*CONTACT PAIR
SPHERE,ERODE
*NSET, NSET=N1
599991,
*NODE HISTORY, NSET=N1
V,
*FILE OUTPUT, NUM=2, TIMEMARKS=YES
*NODE FILE
U,V
*EL FILE
S,PEEQ
STATUS,
*EL FILE
ELEN,
*END STEP
2-757
2.1.4 Tennis racket and ball

This example simulates the oblique impact of a tennis ball onto a racket at 6.706 m/sec (264 in/sec).
The example illustrates contact between a deforming surface and a node set, the definition of initial
stresses via the *INITIAL CONDITIONS option, and modeling of the compressible gas inside the ball
with hydrostatic fluid cavity elements.
Problem description
The strings on the tennis racket are modeled using T3D2 truss elements. They are assumed to be linear
elastic, with Young's modulus of 6.895 GPa (1.0 ´ 106 psi), Poisson's ratio of 0.3, and density of 1143
kg/m3 (1.07 ´ 10-4 lb sec2in-4). The strings are under an initial tension of 44.48 N (10 lb), which is
specified with the *INITIAL CONDITIONS option.
The frame is assumed to be rigid and is modeled using R3D4 elements. The nodes of the strings (truss
elements) around the perimeter are the same nodes as those used for the R3D4 elements. The reference
node for the rigid frame has boundary conditions applied to constrain all six degrees of freedom on the
rigid body so that the frame does not move.
The tennis ball is modeled as a sphere, using 150 S4R shell elements. It is assumed to be made of
rubber, modeled with the *HYPERELASTIC option as a Mooney-Rivlin material with the constants
C10 = 0.690 MPa (100 lb/in 2) and C01 = 0.173 MPa (25 lb/in 2). ABAQUS/Explicit requires some
compressibility for hyperelastic materials. In the results shown here, D1 = 0.0145 MPa -1 (10-4 psi-1).
This gives an initial bulk modulus ( K0 = 2=D1 ) that is 80 times the initial shear modulus
2(C10 + C01 ). This ratio is lower than the ratio for typical rubbers, but the results are not particularly
sensitive to this value in this case because the rubber is unconfined. A more accurate representation of
the material's compressibility would be needed if the rubber were confined by stiffer adjacent
components or reinforcement. Decreasing D1 by an order of magnitude (thus increasing the initial bulk
modulus by a factor of 10) has little effect on the overall results but causes a reduction in the stable
p
time increment by a factor of 10 due to the increase in the bulk modulus. The density of the tennis
ball is 1068 kg/m 3 (1.07 ´ 10-4 lb sec2in-4).
The tennis ball is under an initial internal pressure of 41 kPa (6 psi) in addition to the ambient
atmospheric pressure of 100 kPa (14.7 psi). Hydrostatic fluid elements of type F3D4 are used to model
the gas in the tennis ball. Since the ball is impermeable to gas, the pressure of the gas will rise when
the volume of the ball decreases, and vice versa. The fluid density is arbitrarily chosen to be one-tenth
of that of rubber under an ambient pressure of 100 kPa (14.7 psi). Static equilibrium gives the value of
the initial biaxial membrane stresses in the shell elements of the sphere as pr=2t = 155 kPa (22.5 psi)
to balance the internal pressure (here p is the internal gas pressure, r is the radius of the sphere, and t
is the tennis ball thickness). This initial state of stress in the ball is defined with the *INITIAL
CONDITIONS option.
A coefficient of friction of 0.1 is specified between the ball and the strings. The ball impacts on the
strings at 6.706 m/sec (264 in/sec) at an angle of 15°.
2-758
No attempt has been made to generate an accurate model of the ball and strings: the model parameters
are chosen simply to provide a "soft" ball relative to the strings to illustrate contact effects.
The complete model is shown in Figure 2.1.4-1. There are 2241 degrees of freedom in the model.
An element-based surface is defined on the tennis ball using the *SURFACE, TYPE=ELEMENT
option. Since the truss elements are line elements, they do not form a planar surface. A node-based
surface is defined that contains all the nodes of the strings using the *SURFACE, TYPE=NODE
option. The *CONTACT PAIR option is then used to define contact between the element-based
surface of the ball and any of the nodes defined in the node-based surface.

Figure 2.1.4-2 shows the position of the ball with respect to the strings in the undeformed
configuration. The deformed shapes at different stages of the analysis are shown in Figure
2.1.4-3through Figure 2.1.4-7. Figure 2.1.4-8shows a time history of the energies for the model. These
include the total internal energy (ALLIE), the kinetic energy (ALLKE), the viscous dissipation
(ALLVD), the energy dissipated by friction (ALLFD), the external work (ALLWK), and the total
energy balance for the model (ETOTAL). The total energy is seen to remain almost constant during the
analysis, as it should. Figure 2.1.4-9and Figure 2.1.4-10 give the history of pressure inside the ball and
the history of the actual volume of the ball. It can be seen that both the gas pressure inside the ball and
the ball volume stabilize after 10 msec.
Input files
tennis.inp
Input data used in this analysis.
tennis_ef1.inp
tennis_ef2.inp
tennis_ef3.inp
Figures
Figure 2.1.4-1 Undeformed mesh.
2-759
Figure 2.1.4-2 Original position of ball and strings.
Figure 2.1.4-3 Deformed shape at 2.5 milliseconds.
Figure 2.1.4-4 Deformed shape at 5 milliseconds.
2-760
Figure 2.1.4-5 Deformed shape at 7.5 milliseconds.
2-761
Figure 2.1.4-8 Energy histories.
Figure 2.1.4-9 History of the gas pressure inside the tennis ball.
2-762
Figure 2.1.4-10 History of the ball volume.
Sample listings
2-763
Listing 2.1.4-1
*HEADING
RACKET AND BALL IMPACT
********************
**
** Strings.
**
********************
*NODE
** Bottom curve.
104, -2.700,-6.625,0.
109, 0.000,-8.500,0.
114, 2.700,-6.625,0.
** Left side curve.
501, -4.020,-3.750,0.
1201, -4.520, 0.000,0.
1901, -3.780, 3.750,0.
** Right side curve.
517, 4.020,-3.750,0.
1217, 4.520, 0.000,0.
1917, 3.780, 3.750,0.
** Top curve.
2103, -3.240, 4.500,0.
2109, 0.000, 5.500,0.
2115, 3.240, 4.500,0.
** Rectangular portion.
202, -3.780,-6.625,0.
216, 3.780,-6.625,0.
502, -3.780,-3.750,0.
516, 3.780,-3.750,0.
1902, -3.780, 3.750,0.
1916, 3.780, 3.750,0.
2002, -3.780, 4.500,0.
2016, 3.780, 4.500,0.
*NGEN,NSET=N200
202,216,1
*NGEN,NSET=N500
502,516,1
*NGEN,NSET=N1900
1902,1916,1
*NGEN
2002,2016,1
2-764
*NFILL,NSET=STRINGS
N200,N500,3,100
N500,N1900,14,100
*NGEN,NSET=STRINGS,LINE=P
104,114,1,109
517,1917,100,1217
501,1901,100,1201,0.,0.,1.
2103,2115,1,2109,0.,0.,1.
*NSET,NSET=MIDPLANE,GEN
105,113,1
2104,2114,1
501,1801,100
517,1817,100
*NSET,NSET=MIDPLANE
204,303,402,214,315,416
1902,2003,1916,2015
**
*ELEMENT,TYPE=T3D2,ELSET=STRINGS
1, 204, 205
11, 303, 304
23, 402, 403
247,2003,2004
1001, 402, 502
1016, 303, 403
1033, 204, 304
1231, 315, 415
1248, 416, 516
2001, 501, 502
2015, 516, 517
2029, 105, 205
2038,2004,2104
*ELGEN,ELSET=STRINGS
1, 10,1,1
11, 12,1,1
23, 14,1,1, 16,100,14
247, 12,1,1
1001, 15,100,1
1016, 17,100,1
1033, 18,100,1, 11,1,18
1231, 17,100,1
1248, 15,100,1
2001, 14,100,1
2015, 14,100,1
2-765
2029, 9,1,1
2038, 11,1,1
*SOLID SECTION,ELSET=STRINGS,MATERIAL=STRING
** Diameter of strings is 0.1 in.
7.854E-3,
**
*NSET, NSET=NCONTACT, GENERATE
205,213,1
304,314,1
403,415,1
502,516,1
602,616,1
702,716,1
802,816,1
902,916,1
1002,1016,1
1102,1116,1
1202,1216,1
1302,1316,1
1402,1416,1
1502,1516,1
1602,1616,1
1702,1716,1
1802,1816,1
1903,1915,1
2004,2014,1
********************
**
** Frame.
**
********************
*NCOPY,CHANGE NUMBER=10000,OLD SET=MIDPLANE,SHIFT,
NEW SET=FRONT
0.,0.,.25
0.,0.,0., 0.,0.,1., 0.0
*NCOPY,CHANGE NUMBER=20000,OLD SET=MIDPLANE,SHIFT,
NEW SET=BACK
0.,0.,-.25
0.,0.,0., 0.,0.,1., 0.0
*NODE
10000, 0,0.,0.
*ELEMENT,TYPE=R3D4,ELSET=FRAME
3001, 105, 106, 10106, 10105
2-766
3009, 113, 214, 10214, 10113

3010, 214, 315, 10315, 10214
3011, 315, 416, 10416, 10315
3012, 416, 517, 10517, 10416
3013, 517, 617, 10617, 10517
3036, 1817, 1916, 11916, 11817
3037, 1916, 2015, 12015, 11916
3038, 2015, 2114, 12114, 12015
3039, 2105, 2104, 12104, 12105
3049, 2104, 2003, 12003, 12104
3050, 2003, 1902, 11902, 12003
3051, 1902, 1801, 11801, 11902
3052, 601, 501, 10501, 10601
3065, 501, 402, 10402, 10501
3066, 402, 303, 10303, 10402
3077, 303, 204, 10204, 10303
3078, 204, 105, 10105, 10204
4001, 105, 106, 20106, 20105
4009, 113, 214, 20214, 20113
4010, 214, 315, 20315, 20214
4011, 315, 416, 20416, 20315
4012, 416, 517, 20517, 20416
4013, 517, 617, 20617, 20517
4036, 1817, 1916, 21916, 21817
4037, 1916, 2015, 22015, 21916
4038, 2015, 2114, 22114, 22015
4039, 2105, 2104, 22104, 22105
4049, 2104, 2003, 22003, 22104
4050, 2003, 1902, 21902, 22003
4051, 1902, 1801, 21801, 21902
4052, 601, 501, 20501, 20601
4065, 501, 402, 20402, 20501
4066, 402, 303, 20303, 20402
4077, 303, 204, 20204, 20303
4078, 204, 105, 20105, 20204
**
*ELGEN,ELSET=FRAME
3001, 8,1,1
3013, 13,100,1
3039, 10,1,1
3052, 13,100,1
4001, 8,1,1
4013, 13,100,1
2-767
4039, 10,1,1
4052, 13,100,1
*NODE,NSET=SPHERE,INPUT=tennis_ef1.inp
**
*ELEMENT,TYPE=S4R,ELSET=SPHERE,
INPUT=tennis_ef2.inp
*SHELL SECTION,ELSET=SPHERE,MATERIAL=RUBBER
.2,3
**
*ELEMENT,TYPE=F3D4,ELSET=CAVITY,
INPUT=tennis_ef3.inp
*FLUID PROPERTY,ELSET=CAVITY,REF NODE=30001,
AMBIENT=14.7
*FLUID DENSITY
0.1E-4,
********************
**
** Material definitions.
**
********************
*DENSITY
1.E-4,
*HYPERELASTIC,N=1
100.,25.,1.E-4
*MATERIAL,NAME=STRING
** Nylon type 6 general purpose.
*DENSITY
1.07E-4,
*ELASTIC
1.E6,
********************
**
** Initial conditions, boundary conditions.
**
********************
*INITIAL CONDITIONS,TYPE=STRESS
** Tension in strings is 10 lb.
STRINGS,1273.
SPHERE,22.5,22.5
** Initial velocity of ball is 22 fps at a
2-768
** 15 degree angle of attack.

SPHERE, 1, 68.3
SPHERE, 2, 0.
SPHERE, 3,-255.
*BOUNDARY
10000,1,6
*INITIAL CONDITIONS,TYPE=FLUID PRESSURE
30001,6.
**
********************
**
*SURFACE,TYPE=ELEMENT,NAME=SPHERE
SPHERE,SPOS
*SURFACE,TYPE=NODE,NAME=STRINGS
NCONTACT,
*RIGID BODY,ELSET=FRAME,REF NODE=10000
*STEP
*DYNAMIC,EXPLICIT
,15.E-3
********************
**
** Contact definitions.
**
********************
*SURFACE INTERACTION,NAME=SPH_STRING
*FRICTION
0.1,
*CONTACT PAIR,INTERACTION=SPH_STRING
SPHERE,STRINGS
********************
**
** Output requests.
**
********************
*FILE OUTPUT, NUM=2, TIMEMARKS=YES
*EL FILE
ELEN,
S,LE,ERV
*NODE FILE
U,V
*OUTPUT,FIELD,OP=NEW,NUMBER INTERVAL=5,
TIMEMARKS=NO
2-769
*ELEMENT OUTPUT
S,
*NODE OUTPUT
U,
*OUTPUT,HISTORY,OP=NEW,TIME INTERVAL=3.E-5
*ENERGY OUTPUT
*END STEP
2.1.5 Pressurized fuel tank with variable shell thickness

This problem demonstrates the variable shell thickness capability in ABAQUS. The example is based
on an analysis conducted by SOLVAY RESEARCH & TECHNOLOGY (see reference) of a
blow-molded, plastic fuel tank with dimensions similar to those considered here.
Geometry and model

The mesh shown in Figure 2.1.5-1 is used in this example to model a fuel tank and its support straps.
The mesh uses 2812 3-node shell elements ( S3R), with the support straps modeled with 32 2-node
beam elements (B31). Depending on the desired accuracy and detail of the solution, the analyst may
identify some regions of the mesh in which additional refinement, or second-order elements, would be
appropriate. The fuel tank would fit within a box of dimensions 450 mm ´ 200 mm ´ 680 mm. An
internal pressure of 7 ´ 10-3 MPa is applied statically to the tank.
Analyses are conducted for a uniform shell thickness of 5 mm and for a spatially varying shell
thickness in the range 1.38 mm to 9.35 mm (see Figure 2.1.5-2), which is a more accurate
representation of the tank. The uniform thickness analysis provides a comparison to judge the effects
of variable thickness. The overall volume of plastic modeled in the variable thickness analysis is about
93% of that in the uniform thickness analysis. For the variable thickness analysis the presence of the
NODAL THICKNESS parameter on the *SHELL SECTION option indicates that the shell thickness is
to be interpolated from nodal values specified with the *NODAL THICKNESS option. For elements
with more than one integration point, this approach results in a thickness that can vary over the
element.
The materials are modeled as isotropic elastic. The plastic fuel tank has a Young's modulus of 0.6 GPa
and a Poisson's ratio of 0.3. The steel support straps have a Young's modulus of 206.8 GPa and a
Poisson's ratio of 0.29. Geometrically nonlinear effects are significant in this example, so the
NLGEOM parameter is included on the *STEP option.

Contour plots of the Mises stress at the inner surface of the fuel tank (section point 1 in the shell
elements) for the variable shell thickness and uniform shell thickness analyses are shown in Figure
2.1.5-3 and Figure 2.1.5-4. The ratio of the maximum Mises stress found in the variable thickness
2-770
analysis to that found in the uniform thickness analysis is 1.5. For the variable thickness analysis, the
maximum Mises stress occurs at a location where the fuel tank skin is relatively thin (see Figure
2.1.5-2 and Figure 2.1.5-3).
Examination of the y-component of displacement shows that the overall expansion of the tank in the
y-direction is about 1.5% greater in the variable thickness analysis.
Input files
pressfueltank_variablethick.inp
Example using variable shell thickness.
pressfueltank_uniformthick.inp
Example using uniform shell thickness.
pressfueltank_node.inp
Nodal coordinate data for both models.
pressfueltank_shellelement.inp
Shell element connectivity data for both models.
pressfueltank_beamelement.inp
Beam element connectivity data for both models.
pressfueltank_shellthickness.inp
Shell thickness data for the variable shell thickness model.
Reference
· SOLVAY RESEARCH & TECHNOLOGY, Plastic Processing Department, Rue de Ransbeek,
310, B-1120 Brussels, Belgium.
Figures
Figure 2.1.5-1 Fuel tank mesh with S3R and B31 elements.
2-771
Figure 2.1.5-2 Shell thickness for variable thickness analysis.
2-772
Figure 2.1.5-3 Mises stress solution for variable shell thickness analysis.
2-773
Figure 2.1.5-4 Mises stress solution for uniform shell thickness analysis.
2-774
Sample listings
2-775
Listing 2.1.5-1
*HEADING
FUEL TANK - VARIABLE THICKNESS
*PREPRINT,MODEL=NO
**
**READ MESH DATA FROM SEPARATE FILES
*NODE, INPUT=pressfueltank_node.inp
*ELEMENT, TYPE=S3R, ELSET=TANK,
INPUT=pressfueltank_shellelement.inp
*ELEMENT, TYPE=B31, ELSET=STRAPS,
INPUT=pressfueltank_beamelement.inp
**
**READ VARIABLE SHELL THICKNESS DATA FROM
**SEPARATE FILE
*SHELL SECTION,ELSET=TANK,MATERIAL=MTANK,
NODAL THICKNESS
5. , 3
*NODAL THICKNESS,
INPUT=pressfueltank_shellthickness.inp
**
*BEAM SECTION,SECTION=RECT,ELSET=STRAPS,
MATERIAL=MSTRAP
40.,1.5
0.,0.,-1.
**
*MATERIAL,NAME=MTANK
*ELASTIC
600., 0.3
**
*MATERIAL,NAME=MSTRAP
*ELASTIC
2.068E+05, 0.29
**
*NSET,NSET=APPUI1
112, 164, 168, 169, 170, 171, 172, 173, 177, 180,
214, 233, 236, 237, 238, 239, 240, 241, 242, 245,
271, 272, 279, 284, 285, 288, 310, 311, 312, 313,
342
*NSET,NSET=APPUI2
1455, 1458, 1459, 1512, 1513, 1514, 1515, 1518,
1519, 1520, 1579, 1580, 1583, 1584, 1585, 1586,
1639, 1640
2-776
*NSET,NSET=ENDSTR
1740, 1743, 1746, 1747
**
*BOUNDARY
ENDSTR,1,6,0.
APPUI1,2,2,0.
APPUI2,2,2,0.
**
*ELSET,ELSET=SAMPLE,GENERATE
200,1400,200
**
**
**
*STEP,NLGEOM
*STATIC
0.1,1.
*DLOAD,OP=NEW
TANK, P, 7.E-3
*EL FILE,ELSET=SAMPLE
STH,
SINV,
*OUTPUT,FIELD,VAR=PRESELECT,FREQ=10
*ELEMENT OUTPUT
STH,
*OUTPUT,FIELD
*ELEMENT OUTPUT,ELSET=SAMPLE
STH,
SINV,
*OUTPUT,HISTORY
STH,
SINV,
*END STEP
2-777
Listing 2.1.5-2
*HEADING
FUEL TANK - UNIFORM THICKNESS
*PREPRINT,MODEL=NO
**
**READ MESH DATA FROM SEPARATE FILES
*NODE, INPUT=pressfueltank_node.inp
*ELEMENT, TYPE=S3R, ELSET=TANK,
INPUT=pressfueltank_shellelement.inp
*ELEMENT, TYPE=B31, ELSET=STRAPS,
INPUT=pressfueltank_beamelement.inp
**
*SHELL SECTION,ELSET=TANK,MATERIAL=MTANK
5. , 3
**
*BEAM SECTION,SECTION=RECT,ELSET=STRAPS,
MATERIAL=MSTRAP
40.,1.5
0.,0.,-1.
**
*MATERIAL,NAME=MTANK
*ELASTIC
600., 0.3
**
*MATERIAL,NAME=MSTRAP
*ELASTIC
2.068E+05, 0.29
**
*NSET,NSET=APPUI1
112, 164, 168, 169, 170, 171, 172, 173, 177, 180,
214, 233, 236, 237, 238, 239, 240, 241, 242, 245,
271, 272, 279, 284, 285, 288, 310, 311, 312, 313,
342
*NSET,NSET=APPUI2
1455, 1458, 1459, 1512, 1513, 1514, 1515, 1518,
1519, 1520, 1579, 1580, 1583, 1584, 1585, 1586,
1639, 1640
*NSET,NSET=ENDSTR
1740, 1743, 1746, 1747
**
*BOUNDARY
ENDSTR,1,6,0.
2-778
APPUI1,2,2,0.
APPUI2,2,2,0.
**
*ELSET,ELSET=SAMPLE,GENERATE
200,1400,200
**
**
**
*STEP,NLGEOM
*STATIC
0.1,1.
*DLOAD,OP=NEW
TANK, P, 7.E-3
*EL FILE,ELSET=SAMPLE
STH,
SINV,
*OUTPUT,FIELD,VAR=PRESELECT,FREQ=10
*ELEMENT OUTPUT
STH,
*OUTPUT,FIELD
STH,
SINV,
*OUTPUT,HISTORY
STH,
SINV,
*END STEP
2.1.6 Modeling of an automobile suspension

This example illustrates the use of JOINTC elements. It is repeated to illustrate the use of connector
elements. JOINTC elements (``Flexible joint element,'' Section 17.4.1 of the ABAQUS/Standard User's
Manual) can be used to model the interaction between two nodes that are almost coincident
geometrically and that represent a joint that has internal stiffness and/or damping. The behavior of the
joint is defined in a local coordinate system, defined by an *ORIENTATION option (`Òrientations,''
Section 2.2.4 of the ABAQUS/Standard User's Manual). This system rotates with the motion of the
first node of the element and may consist of linear or nonlinear springs and dashpots arranged in
parallel, coupling the corresponding components of relative displacement and of relative rotation in the
2-779
joint. This feature can be used to model, for example, a rubber bearing in a car suspension.
In the connector element model the JOINTC elements are replaced by connector elements (see
``Connectors,'' Section 17.1.1 of the ABAQUS/Standard User's Manual) with connection types
CARTESIAN to define the translational behavior and ROTATION to define the rotational behavior.
These connection types allow linear or nonlinear spring and dashpot behavior to be defined in a local
coordinate system that rotates with the first node on the element. Several different connection types
can be used to model the finite rotational response. See ``Connection-type library,'' Section 17.1.3 of
the ABAQUS/Standard User's Manual, for connection types using different finite rotation
parametrizations. In this model the rotation magnitudes are assumed small. Hence, a rotation vector
parametrization of the joint using ROTATION is appropriate.
The primary objective of this example is to verify the accuracy of JOINTC and connector elements in a
structure undergoing rigid rotation motions. A secondary objective of this example is to demonstrate
the use of equivalent rigid body motion output variables in ABAQUS/Standard.
Geometry and model

The structure analyzed is an automobile's left front suspension subassembly (see Figure 2.1.6-1). The
physical components included in the assemblage are the tire, the wheel, the axle (hub), the A-arm
(wishbone), the coil spring, and the frame. The tire is modeled with a JOINTC or connector element; a
curved bar element has been attached for visualization. (The JOINTC or connector element is used
because it is a convenient way of defining the tire's nonlinear stiffness in a local coordinate system.)
The vertical stiffness of the wheel is represented by a beam element. The axle and A-arm are both
modeled with beam elements. The axle is connected to the A-arm by a pin-type MPC in the JOINTC
model or with connection type JOIN in the connector model. The coil spring is modeled by a
SPRINGA element in the JOINTC model or with connection type AXIAL in the connector model. The
automobile frame is represented by a MASS element. The top of the coil spring is connected directly to
the frame, while the A-arm is connected to the frame by two JOINTC elements or two connector
elements with connection types CARTESIAN and ROTATION (representing the A-arm bushings). The
initial position represents a fully weighted vehicle, and the tire and coil spring have a corresponding
initial preload.
The first step in the analysis allows the suspension system to reach equilibrium. The second step
models the tire moving over a bump in the road. The bump is idealized as a triangular shape 100 mm
high by 400 mm long, and the vehicle is assumed to travel at 5 km/hr. A second input file is used to
show the effects of large rotation on the suspension response. This file includes an initial rotation step,
which rigidly rotates the model by 90° about the vertical axis but is otherwise identical to the first
input file. We expect the response from the two analyses to be the same.
In this example we are primarily interested in the equivalent rigid body motion of the A-arm and, in
particular, in the average displacement and rotation.

The results from the connector element models are qualitatively and quantitatively equivalent to the
JOINTC models. Hence, only the JOINTC results are discussed further.
2-780
The vertical displacement histories in Step 2 are shown in Figure 2.1.6-2for the contact point of the tire
with the ground, the wheel center, and the frame. Figure 2.1.6-3 shows a series of overlaid displaced
plots as the tire rolls up the bump.
To verify the behavior of the JOINTC elements with large rotations, the second model rigidly rotates
the entire structure by 90° before applying the bump excitation. Figure 2.1.6-4shows the displacement
time histories from the two models overlaid on the same plot. They are nearly identical.
In the analysis of deformable bodies undergoing large motions it is convenient to obtain information
about the equivalent rigid body motions: average displacement and rotation, as well as linear and
angular momentum about the center of mass. For this purpose ABAQUS provides a set of equivalent
rigid body output variables. As indicated above, a secondary objective of this example is to
demonstrate the use of these output variables, which represent the average motion of the specified
element set. The variables are requested using the *EL PRINT and *EL FILE commands. If no element
set is specified, the average motion of the entire model is given. This type of output can only be
requested in a *DYNAMIC analysis, and only elements that have a mass will contribute to the
equivalent rigid body motion. For a precise definition of the equivalent rigid body motion of a
deformable body, see `Èquivalent rigid body dynamic motion,'' Section 2.4.4 of the ABAQUS Theory
Manual.
Figure 2.1.6-5 and Figure 2.1.6-6have been generated using the equivalent rigid body output variables.
Figure 2.1.6-5 shows the vertical motion of node 5001 (bearing point "A" of the A-arm), node 5080
(point of A-arm nearest the tire), and the average vertical motion of the A-arm (output variable UC3).
As expected, the displacement of the center of mass of the component lies between the displacements
of its two ends. Figure 2.1.6-6shows the average rigid body rotation of the A-arm component about its
center of mass. Rigid body rotations are available about the three global axes. Here we are interested in
the rotation about the global X-axis (output variable URC1). The standard output for rotations is
radians, but the results have been scaled to plot the rotation in degrees.
Input files
jointcautosuspension.inp
Suspension analysis with JOINTC elements.
jointcautosuspension_rotated.inp
Rotated suspension analysis with JOINTC elements. This input file includes one extra (rotation)
step but is otherwise identical to jointcautosuspension.inp.
jointcautosuspension_depend.inp
Identical to jointcautosuspension.inp, except that field-variable-dependent linear and nonlinear
spring properties are used in the JOINTC elements.
connautosuspension.inp
Suspension analysis with connector elements.
connautosuspension_rotated.inp
2-781
Rotated suspension analysis with connector elements. This input file includes one extra (rotation)
step but is otherwise identical to connautosuspension.inp.
connautosuspension_depend.inp
Identical to connautosuspension.inp, except that field-variable-dependent linear and nonlinear
spring properties are used in the connector elements with connection types CARTESIAN,
ROTATION, and AXIAL.
Figures
Figure 2.1.6-1 Left front automobile suspension.
Figure 2.1.6-2 Displacement histories of tire, wheel center, and frame.
2-782
Figure 2.1.6-3 Displaced shapes during positive vertical tire motion.
Figure 2.1.6-4 Overlay of unrotated and rotated suspension analyses.
Figure 2.1.6-5 Vertical motion of the A-arm: average and nodal motions.
2-783
Figure 2.1.6-6 Average rotation of the A-arm about its center of mass.
Sample listings
2-784
Listing 2.1.6-1
*HEADING
JOINTC ELEMENTS; MODELING AN AUTOMOBILE
SUSPENSION
**
** Auto suspension sub-assembly modeled in
** sprung condition. Coil spring and tire
** spring have initial preload as given by
** their respective spring load-deflection
** curves. Weight of frame (by gravity load)
** rests on the coil spring top and the two
** wishbone bearing (pivot) points.
**
** FRAME
**
** Model frame as lumped mass at auto centerline
*NODE, NSET=FRAME
1001, 0.00, 0.00, 0.00
*ELEMENT, TYPE=MASS, ELSET=FRAME
1001, 1001
** Mass used is one quarter of total auto mass
*MASS, ELSET=FRAME
0.423,
*BOUNDARY
1001, 1,2, 0.0
1001, 4,6, 0.0
**
** WHEEL
**
** 3001 is lower rim, 3002 is wheel center
*NODE, NSET=WHEEL
3001, 0.00, -779.50, -315.00
3002, 0.00, -779.50, 0.00
*ELEMENT, TYPE=B31, ELSET=WHEEL
3001, 3001, 3002
*BEAM SECTION, SECTION=CIRC, MATERIAL=STEEL,
ELSET=WHEEL
24.0,
0.0, 1.0, 0.0
*BOUNDARY
WHEEL, 6,6, 0.0
2-785
**
** TIRE
**
*NODE, NSET=GROUND
2000, 0.00, 0.00, -335.00
*NODE, NSET=TIRE
2001, 0.00, -779.50, -335.00
2091, 167.50, -779.50, -290.11
2092, -167.50, -779.50, -290.11
*ORIENTATION, NAME=XYZ
1.0, 0.,0.,0.,1.,0.
3, 0.0
*ELEMENT, TYPE=JOINTC, ELSET=TIRE
2001, 3001, 2001
*JOINT, ELSET=TIRE, ORIENTATION=XYZ
*SPRING
1,
175.,
*SPRING
2,
175.,
*SPRING, NONLINEAR
3,
-10000., -24.21
0., -19.21
4150., 0.00
10630., 30.00
** for visualization only
*ELEMENT, TYPE=T3D3, ELSET=PLOTA
2091, 2091, 2001, 2092
*SOLID SECTION, MATERIAL=STEEL, ELSET=PLOTA
0.0005,
*MPC
BEAM, 2091, 2001
BEAM, 2092, 2001
**
** Tie bottom of tire to ground (makes global
** rotation possible)
*MPC
BEAM, 2001, 2000
**
** AXLE
**
2-786
*NODE, NSET=AXLE
4001, 0.00, -705.30, 0.00
4002, 0.00, -736.30, 0.00
4003, 2.56, -750.00, -116.00
4004, 152.00, -695.61, -13.81
4005, 152.00, -695.61, -13.81
*ELEMENT, TYPE=B31, ELSET=AXLE1
4020, 3002, 4001
4021, 4001, 4002
ELSET=AXLE1
13.0,
4022, 4002, 4003
ELSET=AXLE2
15.0,
1.0, 0.0, 0.0
4023, 4002, 4004
*BEAM GENERAL SECTION, SECTION=GENERAL,
ELSET=AXLE3, DENSITY=7.8E-9
391.0, 17000.0, 0.0, 9400.0,22600.0
0.0, 0.0, -1.0
2.1E+05, 8.077E+04
*BOUNDARY
AXLE, 5,5, 0.0
**
** WISHBONE
**
*NODE, NSET=WBONE
5001, -73.00, -351.00, -90.00
5002, 325.00, -351.00, -90.00
5011, -17.00, -540.00, -128.00
5012, -73.00, -398.60, -110.70
5013, 325.00, -398.60, -110.70
5080, 2.56, -750.00, -116.00
*ELEMENT, TYPE=B31, ELSET=WB1
5001, 5080, 5011
5002, 5011, 5012
5003, 5011, 5013
ELSET=WB1, DENSITY=7.8E-9
2-787
542.0, 449591.0, 0.0, 196281.0, 639175.0

0.0, 0.0, -1.0
2.1E+05, 8.077E+04
5033, 5012, 5001
5034, 5013, 5002
ELSET=WB2, DENSITY=7.8E-9
342.0, 81077.0, 0.0, 81077.0, 162155.0
0.0, 0.0, -1.0
2.1E+05, 8.077E+04
5035, 5012, 5013
ELSET=WB3
13.0,
**
*ELSET, ELSET=WBONE
WB1, WB2, WB3
** Pin axle and wishbone together
*MPC
PIN, 4003, 5080
**
** COIL SPRING
**
*NODE, NSET=CTOP
6001, -8.50, -514.00, 214.00
*ELEMENT, TYPE=SPRINGA, ELSET=CSPRG
6001, 5011, 6001
*SPRING, ELSET=CSPRG, NONLINEAR
-13600.0, -100.00
-8900.0, 0.00
-4600.0, 100.00
** Tie top of spring to frame
*MPC
BEAM, 6001, 1001
** for visualization only
*ELEMENT, TYPE=T3D2, ELSET=PLOTB
6091, 5011, 6001
*SOLID SECTION, MATERIAL=STEEL, ELSET=PLOTB
0.0005,
**
2-788
** BEARING POINT A
**
*NODE, NSET=A-PVT
7001, -73.00, -351.00, -90.00
*ELEMENT, TYPE=JOINTC, ELSET=APVT
7001, 7001, 5001
** Tie fixed side of jointc to frame
*MPC
BEAM, 7001, 1001
*JOINT, ELSET=APVT, ORIENTATION=XYZ
*SPRING, NONLINEAR
1,
-16000.0, -4.8
-10000.0, -3.8
-6000.0, -2.6
-3000.0, -1.1
0.0, 0.0
3000.0, 1.1
6000.0, 2.6
10000.0, 3.8
16000.0, 4.8
*SPRING, NONLINEAR
2,
-10000.0, -1.80
-6000.0, -1.25
-2000.0, -0.53
0.0, 0.00
2000.0, 0.53
6000.0, 1.25
10000.0, 1.80
*SPRING, NONLINEAR
3,
-10000.0, -2.40
-5000.0, -1.45
-2000.0, -0.68
0.0, 0.00
2000.0, 0.68
5000.0, 1.45
10000.0, 2.40
*SPRING
4,
1.142E5,
*SPRING
2-789
5,
2.8568E6,
*SPRING
6,
2.8568E6,
**
** BEARING POINT B
**
*NODE, NSET=B-PVT
7002, 325.00, -351.00, -90.00
*ELEMENT, TYPE=JOINTC, ELSET=BPVT
7002, 7002, 5002
** Tie fixed side of jointc to frame
*MPC
BEAM, 7002, 1001
*JOINT, ELSET=BPVT, ORIENTATION=XYZ
*SPRING, NONLINEAR
1,
-1600.0, -6.00
-1000.0, -2.50
-700.0, -1.50
-500.0, -1.00
-300.0, -0.50
0.0, 0.00
300.0, 0.50
500.0, 1.00
700.0, 1.50
1000.0, 2.50
1600.0, 6.00
*SPRING, NONLINEAR
2,
-6800.0, -4.00
-4800.0, -3.75
-3600.0, -3.50
-2800.0, -3.25
-1700.0, -2.75
-1000.0, -2.25
-700.0, -1.75
0.0, 0.00
700.0, 1.75
1000.0, 2.25
1700.0, 2.75
2800.0, 3.25
2-790
3600.0, 3.50
4800.0, 3.75
6800.0, 4.00
*SPRING,NONLINEAR
3,
-6400.0, -2.00
-5200.0, -1.75
-4160.0, -1.50
-2400.0, -1.00
1000.0, 0.00
4400.0, 1.00
6160.0, 1.50
7200.0, 1.75
8400.0, 2.00
*SPRING
4,
4.3082E4,
*SPRING
5,
1.1467E6,
*SPRING
6,
3.1311E5,
**
*ELASTIC
2.1E+05,0.3
*DENSITY
7.8E-9,
*ELSET, ELSET=ALL
1001,
**
** Idealized triangular speed bump 400mm long
** and 100mm high. Time data based on 5km/hr
** auto speed
*AMPLITUDE, NAME=BUMP, TIME=STEP TIME
0.0, 0.0, 0.05, 0.0, 0.20, 100.0, 0.35, 0.0
0.40, 0.0
**
**
Auto Weight/Tire Compression/Coil spring
Equilibrium
2-791
*STATIC
0.2, 0.2
*BOUNDARY
2000, 1,6, 0.0
*DLOAD
ALL, GRAV, 9815.0, 0.0, 0.0, -1.0
*NSET, NSET=PLOT
2000, 3002, 1001
*NODE FILE, NSET=PLOT
U,
*END STEP
**
**
Apply "bump" to tire bottom with boundary
condition
*DYNAMIC, HAFTOL=5000
5.0E-4, 0.40, , 0.01
*BOUNDARY, AMPLITUDE=BUMP
2000, 3,3, 1.0
*EL FILE, ELSET=WBONE
XC,
UC,
VC,
HC,
HO,
RI,
MASS, VOL
*END STEP
2.1.7 Explosive pipe closure

This problem illustrates the following concepts: large deformation kinematics, equations of state,
elastic-plastic material, transformations, detonation points.
Problem description
The units used in this analysis are referred to as c.g. ¹sec. Using these units, length is given in
centimeters (cm), mass in grams (gm), and time is measured in microseconds (¹sec). The stresses have
units of mega bars (M bar). These units are commonly used in shock wave physics applications
because the pressures tend to have values on the order of unity.
2-792
In this example problem two concentric pipes have the annulus between them filled with high
explosive (HE). The inside radius of the inner pipe is 10 mm. The inside radius of the outer pipe is 20
mm. Both pipes are steel with a wall thickness of 2 mm. Each pipe is modeled with 6 elements in the
radial direction, while the HE is modeled with 24 elements in the radial direction.
The steel pipe is an elastic, perfectly plastic material with Young's modulus of 221.1 GPa (.02211 M
bar), Poisson's ratio of 0.279, yield strength of 430 MPa (.0043 M bar), and density of 7846 kg/m2
(7.846 gm/cm3).
The explosive material is modeled using the JWL equation of state with detonation wave speed = 7596
m/sec (.7596 cm/microsecond), A = 520.6 GPa (5.206 M bar), B = 5.3 GPa (0.053 M bar), R1 = 4.1,
R2 = 1.2, !=.35, density of 1900 kg/m3 (1.9 gm/cm3), and initial specific energy of 3.63 Joule/kg
(0.0363 T erg/gm). The tension cutoff pressure is assumed to be zero and is specified using the
*TENSILE FAILURE option. Refer to `Èquation of state,'' Section 9.5.1 of the ABAQUS/Explicit
User's Manual, for a description of this material model.
The explosive material is detonated at four points around the circumference of the cylinder. Because of
the symmetry in this problem, only one-eighth of the pipe is modeled. Figure 2.1.7-1 shows the
original geometry and the location of the detonation point for the model. A transformed coordinate
system is used to define the symmetry conditions along the sloping boundary.
This analysis is run in two steps to reduce the amount of output written to the restart file. In the early
part of the analysis, the deformations are not of much interest. Hence, the first step has a duration of 6
¹sec and requests only 1 restart interval (*RESTART, WRITE, NUMBER INTERVAL=1). After 6
¹sec the deformations are becoming significant. The second step has a duration of 1.5 ¹sec and
requests 3 restart intervals (*RESTART, WRITE, NUMBER INTERVAL=3).
This analysis is run as both a two-dimensional case using CPE4R elements and as a three-dimensional
case using C3D8R elements. In the three-dimensional case the displacements are constrained to be
zero in the out-of-plane direction.

Figure 2.1.7-2 through Figure 2.1.7-5show a sequence of the deformed shapes computed by
ABAQUS/Explicit for the two-dimensional case. Although not shown here, the results of the
three-dimensional analysis are indistinguishable from those of the two-dimensional analysis.
This problem tests the features listed, but it does not provide independent verification of them.
Input files
eoscyl2d.inp
Two-dimensional case.
eoscyl3d.inp
Three-dimensional case.
Figures
2-793
Figure 2.1.7-1 Original geometry.
Figure 2.1.7-2 Deformed configuration after 6.0 ¹sec.
2-794
2-795
Sample listings
2-796
Listing 2.1.7-1
*HEADING
EXPLOSIVE PIPE CLOSURE
*NODE
1,1.0,0.
7,1.2,0.
31,2.0,0.
37,2.2,0.
3001, .707107, .707107
3007, .848528, .848528
3031,1.414213,1.414213
3037,1.555634,1.555634
*NGEN,LINE=C,NSET=A
1,3001,100,,0.,0.,0.
*NGEN,LINE=C,NSET=B
7,3007,100,,0.,0.,0.
*NGEN,LINE=C,NSET=C
31,3031,100,,0.,0.,0.
*NGEN,LINE=C,NSET=D
37,3037,100,,0.,0.,0.
*NFILL
A,B,6,1
B,C,24,1
C,D,6,1
*NSET,NSET=XAXIS,GEN
1,37,1
*NSET,NSET=TAXIS,GEN
3001,3037,1
*TRANSFORM,NSET=TAXIS
.707107,.707107,0., -.707107,.707107,0.
*ELEMENT,TYPE=CPE4R,ELSET=PIPE
1, 1, 2,102,101
901,31,32,132,131
*ELGEN,ELSET=PIPE
1, 6,1,1, 30,100,6
901, 6,1,1, 30,100,6
*SOLID SECTION,ELSET=PIPE,MATERIAL=STEEL
*ELEMENT,TYPE=CPE4R,ELSET=HE
181, 7,8,108,107
*ELGEN,ELSET=HE
181, 24,1,1, 30,100,24
*SOLID SECTION,ELSET=HE,MATERIAL=HE
2-797
*DENSITY
7.846,
*ELASTIC
2.211E-2,.279
*PLASTIC
.0043,0.
*MATERIAL,NAME=HE
*DENSITY
1.9,
*EOS,TYPE=JWL
0.7596,5.206,.053,.35,4.1,1.2,3.63E-2
*DETONATION POINT
1.414213,1.414213,0.,0.
*TENSILE FAILURE,ELEMENT DELETION=NO,
PRESSURE=BRITTLE,SHEAR=BRITTLE
0.,
**
*BOUNDARY
XAXIS,YSYMM
TAXIS,YSYMM
*STEP
*DYNAMIC,EXPLICIT
,2.
*EL FILE
PRESS,S,LE
*NODE FILE
U,
*ENERGY FILE
*ENERGY HISTORY
ALLKE,ALLIE,ALLSE,ALLAE,ALLVD,ALLCD,ALLPD,
ALLFD,ALLWK,ETOTAL,DT
*END STEP
*STEP
*DYNAMIC,EXPLICIT
,0.4
*FILE OUTPUT,NUM=2, TIMEMARKS=YES
*EL FILE
PRESS,S,LE
2-798
*NODE FILE
U,
*ENERGY FILE
*ENERGY HISTORY
ALLKE,ALLIE,ALLSE,ALLAE,ALLVD,ALLCD,ALLPD,
ALLFD,ALLWK,ETOTAL,DT
*END STEP
2.1.8 Knee bolster impact with double-sided surface contact

This example illustrates the use of the double-sided surface contact capability in a simulation
involving large relative motion between potentially contacting surfaces. Double-sided surface contact
is required because large deformations during the course of the analysis cause several slave nodes to
displace around and then behind their originally opposing master surfaces.
Problem description
This model represents an automobile knee bolster assembly--the portion of the instrument panel that
the occupant's legs impact in the event of a crash. The assembly consists of a hard plastic cover (the
knee bolster) supported by a stiff steel substructure. Proper design of this assembly ensures that the
occupant's energy is dissipated with a minimum of injury causing forces. In this simulation the legs
approach the knee bolster at 6 m/s, representing unrestrained motion following a 15 mph to dead stop
crash event.
The components of the instrument panel are modeled using S3R and S4R shell elements. The bolster is
made up of 2690 shell elements, with the material modeled as a von Mises elastic strain hardening
plastic material with a Young's modulus of 2.346 GPa, a Poisson's ratio of 0.4, a density of 1140
kg/m3, and a yield stress of 11.7 MPa. The steel substructure is made up of 1648 elements, with the
material modeled as a strain hardening steel with a Young's modulus of 207 GPa, a Poisson's ratio of
0.3, a density of 7700 kg/m3, and a yield stress of 207 MPa. Contact between the instrument panel
assembly components is modeled by defining double-sided surfaces on 2536 shell elements in the
bolster and on 213 shell elements on the steel substructure. Figure 2.1.8-1 shows the model geometry
from the rear of the knee bolster prior to impact, and Figure 2.1.8-2shows where a double-sided
surface definition is needed because of large motions of the bolster structure. Figure 2.1.8-3 shows the
knee bolster and knee/leg assembly from a position outboard and behind the driver prior to impact.
The legs are represented as structural members with a surrounding rigid surface. The structural
members, representing the bones, are modeled with B31 beam elements and T3D2 truss elements, with
the material modeled as elastic with a Young's modulus of 207 GPa, a Poisson's ratio of 0.3, and a
density of 7.7 kg/m3. The rigid surfaces, representing the knee and shin, are modeled with R3D4 rigid
elements. A single-sided surface definition is used in this case. The body mass is modeled by
distributing mass elements at various locations among the nodes of the structural elements. Contact
between the legs and the instrument panel is defined using node-based surfaces on the bolster and rigid
2-799
surfaces to represent the knees and shins.

Initial velocities are defined on the leg components to approximate a 15 mph (6 m/s) crash condition.
The hips are constrained to translate in the plane of the seat. The ankles are constrained consistent with
fixed planting of the feet on the floor of the car. The dashboard substructure is fixed at locations where
it would be welded to the automobile frame; deformations due to this impact are assumed to be
confined to the explicitly modeled structure.

Figure 2.1.8-4 shows the deformed shape of the bolster assembly after 30.0 ms.
This analysis would not be possible with single-sided surfaces because nodes move around the edge of
the steel structure and shift from the positive side of the surface to the negative side as the surfaces
deform. Such motions are allowed with double-sided surfaces but not with single-sided surfaces.
Figure 2.1.8-5 shows the energy time history of the whole model: internal energy, kinetic energy,
recoverable strain energy, and plastic dissipation. This figure shows that roughly one-half of the body's
initial kinetic energy has been transferred by the end of this simulation. Of this transferred amount
roughly one-quarter has been transferred to elastic deformations in the instrument panel structure and
bones, and the balance is lost to plastic dissipation; the crash event cannot be considered complete at
30 ms.
Figure 2.1.8-6 shows the total knee and shin contact forces measured against the displacement into the
bolster. Consistent with the observations of the energy quantities, it is not clear that these forces have
peaked or that the crash event is complete.
Acknowledgment
HKS would like to thank GE Plastics for supplying the model used in this example.
Input files
knee_bolster.inp
knee_bolster_ef1.inp
External file referenced by this analysis.
Figures
2-800
Figure 2.1.8-1 Initial configuration of the knee bolster model (view from behind the bolster).
Figure 2.1.8-2 Location of node that moves from initially opposing one side of a shell element in the
steel substructure to opposing the opposite side following large deformation of the bolster structure.
Figure 2.1.8-3 Initial configuration of the knee bolster model (view from outboard and behind the
driver).
2-801
Figure 2.1.8-4 Deformed shape after 30 ms.
Figure 2.1.8-5 Time histories of the whole model: internal energy, kinetic energy, recoverable strain
energy, and plastic dissipation.
2-802
Figure 2.1.8-6 Front leg reaction forces measured against impact displacement.
Sample listings
2-803
Listing 2.1.8-1
*HEADING
KNEE BOLSTER IMPACT SIMULATION
**
**This example simulates the impact of a knee
**and leg structure into a knee bolster assembly
**
** READ NODE AND NODE SET DEFINITIONS
** FROM EXTERNAL FILE:
*INCLUDE,INPUT=knee_bolster_ef1.inp
**
** READ ELEMENT AND ELSET DEFINITIONS
**
** Material definitions
**
*DENSITY
7.7E-09,
*ELASTIC, TYPE=ISO
207000.0, 0.3
*PLASTIC
207.0, 0.0
276.0, 0.22
**
*MATERIAL, NAME=BONE
*DENSITY
7.7E-12,
*ELASTIC, TYPE=ISO
207000.0, 0.3
**
*MATERIAL,NAME=MC8002
*ELASTIC
.2346E+04, .4000
*PLASTIC
11.7, 0.00
32.8, 0.00114
47.0, 0.0052
53.5, 0.0227
*DENSITY
1.14E-09,
2-804
**
*SHELL SECTION,ELSET=STEEL ,MATERIAL=STEEL
1.50000, 5
*SHELL SECTION,ELSET=TOPBRKT ,MATERIAL=STEEL
1.50000, 5
*SHELL SECTION,ELSET=FLRBRKT ,MATERIAL=STEEL
1.50000, 5
*SHELL SECTION,ELSET=SDBRKT ,MATERIAL=STEEL
1.50000, 5
*SHELL SECTION,ELSET=FRONT ,MATERIAL=MC8002
3.00000, 5
*SHELL SECTION,ELSET=CORR ,MATERIAL=MC8002
2.50000, 5
*SHELL SECTION,ELSET=RIBS ,MATERIAL=MC8002
1.88000, 5
**
** READ BOUNDARY CONDITION DEFINITIONS
**
*MPC
BEAM, 6447, 3253
BEAM, 6264, 2540
BEAM, 6366, 2555
**
** Below are beam elements for lower torso.
**
*BEAM SECTION,ELSET=TIBCHK,MATERIAL=BONE,
SECTION=CIRC
25.4,
*ELEMENT,TYPE=B31,ELSET=TIBCHK
** Left Tibia
8000, 9000, 9001
** Right Tibia
8100, 9100, 9101
** Left Hip
8400, 9200, 9400
** Right Hip
8500, 9400, 9300
*SOLID SECTION,ELSET=FEMUR,MATERIAL=BONE
2027.0,
*ELEMENT, TYPE=T3D2,ELSET=FEMUR
** Left Femur
2-805
8200, 9200, 9001

** Right Femur
8300, 9300, 9101
**
**
** Initial Conditions and Definitions.
**
**
*MASS,ELSET=WT
.0082,
**
** Left Shin
**
*NSET, NSET=LSHIN, GENERATE
8000, 8121, 1
*ELEMENT,TYPE=MASS,ELSET=WT
10000, 9001
*NSET,NSET=LKNE
9001,
*NSET, NSET=RSHIN, GENERATE
8200, 8321, 1
10001, 9101
*NSET,NSET=RKNE
9101,
*NSET,NSET=KNEES
9001, 9101
KNEES, 1, -5364.48
KNEES, 3, 2839.18
KNEES, 5, -14.1725
**
** Left hip mass
**
*NSET,NSET=M3
9200,
10002, 9200
*BOUNDARY
M3, 3, , 0.
M3, 1, -6593.58
**
2-806
** Center mass
**
*NSET,NSET=M4
9400,
10003, 9400
*BOUNDARY
M4, 2, 3, 0.
M4, 1, -6593.58
**
** Right torso mass
**
*NSET,NSET=M5
9300,
10004, 9300
*BOUNDARY
M5, 3,, 0.
M5, 1, -6593.58
**
** Ankles
**
*NSET,NSET=ANKLE
9000, 9100
*TRANSFORM,NSET=ANKLE
142.200, 0.0, 393.900, 0.0, 100.0, 0.0
*BOUNDARY
ANKLE, 1,3, 0.0
ANKLE, 4,, 0.0
ANKLE, 5, -14.1725
**
** History Section
**
*SURFACE,TYPE=NODE,NAME=LNODE
LNODE,
*SURFACE,TYPE=ELEMENT, NAME=RMETAL
RMETAL,
*SURFACE,TYPE=ELEMENT, NAME=CORR
CORR,
2-807
*SURFACE,TYPE=ELEMENT,NAME=LK
LKSHIN2,SPOS
*SURFACE,TYPE=NODE,NAME=RNODE
RNODE,
*SURFACE,TYPE=ELEMENT, NAME=FRONT
FRONT,
*SURFACE,TYPE=ELEMENT,NAME=RK
RKSHIN2,SPOS
*SURFACE,TYPE=ELEMENT, NAME=LMETAL
LMETAL,
*RIGID BODY,ELSET=LKSHIN2,REF NODE=9001
*RIGID BODY,ELSET=RKSHIN2,REF NODE=9101
*STEP
*DYNAMIC, EXPLICIT
,.030
*CONTACT CONTROLS,GLOBTRKINC=1000,CPSET=CPAIR
**
*FILE OUTPUT, NUMBER INTERVAL = 2
*EL FILE, ELSET = FRONT
S,
*ENERGY FILE
ALLIE, ALLKE, ETOTAL, ALLWK, ALLAE, ALLSE, ALLPD
*NODE FILE, NSET = KNEES
U,V
*NODE FILE, NSET = M3
U,V
U,V
U,V
**
** Surface Definition
**
*CONTACT PAIR, INTERACTION=FRIC,CPSET=CPAIR
LK, LNODE
RK, RNODE
**
FRONT, CORR
FRONT, RMETAL
FRONT, LMETAL
CORR, RMETAL
CORR, LMETAL
**
2-808

*FRICTION
0.2,
**
** Output
**
RF,BONDSTAT,BONDLOAD
*NODE HISTORY,NSET=LSHIN
RF,
*NODE HISTORY,NSET=RSHIN
RF,
*NODE HISTORY,NSET=LKNE
U,V,A
*NODE HISTORY,NSET=RKNE
U,V,A
*NODE HISTORY,NSET=M4
U,V,A
*ENERGY HISTORY
ALLIE, ALLKE, ETOTAL, ALLWK, ALLAE, ALLSE, ALLPD
*MONITOR, NODE=9400,DOF=1
*ENDSTEP
2.1.9 Cask drop with foam impact limiter

A containment cask is partially filled with fluid and a foam impact limiter. The complete package is
dropped a distance of 9.09 m (30 ft) onto a rigid surface, which results in an impact speed of 13.35
m/sec (525.3 in/sec). The problem illustrates the use of an initial velocity condition and the analysis of
a structure containing liquid and incorporating crushable foam to absorb impact energy. Experimental
and numerical results for this problem have been reported by Sauvé et al. (1993). The numerical results
given in the reference were obtained using a relatively coarse finite element mesh. In this example
results are presented for the same coarse mesh as the reference and also for a more refined mesh.
Model description
The containment cask shown in Figure 2.1.9-2consists of two compartments. The upper compartment
surrounds the fluid and is made of stainless steel ( 304L). It has a height of 580 mm (22.8 in), a
diameter of 300 mm (11.8 in), and a wall thickness of 4.76 mm (0.187 in). The top mild steel cover has
a thickness of 9.52 mm (0.375 in). The water is filled to a depth of 522 mm (20.55 in), which is 90% of
the container's capacity. Figure 2.1.9-3shows the original, coarse mesh of C3D8R elements used to
model the fluid. Contact conditions are defined between the fluid and the inside of the upper
compartment.
2-809
An impact limiter made of polyurethane foam is contained within the bottom mild steel compartment
of the cask. The height of the foam impact limiter is 127.3 mm (5.01 in). Figure 2.1.9-4shows the
coarse mesh used to model the foam. Contact conditions are defined between the foam and the inside
of the bottom compartment of the cask. The foam impact limiter and the fluid/stainless steel liner are
separated by a mild steel bulkhead with a thickness of 12.7 mm (0.5 in). A 12.7 mm (0.5 in) air gap
exists between the top of the foam surface and this bulkhead.
In the experiment a pressure transducer is located in the polyurethane foam on the centerline of the
cask at the top of the impact limiter. This result is compared with vertical stress-time histories taken
from the element at the top of the foam model on the centerline.
Both axisymmetric and three-dimensional models are analyzed. Figure 2.1.9-5shows the
three-dimensional model formed by assembling the parts shown in Figure 2.1.9-2through Figure
2.1.9-4. The equivalent axisymmetric model is shown in Figure 2.1.9-6.
Contact pairs are defined between the solids and the shells. Element-based surfaces are defined on the
shells, and node-based surfaces are defined containing the nodes on the outer surfaces of the solid
elements. The shell thickness was not taken into account when the original meshes were designed, and
the outer surface of the solids usually coincides with the midsurface of the enclosing shell. This would
lead to an initial overclosure of one-half the shell thickness, unless the NO THICK parameter is used
to enforce contact at the midsurface of the shell, as if it had zero thickness. The use of a node-based
surface implies a pure master-slave relationship for the contact pair. This is important in this problem
because the default in ABAQUS/Explicit when contact is defined between shells and solids is to define
a pure master-slave relationship with the solids as the master and the shells as the slave. In this case
the shell structures are much stiffer than the fluid and foam structures, so the master-slave roles must
be reversed.
For the axisymmetric model two cases using different section control options for the foam and fluid
elements are analyzed. The first case uses the COMBINED hourglass control option; the second case
uses the default section control options (the RELAX STIFFNESS hourglass control). The
three-dimensional model also has two cases with different section control options for the foam and
fluid elements. The first case uses the ORTHOGONAL kinematic option and COMBINED hourglass
control; the second three-dimensional case uses the default section control options (the AVERAGE
STRAIN kinematic option and the RELAX STIFFNESS hourglass control). The options used are
summarized in Table 2.1.9-4. Coarse and refined meshes are used for all analysis cases.
Material description
The general material properties are listed in Table 2.1.9-1. The material models for the water and foam
are further described below.
Water:
The water is treated as a simple hydrodynamic material model. This provides zero shear
strength and a bulk response given by
p = K"vol ;
2-810
where K is the bulk modulus with a value 2068 MPa (300000 psi). This model is defined
using the linear Us ¡ Up equation of state model provided in ABAQUS/Explicit. The linear
Us ¡ Up Hugoniot form, p = f (½) + g (½)Em , is
½0 c20 ´ ¡0 ´
p= 2
(1 ¡ ) + ¡0 ½0 Em ;
(1 ¡ s´) 2
where ´ = 1 ¡ ½0 =½ is the same as the nominal volumetric strain measure, "vol . Since
K = ½0 c20 , setting the parameters s = 0:0 and ¡0 = 0:0 gives the simple hydrostatic bulk
response defined earlier. In this analysis c0 = 1450.6 m/sec (57100 in/sec) and ½0 = 983.2
kg/m3 (0.92 ´ 10-4 lb sec2in-4). The tension cutoff pressure is assumed to be zero and is
specified using the *TENSILE FAILURE option. Refer to `Èquation of state,'' Section 9.5.1
of the ABAQUS/Explicit User's Manual, for a description of this material model.
Foam:
The crushable foam model is used for the polyurethane foam. In this model the flow potential,
h, is chosen as
r
9 2
h= p + q2 ;
2
where q is the Mises equivalent stress and p is the hydrostatic pressure. The yield surface is
defined as
"µ ¶2 ³ ´ # 12
pt ¡ pc q 2 pc + pt
+p + = :
2 M 2
Sauvé et al. use the "soils and crushable foams" model, which was originally defined in an
unpublished report by Krieg (1978) and is based upon a Mises plasticity model in which the
yield stress depends upon the mean volumetric pressure. The volumetric deformation allows
for plastic behavior, defined by tabular data defining pressure versus volume strain. This
model is easy to implement in an explicit dynamics algorithm and useful because the
deviatoric and volumetric terms are only loosely coupled. However, it requires an experienced
analyst to ensure that meaningful results are obtained, mainly because the model does not
match physical behavior well under deviatoric straining.
To define the initial shape of the yield surface, the ABAQUS/Explicit crushable foam model
requires the initial yield surface position, "pl
vol j0 ; the initial yield stress in uniaxial
compression, ¾0 ; and the magnitude of the strength in hydrostatic tension, pt . Sauvé et al.
define the pressure-dependent yield surface for the foam model as
¾y2 = 3:18 + 2:06p;
2-811
where the units of stress are MPa and pressure is positive in compression. To calibrate the
ABAQUS/Explicit crushable foam model to this pressure-dependent data, we observe that
p = 13 ¾y for the uniaxial compression case. Substituting this value for p in the above equation
and solving for ¾y gives ¾0 =2.16 MPa (313.3 psi). The value of pt is obtained by solving the
above equation for ¾y = 0, giving pt = 1.54 MPa (223.8 psi). The value of "pl vol j0 is chosen to
be zero because tensile experimental data are not given in the reference and the foam is almost
always under hydrostatic compression. If the foam were subjected to triaxial tension, it would
be necessary to define pressure-plastic volumetric strain data to account for strain softening, as
discussed in ``Crushable foam plasticity model,'' Section 10.3.3 of the ABAQUS/Explicit
User's Manual.
The pressure/volumetric strain data in the reference are given in Table 2.1.9-2. Table
2.1.9-3shows the values as converted to the form required for ABAQUS/Explicit. Each form
of the data is plotted in Figure 2.1.9-1.

The deformed geometries for the three-dimensional and the axisymmetric models at 5 msec are shown
in Figure 2.1.9-7and Figure 2.1.9-8. The axisymmetric model is analyzed using COMBINED hourglass
control. The three-dimensional model uses the ORTHOGONAL kinematics and COMBINED
hourglass control. Figure 2.1.9-9 shows plots of the vertical stress versus time for the element located
at the pressure transducer in the foam; results from the models with the previous section control
options, as well as results from analyses using the default section control options, are reported for
comparison (see Table 2.1.9-4). Axisymmetric and three-dimensional results are compared to the
experimental pressure trace. The time origin of the experimental curve is not defined in the reference;
therefore, the experimental curve is shifted so that the time when pressure in the transducer changes to
a positive value is assumed to be the time at which impact occurs. The numerical pressure results show
significant oscillations about the experimental results during the first 2 msec of the response. This is
partly because the meshes are quite coarse and partly because pressure transducers in experiments
exhibit inertia in their response and will not report sharp gradients in time. During the next 3 msec the
numerical results correspond more closely with the experimental results. The analyses run with
different section control options compare very well.
A more refined three-dimensional mesh is shown in Figure 2.1.9-10. The refined axisymmetric model
is the same model used in the r-z plane. The deformed geometries for these models are shown in
Figure 2.1.9-11(using the ORTHOGONAL kinematics and COMBINED hourglass control) and Figure
2.1.9-12(using the COMBINED hourglass control). The vertical stress histories for the refined models
are shown in Figure 2.1.9-13for the same options used for the coarse meshes (see Table 2.1.9-4). The
numerical results show less oscillation about the experimental results than those obtained with the
coarse mesh. They compare well with the experimental results during the following 3 msec of the
response. In addition, Figure 2.1.9-11shows that the refined mesh eliminates much of the fluid's
hourglass-like response due to its zero shear strength.
Input files
2-812
cask_drop_axi_cs.inp
Coarse axisymmetric model using the COMBINED hourglass control.
cask_drop_3d_ocs.inp
Coarse three-dimensional model using the ORTHOGONAL kinematic and the COMBINED
hourglass control.
cask_drop_axi.inp
Coarse axisymmetric mesh using the default section controls.
cask_drop_3d.inp
Coarse three-dimensional mesh using the default section controls.
cask_drop_axi_r_cs.inp
Refined axisymmetric model using the COMBINED hourglass control.
cask_drop_3d_r_ocs.inp
Refined three-dimensional model using the ORTHOGONAL kinematic and the COMBINED
hourglass control.
cask_drop_axi_r.inp
Refined axisymmetric mesh using the default section controls.
cask_drop_3d_r.inp
Refined three-dimensional mesh using the default section controls.
References
· Krieg, R. D., `À Simple Constitutive Description for Soils and Crushable Foams ,''
SC-DR-72-0883, Sandia National Laboratories, Albuquerque, NM , 1978.
· Sauvé, R. G., G. D. Morandin, and E. Nadeau, `Ìmpact Simulation of Liquid-Filled Containers

Including Fluid-Structure Interaction,'' Journal of Pressure Vessel Technology, vol. 115, pp.
68-79, 1993.
Tables
Table 2.1.9-1 Material properties.

Properties A36 304L Liqui Foa
d m
Density, ½ (kg m-3) 8032 8032 983 305
Young's modulus, E (GPa) 193. 193. .129
1 1
Poisson's ratio, º 0.28 0.28 0
Yield stress, ¾y0 (MPa) 206. 305.
2-813
8 4
Bulk modulus, K (GPa) 2.07
Hardening modulus, Et 0 1.52
(GPa)
Table 2.1.9-2 Pressure-volumetric strain data.

²v 0 .01 .02 .03 .04 .05 .06 .385 .48 .53 .55
p 0 2.7 4.1 5.1 5.5 5.8 6.2 10.3 19.3 39.3 82.7
(MPa) 6 4 7 2 6 1 4 1 0 4
Table 2.1.9-3 Modified pressure-volumetric strain data.

²pv 0 .01 .02 .345 .44 .49 .51 2
pc + pt (MPa) 7.0 7.4 7.7 11.8 20.8 40.8 84.2 5516
6 0 5 9 5 4 8 .
Table 2.1.9-4 Analysis options tested.

Analysis Section Controls
Label Kinematic Hourglass
AXI n/a relax
AXI CS n/a combined
3D average relax
3D OCS orthogonal combined
Figures
Figure 2.1.9-1 Foam hardening curves.
Figure 2.1.9-2 Containment structure mesh in the three-dimensional model (coarse mesh).
2-814
Figure 2.1.9-3 Fluid mesh in the three-dimensional model (coarse mesh).
Figure 2.1.9-4 Foam mesh in the three-dimensional model (coarse mesh).
Figure 2.1.9-5 The complete three-dimensional model (coarse mesh).
2-815
Figure 2.1.9-6 Axisymmetric model (coarse mesh).
Figure 2.1.9-7 Three-dimensional deformed geometry using ORTHOGONAL element kinematics and
COMBINED hourglass control (coarse mesh).
2-816
Figure 2.1.9-8 Axisymmetric deformed geometry using COMBINED hourglass control (coarse mesh).
Figure 2.1.9-9 Vertical stress history in the foam (coarse mesh).
Figure 2.1.9-10 Refined mesh for the three-dimensional model.
2-817
Figure 2.1.9-11 Three-dimensional deformed geometry using ORTHOGONAL element kinematics

and COMBINED hourglass control (refined mesh).
Figure 2.1.9-12 Axisymmetric deformed geometry using COMBINED hourglass control (refined
mesh).
2-818
Figure 2.1.9-13 Vertical stress history in the foam (refined mesh).
Sample listings
2-819
Listing 2.1.9-1
*HEADING
CASK DROP PROBLEM
AXISYMMETRIC MODEL (coarse mesh)
SECTION CONTROLS USED (HOURGLASS=COMBINED)
**
** Create liner mesh
**
*NODE,NSET=NSHE1
1, 0.0,28.3461
2, 2.3622,28.3461
3, 4.7244,28.3461
4, 7.0865,28.3461
5, 9.4487,28.3461
6, 11.8109,28.3461
7, 11.8109,27.1650
8, 11.8109,25.1965
9, 11.8109,23.2281
10, 11.8109,21.2596
11, 11.8109,19.2911
12, 11.8109,17.3226
13, 11.8109,15.3541
14, 11.8109,13.3857
15, 11.8109,11.4172
16,11.8109,9.4487
17,11.8109,8.4645
18,11.8109,7.4802
19,11.8109,6.4960
20,11.8109,5.5117
21,9.4487,5.5117
22,7.0865,5.5117
23,4.7244,5.5117
24,2.3622,5.5117
25, 0.0,5.5117
26,11.8109,4.4094
27,11.8109,3.3070
28,11.8109,2.2047
29,11.8109,1.1023
30,11.8109,0.0
31,9.4487,0.0
32,7.0865,0.0
33,4.7244,0.0
2-820
34,2.3622,0.0
35,0.0,0.0
*ELEMENT, TYPE=SAX1
1, 1,2
6, 6,7
20, 20,26
21, 26,27
30, 20,21
*ELGEN, ELSET=S1
1,5,1,1
*ELGEN, ELSET=S2
6,14,1,1
*ELGEN, ELSET=S3
30, 5,1,1
*ELSET,ELSET=S4
20,
*ELGEN, ELSET=S4
21,9,1,1
*ELSET,ELSET=SSTEEL
S1,S2,S3,
*ELSET,ELSET=MSTEEL
S4,
**
** Shell material definition
**
*SHELL SECTION,MATERIAL=A2,ELSET=S1,
0.3748,5
0.1874,5
0.5,5
0.0591,5
**
** Stainless steel (304L)
**
*MATERIAL,NAME=A1
*DENSITY
0.00075163,
2-821
*ELASTIC
28E6,0.28
*PLASTIC
44300.0,0.
66300.0,0.1
**
** Mild steel (A36)
**
*MATERIAL,NAME=A2
*DENSITY
0.00075163,
*ELASTIC
28E6,0.28
*PLASTIC
30000.0,0.
**
** Create fluid mesh
**
*NODE, NSET=WTOP
101, 0.0,26.0627
102, 2.36218,26.0627
103, 4.72436,26.0627
104, 7.08654,26.0627
105, 9.44872,26.0627
106, 11.8109,26.0627
*NODE, NSET=WMID
155, 0.0,9.2482
156, 2.36218,9.2482
157, 4.72436,9.2482
158, 7.08654,9.2482
159, 9.44872,9.2482
160, 11.8109,9.2482
*NODE, NSET=WBOT
179, 0.0,5.5117
180, 2.36218,5.5117
181, 4.72436,5.5117
182, 7.08654,5.5117
183, 9.44872,5.5117
184, 11.8109,5.5117
*NFILL,NSET=NWATER
WTOP,WMID,9,6
*NFILL,NSET=NWATER
WMID,WBOT,4,6
2-822
101, 102,101,107,108
*ELGEN, ELSET=WATER
101, 5,1,1,13,6,5
*MATERIAL, NAME=MWATER
*DENSITY
9.2E-5,
*EOS, TYPE=USUP
57100.874,0.0,0.0
0.,
*SOLID SECTION, ELSET=WATER, MATERIAL=MWATER,
CONTROL=B
*SECTION CONTROLS, HOURGLASS=COMBINED, NAME=B
**
** Create foam mesh
**
*NODE, NSET=FTOP
201, 0.0,5.0117
202, 2.36218,5.0117
203, 4.72436,5.0117
204, 7.08654,5.0117
205, 9.44872,5.0117
206, 11.8109,5.0117
*NODE,NSET=FBOT
231, 0.0,0.0
232, 2.36218,0.0
233, 4.72436,0.0
234, 7.08654,0.0
235, 9.44872,0.0
236, 11.8109,0.0
*NFILL, NSET=NFOAM
FTOP,FBOT,5,6
201, 202,201,207,208
*ELGEN, ELSET=FOAM
201, 5,1,1,5,6,5
*MATERIAL, NAME=MFOAM
*ELASTIC
18738,0.0
*DENSITY
2.85E-5,
2-823
*FOAM
0.0,223.84,313.35
*FOAM HARDENING
1023.84,0.00
1073.84,0.01
1123.84,0.02
1723.84,0.345
3023.84,0.44
5923.84,0.49
12223.84,0.51
8E5,2.00
*SOLID SECTION, ELSET=FOAM, MATERIAL=MFOAM,
CONTROL=B
**
** Define fluid and foam contact node set
**
*NSET, NSET=WCON1
101,102,103,104,105,106,112,118,124,130,136,142,
148,154,160,166,172,178,179,180,181,182,183,184
*NSET,NSET=FCON1
201,202,203,204,205,206,212,218,224,230,231,232,
233,234,235,236
**
** Define boundary condition
**
*NSET, NSET=FEND
30,31,32,33,34,35
*NSET, NSET=WAXI, GENERATE
101,179,6
*NSET, NSET=FAXI, GENERATE
201,231,6
*BOUNDARY
FEND,2,2
35,XSYMM
25,XSYMM
1,XSYMM
WAXI,XSYMM
FAXI,XSYMM
**
** Define initial condition
**
*NSET,NSET=NALL, GENERATE
1,29,1
2-824
101,184,1
201,236,1
NALL, 2, -523.3
**
** Define history output element set
**
*ELSET,ELSET=FLUID
161,
*ELSET,ELSET=SOLID
201,
*ELSET,ELSET=MIDEL
161,201
*ELSET,ELSET=SHELL
22,
*ELSET,ELSET=ELOUT
MIDEL,
SHELL,
*NSET, NSET=NOUT
27,201
*RESTART, WRITE, NUM=1, TIMEMARKS=NO
*SURFACE,TYPE=NODE,NAME=FCON1
FCON1,
*SURFACE,TYPE=NODE,NAME=WCON1
WCON1,
*SURFACE,TYPE=ELEMENT, NAME=SSIDE1, NO THICK
S1,SNEG
S2,SNEG
S3,SNEG
S3,SPOS
S4,SNEG
*STEP
*DYNAMIC, EXPLICIT
,0.005
*CONTACT PAIR
SSIDE1,WCON1
*CONTACT PAIR
SSIDE2,FCON1
*FILE OUTPUT, NUM=4, TIMEMARKS=NO
*EL FILE, ELSET=FLUID
PRESS,
*EL FILE, ELSET=SOLID
2-825
PEEQ,PE,LE
MISES,PRESS
*EL FILE, ELSET=SHELL
PEEQ,PE,LE
MISES,PRESS
SF,STH
U,V
*ENERGY HISTORY
*NODE HISTORY, NSET=NOUT
U,V
*EL HISTORY,ELSET=MIDEL
S22,PEEQ,MISES,ERV
*EL HISTORY,ELSET=ELOUT
PRESS,
*EL HISTORY,ELSET=SHELL
SF,STH
*EL HISTORY,ELSET=SHELL,SECTION POINT=1
S22,PEEQ,MISES
S22,PEEQ,MISES
S22,PEEQ,MISES
*OUTPUT,FIELD,OP=NEW,NUM=4,TIMEMARKS=NO
*ELEMENT OUTPUT,ELSET=FLUID
PRESS,
*ELEMENT OUTPUT,ELSET=SOLID
PEEQ,PE,LE
MISES,PRESS
PEEQ,PE,LE
MISES,PRESS
SF,STH
*NODE OUTPUT
U,V
*OUTPUT,HISTORY,OP=NEW,TIME INTERVAL=0.0
*ENERGY OUTPUT
U,V
*ELEMENT OUTPUT,ELSET=MIDEL
2-826
S22,PEEQ,MISES,ERV
*ELEMENT OUTPUT,ELSET=ELOUT
PRESS,
SF,STH
S22,PEEQ,MISES
S22,PEEQ,MISES
S22,PEEQ,MISES
*END STEP
2-827
Listing 2.1.9-2
*HEADING
CASK DROP PROBLEM
3D MODEL (coarse mesh)
SECTION CONTROLS USED (KINEMA=ORTHOGONAL,
HOURGLASS=COMBINED)
**
** Create liner mesh
**
** Top shell mesh
*NODE,NSET=N1
1, 0.0,0.0,28.3461
2, 2.3622,0.0,28.3461
3, 0.0,2.3622,28.3461
4, 2.3622,2.3622,28.3461
*NODE,NSET=N2
5, 4.7244,0.0,28.3461
6, 4.7244,2.3622,28.3461
7, 4.7244,4.7244,28.3461
8, 2.3622,4.7244,28.3461
9, 0.0,4.7244,28.3461
*NODE,NSET=N3
20, 11.8109,0.0,28.3461
21,10.911849,4.5198358,28.3461
22,8.3515675,8.3515675,28.3461
23,4.5198358,10.911849,28.3461
24, 0.0,11.8109,28.3461
*NFILL,NSET=N31
N2,N3,3,5
*ELEMENT,TYPE=S4R,ELSET=ETSHELL
1, 1,2,4,3
2, 2,5,6,4
3, 3,4,8,9
4, 4,6,7,8
5, 5,10,11,6
*NSET,NSET=NTSHELL
N1,N2,N3,N31
*ELGEN, ELSET=ETSHELL
5,4,1,1,3,5,4
** First shell wall and middle shell mesh
*NCOPY,OLD SET=N3, NEW SET=N4, CHANGE NUMBER=5,
SHIFT
2-828
0.0,0.0,-1.1811
0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0
SHIFT
0.0,0.0,-17.7163
0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0
*NCOPY,OLD SET=NTSHELL, NEW SET=NMSHELL,
CHANGE NUMBER=89, SHIFT
0.0,0.0,-22.8344
0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0
*NFILL
N4,N5,9,5
SHIFT
0.0,0.0,-20.669
0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0
*NFILL
N5,N6,3,5
** Second shell wall and bottom shell mesh
*NCOPY,OLD SET=NTSHELL, NEW SET=NBSHELL,
0.0,0.0,-28.3461
0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0
SHIFT
0.0,0.0,-26.0627
0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0
*NSET,NSET=N61
109,110,111,112,113
*NFILL
N61,N7,4,5
*ELEMENT,TYPE=S4R
17, 25,26,21,20
69, 109,110,86,85
89, 114,115,110,109
105,153,154,130,129
*ELGEN, ELSET=W1SHELL
17, 4,1,1,13,5,4
*ELGEN, ELSET=W1SHELL
69, 4,1,1
*ELCOPY, ELEMENT SHIFT=72, OLD SET=ETSHELL,
SHIFT NODE=89, NEW SET=EMSHELL
*ELCOPY, ELEMENT SHIFT=36, OLD SET=EMSHELL,
2-829
SHIFT NODE=44, NEW SET=EBSHELL

*ELGEN,ELSET=W2SHELL
89,4,1,1,4,5,4
*ELGEN,ELSET=W2SHELL
105, 4,1,1
*ELSET,ELSET=SHELL1
ETSHELL,W1SHELL,EMSHELL,W2SHELL,EBSHELL
**
** Shell material definition
**
*SHELL SECTION,MATERIAL=A2,ELSET=ETSHELL,
0.3748,5
*SHELL SECTION,MATERIAL=A1,ELSET=W1SHELL,
0.1874,5
*SHELL SECTION,MATERIAL=A2,ELSET=EMSHELL,
0.5,5
*SHELL SECTION,MATERIAL=A2,ELSET=W2SHELL,
0.0591,5
*SHELL SECTION,MATERIAL=A2,ELSET=EBSHELL,
0.0591,5
**
** Stainless steel (304L)
**
*MATERIAL,NAME=A1
*DENSITY
0.00075163,
*ELASTIC
28E6,0.28
*PLASTIC
44300.0,0.
66300.0,0.1
**
** Mild steel (A36)
**
*MATERIAL,NAME=A2
*DENSITY
0.00075163,
*ELASTIC
2-830
28E6,0.28
*PLASTIC
30000.0,0.
**
** Create fluid mesh
**
*NCOPY,OLD SET=NTSHELL,CHANGE NUMBER=200,
NEW SET=WTOP,SHIFT
0.0,0.0,-2.2834
0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0
*NCOPY,OLD SET=WTOP, NEW SET=WMID,
0.0,0.0,-16.8145
0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0
*NCOPY,OLD SET=WTOP, NEW SET=WBOT,
0.0,0.0,-20.551
0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0
*NFILL
WTOP,WMID,9,24
*NFILL
WMID,WBOT,4,24
*ELEMENT,TYPE=C3D8R,ELSET=E1
201, 225,226,228,227,201,202,204,203
202, 226,229,230,228,202,205,206,204
203, 227,228,232,233,203,204,208,209
204, 228,230,231,232,204,206,207,208
205, 229,234,235,230,205,210,211,206
206, 230,235,236,231,206,211,212,207
207, 231,236,237,232,207,212,213,208
208, 232,237,238,233,208,213,214,209
209, 234,239,240,235,210,215,216,211
210, 235,240,241,236,211,216,217,212
211, 236,241,242,237,212,217,218,213
212, 237,242,243,238,213,218,219,214
213, 239,244,245,240,215,220,221,216
214, 240,245,246,241,216,221,222,217
215, 241,246,247,242,217,222,223,218
216, 242,247,248,243,218,223,224,219
*ELCOPY,OLD SET=E1,ELEMENT SHIFT=16,
SHIFT NODE=24
SHIFT NODE=48
2-831

SHIFT NODE=72
SHIFT NODE=96
SHIFT NODE=120
SHIFT NODE=144
SHIFT NODE=168
SHIFT NODE=192
SHIFT NODE=216
SHIFT NODE=240
SHIFT NODE=264
SHIFT NODE=288
*ELSET,ELSET=WATER,GENERATE
201,408,1
*MATERIAL, NAME=MWATER
*DENSITY
9.2E-5,
*EOS, TYPE=USUP
57100.874,0.0,0.0
0.,
*SOLID SECTION, ELSET=WATER, MATERIAL=MWATER,
CONTROL=C
*SECTION CONTROLS, KINEMA=ORTHOGONAL,
HOURGLASS=COMBINED, NAME=C
**
** Create foam mesh
**
*NCOPY,OLD SET=NMSHELL,CHANGE NUMBER=511,
NEW SET=FTOP,SHIFT
0.0,0.0,-0.5
0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0
*NCOPY,OLD SET=NBSHELL,CHANGE NUMBER=587,
NEW SET=FBOT,SHIFT
2-832
0.0,0.0,0.0
0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0
*NFILL
FTOP,FBOT,5,24
SHIFT NODE=400,NEW SET=E2
SHIFT NODE=24
SHIFT NODE=48
SHIFT NODE=72
SHIFT NODE=96
*ELSET,ELSET=FOAM,GENERATE
501,580,1
*MATERIAL, NAME=MFOAM
*ELASTIC
18738,0.0
*DENSITY
2.85E-5,
*FOAM
0.0,223.84,313.35
*FOAM HARDENING
1023.84,0.00
1073.84,0.01
1123.84,0.02
1723.84,0.345
3023.84,0.44
5923.84,0.49
12223.84,0.51
8E5,2.00
*SOLID SECTION, ELSET=FOAM, MATERIAL=MFOAM,
CONTROL=C
**
** Define fluid and foam contact node set
**
*NSET, NSET=WCONT, GENERATE
201,224,1
220,532,24
221,533,24
222,534,24
223,535,24
2-833
224,536,24
513,536,1
*NSET, NSET=FCONT, GENERATE
601,624,1
620,740,24
621,741,24
622,742,24
623,743,24
624,744,24
721,744,1
**
** Define boundary condition
**
*NSET,NSET=FEND, GENERATE
134,157,1
*NSET,NSET=YANODES
1,2,5,10,15,20,25,30,35,40,45,50,55,60,65,70,
75,80,85,90,91,94,99,104,109,114,119,124,129,
134,135,138,143,148,153,513,514,517,522,527,
532,721,722,725,730,735,740,
*NSET,NSET=YANODES,GENERATE
201,513,24
202,514,24
205,517,24
210,522,24
215,527,24
220,532,24
601,721,24
602,722,24
605,725,24
610,730,24
615,735,24
620,740,24
*NSET,NSET=XANODES
1,3,9,14,19,24,29,34,39,44,49,54,59,64,69,74,
79,84,89,90,92,98,103,108,113,118,123,128,133,
134,136,142,147,152,157,513,515,521,526,531,
536,721,723,729,734,739,744,
*NSET,NSET=XANODES,GENERATE
201,513,24
203,515,24
209,521,24
214,526,24
2-834
219,531,24
224,536,24
601,721,24
603,723,24
609,729,24
614,734,24
619,739,24
624,744,24
*BOUNDARY
FEND,ZSYMM
XANODES,XSYMM
YANODES,YSYMM
**
** Define initial condition
**
*NSET,NSET=NALL,GENERATE
1,133,1
201,536,1
601,720,1
NALL, 3, -523.3
**
** Define history output element set
**
*ELSET,ELSET=FLUID
393,
*ELSET,ELSET=SOLID
501,
*ELSET,ELSET=MIDEL
393,501
*ELSET,ELSET=SHELL
97,
*ELSET,ELSET=ELOUT
MIDEL,
SHELL,
*NSET, NSET=NOUT
119,601
*NSET, NSET=NOUT, ELSET=ELOUT
*RESTART, WRITE, NUM=1, TIMEMARKS=NO
ETSHELL,SNEG
W1SHELL,SNEG
EMSHELL,SPOS
2-835
*SURFACE,TYPE=NODE,NAME=FCONT
FCONT,
EMSHELL,SNEG
W2SHELL,SNEG
EBSHELL,SPOS
*SURFACE,TYPE=NODE,NAME=WCONT
WCONT,
*STEP
*DYNAMIC, EXPLICIT
,0.005
*CONTACT PAIR
SSIDE1,WCONT
*CONTACT PAIR
SSIDE2,FCONT
*FILE OUTPUT, NUM=4, TIMEMARKS=NO
*EL FILE, ELSET=FLUID
PRESS,
*EL FILE, ELSET=SOLID
PEEQ,PE,LE
MISES,PRESS
*EL FILE, ELSET=SHELL
PEEQ,PE,LE
MISES,PRESS
SF,STH
U,V
*ENERGY HISTORY
*NODE HISTORY, NSET=NOUT
U,V
*EL HISTORY,ELSET=MIDEL
S33,PEEQ,MISES,ERV
*EL HISTORY,ELSET=ELOUT
PRESS,
*EL HISTORY,ELSET=SHELL
SF,STH
S33,PEEQ,MISES
S33,PEEQ,MISES
2-836
S33,PEEQ,MISES
*OUTPUT,FIELD,VARIABLE=PRESELECT,OP=NEW,NUM=1,
TIMEMARKS=NO
*OUTPUT,FIELD,OP=NEW,NUM=4,TIMEMARKS=NO
*ELEMENT OUTPUT,ELSET=FLUID
PRESS,
*ELEMENT OUTPUT,ELSET=SOLID
PEEQ,PE,LE
MISES,PRESS
PEEQ,PE,LE
MISES,PRESS
SF,STH
U,V
*OUTPUT,HISTORY,OP=NEW,TIME INTERVAL=0.0
*ENERGY OUTPUT
U,V
*ELEMENT OUTPUT,ELSET=MIDEL
S33,PEEQ,MISES,ERV
*ELEMENT OUTPUT,ELSET=ELOUT
PRESS,
SF,STH
S33,PEEQ,MISES
S33,PEEQ,MISES
S33,PEEQ,MISES
*END STEP
2.1.10 Oblique impact of a copper rod

This example simulates a high velocity, oblique impact of a copper rod into a rigid wall. Extremely
high plastic strains develop at the crushed end of the rod, resulting in severe local mesh distortion.
Adaptive meshing is used to reduce element distortion and to obtain an accurate and economical
solution to the problem.
Problem description
2-837
The model geometry is depicted in Figure 2.1.10-1. A cylindrical rod, measuring 32.4 ´ 3.2 mm,
impacts a rigid wall with an initial velocity of vy =340 m/sec. The wall is perpendicular to the x-z plane
and makes an angle of 30° with the x-y plane. The half-symmetric finite element model is shown in
Figure 2.1.10-2. Symmetry boundary conditions are applied at the y=0 plane. The rod is meshed with
CAX4R elements, and the wall is modeled as an analytical rigid surface using the *SURFACE,
TYPE=CYLINDER option in conjunction with the *RIGID BODY option. Coulomb friction is
assumed between the rod and the wall, with a friction coefficient of 0.2. The analysis is performed for
a period of 120 microseconds.
The rod is modeled as a Johnson-Cook, elastic-plastic material with a Young's modulus of 124 GPa, a
Poisson's ratio of 0.34, and a density of 8960 kg/m 3. The Johnson-Cook model is appropriate for
modeling high-rate impacts involving metals. The Johnson-Cook material parameters are taken from
Johnson and Cook (1985) in which the following constants are used: A = 90 MPa, n = 0.31, m =
1.09, C = 0.025, and "_o = 1 s-1. Furthermore, the melting temperature is 1058°C, and the transition
temperature is 25°C. Adiabatic conditions are assumed with a heat fraction of 50%. The specific heat
of the material is 383 J/Kg°C, and the thermal expansion coefficient is 0.00005°C -1.
Adaptive meshing
A single adaptive mesh domain that incorporates the entire rod is defined. Symmetry boundary
conditions are defined as Lagrangian surfaces (the default), and contact surfaces are defined as sliding
contact surfaces (the default). Because the impact phenomenon modeled in this example is an
extremely dynamic event with large changes in geometry occurring over a relatively small number of
increments, it is necessary to increase the frequency and intensity of adaptive meshing. The frequency
value is reduced to 5 increments from a default value of 10, and the number of mesh sweeps used to
smooth the mesh is increased to 3 from the default value of 1. The default values are used for all other
adaptive mesh controls.

Deformed shape plots at 40, 80, and 120 microseconds are shown in Figure 2.1.10-3, Figure 2.1.10-4,
and Figure 2.1.10-5, respectively. The rod rebounds from the wall near the end of the analysis.
High-speed collisions such as these result in significant amounts of material flow in the impact zone. A
pure Lagrangian analysis of this finite element model fails as a result of excessive distortions.
Continuous adaptive meshing allows the analysis to run to completion while retaining a high-quality
mesh. The kinetic and internal energy histories are plotted in Figure 2.1.10-6. Most of the initial
kinetic energy is converted to internal energy as the rod is plastically deformed. Both energy curves
plateau as the rod rebounds from the wall.
Input files
ale_rodimpac_inclined.inp
Analysis using adaptive meshing.
ale_rodimpac_inclined_nodelem.inp
2-838
Reference
· Johnson, G. R. and W. H. Cook, "Fracture Characteristics of Three Metals Subjected to Various
Strains, Strain Rates, Temperatures and Pressures, " Engineering Fracture Mechanics, 21, pp.
31-48, 1985.
Figures
Figure 2.1.10-1 Model geometry.
Figure 2.1.10-2 Initial configuration.
2-839
Figure 2.1.10-3 Deformed configuration at 40 microseconds.
2-840
Figure 2.1.10-6 Time history of kinetic and internal energies of the rod.
Sample listings
2-841
Listing 2.1.10-1
*HEADING
OBLIQUE IMPACT OF A COPPER ROD(3D MODEL)
Units - N, m ,second
**
*INCLUDE,INPUT=ale_rodimpac_inclined_nodelem.inp
**
** Representative material properties for copper
**
*MATERIAL,NAME=COPPER
*ELASTIC
124.E9, 0.34
** Young's modulus unit: Pa
*PLASTIC,HARDENING=JOHNSON COOK
90.E6, 292.E6, 0.31, 1.09, 1058., 25.
** A & B unit: Pa
*RATE DEPENDENT,TYPE=JOHNSON COOK
0.025, 1.0
**unit:reference strain rate has a unit of s^(-1)
0.5,
** dimensionless
*SPECIFIC HEAT
383.,
** unit: J/kgK
*EXPANSION
5.E-5,
** unit: K^(-1)
*DENSITY
8.96E3,
** unit: kg/m^3
*SOLID SECTION,ELSET=ROD,MATERIAL=COPPER,
CONTROLS=SECT
*SECTION CONTROLS,NAME=SECT,
KINEMATICS=ORTHOGONAL,HOURGLASS=VISCO
*BOUNDARY
SPLANE1, 2
200000, 1, 6
ROD0, 25.
2-842
BODY, 3, -340.
**
** Run the problem for 120 microseconds
**
*SURFACE,TYPE=CYLINDER,NAME=RSURF
-20.E-3,0.,-13.3945E-3,20.E-3,0.,-13.3945E-3
-20.E-3,1.,-13.3945E-3
START, 0., 0.
LINE, 40.E-3, 23.094E-3
*SURFACE,TYPE=ELEMENT,NAME=RODSURF,
REGION TYPE=SLIDING
ROD,
*RIGID BODY,REFNODE=200000,
ANALYTICAL SURFACE =RSURF
*STEP
*DYNAMIC,EXPLICIT,ADIABATIC
,120.E-6
*CONTACT PAIR,INTERACTION=INTER
RSURF,RODSURF
*SURFACE INTERACTION,NAME=INTER
*FRICTION
0.2,
*EL FILE
PEEQ,MISES,
*NODE FILE
U,
*ENERGY FILE
**
*ADAPTIVE MESH,ELSET=ROD,FREQUENCY=5,
MESH SWEEPS=3
**
*END STEP
2.1.11 Water sloshing in a baffled tank

This example illustrates the use of adaptive meshing to model an inviscid fluid sloshing inside a
baffled tank. The overall structural response resulting from the coupling between the water and tank,
rather than a detailed solution in the fluid, is sought. Adaptive meshing permits the investigation of
this response over longer time periods than a pure Lagrangian approach would because the mesh
entanglement that occurs in the latter case is prevented.
2-843
Problem description
The geometry for the problem is shown in Figure 2.1.11-1. The model consists of a baffled tank filled
with water. The baffle, which is attached to the sides and top of the tank, does not penetrate the entire
depth of the water. The tank measures 508 ´ 152.4 ´ 152.4 mm (20 ´ 6 ´ 6 inches), and the baffle
measures 3.048 ´ 152.4 ´ 121.92 mm (0.12 ´ 6 ´ 4.8 inches). The tank is filled with 101.6 mm (4
inches) of water.
A cutaway view of the finite element model that displays the baffle and the water is shown in Figure
2.1.11-2. The top of the tank is not modeled because the water is not expected to come into contact
with it. The tank is modeled as a rigid body and is meshed with R3D4 elements. The baffle is modeled
as a deformable body and is meshed with S4R elements. A graded mesh of C3D8R elements is used
for the water, with more refinement adjacent to the baffle where significant deformations are expected.
In sloshing problems water can be considered an incompressible and inviscid material. An effective
method for modeling water in ABAQUS/Explicit is to use a simple Newtonian viscous shear model
and a linear Us ¡ Up equation of state for the bulk response. The bulk modulus functions as a penalty
parameter for the incompressible constraint. Since sloshing problems are unconfined, the bulk modulus
chosen can be two or three orders of magnitude less than the actual bulk modulus and the water will
still behave as an incompressible medium. The shear viscosity also acts as a penalty parameter to
suppress shear modes that could tangle the mesh. The shear viscosity chosen should be small because
water is inviscid; a high shear viscosity will result in an overly stiff response. An appropriate value for
the shear viscosity can be calculated based on the bulk modulus. To avoid an overly stiff response, the
internal forces arising due to the deviatoric response of the material should be kept several orders of
magnitude below the forces arising due to the volumetric response. This can be done by choosing an
elastic shear modulus that is several orders of magnitude lower than the bulk modulus. If the
Newtonian viscous deviatoric model is used, the shear viscosity specified should be on the order of an
equivalent shear modulus, calculated as mentioned earlier, scaled by the stable time increment. The
expected stable time increment can be obtained from a datacheck analysis of the model. This method is
a convenient way to approximate a shear strength that will not introduce excessive viscosity in the
material.
In addition, if a shear model is defined, the hourglass control forces are calculated based on the shear
stiffness of the material. Thus, in materials with extremely low or zero shear strengths such as inviscid
fluids, the hourglass forces calculated based on the default parameters are insufficient to prevent
spurious hourglass modes. Therefore, a sufficiently high hourglass scaling factor is used to increase the
resistance to such modes. This analysis methodology is verified in ``Water sloshing in a pitching tank,''
Section 1.11.6 of the ABAQUS Benchmarks Manual.
For this example the linear Us ¡ Up equation of state is used with a wave speed of 45.85 m/sec (1805
in/sec) and a density of 983.204 kg/m 3 (0.92 ´ 10-4 lb sec2/in4). The wave speed corresponds to a bulk
modulus of 2.07 MPa (300 psi), three orders of magnitude less than the actual bulk modulus of water,
2.07 GPa (3.0 ´ 105 psi). The shear viscosity is chosen as 1.5 ´ 10-8 sec-1. The baffle is modeled as a
Mooney-Rivlin elastomeric material with hyperelastic constants C10 = 689480 Pa (100 psi) and C01 =
172370 Pa (25 psi) and a density of 10900.74 kg/m 3 (1.02 ´ 10-3 lb sec2/in4).
2-844
Pure master-slave contact is defined between the tank and the water; balanced master-slave contact is
defined between the baffle and the water. The bottom edge of the baffle has nodes in common with the
underlying water surface. This prevents relative slip between the bottom edge of the baffle and the
water immediately below it. The motion of the other edges of the baffle coincides with that of the
tank.
The water is subjected to gravity loading. Consequently, an initial geostatic stress field is defined to
equilibrate the stresses caused by the self-weight of the water. A velocity pulse in the form of a sine
wave with an amplitude of 63.5 mm (2.5 inches) and a period of 2 seconds is prescribed for the tank in
both the x- and y-directions simultaneously. All remaining degrees of freedom for the tank are fully
constrained. The sloshing analysis is performed for two seconds.
Adaptive meshing
A single adaptive mesh domain that incorporates the water is defined. Sliding boundary regions are
used for all contact surface definitions on the water (the default). Because the sloshing phenomenon
modeled in this example results in large mesh motions, it is necessary to increase the frequency and
intensity of adaptive meshing. The frequency value is reduced to 5 increments from a default value of
10, and the number of mesh sweeps used to smooth the mesh is increased to 3 from a default value of
1. The SMOOTHING OBJECTIVE parameter on the *ADAPTIVE MESH CONTROLS option is set
to GRADED so that the initial mesh gradation of the water is preserved while continuous adaptive
meshing is performed. The default values are used for all other parameters and controls.

Figure 2.1.11-3, Figure 2.1.11-4, and Figure 2.1.11-5 show the deformed mesh configuration at t = 1.2
s, t = 1.6 s, and t = 2.0 s, respectively. Four time histories of the vertical displacement of the water
level are shown in Figure 2.1.11-6; these correspond to the water level at the baffle in the front and
back of the left and right bays. The locations at which the time histories are measured are denoted A,
B, C, and D in Figure 2.1.11-1. An analysis such as this could be used to design a baffle that attenuates
sloshing at certain frequencies. Using adaptivity in ABAQUS/Explicit is appropriate for sloshing
problems in which the structural response is of primary interest. It is generally not possible to model
such flow behaviors as splashing or complex free surface interactions. Furthermore, surface tension is
not modeled.
Input file
ale_water_sloshing.inp
ale_water_sloshingel.inp
Element data.
Figures
2-845
Figure 2.1.11-1 Model geometry.
Figure 2.1.11-2 Initial configuration (front of rigid tank is not shown).
Figure 2.1.11-3 Deformed configuration of the water and the baffle at 1.2 seconds.
2-846
Figure 2.1.11-6 Time histories of the vertical displacement of the water at the baffle at both the front
and back of the left and right bays.
2-847
Sample listings
2-848
Listing 2.1.11-1
*HEADING
WATER SLOSHING IN A BAFFLED TANK
Unit lb, in, sec
*NODE
1, 10.00, 0.0, 0.0
21, 20.00, 0.0, 0.0
127, 10.00, 0.0, 1.2
147, 20.00, 0.0, 1.2
316, 10.00, 0.0, 4.0
336, 20.00, 0.0, 4.0
5041,10.00, 6.0, 0.0
5061,20.00, 6.0, 0.0
5167,10.00, 6.0, 1.2
5187,20.00, 6.0, 1.2
5356,10.00, 6.0, 4.0
5376,20.00, 6.0, 4.0
75000, 25.00, 20.0
*NGEN, NSET=A1
1, 127,21
*NGEN, NSET=A2
5041, 5167,21
*NFILL, NSET=A
A1,A2,15,336
*NGEN, NSET=B1
21,147,21
*NGEN, NSET=B2
5061,5187,21
*NFILL, NSET=B
B1,B2,15,336
*NFILL,NSET=RIGHT1,BIAS=.9
A, B, 20, 1
*NGEN, NSET=C1
127, 316,21
*NGEN, NSET=C2
5167, 5356,21
*NFILL, NSET=C
C1,C2,15,336
*NGEN, NSET=D1
147,336,21
2-849
*NGEN, NSET=D2
5187,5376,21
*NFILL, NSET=D
D1,D2,15,336
*NFILL,NSET=RIGHT2,BIAS=.9
C, D, 20, 1
*NSET,NSET=RIGHT
RIGHT1,RIGHT2
*ELEMENT, TYPE=C3D8R
1,1,2,338,337,22,23,359,358
*ELGEN, ELSET=WATER1
1,20,1,1,15,21,20,15,336,300,
*****************************************
*NCOPY,CHANGE NUMBER=5376,OLD SET=RIGHT,
NEW SET=LEFT,REFLECT=MIRROR
10.,0., 20.,10.,3., 20.
10.,0., -5.,
*ELCOPY,ELEMENT SHIFT=4500,OLD SET=WATER1,
SHIFT NODES=5376,REFLECT,NEW SET=WATER2
*ELSET,ELSET=WATER
WATER1,WATER2
*NODE
10753, 0.0,0.0,0.0
10754,20.0,0.0,0.0
10755,20.0,0.0,6.0,
10756,0.00,0.0,6.0,
10757, 0.0,6.0,0.0,
10758,20.0,6.0,0.0,
10759,20.0,6.0,6.0,
10760,0.00,6.0,6.0,
*ELEMENT, TYPE=R3D4, ELSET=RIGID
9001,10756,10755,10754,10753
9002,10757,10758,10759,10760
9003,10755,10759,10758,10754
9004,10753,10757,10760,10756
9005,10753,10754,10758,10757
*NODE
10761,10.00,-0.1,1.2
10776,10.00,6.1,1.2
11001,10.00,-0.1,6.0
11016,10.00,6.1,6.0
*NGEN, NSET=S1
10761,10776,1
2-850
*NGEN, NSET=S2
11001,11016,1
*NFILL,NSET=S
S1,S2,15, 16
*NSET,NSET=EDGES, GEN
10777,11001,16
10792,11016,16
11001,11016,1
*NSET,NSET=OUT
316,5356,5692,10732
*ELEMENT, TYPE=S4R
9006,10761,10777,10778,10762
*ELGEN, ELSET=BAFFLE
9006,15,1,1,15,16,15
*INCLUDE, INPUT=ale_water_sloshingel.inp
*SHELL SECTION, ELSET=BAFFLE, MATERIAL=BAFFLE
.12,
*MATERIAL, NAME=BAFFLE
*HYPERELASTIC,POLYNOMIAL, N=1
100., 25.
*DENSITY
1.07E-3,
*SOLID SECTION, ELSET=WATER, MATERIAL=WATER,
CONTROLS=SECT
*SECTION CONTROLS, NAME=SECT,
KINEMATICS=ORTHOGONAL, HOURGLASS=STIFFNESS
10000.,
***K=300,K/G=1.E+6, (Actual K of water = 300000)
*MATERIAL, NAME=WATER
*EOS, TYPE=USUP
1805.7878,0.,0.
*EOS SHEAR, TYPE=VISCOUS
1.5E-8,
*DENSITY
.92E-4,
*BOUNDARY
75000,3,6,0.0
EDGES,3,6,0.0
*AMPLITUDE,NAME=AMP,DEFINITION=SMOOTH STEP
0.,0.,.25,1.76777,.5,2.5,.75,1.76777,
1.0,0.0,10.0,0.0
*INITIAL CONDITIONS, TYPE=STRESS, GEOSTATIC
WATER,0.,4.,-.1420,0.,1.,1.,
2-851
*SURFACE,TYPE=ELEMENT, NAME=BAFFLESURF1,NOTHICK
BAFFLE,SNEG
*SURFACE,TYPE=ELEMENT, NAME=BAFFLESURF2,NOTHICK
BAFFLE,SPOS
*SURFACE,TYPE=ELEMENT, NAME=RSURF
RIGID,SPOS
*SURFACE,TYPE=ELEMENT, NAME=WATSURF1
WATER,
OUT1,S6
OUT11,S2
OUT2,S3
OUT21,S2
POSITION=CENTER OF MASS,ELSET=RIGID
*STEP
*DYNAMIC, EXPLICIT
,2.0
*DLOAD
WATER, GRAV, 386., 0., 0.,-1.
*BOUNDARY,TYPE=VELOCITY,AMPLITUDE=AMP
75000,1,2,1.
EDGES,1,2,1.
*CONTACT PAIR,INTERACTION=WATRIG
RSURF,WATSURF1
*CONTACT PAIR,INTERACTION=WATBAFFLE,WEIGHT=1.0,MECHANICAL CONSTRAINT=PENA
BAFFLESURF1,WATSURF2
*CONTACT PAIR,INTERACTION=WATBAFFLE,WEIGHT=1.0,MECHANICAL CONSTRAINT=PENA
BAFFLESURF2,WATSURF3
*SURFACE INTERACTION, NAME=WATRIG
*SURFACE INTERACTION, NAME=WATBAFFLE
*FILE OUTPUT, NUMBER=4, TIMEMARKS=YES
*EL FILE,ELSET=OUT
S,
PRESS,
*NODE FILE,NSET=EDGES
U,V
*NODE HISTORY,NSET=EDGES
U,V
*NODE HISTORY, NSET=OUT
U,V
2-852
*EL HISTORY,ELSET=OUT
S,
PRESS,
SHRMOD, DILMOD
*ENERGY HISTORY
DT,
*ADAPTIVE MESH,ELSET=WATER,CONTROLS=CONTROLS,
FREQUENCY=5,MESH SWEEPS=3
*ADAPTIVE MESH CONTROLS,NAME=CONTROLS,
SMOOTHING=GRADED,TRANSITION FEATURE ANGLE=0.0
*END STEP
2.1.12 Rigid multi-body mechanism

This example illustrates the use of connector elements to model kinematic constraints between rigid
bodies in a multi-body mechanism.
Problem description
The crank mechanism considered here transmits a rotational motion through two universal joints and
then converts the rotation into translational motion of two slides. The mechanism is modeled using
nine rigid components attached with eight connector elements. The various kinematic constraints
modeled with connector elements include TRANSLATOR, which allows relative translation along a
line but no rotations; HINGE, which allows one relative rotation and fixes relative translations;
CYLINDRICAL, which allows relative translation along a line and relative rotation about that line;
JOIN, which fixes relative translations but leaves the rotations free; PLANAR, which keeps a point on
a plane and allows only relative rotations about the normal to that plane; and UJOINT, which fixes the
relative translations and enforces a universal constraint on the relative rotations. The complete model
is shown in Figure 2.1.12-1.
The axes of rotation of the small and large disks are parallel but offset. A constant angular velocity of
the small disk is specified about its axis with a velocity boundary condition on its rigid body reference
node. All other degrees of freedom of the rigid body reference node are fixed. The rotational motion of
the small disk is transmitted to the large disk through two UJOINT connections and a rigid link. A
UJOINT connection, or a universal rotation constraint with shared translational degrees of freedom,
between two nonaligned shafts will not transmit constant angular velocity. However, two
symmetrically placed universal constraints, as here, will produce constant angular velocity coupling
between the two disks. The large disk is connected to a rigid circular rod with a JOIN connection. A
JOIN connection is equivalent to a ball-and-socket or a spherical joint. The circular rod connects
through a sleeve to a flat block. The rod and sleeve constraint is modeled with a CYLINDRICAL
connection, which allows the sleeve to translate along and rotate about the rod. The attachment of the
circular rod to the flat block is a HINGE connection, which allows only a single relative rotation about
2-853
the shared hinge axis. The flat block, in turn, is assumed to slide between two fixed parallel plates.
This sliding constraint is modeled with a PLANAR connection. The sleeve on the circular rod is
connected to a square-section sleeve on the square rod with a HINGE connection. The square rod is
fixed in space. The square-section sleeve slides along the square bar without rotating. This sliding
constraint is modeled with a TRANSLATOR connection.

Figure 2.1.12-2 shows the position of the mechanism at various times. By visual inspection it can be
observed that the connector elements are enforcing the correct kinematic constraints.
In this model there are nine rigid bodies with 6 degrees of freedom each, accounting for 54 rigid body
degrees of freedom. The eight connector elements eliminate 33 rigid body degrees of freedom through
kinematic constraints (enforced via Lagrange multipliers) as itemized in Table 2.1.12-1. Hence, the
model has 21 rigid body degrees of freedom to be specified as boundary conditions or determined by
the solution. In this case all remaining rigid body degrees of freedom are specified as boundary
conditions, with the z-component of angular velocity specified for the small disk and 20 additional
fixed boundary conditions used.
Input files
rigmultimech_exp.inp
ABAQUS/Explicit analysis.
rigmultimech_std.inp
ABAQUS/Standard analysis.
rigmultimech_bulk.inp
Node and element bulk data for the rigid bodies.
Table
Table 2.1.12-1 Rigid body degrees of freedom eliminated by kinematic constraints.

Connection type Number of Total rigid body dofs
kinematic eliminated from
constraints model
UJOINT (2) 4 8
JOIN 3 3
CYLINDRICAL 4 4
HINGE (2) 5 10
PLANAR 3 3
TRANSLATOR 5 5
Total eliminated: 33
Figures
2-854
Figure 2.1.12-1 Rigid mechanism model.
Figure 2.1.12-2 Time history of the motion of the mechanism during the first revolution.
2-855
Sample listings
2-856
Listing 2.1.12-1
*HEADING
Rigid, multi-body mechanism - connector elements
** node and element definitions:
*INCLUDE, INPUT=rigmultimech_bulk.inp
*****************************
*ELSET, ELSET=SQUAREROD, GEN
1,4,1
*ELSET, ELSET=SQUARESLEAVE, GEN
5,92,1
*ELSET, ELSET=ROUNDROD, GEN
93,172,1
*ELSET, ELSET=ROUNDSLEAVE, GEN
173,260,1
*ELSET, ELSET=SLIDINGBLOCK, GEN
261,272,1
*ELSET, ELSET=RACEWAY, GEN
273,278,1
*ELSET, ELSET=LARGEDISK, GEN
1001,1174,1
*ELSET, ELSET=DISKLINK, GEN
1175, 1194,1
*ELSET, ELSET=SMALLDISK, GEN
1195, 1368,1
*****************************
** Square rod
*NSET, NSET=ADD1
19
*RIGID BODY,ELSET=SQUAREROD,REF NODE=10000,
POSITION=CENTER OF MASS,TIE NSET=ADD1
** Square sleave
*NSET, NSET=ADD2
20,218
*RIGID BODY,ELSET=SQUARESLEAVE,REF NODE=20000,
** Round rod
*NSET, NSET=ADD3
385,746,1034
*RIGID BODY,ELSET=ROUNDROD,REF NODE=30000,
** Round sleave
*NSET, NSET=ADD4
2-857
219,386
*RIGID BODY,ELSET=ROUNDSLEAVE,REF NODE=40000,
** Large disk
*NSET, NSET=ADD5
747,768
*RIGID BODY,ELSET=LARGEDISK,REF NODE=70000,
** Sliding block
*NSET, NSET=ADD6
1035, 1040
*RIGID BODY,ELSET=SLIDINGBLOCK,REF NODE=50000,
** Raceway
*NSET,NSET=ADD7
1041,
*RIGID BODY,ELSET=RACEWAY,REF NODE=60000,
** DiskLink
*NSET,NSET=ADD8
795,798
*RIGID BODY,ELSET=DISKLINK,REF NODE=80000,
** Smalldisk
*NSET,NSET=ADD9
821,
*RIGID BODY,ELSET=SMALLDISK,REF NODE=90000,
**
*NSET, NSET=REF
10000,20000,30000,40000,50000,60000,70000,80000,
90000
*****************************
** Square rod - square sleave
*****************************
*ELEMENT, TYPE = CONN3D2, ELSET=SR-SS
5001, 19, 20
*ORIENTATION, DEFINITION=COORDINATES, NAME=SR-SS
1, 0, 0, 0, 1, 0
*CONNECTOR SECTION, ELSET=SR-SS
translator,
SR-SS,
*****************************
2-858
** Square sleave - round sleave

*****************************
*ELEMENT, TYPE = CONN3D2, ELSET=SS-RS
5002, 218, 219
*ORIENTATION, DEFINITION=COORDINATES, NAME=SS-RS
0, 0, 1, 1, 0, 0
*CONNECTOR SECTION, ELSET=SS-RS
hinge,
SS-RS,
*****************************
** Round rod - round sleave
*****************************
*ELEMENT, TYPE = CONN3D2, ELSET=RR-RS
5003, 385, 386
*ORIENTATION, DEFINITION=COORDINATES, NAME=RR-RS
0, 1, 0, 0, 0, 1
*CONNECTOR SECTION, ELSET=RR-RS
cylindrical,
RR-RS
*****************************
** Round rod - large disk
*****************************
*ELEMENT, TYPE = CONN3D2, ELSET=RR-LD
5004, 746, 747
*ORIENTATION, DEFINITION=COORDINATES, NAME=RR-LD
1, 0, 0, 0, 1, 0
*CONNECTOR SECTION, ELSET=RR-LD
join,
RR-LD
*****************************
** Round rod - sliding block
*****************************
*ELEMENT, TYPE = CONN3D2, ELSET=RR-SB
5005, 1034, 1035
*ORIENTATION, DEFINITION=COORDINATES, NAME=RR-SB
0, 0, 1, 1, 0, 0
*CONNECTOR SECTION, ELSET=RR-SB
hinge,
RR-SB
*****************************
** Sliding block - raceway
*****************************
*ELEMENT, TYPE = CONN3D2, ELSET=SB-R
2-859
5006, 1040, 1041

*ORIENTATION, DEFINITION=COORDINATES, NAME=SB-R
1, 0, 0, 0, 1, 0
*CONNECTOR SECTION, ELSET=SB-R
slideplane, revolute
SB-R,
*****************************
** Large disk - disklink
*****************************
*ELEMENT, TYPE = CONN3D2, ELSET=LD-DL
5007, 768, 795
*ORIENTATION, DEFINITION=COORDINATES, NAME=LD-DL
-1, 0, 0, 0, 0, 1
*ORIENTATION, DEFINITION=COORDINATES, NAME=DL-LD
-.8660254,0.,.5,.5,0.,.8660254
*CONNECTOR SECTION, ELSET=LD-DL
ujoint,
LD-DL,DL-LD
*****************************
** Small Disk - Disk Link
*****************************
*ELEMENT, TYPE = CONN3D2, ELSET=SD-DL
5008, 821, 798
*ORIENTATION, DEFINITION=COORDINATES, NAME=SD-DL
-1, 0, 0, 0, 0, 1
*ORIENTATION, DEFINITION=COORDINATES, NAME=DL-SD
-.8660254,0.,.5,.5,0.,.8660254
*CONNECTOR SECTION, ELSET=SD-DL
ujoint,
SD-DL,DL-SD
*BOUNDARY
10000,1,2,0.0
10000,4,6,0.0
70000,1,5,0.0
90000,1,2,0.0
90000,4,5,0.0
60000,1,6,0.0
*AMPLITUDE, NAME=AMP, DEFINITION=SMOOTH STEP
0.0,0.0,.00375,23.26,3.0,23.26
***********************************
*NSET, NSET=CONNECTIONNODES
746,747,386,385,20,19,
219,218,1034,1035,1040,1041,
2-860
768,795,798,821
*NSET,NSET=OUT
CONNECTIONNODES,REF
*STEP,NLGEOM=YES
*DYNAMIC,EXPLICIT,DIRECT
1.8E-4,0.18
*BOUNDARY, TYPE=VELOCITY, AMPLITUDE=AMP
90000,6,6,5.0
***************************************
*FILE OUTPUT,TIMEMARKS=NO,NUM=10
*NODE FILE,NSET=OUT
U,V,A
*ENERGY FILE
*OUTPUT,FIELD,NUM=300
*NODE OUTPUT
U,V,A
*NODE OUTPUT,NSET=OUT
U,V,A
*END STEP
2.2 Mode-based dynamic analyses

2.2.1 Analysis of a rotating fan using superelements and cyclic
symmetry model
This example illustrates the single and multi-level substructure capability of ABAQUS for problems
where the part being modeled consists of repeated structures. It also demonstrates the capability of
ABAQUS to analyze cyclic symmetric models using the *CYCLIC SYMMETRY MODEL option.
Some of the limitations of modeling a structure using superelements or using cyclic symmetry are also
discussed.
Geometry and material

The structure is a fan consisting of a central hub and four blades, as shown in Figure 2.2.1-1. The
blades and the hub are made up of S4R shell elements. The material is elastic, with a Young's modulus
of 2 ´ 1011 Pa and a Poisson's ratio of 0.29. The density of the material is 7850 kg/m 3. All nodes along
the mounting hole in the hub are fixed.
Models
Four different models are considered, as follows:
1. The fan is modeled as a single structure (no superelements).
2-861
2. One quadrant of the fan, consisting of a quarter of the hub and a single blade, is reduced to a
superelement. The fan is then modeled with four superelements (a single-level superelement
structure). During superelement generation all degrees of freedom are retained for the nodes along
the edges of the hub in each quadrant as well as one node at the blade tip (see Figure 2.2.1-2).
3. A single fan blade is reduced to a superelement, which is then combined with one-quarter of the
hub to form a higher level superelement. Four of these superelements are then combined to form
the fan (similar to the single-level superelement), thus forming a multi-level superelement
structure. Nodes along the base of the fan blade and one node at the tip of the blade have all their
degrees of freedom retained during generation of the fan blade superelement as shown in Figure
2.2.1-2. At the higher level superelement generation stage, nodes along the edge of the hub in each
quadrant as well as the node at the blade tip have their degrees of freedom retained.
4. One quadrant of the fan, consisting of a quarter of the hub and a single blade, is modeled with and
without superelements as a datum sector for the *CYCLIC SYMMETRY MODEL option. Two
surfaces, which are at 90° to each other, are chosen to serve as the slave and master surfaces for
the *TIE, CYCLIC SYMMETRY option. The finite element mesh contains matching nodes on the
symmetry surfaces; therefore, both surfaces are defined with the *SURFACE, TYPE=NODE
option. The axis of cyclic symmetry is parallel to the global z-axis and passes through the point on
the x-y plane with coordinates (3.0, 3.0). The cyclic symmetry model is shown in Figure 2.2.1-3.
The entire model consists of four repetitive sectors.
Both a frequency analysis and a static analysis are performed on the first three models. Stress- and
load-stiffening effects due to the centrifugal loading on the fan are built into the superelement stiffness
during generation using the *PRELOAD HISTORY option with the NLGEOM parameter included on
the *STEP option. To get the proper stress stiffening in the hub of the multi-level superelement, the
centrifugal load defined in the lowest level superelement (the blade) needs to be captured with the
*SLOAD CASE option and must be applied as a preload with the *SLOAD option in the next level
superelement.
The reduced mass matrix for each superelement is generated by including the *SUPER MASS option
during the superelement generation. To improve the representation of the superelement's dynamic
behavior in the global analysis, dynamic modes are extracted by including the MODES=m parameter
on the *SUPER MASS option and running a *FREQUENCY preload step that extracts at least m
frequencies. The reduced mass matrix obtained by the default value of m = 0 corresponds to the
Guyan reduction technique, while m > 0 corresponds to the restrained mode addition technique. In the
"Results and discussion" section below the solution obtained for the model without superelements (the
"full model") is used as the reference solution.
For the cyclic symmetry model without superelements, the eigenvalue extraction procedure was
performed on the preloaded structure. The nonlinear static step has the centrifugal load applied to the
blade. Fifty eigenvalues were requested using the Lanczos eigenvalue solver, which is the only
eigensolver that can be used for *FREQUENCY analysis with the *CYCLIC SYMMETRY MODEL
option. The *SELECT CYCLIC SYMMETRY MODES option is omitted; therefore, the eigenvalues
are being extracted for all possible (three) cyclic symmetry modes. In the discussion that follows the
2-862
solution obtained for the cyclic symmetry model is compared with the solution for the entire 360°
model as the reference solution. A similar simulation was performed for the cyclic symmetry model
with superelements but without the preload step. Twenty eigenvalues were extracted and compared
with the reference solution obtained for the entire 360° model with superelements.

Results for the frequency analysis and the static analysis appear below.
Frequency analysis for models with superelements

Frequencies corresponding to the 15 lowest eigenvalues have been extracted and are tabulated in Table
2.2.1-1 for each model. To study the effect of retaining dynamic modes during superelement
generation, the superelement models were run after extracting 0, 5, and 20 dynamic modes during
superelement generation.
While the Guyan reduction technique (m =0) yields frequencies that are reasonable compared to those
of the full model, the values obtained with 5 retained modes are much closer to full model predictions,
especially for the higher eigenvalues. Increasing the number of retained modes to 20 does not yield a
significant improvement in the results, consistent with the fact that in the Guyan reduction technique
the choice of retained degrees of freedom affects accuracy, while for the restrained mode addition
technique the modes corresponding to the lowest frequencies are by definition optimal.
When superelements are used in an eigenfrequency analysis, it is to be expected that the lowest
eigenfrequency in the superelement model is higher than the lowest eigenfrequency in the
corresponding model without superelements. This is indeed the case for the single-level superelement
analysis, but for the multi-level superelement analysis the lowest eigenfrequency is below the one for
the full model. This occurs because the stress and load stiffness for the lowest level superelement (the
blade) are generated with the root of the blade fixed, whereas in the full model the root of the blade
will move radially due to the deformation of the hub under the applied centrifugal load. Hence, the
superelement stiffness is somewhat inaccurate. Since the radial displacements at the blade root are
small compared to the overall dimensions of the model (of order 10 -3), the resulting error should be
small, as is observed from the results.
Table 2.2.1-2 shows what happens if the NLGEOM parameter is omitted during the preloading steps. It
is clear that the results are significantly different from the ones that take the effect of the preload on the
stiffness into account. Note that in this case the lowest eigenfrequency in the superelement models is
indeed above the lowest eigenfrequency in the model without superelements.
Static analysis for models with superelements

A static analysis of the fan is carried out about the preloaded base state by applying a pressure load of
105 Pa normal to the blades of the fan. The axial displacement of the outer edge of the fan blade due to
the pressure load is monitored at nodes along path AB, as shown in Figure 2.2.1-1. The results are
shown in Figure 2.2.1-4; there is good agreement between the solutions for the superelement models
and the full model.
While superelements can be generated from models that exhibit nonlinear response, it must be noted
2-863
that, once created, a superelement always exhibits linear response at the usage level. Hence, a
preloaded superelement will produce a response equivalent to that of the response to a linear
perturbation load on a preloaded full model. Consequently, the full model is analyzed by applying the
centrifugal preload in a general step and the pressure load in a linear perturbation step. Since an
analysis using superelements is equivalent to a perturbation step, the results obtained do not
incorporate the preload deformation. Thus, if the total displacement of the structure is desired, the
results of this perturbation step need to be added to the base state solution of the structure.
Static analysis and eigenvalue extraction for the cyclic symmetry

model
In the general static step that includes nonlinear geometry, the centrifugal load is applied to the datum
sector. Only symmetric loads can be applied in general static steps with the *CYCLIC SYMMETRY
MODEL option. The computed eigenvalues are identical with those obtained for the entire 360°
model, as shown in the Table 2.2.1-1. The additional information obtained during the eigenvalue
extraction is the cyclic symmetry mode number associated with each eigenvalue. In the case of 4
repetitive sectors, all the eigenvalues corresponding to cyclic symmetry mode 1 appear in pairs; the
eigenvalues corresponding to modes 0 and 2 are single. The lowest first two eigenvalues correspond to
cyclic symmetry mode 1, followed by the single eigenvalues corresponding to cyclic symmetry modes
2 and 0. For a comparison with the cyclic symmetry model option, the problem is also modeled with
*MPC type CYCLSYM (see fansuperelem_mpc.inp). To verify the use of superelements with the
cyclic symmetry model, it was determined that the results obtained with fansuperelem_cyclic.inp were
identical to the results obtained with fansuperelem_1level_freq.inp.
Input files
fan_cyclicsymmodel.inp
Cyclic symmetry model with static and eigenvalue extraction steps.
fansuperelem_1level_freq.inp
Single-level superelement usage analysis with a frequency extraction step.
fansuperelem_1level_static.inp
Single-level superelement usage analysis with a static step.
fansuperelem_multi_freq.inp
Multi-level superelement usage analysis with a frequency extraction step.
fansuperelem_multi_static.inp
Multi-level superelement usage analysis with a static step.
fansuperelem_freq.inp
Frequency extraction without superelements.
fansuperelem_static.inp
2-864
Static analysis without superelements.

fansuperelem_mpc.inp
Single-level usage analysis demonstrating the use of cyclic symmetry MPCs.
fansuperelem_gen1.inp
Superelement generation for a single blade used in the multi-level superelement generation file
fansuperelem_gen2.inp.
Multi-level superelement generation used in fansuperelem_multi_freq.inp and
fansuperelem_multi_static.inp.
Single-level superelement generation used infansuperelem_1level_freq.inp,
fansuperelem_1level_static.inp, and fansuperelem_mpc.inp.
fansuperelem_cyclic.inp
Single-level superelement with the cyclic symmetry model used in a frequency analysis.
Tables
Table 2.2.1-1 Comparison of natural frequencies for single-level and multi-level superelements with
the values for the model without superelements.
With substructuring: 1 With substructuring: 2
Eigenvalue Full
level levels
no. cycles/sec model
m=0 m=5 m=20 m=0 m=5 m=20
1 6.9464 6.7893 6.7882 6.9191 6.7665 6.7654 6.7881
2 6.9464 6.7893 6.7882 6.9191 6.7665 6.7654 6.7881
3 8.0024 7.7148 7.7139 8.0082 7.7228 7.7219 7.7139
4 8.2007 7.8817 7.8810 8.2079 7.8909 7.8903 7.8810
5 11.343 11.021 11.010 11.308 10.986 10.976 11.009
6 11.343 11.021 11.010 11.308 10.986 10.976 11.009
7 12.513 11.916 11.897 12.291 11.760 11.741 11.895
8 14.683 14.354 14.301 14.671 14.303 14.252 14.303
9 17.862 14.432 14.432 17.745 14.470 14.470 14.432
10 18.921 14.776 14.772 18.913 14.814 14.810 14.771
11 21.150 14.776 14.772 21.010 14.814 14.810 14.771
12 21.150 15.990 15.952 21.010 16.001 15.963 15.948
13 28.449 17.773 17.696 28.417 17.652 17.575 17.696
14 28.986 19.029 19.012 29.001 19.030 19.013 19.001
15 28.986 21.234 21.077 29.001 21.082 20.928 21.075
Table 2.2.1-2 Comparison of natural frequencies for single-level and two-level superelements with
2-865
the full model values without the use of the NLGEOM parameter.
With
Eigenvalue Full
substructuring
no. cycles/sec model
1 level 2 levels
1 4.4811 4.4811 4.4809
2 4.4811 4.4811 4.4809
3 4.5489 4.5489 4.5487
4 4.8916 4.8916 4.8914
5 9.5519 9.5519 9.5423
6 9.5519 9.5519 9.5423
7 9.7893 9.7894 9.7758
8 12.611 12.611 12.570
9 14.006 14.006 14.003
10 14.332 14.332 14.325
11 14.332 14.332 14.325
12 15.475 15.475 15.455
13 16.962 16.963 16.897
14 18.244 18.245 18.220
15 19.040 19.041 18.933
Figures
Figure 2.2.1-1 Mesh used for the complete fan model.
Figure 2.2.1-2 Superelements generated.
2-866
Figure 2.2.1-3 Datum sector for cyclic symmetry model.
2-867
Figure 2.2.1-4 Displacements due to pressure loading along path AB.
2-868
Sample listings
2-869
Listing 2.2.1-1
*HEADING
Superelement analysis of a fan
Frequency analysis : One level superelement.
Usage level: FAN=HUB+4xBLADES
**
Z100: a superelement for a single blade;
Z200: a superelement which contains the 1/4 hub
and a superelement blade (Z100);
Z300: a superelement which contains the 1/4 hub
and a full blade.
Requires superelement generation file
**
*RESTART,WRITE,FRE=1
**
*NODE
1, 3.50000, 3.00000, 4.00000
3, 3.46638, 3.18024, 4.00000
7, 3.18024, 3.46638, 4.00000
9, 3.00000, 3.50000, 4.00000
**
31, 6.00000, 3.00000, 4.00000
33, 5.79830, 4.08144, 4.00000
37, 4.08144, 5.79830, 4.00000
39, 3.00000, 6.00000, 4.00000
**
91, 6.00000, 3.00000, 0.00000
93, 5.79830, 4.08144, 0.00000
97, 4.08144, 5.79830, 0.00000
99, 3.00000, 6.00000, 0.00000
**
100, 7.89643, 8.00249, 2.59810
9998, 3.00000, 3.00000, 4.00000
9999, 3.00000, 3.00000, 0.00000
*NGEN,LINE=C,NSET=NHUB
1,3,1,9998
3,7,1,9998
7,9,1,9998
*NGEN,LINE=C,NSET=RINGF
31,33,1,9998
33,37,1,9998
2-870
37,39,1,9998
*NGEN,LINE=C,NSET=RINGB
91,93,1,9999
93,97,1,9999
97,99,1,9999
*NFILL,NSET=HUB
NHUB,RINGF,3,10
RINGF,RINGB,6,10
**
*NSET,NSET=PART1,GENERATE
1,91,10
9,99,10
2,8,1
32,38,1
100,100,1
**
*NCOPY,CHANGE NUMBER=100,OLD SET=PART1,SHIFT
0.,0.,0.
3.,3.,-1.,3.,3.,1.,90.
0.,0.,0.
3.,3.,-1.,3.,3.,1.,180.
0.,0.,0.
3.,3.,-1.,3.,3.,1.,270.
**
*NSET,NSET=PART1X,GENERATE
1,91,10
*NSET,NSET=PART1Y,GENERATE
9,99,10
109,199,10
101,191,10
201,291,10
209,299,10
309,399,10
301,391,10
*MPC
2-871
TIE,PART4X,PART1X
TIE,PART2Y,PART1Y
TIE,PART2X,PART3X
TIE,PART4Y,PART3Y
**
*NSET,NSET=NODEHUB,GENERATE
1, 9, 1
102, 108, 1
201, 209, 1
302, 308, 1
**
*ELEMENT,TYPE=Z300,ELSET=P1,FILE=FAN
901,1,2,3,4,5,6,7,8,9,11,19,21,29,31,32,
33,34,35,36,37,38,39,41,49,51,59,61,69,71,79,81,
89,91,99,100
*ELCOPY,ELEMENT SHIFT=1,OLD SET=P1,
SHIFT NODES=100,NEW SET=P2
**
*SUPER PROPERTY,ELSET=P1
0.,0.,0.
0.,0.,0.
3.,3.,0.,3.,3.,3.,90.
0.,0.,0.
3.,3.,0.,3.,3.,3.,180.
0.,0.,0.
3.,3.,0.,3.,3.,3.,270.
**
*STEP
Step 1: Eigenfrequency extraction
*FREQUENCY
20,
*BOUNDARY
NODEHUB,ENCASTRE
**
*EL PRINT,FREQ=0
*EL FILE,FREQ=0
2-872
*NODE PRINT,FREQ=0
*MODAL FILE,FREQ=99
**
*END STEP
2.2.2 Linear analysis of the Indian Point reactor feedwater line

This example concerns the linear analysis of an actual pipeline from a nuclear reactor and is intended
to illustrate some of the issues that must be addressed in performing seismic piping analysis. The
pipeline is the Indian Point Boiler Feedwater Pipe fitted with modern supports, as shown in Figure
2.2.2-1. This pipeline was tested experimentally in EPRI's full-scale testing program. The model
corresponds to Configuration 1 of the line in Phase III of the testing program. The experimental results
are documented in EPRI Report NP-3108 Volume 1 (1983).
We first verify that the geometric/kinematic model is adequate to simulate the dynamic response
accurately. For this purpose we compare predictions of the natural frequencies of the system using a
coarse model and a finer model, as well as two substructure models created from the coarse mesh.
These analyses are intended to verify that the models used in subsequent runs provide accurate
predictions of the lower frequencies of the pipeline. We then perform linear dynamic response analysis
in the time domain for one of the "snap-back" loadings applied in the physical test ( EPRI NP-3108,
1983) and compare the results with the experimental measurements. The linear dynamic response
analysis results are also compared with the results of direct integration analysis (integration of all
variables in the entire model, as would be performed for a generally nonlinear problem). This is done
primarily for cross-verification of the two analysis procedures. These snap-back response analyses
correspond to a load of 31136 N (7000 lb) applied at node 25 in the z-direction, with the pipe filled
with water. This load case is referred to as test S138R1SZ in EPRI NP-3108.
We also compute the pipeline's response in the frequency domain to steady excitation at node 27 in the
z-direction. Experimental data are also available for comparison with these results.
Geometry and model

Geometrical and material properties are taken from EPRI NP-3108 (1983). The supports are assumed
to be linear springs for the purpose of these linear analyses, although their actual response is probably
nonlinear. The spring stiffness values are those recommended by Tang et al. (1985). The pipe is
assumed to be completely restrained in the vertical direction at the wall penetration.
In the experimental snap-back test used for the comparison (test S138R1SZ), the pipe is full of water.
The DENSITY parameter on the *BEAM GENERAL SECTION option is, therefore, adjusted to
account for the additional mass of the water by computing a composite (steel plus water) mass per unit
length of pipe.
The pipeline is modeled with element type B31. This is a shear flexible beam element that uses linear
interpolation of displacement and rotation between two nodes, with transverse shear behavior modeled
according to Timoshenko beam theory. The element uses a lumped mass matrix because this provides
2-873
more accurate results in test cases.

The coarse finite element model uses at least two beam elements along each straight run, with a finer
division around the curved segments of the pipe to describe the curvature of the pipe with reasonable
accuracy. Separate nodes are assigned for all spring supports, external loading locations, and all the
points where experimental data have been recorded. The model is shown in Figure 2.2.2-2. This mesh
has 74 beam elements.
In typical piping systems the elbows play a dominant role in the response because of their flexibility.
This could be incorporated in the model by using the ELBOW elements. However, ELBOW elements
are intended for applications that involve nonlinear response within the elbows themselves and are an
expensive option for linear response of the elbows, which is the case for this study. Therefore, instead
of using elbow elements, we modify the geometrical properties of beam elements to model the elbows
with correct flexibility. This is done by calculating the flexibility factor, k, for each elbow and
modifying the moments of inertia of the beam cross-sections in these regions. The flexibility factor for
an elbow is a function of two parameters. One is a geometric parameter, ¸, defined as
tR
¸= p ;
r2 1 ¡ º 2
where t is the wall thickness of the curved pipe, R is the bend radius of the centerline of the curved
pipe, r is the mean cross-sectional radius of the curved pipe, and º is Poisson's ratio. The other
parameter is an internal pressure loading parameter, Ã. For thick sections (like the ones used in this
pipe), Ã has negligible effect unless the pressures are very high and the water in this case is not
pressurized. Consequently, the flexibility factor is a function of ¸ only.
For the elbows in this pipeline ¸ = 0.786 for the 203 mm (8 in) section and ¸ = 0.912 for the 152 mm
(6 in) section. The corresponding flexibility factors obtained from Dodge and Moore (1972) are 2.09
and 1.85. These are implemented in the model by modifying the moments of inertia of the beam
cross-sections in the curved regions of the pipeline.
ABAQUS provides two different options for introducing geometrical properties of a beam
cross-section. One is the *BEAM GENERAL SECTION option, in which all geometric properties
(area, moments of inertia) can be given without specifying the shape of the cross-section. The material
data, including the density, are given on the same option. Alternatively, the geometrical properties of
the cross-section can be given by using the *BEAM SECTION option. With this option the
cross-section dimensions are given, and ABAQUS calculates the corresponding cross-sectional
behavior by numerical integration, thus allowing for nonlinear material response in the section. When
this option is used, the material properties--including density and damping coefficients--are introduced
in the *MATERIAL option associated with the section. This approach is more expensive for systems
in which the cross-sectional behavior is linear, since numerical integration over the section is required
each time the stress must be computed. Thus, in this case we use the *BEAM GENERAL SECTION
option.
To verify that the mesh will provide results of adequate accuracy, the natural frequencies predicted
with this model are compared with those obtained with another mesh that has twice as many elements
2-874
in each pipe segment. Table 2.2.2-1 shows that these two meshes provide results within 2% for the first
six modes and generally quite similar frequencies up to about 30 Hz. Based on this comparison the
smaller model, with 74 beam elements, is used for the remaining studies (although the larger model
would add little to the cost of the linear analyses, which for either case would be based on the same
number of eigenmodes: only in direct integration would the cost increase proportionally with the
model size).
Superelement models
In ABAQUS the dynamic response of a superelement is defined by a combination of Guyan reduction
and the inclusion of some natural modes of the fully restrained superelement. Guyan reduction consists
of choosing additional physical degrees of freedom to retain in the dynamic model that are not needed
to connect the superelement to the rest of the mesh. In this example we use only Guyan reduction since
the model is small and it is easy to identify suitable degrees of freedom to retain. A critical modeling
issue with this method is the choice of retained degrees of freedom: enough degrees of freedom must
be retained so that the dynamic response of the substructure is modeled with sufficient accuracy. The
retained degrees of freedom should be such as to distribute the mass evenly in each substructure so that
the lower frequency response of each substructure is modeled accurately. Only frequencies up to 33 Hz
are generally considered important in the seismic response of piping systems such as the one studied in
this example, so the retained degrees of freedom must be chosen to provide accurate modeling of the
response up to that frequency.
In this case the pipeline naturally divides into three segments in terms of which kinematic directions
participate in the dynamic response, because the response of a pipeline is generally dominated by
transverse displacement. The lower part of the pipeline, between nodes 1 and 23, is, therefore, likely to
respond predominantly in degrees of freedom 1 and 2; the middle part, between nodes 23 and 49,
should respond in degrees of freedom 2 and 3; and the top part, above node 49, should respond in
degrees of freedom 1 and 3. Comparative tests (not documented) have been run to verify these
conjectures, and two superelement models have been retained for further analysis: one in which the
entire pipeline is treated as a single superelement, and one in which it is split into three superelements.
In the latter case all degrees of freedom must be retained at the interface nodes to join the
superelements correctly. At other nodes only some translational degrees of freedom are retained, based
on the arguments presented above.
The choice of which degrees of freedom to retain can be investigated inexpensively in a case such as
this by numerical experiments--extracting the modes of the reduced system for the particular set of
retained degrees of freedom and comparing these modes with those of the complete model. The
choices made in the superelement models used here are based somewhat on such tests, although
insufficient tests have been run to ensure that they are close to the optimal choice for accuracy with a
given number of retained variables. For linear analysis of a model as small as this one, achieving an
optimal selection of retained degrees of freedom is not critical because computer run times are short: it
becomes more critical when the reduced model is used in a nonlinear analysis or where the underlying
model is so large that comparative eigenvalue tests cannot be performed easily. In such cases the
inclusion of natural modes of the superelement is desirable. The superelement models are shown in
Figure 2.2.2-2.
2-875
Damping
"Damping" plays an essential role in any practical dynamic analysis. In nonlinear analysis the
"damping" is often modeled by introducing dissipation directly into the constitutive definition as
viscosity or plasticity. In linear analysis equivalent linear damping is used to approximate dissipation
mechanisms that are not modeled explicitly.
Experimental estimates of equivalent linear damping, based on three different methods, are found in
EPRI NP-3108 (see Table 7-6, Table 7-7, and Figure 7-15 of that report). For the load case and pipe
configuration analyzed here, those results suggest that linear damping corresponding to 2.8% of critical
damping in the lowest mode of the system matches the measured behavior of the structure, with the
experimental results also showing that the percentage of critical damping changes from mode to mode.
In spite of this all the numerical analyses reported here assume the same damping ratio for all modes
included in the model, this choice being made for simplicity only.
For linear dynamic analysis based on the eigenmodes, ABAQUS allows damping to be defined as a
percentage of critical damping in each mode, as structural damping (proportional to nodal forces), or as
Rayleigh damping (proportional to the mass and stiffness of the structure). Only the last option is
possible when using direct integration, although other forms of damping can be added as discrete
dashpots or in the constitutive models. In this case, results are obtained for linear dynamic analysis
with modal and Rayleigh damping and for direct integration with Rayleigh damping.

Results are shown for four geometric models: the "coarse" (74 element) model, which has a total of
435 degrees of freedom; a finer (148 element) model, which has a total of 870 degrees of freedom; a
model in which the pipeline is modeled as a single superelement (made from the coarse model), with
59 retained degrees of freedom; and a model in which the pipeline is modeled with three
superelements (made from the coarse model), with 65 retained degrees of freedom.
The first comparison of results is the natural frequencies of the system, as they are measured and as
they are predicted by the various models. The first 24 modes are shown in Table 2.2.2-1. These modes
span the frequency range from the lowest frequency (about 4.3 Hz) to about 43 Hz. In typical seismic
analysis of systems such as this, the frequency range of practical importance is up to 33 Hz; on this
basis these modes are more than sufficient.
Only the first six modes of the actual system have been measured, so any comparison at higher
frequencies is between the numerical calculations reported here and other similar computations. The
results obtained with the four models correlate quite well between themselves, suggesting that the
mesh and the choices of retained degrees of freedom in the superelement models are reasonable. It is
particularly noteworthy that the results for the superelement models correspond extremely well with
those provided by the original model, considering the large reduction in the number of degrees of
freedom for the substructures. The results also correlate roughly with the analysis results obtained by
EDS and reported in EPRI NP-3108: except for modes 3 and 4 the frequencies are within 10% of the
EDS numbers. For the first three modes the ABAQUS results are lower than those reported by EDS.
This suggests the possibility that the ABAQUS model may be too flexible. The SUPERPIPE values are
2-876
significantly higher than any of the other data for most modes, and the ABAQUS and the EDS results
diverge from the test results after the first four modes.
The results of the time history analyses are summarized in Table 2.2.2-2to Table 2.2.2-5. These
analyses are based on using all 24 modes of the coarse model. Typical predicted response plots are
shown in Figure 2.2.2-3 to Figure 2.2.2-7. In many cases of regular, beam-type, one-dimensional
structures, the first few modes will generally establish the dynamic behavior. Although the pipeline has
an irregular shape, it is worth checking how much the higher modes influence the results. This is done
in this case by comparing the results using the first six modes only with the results obtained with 24
modes. The highest discrepancy (20%) is found in the predicted accelerations at certain degrees of
freedom. All other results show at most 5-10% differences (see Figure 2.2.2-3and Figure 2.2.2-4). This
conclusion is also supported by the steady-state results.
All the ABAQUS results are reasonably self-consistent, in the sense that Rayleigh and modal damping
and modal dynamics and direct integration all predict essentially the same values. The choice of 2.8%
damping seems reasonable, in that oscillations caused by the snap-back are damped out almost
completely in 10 seconds, which corresponds to the measurements.
Unfortunately there is poor correlation between predicted and measured support reactions and
maximum recorded displacements. The test results and the corresponding computations are shown in
Table 2.2.2-2and Table 2.2.2-3. All the models give essentially the same values. The initial reactions
and displacements are computed for a snap-back load of 31136 N (7000 lb) applied at node 25 (node
417 in EPRI report NP-3108) in the z-direction. The maximum recorded displacements occur at node
27 (node 419 in EPRI report NP-3108) in the y- and z-directions. It is assumed that the supports are in
the positions relative to the pipe exactly as shown in Figure 2.2.2-1. The scatter in the experimental
measurements makes it difficult to assess the validity of the stiffness chosen for the spring supports.
The maximum displacement predicted at node 27 in the z-direction is almost twice that measured. This
again implies the possibility that, at least in the area near this node, the model is too flexible.
The generally satisfactory agreement between the natural frequency predictions and poor agreement
between the maximum displacements and reactions suggests that improved modeling of the supports
may be necessary. In this context it is worthwhile noting that the experimental program recorded
significantly different support parameters in different tests on the pipeline system.
Table 2.2.2-4 shows the results for displacement and acceleration for node 27 (which has the largest
displacement). All the computed results are higher than the experimental values. The largest
discrepancies between the measurements and the analysis results are in the predictions of peak forces
in the springs, summarized in Table 2.2.2-5. Results obtained with the various models differ by less
than 10%: these differences are caused by the differences in the models, different types of damping,
and--for the direct integration results--errors in the time integration (for the modal dynamic procedure
the time integration is exact). The principal cause of the discrepancies between the measurements and
the computed values is believed to be the assumption of linear response in the springs in the numerical
models. In reality the spring supports are either rigid struts or mechanical snubbers (Configuration 2).
Especially when snubbers are used, the supports perform as nonlinear elements and must be modeled
as such to reflect the support behavior accurately. Interestingly, even with the assumption of linear
support behavior, the character of the oscillation is well-predicted for many variables.
2-877
The last group of numerical results are frequency domain calculations obtained using the *STEADY
STATE DYNAMICS linear dynamic response option. The response corresponds to steady harmonic
excitation at node 27 in the z-direction by a force with a peak amplitude of 31136 N (7000 lb). Such
frequency domain results play a valuable role in earthquake analysis because they define the frequency
ranges in which the structure's response is most amplified by the excitation. Although it is expected
that the first few natural frequencies will be where the most amplification occurs, the results show
clearly that some variables are strongly amplified by the fifth and sixth modes. This is observed both in
the simulations and in the experimental measurements. Measured experimental results are available for
the acceleration of node 33 (node 419 in EPRI NO-3108) in the z-direction and for the force in spring
FW-R-21. The character of curves obtained with ABAQUS agrees well with the experimental results
(see Figure 2.2.2-8and Figure 2.2.2-9), but the values differ significantly, as in the time domain results.
The peak acceleration recorded is 2.0 m/s 2 (78.47 in/s 2), at the first natural frequency, while the
analysis predicts 4.0 m/s 2 (157.5 in/s 2). Likewise, the peak force value recorded is 2.0 kN (450 lb),
compared to 5.9 kN (1326 lb) predicted. The discrepancies are again attributed to incorrect estimates
of the support stiffness or to nonlinearities in the supports.
Input files
indianpoint_modaldyn_coarse.inp
*MODAL DYNAMIC analysis with modal damping using the coarse model.
indianpoint_modaldyn_3sub.inp
*MODAL DYNAMIC analysis using the three substructure model.
indianpoint_3sub_gen1.inp
First superelement generation referenced by the analysis indianpoint_modaldyn_3sub.inp.
Second superelement generation referenced by the analysis indianpoint_modaldyn_3sub.inp.
Third superelement generation referenced by the analysis indianpoint_modaldyn_3sub.inp.
indianpoint_sstate_sinedwell.inp
*STEADY STATE DYNAMICS analysis corresponding to the sine dwell test performed
experimentally using the coarse model.
indianpoint_direct_beam_coarse.inp
Direct integration analysis using the coarse model with the *BEAM SECTION option.
indianpoint_sstate_modaldamp.inp
*STEADY STATE DYNAMICS analysis with modal damping, covering a range of frequencies
using the coarse model.
indianpoint_modaldyn_1sub.inp
2-878
*MODAL DYNAMIC analysis with one substructure.

Superelement generation referenced by the analysis indianpoint_modaldyn_1sub.inp.
indianpoint_direct_beamgensect.inp
Direct integration using the coarse model with *BEAM GENERAL SECTION instead of *BEAM
SECTION, which, thus, runs faster on the computer since numerical integration of the
cross-section is avoided.
indianpoint_modaldyn_elmatrix1.inp
*MODAL DYNAMIC analysis that reads and uses the superelement matrix written to the results
file in indianpoint_3sub_gen1.inp, indianpoint_3sub_gen2.inp, and indianpoint_3sub_gen3.inp.
Reads and uses the element matrix written to the results file in indianpoint_modaldyn_3sub.inp.
Reads and uses the superelement matrix written to the results file in indianpoint_1sub_gen1.inp.
Reads and uses the element matrix written to the results file in indianpoint_modaldyn_1sub.inp.
indianpoint_modaldamp_rayleigh.inp
*MODAL DAMPING analysis with modal Rayleigh damping using the coarse mesh with the
*BEAM SECTION option.
indianpoint_dyn_rayleigh_3sub.inp
*DYNAMIC analysis with Rayleigh damping using the three substructure model.
indianpoint_rayleigh_3sub_gen1.inp
First superelement generation referenced by the analysis indianpoint_dyn_rayleigh_3sub.inp.
Second superelement generation referenced by the analysis indianpoint_dyn_rayleigh_3sub.inp.
Third superelement generation referenced by the analysis indianpoint_dyn_rayleigh_3sub.inp.
indianpoint_modaldyn_unsorted.inp
One substructure *MODAL DYNAMIC analysis with unsorted node sets and unsorted retained
degrees of freedom.
indianpoint_unsorted_gen1.inp
Superelement generation with unsorted node sets and unsorted retained degrees of freedom
referenced by the analysis indianpoint_modaldyn_unsorted.inp.
2-879
indianpoint_lanczos.inp
Same as indianpoint_modaldyn_coarse.inp, except that it uses the Lanczos solver and the
eigenvectors are normalized with respect to the generalized mass.
indianpoint_restart_normdisp.inp
Restarts from indianpoint_lanczos.inp and continues the eigenvalue extraction with the
eigenvectors normalized with respect to the maximum displacement.
indianpoint_restart_bc.inp
Restarts from indianpoint_lanczos.inp and continues the eigenvalue extraction with modified
boundary conditions.
indianpoint_overlapfreq.inp
Contains two steps, which extract eigenvalues with overlapping frequency ranges.
References
· Consolidated Edison Company of New York, Inc., EDS Nuclear, Inc., and Anco Engineers, Inc.,
Testing and Analysis of Feedwater Piping at Indian Point Unit 1, Volume 1: Damping and
Frequency, EPRI NP-3108, vol. 1, July 1983.
· Dodge, W. G., and S. E. Moore, "Stress Indices and Flexibility Factors for Moment Loadings in
Elbows and Curved Pipe," WRC Bulletin, no. 179, December 1972.
· Tang, Y. K., M. Gonin, and H. T. Tang, "Correlation Analysis of In-situ Piping Support
Reactions," EPRI correspondence with HKS, May 1985.
Tables
Table 2.2.2-1 Comparison of natural frequencies (Hz).

Anco EDS SUPE ABAQUS
Mod R
(experiment PIPE coarse finer single three
e mesh mes super super
)
h s
1 4.20 4.30 5.30 4.25 4.26 4.25 4.25
2 6.80 6.80 8.10 6.27 6.25 6.27 6.27
3 8.30 8.80 12.00 7.29 7.29 7.30 7.30
4 12.60 10.6 13.30 12.80 12.6 12.87 12.86
0 6
5 15.40 13.0 14.40 13.18 13.1 13.19 13.20
0 4
6 16.70 14.5 15.90 13.90 13.7 13.91 13.92
0 5
7 16.2 18.30 15.11 15.9 14.34 14.39
2-880
0 8
8 19.40 16.30 16.0 16.24 16.31
7
9 20.20 16.89 16.8 16.43 16.43
1
10 22.20 17.43 17.8 17.17 17.20
2
11 18.02 19.0 18.10 18.10
7
12 19.58 20.1 20.05 20.01
0
13 23.43 21.4 23.98 24.00
5
14 23.99 22.1 24.47 24.47
3
15 24.27 23.5 24.97 24.96
8
16 24.80 24.1 25.34 25.28
5
17 26.82 26.8 27.63 27.56
4
18 29.53 30.1 30.31 30.55
8
19 30.61 30.6 31.08 31.06
0
20 30.95 32.5 31.43 31.43
8
21 31.52 33.1 32.00 31.98
1
22 33.50 35.0 33.76 33.77
8
23 39.09 39.6 39.75 39.97
5
24 39.86 43.2 42.98 42.97
5
Table 2.2.2-2 Comparison of initial support reactions. Snap-back Test No. S138R1SZ; 31136 N
(7000 lb) at node 25, z-direction.
NOD SUPPOR Anco TEST ABAQUS
E T N (lb) N (lb)
15 FW-R-11 -8000 (-1798.6 -11712 (-2633)
)
22 FW-R-13 30000 (6744.6) 29352 (6599)
23 FW-R-14 -252 (-56.7) -3754 (-844)
35 FW-R-17 23625 (5311.4) 102 (-22.8)
2-881
35 FW-R-18 10025 (2553.8) -18468 (-4152)

39 FW-R-20 24000 (5395.7) 4212 ( 947)
39 FW-R-21 -2450 (-5508.1 25016 (5624)
0 )
49 FW-R-23 8000 (1798.6) 24348 (5474)
53 FW-R-24 -4324 (-972.1) -4057 (-912)
53 FW-R-25 2000 (449.6) -816 (-183)
56 FW-R-27 432 (97.1) -1801 (-405)
56 FW-R-28 156 (35.1) -799 (-180)
Table 2.2.2-3 Comparison of maximum displacements.

ABAQUS Anco Measured ABAQUS
NODE NODE
No. No. mm (in) mm (in)
27 419-Y -16.0 (-.630) -26.85 (-1.057)
27 419-Z 37.81 (1.49) 65.72 (2.587)
Table 2.2.2-4 Peak displacement and acceleration values at node 27.

Variable Measured ABAQUS
(Anco) Modal, Modal, Direct
2.8% modal Rayleigh integration
damping damping
uy (mm) -0.024/0.024 -0.029/0.029 -0.031/0.03 -0.031/0.031
1
uz (mm) -0.038/0.038 -0.058/0.066 -0.062/0.05 -0.063/0.068
9
Äz (m/s2)
u -47.6/40.9 -42.1/50.8 -49.6/49.9 -83.8/91.0
The high acceleration amplitude reported for the ABAQUS direct integration
analysis
occurs only during the first few increments, after which it reduces to
-31.6/48.6 m/s 2.
Table 2.2.2-5 Peak reaction forces at supports (in kN).

Support Measured ABAQUS
number (Anco) Modal, Modal, Direct
2.8% modal Rayleigh integration
damping damping
FW-R-11 -16.44/19.22 -19.80/13.42 -19.90/14.26 -21.82/15.76
FW-R-13 -15.10/29.91 -18.94/24.45 -19.46/23.61 -28.50/21.98
FW-R-14 -7.22/12.00 -9.34/7.35 -10.23/10.00 -12.54/9.03
FW-R-17 34.40/26.20 -7.50/10.59 -.17/9.25 -7.91/10.97
FW-R-18 -14.30/14.40 -33.26/32.06 -33.58/31.61 -33.46/32.63
FW-R-20 -25.60/26.90 -7.54/8.79 -7.98/8.50 -8.07/10.60
FW-R-21 -24.50/23.80 -25.55/24.47 -26.38/25.30 -27.78/25.26
FW-R-23 -15.30/16.00 -25.39/24.63 -26.06/25.36 -25.40/24.35
2-882
FW-R-24 -9.61/7.30 -7.17/6.87 -7.69/7.20 -8.23/8.64

FW-R-25 -6.77/6.21 -3.48/4.36 -3.34/4.55 -7.13/4.71
FW-R-27 -3.76/3.04 -4.12/4.00 -3.78/3.80 -4.29/4.43
FW-R-28 -1.10/1.82 -1.53/1.08 -1.62/1.15 -1.79/1.44
Figures
Figure 2.2.2-1 Indian Point boiler feedwater line: modern supports, Configuration 1.
Figure 2.2.2-2 Basic mesh and superelement models.
2-883
Figure 2.2.2-3 z-displacement at node 27, modal analysis with 24 modes.
Figure 2.2.2-4 z-displacement at node 27, modal analysis with 6 modes.
2-884
Figure 2.2.2-5 z-displacement at node 27, direct integration analysis.
Figure 2.2.2-6 z-direction acceleration at node 27, modal analysis with 24 modes.
2-885
Figure 2.2.2-7 Force in spring support FW-R-11, modal analysis with 24 modes.
Figure 2.2.2-8 Comparison of z-direction acceleration at node 33 between experimental steady-state

results (solid line) and ABAQUS (dashed line).
2-886
Figure 2.2.2-9 Comparison of force in spring support FW-R-21, between experimental steady-state
results (solid line) and ABAQUS (dashed line).
2-887
Sample listings
2-888
Listing 2.2.2-1
*HEADING
INDIAN POINT FEEDWATER LINE WITH SPRING SUPPORTS
** BEAM ELEMENTS WITH MODAL DYNAMICS,
** MODAL DAMPING
*NODE
1, 0., 423., -234.96
3, 0., 423., -150.96
5, 0., 435., -138.96
6, 0., 474., -138.96
8, 0., 486., -126.96
10, 0., 486., -75.96
11, 0., 486., -51.96
12, 0., 486., -18.00
13, 0., 486., 9.00
15, 0., 486., 144.5
16, 0., 486., 159.
18, 8.484, 494.484, 171.
19, 8.484,494.484,171.
21, 16.93, 497.96, 171.
22, 19.8125, 497.96 , 171.
23, 29.125 ,497.96, 171.
25, 200.72, 497.96, 171.00
27, 260.72, 497.96, 171.
29, 272.72, 509.96, 171.00
31, 272.72, 569.964, 171.00
33, 280.44, 581.96, 180.19
35, 330.1 , 581.96, 239.3
36, 335.21, 581.96, 245.46
38, 342.91, 593.96, 254.65
39, 342.91 , 628. , 254.65
40, 342.91, 660., 254.65
42, 342.91, 706., 254.65
44, 340.22, 714.48, 256.91
46, 296.57, 771.47, 293.54
48, 282.36, 779.95, 289.80
49, 278.50, 779.95, 285.20
50,274.644, 779.95, 280.61
52, 266.93, 791.95, 271.42
53, 266.93, 801., 271.42
54, 266.93, 876.00, 271.42
56, 266.93, 990.96, 271.42
2-889
57, 266.93, 1000.27, 271.42

59, 278.88, 1012.27, 272.46
61, 335.26, 1012.27, 277.39
63, 343.40, 1012.27, 281.64
64, 366.97, 1012.27, 309.73
65, 369.52, 1012.27, 312.76
66, 379.16, 1012.27, 324.25
67, 388.8, 1012.27, 335.74
68, 389.11, 1012.27, 336.11
70, 396.83, 1024.27, 345.3
71, 396.83, 1027.27, 345.3
72, 396.83, 1033.27, 345.3
73, 396.83, 1040.95, 345.3
75, 389.93, 1049.95, 351.08
76, 380.74, 1049.95, 358.8
*NGEN
1,3
8,10
13,15
23,25
25,27
29,31
33,35
40,42
44,46
54,57
59,61
*NGEN,LINE=C
3,5,1,, 0., 435., -150.96
6,8,1,, 0., 474., -126.96
16,18,1,, 8.484, 494.48, 159.00
19, 21,1,, 16.932, 485.96, 171.00
27,29,1,, 260.724, 509.96, 171.00
31,33,1,, 280.44, 569.96, 180.19
36,38,1,, 335.21, 593.96, 245.46
42,44,1,, 333.71, 706.00, 262.37
46,48,1,, 288.85, 771.47, 284.35
50,52,1,, 274.64, 791.95, 280.61
57,59,1,, 278.88, 1000.27, 272.46
61,63,1,, 334.21, 1012.27, 289.34
68,70,1,, 389.11, 1024.27, 336.11
73,75,1,, 389.93, 1040.95, 351.08
**
2-890
** SPRING DEFINITIONS
**
*NODE,NSET=SPRS
115, 24.91, 475.93, 144.5
122, 19.81, 497.96, 219.5
123, 29.13, 597.41, 160.55
135, 299.94, 555., 239.30
235, 330.10, 599.46, 239.30
139, 364.74, 628.00, 220.25
239, 359.58, 628.00, 291.83
149,278.50, 792.45, 285.20
153, 321.48, 801.00, 318.12
253, 314.43, 801.00, 212.09
156, 311.81, 990.96, 215.36
256,290.46, 1008.88, 299.46
*ELEMENT,TYPE=SPRINGA,ELSET=FWR11
1001,15,115
1002,22,122
1003,23,123
1004,35,135
1005,35,235
1006,39,139
1007,39,239
1008,49,149
1009,53,153
1010,53,253
1011,56,156
1012,56,256
*ELSET,ELSET=SPRINGS
FWR11, FWR13, FWR14, FWR17, FWR18, FWR20,
FWR21, FWR23, FWR24, FWR25, FWR27, FWR28
*SPRING ,ELSET=FWR11
2-891
17700. ,
*SPRING,ELSET=FWR13
119600.,
403000.,
97900.,
*SPRING,ELSET=FWR18
228000.,
86300.,
*SPRING,ELSET=FWR21
86300.,
*SPRING,ELSET=FWR23
319000.,
*SPRING,ELSET=FWR24
56800.,
*SPRING,ELSET=FWR25
39100.,
*SPRING,ELSET=FWR27
55500.,
*SPRING,ELSET=FWR28
68000.,
**
** PIPE DEFINITIONS
**
*ELEMENT,TYPE=B31
1,1,2
14,13,14
20,19,20
28,25,26
2-892
53,49,50
*ELEMENT,TYPE=B31,ELSET=ONE
24,22,23
*ELGEN
1,12
14,5
20,3
24,3
28,24
53,27
*MPC
BEAM,18,19
*ELSET,ELSET=D8
1,2,5,8,9,10,11,12,14,16,22 ,15
24,25,26,28,29,32,33,36,37,38,41,42,43,44,47,48
51,53,56,57,58,59,60,63,64,67,68,71
74,
*ELSET,ELSET=D8E
3, 4, 6, 7,17,18,20,21,30,31,34,35,39,40
45,46,49,50,54,55,61,62,65,66,72,73
*ELSET,ELSET=BF57
69,70
*ELSET,ELSET=BWR
75,
*ELSET,ELSET=D6
76,79
*ELSET,ELSET=D6E
77,78
*BEAM GENERAL SECTION,ELSET=D8 ,SECTION=GENERAL,
DENSITY= .0010691
12.763 , 105.317 , , 105.317 , 210.635
1. , , -1.
27.9E6 , 10.73E6
*BEAM GENERAL SECTION,ELSET=D8E,SECTION=GENERAL,
DENSITY=.0010691
12.763 , 50.439 , , 50.439 , 210.635
1. , , -1.
27.9E6 , 10.73E6
DENSITY=0.00102423
8.405 , 40.295 , , 40.295 , 80.589
1.,
27.9E6 , 10.73E6
2-893

DENSITY=0.00102423
8.405 , 21.828, ,21.828, 80.589
1.,
27.9E6 , 10.73E6
*BEAM GENERAL SECTION,ELSET=BWR,SECTION=GENERAL,
DENSITY=0.0013266
10.4806 , 67.143 , , 67.143 , 134.29
1. , , -1.
27.9E6 , 10.73E6
*BEAM GENERAL SECTION,ELSET=BF57,SECTION=GENERAL,
DENSITY=0.002591
30.631 ,161.77 , , 161.77 , 323.53
1. , , -1.
27.9E6 , 10.73E6
*NSET,NSET=SMALL
1,25,27,45,55
*BOUNDARY
1,1,6
76,1,6
70,1
70,2
10,2
SPRS,1,3
*STEP
*FREQUENCY
24,
*EL PRINT,ELSET=SPRINGS,FREQUENCY=1
S11,E11
*EL FILE,ELSET=SPRINGS
ELEN,
*EL PRINT,ELSET=ONE,FREQUENCY=1
SENER,
ELSE,
*EL FILE,ELSET=ONE
ENER,
ELEN,
*OUTPUT,FIELD
*ELEMENT OUTPUT,ELSET=SPRINGS
ELEN,
*ELEMENT OUTPUT,ELSET=ONE
2-894
ENER,
ELEN,
*NODE OUTPUT
*END STEP
*STEP,PERTURBATION
SNAP LOAD---APPLIED STATICALLY
*STATIC
S,E
*CLOAD
25,3,7000.
*NODE PRINT
U,
RF,
*NODE FILE,NSET=SMALL,FREQUENCY=1
U,RF
*OUTPUT,FIELD, FREQUENCY=1
*ELEMENT OUTPUT, ELSET=SPRINGS
ELEN,
ENER,
S,E
*ELEMENT OUTPUT,ELSET=ONE
ELEN,
ENER,
S,E
*NODE OUTPUT,NSET=SMALL
U,
RF,
U,RF
*END STEP
*STEP
RELEASE LOAD
*MODAL DYNAMIC,CONTINUE=YES
0.02,10.0
*MODAL DAMPING,MODAL=DIRECT
1,24,0.028
*PRINT,FREQUENCY=100
*NODE PRINT,NSET=SMALL,FREQUENCY=100
U,
U,V,A,RF
2-895
*EL FILE,ELSET=SPRINGS,FREQUENCY=100
S,
S,
U,V,A,RF
S,
U,V,A,RF
S,
*END STEP
2-896
Listing 2.2.2-2
*HEADING
INDIAN POINT FEEDWATER LINE WITH SPRING SUPPORTS
** BEAM ELEMENTS WITH STEADY STATE SINE DWELL
*NODE
1, 0., 423., -234.96
3, 0., 423., -150.96
5, 0., 435., -138.96
6, 0., 474., -138.96
8, 0., 486., -126.96
10, 0., 486., -75.96
11, 0., 486., -51.96
12, 0., 486., -18.00
13, 0., 486., 9.00
15, 0., 486., 144.5
16, 0., 486., 159.
18, 8.484, 494.484, 171.
19, 8.484,494.484,171.
21, 16.93, 497.96, 171.
22, 19.8125, 497.96 , 171.
23, 29.125 ,497.96, 171.
25, 200.72, 497.96, 171.00
27, 260.72, 497.96, 171.
29, 272.72, 509.96, 171.00
31, 272.72, 569.964, 171.00
33, 280.44, 581.96, 180.19
35, 330.1 , 581.96, 239.3
36, 335.21, 581.96, 245.46
38, 342.91, 593.96, 254.65
39, 342.91 , 628. , 254.65
40, 342.91, 660., 254.65
42, 342.91, 706., 254.65
44, 340.22, 714.48, 256.91
46, 296.57, 771.47, 293.54
48, 282.36, 779.95, 289.80
49, 278.50, 779.95, 285.20
50,274.644, 779.95, 280.61
52, 266.93, 791.95, 271.42
53, 266.93, 801., 271.42
54, 266.93, 876.00, 271.42
56, 266.93, 990.96, 271.42
57, 266.93, 1000.27, 271.42
2-897
59, 278.88, 1012.27, 272.46

61, 335.26, 1012.27, 277.39
63, 343.40, 1012.27, 281.64
64, 366.97, 1012.27, 309.73
65, 369.52, 1012.27, 312.76
66, 379.16, 1012.27, 324.25
67, 388.8, 1012.27, 335.74
68, 389.11, 1012.27, 336.11
70, 396.83, 1024.27, 345.3
71, 396.83, 1027.27, 345.3
72, 396.83, 1033.27, 345.3
73, 396.83, 1040.95, 345.3
75, 389.93, 1049.95, 351.08
76, 380.74, 1049.95, 358.8
*NGEN
1,3
8,10
13,15
23,25
25,27
29,31
33,35
40,42
44,46
54,57
59,61
*NGEN,LINE=C
3,5,1,, 0., 435., -150.96
6,8,1,, 0., 474., -126.96
16,18,1,, 8.484, 494.48, 159.00
19, 21,1,, 16.932, 485.96, 171.00
27,29,1,, 260.724, 509.96, 171.00
31,33,1,, 280.44, 569.96, 180.19
36,38,1,, 335.21, 593.96, 245.46
42,44,1,, 333.71, 706.00, 262.37
46,48,1,, 288.85, 771.47, 284.35
50,52,1,, 274.64, 791.95, 280.61
57,59,1,, 278.88, 1000.27, 272.46
61,63,1,, 334.21, 1012.27, 289.34
68,70,1,, 389.11, 1024.27, 336.11
73,75,1,, 389.93, 1040.95, 351.08
**
** SPRING DEFINITIONS
2-898
**
*NODE,NSET=SPRS
115, 24.91, 475.93, 144.5
122, 19.81, 497.96, 219.5
123, 29.13, 597.41, 160.55
135, 299.94, 555., 239.30
235, 330.10, 599.46, 239.30
139, 364.74, 628.00, 220.25
239, 359.58, 628.00, 291.83
149,278.50, 792.45, 285.20
153, 321.48, 801.00, 318.12
253, 314.43, 801.00, 212.09
156, 311.81, 990.96, 215.36
256,290.46, 1008.88, 299.46
*NSET,NSET=NPDR
25,27,33,36,42,1,76,15,22,23,35,39,49,53,56
1001,15,115
1002,22,122
1003,23,123
1004,35,135
1005,35,235
1006,39,139
1007,39,239
1008,49,149
1009,53,153
1010,53,253
1011,56,156
1012,56,256
*ELSET,ELSET=SPRINGS
FWR11, FWR13, FWR14, FWR17, FWR18, FWR20,
FWR21, FWR23, FWR24, FWR25, FWR27, FWR28
2-899
17700. ,
*SPRING,ELSET=FWR13
119600.,
403000.,
97900.,
*SPRING,ELSET=FWR18
228000.,
86300.,
*SPRING,ELSET=FWR21
86300.,
*SPRING,ELSET=FWR23
319000.,
*SPRING,ELSET=FWR24
56800.,
*SPRING,ELSET=FWR25
39100.,
*SPRING,ELSET=FWR27
55500.,
*SPRING,ELSET=FWR28
68000.,
**
** PIPE DEFINITIONS
**
*ELEMENT,TYPE=B31
1,1,2
14,13,14
20,19,20
2-900
24,22,23
28,25,26
53,49,50
*ELGEN
1,12
14,5
20,3
24,3
28,24
53,27
*MPC
BEAM,18,19
*ELSET,ELSET=D8
1,2,5,8,9,10,11,12,14,16,22 ,15
24,25,26,28,29,32,33,36,37,38,41,42,43,44,47,48
51,53,56,57,58,59,60,63,64,67,68,71
74,
*ELSET,ELSET=D8E
3, 4, 6, 7,17,18,20,21,30,31,34,35,39,40
45,46,49,50,54,55,61,62,65,66,72,73
*ELSET,ELSET=BF57
69,70
*ELSET,ELSET=BWR
75,
*ELSET,ELSET=D6
76,79
*ELSET,ELSET=D6E
77,78
DENSITY= .0010691
12.763 , 105.317 , , 105.317 , 210.635
1. , , -1.
27.9E6 , 10.73E6
DENSITY=.0010691
12.763 , 50.439 , , 50.439 , 210.635
1. , , -1.
27.9E6 , 10.73E6
DENSITY=0.00102423
8.405 , 40.295 , , 40.295 , 80.589
1.,
27.9E6 , 10.73E6
2-901

DENSITY=0.00102423
8.405 , 21.828, ,21.828, 80.589
1.,
27.9E6 , 10.73E6
*BEAM GENERAL SECTION,ELSET=BWR,SECTION=GENERAL,
DENSITY=0.0013266
10.4806 , 67.143 , , 67.143 , 134.29
1. , , -1.
27.9E6 , 10.73E6
*BEAM GENERAL SECTION,ELSET=BF57,SECTION=GENERAL,
DENSITY=0.002591
30.631 ,161.77 , , 161.77 , 323.53
1. , , -1.
27.9E6 , 10.73E6
*NSET,NSET=SMALL
25,27,33
*BOUNDARY
1,1,6
76,1,6
70,1
70,2
10,2
SPRS,1,3
*AMPLITUDE,NAME=AMP
0.0,1.0,2.0,18.46,2.4,26.54,2.8,36.16,
3.0,41.55,3.4,53.35,3.8,66.64,4.0,73.84,
4.1,77.58,4.2,81.44,4.3,85.34,4.4,89.35,
4.6,97.66,4.8,106.34,5.0,115.38,5.2,124.79,
5.4,134.58,5.6,144.73,5.8,155.26,6.0,166.15,
6.1,171.73,6.2,177.41,6.35,186.1,6.4,189.04,
6.6,201.04,6.8,213.41,7.0,226.15,7.2,239.26,
7.4,252.73,7.8,280.79,8.0,295.38
*STEP
*FREQUENCY
24,
*EL PRINT,ELSET=SPRINGS
S11,E11
*END STEP
*STEP
*STEADY STATE DYNAMICS,FREQUENCY SCALE=LINEAR
0.02,2.0,3,,
2-902
2.4,8.0,27,1.0
*CLOAD,AMPLITUDE=AMP
27,3,1.0
1,24,0.028
S11,E11
*NODE PRINT,NSET=SMALL,FREQUENCY=10
U,
U,V,A
*MODAL PRINT,FREQUENCY=10
GU,
GA,
GPU,
*EL FILE,ELSET=SPRINGS,FREQUENCY=10
S,
U,V,A
S,
U,V,A
S,
*END STEP
2.2.3 Response spectra of a three-dimensional frame building

The purpose of this example is to verify the different summation methods for natural modes in the
*RESPONSE SPECTRUM procedure. To compare the five different methods that are available in
ABAQUS, a three-dimensional model with closely spaced modes is examined.
Geometry and model

A four-story steel-frame building is analyzed. All columns in the building have the same geometric
properties. However, as shown in Figure 2.2.3-1, the properties of the beams in Frames 1 and 2 are
different, as compared to those in Frames 3 and 4, to move the center of mass of the structure away
from its geometric center. Eigenvalue extraction performed on the model shows that many of the 30
modes that cover the frequency range up to 40 Hz are closely spaced. An acceleration spectrum based
on the El Centro earthquake record is applied in the x-y plane. The FORTRAN program given in
2-903
frameresponsespect_acc.f is used to generate the spectrum. The frequency range is chosen between 0.1
Hz and 40 Hz, and the number of points at which the spectrum is calculated is set at 501. Only one
spectrum curve is requested for 2% damping.

As described in ``Linear analysis of a rod under dynamic loading,'' Section 1.4.9 of the ABAQUS
Benchmarks Manual, for structures with well-separated modes the TENP and the CQC methods reduce
to the SRSS method, while the NRL and the ABS methods give similar results. Hence, for such
structures, two summation rules would suffice, with ABS providing the more conservative results.
However, when structures with closely spaced modes are analyzed, all five summation rules can yield
very different results. This is even more apparent in three-dimensional problems. In the present
example, the plane of the earthquake motion lies along the x-axis, so we expect that the structural
response will be dominated by Frames 1 and 3 and will result in a significant base shear in the
x-direction. All five methods are compared against a modal time history response using the same El
Centro acceleration record in Table 2.2.3-1, where the base shear forces are summed up in the plane of
each frame Si , where i is the frame number. This comparison shows that the best approximation is
generated by the CQC method. The other methods overestimate the shear in the y-direction, and some
of them underestimate the base shear in the x-direction. The CQC method is generally recommended
for asymmetrical three-dimensional problems with closely spaced structural modes. This method takes
into account the sign of the mode shapes through cross-modal correlation factors and can correctly
predict the response in directions perpendicular to the direction of excitation.
Input files
frameresponsespect_freq.inp
*FREQUENCY analysis.
frameresponsespect_rs.inp
*RESPONSE SPECTRUM analysis.
frameresponsespect_modal.inp
*MODAL DYNAMIC analysis.
frameresponsespect_acc.f
FORTRAN program that will produce the acceleration spectrum needed to run
frameresponsespect_rs.inp.
Table
Table 2.2.3-1 Comparison of base shear forces for different summation methods.
Method S1 (kip) S2 (kip) S3 (kip) S4 (kip)
Time history -25.5 14.0 -37.0 -22.8
ABS 52.5 52.5 69.6 69.6
2-904
SRSS 20.9 20.9 26.8 26.8

TENP 33.1 33.1 38.8 38.8
NRL 29.3 29.3 37.2 37.2
CQC 26.6 14.6 31.6 22.1
Figure
Figure 2.2.3-1 Three-dimensional frame system.
Sample listings
2-905
Listing 2.2.3-1
*HEADING
3-D BUILDING SUBJECTED TO EARTHQUAKE
RESP. SPECTRUM
*NODE,NSET=BOT
1,
5,200.,0.,0.,0.
9,350.,0.,0.,0.
13,550.,0.,0.,0.
17,550.,200.,0.
21,550.,350.,0.
25,550.,550.,0.
29,350.,550.,0.
33,200.,550.,0.
37,0.,550.,0.
41,0.,350.,0.
45,0.,200.,0.
*NCOPY,OLD SET=BOT,NEW SET=TOP,SHIFT,
CHANGE NUMBER=16000
0.,0.,400.
0.,0.,0.,0.,0.,10.,0.
*NFILL
BOT,TOP,16,1000
*NGEN
4021,4025,1
*NGEN,NSET=B1
4001,4005,1
4005,4009,1
4009,4013,1
*NGEN,NSET=B3
4013,4017,1
4017,4021,1
4021,4024,1
*NCOPY,OLD SET=B1,NEW SET=B2,REFLECT=POINT,
CHANGE NUMBER=24
275.,275,100.
*NCOPY,OLD SET=B3,NEW SET=B4,REFLECT=POINT,
CHANGE NUMBER=24
275.,275,100.
*NSET,NSET=BOT1
B1,B2,B3,B4
2-906
*NCOPY,OLD SET=BOT1,NEW SET=TOP1,SHIFT,

CHANGE NUMBER=12000
0.,0.,300.
0.,0.,0.,0.,0.,10.,0.
*NFILL
BOT1,TOP1,3,4000
**
** NODE USED TO DEFINE THE SECTION ORIENTATION
** OF ELSET C1
*NODE
50001,1000,0
**
**CREATE ALL COLUMNS
*ELEMENT,TYPE=B32,ELSET=C1
1,1,1001,2001,50001
*ELGEN,ELSET=C1
1,8,2000,1,12,4,8
**CREATE BEAMS IN FRAME 1
*ELEMENT,TYPE=B32,ELSET=B1
101,4001,4002,4003
*ELGEN,ELSET=B1
101,4,4000,1,6,2,4
125,4013,4014,4015
*ELGEN,ELSET=B2
125,4,4000,1,6,2,4
149,4025,4026,4027
*ELGEN,ELSET=B3
149,4,4000,1,6,2,4
173,4037,4038,4039
*ELGEN,ELSET=B4
173,4,4000,1,5,2,4
193,4047,4048,4001
194,8047,8048,8001
195,12047,12048,12001
196,16047,16048,16001
*ELSET,ELSET=THREE
2-907
1,9,104
*ELASTIC
30.E6,
*DENSITY
0.000728,
*BEAM SECTION,SECTION=BOX,MATERIAL=STEEL,ELSET=C1
14.,14.,1.5,1.5,1.5,1.5
*BEAM SECTION,SECTION=BOX,MATERIAL=STEEL,ELSET=B1
8.0,10.,1.0,1.0,1.0,1.0
0.,1.,0.
8.0,10.,1.0,1.0,1.0,1.0
-1.,0.,0.
12.,14.,1.2,1.2,1.2,1.2
0.,-1.,0.
12.,14.,1.2,1.2,1.2,1.2
1.,0.,0.
*BOUNDARY
BOT,1,6
*STEP
*FREQUENCY
30,
*NODE PRINT,NSET=TOP
U,
*EL FILE,ELSET=THREE
SF,
*OUTPUT,FIELD
*ELEMENT OUTPUT,ELSET=THREE
SF,
*NODE FILE,NSET=TOP
U,
*OUTPUT,FIELD
*NODE OUTPUT,NSET=TOP
U,
*MODAL FILE
*MODAL OUTPUT
*END STEP
2-908
Listing 2.2.3-2
*HEADING
RESPONSE SPECTRUM FOR 3-D
BUILDING
*RESTART,READ,STEP=1,INC=1,WRITE,FREQUENCY=0
*ELSET,ELSET=SMALL
1,9,104,112,116,148,152,160,172,176
*SPECTRUM,TYPE=ACCELERATION,INPUT=SPECTRUM.ACC,
NAME=SPEC
*STEP
*RESPONSE SPECTRUM,SUM=ABS,COMP=ALGEBRAIC
SPEC,1.,0.,0.,1.
1,30, 0.02
*EL PRINT,ELSET=SMALL
SF,
*NODE PRINT,NSET=TOP
U,
*NODE PRINT
RF,
*NODE FILE,NSET=TOP
U,
*NODE FILE
RF,
*EL FILE,ELSET=SMALL
SF,
*END STEP
*STEP
*RESPONSE SPECTRUM,SUM=SRSS,COMP=ALGEBRAIC
SPEC,1.,0.,0.,1.
1,30, 0.02
*END STEP
*STEP
*RESPONSE SPECTRUM,SUM=SRSS,COMP=SRSS
SPEC,1.,0.,0.,1.
1,30, 0.02
*END STEP
*STEP
*RESPONSE SPECTRUM,SUM=CQC,COMP=ALGEBRAIC
SPEC,1.,0.,0.,1.
2-909
1,30, 0.02
*END STEP
*STEP
*RESPONSE SPECTRUM,SUM=NRL,COMP=ALGEBRAIC
SPEC,1.,0.,0.,1.
1,30, 0.02
*END STEP
*STEP
*RESPONSE SPECTRUM,SUM=NRL,COMP=SRSS
SPEC,1.,0.,0.,1.
1,30, 0.02
*END STEP
*STEP
*RESPONSE SPECTRUM,SUM=TENP,COMP=SRSS
SPEC,1.,0.,0.,1.
1,30, 0.02
*END STEP
*STEP
*RESPONSE SPECTRUM,SUM=TENP,COMP=ALGEBRAIC
SPEC,1.,0.,0.,1.
1,30, 0.02
*END STEP
****************************
**multidirectional spectra**
****************************
*STEP
*RESPONSE SPECTRUM,SUM=SRSS,COMP=ALGEBRAIC
SPEC,1.,0.,0.,1.
SPEC,0.,1.,0.,1.
1,30, 0.02
*END STEP
*STEP
*RESPONSE SPECTRUM,SUM=SRSS,COMP=SRSS
SPEC,1.,0.,0.,1.
SPEC,0.,1.,0.,1.
1,30, 0.02
2-910
*END STEP
*STEP
*RESPONSE SPECTRUM,SUM=TENP,COMP=SRSS
SPEC,1.,0.,0.,1.
SPEC,0.,1.,0.,1.
1,30, 0.02
*END STEP
*STEP
*RESPONSE SPECTRUM,SUM=CQC,COMP=SRSS
SPEC,1.,0.,0.,1.
SPEC,0.,1.,0.,1.
SPEC,0.,0.,1.,1.
1,30, 0.02
*END STEP
2-911
Listing 2.2.3-3
*HEADING
3-D BUILDING SUBJECTED TO EARTHQUAKE RECORD
*RESTART,READ,STEP=1,INC=1,WRITE,FREQUENCY=0
*AMPLITUDE,VALUE=ABSOLUTE,TIME=STEP TIME,
INPUT=QUAKE.AMP,NAME=EQ
*ELSET,ELSET=ALL
104,112,116,120,124,128,136,
148,152,160,172,176,184,196
*STEP
*MODAL DYNAMIC
0.01,10.
1,30,0.02
*BASE MOTION,AMPLITUDE=EQ,DOF=1,SCALE=386.09
*NODE PRINT,NSET=TOP,FREQUENCY=100
U,
*EL PRINT,ELSET=ALL,FREQUENCY=100
SF,
RF,
** IN ORDER TO GET RESULTS FOR TABLE 3.1.16-1,
** THE NODE FILE SHOULD BE WRITTEN WITH A
** FREQUENCY=2
**NODE FILE,FREQUENCY=2
RF,
*END STEP
2-912
Tire Analyses
3. Tire Analyses
3.1 Tire analyses
3.1.1 Symmetric results transfer for a static tire analysis
This example illustrates the use of the *SYMMETRIC RESULTS TRANSFER option as well as the
*SYMMETRIC MODEL GENERATION option to model the static interaction between a tire and a
flat rigid surface.
The *SYMMETRIC MODEL GENERATION option (``Symmetric model generation,'' Section 7.7.1 of
the ABAQUS/Standard User's Manual) can be used to create a three-dimensional model by revolving
an axisymmetric model about its axis of revolution or by combining two parts of a symmetric model,
where one part is the original model and the other part is the original model reflected through a line or
a plane. Both model generating techniques are demonstrated in this example.
The *SYMMETRIC RESULTS TRANSFER option (``Transferring results from a symmetric mesh to a
three-dimensional mesh,'' Section 7.7.2 of the ABAQUS/Standard User's Manual) allows the user to
transfer the solution obtained from an axisymmetric analysis onto a three-dimensional model with the
same geometry. It also allows the transfer of a symmetric three-dimensional solution to a full
three-dimensional model. Both these results transfer features are demonstrated in this example. The
results transfer capability can significantly reduce the analysis cost of structures that undergo
symmetric deformation, followed by nonsymmetric deformation later during the loading history.
The purpose of this example is to obtain the footprint solution of a 175 SR14 tire subjected to an
inflation pressure and a concentrated load on the axle, which represents the weight of the vehicle. The
footprint solution is used as a starting point in ``Steady-state rolling analysis of a tire,'' Section 3.1.2,
where the free rolling state of the tire rolling at 10 km/h is determined, and in ``Subspace-based
steady-state dynamic tire analysis,'' Section 3.1.3, where a frequency response analysis is performed.
Problem description
The different components of the tire are shown in Figure 3.1.1-1. The tread and sidewalls are made of
rubber, and the belts and carcass are constructed from fiber reinforced rubber composites. The rubber
is modeled as an incompressible hyperelastic material, and the fiber reinforcement is modeled as a
linear elastic material. A small amount of skew symmetry is present in the geometry of the tire due to
the placement and §20.0° orientation of the reinforcing belts.
Two simulations are performed in this example. The first simulation exploits the symmetry in the tire
model and utilizes the results transfer capability; the second simulation does not use the results transfer
capability. Comparisons between the two methodologies are made.
The first simulation is broken down into three separate analyses. In the first analysis the inflation of the
tire by a uniform internal pressure is modeled. Due to the anisotropic nature of the tire construction,
the inflation loading gives rise to a circumferential component of deformation. The resulting stress
field is fully three-dimensional, but the problem remains axisymmetric in the sense that the solution
3-913
Tire Analyses
does not vary as a function of position along the circumference. ABAQUS provides axisymmetric
elements with twist (CGAX) for such situations. These elements are used to model the inflation
loading. Only half the tire cross-section is needed for the inflation analysis due to a reflection
symmetry through the vertical line that passes through the tire axle (see Figure 3.1.1-2). We refer to
this model as the axisymmetric model.
The second part of the simulation entails the computation of the footprint solution, which represents
the static deformed shape of the pressurized tire due to a vertical dead load (modeling the weight of a
vehicle). A three-dimensional model is needed for this analysis. The finite element mesh for this model
is obtained by revolving the axisymmetric cross-section about the axis of revolution. A nonuniform
discretization along the circumference is used as shown in Figure 3.1.1-3. In addition, the
axisymmetric solution is transferred to the new mesh where it serves as the initial or base state in the
footprint calculations. As with the axisymmetric model, only half of the cross-section is needed in this
simulation, but skew-symmetric boundary conditions must be applied along the mid-plane of the
cross-section to account for antisymmetric stresses that result from the inflation loading and the
concentrated load on the axle. We refer to this model as the partial three-dimensional model.
In the last part of this analysis the footprint solution from the partial three-dimensional model is
transferred to a full three-dimensional model and brought into equilibrium. This full three-dimensional
model is used in the steady-state transport example that follows. The model is created by combining
two parts of the partial three-dimensional model, where one part is the mesh used in the second
analysis and the other part is the partial model reflected through a line. We refer to this model as the
full three-dimensional model.
A second simulation is performed in which the same loading steps are repeated, except that the full
three-dimensional model is used for the entire analysis. Besides being used to validate the results
transfer solution, this second simulation allows us to demonstrate the computational advantage
afforded by the ABAQUS results transfer capability in problems with rotational and/or reflection
symmetries.
Model definition
In the first simulation the inflation step is performed on the axisymmetric model and the results are
stored in the results files ( .res, .mdl, .stt, and .prt). The axisymmetric model is discretized
with CGAX4H and CGAX3H elements. The belts and carcass are modeled by defining rebar in the
continuum elements, and the road is defined as an analytical rigid surface. The axisymmetric results
are read into the subsequent footprint analysis, and the partial three-dimensional model is generated by
ABAQUS by revolving the axisymmetric model cross-section about the rotational symmetry axis. The
*SYMMETRIC MODEL GENERATION, REVOLVE option is used for this purpose. The road is
defined in the partial three-dimensional model. The results of the footprint analysis are read into the
final equilibrium analysis, and the full three-dimensional model is generated by reflecting the partial
three-dimensional model through a vertical line using the *SYMMETRIC MODEL GENERATION,
REFLECT=LINE option. The line used in the reflection is the vertical line in the symmetry plane of
the tire, which passes through the axis of rotation. The REFLECT=LINE parameter is used, as opposed
to the REFLECT=PLANE parameter, to take into account the skew symmetry of the tire. The
analytical rigid surface as defined in the partial three-dimensional model is transferred to the full
3-914
Tire Analyses
model without change. The three-dimensional finite element mesh of the full model is shown in Figure
3.1.1-4.
In the second simulation a datacheck analysis is performed to write the axisymmetric model
information to the results files. The full tire cross-section is meshed in this model. No analysis step is
performed. The axisymmetric model information is read in a subsequent run, and a full
three-dimensional model is generated by ABAQUS by revolving the cross-section about the rotational
symmetry axis. The *SYMMETRIC MODEL GENERATION, REVOLVE option is again used for this
purpose. The road is defined in the full model. The three-dimensional finite element mesh of the full
model is identical to the one generated in the first analysis. However, the inflation load and
concentrated load on the axle are applied to the full model without making use of the results transfer
capability.
During the model generation from the axisymmetric to the three-dimensional meshes in both analyses,
the axisymmetric CGAX4H and CGAX3H elements are converted into C3D8H and C3D6H elements,
respectively.
The footprint calculations in both analyses are performed with a friction coefficient of zero in
anticipation of eventually performing a steady-state rolling analysis of the tire using the *STEADY
STATE TRANSPORT option, as explained in ``Steady-state rolling analysis of a tire,'' Section 3.1.2.
Since the results from the static analyses performed in this example are used in a subsequent
time-domain dynamic example, there are a few features in the input files that would not ordinarily be
included for purely static analyses. It is instructive to point out and to discuss these features
briefly.
The TRANSPORT parameter is included with the *SYMMETRIC MODEL GENERATION option to
define streamlines in the model, which are needed by ABAQUS to perform streamline calculations
during the *STEADY STATE TRANSPORT analysis in the next example problem. The TRANSPORT
parameter is not required for any other analysis type except for *STEADY STATE TRANSPORT.
The hyperelastic material, which models the rubber, has a *VISCOELASTIC, TIME=PRONY option
included. This enables us to model viscoelasticity in the steady-state transport example that follows.
As a consequence of defining a time-domain viscoelastic material property, the *HYPERELASTIC
option includes the LONG TERM parameter to indicate that the elastic properties defined in the
associated data lines define the long-term behavior of the rubber. In addition, all *STATIC steps
include the LONG TERM parameter to ensure that the static solutions are based upon the long-term
elastic moduli.
Loading
As discussed in the previous sections, the loading on the tire is applied over several steps. In the first
simulation the inflation of the tire to a pressure of 200.0 kPa is modeled using the axisymmetric tire
model (tiretransfer_axi_half.inp) with a *STATIC analysis procedure. The results from this
axisymmetric analysis are then transferred to the partial three-dimensional model
(tiretransfer_symmetric.inp) in which the footprint solution is computed in two sequential *STATIC
steps. The first of these static steps establishes the initial contact between the road and the tire by
3-915
Tire Analyses
prescribing a vertical displacement of 0.02 m on the rigid body reference node. Since this is a static
analysis, it is recommended that contact be established with a prescribed displacement, as opposed to a
prescribed load, to avoid potential convergence difficulties that might arise due to unbalanced forces.
The prescribed boundary condition is removed in the second static step, and a vertical load of N =
1.65 kN is applied to the rigid body reference node. The 1.65 kN load in the partial three-dimensional
model represents a 3.3 kN load in the full three-dimensional model. The transfer of the results from the
axisymmetric model to the partial three-dimensional model is accomplished by using the
*SYMMETRIC RESULTS TRANSFER option. Once the static footprint solution for the partial
three-dimensional model has been established, the *SYMMETRIC RESULTS TRANSFER option is
used once again to transfer the solution to the full three-dimensional model ( tiretransfer_full.inp),
where the footprint solution is brought into equilibrium in a single *STATIC increment. The results
transfer sequence is illustrated in Figure 3.1.1-5.
It is important to note that boundary conditions and loads are not transferred with the *SYMMETRIC
RESULTS TRANSFER option; they must be carefully redefined in the new analysis to match the loads
and boundary conditions from the transferred solution. Due to numerical and modeling issues the
element formulation for the two-dimensional and three-dimensional elements are not identical. As a
result, there may be slight differences between the equilibrium solutions generated by the two- and
three-dimensional models. In addition, small numerical differences may occur between the symmetric
and full three-dimensional solutions because of the presence of symmetry boundary conditions in the
symmetric model that are not used in the full model. Therefore, it is advised that in a results transfer
simulation an initial step be performed where equilibrium is established between the transferred
solution and loads that match the state of the model from which the results are transferred. Since the
transferred solution is applied in full at time t = 0, the external loads must also be applied in full at the
beginning of the initial step. There is no benefit in reducing the magnitude of the loads to overcome
convergence problems. To ensure that ABAQUS does not waste computational time by attempting
smaller time increments if equilibrium cannot be attained, it is recommended that the initial step
should consist of a *STATIC, DIRECT procedure with the initial time increment set to the total step
time.
In the second simulation identical inflation and the footprint steps are repeated. The only difference is
that the entire analysis is performed on the full three-dimensional model
(tiretransfer_full_footprint.inp). The full three-dimensional model is generated using the restart
information from a datacheck analysis on an axisymmetric model of the full tire cross-section
(tiretransfer_axi_full.inp).
Solution controls
Since the three-dimensional tire model has a small loaded area and, thus, rather localized forces, the
default averaged flux values for the convergence criteria produce very tight tolerances and cause more
iteration than is necessary for an accurate solution. To decrease the computational time required for the
analysis, the *CONTROLS option can be used to override the default values for average forces and
moments. The default controls are used in this example.
3-916
Tire Analyses
The results from the two simulations are essentially identical. The peak Mises stresses and
displacement magnitudes in the two models agree within 0.3% and 0.2%, respectively. The final
deformed shape of the tire is shown in Figure 3.1.1-6. The computational cost of each simulation is
shown in Table 3.1.1-1. The simulation performed on the full three-dimensional model took 2.5 times
longer than the results transfer simulation --clearly demonstrating the computational advantage that can
be attained by exploiting the symmetry in the model using the *SYMMETRIC RESULTS TRANSFER
option.
Input files
tiretransfer_axi_half.inp
Axisymmetric model, inflation analysis (simulation 1).
tiretransfer_symmetric.inp
Partial three-dimensional model, footprint analysis (simulation 1).
tiretransfer_full.inp
Full three-dimensional model, final equilibrium analysis (simulation 1).
tiretransfer_axi_full.inp
Axisymmetric model, datacheck analysis (simulation 2).
tiretransfer_full_footprint.inp
Full three-dimensional model, complete analysis (simulation 2).
tiretransfer_node.inp
Nodal coordinates for both axisymmetric models.
Table
Table 3.1.1-1 Comparison of normalized CPU times to perform the footprint analysis (normalized
with respect to the total "No results transfer" analysis).
Use results transfer No results
and symmetry transfer
conditions
Inflation 0.002(a)+0.039(b) 0.36(e)
Footprint (c)
0.29 +0.061 (d) 0.64(e)
Total 0.39 1.0
(a) axisymmetric model
(b) equilibrium step in partial three-dimensional
model
(c) footprint analysis in partial three-dimensional
model
(d) equlibrium step in full three-dimensional model
3-917
Tire Analyses
(e) full three-dimensional model
Figures
Figure 3.1.1-1 Tire cross-section.
Figure 3.1.1-2 Axisymmetric tire mesh.
Figure 3.1.1-3 Partial three-dimensional tire mesh.
3-918
Tire Analyses
Figure 3.1.1-4 Full three-dimensional tire mesh.
Figure 3.1.1-5 Results transfer analysis sequence.
3-919
Tire Analyses
Figure 3.1.1-6 Deformed three-dimensional tire (Deformations scaled by a factor of 2).
Sample listings
3-920
Tire Analyses
Listing 3.1.1-1
*HEADING
SYMMETRIC RESULTS TRANSFER FOR TIRE MODEL
TIRETRANSFER_AXI_HALF
AXISYMMETRIC HALF TIRE MODEL
STEP 1: INFLATE TIRE TO 200 KPa
UNITS: KG, M
*NODE,NSET=NTIRE,INP=tiretransfer_node.inp
*ELEMENT,TYPE=CGAX4H,ELSET=TREAD
1, 50, 55, 54, 49
2, 45, 50, 49, 44
3, 40, 45, 44, 39
5, 35, 40, 39, 34
7, 31, 35, 34, 30
*ELEMENT,TYPE=CGAX3H,ELSET=TREAD
4, 27, 31, 30
*ELEMENT,TYPE=CGAX4H,ELSET=SIDE
15, 27, 30, 28, 25
16, 24, 27, 25, 22
17, 21, 24, 22, 19
18, 18, 21, 19, 16
19, 15, 18, 16, 13
20, 12, 15, 13, 10
21, 30, 34, 32, 28
29, 9, 12, 10, 7
30, 6, 9, 7, 4
31, 3, 6, 4, 1
*ELEMENT,TYPE=CGAX4H,ELSET=BELT
35, 49, 54, 51, 46
36, 44, 49, 46, 41
37, 39, 44, 41, 36
38, 34, 39, 36, 32
*REBAR,ELEMENT=CONTINUUM,MATERIAL=BELT,
GEOMETRY=ISO,NAME=BELT1
BELT, 0.2118683E-6, 1.16E-3, 70.0, 0.50, 3
*REBAR,ELEMENT=CONTINUUM,MATERIAL=BELT,
GEOMETRY=ISO,NAME=BELT2
BELT, 0.2118683E-6, 1.16E-3, 110.0, 0.83, 3
*REBAR,ELEMENT=CONTINUUM,MATERIAL=CARCASS,
GEOMETRY=ISO,NAME=CARCASS
BELT, 0.4208352E-6, 1.00E-3, 0.0, 0.0, 3
3-921
Tire Analyses
SIDE, 0.4208352E-6, 1.00E-3, 0.0, 0.0, 3

*SOLID SECTION,ELSET=TREAD,MATERIAL=RUBBER
*SOLID SECTION,ELSET=SIDE,MATERIAL=RUBBER
*SOLID SECTION,ELSET=BELT,MATERIAL=RUBBER
*HYPERELASTIC,N=1,MODULI=LONG TERM
1.0e6,
*VISCOELASTIC,TIME=PRONY
0.3, 0.0, 0.1
*DENSITY
1100.,
*MATERIAL,NAME=BELT
*ELASTIC,TYPE=ISO
172.2E+09, 0.3
*DENSITY
5900.,
*MATERIAL,NAME=CARCASS
*ELASTIC,TYPE=ISO
9.87E+9, 0.3
*DENSITY
1500.,
*NSET,NSET=RIM
6, 3, 1
*NSET,NSET=SYM
51, 54, 55
*ELSET,ELSET=SOLID,GENERATE
1, 76,1
*SURFACE,NAME=INSIDE
BELT, S3
SIDE, S3
***********************************
*STEP,INC=100,NLGEOM=YES
1: INFLATION
*STATIC, LONG TERM
0.25, 1.0
*BOUNDARY
RIM, 1, 2
RIM, 5,
SYM, 2,
SYM, 5,
*DSLOAD
INSIDE, P, 200.E3
*EL FILE,FREQUENCY=50,POSITION=NODES
3-922
Tire Analyses
S, E
U
*EL PRINT,FREQUENCY=50,REBAR,ELSET=SOLID
S
*NODE PRINT,FREQUENCY=100,TOTAL=YES
U, RF
*OUTPUT,FIELD,VARIABLE=PRESELECT,FREQ=50
*END STEP
3-923
Tire Analyses
Listing 3.1.1-2
*HEADING
SYMMETRIC RESULTS TRANSFER FOR TIRE MODEL
3D HALF TIRE MODEL
STEP 0: TRANSFER TIRE INFLATION RESULTS FROM
tiretransfer_axi_half AND GENERATE
MODEL USING *SYMMETRIC MODEL GENERATION
STEP 1: BRING TRANSFERRED AXISYMMETRIC RESULTS
TO EQUILIBRIUM
STEP 2: FOOTPRINT ANALYSIS (DISPLACEMENT CONTROL)
STEP 3: FOOTPRINT ANALYSIS (LOAD CONTROL)
UNITS: KG, M
*NODE,NSET=ROAD
9999, 0.0, 0.0, -0.02
*SYMMETRIC MODEL GENERATION,REVOLVE,ELEMENT=200,
NODE=200,TRANSPORT
0.0, 0.0, 0.0, 0.0, 1.0, 0.0
0.0, 0.0, 1.0
90.0, 3
70.0, 3
15.0, 7
10.0, 4
15.0, 7
70.0, 3
90.0, 3
*SYMMETRIC RESULTS TRANSFER, STEP=1, INC=4
*ELSET,ELSET=FOOT,GEN
1001, 4801, 200
1002, 4802, 200
1003, 4803, 200
1004, 4804, 200
1005, 4805, 200
1007, 4807, 200
*SURFACE,TYPE=CYLINDER,NAME=SROAD
0., 0.,-0.31657, 1., 0.,-0.31657
0., 1.,-0.31657
START, -0.3, 0.
LINE, 0.3, 0.
*RIGID BODY,REF NODE=9999,ANALYTICAL SURFACE=SROAD
*SURFACE,NAME=STREAD
FOOT, S3
3-924
Tire Analyses
*CONTACT PAIR,INTERACTION=SRIGID
STREAD, SROAD
*SURFACE INTERACTION,NAME=SRIGID
*FRICTION
0.0
*ELSET,ELSET=SECT,GENERATE
2800, 3200, 1
*NSET,NSET=SECT,GENERATE
2800, 3400, 1
*NSET,NSET=FOOT,ELSET=FOOT
*NSET,NSET=NOUTP,GENERATE
1055, 5055, 200
*NSET,NSET=SYM1
51,54,55,3051,3054,3055
**
** NODE SETS ASYMA,ASYMB,ASYMC, and ASYMD
** USED FOR ANTI_SYMMETRY BC's
**
*NSET,NSET=ASYMA,GENERATE,UNSORTED
255, 2855, 200
254, 2854, 200
251, 2851, 200
*NSET,NSET=ASYMB,GENERATE,UNSORTED
5855, 3255, -200
5854, 3254, -200
5851, 3251, -200
*NSET,NSET=ASYMC,GENERATE,UNSORTED
255, 1055, 200
254, 2854, 200
251, 2851, 200
*NSET,NSET=ASYMD,GENERATE,UNSORTED
5855, 5055, -200
5854, 3254, -200
5851, 3251, -200
*EQUATION
2
ASYMA, 1, 1.0, ASYMB, 1, 1.0
*EQUATION
2
ASYMA, 2, 1.0, ASYMB, 2, 1.0
*EQUATION
2
ASYMC, 3, 1.0, ASYMD, 3, -1.0
3-925
Tire Analyses

********************************************
1: BRING TRANSFERRED RESULTS TO EQUILIBRIUM
*STATIC, LONG TERM
1.0, 1.0
*BOUNDARY,OP=NEW
RIM, 1, 3
ROAD, 1, 6
SYM1, 1, 2
*DSLOAD,OP=NEW
INSIDE, P, 200.E3
*NODE PRINT,NSET=ROAD,FREQ=100
U,
RF,
*EL PRINT,FREQ=0
*NODE FILE,NSET=ROAD
U, RF
*OUTPUT,FIELD,VARIABLE=PRESELECT
*END STEP
******************************************
2: FOOTPRINT (Displacement controlled)
*STATIC, LONG TERM
0.2, 1.0
*BOUNDARY,OP=NEW
RIM, 1, 3
ROAD, 1, 2
ROAD, 4, 6
SYM1, 1, 2
ROAD, 3, , 0.02
U,
RF,
*EL PRINT,FREQ=0
U, RF
*NODE FILE,NSET=FOOT,FREQ=100
*NODE OUTPUT, NSET=ROAD
U, RF
*END STEP
3-926
Tire Analyses
****************************************
3: FOOTPRINT (Load controlled)
*STATIC, LONG TERM
1.0, 1.0
*BOUNDARY,OP=NEW
RIM, 1, 3
ROAD, 1, 2
ROAD, 4, 6
SYM1, 1, 2
*CLOAD, OP=NEW
ROAD, 3, 1650.
*END STEP
3.1.2 Steady-state rolling analysis of a tire

This example illustrates the use of the *STEADY STATE TRANSPORT option in ABAQUS
(``Steady-state transport analysis,'' Section 6.4.1 of the ABAQUS/Standard User's Manual) to model
the steady-state dynamic interaction between a rolling tire and a flat rigid surface. A steady-state
transport analysis uses a moving reference frame in which rigid body rotation is described in an
Eulerian manner and the deformation is described in a Lagrangian manner. This kinematic description
converts the steady moving contact problem into a pure spatially dependent simulation. Thus, the mesh
need be refined only in the contact region--the steady motion transports the material through the mesh.
Frictional effects, inertia effects, and history effects in the material can all be accounted for in a
*STEADY STATE TRANSPORT analysis.
The purpose of this analysis is to obtain free rolling equilibrium solutions of a 175 SR14 tire traveling
at a ground velocity of 10.0 km/h (2.7778 m/s) at different slip angles. The slip angle is the angle
between the direction of travel and the plane normal to the axle of the tire. Straight line rolling occurs
at a 0.0° slip angle.
An equilibrium solution for the rolling tire problem that has zero torque, T , applied around the axle is
referred to as a free rolling solution. An equilibrium solution with a nonzero torque is referred to as
either a traction or a braking solution depending upon the sense of T . Braking occurs when the angular
velocity of the tire is small enough such that some or all of the contact points between the tire and the
road are slipping and the resultant torque on the tire acts in an opposite sense from the angular velocity
of the free rolling solution. Similarly, traction occurs when the angular velocity of the tire is large
enough such that some or all of the contact points between the tire and the road are slipping and the
resultant torque on the tire acts in the same sense as the angular velocity of the free rolling solution.
Full braking (traction) occurs when all of the contact points between the tire and the road are
slipping.
A wheel in free rolling, traction, or braking will spin at different angular velocities, !, for the same
ground velocity, v0 : Usually the combination of ! and v0 that results in free rolling is not known in
3-927
Tire Analyses
advance. Since the steady-state transport analysis capability requires that both the rotational spinning
velocity, !, and the traveling ground velocity, v0 , be prescribed, the free rolling solution must be
found in an indirect manner. One such indirect approach is illustrated in this example.
A finite element analysis of this problem, together with experimental results, have been published by
Koishi et al. (1997).
Problem description and model definition

A description of the tire and finite element model has been given in ``Symmetric results transfer for a
static tire analysis,'' Section 3.1.1. To take into account the effect of the skew symmetry of the actual
tire in the dynamic analysis, the steady-state rolling analysis will be performed on the full
three-dimensional model, also referred to as the full model. Inertia effects are ignored since the rolling
speed is low (v0 = 10 km/h).
As stated earlier, the *STEADY STATE TRANSPORT capability in ABAQUS uses a mixed
Eulerian/Lagrangian approach in which, to an observer in the moving reference frame, the material
appears to flow through a stationary mesh. The paths that the material points follow through the mesh
are referred to as streamlines and must be computed before a steady-state transport analysis can be
performed. As discussed in ``Symmetric results transfer for a static tire analysis,'' Section 3.1.1, the
streamlines needed for the steady-state transport analyses in this example were computed using the
*SYMMETRIC MODEL GENERATION, REVOLVE, TRANSPORT option. This option generated
the three-dimensional mesh by revolving the two-dimensional tire cross-section about the symmetry
axis so that the streamlines followed the mesh lines.
The incompressible hyperelastic material used to model the rubber in this example includes a
time-domain viscoelastic component, which is enabled by the *VISCOELASTIC, TIME=PRONY
option. A simple 1-term Prony series model is used. For an incompressible material a 1-term Prony
series in ABAQUS is defined by providing a single value for the shear relaxation modulus ratio, g¹1P ,
and its associated relaxation time, ¿1 . In this example g¹1P = 0:3 and ¿1 = 0:1. The viscoelastic--i.e.,
material history--effects are included in a *STEADY STATE TRANSPORT step unless the LONG
TERM parameter is used. See ``Time domain viscoelasticity,'' Section 10.6.1 of the
ABAQUS/Standard User's Manual, for a more detailed discussion on modeling time-domain
viscoelasticity in ABAQUS.
Loading
As discussed in ``Symmetric results transfer for a static tire analysis,'' Section 3.1.1, it is recommended
that the footprint analyses be obtained with a friction coefficient of zero (so that no frictional forces are
transmitted across the contact surface). The frictional stresses for a rolling tire are very different from
the frictional stresses in a stationary tire, even if the tire is rolling at very low speed; therefore,
discontinuities may arise in the solution between the last *STATIC analysis and the first *STEADY
STATE TRANSPORT analysis. Furthermore, varying the friction coefficient from zero at the
beginning of the steady-state transport step to its final value at the end of the steady-state transport step
ensures that the changes in frictional forces reduce with smaller load increments. This is important if
ABAQUS must take a smaller load increment to overcome convergence difficulties while trying to
obtain the steady-state rolling solution.
3-928
Tire Analyses
Once the static footprint solution for the tire has been computed, the steady-state rolling contact
problem can be solved using the *STEADY STATE TRANSPORT option. The objective of the first
simulation in this example is to obtain the straight-line, steady-state rolling solutions, including full
braking and full traction, at different spinning velocities. We also compute the straight-line, free rolling
solution. In the second simulation, free rolling solutions at different slip angles are computed. In the
first and second simulations, material history effects are ignored by including the LONG TERM
parameter on the *STEADY STATE TRANSPORT steps. The third simulation repeats a portion of the
straight-line, steady-state rolling analysis from the first simulation; however, material history effects
are included by omitting the LONG TERM parameter. A steady ground velocity of 10.0 km/h is
maintained for all three simulations.
In simulation 1 (rollingtire_brake_trac.inp) the full traction solution is obtained in the first *STEADY
STATE TRANSPORT step by setting the friction coefficient, ¹, to its final value of 1.0 using the
*CHANGE FRICTION option and applying the translational ground velocity together with a spinning
angular velocity that will result in full braking. The *TRANSPORT VELOCITY and *MOTION
options are used for this purpose. An estimate of the angular velocity corresponding to full braking is
obtained as follows. A free rolling tire generally travels farther in one revolution than determined by its
center height, H, but less than determined by the free tire radius. In this example the free radius is
316.2 mm and the vertical deflection is approximately 20.0 mm, so H = 294.2 mm. Using the free
radius and the effective height, it is estimated that free rolling occurs at an angular velocity between
! = 8.78 rad/s and ! = 9.44 rad/s. Smaller angular velocities would result in braking, and larger
angular velocities would result in traction. We use an angular velocity ! = 8.0 rad/s to ensure that the
solution in the first steady-state transport step is a full braking solution (all contact points are slipping,
so the magnitude of the total frictional force across the contact surface is ¹N ).
In the second steady-state transport analysis step of the full model, the angular velocity is increased
gradually to ! = 10.0 rad/s while the ground velocity is held constant. The solution at each load
increment is a steady-state solution to the loads acting on the structure at that instant so that a series of
steady-state solutions between full braking and full traction is obtained.
In the second simulation (rollingtire_slipangles.inp) the free rolling solutions at different slip angles
are computed. The slip angle, µ, is the angle between the direction of travel and the plane normal to the
axle of the tire. In the first step the straight-line free rolling solution from the first simulation is brought
into equilibrium. This step is followed by a *STEADY STATE TRANSPORT step where the slip
angle is gradually increased from µ =0.0° at the beginning of the step to µ =3.0° at the end of the step,
so a series of steady-state solutions at different slip angles are obtained. This is accomplished by
prescribing a traveling velocity vector with components vx = v0 cos µ and vy = v0 sin µ on the
*MOTION option, where µ =0.0° in the first steady-state transport step and µ =3.0° at the end of the
second steady-state transport step.
The final simulation in this example (rollingtire_materialhistory.inp) includes a series of steady-state
solutions between full braking and full traction in which the material history effects are included.

Figure 3.1.2-1 and Figure 3.1.2-2 show the reaction force parallel to the ground (referred to as rolling
3-929
Tire Analyses
resistance) and the torque, T , on the tire axle at different angular spinning velocities. The figures show
that free rolling, T = 0.0, occurs at an angular velocity of approximately 9.0 rad/s. Full braking occurs
at spinning velocities smaller than 8.2 rad/s, and full traction occurs at velocities larger than 9.7 rad/s.
At these spinning velocities all contact points are slipping, and the rolling resistance reaches the
limiting value ¹N:
Figure 3.1.2-3 and Figure 3.1.2-4 show shear stress along the centerline of the tire surface in the free
rolling and full traction states, respectively. The distance along the centerline is measured as an angle
with respect to a plane parallel to the ground passing through the tire axle. The dashed line is the
maximum or limiting shear stress, ¹p, that can be transmitted across the surface, where p is the contact
pressure. The figures show that all contact points are slipping during full traction. During free rolling
all points stick.
A better approximation to the angular velocity that corresponds to free rolling can be made by using
the results generated by rollingtire_brake_trac.inp to refine the search about an angular velocity of 9.0
rad/s. The file rollingtire_trac_res.inp restarts the previous analysis from Step 3, Increment 11
(corresponding to an angular velocity of 9.006 rad/s) and performs a refined search up to 9.04 rad/s.
Figure 3.1.2-5 shows the torque, T , on the tire axle computed in the refined search, which leads to a
more precise value for the free rolling angular velocity of approximately 9.025 rad/s. This result is
used for the model where the free rolling solutions at different slip angles are computed.
Figure 3.1.2-6 shows the transverse force (force along the tire axle) measured at different slip angles.
The figure compares the steady-state transport analysis prediction with the result obtained from a pure
Lagrangian analysis. The Lagrangian solution is obtained by performing an explicit transient analysis
using ABAQUS/Explicit. With this analysis technique a prescribed constant traveling velocity is
applied to the tire, which is free to roll along the rigid surface. Since more than one revolution is
necessary to obtain a steady-state configuration, fine meshing is required along the full circumference;
hence, the Lagrangian solution is much more costly than the steady-state solutions shown in this
example. The figure shows good agreement between the results obtained from the two analysis
techniques.
Figure 3.1.2-7 compares the free rolling solutions with and without material history effects included.
The solid lines in the diagram represent the rolling resistance (force parallel to the ground along the
traveling direction); and the broken lines, the torque (normalized with respect to the free radius) on the
axle. The figure shows that free rolling, marked with bullet points, occurs at a higher angular velocity
when history effects are included. It also shows that the rolling resistance increases when history
effects are included. The influence of material history effects on a steady-state rolling solution is
discussed in detail in ``Steady-state spinning of a disk in contact with a foundation, '' Section 1.5.2 of
the ABAQUS Benchmarks Manual.
Acknowledgments
HKS gratefully acknowledges Hankook Tire and Yokohama Rubber Company for their cooperation in
developing the steady-state transport capability used in this example. HKS thanks Dr. Koishi of
Yokohama Rubber Company for supplying the geometry and material properties used in this example.
3-930
Tire Analyses
Input files
rollingtire_brake_trac.inp
Three-dimensional full model for the full braking and traction analyses.
rollingtire_trac_res.inp
Three-dimensional full model for the refined braking and traction analyses.
rollingtire_slipangles.inp
Three-dimensional full model for the slip angle analysis.
rollingtire_materialhistory.inp
Three-dimensional full model with material history effects.
Reference
· Koishi, M., K. Kabe, and M. Shiratori, "Tire Cornering Simulation using Explicit Finite Element
Analysis Code," 16th annual conference of the Tire Society at the University of Akron, 1997.
Figures
Figure 3.1.2-1 Rolling resistance at different angular velocities.
Figure 3.1.2-2 Torque at different angular velocities.
3-931
Tire Analyses
Figure 3.1.2-3 Shear stress along tire center (free rolling).
Figure 3.1.2-4 Shear stress along tire center (full traction).
3-932
Tire Analyses
Figure 3.1.2-5 Torque at different angular velocities (refined search).
Figure 3.1.2-6 Transverse force as a function of slip angle.
3-933
Tire Analyses
Figure 3.1.2-7 Rolling resistance and normalized torque as a function of angular velocity ( R=0.3162
m).
Sample listings
3-934
Tire Analyses
Listing 3.1.2-1
*HEADING
STEADY-STATE ROLLING ANALYSIS OF A TIRE:
ROLLINGTIRE_BRAKE_TRAC
3D FULL TIRE MODEL
STEP 0: TRANSFER TIRE INFLATION FOOTPRINT RESULTS
FROM TIRETRANSFER_FULL
STEP 1: FULL BRAKING ANAYLSIS
STEP 2: FULL TRACTION ANALYSIS
UNITS: KG, M
*RESTART,READ,STEP=1,INC=1
******************************************
*STEP,INC=300,NLGEOM=YES,UNSYMM=YES
1: STRAIGHT LINE ROLLING (Full braking)
*STEADY STATE TRANSPORT, LONG TERM
0.5, 1.0
*CHANGE FRICTION,INTERACTION=SRIGID
*FRICTION,SLIP=0.01
1.0
*TRANSPORT VELOCITY
NTIRE, 8.0
*MOTION,TYPE=VELOCITY,TRANSLATION
NTIRE, 1, , 2.7778
*CONTACT PRINT,FREQ=100,NSET=NOUTP
*CONTACT FILE,NSET=NOUTP,FREQ=100
*NODE FILE,NSET=NOUTP,FREQ=100
V, COORD
U, RF
*OUTPUT,HISTORY,FREQ=100,OP=ADD
*NODE OUTPUT,NSET=NOUTP
V, COORD
*CONTACT OUTPUT,NSET=NOUTP
CSTRESS,
*END STEP
*******************************************
*STEP,INC=300,NLGEOM=YES,UNSYMM=YES
2: STRAIGHT LINE ROLLING (Full traction)
0.1, 1.0, , 0.1
3-935
Tire Analyses
*TRANSPORT VELOCITY
NTIRE, 10.00
*END STEP
3-936
Tire Analyses
Listing 3.1.2-2
*HEADING
STEADY-STATE ROLLING ANALYSIS OF A TIRE:
STEP 0: RESTART FROM rollingtire_trac_res
STEP 1: GET EQUILIBRIUM
STEP 2: FREE ROLLING AT DIFFERENT SLIP ANGLES
UNITS: KG, M
*RESTART,READ,STEP=5,INC=2,END STEP
******************************************
1: STRAIGHT LINE FREE ROLLING
1.0, 1.0,
*TRANSPORT VELOCITY
NTIRE, 9.023
*END STEP
******************************************
2: SLIP (3 degrees)
0.1, 1.0, , 0.1
*MOTION,TYPE=VELOCITY,TRANSLATION
NTIRE, 1, , 2.774
NTIRE, 2, , 0.14538
*END STEP
3.1.3 Subspace-based steady-state dynamic tire analysis

This example illustrates the use of the *STEADY STATE DYNAMICS, SUBSPACE PROJECTION
option to model the frequency response of a tire about a static footprint solution.
The *STEADY STATE DYNAMICS, SUBSPACE PROJECTION option (``Subspace-based
steady-state dynamic analysis,'' Section 6.3.7 of the ABAQUS/Standard User's Manual) is an analysis
procedure that can be used to calculate the steady-state dynamic response of a system subjected to
harmonic excitation. It does so by the direct solution of the steady-state dynamic equations projected
onto a reduced-dimensional subspace spanned by a set of eigenmodes of the undamped system. If the
dimension of the subspace is small compared to the dimension of the original problem--i.e., if a
relatively small number of eigenmodes are used, the subspace method can offer a very cost-effective
alternative to a direct-solution steady-state analysis.
The purpose of this analysis is to obtain the frequency response of a 175 SR14 tire subjected to a
3-937
Tire Analyses
harmonic load excitation about the footprint solution discussed in ``Symmetric results transfer for a
static tire analysis,'' Section 3.1.1). The *SYMMETRIC RESULTS TRANSFER and *SYMMETRIC
MODEL GENERATION options are used to generate the footprint solution, which serves as the base
state in the steady-state dynamics calculations.
Problem description
A description of the tire being modeled has been given in ``Symmetric results transfer for a static tire
analysis,'' Section 3.1.1. In this example we exploit the symmetry in the tire model and utilize the
results transfer capability in ABAQUS to compute the footprint solution for the full three-dimensional
model in a manner identical to that discussed in ``Symmetric results transfer for a static tire analysis,''
Section 3.1.1.
Once the footprint solution has been computed, several steady-state dynamic steps are performed. Both
the *STEADY STATE DYNAMICS, DIRECT and the *STEADY STATE DYNAMICS, SUBSPACE
PROJECTION options are used. Besides being used to validate the subspace projection results, the
direct steady-state procedure allows us to demonstrate the computational advantage afforded by the
subspace projection capability in ABAQUS.
Model definition
The model used in this analysis is essentially identical to that used in the first simulation discussed in
``Symmetric results transfer for a static tire analysis,'' Section 3.1.1, with CGAX4H and CGAX3H
elements used in the axisymmetric model and rebar in the continuum elements for the belts and
carcass. However, since no *STEADY STATE TRANSPORT steps are performed in this example, the
TRANSPORT parameter is not needed during the symmetric model generation phase. In addition,
instead of using a nonuniform discretization about the circumference, the uniform discretization shown
in Figure 3.1.3-1 is used.
The analysis procedures available with the *STEADY STATE DYNAMICS option are all
frequency-domain procedures. In contrast, the *STEADY STATE TRANSPORT option discussed in
``Steady-state rolling analysis of a tire,'' Section 3.1.2, is a time-domain procedure.
The incompressible hyperelastic material used to model the rubber in this example includes a
frequency-domain viscoelastic model, which is activated by the *VISCOELASTIC,
FREQUENCY=TABULAR option. This is different from the time-domain viscoelastic model
(*VISCOELASTIC, TIME=PRONY) that was used in the steady-state transport example. The
FREQUENCY=TABULAR option requires the user to provide tabular values of !<(g¤ ) and !=(g¤ )
as functions of frequency, where g ¤ is the Fourier transform of the nondimensional shear relaxation
function g (t) = GGR1(t) ¡ 1 . For consistency, the 1-term Prony series viscoelastic model used in the
steady-state rolling example was converted into an equivalent frequency-domain viscoelastic model for
this example using the following relationships:
g¹1P !¿1
!<(g¤ ) = P 1 + !2 ¿ 2
;
1 ¡ g¹1 1
3-938
Tire Analyses
g¹1P ! 2 ¿12
!=(g¤ ) = ¡ ;
1 ¡ g¹1P 1 + ! 2 ¿12
where g¹1P = 0:3 is the modulus ratio for the first term in the Prony series expansion for the shear
relaxation modulus and ¿1 = 0:1 is the relaxation time for the first term in the Prony series expansion.
Since the material is incompressible, no viscoelastic data are needed for the volumetric behavior. See
``Frequency domain viscoelasticity,'' Section 10.6.2 of the ABAQUS/Standard User's Manual, for a
more detailed discussion on frequency-domain viscoelasticity.
Loading
The loading sequence for computing the footprint solution is identical to that discussed in ``Symmetric
results transfer for a static tire analysis,'' Section 3.1.1, with the axisymmetric model contained in
tiredynamic_axi_half.inp, the partial three-dimensional model in tiredynamic_symmetric.inp, and the
full three-dimensional model in tiredynamic_freqresp.inp. Since the NLGEOM=YES parameter is
active for the *STATIC steps used in computing the footprint solution, the steady-state dynamic
analyses, which are linear perturbation procedures, are performed about the nonlinear deformed shape
of the footprint solution.
The first frequency response analyses of the tire are performed using the *STEADY STATE
DYNAMICS, SUBSPACE PROJECTION option. The excitation is due to a harmonic vertical load of
200 N, which is applied to the analytical rigid surface through its reference node. The frequency is
swept from 80 Hz to 130 Hz. The rim of the tire is held fixed throughout the analysis. Prior to the
subspace analysis being performed, the eigenmodes that are used for the subspace projection are
computed in a *FREQUENCY step. In the frequency step the first 20 eigenpairs are extracted, for
which the computed eigenvalues range from 50 to 185 Hz.
The accuracy of the subspace analysis can be improved by including some of the stiffness associated
with frequency-dependent material properties--i.e., viscoelasticity--in the eigenmode extraction step.
This is accomplished by using the PROPERTY EVALUATION parameter with the *FREQUENCY
option. In general, if the material response does not vary significantly over the frequency range of
interest, the value for the PROPERTY EVALUATION parameter can be set to the center of the
frequency span. Otherwise, more accurate results will be obtained by running several separate
frequency analyses over smaller frequency ranges with appropriate settings for the PROPERTY
EVALUATION parameter. In this example a single frequency sweep is performed with PROPERTY
EVALUATION=105 Hz.
The main advantage that the subspace projection method offers over mode-based techniques
(``Mode-based steady-state dynamic analysis,'' Section 6.3.6 of the ABAQUS/Standard User's Manual)
is that it allows frequency-dependent material properties, such as viscoelasticity, to be included
directly in the analysis. However, there is a cost involved in assembling the projected equations, and
this cost must be taken into account when deciding between a subspace solution and a direct solution.
ABAQUS offers four different parameter values that may be assigned to the SUBSPACE
PROJECTION option to control how often the projected subspace equations are recomputed. These
values are ALL FREQUENCIES, in which new projected equations are computed for every frequency
in the analysis; EIGENFREQUENCY, in which projected equations are recomputed only at the
3-939
Tire Analyses
eigenfrequencies; PROPERTY CHANGE, in which projected equations are recomputed when the
stiffness and/or damping properties have changed by a user-specified percentage; and CONSTANT,
which computes the projected equations only once at the center of the frequency range specified in the
data lines of the *STEADY STATE DYNAMICS option. Setting SUBSPACE PROJECTION=ALL
FREQUENCIES is, in general, the most accurate option; however, the computational overhead
associated with recomputing the projected equations at every frequency can significantly reduce the
cost benefit of the subspace method versus a direct solution. The SUBSPACE
PROJECTION=CONSTANT option is the most inexpensive choice, but it should be chosen only when
the material properties do not depend strongly on frequency. In general, the accuracy and cost
associated with the four SUBSPACE PROJECTION parameter values are strongly problem dependent.
In this example problem the results and computational expense of all four parameter values for
SUBSPACE PROJECTION are discussed.
The results from the various subspace analyses are compared to the results from a *STEADY STATE
DYNAMICS, DIRECT analysis.

Each of the subspace analyses utilizes all 20 modes extracted in the *FREQUENCY step. Figure
3.1.3-2 shows the frequency response plots of the vertical displacements of the road's reference node
for the direct solution along with the four subspace solutions using each of the SUBSPACE
PROJECTION parameter values discussed above. Similarly, Figure 3.1.3-3 shows the frequency
response plots of the horizontal displacement of a node on the tire's sidewall for the same five
analyses. As illustrated in Figure 3.1.3-2 and Figure 3.1.3-3, all four of the subspace projection
methods yield almost identical solutions; except for small discrepancies in the vertical displacements
at 92 and 120 Hz, the subspace projection solutions closely match the direct solution as well. Timing
results shown in Table 3.1.3-1 show that the SUBSPACE PROJECTION method results in savings in
CPU time versus the direct solution.
Input files
tiredynamic_axi_half.inp
Axisymmetric model, inflation analysis.
tiredynamic_symmetric.inp
Partial three-dimensional model, footprint analysis.
tiredynamic_freqresp.inp
Full three-dimensional model, steady-state dynamic analyses.
tiretransfer_node.inp
Nodal coordinates for axisymmetric model.
Table
3-940
Tire Analyses
Table 3.1.3-1 Comparison of normalized CPU times (normalized with respect to the DIRECT
analysis) to perform frequency sweep from 80 Hz to 130 Hz and the *FREQUENCY step.
Normalized
CPU time
SUBSPACE PROJECTION=ALL FREQUENCIES 0.89
SUBSPACE PROJECTION=EIGENFREQUENCY 0.54
SUBSPACE PROJECTION=PROPERTY CHANGE 0.49
SUBSPACE PROJECTION=CONSTANT 0.36
DIRECT 1.0
*FREQUENCY 0.073
Figures
Figure 3.1.3-1 Uniform three-dimensional tire mesh.
Figure 3.1.3-2 Frequency response of the vertical road displacement due to a vertical harmonic point
load of 200 N applied to the reference node.
3-941
Tire Analyses
Figure 3.1.3-3 Frequency response of the horizontal sidewall displacement due to a vertical harmonic
point load of 200 N applied to the reference node.
Sample listings
3-942
Tire Analyses
Listing 3.1.3-1
*HEADING
SYMMETRIC RESULTS TRANSFER FOR STATIC TIRE
ANALYSIS: TIREDYNAMIC_FREQRESP
3D FILE TIRE MODEL
STEP 0: TRANSFER TIRE FOOTPRINT RESULTS FROM
TIREDYNAMIC_SYMMETRIC AND GENERATE MODEL
USING *SYMMETRIC MODEL GENERATION
STEP 1: BRING TRANSFERRED RESULTS TO EQUILIBRIUM
STEP 2: FREQUENCY ANALYSIS
STEP 3: STEADY STATE DYNAMICS,
SUBSPACE PROJECTION=ALL FREQUENCIES
SUBSPACE PROJECTION=EIGENFREQUENCY
SUBSPACE PROJECTION=CONSTANT
SUBSPACE PROJECTION=PROPERTY CHANGE
STEP 7: STEADY STATE DYNAMICS,DIRECT
UNITS: KG, M
*SYMMETRIC MODEL GENERATION,REFLECT=LINE,
ELEMENT=100000,NODE=100000
0.0, 0.0, 0.0, 0.0, 0.0, 1.0
*SYMMETRIC RESULTS TRANSFER,STEP=3,INC=1
*NSET,NSET=SIDE
6015,
***********************************************
1: BRING TRANSFERRED RESULTS TO EQUILIBRIUM
*STATIC
1.0, 1.0
*BOUNDARY,OP=NEW
RIM, 1, 3
ROAD, 1, 2
ROAD, 4, 6
*DSLOAD,OP=NEW
INSIDE, P, 200.E3
*CLOAD,OP=NEW
ROAD, 3, 3300.
3-943
Tire Analyses
U,
RF,
*EL PRINT,FREQ=0
*NODE FILE,NSET=ROAD,FREQ=1
U,RF
*EL FILE,FREQ=0
*NODE OUTPUT, NSET=ROAD
U,RF
*END STEP
************************************************
*STEP
2: FREQUENCY ANALYSIS
*FREQUENCY,EIGENSOLVER=LANCZOS,
PROPERTY EVALUATION=105
20,0.,200.,1.0
** 50,0.,200.,1.0
** For QA use 20,0.,200.,1.0
*EL FILE, FREQ=0
*NODE FILE, FREQ=0
*EL PRINT, FREQ=0
*NODE PRINT, FREQ=0
*END STEP
************************************************
*STEP
3: FREQUENCY RESPONSE: STEADY STATE DYNAMICS,
SUBSPACE
SUBSPACE PROJECTION=ALL FREQUENCIES
*STEADY STATE DYNAMICS,
SUBSPACE PROJECTION=ALL FREQUENCIES,
INTERVAL=EIGENFREQUENCY,FREQUENCY SCALE=LINEAR
80,130,3
** 80,130,7,
** For QA use 80,130,3
*CLOAD
ROAD,3,200.
*EL PRINT, FREQ=0
*EL FILE, FREQ=0
*NODE PRINT,FREQ=0
*NODE FILE,NSET=SIDE,FREQ=1
U,
3-944
Tire Analyses
*NODE FILE,NSET=ROAD,FREQ=1
U,RF
*NODE OUTPUT,NSET=ROAD
U,RF
*NODE OUTPUT,NSET=SIDE
U,
*END STEP
*************************************************
*STEP
SUBSPACE
SUBSPACE PROJECTION=EIGENFREQUENCY
SUBSPACE PROJECTION=EIGENFREQUENCY,
80,130,3
** 80,130,7,
** For QA use 80,130,3
*CLOAD
ROAD,3,200.
*END STEP
**************************************************
*STEP
SUBSPACE
SUBSPACE PROJECTION=CONSTANT
SUBSPACE PROJECTION=CONSTANT,
80,130,3
** 80,130,7,
** For QA use 80,130,3
*CLOAD
ROAD,3,200.
*END STEP
**************************************************
*STEP
SUBSPACE
SUBSPACE PROJECTION=PROPERTY CHANGE
3-945
Tire Analyses
SUBSPACE PROJECTION=PROPERTY CHANGE,

80,130,3
** 80,130,7,
** For QA use 80,130,3
*CLOAD
ROAD,3,200.
*END STEP
**************************************************
*STEP,NLGEOM=YES
DIRECT
*STEADY STATE DYNAMICS,DIRECT,
80,130,3
** 80,130,7,
** For QA use 80,130,3
*CLOAD
ROAD,3,200.
*END STEP
3-946
Heat Transfer and Thermal-Stress Analyses
4. Heat Transfer and Thermal-Stress Analyses

4.1 Heat transfer and thermal-stress analyses
4.1.1 Thermally coupled analysis of a disc brake
Disc brakes operate by pressing a set of composite material brake pads against a rotating steel disc: the
frictional forces cause deceleration. The dissipation of the frictional heat generated is critical for
effective braking performance. Temperature changes of the brake cause axial and radial deformation;
and this change in shape, in turn, affects the contact between the pads and the disc. Thus, the system
should be analyzed as a fully coupled thermo-mechanical system.
In this example two thermally coupled disc brake analysis examples are discussed. The first example is
an axisymmetric model in which the brake pads and the frictional heat generated by braking are
"smeared" out over all 360° of the model. This problem is solved using only ABAQUS/Standard. The
heat generation is supplied by user subroutine FRIC, and the analysis models a linear decrease in
velocity as a result of braking.
The second example is a three-dimensional model of the entire disc with pads touching only part of the
circumference. The disc is rotated so that the heat is generated by friction. This problem is solved
using both ABAQUS/Standard and ABAQUS/Explicit.
It is also possible to perform uncoupled analysis of a brake system. The heat fluxes can be calculated
and applied to a thermal model, and then the resulting temperatures can be applied to a stress analysis.
However, since the thermal and stress analyses are uncoupled, this approach does not account for the
effect of the thermal deformation on the contact which, in turn, affects the heat generation.
Another type of geometrical model for a disc brake is used by Gonska and Kolbinger (1993). They
model a "vented" disc brake (Figure 4.1.1-1) and take advantage of radial repetition by modeling a
pie-slice segment (Figure 4.1.1-2). Like the axisymmetric model, this requires the effect of the pads to
be smeared, but it allows the modeling of radial cooling ducts while still reducing the model size
relative to a full model.
Geometry and model

Both models analyzed in this example have solid discs, which allows the models to use coarser meshes
than would be required to model the detail of a typical disc brake that has complicated geometrical
features such as cooling ducts and bolt holes. The first example further simplifies the model by
considering the pads to be "smeared" around the entire 360° so that the system is axisymmetric. The
second example is a full three-dimensional model of the entire annular disc with pads touching only
part of the circumference. However, the geometry of the disc has been simplified by making it
symmetrical about a plane normal to the axis. Therefore, only half of the disc and one brake pad is
modeled, and symmetry boundary conditions are applied.
The dimensions of the axisymmetric model are taken from a typical car disc brake. The disc has a
thicker friction ring connected to a conical section that, in turn, connects to an inner hub. The inner
4-947
radius of the friction ring is 100.0 mm, the outer radius is 135.0 mm, and it is 10.0 mm thick. The
conical section is 32.5 mm deep and 5.0 mm thick. The hub has an inner radius of 60.0 mm, an outer
radius of 80.0 mm, and is 5.0 mm thick. The pads are 20.0 mm thick and initially cover the entire
friction ring surfaces.
The pads and disc of the axisymmetric model are modeled with CAX4T elements. Frictional contact
between the pads and the disc is modeled by contact pairs between surfaces defined on the element
faces in the contact region. Small sliding is assumed. The mesh is shown in Figure 4.1.1-3, with the
pads drawn in a darker gray than the disc. There are six elements through the thickness of the friction
ring and four elements through the thickness of each of the pads. The mesh is somewhat coarse but is
optimized by using thinner elements near the surfaces of the disc and pads where contact occurs for
better resolution of the thermal gradients in these areas.
The disc for the three-dimensional model has an outer radius of 135.0 mm, an inner radius of 90.0 mm,
and a thickness of 10.0 mm (the half-model has a thickness of 5.0 mm). The ring has a thinner section
out to a radius of 100.0 mm, which has a thickness of 6.0 mm (the half-model has a thickness of 3.0
mm). The pad is 10.0 mm thick and covers a little less than one-tenth the circumference. The pad does
not quite reach to the edge of the thicker part of the friction ring.
The pad and disc of the three-dimensional model are modeled with C3D8T elements in
ABAQUS/Standard and with C3D8RT elements in ABAQUS/Explicit; the contact and friction
between the pad and the disc are modeled by contact pairs between surfaces defined on the element
faces in the contact region. The same mesh is used in both ABAQUS/Standard and ABAQUS/Explicit.
It is shown in Figure 4.1.1-4, with the pad drawn in a darker gray than the disc. The disc is a simple
annulus with a thinner inner ring. This mesh is also rather coarse with only three elements through the
thickness of the disc and three elements through the pad. The elements on the contact sides are thinner
since they will be in the areas of higher thermal gradients. There are 36 elements in the circumferential
direction of the disc.
Material properties
The thermal mechanical properties for the axisymmetric model were taken from a paper by Day and
Newcomb (1984) describing the analysis of an annular disc brake. The pad is made of a resin-bonded
composite friction material, and the disc is made of steel. Although Day and Newcomb note that
material changes occur in the pad material because of thermal degradation, the pad in the axisymmetric
model has the properties of the unused pad material. For the axisymmetric model the modulus, density,
conductivity, and coefficient of friction are divided by 18 since the pads actually cover only a 20°
section of the disc, even though they are modeled as being smeared around the entire circumference.
The pad for the three-dimensional model is also a resin-bonded composite friction material whose
thermal mechanical properties are listed in Table 4.1.1-1 and coefficient of friction is listed in Table
4.1.1-2. The properties were taken from a paper by Day (1984). It is noted that above certain
temperatures, approximately 400°C, the pad material becomes thermally degraded and ¹ is assumed
constant from this point on.
It is assumed that all the frictional energy is dissipated as heat and distributed equally between the disc
and the pad; therefore, the *GAP HEAT GENERATION option is set to 1.0, and the default
4-948
distribution is used. The *GAP HEAT GENERATION option allows the user to specify an unequal
distribution, which is particularly important if the heat conduction across the interface is poor. In this
example the conductivity value specified with the *GAP CONDUCTANCE option is quite high;
hence, the results are not very sensitive to changes in distribution. In ABAQUS/Explicit arbitrarily
high gap conductivity values may cause the stable time increment associated with the thermal part of
the problem to control the time incrementation, possibly resulting in a very inefficient analysis. In this
problem the gap conductivity value used in the ABAQUS/Explicit simulation is 20 times smaller than
the one used in the ABAQUS/Standard simulation. This allows the stable time increment associated
with the mechanical part of the problem to control the time incrementation, thus permitting a more
efficient solution while hardly affecting the results.
Loading
The pads of the axisymmetric model are first pressed against the disc. The magnitude of the load is
divided by 18 since the pads are not actually axisymmetric. The frictional forces are then applied
through user subroutine FRIC to simulate a linear decrease in velocity of the disc relative to the pads.
The braking is done over three steps; then, when the velocity is zero, a final step shows the continued
heat conduction through the model.
The pad of the three-dimensional model is fixed in the nonaxial degrees of freedom and is pressed
against the disc with a distributed load applied to the back of the pad. In ABAQUS/Standard the disc is
then rotated by 60° using an applied boundary condition to the center ring. In ABAQUS/Explicit this
boundary condition is prescribed using the *AMPLITUDE, TYPE=SMOOTH STEP option to
minimize the effects of centrifugal forces at the beginning and end of the step. Frictional forces
between the surfaces generate heat in the brake.
The initial temperature of both models is 20°C.
Solution controls (ABAQUS/Standard only)

Since the three-dimensional model has a small loaded area and, thus, rather localized forces and heat
fluxes, the default averaged flux values for the convergence criteria produce very tight tolerances and
cause more iteration than is necessary for an accurate solution. To decrease the computational time
required for the analysis, the *CONTROLS option is used to override the automatic calculation of the
average forces and heat fluxes. The option is first used with the FIELD=DISPLACEMENT parameter.
The convergence criterion ratio is set to 1%, and the time-average and average fluxes are set to a
typical nodal force (displacement flux):
fd = pA = (1:7E6)(1:77E-4) ¼ 300;
where p is the pressure and A is the area of a typical pad element. The option is next used with the
FIELD=TEMPERATURE parameter. The convergence criterion ratio is set to 1%, and the
time-average and average fluxes are set to the nodal heat flux (temperature flux) for a typical pad
element. The heat flux density generated by an interface element due to frictional heat generation is
qg = ´¿ v, where ´ is the gap heat generation factor, ¿ is the frictional stress, and v is the velocity.
Therefore, the nodal heat flux is
4-949
ft = qg A = ´ (¹p)(!r )A;
where A is the contact area of a typical pad element, ¹ is the friction coefficient, and p is the contact
pressure. The angular velocity, !, is obtained as the total rotation, ¼=3, divided by the total time, 0.015
sec. The radius, r, is set to 0.120 m, which is the distance from the axis to a point approximately in the
middle of the pad surface. This yields
(1:0)(:37)(1:7E6)( ¼3 )(0:12)(1:77E-4)
ft = ¼ 900:
0:015
Additional solution controls can reduce the solver cost for an increment by improving the initial
solution guess, solving thermal and mechanical equations separately, and reducing the wavefront of
three-dimensional finite-sliding contact analysis. These features are discussed below. The impact of
combining these features is also discussed.
When the default convergence controls are used, it is possible to obtain faster convergence by setting
the EXTRAPOLATION parameter on the *STEP option to PARABOLIC. For the three-dimensional
model the use of this feature yields a 14% enhancement in computational speed per increment.
The coupling between the thermal and mechanical fields in this problem is relatively weak. It is,
therefore, possible to obtain a more efficient solution by specifying separate solutions for the thermal
and mechanical equations each increment. This technique, which is specified by using the
*SOLUTION TECHNIQUE, TYPE=SEPARATED option, results in faster per-iteration solution times
at the expense of poorer convergence when a strong interfield coupling is present. Use of this
technique also permits the use of the symmetric solver and storage scheme, which is invoked by
specifying UNSYMM=NO on the *STEP option. The resulting symmetric approximation of the
mechanical equations was also found to be cost effective for this problem, when combined with a
quality initial solution guess obtained by specifying EXTRAPOLATION=PARABOLIC on the *STEP
option. Neither of these approximations impacts solution accuracy. For the three-dimensional model
the use of *SOLUTION TECHNIQUE, TYPE=SEPARATED and *STEP, UNSYMM=NO,
EXTRAPOLATION=PARABOLIC yields a 50% decrease in the total solution time.
In the three-dimensional model the deformable master surface is defined from a large number of
connecting elements resulting in a large wavefront. By default, ABAQUS/Standard employs an
automated contact patch algorithm to reduce the wavefront and solution time. For instance, in the
coupled thermal-mechanical analysis a substantial savings in solution time (a 30% to 50% decrease) is
obtained when the automatic contact patch algorithm is employed compared to an analysis that uses a
fixed contact patch encompassing the entire master surface. The reduction in solution time is system
dependent and depends on several factors, such as CPU type, system memory, and IO speed. This
solution time savings is in addition to any of the other savings discussed in this section. The additional
savings is, therefore, realized when the separated solution scheme and parabolic extrapolation are also
specified.
4-950
The temperature distribution of the axisymmetric model at an early time increment is shown in Figure
4.1.1-5. The temperature is greatest at the interfaces between the disc and pads, and the heat has just
started to conduct into the disc. Figure 4.1.1-6 shows the temperature distribution at the end of the
analysis when the velocity is zero. The heat has conducted through the friction ring of the disc. Figure
4.1.1-7 is a displaced plot of the model at the end of the analysis and shows the characteristic conical
deformation due to thermal expansion. The displacement has been magnified by a factor of 128 to
show the deformation more clearly.
The temperature distribution of the disc surface of the three-dimensional model after a rotation of 60°
is shown in Figure 4.1.1-8 (ABAQUS/Standard) and Figure 4.1.1-9 (ABAQUS/Explicit). The
agreement between the two results is excellent. The hottest region is the area under the pad, while the
heat in the regions that the pad has passed over has dissipated somewhat. Figure 4.1.1-10 shows the
temperature distribution of the inside of the brake pad predicted by ABAQUS/Standard, while Figure
4.1.1-11 shows the same result obtained with ABAQUS/Explicit. Again excellent agreement between
the two results is noted. Figure 4.1.1-12 shows the temperature distribution in the disc predicted by
ABAQUS/Standard with the thickness magnified by a factor of 20. The heat has conducted into the
disc in the regions that the pad has passed over.
The stresses predicted by ABAQUS/Standard do not account for the effects of centrifugal loads
(*COUPLED TEMPERATURE-DISPLACEMENT is a quasi-static procedure), while the stresses
predicted by ABAQUS/Explicit do. These effects can be significant, especially during the early
transient portion of the simulation when the initially stationary disc is brought up to speed. To compare
the stress results between ABAQUS/Standard and ABAQUS/Explicit, we gradually initiated and ended
the disc rotation in the ABAQUS/Explicit simulation; thus, in ABAQUS/Explicit, the centrifugal
stresses at the beginning and end of the step are small compared with the thermal stresses. At points in
between, however, the effects of centrifugal loading are more pronounced and differences between the
stress states predicted by ABAQUS/Standard and ABAQUS/Explicit are observed. The overall effect
on the thermal response, however, is negligible.
The ABAQUS/Explicit analysis did not include mass scaling because its presence would artificially
scale the stresses due to the centrifugal loads. It is possible to include mass scaling to make the
analysis more economical, but any results obtained with mass scaling must be interpreted carefully in
this problem.
Input files

discbrake_axi.inp
Axisymmetric model.
discbrake_axi.f
User subroutine FRIC used in discbrake_axi.inp.
discbrake_3d.inp
Three-dimensional model.
4-951
discbrake_postoutput.inp
*POST OUTPUT analysis of the three-dimensional model.
discbrake_3d_extrapara.inp
Three-dimensional model with the second step run with *STEP,
EXTRAPOLATION=PARABOLIC and with the default *CONTROLS option.
discbrake_3d_extrapara_300c.inp
Three-dimensional model with the second step run with *STEP,
EXTRAPOLATION=PARABOLIC. It is assumed that several revolutions occurred and the initial
temperature for the disc brake and pad is 300°C.
discbrake_3d_separated.inp
Three-dimensional model run using the *SOLUTION TECHNIQUE, TYPE=SEPARATED option.
ABAQUS/Explicit input file

discbrake_3d_xpl.inp
Three-dimensional model.
References
· Day, A. J., "An Analysis of Speed, Temperature, and Performance Characteristics of Automotive
Drum Brakes," Journal of Tribology, vol. 110, pp. 295-305, 1988.
· Day, A. J., and T. J. Newcomb, "The Dissipation of Frictional Energy from the Interface of an
Annular Disc Brake," Proc. Instn. Mech. Engrs, vol. 198D, no. 11, pp. 201-209, 1984.
· Gonska, H. W., and H. J. Kolbinger, "ABAQUS Application Example: Temperature and

Deformation Calculation of Passenger Car Brake Disks," ABAQUS Users' Conference
Proceedings, 1993.
Tables
Table 4.1.1-1 Thermal-mechanical properties.

Temperature of property measurement 20 100 200 300
(°C)
Young's modulus, E (N/mm2) 2200 1300 530 320
Poisson's ratio, º 0.25 0.25 0.25 0.25
Density, (kg/m )
½ 3 1550 1550 1550 1550
Thermal expansion coefficient (K -1) 10e-6 - 30e-6 -
Thermal conductivity, ∙ (w/mK) 0.5 0.5 0.5 0.5
Specific heat, Cp (J/kgK) 1200 1200 1200 1200
4-952
Table 4.1.1-2 Brake lining temperature characteristic.

Temperature of property measurement 100 200 300 400
(°C)
Friction coefficient, ¹ 0.38 0.41 0.42 0.24
Figures
Figure 4.1.1-1 A vented brake disc design.
Figure 4.1.1-2 Modeling a segment of a brake disc.
Figure 4.1.1-3 Mesh for the axisymmetric model, ABAQUS/Standard.
4-953
Figure 4.1.1-4 Mesh for the three-dimensional model.
Figure 4.1.1-5 Isotherms of the axisymmetric model at t =0.675, ABAQUS/Standard.
4-954
Figure 4.1.1-6 Isotherms of the axisymmetric model when braking has ended, ABAQUS/Standard.
Figure 4.1.1-7 Deformation of the axisymmetric disc, displacement magnified by 128,

ABAQUS/Standard.
Figure 4.1.1-8 Isotherms of the disc surface, ABAQUS/Standard.
4-955
Figure 4.1.1-9 Isotherms of the disc surface, ABAQUS/Explicit.
Figure 4.1.1-10 Isotherms of the inside of the brake pad, ABAQUS/Standard.
4-956
Figure 4.1.1-11 Isotherms of the inside of the brake pad, ABAQUS/Explicit.
Figure 4.1.1-12 Isotherms of the disc with the thickness magnified 20 times, ABAQUS/Standard.
4-957
Sample listings
4-958
Listing 4.1.1-1
*HEADING
AXISYMMETRIC DISC BRAKE ANALYSIS
*NODE,NSET=CENTER
1, 60.E-3, 30.E-3
2, 60.E-3, 27.5E-3
3, 60.E-3, 25.E-3
*NODE
801, 80.E-3, 30.E-3
802, 80.E-3, 27.5E-3
803, 80.E-3, 25.E-3
611, 85.E-3, 2.5E-3
612, 85.E-3, 0.E-3
613, 85.E-3, -2.5E-3
1011, 95.E-3, 2.5E-3
1012, 95.E-3, 0.E-3
1013, 95.E-3, -2.5E-3
*NODE,NSET=INNER
1204,100.E-3, 15.E-3
1205,100.E-3, 10.E-3
1206,100.E-3, 7.E-3
1207,100.E-3, 5.5E-3
1208,100.E-3, 5.E-3
**
1209,100.E-3, 5.E-3
1210,100.E-3, 4.5E-3
1211,100.E-3, 3.E-3
1212,100.E-3, 0.E-3
1213,100.E-3, -3.E-3
1214,100.E-3, -4.5E-3
1215,100.E-3, -5.E-3
**
1216,100.E-3, -5.E-3
1217,100.E-3, -5.5E-3
1218,100.E-3, -7.E-3
1219,100.E-3,-10.E-3
1220,100.E-3,-15.E-3
*NODE,NSET=OUTER
2404,135.E-3, 15.E-3
2405,135.E-3, 10.E-3
2406,135.E-3, 7.E-3
2407,135.E-3, 5.5E-3
4-959
2408,135.E-3, 5.E-3
**
2409,135.E-3, 5.E-3
2410,135.E-3, 4.5E-3
2411,135.E-3, 3.E-3
2412,135.E-3, 0.E-3
2413,135.E-3, -3.E-3
2414,135.E-3, -4.5E-3
2415,135.E-3, -5.E-3
**
2416,135.E-3, -5.E-3
2417,135.E-3, -5.5E-3
2418,135.E-3, -7.E-3
2419,135.E-3,-10.E-3
2420,135.E-3,-15.E-3
*NGEN,NSET=ALL
1, 801,100
2, 802,100
3, 803,100
611,1011,100
612,1012,100
613,1013,100
1011,1211,100
1012,1212,100
1013,1213,100
603, 611, 1
703, 711, 1
803, 811, 1
*NFILL,NSET=ALL
INNER, OUTER, 12, 100
*NSET, NSET=PADS, GENERATE
1204, 2404, 100
1205, 2405, 100
1206, 2406, 100
1207, 2407, 100
1208, 2408, 100
1216, 2416, 100
1217, 2417, 100
1218, 2418, 100
1219, 2419, 100
1220, 2420, 100
*NSET, NSET=PADBACK, GENERATE
1204, 2404, 100
4-960
1220, 2420, 100

*ELEMENT, TYPE=CAX4T
1, 1, 2, 102, 101
1209,1209,1210,1310,1309
*ELGEN,ELSET=DISK
1, 2,1,1, 8,100,100
601,12,1,1, 2,100,100
611, 2,1,1, 6,100,100
1209, 6,1,1,12,100,100
*ELSET,ELSET=SURF1
611, 612, 1209, 1210, 1213, 1214
*ELSET,ELSET=SURF2, GENERATE
612, 1112, 100
1214, 2314, 100
701, 710, 1
2309, 2314, 1
1, 701, 100
811, 1111, 100
1209, 2309, 100
*ELEMENT, TYPE=CAX4T
1204,1204,1205,1305,1304
1216,1216,1217,1317,1316
*ELGEN,ELSET=PAD1
1204, 4,1,1,12,100,100
*ELGEN,ELSET=PAD2
1216, 4,1,1,12,100,100
*ELSET,ELSET=PADS
PAD1,PAD2
*ELSET,ELSET=PADBACK1,GENERATE
1204, 2304, 100
*ELSET,ELSET=PADBACK2,GENERATE
1219, 2319, 100
*ELSET,ELSET=FRIC1, GENERATE
1209, 2309,100
*ELSET, ELSET=FRIC2,GENERATE
1214,2314,100
*ELSET, ELSET=PAD1EL,GENERATE
1207,2307,100
*ELSET, ELSET=PAD2EL,GENERATE
1216,2316,100
*SOLID SECTION,ELSET=DISK,MATERIAL=STEEL
4-961
*ELASTIC
210.E9, .3
*EXPANSION
11.E-6,
*DENSITY
7800.,
*SPECIFIC HEAT
452.,
*CONDUCTIVITY
48.,
*SOLID SECTION,ELSET=PADS,MATERIAL=PAD
*MATERIAL, NAME=PAD
** THE MODULUS, DENSITY AND CONDUCTIVITY HAVE BEEN
** DIVIDED BY 18 TO ACCOUNT FOR THE FACT THAT THE
** PADS ARE NOT AXISYMMETRIC
*ELASTIC
16.66667E6, .25
*EXPANSION
57.E-6,
*DENSITY
125.,
*SPECIFIC HEAT
1000.,
*CONDUCTIVITY
0.05,
*surface, NAME=RING1
FRIC1,S4
*surface, NAME=RING2
FRIC2,S2
*surface, NAME=PAD1SURF
PAD1EL,S2
*surface, NAME=PAD2SURF
PAD2EL,S4
*SURFACE INTERACTION, NAME=INT
*FRICTION,USER,PROPERTIES=1
0.37,
*GAP CONDUCTANCE
1.0E9, 0.
1.0E9, 1.
1.0,0.5
*CONTACT PAIR, INTERACTION=INT,SMALL SLIDING
4-962
**PAD1SURF,RING1
**PAD2SURF,RING2
RING1,PAD1SURF
RING2,PAD2SURF
ALL,20.
*****
**
*STEP, AMPLITUDE=RAMP
START BRAKING: VELOCITY DECREASES LINEARLY OVER
TIME IN 4.5 SECS
0.005, 0.01
*BOUNDARY
CENTER, 1, 2
PADBACK, 1
** PRESSURE LOAD DIVIDED BY 18 AS WELL
*DLOAD
PADBACK1, P4, 0.09444444E6
PADBACK2, P2, 0.09444444E6
*FILM
SURF1, F1, 20., 100.
SURF2, F2, 20., 100.
SURF3, F3, 20., 100.
SURF4, F4, 20., 100.
*PRINT, CONTACT=YES
S,
E,
*NODE PRINT, NSET=PADS, FREQUENCY=100
U1, U2, NT11, RF1, RF2
*NODE FILE, NSET=PADS, FREQUENCY=100
U, NT, RF
*END STEP
**
*STEP, INC=100, AMPLITUDE=RAMP
CONTINUE BRAKING: 0.01 TO 0.04 SECS
0.01, 0.04
*END STEP
**
4-963
CONTINUE BRAKING: 0.04 TO 4.5 SECS

0.1, 4.0, , .4
*END STEP
**
HEAT CONDUCTION WITH ZERO VELOCITY
1., 20.0, , 4.
*END STEP
4-964
Listing 4.1.1-2
*HEADING
THERMALLY COUPLED ANALYSIS OF A SYMMETRIC
DISC BRAKE
*NODE
1, 90.0000000E-3, 0.00000000E-3, 0.0E-3
36, 88.6326978E-3, -15.62833599E-3, 0.0E-3
101, 100.0000000E-3, 0.00000000E-3, 0.0E-3
136, 98.4807753E-3, -17.36481777E-3, 0.0E-3
401, 135.0000000E-3, 0.00000000E-3, 0.0E-3
436, 132.9490467E-3, -23.44250399E-3, 0.0E-3
1001, 90.0000000E-3, 0.00000000E-3, 3.0E-3
1036, 88.6326978E-3, -15.62833599E-3, 3.0E-3
1101, 100.0000000E-3, 0.00000000E-3, 3.0E-3
1136, 98.4807753E-3, -17.36481777E-3, 3.0E-3
1401, 135.0000000E-3, 0.00000000E-3, 3.0E-3
1436, 132.9490467E-3, -23.44250399E-3, 3.0E-3
2101, 100.0000000E-3, 0.00000000E-3, 4.5E-3
2136, 98.4807753E-3, -17.36481777E-3, 4.5E-3
2401, 135.0000000E-3, 0.00000000E-3, 4.5E-3
2436, 132.9490467E-3, -23.44250399E-3, 4.5E-3
3101, 100.0000000E-3, 0.00000000E-3, 5.0E-3
3136, 98.4807753E-3, -17.36481777E-3, 5.0E-3
3401, 135.0000000E-3, 0.00000000E-3, 5.0E-3
3436, 132.9490467E-3, -23.44250399E-3, 5.0E-3
*NODE, NSET=CENTER
1000, 0., 0., 0.,
*NGEN, LINE=C, NSET=SUPPORT
1, 36, 1, , 0., 0., 0.0E-3, 0., 0., 100.E-3
1001, 1036, 1, , 0., 0., 3.0E-3, 0., 0., 100.E-3
*NGEN, LINE=C, NSET=INNER1
101, 136, 1, , 0., 0., 0.0E-3, 0., 0., 100.E-3
1101, 1136, 1, , 0., 0., 3.0E-3, 0., 0., 100.E-3
2101, 2136, 1, , 0., 0., 4.5E-3, 0., 0., 100.E-3
3101, 3136, 1, , 0., 0., 5.0E-3, 0., 0., 100.E-3
*NSET, NSET=INNER
INNER1, INNER2
*NGEN, LINE=C, NSET=OUTER
401, 436, 1, , 0., 0., 0.0E-3, 0., 0., 100.E-3
1401, 1436, 1, , 0., 0., 3.0E-3, 0., 0., 100.E-3
2401, 2436, 1, , 0., 0., 4.5E-3, 0., 0., 100.E-3
4-965
3401, 3436, 1, , 0., 0., 5.0E-3, 0., 0., 100.E-3

*NFILL, NSET=DISK
*NSET, NSET=SYMM, GENERATE
101, 436
*ELEMENT, TYPE=C3D8T, ELSET=SUPPORT
1, 1, 101, 102, 2, 1001, 1101, 1102, 1002
36, 36, 136, 101, 1, 1036, 1136, 1101, 1001
*ELEMENT, TYPE=C3D8T, ELSET=DISK
101, 101, 201, 202, 102, 1101, 1201, 1202, 1102
136, 136, 236, 201, 101, 1136, 1236, 1201, 1101
*ELGEN, ELSET=SUPPORT
1, 35, 1, 1
*ELGEN, ELSET=DISK
101, 35, 1, 1, 3, 100, 100, 3, 1000, 1000
136, 1, 1, 1, 3, 100, 100, 3, 1000, 1000
*ELSET, ELSET=FULLDISK
SUPPORT, DISK
*ELSET, ELSET=DISKSURF, GENERATE
2101, 2436
*SOLID SECTION, ELSET=FULLDISK, MATERIAL=STEEL,
ORIENTATION=CYL
*ELASTIC
209.E9, .3
*EXPANSION
11.E-6,
*DENSITY
7800.,
*SPECIFIC HEAT
452.,
*CONDUCTIVITY
48.,
*ORIENTATION, SYSTEM=CYLINDRICAL, NAME=CYL
0., 0., 0., 0., 0., 1.
3, 0.
*MPC
BEAM, SUPPORT, 1000
*NODE
10001, 29.E-3, 97.26895702E-3, 5.0E-3
10005, -29.E-3, 97.26895702E-3, 5.0E-3
10301, 38.E-3, 127.4558747E-3, 5.0E-3
10305, -38.E-3, 127.4558747E-3, 5.0E-3
4-966
11001, 29.E-3, 97.26895702E-3, 5.5E-3

11005, -29.E-3, 97.26895702E-3, 5.5E-3
11301, 38.E-3, 127.4558747E-3, 5.5E-3
11305, -38.E-3, 127.4558747E-3, 5.5E-3
12001, 29.E-3, 97.26895702E-3, 7.0E-3
12005, -29.E-3, 97.26895702E-3, 7.0E-3
12301, 38.E-3, 127.4558747E-3, 7.0E-3
12305, -38.E-3, 127.4558747E-3, 7.0E-3
13001, 29.E-3, 97.26895702E-3, 10.0E-3
13005, -29.E-3, 97.26895702E-3, 10.0E-3
13301, 38.E-3, 127.4558747E-3, 10.0E-3
13305, -38.E-3, 127.4558747E-3, 10.0E-3
14001, 29.E-3, 97.26895702E-3, 15.0E-3
14005, -29.E-3, 97.26895702E-3, 15.0E-3
14301, 38.E-3, 127.4558747E-3, 15.0E-3
14305, -38.E-3, 127.4558747E-3, 15.0E-3
*NGEN, LINE=C, NSET=INPAD
10001,10005,1,,0.,0.,5.0E-3,0.,0.,100.E-3
11001,11005,1,,0.,0.,5.5E-3,0.,0.,100.E-3
12001,12005,1,,0.,0.,7.0E-3,0.,0.,100.E-3
13001,13005,1,,0.,0.,10.0E-3,0.,0.,100.E-3
14001,14005,1,,0.,0.,15.0E-3,0.,0.,100.E-3
*NGEN, LINE=C, NSET=OUTPAD
10301,10305,1,,0.,0.,5.0E-3,0.,0.,100.E-3
11301,11305,1,,0.,0.,5.5E-3,0.,0.,100.E-3
12301,12305,1,,0.,0.,7.0E-3,0.,0.,100.E-3
13301,13305,1,,0.,0.,10.0E-3,0.,0.,100.E-3
14301,14305,1,,0.,0.,15.0E-3,0.,0.,100.E-3
*NFILL, NSET=NPAD
INPAD, OUTPAD, 3, 100
*NSET, NSET=NPADBACK, GENERATE
14001, 14305
*NSET, NSET=ALL
DISK, NPAD, SUPPORT
*ELEMENT, TYPE=C3D8T
10001, 10001, 10101, 10102, 10002, 11001, 11101,
11102, 11002
*ELGEN, ELSET=EPAD
10001, 4, 1, 1, 3, 100, 100, 4, 1000, 1000
*ELSET, ELSET=EPADBACK, GENERATE
13001, 13204
*SOLID SECTION, ELSET=EPAD, MATERIAL=PADMAT,
ORIENTATION=CYL
4-967
*MATERIAL, NAME=PADMAT
*ELASTIC
2200.E6, .25, 20.0
1300.E6, .25, 100.0
530.E6, .25, 200.0
320.E6, .25, 300.0
*EXPANSION
10.E-6, 20.0
30.E-6,200.0
*DENSITY
1550.,
*SPECIFIC HEAT
1200.,
*CONDUCTIVITY
0.9,
**
** DEFINE SURFACES AND CONTACT PAIR
**
*SURFACE, NAME=MASTER
DISKSURF, S2
*SURFACE, NAME=SLAVE
10001, S1
10002, S1
10003, S1
10004, S1
10101, S1
10102, S1
10103, S1
10104, S1
10201, S1
10202, S1
10203, S1
10204, S1
*CONTACT PAIR, INTERACTION=INTERF
SLAVE,MASTER
*SURFACE INTERACTION, NAME=INTERF
*FRICTION
0.37,,,20.0
0.38,,,100.0
0.41,,,200.0
0.39,,,300.0
0.24,,,400.0
*GAP CONDUCTANCE
4-968
1.0E9, 0.
1.0E9, 1.
1.0,
ALL,20.
*****
*STEP,NLGEOM,AMPLITUDE=RAMP,UNSYMM=YES
PRESS THE PAD AGAINST THE DISK
0.0005, 0.001
*BOUNDARY
SYMM, 3, 3
CENTER, 1, 6
NPADBACK, 1, 2
*DLOAD
EPADBACK, P2, 1.7E6
*FILM
DISKSURF, F2, 20., 100.
**
*PRINT, CONTACT=YES
*CONTACT FILE, FREQUENCY=5
S,
E,
*NODE PRINT, NSET=CENTER, FREQUENCY=1
UR3,RF
*NODE PRINT, NSET=NPAD, FREQUENCY=100
U1, U2, U3, NT11, RF1, RF2, RF3
*NODE FILE, NSET=CENTER, FREQUENCY=1
U, RF
*NODE FILE, NSET=NPAD, FREQUENCY=100
U, NT, RF
*END STEP
*****
*STEP, NLGEOM,INC=100,AMPLITUDE=RAMP,UNSYMM=YES
ROTATE THE DISK BY 60 DEGREES
*COUPLED TEMPERATURE-DISPLACEMENT, DELTMX=100.
0.001, 0.015, , .004
*CONTROLS,PARAMETERS=FIELD,FIELD=DISPLACEMENT
0.01,,300.,300.
4-969
*CONTROLS,PARAMETERS=FIELD,FIELD=TEMPERATURE
0.01,,900.,900.
*BOUNDARY
CENTER, 6, 6, 1.04717
*END STEP
4-970
Listing 4.1.1-3
*HEADING
THERMALLY COUPLED ANALYSIS OF A SYMMETRIC DISC
BRAKE (EXPLICIT)
RIGID BODY (PIN NSET)
*NODE
1, 90.0000000E-3, 0.00000000E-3, 0.0E-3
36, 88.6326978E-3, -15.62833599E-3, 0.0E-3
101, 100.0000000E-3, 0.00000000E-3, 0.0E-3
136, 98.4807753E-3, -17.36481777E-3, 0.0E-3
401, 135.0000000E-3, 0.00000000E-3, 0.0E-3
436, 132.9490467E-3, -23.44250399E-3, 0.0E-3
1001, 90.0000000E-3, 0.00000000E-3, 3.0E-3
1036, 88.6326978E-3, -15.62833599E-3, 3.0E-3
1101, 100.0000000E-3, 0.00000000E-3, 3.0E-3
1136, 98.4807753E-3, -17.36481777E-3, 3.0E-3
1401, 135.0000000E-3, 0.00000000E-3, 3.0E-3
1436, 132.9490467E-3, -23.44250399E-3, 3.0E-3
2101, 100.0000000E-3, 0.00000000E-3, 4.5E-3
2136, 98.4807753E-3, -17.36481777E-3, 4.5E-3
2401, 135.0000000E-3, 0.00000000E-3, 4.5E-3
2436, 132.9490467E-3, -23.44250399E-3, 4.5E-3
3101, 100.0000000E-3, 0.00000000E-3, 5.0E-3
3136, 98.4807753E-3, -17.36481777E-3, 5.0E-3
3401, 135.0000000E-3, 0.00000000E-3, 5.0E-3
3436, 132.9490467E-3, -23.44250399E-3, 5.0E-3
*NODE, NSET=CENTER
1000, 0., 0., 0.,
*NGEN, LINE=C, NSET=SUPPORT
1, 36, 1, , 0., 0., 0.0E-3, 0., 0., 100.E-3
1001, 1036, 1, , 0., 0., 3.0E-3, 0., 0., 100.E-3
101, 136, 1, , 0., 0., 0.0E-3, 0., 0., 100.E-3
1101, 1136, 1, , 0., 0., 3.0E-3, 0., 0., 100.E-3
2101, 2136, 1, , 0., 0., 4.5E-3, 0., 0., 100.E-3
3101, 3136, 1, , 0., 0., 5.0E-3, 0., 0., 100.E-3
*NSET, NSET=INNER
INNER1, INNER2
*NGEN, LINE=C, NSET=OUTER
401, 436, 1, , 0., 0., 0.0E-3, 0., 0., 100.E-3
1401, 1436, 1, , 0., 0., 3.0E-3, 0., 0., 100.E-3
4-971
2401, 2436, 1, , 0., 0., 4.5E-3, 0., 0., 100.E-3

3401, 3436, 1, , 0., 0., 5.0E-3, 0., 0., 100.E-3
*NFILL, NSET=DISK
*NSET, NSET=SYMM, GENERATE
101, 436
*ELEMENT, TYPE=C3D8RT, ELSET=SUPPORT
1, 1, 101, 102, 2, 1001, 1101, 1102, 1002
36, 36, 136, 101, 1, 1036, 1136, 1101, 1001
*ELEMENT, TYPE=C3D8RT, ELSET=DISK
101, 101, 201, 202, 102, 1101, 1201, 1202, 1102
136, 136, 236, 201, 101, 1136, 1236, 1201, 1101
*ELGEN, ELSET=SUPPORT
1, 35, 1, 1
*ELGEN, ELSET=DISK
101, 35, 1, 1, 3, 100, 100, 3, 1000, 1000
136, 1, 1, 1, 3, 100, 100, 3, 1000, 1000
*ELSET, ELSET=FULLDISK
SUPPORT, DISK
*ELSET, ELSET=DISKSURF, GENERATE
2101, 2436
*SOLID SECTION,ELSET=FULLDISK,MATERIAL=STEEL,
ORIENTATION=CYL
*ELASTIC
209.E9, .3
*EXPANSION
11.E-6,
*DENSITY
7800.,
*SPECIFIC HEAT
452.,
*CONDUCTIVITY
48.,
*ORIENTATION, SYSTEM=CYLINDRICAL, NAME=CYL
0., 0., 0., 0., 0., 1.
3, 0.
**
*RIGID BODY, REF NODE=1000, PIN NSET=SUPPORT
**
*NODE
10001, 29.E-3, 97.26895702E-3, 5.0E-3
10005, -29.E-3, 97.26895702E-3, 5.0E-3
4-972
10301, 38.E-3, 127.4558747E-3, 5.0E-3

10305, -38.E-3, 127.4558747E-3, 5.0E-3
11001, 29.E-3, 97.26895702E-3, 5.5E-3
11005, -29.E-3, 97.26895702E-3, 5.5E-3
11301, 38.E-3, 127.4558747E-3, 5.5E-3
11305, -38.E-3, 127.4558747E-3, 5.5E-3
12001, 29.E-3, 97.26895702E-3, 7.0E-3
12005, -29.E-3, 97.26895702E-3, 7.0E-3
12301, 38.E-3, 127.4558747E-3, 7.0E-3
12305, -38.E-3, 127.4558747E-3, 7.0E-3
13001, 29.E-3, 97.26895702E-3, 10.0E-3
13005, -29.E-3, 97.26895702E-3, 10.0E-3
13301, 38.E-3, 127.4558747E-3, 10.0E-3
13305, -38.E-3, 127.4558747E-3, 10.0E-3
14001, 29.E-3, 97.26895702E-3, 15.0E-3
14005, -29.E-3, 97.26895702E-3, 15.0E-3
14301, 38.E-3, 127.4558747E-3, 15.0E-3
14305, -38.E-3, 127.4558747E-3, 15.0E-3
*NGEN, LINE=C, NSET=INPAD
10001,10005,1,,0.,0.,5.0E-3,0.,0.,100.E-3
11001,11005,1,,0.,0.,5.5E-3,0.,0.,100.E-3
12001,12005,1,,0.,0.,7.0E-3,0.,0.,100.E-3
13001,13005,1,,0.,0.,10.0E-3,0.,0.,100.E-3
14001,14005,1,,0.,0.,15.0E-3,0.,0.,100.E-3
*NGEN, LINE=C, NSET=OUTPAD
10301,10305,1,,0.,0.,5.0E-3,0.,0.,100.E-3
11301,11305,1,,0.,0.,5.5E-3,0.,0.,100.E-3
12301,12305,1,,0.,0.,7.0E-3,0.,0.,100.E-3
13301,13305,1,,0.,0.,10.0E-3,0.,0.,100.E-3
14301,14305,1,,0.,0.,15.0E-3,0.,0.,100.E-3
*NFILL, NSET=NPAD
INPAD, OUTPAD, 3, 100
*NSET, NSET=NPADBACK, GENERATE
14001, 14305
*NSET, NSET=ALL
DISK, NPAD, SUPPORT
*ELEMENT, TYPE=C3D8RT
10001, 10001, 10101, 10102, 10002, 11001, 11101,
11102, 11002
*ELGEN, ELSET=EPAD
10001, 4, 1, 1, 3, 100, 100, 4, 1000, 1000
*ELSET, ELSET=EPADBACK, GENERATE
13001, 13204
4-973
*SOLID SECTION, ELSET=EPAD, MATERIAL=PADMAT,

ORIENTATION=CYL
*MATERIAL, NAME=PADMAT
*ELASTIC
2200.E6, .25, 20.0
1300.E6, .25, 100.0
530.E6, .25, 200.0
320.E6, .25, 300.0
*EXPANSION
10.E-6, 20.0
30.E-6,200.0
*DENSITY
1550.,
*SPECIFIC HEAT
1200.,
*CONDUCTIVITY
0.9,
ALL,20.
*ELSET,ELSET=EALL
EPAD,FULLDISK
**
** DEFINE SURFACES
**
*SURFACE, NAME=MASTER
DISKSURF, S2
*SURFACE, NAME=SLAVE
10001, S1
10002, S1
10003, S1
10004, S1
10101, S1
10102, S1
10103, S1
10104, S1
10201, S1
10202, S1
10203, S1
10204, S1
**
*STEP
PRESS THE PAD AGAINST THE DISK
4-974
, 0.001
*BOUNDARY
SYMM, 3, 3
CENTER, 1, 6
NPADBACK, 1, 2
**
** DEFINE CONTACT PAIR
**
*CONTACT PAIR, INTERACTION=INTERF,WEIGHT=0.0
SLAVE,MASTER
*SURFACE INTERACTION, NAME=INTERF
*FRICTION
0.37,,,20.0
0.38,,,100.0
0.41,,,200.0
0.39,,,300.0
0.24,,,400.0
*GAP CONDUCTANCE, PRESSURE
5.0E7,0.0
5.0E7,1000.0
1.0,
**
0.0,0.0,0.001,1.0
*DLOAD,AMP=RAMP
EPADBACK, P2, 1.7E6
*FILM,AMP=RAMP
DISKSURF, F2, 20., 100.
**
*FILE OUTPUT,NUM=1
*NODE FILE
NT,
*EL FILE
S,
*END STEP
*****
*STEP
ROTATE THE DISK BY 60 DEGREES
, 0.015
*AMPLITUDE,NAME=RAMP2,DEFINITION=SMOOTH STEP
0.,0.,0.015,1.0
4-975
*BOUNDARY, TYPE=DISPLACEMENT,AMP=RAMP2
CENTER, 6, 6, 1.04717
*END STEP
4.1.2 Exhaust manifold assemblage

Engine exhaust manifolds are commonly subject to severe thermal cycles during operation and upon
shutdown. Thermal expansion and contraction of the manifold is constrained by its interaction with the
engine head to which it is bolted. These constraints govern the thermo-mechanical fatigue life of the
manifold.
The initial assembly procedure consists of bolting the flanges of the manifold to the engine head with
prescribed bolt forces that produce uniform axial bolt stresses. Under subsequent operating conditions
such as thermal cycling and creep, these bolt forces may increase or relax, possibly changing normal
pressures and resulting in lateral slippage between the engine head and the manifold flanges. Thus, the
boundary constraints on the manifold flanges are a function of the response of the entire assembly to
its operating conditions. As such, these boundary constraints cannot be prescribed a priori. This
example shows how to simulate these varying boundary constraints with the prescribed assembly load
capability of ABAQUS.
The problem scenario consists of three steps:
1. Apply prescribed bolt loads to fasten the exhaust manifold to the engine head.
2. Subject the assembly to the steady-state operating temperature distribution.
3. Return the assembly to ambient temperature conditions.
Geometry and model

The exhaust manifold assemblage being analyzed is depicted in Figure 4.1.2-1. It consists of a four
tube exhaust manifold with three flanges, bolted with seven bolts to a small section of the engine
head.
The manifold is cast from gray iron with a Young's modulus of 138 GPa, a Poisson's ratio of 0.283,
and a coefficient of thermal expansion of 13.8 ´ 10-6 per °C. In this example the region of the manifold
where the hot exhaust gases converge is subject to temperatures ranging from an initial value of 300 K
to an extreme of 980 K. The elastic-plastic response of gray cast iron varies greatly over this range of
temperatures, so the temperature-dependent plasticity curves shown in Figure 4.1.2-2 are used for the
manifold material. Gray cast iron exhibits different behavior in tension and compression; therefore,
these curves represent the average response. The Mises metal plasticity model with isotropic hardening
is used. The three manifold flanges contain a total of seven bolt holes. The 9.0 mm diameter of these
bolt holes is slightly greater than the 8.0 mm diameter of the bolt shanks to allow for some
unobstructed lateral motion of the manifold.
For simplicity, only a portion of engine head directly beneath the manifold flanges is modeled. The
4-976
head is made from aluminum, with a Young's modulus of 69 GPa, a Poisson's ratio of 0.33, and a
coefficient of thermal expansion of 22.9 ´ 10-6 per °C. The head has four exhaust ports leading into the
manifold tubes. It has seven bolt holes used to secure the manifold.
Seven bolts fasten the manifold to the head. The bolts are made from steel, with a Young's modulus of
207 GPa, a Poisson's ratio of 0.3, and a coefficient of thermal expansion of 13.8 ´ 10-6 per °C. The bolt
shanks have a diameter of 8 mm. The bolt head diameters are 16 mm.
Three-dimensional, deformable-to-deformable, small-sliding contact conditions apply to the model.
The bottom of the bolt heads form contact bearing surfaces, with the top surfaces of the manifold
flanges lying directly beneath them. In addition, the bottoms of the manifold flanges form contact
bearing surfaces with the top of the engine head. Each of these surfaces is defined in ABAQUS with
the *SURFACE option. Respective mating surfaces are paired together with the *CONTACT PAIR
option. Normal pressures will be transmitted through these contact pairs as a result of the bolt
tightening forces in Step 1. The forces carried by the bolts will vary as they respond to the thermal
cycling of the assembly in subsequent steps. These fluctuations in bolt loads will result in varying
normal pressures transmitted across the contact pairs. Lateral slip of the mating components will occur
if the critical frictional shear stress limit is surpassed by lateral forces developed in the system. A
friction coefficient of 0.2 is used between all contacting surfaces. Contact conditions are not necessary
between the bolt shanks and the holes in the manifold flanges because of the design clearance between
them. Contact between the bolt shanks and the holes in the engine head is not modeled.
All three structural components (manifold, head, and bolts) are modeled with three-dimensional
continuum elements. The model consists of 7450 first-order brick elements with incompatible
deformation modes, C3D8I, and 282 first-order prism elements, C3D6. The C3D6 elements are used
only where the complex geometry precludes the use of C3D8I elements. The C3D8I elements are
selected to represent the bending of the manifold walls with only one element through the thickness of
the tube walls.
Loading and boundary constraints

It is assumed that the engine head is securely fixed to a stiff and bulky engine block, so the nodes
along the base of the head are secured in the direction normal to the base (the global x-direction) but
are free to move in the two lateral directions to account for thermal expansion. It is also assumed that
the bolts are threaded tightly into the engine head, with the bolt threads beginning directly beneath the
section of engine head modeled. Therefore, the nodes at the bottom of the bolt shanks are shared with
the nodes of the surrounding engine head elements and are also secured in the global x-direction. The
manifold flanges are sandwiched between the top of the engine head and the base of the bolt heads
using the *CONTACT PAIR option. The line of action of the bolt forces (bolt shank axes) is along the
global x degree of freedom. Soft springs acting in the global y- and z-directions are attached to the
outlet end of the manifold and to the two ends of the head to suppress rigid body motions of the
manifold and head, respectively. These springs have no influence on the solution.
In the first step of the analysis each of the seven bolts is tightened to a uniform bolt force of 20 kN. In
subsequent steps the variation of the bolt loads is monitored as the bolts respond to the thermal loading
on the assembly as a whole. The "prescribed assembly load" capability of ABAQUS is used. For each
4-977
bolt we define a "cut," or pre-tension section, and subject the section to a specified tensile load. As a
result, the length of the bolt at the pre-tension section will change by the amount necessary to carry the
prescribed load, while accounting for the compliance of the rest of the system. In the next step the
prescribed bolt loads are replaced by the condition that the length changes calculated in the previous
step remain fixed. The remainder of the bolt is free to deform. Then, during further external loading of
the assembly the total force across each pre-tension section can be monitored as the reaction force
required to hold the pre-tension section length change constant.
The same procedure is used for all seven bolts. First, pre-tension sections are defined as "cuts" that are
perpendicular to the bolt shank axes by using the *SURFACE option on the faces of a group of
elements within each bolt shank, as shown in Figure 4.1.2-3. The line of action of the bolt force is in
the direction that is normal to this surface. Next, each bolt is assigned an arbitrary, independent node
that possesses one degree of freedom (dof 1), to which the bolt force will be applied. These nodes are
called the "pre-tension nodes" (all seven bolt pre-tension nodes are placed into a node set named
BOLTS). The spatial position of a pre-tension node is irrelevant. Finally, each surface is associated
with the appropriate pre-tension node using the *PRE-TENSION SECTION option.
A portion of the ABAQUS model definition section defining the pre-tension section is shown below:
*ELSET, ELSET=BCUT1, GENERATE
19288,19307
*SURFACE, NAME=BOLT1
BCUT1,S2
*NODE, NSET=BOLTS
99991, 21.964 , -139.80 , -12.425
...
99997, 21.964 , 137.38 , -12.226
*PRE-TENSION SECTION, SURFACE=BOLT1, NODE=99991
In Step 1 of the analysis a concentrated clamping load of 20 kN is applied to each of the pre-tension
nodes in node set BOLTS. In Step 2 the concentrated load from Step 1 is removed and replaced by a
"fixed" boundary condition on each pre-tension node that will hold the pre-tension section length
changes from Step 1 fixed. The time period of this step is small so that it will not appear in time
history plots of bolt loads. It is not a requirement that a separate step be used to replace the
concentrated bolt force with a fixed boundary constraint; the replacement procedure can, in principle,
be performed during the thermal load step. However, over the course of a step in which a load is
replaced by a boundary condition, CF1 is ramped down, while RF1 is ramped up to replace it.
Therefore, the total force across the bolt is the sum of the concentrated force ( CF1) and the reaction
force (RF1) on the pre-tension node.
In Step 3 of the analysis nodal temperatures depicting the steady-state temperature distribution in the
manifold are read from an external file. The temperature distribution is shown in Figure 4.1.2-4. These
nodal temperatures can be generated by an ABAQUS heat transfer analysis. Each of the nodes in the
model has its temperature ramped up from the initial ambient temperature of 300 K to its final
steady-state temperature. These nodal temperatures are interpolated to the element integration points so
that the correct temperature-dependent plasticity data can be used in the constitutive calculations.
4-978
Finally, in Step 4 the nodal temperatures are ramped back down to the initial ambient temperature of
300 K.

The analysis is performed as a small-displacement analysis. The nonlinearities in the problem are the
result of changing contact conditions, frictional slip and stick, and temperature-dependent plasticity.
Figure 4.1.2-5 shows the lateral displacement of the bottom surface of the flange at the end of the
heat-up step. As a result of frictional sticking, the ends of the two outer manifold flanges have
expanded outward relative to one another by only about 0.75 mm. Plastic yielding conditions result
since thermal expansion of the remainder of the manifold is constrained by this limited lateral flange
motion. A separate thermal-stress analysis of the manifold only, with no bolt constraints included,
produced relative lateral expansions of about 1.1 mm and very little plasticity.
Figure 4.1.2-6 is a plot of the forces carried by each of the seven bolts throughout the load history.
This plot can be obtained with the X-Y plotting capabilities in ABAQUS/Viewer. The curves contain
the values of the concentrated forces (CF1) for the pre-tension nodes in node set BOLTS for the first
step of the analysis and the reaction forces (RF1) for subsequent steps. The loads carried by the bolts
increase significantly during the heat-up step. The loads do not return precisely to the original bolt load
specification upon cool down because of the residual stresses, plastic deformation, and frictional
dissipation that developed in the manifold.
Input files
manifold.inp
manifold_node_elem.inp
Node and element definitions.
manifold_nodaltemp.inp
Nodal temperature data.
Figures
Figure 4.1.2-1 Manifold assemblage.
4-979
Figure 4.1.2-2 Gray cast iron temperature-dependent plasticity curves.
Figure 4.1.2-3 Pre-tension section.
4-980
Figure 4.1.2-4 Steady-state temperature distribution.
Figure 4.1.2-5 Lateral expansion of manifold footprint.
Figure 4.1.2-6 History of bolt forces.
4-981
Sample listings
4-982
Listing 4.1.2-1
*HEADING
BOLT DOWN AND THERMAL CYCLE OF EXHAUST MANIFOLD
*PREPRINT, ECHO=NO
**
** read nodes and elements from external file
** 14682 nodes in nset NALL
** 7450 C3D8I elements
** 282 C3D6 elements
** create element sets: MANIFOLD, BOLTS, HEAD
** and HDTOP
** create node sets: MANI-N, BOLT-N, HEAD-N
** and HDBASE
**
*INCLUDE, INPUT=manifold_node_elem.inp
**
*NODE,NSET=BOLTS
99991, 21.964 , -139.80 , -12.425
99992, 21.964 , -87.632 , 23.730
99993, 21.964 , -54.143 , 24.577
99994, 21.964 , -6.2972 , -25.424
99995, 21.964 , 51.885 , 23.802
99996, 21.964 , 85.185 , 24.583
99997, 21.964 , 137.38 , -12.226
*ELSET,ELSET=BCUT1,GEN
19288,19307
*SURFACE,NAME=BOLT1
BCUT1,S2
*PRE-TENSION SECTION,SURFACE=BOLT1,NODE=99991
19248,19267
*SURFACE,NAME=BOLT2
BCUT2,S2
19228,19247
*SURFACE,NAME=BOLT3
BCUT3,S2
19268,19287
*SURFACE,NAME=BOLT4
4-983
BCUT4,S2
19308,19327
*SURFACE,NAME=BOLT5
BCUT5,S2
19190,19208
*SURFACE,NAME=BOLT6
BCUT6,S2
19209,19227
*SURFACE,NAME=BOLT7
BCUT7,S2
**
** SOFT SPRINGS ON MANIFOLD AND HEAD
** TO ELIMINATE RIGID BODY MOTION
**
*ELEMENT, TYPE=SPRING1, ELSET=SOFT2
50002, 3550
60002, 3410
70002, 28794
80002, 28612
*ELEMENT, TYPE=SPRING1, ELSET=SOFT3
50003, 3550
60003, 3410
70003, 28794
80003, 28612
*SPRING, ELSET=SOFT2
2,
1.0,
*SPRING, ELSET=SOFT3
3,
1.0,
**
*ELSET, ELSET = B1, GEN
18810,18822
*SURFACE, NAME=B1S
B1,S2
*ELSET, ELSET = F1, GEN
4-984
11621,11633
*SURFACE, NAME=F1S
F1,S1
*CONTACT PAIR, SMALL SLIDING,INTERACTION=BTOF,
ADJUST=0.01
F1S, B1S
**
18797,18809
*SURFACE, NAME=B2S
B2,S2
11608,11620
*SURFACE, NAME=F2S
F2,S1
ADJUST=0.01
F2S, B2S
**
18784,18796
*SURFACE, NAME=B3S
B3,S2
11562,11574
*SURFACE, NAME=F3S
F3,S1
ADJUST=0.01
F3S, B3S
**
18771,18783
*SURFACE, NAME=B4S
B4,S2
11549,11561
*SURFACE, NAME=F4S
F4,S1
ADJUST=0.01
F4S, B4S
**
4-985

18758,18770
*SURFACE, NAME=B5S
B5,S2
11521,11533
*SURFACE, NAME=F5S
F5,S1
ADJUST=0.01
F5S, B5S
**
18745,18757
*SURFACE, NAME=B6S
B6,S2
11492,11504
*SURFACE, NAME=F6S
F6,S1
ADJUST=0.01
F6S, B6S
**
18732,18744
*SURFACE, NAME=B7S
B7,S2
11472,11484
*SURFACE, NAME=F7S
F7,S1
ADJUST=0.01
F7S, B7S
**
*SURFACE INTERACTION, NAME=BTOF
*FRICTION
0.2,
**
*SURFACE, NAME=HS
HDTOP,S1
*ELSET, ELSET = FL1, GEN
4-986
11766, 11795, 1
11874, 11899, 1
12044, 12073, 1
12076, 12101, 1
*SURFACE, NAME=FL1S
FL1,S2
*CONTACT PAIR, SMALL SLIDING,INTERACTION=FTOH,
ADJUST=0.01
FL1S, HS
**
11703, 11765, 1
11822, 11873, 1
11957, 12017, 1
12019, 12020, 1
12023, 12043, 1
12102, 12114, 1
*SURFACE, NAME=FL2S
FL2,S2
ADJUST=0.01
FL2S, HS
**
11679, 11702, 1
11796, 11821, 1
11900, 11956, 1
*SURFACE, NAME=FL3S
FL3,S2
ADJUST=0.01
FL3S, HS
**
*SURFACE INTERACTION, NAME=FTOH
*FRICTION
0.2,
**
*SOLID SECTION,ELSET=MANIFOLD,MATERIAL=CASTFE
*SOLID SECTION,ELSET=BOLTS,MATERIAL=STEEL
*SOLID SECTION,ELSET=HEAD,MATERIAL=ALUM
*MATERIAL, NAME=CASTFE
*ELASTIC, TYPE=ISO
0.1380E+06, .2830
4-987
*PLASTIC
325., 0., 293.
375., 0.000192, 293.
400., 0.000351, 293.
425., 0.000920, 293.
437.5,0.0018, 293.
**
262.5,0., 573
300., 0.000076, 573.
325., 0.000395, 573.
350., 0.0019, 573.
**
225., 0., 773.
250., 0.000188, 773.
262.5,0.0005978,773.
275., 0.001257, 773.
281.25,0.00196, 773.
**
25., 0.0, 973.
37.5, 0.000478, 973.
50., 0.0011376, 973.
51.5, 0.003627, 973.
**
12.5, 0.0,1173.
18.75, 0.0004889,1173.
25., 0.00082, 1173.
31.25, 0.00177, 1173.
*EXPANSION, TYPE=ISO
0.0000138,
**
*ELASTIC, TYPE=ISO
0.207E+06, 0.30
0.0000138,
**
*MATERIAL, NAME=ALUM
*ELASTIC, TYPE=ISO
0.6900E+5, 0.33
0.0000229,
**
*INITIAL CONDITION, TYPE=TEMPERATURE
4-988
NALL,300.
**
*BOUNDARY
HDBASE,1
**
*STEP
STEP 1.) APPLY PRESCRIBED BOLT LOADS
*STATIC
0.5,1.0
*CLOAD
BOLTS,1,20000.
*NODE FILE,NSET=BOLTS,FREQ=1
U,CF,RF
*EL PRINT,FREQ = 0
S,
PE,
*NODE PRINT,FREQ = 0
U,
RF,
*PRINT,CONTACT=YES
**
** OUTPUT FOR ABAQUS/Safe
**
*ELEMENT OUTPUT, ELSET=MANIFOLD, POSITION=NODES
S, TEMP
*NODE OUTPUT
U,
**
*END STEP
**
*STEP, INC=1
STEP 2.) SUBSTITUTE BOLT LOAD WITH RESULTANT
LENGTH CHANGE
*STATIC
1.0E-10,1.0E-10,1.0E-10,1.0E-10
*BOUNDARY, FIXED
BOLTS,1
*CLOAD, OP=NEW
*ENDSTEP
**
*STEP, INC=200
STEP 3.) APPLY PRESCRIBED THERMAL LOAD
4-989
*STATIC
0.1, 1.0, 0.00001
*TEMPERATURE, INPUT=manifold_nodaltemp.inp
*EL FILE, FREQ=100, ELSET=MANIFOLD,
POSITION=CENTROID
SP,
*ENDSTEP
**
*STEP, INC=200
STEP 4.) RETURN TO AMBIENT TEMPERATURES
*STATIC
0.2, 1.0, 0.00001
*TEMPERATURE
NALL,300.
*ENDSTEP
4.1.3 Coolant manifold cover gasketed joint

Engine gaskets are used to seal the mating surfaces of engine components to maintain the integrity of
the closed system throughout a wide range of operating loads and environmental conditions.
Inadequate gasket performance leads to diminished engine pressure and fluid leakage, resulting in
degradation of engine performance and potential engine damage. The gasket, the engine component
flanges, and the fasteners--collectively referred to as a gasketed joint--must be considered as a unit
when determining the system sealing performance because most gasketed joints do not obtain a
uniform contact stress distribution due to nonuniform bolt spacing and flange distortion during
assembly and subsequent operational loading.
Engine gaskets are often complicated geometric constructs of various engineering materials and are
subject to large compressive strains. The compressive response of the gasket is highly nonlinear. Such
complexities make detailed modeling of gaskets with continuum elements difficult and impractical
when analyzing complete assemblies.
ABAQUS has a dedicated class of elements, referred to as gasket elements, that simplify the modeling
of such components while maintaining the essential ingredients of the nonlinear response. Typical use
of these gasket elements involves a tabular representation of the pressure versus closure relationship in
the thickness direction of the gasket. The pressure versus closure models available in ABAQUS allow
the modeling of very complex gasket behaviors, including nonlinear elasticity, permanent plastic
deformation, and loading/unloading along different paths. These behaviors are usually calibrated
directly from test data. In this manner a complex gasket can be modeled effectively using a single
gasket element in the thickness direction.
In this example a paper foam gasket with a silkscreened silicone bead is compressed between the lower
engine intake manifold and the coolant manifold cover. The coolant manifold cover seals the lower
intake manifold coolant passages so that the coolant can be distributed to the cylinder heads. An
4-990
exploded view of the gasketed joint model is shown in Figure 4.1.3-1. It consists of two steel bolts, an
aluminum coolant manifold cover, a paper foam gasket with a silicone bead, and--for simplicity--only
a portion of the lower intake manifold, which is composed of steel. Symmetry conditions reduce the
structure to a half model. The gasketed joint is subjected to the following mechanical and
environmental load conditions:
1. Simulate the bolt loading sequence to fasten the joint.
2. Heat the assembly to the maximum operating temperature and apply interior cavity pressure.
3. Cool the assembly to the minimum operating temperature while maintaining interior cavity
pressure.
4. Return the assembly to ambient conditions with the interior pressure removed.
5. Disassemble the gasketed joint.

The portion of the lower intake manifold that is modeled has two passages. Coolant flows from one
passage into the manifold cover and back out through the other passage. Two steel bolts secure the
cover to the manifold. The bolt shanks have a diameter of 6.0 mm, and the bolt heads have a diameter
of 11.8 mm. The bolts and the lower intake manifold are assigned a Young's modulus of 2.0 ´ 105
MPa, a Poisson's ratio of 0.28, and a coefficient of thermal expansion of 1.6 ´ 10-5 per °C. The
aluminum coolant manifold cover has a Young's modulus of 7.1 ´ 104 MPa, a Poisson's ratio of 0.33,
and a coefficient of thermal expansion of 2.3 ´ 10-5 per °C.
The metal components (bolts, cover, and intake manifold) are modeled with three-dimensional
continuum elements: 1304 first-order brick elements with incompatible deformation modes ( C3D8I)
and 208 first-order prism elements (C3D6). The C3D8I elements are chosen to capture the bending of
the cover, using only one element through its thickness. The C3D6 elements are used only where
geometric constraints preclude the use of C3D8I elements.
The gasket schematic shown in Figure 4.1.3-2 has two distinct regions. The majority of the gasket is
composed of a 0.79 mm thick, flat, crushable paper foam material. To ensure proper sealing pressures
for this joint, a 0.076 mm thick silicone bead has been silkscreened along the top surface of the gasket
encircling the interior cavity. Placing silicone beads on gaskets results in a change in the load
transmitting characteristics of the gasket, which often improves both the recovery properties of the
gasket and its potential to remain sealed for the long term.
The entire gasket, including the bead, is modeled as a flat sheet with one gasket element through the
thickness (see Figure 4.1.3-3). A fine mesh is used for the gasket to capture the in-plane variation of
the gasket sealing pressure. This creates a mismatched mesh across the contacting surfaces, but
ABAQUS contact definitions do not require one-to-one matching meshes across contact pairs. The
gasket components (silicone bead region and paper foam region) are modeled with 973 first-order
8-node area elements (GK3D8) and 29 first-order 6-node area elements (GK3D6). The physical
thickness of the entire sheet of gasket elements corresponds to the initial combined height of the paper
4-991
foam and the silicone bead, 0.866 mm. The elements in the region of the gasket beneath the silicone
bead are assigned different gasket properties from the rest of the elements in the gasket model. The
paper foam region is initially not in contact with the cover. The initial gap is 0.076 mm. No pressure is
generated in this portion of the gasket until the gap has closed. Gasket region property distinctions,
such as initial gaps and different pressure versus closure relationships, are assigned to corresponding
element sets by referring to different *GASKET SECTION options.
Experimentally determined pressure versus closure curves for the two distinct gasket regions without
the initial gap taken into account are shown in Figure 4.1.3-4. Tabular representations of these curves
are specified using the *GASKET THICKNESS BEHAVIOR option that is associated with the
respective *GASKET SECTION options. Creep/relaxation properties of the gasket and
temperature-dependent pressure versus closure properties, capturing such effects as the glassy
transition temperature of the silicone bead, are not accounted for in this example. Initially, ABAQUS
considers the gasket behavior to be nonlinear elastic, such that loading and unloading occur along the
same user-defined nonlinear path. ABAQUS considers yielding to occur once the slope of the pressure
versus closure curve decreases by at least 10%. In addition to the single loading curve, whose closure
increases monotonically, the user can define any number of unloading curves at different levels of
plastic closure. Yielding occurs at a closure of 0.1118 mm for both regions of the gasket in this
example, after which the gasket stiffness decreases slightly up to a closure of 0.15 mm, the final point
on the loading curve. Beyond the data of the loading curve defined by the user, ABAQUS considers
the gasket to behave with a fully crushed elastic response by linearly extrapolating the last segment of
the last specified unloading curve (alternatively, the user could have specified a piecewise linear
form).
A single unloading curve is defined for each of the two gasket regions: the unloading curve for the
silicone bead region is defined at 0.11 mm of plastic closure, and the unloading curve for the paper
foam region is defined at 0.09 mm of plastic closure. Any unloading of the gasket beyond the yield
point occurs along a curve interpolated between the two bounding unloading curves, which--for this
example--are the initial, nonlinear elastic curve and the single unloading curve.
Gasket materials often have higher coefficients of thermal expansion than most of the metals from
which the bolts and flanges are made. For situations involving wide and rapid temperature fluctuations
resultant differences in relative expansion and contraction can have a significant effect on the sealing
properties of the gasket. The coefficient of thermal expansion for the silicone bead region is 1.2 ´ 10-4
per °C, and for the paper foam region it is 3.0 ´ 10-5 per °C.
In this case, because of the differences in thermal expansion between the aluminum cover and the steel
intake manifold, it is important to account for the membrane and transverse shear properties of the
gasket and to model frictional effects between mating surfaces. For this analysis the silicone bead
region of the gasket is defined to have a membrane stiffness of 75 MPa and a transverse shear stiffness
of 40 MPa. The base foam material is defined with a value of 105 MPa for the membrane stiffness and
a value of 55 MPa for the transverse shear stiffness. A friction coefficient of 0.2 is used between all
mating surfaces.
A separate analysis is included in this example problem using the "thickness-direction only" version of
the gasket elements (GK3D8N and GK3D6N). These elements respond only in the thickness direction
4-992
and have no membrane or transverse shear stiffness properties. They possess only one degree of
freedom per node. As a result, frictional effects cannot be included at the surfaces of these elements.
They are more economical than more general gasket elements that include membrane and transverse
shear responses and may, thus, be preferable in models where lateral response can be considered
negligible.
Loading and boundary constraints

Symmetry boundary constraints are placed along the nodes on the symmetry plane. Furthermore, it is
assumed that the intake manifold is a stiff and bulky component, so nodes along the base of the portion
of the manifold modeled are secured in the normal direction (the global z-direction). Except for a soft
spring constraint to eliminate rigid body motion, these manifold base nodes are free to displace
laterally to allow for thermal expansion. Soft springs are also attached to the cover to eliminate rigid
body motion in the x- and z-directions.
The bottoms of the bolt heads form contact bearing surfaces with the top surface of the cover flange. In
addition, the top of the gasket interacts with the bottom of the cover, while the bottom of the gasket
contacts the top of the manifold. Each of these surfaces is defined with the *SURFACE option. Mating
surfaces are paired together with the *CONTACT PAIR option. Three-dimensional,
deformable-to-deformable, small-sliding contact conditions apply to each of these contact pairs. The
gasket is attached to the manifold base using the *SURFACE BEHAVIOR, NO SEPARATION option,
thus constraining it against rigid body motion in the global z-direction. The gasket membrane is
allowed to stretch, contract, or shear as a result of frictional effects on both sides of the gasket. The
bolts are assumed to be threaded tightly into the base. Therefore, the nodes at the bottom of the bolt
shanks are shared with the intake manifold. Contact between the bolt shanks and the bolt holes is not
modeled.
The "prescribed assembly load" capability is used to define pre-tension loads in each of the bolts. For
each of the two bolts we define a "cut" or pre-tension section and subject the section to a specified
load. As a result, the length of the bolt at the pre-tension section changes by the amount necessary to
carry the prescribed load, while accounting for the compliance of the rest of the joint. Once a bolt has
been pre-tensioned, the applied concentrated bolt load is replaced with a "fixed" boundary condition,
which specifies that the length change of the bolt at the "cut" remains fixed, while the remainder of the
bolt is free to deform. Then, during further external loading of the assembly the total force across each
pre-tension section (the load on the bolt) can be monitored as the reaction force required to hold the
resultant length change of the pre-tension section constant.
The sequence in which the bolts are tightened can have an impact on the distribution of the resultant
contact area stress. A poorly specified bolt sequence can cause excessive distortion of the gasket and
the flanges, which may lead to poor sealing performance. In the first step of the analysis the left bolt is
pre-tensioned to a load of 6000 N using the *PRE-TENSION SECTION option. In the second step the
right bolt is pre-tensioned to 6000 N. An interim step, referred to as Step 1a, is used to replace the
prescribed load on the left bolt with a fixed boundary condition as described above. Since only half of
each bolt is modeled, a total load of 12000 N is carried by each bolt. Following Step 2 another interim
step, referred to as Step 2a, replaces the prescribed load on the pre-tension section of the right bolt
with a fixed boundary condition.
4-993
Step 3 is the beginning of the three-step thermo-mechanical operational cycle. In Step 3 the entire
assembly is heated uniformly to its maximum operating temperature of 150°C, while simultaneously
pressurizing the interior cavity to 0.689 MPa. In Step 4 the system temperature is decreased to the
minimum operating temperature of -40°C while maintaining the interior pressure load of 0.689 MPa.
In Step 5 the gasketed joint is returned to the ambient temperature conditions and the internal cavity
pressure is removed.
The sixth and final step in the analysis simulates disassembly of the gasketed joint by removing the
bolt loads. This process demonstrates the interpolated unloading response for the different regions of
the permanently deformed gasket.

The prime interest in this problem is the variation of bolt forces during the initial assembly and
thermo-mechanical cycle and the resultant distribution and variation of the gasket sealing pressure.
The function of the fasteners in a gasketed joint is to apply and maintain the load required to seal the
joint. The bolt pattern and tension are directly related to the sealing pressure in the clamped gasket. At
the maximum service temperature the bolt loads can be expected to be at their peak as a result of
thermal expansion effects. It is important to ensure that the stress values of the metal engine
components remain below yield and that there is no significant bending of the flanges, which may
cause improper sealing of the gasket. At the minimum operating temperature the bolt loads are
expected to reach a minimum as a result of thermal contraction effects. Hence, it is necessary to assess
that adequate sealing pressure is retained throughout the gasket.
Figure 4.1.3-5 shows the bolt load variation over the course of the six analysis steps. During the first
step the pre-tension section node on the right bolt was prescribed a zero change of length constraint,
which implies that the right bolt has just been placed in position but not torqued tightly. Hence, as the
left bolt is tightened during Step 1, a small reaction load is generated in the right bolt. At the end of the
second step during which the right bolt is tightened to carry a force of 6000 N, the force in the left bolt
increases to 6200 N. In Step 3 the deformation of the assembly causes the bolt forces to increase to
maximum values of 6800 N in the left bolt and 6600 N in the right bolt because of thermal expansion
and interior pressurization. When the assembly is cooled to the minimum operating temperature, the
bolt loads reach their minimum values. Due to thermal cycling and interior cavity pressure inducing
inelastic response in the gasket, the bolt forces at the end of the operational cycle reduce to 6050 N in
the left bolt and 5950 N in the right bolt.
The gasket sealing pressure pattern depends on the rigidity of the flanges. Hence, it is useful to predict
how the structure will deform due to the applied loading. Figure 4.1.3-6 shows the deformed shape of
the coolant manifold cover at a displacement magnification factor of 50. Bowing of the cover from
initial assembly and subsequent operational loads will lead to a nonuniform sealing pressure
distribution in the gasket. Figure 4.1.3-7illustrates the gasket pressure distribution after initial
fastening of the joint.
Figure 4.1.3-8 shows the sealing pressure as a function of position along the perimeter of the silicone
bead at the end of each of the analysis steps. The sealing pressure reaches a minimum at the point
4-994
equidistant from the bolts, making this the critical point in the gasketed joint design. This figure also
reflects the reduction in the sealing pressure near the bolt holes as a result of plastic deformation of the
gasket body during the operational cycle. Figure 4.1.3-9 is a contour plot of the permanent deformation
in the gasket after completion of the thermo-mechanical cycle.
Figure 4.1.3-10 follows the pressure/closure history of one point in the gasket during this analysis in
relation to the user-specified loading/unloading test data. The "mechanical closure" (total closure, E11,
minus thermal closure, THE11) is plotted along the abscissa of this figure. The material point traced
(element 18451, integration point 1) is located along the inside periphery of the silicone bead at the
symmetry plane of the assembly nearest the left bolt. Step 1 shows that this point follows the initial
elastic loading curve up to the closure of 0.1118 mm. After this amount of closure, further loading
causes plastic deformation. In the second step the tightening of the bolt results in a very small amount
of unloading for this material point. For purposes of clarity, this deformation is not shown in the
figure. Step 3 involves heating the system to the maximum operating temperature and pressurizing the
interior cavity so that further yielding of the material point occurs. Step 4 results in the partial
unloading of the point due to the thermal contraction associated with cooling the assembly to the
minimum operating temperature. For this case the unloading path is based on a curve interpolated
between the initial, nonlinear elastic curve and the single unloading curve. The return of the assembly
to ambient conditions partially reloads this point along the same path as the previous unloading;
however, no further yielding of this material point occurs during this step. In the final step the gasket is
unloaded completely.
The analysis using the "thickness-direction only" gasket elements runs in nearly half the CPU time of
the full three-dimensional gasket element model. Minimum gasket sealing pressures in Step 4 of this
analysis are predicted to be about 20% lower because frictional effects are neglected.
Input files
manifoldgasket.inp
manifoldgasket_mesh.inp
Node, element, and surface definitions.
manifoldgasket_thick.inp
"Thickness-direction only" gasket element analysis.
manifoldgasket_thick_mesh.inp
Node, element, and surface definitions for the "thickness-direction only" gasket element analysis.
Reference
· Czernik, D. E., Gasket Handbook, McGraw-Hill, New York, 1996.
Figures
4-995
Figure 4.1.3-1 Coolant manifold assemblage.
Figure 4.1.3-2 Schematic representation of a silicone bead printed on the gasket body.
4-996
Figure 4.1.3-3 Mesh of gasket with silicone bead highlighted.
Figure 4.1.3-4 Pressure versus closure behavior for the gasket and the gasket with silicone bead.
Figure 4.1.3-5 History of bolt force.
4-997
Figure 4.1.3-6 Deformed shape of coolant manifold cover at a displacement magnification factor of
50.
Figure 4.1.3-7 Gasket pressure distribution after initial fastening sequence.
Figure 4.1.3-8 Sealing pressure along inside periphery of silicone bead region of gasket.
4-998
Figure 4.1.3-9 Plastic closure in gasket after operational cycle.
Figure 4.1.3-10 Typical pressure-closure diagram for material point in silicone bead region of gasket.
Sample listings
4-999
Listing 4.1.3-1
*HEADING
BOLT DOWN AND OPERATIONAL CYCLE OF COOLANT
MANIFOLD COVER
Units: N-mm-sec-deg C
**
**
** read nodes, elements, and surfaces
** from external file
** 4639 nodes in nset NALL
** 1304 C3D8I elements
** 208 C3D6 elements
** 973 GK3D8 elements
** 29 GK3D6 elements
** create element sets: COVER1,COVER2,BASE1,BASE2,
** BOLT1,MATL1,BEAD1,COVER,
** BASE,COVERF1,COVERF2,
** COVERF3,COVERF4,COVERF5,
** COVERF6,BASEF1,BASEF3,
** BASEF4,BASEF5,BASEF6
** create node sets: NALL,BASE_BOT
** create surfaces: SCOVER2,SBOLTL,SBOLTR,SCOV1L,
** SCOV1R,SGTOP,SGBOT,SBASE1
**
*INCLUDE,INPUT=manifoldgasket_mesh.inp
**
*****
*GASKET SECTION, ELSET=BEAD1, BEHAVIOR=GBEAD1
*GASKET BEHAVIOR,NAME=GBEAD1
*GASKET ELASTICITY,COMPONENT=MEMBRANE
75.,
*GASKET ELASTICITY,COMPONENT=TRANSVERSE SHEAR
40.0,
*GASKET THICKNESS BEHAVIOR,TYPE=ELASTIC-PLASTIC,
DIRECTION=LOADING
0.0 ,0.0
3.27 ,0.0254
8.40 ,0.0508
18.52,0.0762
34.26,0.1016
41.69,0.1118
4-1000
45.50,0.13
47.80,0.15
DIRECTION=UNLOADING
0.0,0.11,0.11
10.0,0.13,0.11
20.0,0.14,0.11
30.0,0.145,0.11
40.0,0.1475,0.11
47.8,0.15,0.11
*EXPANSION
1.2E-4,
*GASKET SECTION, ELSET=MATL1, BEHAVIOR=GMATL1
,0.076
*GASKET BEHAVIOR,NAME=GMATL1
*GASKET ELASTICITY,COMPONENT=MEMBRANE
105.,
*GASKET ELASTICITY,COMPONENT=TRANSVERSE SHEAR
55.,
DIRECTION=LOADING
0.0 ,0.0
4.83 ,0.0254
11.72,0.0508
25.17,0.0762
43.09,0.1016
52.75,0.1118
57.02,0.13
58.85,0.15
DIRECTION=UNLOADING
0.0,0.09,0.09
10.0,0.108,0.09
20.0,0.122,0.09
30.0,0.131,0.09
40.0,0.142,0.09
50.0,0.147,0.09
58.85,0.15,0.09
*EXPANSION
3.E-5,
*****
*SOLID SECTION, ELSET=COVER, MATERIAL=ALUM
*SOLID SECTION, ELSET=BASE, MATERIAL=STEEL
4-1001
*SOLID SECTION, ELSET=BOLT1, MATERIAL=STEEL

*****
*ELASTIC
200000.,.28
*EXPANSION
1.6E-5,
*****
*MATERIAL, NAME=ALUM
*ELASTIC
71000.,.33
*EXPANSION
2.3E-5,
*****
**
** INITIAL TEMPERATURE -- 20 DEGREES CELSIUS
**
NALL,20.0
**
**
** CONTACT DEFINITIONS
**
*CONTACT PAIR,INTERACTION=IBOLT,SMALL SLIDING,
ADJUST=.01
SBOLTL,SCOV1L
SBOLTR,SCOV1R
*CONTACT PAIR,INTERACTION=ITOP,SMALL SLIDING,
ADJUST=.01
SGTOP,SCOVER2
*CONTACT PAIR,INTERACTION=IBOT,SMALL SLIDING,
ADJUST=.01
SGBOT,SBASE1
*SURFACE INTERACTION,NAME=IBOLT
*FRICTION
0.20,
*SURFACE INTERACTION,NAME=ITOP
*FRICTION
0.20,
*SURFACE INTERACTION,NAME=IBOT
*FRICTION
0.20,
4-1002
**
** PRE-TENSION SECTION DEFINITIONS
**
*ELSET,ELSET=RCUT
3701,3702,3703,3708
*ELSET,ELSET=LCUT
3709,3710,3711,3716
*SURFACE,NAME=PREBOLTL
LCUT, S1
*SURFACE,NAME=PREBOLTR
RCUT, S1
*PRE-TENSION SECTION,SURFACE=PREBOLTL,NODE=1000001
*PRE-TENSION SECTION,SURFACE=PREBOLTR,NODE=1000002
**
** SOFT SPRINGS ON COVER, BASE AND GASKET
** TO ELIMINATE RIGID BODY MOTION
** - Nodes 643 and 2012 are located on COVER
** - Nodes 5560 and 5689 are located on BASE
** - Nodes 15827 and 15737 are located on GASKET
**
*ELEMENT,TYPE=SPRING1,ELSET=SP1
999001,643
999002,2012
999003,5560
999004,5689
999005,15827
999006,15737
*ELEMENT,TYPE=SPRING1,ELSET=SP3
999007,643
999008,2012
*SPRING,ELSET=SP1
1,
1.,
*SPRING,ELSET=SP3
3,
1.,
**
** DISPLACEMENT BOUNDARY CONDITIONS
**
*BOUNDARY
BASE_BOT,3
YSYMM,2
**
4-1003
** NODE SETS FOR OUTPUT

**
*NSET,NSET=REFNODE
1000001,1000002
*****************************************
**
** STEP 1 -- PRE-TENSION LEFT BOLT
**
*STEP,INC=20
*STATIC
0.01,1.0
*CLOAD
1000001,1,6000.
*BOUNDARY
1000002,1,,0.0
**
*NODE FILE,NSET=REFNODE,FREQ=1
U,RF,CF
*EL FILE,ELSET=IBEAD,FREQ=1
S,E,THE
*EL PRINT,FREQ=0
*NODE PRINT,FREQ=0
*END STEP
**
** STEP 1A -- SUBSTITUTE LEFT BOLT LOAD
** WITH RESULTANT LENGTH CHANGE
**
*STEP,INC=1
*STATIC
1.E-10,1.0E-10,1.0E-10,1.0E-10
*BOUNDARY,FIXED
1000001,1
*CLOAD,OP=NEW
*ENDSTEP
**
** STEP 2 -- PRE-TENSION RIGHT BOLT
**
*STEP,INC=20
*STATIC
0.01,1.0
*CLOAD,OP=NEW
4-1004
1000002,1,6000.
*BOUNDARY,OP=NEW,FIXED
1000001,1
*BOUNDARY,OP=NEW
BASE_BOT,3
YSYMM,2
*END STEP
**
** STEP 2A -- SUBSTITUTE RIGHT BOLT LOAD
** WITH RESULTANT LENGTH CHANGE
**
*STEP,INC=1
*STATIC
1.E-10,1.0E-10,1.0E-10,1.0E-10
*BOUNDARY,FIXED
1000002,1
*CLOAD,OP=NEW
*ENDSTEP
**
** STEP 3 -- INCREASE TEMPERATURE TO 150
** DEGREES CELSIUS and PRESSURIZE
** INTERIOR CAVITY TO 0.6895 MPa
**
*STEP,INC=20
*STATIC
0.125,1.0
*TEMPERATURE
NALL,150.
*DLOAD,OP=NEW
coverf1,p1,0.6895
coverf2,p2,0.6895
coverf3,p3,0.6895
coverf4,p4,0.6895
coverf5,p5,0.6895
coverf6,p6,0.6895
basef1 ,p1,0.6895
basef3 ,p3,0.6895
basef4 ,p4,0.6895
basef5 ,p5,0.6895
basef6 ,p6,0.6895
*END STEP
**
4-1005
** STEP 4 -- DECREASE TEMPERATURE TO

** -40 DEGREES CELSIUS
**
*STEP,INC=20
*STATIC
0.125,1.0
*TEMPERATURE
NALL,-40.0
*END STEP
**
** STEP 5 -- INCREASE TEMPERATURE TO
** 20 DEGREES CELSIUS and
** REMOVE INTERNAL PRESSURE LOAD
**
*STEP,INC=20
*STATIC
0.05,1.0
*TEMPERATURE
NALL,20.0
*DLOAD,OP=NEW
*END STEP
**
** STEP 6 -- UNLOAD THE GASKET
**
*STEP,INC=20
*STATIC
0.125,1.0
*BOUNDARY,OP=NEW
BASE_BOT,3
YSYMM,2
1000001,1,,0.005
1000002,1,,0.005
*END STEP
4.1.4 Radiation analysis of a plane finned surface

This example illustrates the ABAQUS capability to solve heat transfer problems including cavity
radiation. We simulate the effects of a fire condition on a plane finned surface. This problem was
proposed by Glass et al. (1989) as a benchmark for thermal radiation. We compare their results with
those obtained using ABAQUS.
The configuration shown in Figure 4.1.4-1 represents a plane wall with a uniform array of parallel
rectangular fins attached. The problem represents three phases in a fire test. The first is the pretest, a
4-1006
steady-state condition where heat is transferred by natural convection from an internal fluid at a fixed
temperature of 100°C to the plane inside wall. Heat is conducted through the wall and dissipated by
radiation and natural convection from the outside wall and fin surfaces to the surrounding medium
which is at a temperature of 38°C. The second phase is a 30-minute fire transient, where heat is
supplied by radiation and forced convection from a hot external fluid at 800°C. After conduction
through the fins and wall, heat is rejected by natural convection to the internal fluid. Finally, the third
phase is a 60-minute cool down period, where heat absorbed during the fire transient is rejected to the
surroundings by the same process as that used to establish the initial steady-state condition.
Geometry and model

The finite element mesh used for the wall and fins is shown in Figure 4.1.4-2. By making use of the
radiation periodic symmetry capability in ABAQUS, we are able to represent the array of fins while
meshing only one fin and corresponding wall section.
The outside ambient is modeled with a single horizontal row of elements at some distance above the
top of the fin (not shown in the figure). The varying ambient temperature is simulated by prescribing
temperatures to the nodes of these elements. The elements representing the outside ambient are also
assigned a surface emissivity of 1.0.
Material and boundary conditions

The thermal conductivity of the wall and fins is 50 W/m°C (k), their specific heat is 500 J/kg°C (c),
and the density is 7800 kg/m 3 (½). The surface emissivity of the wall and fins is 0.8, the
Stefan-Boltzmann radiation constant is 5.6697 ´ 10-8 W/m2°K4, and the temperature of absolute zero is
-273°C.
The natural convection between the internal fluid and the inside of the wall is modeled with a film
boundary condition where the film coefficient is given as 500( µw ¡ µf )1/3 W/m2°C, where µw is the
inside wall temperature and µf is the temperature of the internal fluid. The film boundary condition
user subroutine is used for this purpose since the film condition is temperature dependent.
The natural convection between the outside finned surface and its surroundings is modeled with a film
boundary condition where the film coefficient is given as 2(µs ¡ µa )1/3 W/m2°C, where µs is the
temperature of the finned surface and µa is the outside ambient temperature. Again, the film boundary
condition user subroutine is employed. The forced convection between the hot surroundings and the
finned surface is modeled with a constant film coefficient of 10 W/m 2°C.
Loading
The first simulation step is a steady-state heat transfer analysis to establish the initial pretest
conditions. This is followed by a 30-minute transient heat transfer analysis during which time the
ambient fire temperature is 800°C. Finally, a second transient heat transfer step is performed to
simulate the 60-minute cool down period.
The integration procedure used in ABAQUS for transient heat transfer analysis procedures introduces
a relationship between the minimum usable time increment and the element size and material
properties. The guideline given in the User's Manual is
4-1007
½c 2
¢t > ¢l ;
6k
where ¢l is the element size. This suggests that an initial time increment of 10 seconds is appropriate
for the transient steps of this problem. Automatic time incrementation is chosen for the transient steps
by setting DELTMX to 100°C. DELTMX controls the time integration by limiting the temperature
change allowed at any point during an increment. A DELTMX of 100°C may seem rather coarse in
comparison with temperature ranges of 800°C, but this is not expected to be the limiting factor; the
strong nonlinearity of the radiation conditions is expected to dictate the time incrementation in the
transient steps.

The results published by Glass et al. include those obtained by a number of different heat transfer
codes, all of which give similar results. Since the most details are given for the results obtained with
the program TAU (Johnson, 1987), we have chosen to compare the ABAQUS results to those of TAU.
Figure 4.1.4-3 shows the history of the temperature at the top of the fin (point 1 in Figure 4.1.4-1).
Figure 4.1.4-4 shows the histories of the temperature at the root of the fin (point 2 in Figure 4.1.4-1)
and on the wall inside surface (point 3). In all cases the results obtained with ABAQUS match the
TAU results quite well. In Figure 4.1.4-5 we show the temperature distribution around the fin
perimeter (starting at point 1 and ending at point 2) at the end of the fire transient. Again, the
ABAQUS and TAU results match closely. Finally, temperature contours at the end of the fire transient
are shown in Figure 4.1.4-6.
Input file
radiationfinnedsurf.inp
Fire transient problem.
radiationfinnedsurf.f
User subroutine FILM used in radiationfinnedsurf.inp.
References
· Glass, R. E., et al., "Standard Thermal Problem Set," Proceedings of the Ninth International
Symposium on the Packaging of Radioactive Materials, pp. 275-282, June 1989.
· Johnson, D., "Surface to Surface Radiation in the Program TAU, Taking Account of Multiple
Reflection," United Kingdom Atomic Energy Authority Report ND-R-1444(R), 1987.
Figures
Figure 4.1.4-1 Plane finned surface.
4-1008
Figure 4.1.4-2 Finite element mesh of fin and inner wall.
Figure 4.1.4-3 Temperature history at top of fin.
4-1009
Figure 4.1.4-4 Temperature history at root of fin and inside wall surface.
Figure 4.1.4-5 Temperature distribution along fin perimeter at end of fire transient.
4-1010
Figure 4.1.4-6 Temperature contours at end of fire transient.
Sample listings
4-1011
Listing 4.1.4-1
*HEADING
PLANE FINNED SURFACE UNDER FIRE CONDITION
**
** FIN...
*NODE
1,
5, 0.025D0
7, 0.035D0
11, 0.06D0
1001, 0.D0,0.1D0
1005,0.025D0,0.1D0
1007,0.035D0,0.1D0
1011,0.060D0,0.1D0
3001, 0D0,0.25D0
3005,0.025D0,0.25D0
3007,0.035D0,0.25D0
3011,0.060D0,0.25D0
*NGEN,NSET=N1
1,5
5,7
7,11
*NGEN,NSET=N2
1001,1005
1005,1007
1007,1011
*NGEN,NSET=N3
3001,3005
3005,3007
3007,3011
*NFILL,NSET=NALL
N1,N2,10,100
N2,N3,20,100
*ELEMENT,TYPE=DC2D4
1,1,2,102,101
5000,1005,1006,1106,1105
*ELGEN,ELSET=STRUC
1,10,1,1,10,100,10
5000,2,1,1,20,100,2
**
** AMBIENT...
*NODE
4-1012
5001, 0D0,0.550D0
5011,0.06D0,0.550D0
5101, 0D0,0.555D0
5111,0.06D0,0.555D0
*NGEN,NSET=NAMB
5001,5011
5101,5111
*ELEMENT,TYPE=DC2D4
8000,5001,5002,5102,5101
*ELGEN,ELSET=STRUC
8000,10,1,1
*ELSET,ELSET=EAMB,GENERATE
8000,8009
**
** SURFACES...
*ELSET,ELSET=BOTF1,GENERATE
1,10
*ELSET,ELSET=TOPF3,GENERATE
91,94
97,100
5038,5039
5000,5038,2
5001,5039,2
*surface,NAME=SRFS,PROPERTY=REFL
TOPF3,S3
TOPF4,S4
TOPF2,S2
*surface,NAME=SAMB,PROPERTY=RAMB
EAMB,S1
*SURFACE PROPERTY,NAME=REFL
*EMISSIVITY
0.8D0,
*SURFACE PROPERTY,NAME=RAMB
*EMISSIVITY
1.0D0,
*CAVITY DEFINITION,NAME=CAV2D
SRFS,SAMB
**
*SOLID SECTION,ELSET=STRUC,MATERIAL=MSTRUC
*MATERIAL,NAME=MSTRUC
*DENSITY
4-1013
7800.0D0,
*CONDUCTIVITY
50.0D0,
*SPECIFIC HEAT
500.0D0,
*PHYSICAL CONSTANTS,STEFANBOLTZMANN=5.6697D-8,
ABSOLUTE ZERO=-273.0
**
NALL,77.D0
NAMB,77.D0
*NSET,NSET=THREE
11,1011,3006
**
**
*surface, NAME=BOTF1
BOTF1, S1
*STEP,INC=500
*HEAT TRANSFER,STEADY STATE
1.0,
*BOUNDARY
NAMB,11,,38.D0
*SFILM
BOTF1,FNU
*SFILM
SRFS,FNU
*RADIATION VIEWFACTOR,SYMMETRY=NSYMM
*RADIATION SYMMETRY,NAME=NSYMM
*PERIODIC,TYPE=2D,NR=20
0.0D0,0.555D0, 0.0D0,0.0D0, 0.06D0,0.0D0
*VIEWFACTOR OUTPUT, CAVITY=CAV2D, FREQUENCY=100
*RADIATION PRINT,CAVITY=CAV2D, FREQUENCY=10
RADFL,RADFLA,RADTL,RADTLA,VFTOT,FTEMP
*RADIATION FILE,CAVITY=CAV2D, FREQUENCY=10
RADFL,RADFLA,RADTL,RADTLA,VFTOT,FTEMP
*NODE PRINT,NSET=THREE
NT,
*NODE FILE,NSET=THREE
NT,
*END STEP
*STEP,INC=500
4-1014
*HEAT TRANSFER,DELTMX=100
10,1800
*BOUNDARY
NAMB,11,,800.D0
*END STEP
*STEP,INC=500
*HEAT TRANSFER,DELTMX=100
10,3600
*BOUNDARY
NAMB,11,,38.D0
*END STEP
4-1015
Electrical Analyses
5. Electrical Analyses
5.1 Piezoelectric analyses
5.1.1 Eigenvalue analysis of a piezoelectric transducer
This problem performs an eigenspectrum analysis of a cylindrical transducer consisting of a
piezoelectric material with brass end caps. Various elements are used in the analysis. The elements
range from axisymmetric elements to three-dimensional elements, using both lower- and higher-order
elements. The basis of the piezoelectric capability in ABAQUS is described in ``Piezoelectric
analysis,'' Section 2.10.1 of the ABAQUS Theory Manual.

This problem is identical to the one discussed in a report by Mercer et al. (1987). The structure is
shown in Figure 5.1.1-1and consists of a piezoelectric material PZT4 with brass end caps. The
piezoelectric material is electroded on both the inner and outer surfaces.
The properties for PZT4 in a cylindrical system are:
Elasticity Matrix:
2 3
115:4 74:28 74:28 0 0 0
6 74:28 139:0 77:84 0 0 0 7
6 7
6 74:28 77:84 139:0 0 0 0 7
6 7 GPa
6 0 0 0 25:64 0 0 7
4 5
0 0 0 0 25:64 0
0 0 0 0 0 25:64
Piezoelectric Coupling Matrix (Stress Coefficients):

2 3
15:08 0 0
6 ¡5:207 0 0 7
6 7
6 ¡5:207 0 0 7
6 7 Coulomb/m2
6 0 12:710 0 7
4 5
0 0 12:710
0 0 0
Dielectric Matrix:
2 3
5:872 0 0
4 0 6:752 0 5 10¡9 farad/m
0 0 6:752
The 1-direction is radial, the 2-direction is axial, and the 3-direction is tangential. From these matrices
5-1016
Electrical Analyses
it is seen that the poling direction is radially outwards from the axis of symmetry. (The order of the
stresses in ABAQUS may differ from those typically used in electrical applications. ABAQUS uses the
standard mechanical convention, where the stress components are ordered as f¾11 ¾22 ¾33 ¿12 ¿13 ¿23 g.
See ``Piezoelectric behavior,'' Section 12.6.2 of the ABAQUS/Standard User's Manual.)
The brass is elastic and isotropic with a Young's modulus of 104 GPa and a Poisson's ratio of 0.37.
Models
The transducer is modeled with a variety of elements. It is modeled as an axisymmetric structure
utilizing both the planar, axisymmetric elements and the three-dimensional elements. For the
axisymmetric elements, five meshes employing 4-node, 6-node, and 8-node elements are used in the
finite element discretization. The first two meshes use 4-node elements with two levels of refinement,
the third mesh uses 6-node elements, and the last two meshes use the 8-node elements with two levels
of refinement. Lumped mass matrices are used for the lower-order elements. Consistent mass matrices
are used in the higher-order elements. The meshes used for the 4-node and 6-node axisymmetric
elements are shown in Figure 5.1.1-2. The meshes used for the 8-node axisymmetric elements are
shown in Figure 5.1.1-3.
The three-dimensional model uses a slice of the structure and applies axisymmetric boundary
conditions; 8-node and 20-node brick elements are used. The discretization used for each model is
shown in Figure 5.1.1-3. These models use the *ORIENTATION option to maintain the proper
definitions of the material properties. Also, in order to prescribe the axisymmetric boundary
conditions, the nodal degrees of freedom are transformed into a cylindrical coordinate system.
All the models are considered to be open-circuited. The potentials on the inside surface are restrained
to zero. The frequencies correspond to those for anti-resonance.

The solutions obtained with the various ABAQUS models are shown in Table 5.1.1-1. Even for these
coarse models, the results are quite close to the experimental results. In addition, the results from
ABAQUS for the lower-order axisymmetric elements with lumped mass and the higher-order
axisymmetric elements with consistent mass matrices in the computation of both the resonant and
antiresonant frequencies match well with the numerical results reported in Mercer et al. The first four
mode shapes for the more refined model with CAX8RE elements are shown in Figure 5.1.1-4.
Similar analyses have been performed considering the problem to be closed-circuited to obtain the
resonant frequencies. For this situation, the potentials on both the inner and outer surfaces are set to
zero. The results also compare well with those given in Mercer et al.
Input files
eigenpiezotrans_cax4e_coarse.inp
Coarse mesh with 4-node axisymmetric elements.
eigenpiezotrans_cax4e_fine.inp
5-1017
Electrical Analyses
Refined mesh with 4-node axisymmetric elements.

eigenpiezotrans_cax6e.inp
Mesh with 6-node axisymmetric elements.
eigenpiezotrans_cax8re_coarse.inp
Coarse mesh with 8-node axisymmetric elements.
eigenpiezotrans_cax8re_fine.inp
Refined mesh with 8-node axisymmetric elements.
eigenpiezotrans_c3d8e.inp
Mesh with 8-node three-dimensional elements.
eigenpiezotrans_c3d8e.f
User subroutine ORIENT used in eigenpiezotrans_c3d8e.inp.
eigenpiezotrans_c3d20e.inp
Mesh with 20-node three-dimensional elements.
eigenpiezotrans_c3d20e.f
User subroutine ORIENT used in eigenpiezotrans_c3d20e.f.
eigenpiezotrans_elmatrixout.inp
Data that test the use of *ELEMENT MATRIX OUTPUT with piezoelectric elements.
eigenpiezotrans_usr_element.inp
Data for a job that reads in the matrices output in eigenpiezotrans_elmatrixout.inp and performs an
eigenvalue analysis.
Reference
· Mercer, C. D., B. D. Reddy, and R. A. Eve, "Finite Element Method for Piezoelectric Media,"
University of Cape Town/CSIR Applied Mechanics Research Unit Technical Report No. 92,
April 1987.
Table
Table 5.1.1-1 Piezoelectric transducer eigenvalue estimates.

Model Frequencies (kHz) for mode number
Type # of Elements 1 2 3 4 5
CAX4E 13 14.1 39.1 56.2 66.1 79.3
CAX4E 320 18.6 40.3 57.8 64.2 88.1
CAX6E 10 20.0 43.2 63.2 70.4 98.8
CAX8RE 5 19.6 42.8 61.0 66.9 96.3
5-1018
Electrical Analyses
CAX8RE 80 18.6 40.3 57.6 64.2 87.6

C3D8E 16 19.8 41.8 62.0 68.7 95.2
C3D20E 16 19.7 42.9 60.4 66.5 91.7
Experimental(1) 18.6 35.4 54.2 63.3 88.8
(1): Experimental results obtained from Mercer et al. (1987).
Figures
Figure 5.1.1-1 Piezoelectric transducer.
Figure 5.1.1-2 Meshes used with 4-node and 6-node axisymmetric elements.
5-1019
Electrical Analyses
Figure 5.1.1-3 Meshes used with 8-node axisymmetric and 8-node and 20-node three-dimensional
elements.
5-1020
Electrical Analyses
Figure 5.1.1-4 Mode shapes for 8-node axisymmetric elements.
Sample listings
5-1021
Electrical Analyses
Listing 5.1.1-1
*HEADING
EIGENVALUE ANALYSIS OF CYLINDRICAL TRANSDUCER
PIEZOELECTRIC CERAMIC AND BRASS END CAPS, COARSE MODEL WITH
FOUR-NODE AXISYMMETRIC PIEZOELECTRIC ELEMENTS
*NODE
1, .0110,.0
2, .01175, 0.0
3, .0125, 0.0
13, .0110, .0125
14, .01175,.0125
15, .0125,.0125
16, .0110, .0155
17, .01175,.0155
18, .0125,.0155
19, 0.0,.0125
21,.0073333,.0125
22, 0.0,.0155
24,.0073333,.0155
*NGEN
1,13,3
2,14,3
3,15,3
19,21
22,24
*ELEMENT,TYPE=CAX4E
1, 1,2,5,4
*ELEMENT,TYPE=CAX4,ELSET=BRASS
9, 13,14,17,16
10, 14,15,18,17
11, 19,20,23,22
12, 20,21,24,23
13, 21,13,16,24
*ELGEN,ELSET=PZT4
1,2,1,1,4,3,2
**local orientation matching global system
*ORIENTATION,NAME=RECT
1.0, 0.0, 0.0, 0.0, 1.0, 0.0
1, 0.0
*SOLID SECTION,MATERIAL=PZT4,ELSET=PZT4,ORIENT=RECT
*SOLID SECTION,MATERIAL=BRASS,ELSET=BRASS
*MATERIAL,NAME=PZT4
5-1022
Electrical Analyses
*ELASTIC,TYPE=ORTHOTROPIC
11.54E10,7.428E10,13.90E10,7.428E10,7.784E10,13.90E10,2.564E10,2.564E10
2.564E10,
*PIEZOELECTRIC,TYPE=S
15.08,-5.207,-5.207,0.,0.,0.,0.,0.
0.,12.710,0.,0.,0.,0.,0.,0.,
12.710,0.
*DIELECTRIC,TYPE=ANISO
5.872E-9,0.,6.752E-9,0.,0.,6.752E-9
*DENSITY
7500.,
*MATERIAL,NAME=BRASS
10.4E10,0.37
*DENSITY
8500.,
*NSET,NSET=IN,GENERATE
1,13,3
*NSET,NSET=SYM
1,2,3
*NSET,NSET=ASYM
19,22
*BOUNDARY
SYM,2
ASYM,1
IN,9
*RESTART,WRITE
**
*STEP,PERTURBATION
*FREQUENCY
5,100000.,10.
*EL PRINT
S,E
EFLX,EPG
*NODE PRINT
U,EPOT
RF,RCHG
*ELSET,ELSET=OUTPUT
1,9
*EL FILE,ELSET=OUTPUT
S,E,EFLX,EPG
*NSET,NSET=OUTPUT1
5,8
5-1023
Electrical Analyses
*NODE FILE,NSET=OUTPUT1
U,EPOT
*NSET,NSET=OUTPUT2
1,19
*NODE FILE,NSET=OUTPUT2
RF,RCHG
*OUTPUT,FIELD
*ELEMENT OUTPUT,ELSET=OUTPUT
S,E,EFLX,EPG
*OUTPUT,FIELD
*NODE OUTPUT
U,
*NODE OUTPUT,NSET=OUTPUT1
EPOT,
*OUTPUT,FIELD
*NODE OUTPUT,NSET=OUTPUT2
RF,RCHG
*END STEP
5.2 Joule heating analyses

5.2.1 Thermal-electrical modeling of an automotive fuse
Joule heating arises when the energy dissipated by electrical current flowing through a conductor is
converted into thermal energy. ABAQUS provides a fully coupled thermal-electrical procedure for
analyzing this type of problem.
An overview of the capability is provided in ``Coupled thermal-electrical analysis,'' Section 6.6.2 of
the ABAQUS/Standard User's Manual. This example illustrates the use of the capability to model the
heating of an automotive electrical fuse due to a steady 30 A electrical current. Fuses are the primary
circuit protection devices in automobiles. They are available in a range of different current ratings and
are designed so that, when the operating current exceeds the design current for a period of time,
heating due to electrical conduction causes the metal conductor to melt and--hence--the circuit to
disconnect.
A description of the original problem, as well as experimental measurements, can be found in Wang
and Hilali (1995). The experimental data and some of the material properties were refined subsequent
to this publication. These properties are used here, and the finite element results are compared with the
refined measurements (Hilali, July 1995).
Problem description
An automotive electrical fuse consists of a metal conductor, such as zinc, embedded within a
transparent plastic housing. The plastic housing, which only protects and supports the thin conductor,
5-1024
Electrical Analyses
is not represented in the finite element model. Figure 5.2.1-1 shows front and top sections of the
geometry of the conductor. It consists of two 0.76 mm thick blades, with an S-shaped fuse element
supported between the blades. The blades fit tightly into standard electrical terminals that are built into
the circuit and provide the connection between the electrical circuit and the fuse element. The fuse
element is usually much thinner than the fuse blades (in this case 0.28 mm thick) and is designed to
melt when the operating current exceeds the design current for a period of time. The fuse blades are 8
mm wide and 30.4 mm long. The fuse element is approximately 3.6 mm wide.
The model is discretized (see Figure 5.2.1-1) with 8-node first-order brick elements (element type
DC3D8E), using one element through the thickness. Two 6-node triangular prism elements (element
type DC3D6E) are used to fill regions where the geometry precludes the use of brick elements. No
mesh convergence studies have been performed.
The electrical conductivity of zinc varies linearly between 16.75 ´ 103 1=− mm at 20°C and 12.92 ´
1031=− mm at 100°C. The thermal conductivity varies linearly between 0.1120 W/mm°C at 20°C and
0.1103 W/mm°C at 100°C. The density is 7.14 ´ 10-6 kg/mm3, and the specific heat is 388.9 J/kg°C.
The *JOULE HEAT FRACTION option is used to specify the amount of electrical energy that is
converted into thermal energy. We assume that all electrical energy is converted into thermal energy.
The analysis is done in two steps. In the first step heating of the conductor due to current flow is
considered. Once steady-state conditions are reached, the current is switched off and the fuse is
allowed to cool down to the ambient temperature in a second step. During the first part of the analysis,
the coupled thermal-electrical equations are solved for both temperature and electrical potential at the
nodes using the *COUPLED THERMAL-ELECTRICAL procedure. In the subsequent cool down
period, since there is no longer any electric current in the fuse, an uncoupled *HEAT TRANSFER
analysis (`Ùncoupled heat transfer analysis,'' Section 6.5.2 of the ABAQUS/Standard User's Manual)
is performed. Input files illustrating both steady-state and transient analyses are provided. Steady-state
analysis is obtained by specifying the STEADY STATE parameter on the *COUPLED
THERMAL-ELECTRICAL procedure. No data lines are required. Transient analysis is available by
omitting the STEADY STATE parameter. In this analysis the DELTMX parameter is set to 20°C so
that automatic time incrementation is used. The END parameter is set to SS so that the analysis
terminates when steady-state conditions are reached. Steady state is defined here as the point at which
the temperature rate change is less than 0.1° C/s. This condition is defined on the data line following
the *COUPLED THERMAL-ELECTRICAL option. We specify a total analysis time of 100 s, with an
initial time step size of 0.1 s.
The electrical loading is a steady 30 A current. This is applied as a concentrated current on each of the
nodes on the bottom edge of the left-hand-side terminal. The *CECURRENT option is used for this
purpose. The *SECTION FILE option is used to output the total current and the total heat flux in a
section defined through the fuse element. The electrical potential (degree of freedom 9) is constrained
at the bottom edge of the right-hand-side blade by using a *BOUNDARY option (``Boundary
conditions,'' Section 19.3.1 of the ABAQUS/Standard User's Manual). This option is also used to keep
the bottom edges of the fuse blades at sink temperatures (degree of freedom 11) of 29.4°C and 30.2°C,
respectively.
It is assumed that the exposed metal surfaces lose heat through convection to an ambient temperature
5-1025
Electrical Analyses
of µ0 = 23.3°C. Heat loss from the thin edges is ignored. The film coefficient varies with temperature
according to the empirical relation
h = h0 (µ ¡ µ0 )1=4 ;
where µ is surface temperature (°C); h is the film coefficient (W/mm2°C); and h0 is a constant that
depends on the surface geometry--h0 =4.747 ´ 10-6 W/mm2 for the blade surfaces, and h0 =5.756 ´
10-6 W/mm2 for the fuse element surfaces. This dependence is entered as a table of film property values
using the *FILM PROPERTY option and is referred to on the *FILM option (see ``Thermal loads,''
Section 19.4.3 of the ABAQUS/Standard User's Manual).

Figure 5.2.1-2 shows a contour plot of the magnitude of the electrical current density vector at
steady-state conditions. Since the dissipated electrical energy--and, hence, the thermal energy--is a
function of current density, this figure represents contours of the heat generated. The figure indicates
that most of the heat is generated near the inside curves of the S-shaped fuse element and near the
center hole. The dissipated energy in the fuse blades is negligible compared to that in the fuse
element.
Figure 5.2.1-3 shows a contour plot of the temperature distribution at the end of the first analysis step.
The maximum temperature is reached near the center of the S-shaped fuse element. This area is
expected to fail first when the operating current exceeds the design current. Figure 5.2.1-4 compares
the temperatures at the measuring positions (defined in Figure 5.2.1-1) with the experimental
measurements (Hilali, July 1995). While the results show some discrepancies between the experiment
and analysis, it is clear that the analysis is sufficiently representative that it provides a useful basis for
studying such systems. Figure 5.2.1-5 shows the variation of temperature at measuring position 6
during the heating and subsequent cool-down periods.
Acknowledgments
Mr. Hilali and Dr. Wang of Delphi Packard Electric Systems supplied the geometry of the fuse,
material properties, and experimental results. Delphi Packard assumes no responsibility for the
accuracy of the analysis method or data contained in the analysis.
Input files
thermelectautofuse_steadystate.inp
Steady-state analysis.
thermelectautofuse_transient.inp
Transient analysis.
thermelectautofuse_node.inp
Nodal coordinates for the model.
5-1026
Electrical Analyses
thermelectautofuse_element.inp
Element definitions.
thermelectautofuse_controls.inp
Identical to thermelectautofuse_steadystate.inp, except that it uses the *CONTROLS option for
control of convergence criteria.
References
· Hilali, S. Y., Private communication, July 1995.
· Hilali, S. Y., and B.-J. Wang, "ABAQUS Thermal Modeling for Electrical Assemblies," 1995
ABAQUS Users' Conference, Paris, May 1995, pp. 441-457.
· Wang, B.-J., and S. Y. Hilali, "Electrical-Thermal Modeling Using ABAQUS," 1995 ABAQUS
Users' Conference, Paris, May 1995, pp. 771-785.
Figures
Figure 5.2.1-1 Geometry and finite element discretization.
Figure 5.2.1-2 Contours of the magnitude of the current density vector (A/mm 2).
5-1027
Electrical Analyses
Figure 5.2.1-3 Contours of temperature field (°C).
Figure 5.2.1-4 Temperature (°C) at measuring positions.
5-1028
Electrical Analyses
Figure 5.2.1-5 Variation of temperature (°C) at measuring position 6 with time (s).
Sample listings
5-1029
Electrical Analyses
Listing 5.2.1-1
*HEADING
HEATING OF AUTOMOTIVE ELECTRICAL FUSE
STEADY STATE ANALYSIS
Units: mm; deg. C; Joule; sec; kg; Ampere
**
*NODE, INPUT=thermelectautofuse_node.inp
*NSET, NSET=NLHS
3, 6, 9, 12, 15, 18, 21, 24, 27, 30
*NSET, NSET=NRHS
420, 423, 426, 429, 432, 435, 438, 441, 444, 447
*NSET, NSET=NOUTP
30, 51, 83, 200, 319, 335, 360, 225, 604, 500,
468, 447
*ELEMENT, TYPE=DC3D8E, ELSET=EALL,
INPUT=thermelectautofuse_element.inp
*ELEMENT, TYPE=DC3D6E, ELSET=EALL
135, 381, 98, 96, 382, 104, 102
138, 383, 372, 373, 384, 375, 376
*ELSET,ELSET=ELEMENT,GENERATE
77, 149
226, 227
*ELSET,ELSET=BLADES,GENERATE
1, 76
150, 225
*SOLID SECTION, ELSET=EALL, MATERIAL=ZINC
*SURFACE,NAME=SIDE
88,S5
87,S5
*MATERIAL, NAME=ZINC
*CONDUCTIVITY
0.1121, 20.0
0.1103, 100.0
*ELECTRICAL CONDUCTIVITY
16.75E3, 20.0
12.92E3, 100.0
*JOULE HEAT FRACTION
1.0,
*FILM PROPERTY,NAME=H1
6.57E-6, 25.0
11.6E-6, 40.0
5-1030
Electrical Analyses
14.2E-6, 60.0
15.8E-6, 80.0
17.0E-6,100.0
5.42E-6, 25.0
9.60E-6, 40.0
11.7E-6, 60.0
13.0E-6, 80.0
14.1E-6,100.0
*STEP,INC=100
STEP 1: JOULE HEATING ANALYSIS
*COUPLED THERMAL-ELECTRICAL,STEADY STATE
*BOUNDARY
NRHS, 9, , 0.0
NLHS, 11,, 29.4
NRHS, 11,, 30.2
*CECURRENT
NLHS, 9, 3.0
*FILM
ELEMENT, F1, 23.3, H1
ELEMENT, F2, 23.3, H1
BLADES, F1, 23.3, H2
BLADES, F2, 23.3, H2
*EL PRINT,ELSET=ELEMENT,FREQUENCY=100
JENER, HFLM, ECDM, EPGM
*EL FILE,ELSET=ELEMENT,FREQUENCY=100
HFL, ECD, EPG
*NODE PRINT,NSET=NOUTP,FREQUENCY=100
NT,EPOT
*NODE FILE,NSET=NOUTP,FREQUENCY=100
NT,EPOT
*SECTION PRINT,NAME=SID88,SURFACE=SIDE
SOAREA,SOH,SOE
*SECTION FILE, NAME=SID88,SURFACE=SIDE
SOAREA,SOH,SOE
*END STEP
*STEP,INC=100
STEP 2: COOL DOWN ANALYSIS
*HEAT TRANSFER,STEADY STATE
*END STEP
5-1031
Electrical Analyses
Listing 5.2.1-2
*HEADING
HEATING OF AUTOMOTIVE ELECTRICAL FUSE
TRANSIENT ANALYSIS
Units: mm; deg. C; Joule; sec; kg; Ampere
*NODE, INPUT=thermelectautofuse_node.inp
*NSET, NSET=NLHS
3, 6, 9, 12, 15, 18, 21, 24, 27, 30
*NSET, NSET=NRHS
420, 423, 426, 429, 432, 435, 438, 441, 444, 447
*NSET, NSET=NOUTP
30, 51, 83, 200, 319, 335, 360, 225, 604, 500,
468, 447
*ELEMENT, TYPE=DC3D8E, ELSET=EALL,
INPUT=thermelectautofuse_element.inp
*ELEMENT, TYPE=DC3D6E, ELSET=EALL
135, 381, 98, 96, 382, 104, 102
138, 383, 372, 373, 384, 375, 376
*ELSET,ELSET=ELEMENT,GENERATE
77, 149
226, 227
*ELSET,ELSET=BLADES,GENERATE
1, 76
150, 225
*SOLID SECTION, ELSET=EALL, MATERIAL=ZINC
*MATERIAL, NAME=ZINC
*CONDUCTIVITY
0.1121, 20.0
0.1103, 100.0
*ELECTRICAL CONDUCTIVITY
16.75E3, 20.0
12.92E3, 100.0
*JOULE HEAT FRACTION
1.0,
*DENSITY
7.14E-6,
*SPECIFIC HEAT
389.0,
6.57E-6, 25.0
11.6E-6, 40.0
5-1032
Electrical Analyses
14.2E-6, 60.0
15.8E-6, 80.0
17.0E-6,100.0
5.42E-6, 25.0
9.60E-6, 40.0
11.7E-6, 60.0
13.0E-6, 80.0
14.1E-6,100.0
*SURFACE, NAME=SURF_1
ELEMENT, S1
ELEMENT, S2
BLADES, S1
BLADES, S2
*STEP,INC=100
STEP 1: JOULE HEATING ANALYSIS
*COUPLED THERMAL-ELECTRICAL,DELTMX=20.0,END=SS
0.1, 100.0, , ,0.1
*BOUNDARY
NRHS, 9, , 0.0
NLHS, 11,, 29.4
NRHS, 11,, 30.2
*CECURRENT
NLHS, 9, 3.0
*SFILM
SURF_1, F, 23.3, H1
SURF_2, F, 23.3, H1
SURF_3, F, 23.3, H2
SURF_4, F, 23.3, H2
*ENERGY PRINT,FREQUENCY=100,ELSET=BLADES
*ENERGY PRINT,FREQUENCY=100,ELSET=ELEMENT
*ENERGY FILE,FREQUENCY=100,ELSET=BLADES
*ENERGY FILE,FREQUENCY=100,ELSET=ELEMENT
*EL PRINT,ELSET=ELEMENT,FREQUENCY=100
JENER, HFLM, ECDM, EPGM
*EL FILE,ELSET=ELEMENT,FREQUENCY=100
HFL, ECD, EPG
*NODE PRINT,NSET=NOUTP,FREQUENCY=100
5-1033
Electrical Analyses
NT,EPOT
*NODE FILE,NSET=NOUTP,FREQUENCY=100
NT,EPOT
*END STEP
*STEP,INC=100
STEP 2: COOL DOWN ANALYSIS
*HEAT TRANSFER,DELTMX=20.0,END=SS
0.1, 100.0, , ,0.1
*END STEP
5-1034
Mass Diffusion Analyses
6. Mass Diffusion Analyses

6.1 Mass diffusion analyses
6.1.1 Hydrogen diffusion in a vessel wall section
This one-dimensional problem provides a simple verification of the mass diffusion capability in
ABAQUS. The uncoupled mass diffusion formulation used in ABAQUS is described in ``Mass
diffusion analysis,'' Section 6.8.1 of the ABAQUS/Standard User's Manual, and ``Mass diffusion
The physical problem considered here is that of a pressure vessel shell wall fabricated from 2 1/4 Cr-1
Mo steel alloy base metal with an internal weld overlay of Type 347 stainless steel. These vessels are
typically used at high temperatures and under high pressure conditions. Under such service conditions
hydrogen dissolves into the alloys (Fujii et al., 1982) and during cool-down may cause disbonding of
the weld overlay from the base metal and, possibly, crack initiation and growth in the base metal due to
hydrogen embrittlement. In this example we are concerned with the hydrogen diffusion aspect of the
problem.
Problem description
The problem is shown in Figure 6.1.1-1 and consists of a section of the vessel wall made up of a
200-mm thick base metal and a 5-mm thick weld metal. The problem is one-dimensional, the only
gradient being through the thickness of the wall. The purpose of the analysis is to predict the evolution
of hydrogen concentration through the wall thickness during cooling caused by a shutdown.
Geometry and model

Since the problem is one-dimensional, we use a plane mesh with only one element in the y-direction
(see Figure 6.1.1-2). The mesh is graded, with more elements near the interface between the two
materials because we expect very high concentration gradients in this vicinity.
The material properties of the two metals given by Fujii et al. (1982) are strongly dependent on
temperature and can be written as follows.
Solubility in weld metal:
~
sw = 1288 e¡1078=µ ppm mm N¡1=2
Diffusivity in weld metal:
~
Dw = 9310 e¡6767=µ mm2 =h
Solubility in base metal:
6-1035
~
sb = 4300 e¡3261=µ ppm mm N¡1=2
Diffusivity in base metal:
~
274 e¡1157=µ
Db = mm2 =h;
1 + (1:05 £ 10¡3 e3573=µ~ )
where µ~ is temperature in degrees Kelvin. These temperature-dependent properties are entered in

ABAQUS in tabulated form, as shown in the input listings.
The wall is initially at a uniform temperature of 727.5 K (454.4° C), and during the shutdown schedule
it cools down to 298.15 K (25.0° C) at a constant rate over a period of 21.5 hours.
The boundary conditions are as follows. Under the initial steady-state conditions the exterior of the
weld metal has a hydrogen concentration of 35.85 ppm, which corresponds to a normalized
concentration of 0.1225 N 1/2mm-1. Normalized concentration is used as the primary solution variable
(continuous over the discretized domain) and is given as the concentration divided by the solubility.
The exterior of the base metal has a zero hydrogen concentration. As the cooling period begins, the
hydrogen concentration at the exterior of the weld metal is assumed to drop to zero instantaneously.
Time stepping
The problem is run in two parts. The first part consists of a step in which a single increment of *MASS
DIFFUSION, STEADY STATE analysis is performed with an arbitrary time step to establish the initial
steady-state hydrogen concentration distribution corresponding to the initial temperature.
The hydrogen diffusion during cooling is then analyzed in four subsequent *MASS DIFFUSION
transient analysis steps, using automatic time stepping. This need not be done in four separate steps.
We do it here because the results given by Fujii et al. (1982), with which we compare the ABAQUS
results, are presented at four specific times during the transient: 2.7 h (673.15 K, 400.0°C), 5.2 h
(623.15 K, 350.0°C), 10.2 h (523.15 K, 250.0°C), and 21.5 h (298.15 K, 25.0°C).
The accuracy of the time integration for the *MASS DIFFUSION transient analysis steps, during
which cooling occurs, is controlled by the DCMAX parameter. This parameter specifies the allowable
normalized concentration change per time step. Even in a linear problem such as this, DCMAX
controls the accuracy of the solution because the time integration operator is not exact (the backward
difference rule is used). In this case DCMAX is chosen as 0.01 N 1/2mm-1, which is a very tight value.
This is necessary to obtain an acceptably accurate integration of the concentration because the
solubility of the materials decreases significantly (by more than two orders of magnitude in the base
metal) as the temperature decreases and, therefore, the changes in concentration become larger for a
given change in normalized concentration.
An important issue in transient diffusion problems is the choice of initial time step. As in any transient
problem, the spatial element size and the time step are related to the extent that time steps smaller than
a certain size may lead to spurious oscillations in the solution and, therefore, provide no useful
information. This coupling of the spatial and temporal approximations is always most obvious at the
6-1036
start of diffusion problems, immediately after prescribed changes in the boundary values. For the mass
diffusion case the suggested guideline for choosing the initial time increment (see ``Mass diffusion
analysis,'' Section 6.8.1 of the ABAQUS/Standard User's Manual) is
1
¢t ¸ (¢h)2 ;
6D
where ¢h is a characteristic element size near the disturbance (that is, near the weld metal surface in
our case), and D is the diffusivity of the material. For the weld metal in our model we choose a typical
¢h = 0.125 mm and we have D = 0.85 mm2/h at the initial temperature, which gives ¢t ¸ 0.003 h.
For the base metal in our model we choose a typical ¢h = 1.25 mm and we have D = 4.88 mm2/h at
the initial temperature, which gives ¢t ¸ 0.053 h. Based on these calculations an initial time step of
0.1 h is used, which gives an initial solution with no oscillations, as expected.

Figure 6.1.1-3 shows hydrogen concentration distributions in the weld metal for the initial steady-state
condition and four different times during the cooling period. Figure 6.1.1-4 shows corresponding
hydrogen concentration distributions in the base metal. These results compare very well with those
presented by Fujii et al. (1982) which are not plotted here since they would appear almost
indistinguishable from the ABAQUS results.
It can be observed that, although the primary solution variable (the normalized concentration) remains
continuous across the material interface during the transient, the hydrogen concentration becomes
increasingly discontinuous across the interface. During the cooling process the hydrogen concentration
in the base metal decreases, whereas the hydrogen concentration in the weld metal increases very
significantly, reaching a peak at the weld metal side of the interface.
Input files
hydrodiffvesselwall_2d.inp
Two-dimensional analysis.
hydrodiffvesselwall_3d.inp
Three-dimensional analysis.
hydrodiffvesselwall_fick.inp
Two-dimensional analysis using Fick's law.
hydrodiffvesselwall_nonlinear.inp
Nonlinear version (including concentration dependence on the material properties) of the
two-dimensional analysis.
hydrodiffvesselwall_heat.inp
Heat transfer analysis that writes temperatures to a results file for use in
hydrodiffvesselwall_massdiff.inp.
6-1037
hydrodiffvesselwall_massdiff.inp
Two-dimensional mass diffusion analysis which reads temperatures from the results file written in
hydrodiffvesselwall_heat.inp.
Reference
· Fujii, T., T. Nazama, H. Makajima, and R. Horita, "A Safety Analysis on Overlay Disbonding of
Pressure Vessels for Hydrogen Service," Journal of the American Society for Metals, pp.
361-368, 1982.
Figures
Figure 6.1.1-1 Pressure vessel shell wall section.
Figure 6.1.1-2 Finite element model of shell wall.
6-1038
Figure 6.1.1-3 Hydrogen concentration distribution in weld metal.
Figure 6.1.1-4 Hydrogen concentration distribution in base metal.
Sample listings
6-1039
Listing 6.1.1-1
*HEADING
HYDROGEN DIFFUSION OF A SHELL SECTION
*NODE
1,
2001,,10.
41,5.,
2041,5.,10.
121,205.
2121,205.,10.
*NGEN,NSET=LHEND
1,2001,1000
*NGEN,NSET=MID
41,2041,1000
*NGEN,NSET=RHEND
121,2121,1000
*NFILL,NSET=NALL,TWO STEP,BIAS=1.125
LHEND,MID,40,1
*NFILL,NSET=NALL,TWO STEP,BIAS=.875
MID,RHEND,80,1
*ELEMENT,TYPE=DC2D8
1,1,3,2003,2001,2,1003,2002,1001
*ELGEN,ELSET=WELD
1,20,2,2
*ELEMENT,TYPE=DC2D8
41,41,43,2043,2041,42,1043,2042,1041
*ELGEN,ELSET=BASE
41,40,2,2
*ELSET,ELSET=ELS1
39,41
*SOLID SECTION,MATERIAL=WELD,ELSET=WELD
1.,
*MATERIAL,NAME=WELD
*SOLUBILITY
2.48858D+01, 2.73150D+02
3.46472D+01, 2.98150D+02
4.58298D+01, 3.23150D+02
5.82346D+01, 3.48150D+02
7.16596D+01, 3.73150D+02
8.59120D+01, 3.98150D+02
1.00815D+02, 4.23150D+02
6-1040
1.16210D+02, 4.48150D+02
1.31960D+02, 4.73150D+02
1.47945D+02, 4.98150D+02
1.64063D+02, 5.23150D+02
1.80229D+02, 5.48150D+02
1.96372D+02, 5.73150D+02
2.12432D+02, 5.98150D+02
2.28360D+02, 6.23150D+02
2.44118D+02, 6.48150D+02
2.59672D+02, 6.73150D+02
2.74998D+02, 6.98150D+02
2.90077D+02, 7.23150D+02
2.92704D+02, 7.27550D+02
*DIFFUSIVITY
1.62094D-07,, 2.73150D+02
1.29398D-06,, 2.98150D+02
7.49035D-06,, 3.23150D+02
3.36944D-05,, 3.48150D+02
1.23910D-04,, 3.73150D+02
3.86932D-04,, 3.98150D+02
1.05615D-03,, 4.23150D+02
2.57731D-03,, 4.48150D+02
5.72350D-03,, 4.73150D+02
1.17322D-02,, 4.98150D+02
2.24546D-02,, 5.23150D+02
4.05055D-02,, 5.48150D+02
6.94021D-02,, 5.73150D+02
1.13680D-01,, 5.98150D+02
1.78977D-01,, 6.23150D+02
2.72085D-01,, 6.48150D+02
4.00960D-01,, 6.73150D+02
5.74695D-01,, 6.98150D+02
8.03461D-01,, 7.23150D+02
8.50242D-01,, 7.27550D+02
*SOLID SECTION,MATERIAL=BASE,ELSET=BASE
1.,
*MATERIAL,NAME=BASE
*SOLUBILITY
1.78188D-01, 3.23150D+02
3.67776D-01, 3.48150D+02
6.88842D-01, 3.73150D+02
1.19243D-00, 3.98150D+02
1.93457D-00, 4.23150D+02
6-1041
2.97365D-00, 4.48150D+02
4.36782D-00, 4.73150D+02
6.17278D-00, 4.98150D+02
8.43995D-00, 5.23150D+02
1.12152D+01, 5.48150D+02
1.45379D+01, 5.73150D+02
1.84407D+01, 5.98150D+02
2.29491D+01, 6.23150D+02
2.80820D+01, 6.48150D+02
3.38515D+01, 6.73150D+02
4.02639D+01, 6.98150D+02
4.73200D+01, 7.23150D+02
4.86282D+01, 7.27550D+02
*DIFFUSIVITY
7.85674D-03,, 2.73150D+02
3.34289D-02,, 2.98150D+02
1.13043D-01,, 3.23150D+02
3.17681D-01,, 3.48150D+02
7.65060D-01,, 3.73150D+02
1.61326D-00,, 3.98150D+02
3.02697D-00,, 4.23150D+02
5.12273D-00,, 4.48150D+02
7.92165D-00,, 4.73150D+02
1.13403D+01,, 4.98150D+02
1.52239D+01,, 5.23150D+02
1.93975D+01,, 5.48150D+02
2.37067D+01,, 5.73150D+02
2.80357D+01,, 5.98150D+02
3.23085D+01,, 6.23150D+02
3.64808D+01,, 6.48150D+02
4.05306D+01,, 6.73150D+02
4.44494D+01,, 6.98150D+02
4.82372D+01,, 7.23150D+02
4.88906D+01,, 7.27550D+02
NALL,727.55
*INITIAL CONDITIONS,TYPE=CONCENTRATION
NALL,.1
*AMPLITUDE,NAME=AMP1,TIME=TOTAL TIME,
VALUE=ABSOLUTE
0.,727.55,1.E-5,727.55,21.5,298.15,1000.,298.15
*STEP
STEADY-STATE SOLUTION UNDER OPERATING CONDITIONS
6-1042
*MASS DIFFUSION,STEADY STATE

1.E-5,1.E-5
*BOUNDARY
LHEND,11,11,.12247
RHEND,11,11,0.
CONC,MFL,TEMP
MFL1,MFL2,MFLM
*EL PRINT,FREQUENCY=1000,
POSITION=AVERAGED AT NODES,ELSET=WELD
CONC,
*EL PRINT,FREQUENCY=1000,
POSITION=AVERAGED AT NODES,ELSET=BASE
CONC,
NFLUX,
NNC,NT,RFL
*EL FILE,ELSET=ELS1,FREQUENCY=1000
CONC,MFL
*EL FILE,ELSET=ELS1,POSITION=AVERAGED AT NODES,
FREQUENCY=1000
CONC,
NFLUX,
NNC,RFL
*END STEP
*STEP,INC=100
TRANSIENT ANALYSIS DUE TO SHUTDOWN -- PHASE 1
*MASS DIFFUSION,DCMAX=.01, END=PERIOD
.1,2.7,,,.0001
*BOUNDARY
LHEND,11,11,0.
*TEMPERATURE,AMPLITUDE=AMP1
NALL,
*END STEP
*STEP,INC=100
*MASS DIFFUSION,DCMAX=.01
.1,2.5,,,.0001
*BOUNDARY
6-1043
LHEND,11,11,0.
NALL,
*END STEP
*STEP,INC=100
.1,5.0,,,.0001
*BOUNDARY
LHEND,11,11,0.
NALL,
*END STEP
*STEP,INC=1000
.1,11.3,,,.0001
*BOUNDARY
LHEND,11,11,0.
NALL,
*END STEP
6.1.2 Diffusion toward an elastic crack tip

This simple two-dimensional problem verifies the sequentially coupled, stress-assisted mass diffusion
capability in ABAQUS. The mass diffusion formulation used in ABAQUS is described in ``Mass
diffusion analysis,'' Section 6.8.1 of the ABAQUS/Standard User's Manual, and ``Mass diffusion
A center-cracked plate fabricated from 2 1/4 Cr-1 Mo steel alloy is subjected to end loading in a
hydrogen-rich environment. Hydrogen is drawn to the crack-tip region by high hydrostatic stresses and
may assist in crack growth resulting from hydrogen embrittlement. In this example we are concerned
with the hydrogen diffusion aspect of the problem.
Geometry and model

The problem geometry and boundary conditions are shown in Figure 6.1.2-1. The specimen is 10-mm
thick, 20-mm wide, and 80-mm high, with a 4-mm crack at its center. The mesh near the crack is
focused at the crack tip, with the element size growing as the square of the distance to the crack tip
(*NFILL, SINGULAR=1). A very fine mesh (see Figure 6.1.2-2) is used to capture accurately the
gradients of concentration and stress near the crack tip. Four combinations of stress and mass diffusion
analyses are presented:
6-1044
· Stress analysis with quadratic elements and quarter-point spacing at the crack tip, followed by a
mass diffusion analysis with linear elements.
· Stress analysis with quadratic elements and quarter-point spacing at the crack tip, followed by a
mass diffusion analysis with quadratic elements and quarter-point spacing at the crack tip.
· Stress analysis with quadratic elements (no quarter-point spacing), followed by a mass diffusion
analysis with quadratic elements (no quarter-point spacing).
· Stress analysis with linear elements, followed by a mass diffusion analysis with linear elements.
p
The quarter-point spacing technique is used in fracture mechanics analyses to enforce a 1= r
singularity at the crack tip, where r is the distance from the crack tip.
The sequentially coupled mass diffusion analysis consists of a static stress analysis, followed by a mass
diffusion analysis. Equivalent pressure stresses from the static analysis are written to the results file as
nodal averaged values. Subsequently, these pressures are read in during the course of the mass
diffusion analysis to provide a driving force for mass diffusion.
The material properties for mass diffusion given by Fujii et al. (1982) are as follows.
Solubility:
z
s = 4300 e¡3261=(µ¡µ )
ppm mm N¡1=2
Diffusivity:
z
7611 £ 10¡5 e¡1157=(µ¡µ )
D= mm2 =s;
1 + (1:05 £ 10¡3 e3573=(µ¡µz ) )
where µ is the temperature in degrees Celsius and µz = -273 is the absolute zero temperature.
Stress-assisted diffusion is specified by defining the pressure stress factor, ∙p ; as
V HÁ
∙p = mm N¡1=2 ;
R(µ ¡ µz )
where R =8.31432 Jmol -1K-1 is the universal gas constant, V H =2.0 ´ 103mm3mol-1 is the partial
molar volume of hydrogen in iron-based metals, and Á is the normalized concentration. The
concentration dependence of ∙p is entered in ABAQUS in tabulated form as shown in the input
listings. It is important to note that although ∙p is defined in terms of normalized concentration, Á, the
tabular data must be entered in terms of concentration, c = Ás:
The following properties are also used in the stress analysis: elastic modulus, E =2.0 ´ 105Nmm-2, and
Poisson's ratio, º = 0.3.
The specimen is maintained at a constant temperature of µ ¡ µz = 325 K throughout the analysis.
Under the initial steady-state conditions the specimen has a uniform concentration of 50 ppm, which
6-1045
corresponds to a normalized concentration of 265 N 1/2mm-1. Normalized concentration is used as the

primary solution variable (continuous over the discretized domain) and is given as the concentration
divided by the solubility. The exterior of the specimen has a constant hydrogen concentration equal to
the initial concentration. A 1 MPa distributed pressure is applied to the ends of the specimen, ramped
linearly over the length of the step, and the steady-state distribution of hydrogen is obtained.

The analytical solution for normalized concentration, presented by Liu (1970), has the form
µ ¶
VH p
Á = Áo exp ¡ ;
R(µ ¡ µz )
where Áo is the normalized concentration obtained in the unstressed state and p is the equivalent
pressure stress. This solution dictates that for a crack-tip problem, the concentration follows the
singularity of the stresses.
Figure 6.1.2-3 and Figure 6.1.2-4show the final distribution of equivalent pressure stress and
concentration predicted by the ABAQUS analysis in the region around the crack tip. The results shown
represent the first case described above, using a quadratic, quarter-point mesh for stresses and a linear
mesh for mass diffusion. The shapes of the contours show good agreement, since contours of constant
pressure stress should be contours of constant concentration, as indicated by the analytical solution
above.
Figure 6.1.2-5 and Figure 6.1.2-6show the pressures (in MPa) and concentrations (in ppm) ahead of
the crack tip for all four combinations of stress and mass diffusion analyses. Results are presented as
functions of the ratio of the distance to the crack tip, r, over the crack length, a. For the region
immediately ahead of the crack, linear elastic fracture mechanics yields the analytical solution for
equivalent pressure stress:
KI (1 + º ) (1 + º )¾
p=¡ p ¡ ;
2¼r 3
p
where KI = ¾ ¼a is the stress intensity factor for a Mode I crack of length a and ¾ is the externally
applied distributed load.
As can be seen from the figures, the finite element results for all four combinations of element types
are identical except at the first element, where the results are not expected to be valid. The results
show good agreement with the analytically predicted solutions for both equivalent pressure stress and
concentration as the distance to the crack tip, r, approaches zero. Farther from the crack tip, the
deviation between the analytical solution and the finite element solution increases. This deviation is
consistent with the fact that the linear elastic crack-tip solution is valid only as r approaches zero.
No mesh convergence studies were conducted with respect to the number of elements in the crack-tip
region. For comparison with the solutions presented here, an analysis was conducted with equally
spaced elements approaching the crack tip. The results (not shown here) indicate that biasing the
elements toward the crack tip is necessary to capture the gradients of concentration and equivalent
6-1046
pressure stress adequately. In addition, the equivalent pressure stress results demonstrate that the effect
of using quarter-point positioning of the nodes at the crack tip is insignificant in this problem as long
as the mesh is refined sufficiently.
Differences between the finite element and analytically predicted concentrations are a direct result of
the differences between the finite element and analytically predicted values of pressure stress. If the
analytical values of equivalent pressure stress are used to drive the ABAQUS concentration solution,
the resulting curve is indistinguishable from the analytical concentration shown.
Input files
difftocrack_quarterpstress.inp
Quadratic stress analysis with quarter-point spacing at the crack tip. This analysis writes the results
file used in difftocrack_linearmassdiff1.inp and difftocrack_quarterpmassdiff.inp.
difftocrack_linearmassdiff1.inp
Linear mass diffusion analysis that reads results file data from difftocrack_quarterpstress.inp.
difftocrack_stress.inp
Stress analysis with quadratic elements (no quarter-point spacing). This analysis writes the results
file used in difftocrack_massdiff.inp.
difftocrack_massdiff.inp
Mass diffusion analysis with quadratic elements that reads equivalent pressure stresses from the
results file written in difftocrack_stress.inp.
difftocrack_quarterpmassdiff.inp
Mass diffusion analysis with quadratic elements and quarter-point spacing. This analysis reads
equivalent pressure stresses from the results file written in difftocrack_quarterpstress.inp.
difftocrack_linearstress.inp
Stress analysis with linear elements. This analysis writes the results file used in
difftocrack_linearmassdiff2.inp.
difftocrack_linearmassdiff2.inp
Mass diffusion analysis with linear elements that reads equivalent pressure stresses from the
results file written in difftocrack_linearstress.inp.
difftocrack_node.inp
Node data for all the analyses.
difftocrack_quad_elements.inp
Element data for the analyses using quadratic elements.
difftocrack_linear_elements.inp
Element data for the analyses using linear elements.
6-1047
References
· Fujii, T., T. Nazama, H. Makajima, and R. Horita, "A Safety Analysis on Overlay Disbonding of
Pressure Vessels for Hydrogen Service," Journal of the American Society for Metals, pp.
361-368, 1982.
· Liu, H. W., "Stress-Corrosion Cracking and the Interaction Between Crack-Tip Stress Field and
Solute Atoms," Transactions of the ASME: Journal of Basic Engineering, vol. 92, pp. 633-638,
1970.
Figures
Figure 6.1.2-1 Center crack specimen geometry.
Figure 6.1.2-2 Finite element model of center crack specimen (with 1/4 symmetry) with detail of
crack-tip mesh.
6-1048
Figure 6.1.2-3 Contours of equivalent pressure stress at the crack tip.
Figure 6.1.2-4 Contours of normalized hydrogen concentration at the crack tip.
6-1049
Figure 6.1.2-5 Distribution of pressure stress ahead of the crack tip.
Figure 6.1.2-6 Hydrogen concentration distribution ahead of the crack tip.
6-1050
Sample listings
6-1051
Listing 6.1.2-1
*HEADING
Quadratic stress analysis, singular elements
at crack tip
MODEL BUILT IN mm,N,s
**
*PREPRINT,ECHO=NO,MODEL=NO,HIST=NO
*NODE,NSET=ALL,INPUT=difftocrack_node.inp
**
**
*ELEMENT, TYPE=CPE8R, ELSET=ALL,
INPUT=difftocrack_quad_elements.inp
**
**
*NODE,NSET=TIP
1,2.0,0.
81,2.0,0.
*NGEN,NSET=TIP
1,81,1
**
** tipbound
**
*NSET, NSET=TIPBOUND
5001, 5002, 5003, 5004, 5005,
5006, 5007, 5008, 5009, 5010,
5011, 5012, 5013, 5014, 5015,
5016, 5017, 5018, 5019, 5020,
5021, 5022, 5023, 5024, 5025,
5026, 5027, 5028, 5029, 5030,
5031, 5032, 5033, 5034, 5035,
5036, 5037, 5038, 5039, 5040,
5041, 5042, 5043, 5044, 5045,
5046, 5047, 5048, 5049, 5050,
5051, 5052, 5053, 5054, 5055,
5056, 5057, 5058, 5059, 5060,
5061, 5062, 5063, 5064, 5065,
5066, 5067, 5068, 5069, 5070,
5071, 5072, 5073, 5074, 5075,
5076, 5077, 5078, 5079, 5080,
5081,
*NFILL,NSET=TIPMESH,SINGULAR=1,TWOSTEP
TIP,TIPBOUND,50,100
6-1052
*ELEMENT,TYPE=CPE8R,ELSET=ALL
2001,1,201,203,3,101,202,103,2
*ELGEN,ELSET=ALL
2001,25,200,50,40,2,1
**
** front
**
1,5001,100
*NSET, NSET=FRONT
5001, 5522, 5533, 5554, 5565,
5586, 5597, 5618, 5629, 5650,
5661, 5682, 5693, 5714, 5725,
5746, 5757, 5778, 5789, 5810,
5821, 6802, 6813, 6834, 6845,
6866, 6877, 6898, 6909, 6930,
6941, 6962, 6973, 6994, 7005,
7026, 7037, 7058, 7069, 7090,
7101, 7122, 7133, 7154, 7165,
7186, 7197, 7218, 7229, 7250,
7261, 7282, 7293, 7314, 7325,
7346, 7357, 7378, 7389, 7410,
7421,
**
** RIGHT
**
*NSET, NSET=RIGHT
7421, 7422, 7423, 7424, 7425,
7426, 7427, 7428, 7429, 7430,
7431, 7432, 7433, 7434, 7435,
7436, 7437, 7438, 7439, 7440,
7441, 8642, 8653, 8674, 8685,
8706, 8717, 8738, 8749, 8770,
8781, 8802, 8813, 8834, 8845,
8866, 8877, 8898, 8909, 8930,
8941, 8962, 8973, 8994, 9005,
9026, 9037, 9058, 9069, 9090,
9101, 9122, 9133, 9154, 9165,
9186, 9197, 9218, 9229, 9250,
9261, 9282, 9293, 9314, 9325,
9346, 9357
**
** left
6-1053
**
*NSET, NSET=LEFT
6481, 6782, 6783, 6784, 6785,
6786, 6787, 6788, 6789, 6790,
6791, 6792, 6793, 6794, 6795,
6796, 6797, 6798, 6799, 6800,
6801, 8051, 8071, 8081, 8101,
8111, 8131, 8141, 8161, 8171,
8191, 8201, 8221, 8231, 8251,
8261, 8281, 8291, 8311, 8321,
8341, 8351, 8371, 8381, 8401,
8411, 8431, 8441, 8461, 8471,
8491, 8501, 8521, 8531, 8551,
8561, 8581, 8591, 8611, 8621,
8641, 9387, 9392, 9402, 9407,
9417, 9422, 9432, 9437, 9447,
9452, 9462, 9467, 9477, 9482,
9492, 9497, 9507, 9512, 9522,
9527, 9537, 9542, 9552, 9557,
9567, 9572, 9582, 9587, 9597,
9602, 9612, 9617, 9627, 9632,
9642, 9647, 9657, 9662, 9672,
9677, 9687, 9688, 9689, 9690,
9691, 9692, 9745, 9746
**
** crack
**
*NSET,NSET=CRACK,GEN
81,5081,100
*NSET, NSET=CRACK
5081, 6511, 6531, 6541, 6561,
6571, 6591, 6601, 6621, 6631,
6651, 6661, 6681, 6691, 6711,
6721, 6741, 6751, 6771, 6781,
6801,
**
** top
**
*NSET, NSET=TOP
9357, 9358, 9359, 9360, 9361,
9362, 9363, 9364, 9365, 9366,
9367, 9368, 9369, 9370, 9371,
9372, 9373, 9374, 9375, 9376,
6-1054
9377, 9678, 9679, 9680, 9681,

9682, 9683, 9684, 9685, 9686,
9687,
**
** top
**
*ELSET, ELSET=TOP
1221, 1222, 1223, 1224, 1225,
1226, 1227, 1228, 1229, 1230,
1326, 1327, 1328, 1329, 1330
**
** all
**
*SOLID SECTION, ELSET=ALL, MATERIAL=STEEL
1.,
**
** steel
**
**
**
*ELASTIC, TYPE=ISO
2.E+5, 0.3
*BOUNDARY
LEFT,1,1
FRONT,2,2
*STEP
*STATIC
0.1,1.0
*DLOAD
TOP,P2,-1
*EL PRINT,FREQ=0,position=averaged at nodes
press,
*NODE PRINT,FREQ=0
*NODE FILE,FREQ=0
*RESTART,WRITE,OVERLAY,FREQ=1
*EL FILE, POSITION=AVERAGED AT NODES
SINV,
*END STEP
6-1055
Listing 6.1.2-2
*HEADING
Mass Diffusion Analysis with Linear Elements
Using Quadratic Singular Stress analysis
MODEL BUILT IN mm,N,s
**
*PREPRINT,ECHO=NO,MODEL=NO,HIST=NO
*NODE,NSET=ALL,INPUT=difftocrack_node.inp
**
**
*ELEMENT, TYPE=DC2D4, ELSET=ALL,
INPUT=difftocrack_linear_elements.inp
**
**
*NODE,NSET=TIP
1,2.0,0.
81,2.0,0.
*NGEN,NSET=TIP
1,81,1
**
** tipbound
**
*NSET, NSET=TIPBOUND
5001, 5002, 5003, 5004, 5005,
5006, 5007, 5008, 5009, 5010,
5011, 5012, 5013, 5014, 5015,
5016, 5017, 5018, 5019, 5020,
5021, 5022, 5023, 5024, 5025,
5026, 5027, 5028, 5029, 5030,
5031, 5032, 5033, 5034, 5035,
5036, 5037, 5038, 5039, 5040,
5041, 5042, 5043, 5044, 5045,
5046, 5047, 5048, 5049, 5050,
5051, 5052, 5053, 5054, 5055,
5056, 5057, 5058, 5059, 5060,
5061, 5062, 5063, 5064, 5065,
5066, 5067, 5068, 5069, 5070,
5071, 5072, 5073, 5074, 5075,
5076, 5077, 5078, 5079, 5080,
5081,
*NFILL,NSET=TIPMESH,SINGULAR=1,TWO STEP
TIP,TIPBOUND,50,100
6-1056
*ELEMENT,TYPE=DC2D4,ELSET=ALL
2001,1,201,203,3
*ELGEN,ELSET=ALL
2001,25,200,50,40,2,1
**
** front
**
1,5001,100
*NSET, NSET=FRONT
5001, 5522, 5533, 5554, 5565,
5586, 5597, 5618, 5629, 5650,
5661, 5682, 5693, 5714, 5725,
5746, 5757, 5778, 5789, 5810,
5821, 6802, 6813, 6834, 6845,
6866, 6877, 6898, 6909, 6930,
6941, 6962, 6973, 6994, 7005,
7026, 7037, 7058, 7069, 7090,
7101, 7122, 7133, 7154, 7165,
7186, 7197, 7218, 7229, 7250,
7261, 7282, 7293, 7314, 7325,
7346, 7357, 7378, 7389, 7410,
7421,
**
** RIGHT
**
*NSET, NSET=RIGHT
7421, 7422, 7423, 7424, 7425,
7426, 7427, 7428, 7429, 7430,
7431, 7432, 7433, 7434, 7435,
7436, 7437, 7438, 7439, 7440,
7441, 8642, 8653, 8674, 8685,
8706, 8717, 8738, 8749, 8770,
8781, 8802, 8813, 8834, 8845,
8866, 8877, 8898, 8909, 8930,
8941, 8962, 8973, 8994, 9005,
9026, 9037, 9058, 9069, 9090,
9101, 9122, 9133, 9154, 9165,
9186, 9197, 9218, 9229, 9250,
9261, 9282, 9293, 9314, 9325,
9346, 9357
**
** left
6-1057
**
*NSET, NSET=LEFT
6481, 6782, 6783, 6784, 6785,
6786, 6787, 6788, 6789, 6790,
6791, 6792, 6793, 6794, 6795,
6796, 6797, 6798, 6799, 6800,
6801, 8051, 8071, 8081, 8101,
8111, 8131, 8141, 8161, 8171,
8191, 8201, 8221, 8231, 8251,
8261, 8281, 8291, 8311, 8321,
8341, 8351, 8371, 8381, 8401,
8411, 8431, 8441, 8461, 8471,
8491, 8501, 8521, 8531, 8551,
8561, 8581, 8591, 8611, 8621,
8641, 9387, 9392, 9402, 9407,
9417, 9422, 9432, 9437, 9447,
9452, 9462, 9467, 9477, 9482,
9492, 9497, 9507, 9512, 9522,
9527, 9537, 9542, 9552, 9557,
9567, 9572, 9582, 9587, 9597,
9602, 9612, 9617, 9627, 9632,
9642, 9647, 9657, 9662, 9672,
9677, 9687, 9688, 9689, 9690,
9691, 9692, 9745, 9746
**
** crack
**
*NSET,NSET=CRACK,GEN
81,5081,100
*NSET, NSET=CRACK
5081, 6511, 6531, 6541, 6561,
6571, 6591, 6601, 6621, 6631,
6651, 6661, 6681, 6691, 6711,
6721, 6741, 6751, 6771, 6781,
6801,
**
** top
**
*NSET, NSET=TOP
9357, 9358, 9359, 9360, 9361,
9362, 9363, 9364, 9365, 9366,
9367, 9368, 9369, 9370, 9371,
9372, 9373, 9374, 9375, 9376,
6-1058
9377, 9678, 9679, 9680, 9681,

9682, 9683, 9684, 9685, 9686,
9687,
**
** top
**
*ELSET, ELSET=TOP
1221, 1222, 1223, 1224, 1225,
1226, 1227, 1228, 1229, 1230,
1326, 1327, 1328, 1329, 1330
**
** all
**
*SOLID SECTION, ELSET=ALL, MATERIAL=STEEL
1.,
**
** steel
**
**
**
*ELASTIC, TYPE=ISO
2.E+5, 0.3
*DIFFUSIVITY
34.096361E-6,
*KAPPA,TYPE=PRESS
0.0,0.0
3.921890,1E3
*SOLUBILITY
0.18872283,
*PHYSICAL CONS,ABS=-273
*INITIAL CONDITIONS,TYPE=PRESS
ALL,0.
*INITIAL CONDITIONS,TYPE=CONC
ALL,264.9388
*INITIAL CONDITIONS,TYPE=TEMP
ALL,52.0
*STEP,UNSYMM=YES,AMPLITUDE=STEP
*MASS DIFFUSION,STEADY STATE
0.1,1.0
*BOUNDARY
RIGHT,11,11,264.9388
*PRESSURE STRESS,
6-1059
FILE=difftocrack_quarterpstress
*EL PRINT,FREQ=0
*NODE PRINT,NSET=FRONT
NNC11,
*NODE FILE,NSET=FRONT,FREQ=2
NNC,
*RESTART,WRITE,OVERLAY,FREQ=1
*EL FILE,FREQ=2
CONC,
SOL,
ESOL,
ISOL,
*END STEP
6-1060
Acoustic Analyses
7. Acoustic Analyses
7.1 Acoustic analyses
7.1.1 Coupled acoustic-structural analysis of a car
This example illustrates fully coupled acoustic-structural analysis. Such problems arise when
solid-fluid interaction is fundamental to the overall vibrational behavior of the body or of the acoustic
fluid. Typical examples of such problems include loudspeaker enclosures, fluid-filled tanks, muffler
systems, and vehicle cabin enclosures.
This particular example is a two-dimensional analysis of a car structure and interior and represents a
cross-section of the car cabin cut lengthwise by a vertical plane. The model contains structural
elements to model the car cabin, acoustic elements to model the air interior, and acoustic-structural
interface elements to produce the coupling. All elements have an out-of-plane thickness of 1.0, so all
forces can be interpreted as per unit of thickness of the cross-section.
The analysis begins with natural frequency extractions for the structure alone and for the acoustic
cavity alone. The remaining part of the study obtains the steady-state harmonic response of the fully
coupled system, excited by a point load at one node on the floor of the car, in the range 35-65 Hz. Two
models are used to obtain this response. In one the structure is represented by finite elements. In the
other the structure is represented by some of its natural modes. This latter approach can be quite
cost-effective in some cases (although it is not so in this small example). It is also useful in
applications where the structure is so complex that its harmonic response cannot be predicted
accurately with numerical modeling; instead, the modes and frequencies are obtained experimentally.
This example shows how such numerically or experimentally determined modes can be used in an
analysis. It also includes a study of the sensitivity of the acoustic response to damping in the
structure.
Full model
The models are shown in Figure 7.1.1-1. The structural model is made from beam elements of type
B21, using the properties of various materials making up the structure (steel, glass, and wood). The
acoustic model fills the interior of the structure with 4-node acoustic elements of type AC2D4.
Acoustic interface elements (type ASI2) couple the structure and the acoustic medium. No mesh
convergence study has been done since the example is intended as an illustration only.
The acoustic elements that represent the seat back have a "volumetric drag coefficient" to simulate the
acoustic absorbing properties of the material used in this part of the interior.
The *DAMPING option is used to introduce Rayleigh stiffness proportional damping, governed by the
parameter ¯, into the structural materials in the model. For a given value of ¯ applied to all materials
in the structure, the damping fraction » for a mode ® with natural frequency !® (radians per unit time)
is given by the formula
7-1061
Acoustic Analyses
¯!®
»= :
2
The value of ¯ in the full model was chosen to give approximately 1% critical damping for those
modes of the structure whose natural frequencies are in the range of excitation. This was done by
calculating ¯ to give exactly 1% critical damping at 41.51 Hz (mode 19 of the structure, !19 = 260.81
radians/time), which produces damping fractions ranging from 0.86% at mode 16 to 1.56% at mode 23.
Modal model
In the modal model the structural elements are replaced with a modal representation of the structure as
illustrated in Figure 7.1.1-2. In many practical cases this modal representation is based on experimental
measurements. We do not have such data for this example: instead, we use the modes extracted for the
structure alone. The acoustic elements are defined exactly as in the full model (including volumetric
drag in the seat).
Modal representation means that the physical response uNi in direction i at node N is approximated by
the sum of modal amplitudes along eigenvectors of the structure
M
X
uN
i = a® ÁN
i® ;
®=1
where ÁNi® is the eigenvector of the structural system for mode ®, a® is the modal amplitude of the
response (the "generalized coordinate"), and M is the number of modes used in the representation.
This modal representation of the structure consists of the M independent single degree of freedom
systems coupled to the displacements of the physical nodes through the summation equation above.
Since the modes are orthogonal, the response of each mode, a® ; is that of an independent, one degree
of freedom system (Figure 7.1.1-2), with mass m® , stiffness k® , and viscous damping c® . If the modes
are extracted by ABAQUS, the generalized mass m® and the natural frequency !® (defining
k® = m® =!®2 ) are both available from the output of the *FREQUENCY step. If the modes have been
obtained experimentally, these values are provided as part of the measured response. The damping
value c® for a mode ® is chosen to produce a desired fraction, »® , of the critical damping for that mode
p
and is given by c® = 2»® m® k® :
To couple the displacements of the physical nodes to the generalized coordinates of the modes, a® ,
these generalized coordinates must be present as degrees of freedom in the model. For this purpose a
special, nonphysical node is created for each mode. (The coordinates of these nodes do not affect the
analysis, so for convenience they are all placed at the origin.) The generalized coordinate of a given
mode is represented by displacement in the 1-direction (arbitrarily chosen) at its node. The generalized
mass, damping, and stiffness for a given mode are incorporated as mass, dashpot, and spring elements
at its node.
The physical nodes are needed around the acoustic boundary where we wish to couple the structural
response to the response of the acoustic fluid through ASI-type elements. The summation equation
above is imposed in each direction at each such acoustic boundary node by using the *EQUATION
7-1062
Acoustic Analyses
option to tie the displacements of the physical boundary nodes to the 1-direction displacements of the
nonphysical nodes. Since there are usually many such displacement components, we use a FORTRAN
program to generate the *EQUATION data. acouststructcar_coupled.inp shows that program, using the
ABAQUS results file from the *FREQUENCY analysis of the structure to generate the *EQUATION
data.
The program also extracts m® and k® from the *FREQUENCY analysis results file and calculates c® ;
three files are generated and can be copied directly into an input file to define the mass, spring, and
dashpot coefficients. A damping value, »® , of 1% of critical damping is used in all modes. This modal
damping will not give exactly the same results as the Rayleigh damping used in the full model because
in the full model the fraction of critical damping was exactly 1% at only one frequency. The fraction of
critical damping can be varied in the program by changing the value of the variable FRACTD.

The results for each of the analyses are discussed below.
Natural frequency analysis

The results of the natural frequency analysis of the structure alone are summarized in Table 7.1.1-1 and
illustrated in Figure 7.1.1-3. The eight modes that occur in the frequency range of interest (35-65 Hz)
are shown. The lowest of these modes, mode 16, is at 35.6 Hz when the roof of the car model vibrates
in its third mode. Mode 17, at 37.1 Hz, is the third mode of the back shelf. Mode 18, at 37.7 Hz, is the
fourth mode of the floor. At 47.6 Hz, mode 20, the windshield vibrates in its second mode. The rest of
the modes are higher modes of the floor, the roof, and the back shelf.
The frequency analysis of the acoustic cavity alone is summarized in Table 7.1.1-2 and shown (as
contours of acoustic pressure) in Figure 7.1.1-4. The first nonzero mode of the acoustic cavity is at
50.2 Hz--well above the frequency range of the lowest structural modes. Only this lowest mode falls
into the frequency range of interest, however. Since the acoustic cavity has no boundary conditions on
acoustic pressure when it is modeled alone, there is a zero frequency mode. This requires a small
frequency shift in the *FREQUENCY option to avoid the associated singularity.
Coupled analyses
All of the models are run requesting analysis at 181 frequencies in the range of interest. A coarser
model would result in some of the resonances being missed. A coupled acoustic-structural steady-state
analysis must be performed using the *STEADY STATE DYNAMICS, DIRECT procedure in
ABAQUS.
The coupled analyses are performed as frequency sweeps from 35-65 Hz, with the system excited by a
concentrated force at node 997 (at the location of a rear axle support point), whose magnitude is 1.0 N
in phase and 0.06 N out of phase. The results of the full finite element representation of the structure
are compared to the results obtained with the 25-mode model.
The full model's response is illustrated by the acoustic pressure contours shown in Figure 7.1.1-5. The
contours are shown at four representative frequencies within the range of interest. Figure
7.1.1-6illustrates the response of the 25-mode model for the same frequencies. The differences in the
7-1063
Acoustic Analyses
pressure contours reflect differences in the damping in the two models. The modal model, with 1% of
critical damping at all frequencies, is more damped than the full model at the lower frequencies and
less damped at the higher frequencies.
Figure 7.1.1-7 shows the acoustic pressure at node 271 (about where the driver's head would be
located) and at node 745 (low in the interior, in front of the driver's seat) plotted as a function of
frequency for both the full and 25-mode models. This figure also shows the displacements at nodes
989 and 997 (both on the floor below the seat) for both models.
Effect of volumetric drag

Figure 7.1.1-8 shows the acoustic pressure results for the full model, without Rayleigh damping, both
with and without volumetric drag. The lack of volumetric drag allows the structural resonances at
around 35 Hz and 62 Hz to excite large acoustic pressure amplitudes at these frequencies.
The Rayleigh damping is excluded from the analyses of Figure 7.1.1-8to highlight the effect of
volumetric drag. Comparison of this figure with Figure 7.1.1-7 shows that the Rayleigh damping
dominates volumetric drag effects in this model.
Effect of Rayleigh damping

Figure 7.1.1-9 shows the effect of introducing damping into the full representation of the structure. It
again shows the pressures at nodes 271 and 745 as functions of frequency. As expected, damping in
the structure reduces the amplitude of the resonant response substantially. The volumetric drag was
included in these analyses.
Accuracy of the modal model

The accuracy of the solution using a modal representation of the structure depends on using enough
modes to model the structure properly in the frequency range of interest. The analysis is performed
using 25 modes, then repeated using 50 modes to test the accuracy of the 25-mode solution. It would
be expected that, since the frequency of any mode higher than 25 is well out of the frequency range of
interest, 25 modes would be sufficient to model the structure accurately. This is indeed the case; the
results from the 50-mode model are almost indistinguishable from the 25-mode model and, therefore,
are not shown. The FORTRAN program can generate *EQUATION data for any number of modes by
changing the variable MODES.
Input files
acouststructcar_coupled.inp
Fully coupled steady-state analysis of the full acoustic-structural model, including Rayleigh
damping in the structure.
acouststructcar_structmodes.inp
Extracting the modes of the uncoupled structural model. For use with acouststructcar_equations.f,
the number of modes extracted must match the number of modes desired in the modal analysis of
the system.
7-1064
Acoustic Analyses
acouststructcar_acoustmodes.inp
Extracting the modes of the uncoupled acoustic model.
acouststructcar_equations.f
FORTRAN program used to convert the structural eigenvectors from
acouststructcar_structmodes.inp into *EQUATIONs for the eigenvalue (modal) representation of
the structure.
acouststructcar_eigen25modes.inp
Fully coupled steady-state analysis, which utilizes the eigenvalue representation of the structure
(25 modes), including 1% critical damping of each mode.
acouststructcar_eigen50modes.inp
Fully coupled steady-state analysis, which utilizes the eigenvalue representation of the structure
(50 modes), including 1% critical damping of each mode.
Tables
Table 7.1.1-1 Natural frequencies for the structure alone.

Mode Frequenc Mode Frequenc Mode Frequenc
y, Hz y, Hz y, Hz
1 1.23 11 16.91 21 47.70
2 1.87 12 17.25 22 61.52
3 2.18 13 20.10 23 64.59
4 2.88 14 21.25 24 76.92
5 4.38 15 25.38 25 76.92
6 5.18 16 35.64
7 7.34 17 37.08
8 9.73 18 37.71
9 10.18 19 41.51
10 11.16 20 47.62
Table 7.1.1-2 Natural frequencies for the acoustic cavity alone.

Mode Frequenc
y, Hz
1 50.2
2 95.7
3 104.7
4 141.0
5 164.5
6 190.5
7 207.2
8 228.3
9 231.9
7-1065
Acoustic Analyses
Figures
Figure 7.1.1-1 Two-dimensional model of a car structure and interior.
Figure 7.1.1-2 Modal representation of structure in fully coupled acoustic-structural analysis.
Figure 7.1.1-3 Modes 16-23 of the car structure alone.
7-1066
Acoustic Analyses
Figure 7.1.1-4 Lowest 8 modes of the acoustic medium alone.
7-1067
Acoustic Analyses
Figure 7.1.1-5 Steady-state response of the full model: acoustic pressure plots.
7-1068
Acoustic Analyses
Figure 7.1.1-6 Steady-state response of the 25 mode system: acoustic pressure plots.
Figure 7.1.1-7 Steady-state response: pressure and displacement amplitudes.
7-1069
Acoustic Analyses
Figure 7.1.1-8 Steady-state response of the full model: effect of volumetric drag, without Rayleigh
damping.
7-1070
Acoustic Analyses
Figure 7.1.1-9 Steady-state response of the full model: effect of Rayleigh damping.
7-1071
Acoustic Analyses
Sample listings
7-1072
Acoustic Analyses
Listing 7.1.1-1
*HEADING
Car interior - coupled analysis -
full representation
*RESTART,WRITE
*NODE
101,1.000, 1.100
131,2.600, 1.100
421,0.760, 0.740
581,0.800, 0.520
741,0.600, 0.320
981,0.600,-0.020
989,1.050,-0.020
749,1.050, 0.240
755,1.440, 0.180
435,1.710, 0.730
437,1.810, 0.730
997,1.480,-0.020
1001,1.780,-0.020
921,1.860, 0.120
761,1.880, 0.300
771,2.300, 0.240
451,2.520, 0.730
737,0.250, 0.320
897,0.300, 0.080
979,0.470,-0.020
757,1.630, 0.340
93,0.350, 0.700
415,0.350, 0.700
139,3.380, 0.730
459,3.380, 0.730
*NGEN,NSET=ROOF
101,131
*NGEN,NSET=WINDOW
93,101
415,421
451,459
131,139
*NGEN,NSET=FRONT
101,421,40
421,581,40
581,741,40
7-1073
Acoustic Analyses
741,981,40
*NGEN,NSET=WINDOW
93,421,164
95,341,123
97,261,82
99,181,41
*NGEN,NSET=REAR
131,451,40
451,771,40
761,771
761,921,40
921,1001,40
*NGEN,NSET=FLOOR
979,989
749,989,40
749,755
435,755,40
435,437
437,757,40
757,997,40
997,1001
*NGEN,NSET=TORPED
737,741
737,897,40
897,979,41
*NGEN
181,211
261,291
341,371
421,435
501,515
581,595
661,675
741,749
817,829
837,841
897,909
437,451
517,531
597,611
677,691
757,761
917,921
7-1074
Acoustic Analyses
131,451,40
133,453,40
135,455,40
137,457,160
*ELEMENT,TYPE=AC2D4,ELSET=ACOUSTIC
1,257,218,95,93
2,218,179,97,95
3,179,140,99,97
4,101,99,140,101
5,421,341,218,257
6,341,261,179,218
7,261,181,140,179
8,101,140,181,101
116,213,133,131,211
156,137,135,215,137
157,297,137,215,335
158,457,297,335,455
176,139,137,297,139
177,139,297,457,139
101,183,103,101,181
181,503,423,421,501
189,519,439,437,517
259,819,739,737,817
268,839,759,757,837
*ELGEN,ELSET=ACOUSTIC
101,15,2,1,4,80,20
181, 7,2,1,4,80,20
189, 7,2,1,4,80,20
259, 6,2,1,3,80,20
268, 2,2,1,3,80,20
116, 2,2,20,4,80,1
*ELEMENT,TYPE=AC2D4,ELSET=ACOUSTIC
299,979,899,897,979
137,335,215,213,293
138,455,335,293,373
139,455,373,453,455
*ELEMENT,TYPE=AC2D4,ELSET=ABSORB
188,517,437,435,515
*ELGEN,ELSET=ABSORB
188,4,80,20
*ELEMENT,TYPE=B21
11,103,101
31,95,93
7-1075
Acoustic Analyses
41,457,459
*ELEMENT,TYPE=B21,ELSET=FLOR
81,1001,921
82,999,1001
80,989,997
84,987,989
89,897,979
90,817,897
*ELGEN,ELSET=ROOF
11,15,2
*ELGEN,ELSET=WINDOW
31,4,2
*ELGEN,ELSET=PLYWOOD
41,4,-2
*ELGEN,ELSET=FLOR
82,2,-2
84,5,-2
90,2,-80
*ELSET,ELSET=STEEL
ROOF,FLOR
*BEAM SECTION,SECTION=RECT,MATERIAL=STEEL,
ELSET=STEEL
1,0.9E-3
*DENSITY
7850.,
*ELASTIC
2.1E11,0.3
*DAMPING,BETA=7.668E-5
*BEAM SECTION,SECTION=RECT,MATERIAL=WINDOW,
ELSET=WINDOW
1,2.E-3
*MATERIAL,NAME=WINDOW
*DENSITY
2470.,
*ELASTIC
0.7E11,0.23
*BEAM SECTION,SECTION=RECT,MATERIAL=PLYWOOD,
ELSET=PLYWOOD
1,10.E-3
*MATERIAL,NAME=PLYWOOD
*DENSITY
7-1076
Acoustic Analyses
300.,
*ELASTIC
0.001E11,0.3
*EQUATION
2,
459,1,1,139,1,-1
2,
459,2,1,139,2,-1
2,
459,6,1,139,6,-1
*ELEMENT,TYPE=ASI2,ELSET=INTER
1011,103,101
1031,95,93
1041,457,139
1042,455,457
1081,1001,921
1082,999,1001
1080,989,997
1084,987,989
1089,897,979
1090,817,897
*ELGEN,ELSET=INTER
1011,15,2
1031,4,2
1042,3,-2
1082,2,-2
1084,5,-2
1090,2,-80
*SOLID SECTION,MATERIAL=ACOUSTIC,ELSET=ACOUSTIC
1,
*INTERFACE,ELSET=INTER
1,
*MATERIAL,NAME=ACOUSTIC
*DENSITY
1.293,
*ACOUSTIC MEDIUM,BULK MODULUS
1.183E5,
*SOLID SECTION,ELSET=ABSORB,MATERIAL=ABSORB
1,
*MATERIAL,NAME=ABSORB
*DENSITY
20.,
7-1077
Acoustic Analyses
*ACOUSTIC MEDIUM,BULK MODULUS

5.E6,
*ACOUSTIC MEDIUM,VOLUMETRIC DRAG
30.E2,
*BOUNDARY
101,ENCASTRE
131,ENCASTRE
451,ENCASTRE
737,ENCASTRE
921,ENCASTRE
*NSET,NSET=OUT
271,745, 989,997
*STEP
*STEADY STATE DYNAMICS, DIRECT,
FREQUENCY SCALE=LINEAR
35,65,181,1
*CLOAD
997,2,1.
*CLOAD, LOAD CASE=2
997,2,0.06
S11,PHS11,E11,PHE11
*NODE PRINT,FREQUENCY=10,NSET=OUT
U2,PU2,POR,PPOR
*NODE FILE,NSET=OUT
U,POR,PU,PPOR
*OUTPUT,FIELD
U,POR,PU,PPOR
*OUTPUT,HISTORY
U,POR,PU,PPOR
*END STEP
7.1.2 Fully and sequentially coupled structural acoustics of a

muffler
This example demonstrates the solution of the acoustic field in the vicinity of a muffler in air caused
by the vibrations of the muffler shell. The computations are done using both the fully coupled
(`Àcoustic and coupled acoustic-structural analysis, '' Section 6.9.1 of the ABAQUS/Standard User's
Manual) and sequentially coupled acoustic-solid ( ``Submodeling,'' Section 7.3.1 of the
ABAQUS/Standard User's Manual) interaction procedures in ABAQUS. In the fully coupled case the
7-1078
Acoustic Analyses
solid medium of the muffler is directly coupled to the enclosed and surrounding air in a single analysis.
In the sequentially coupled case the muffler vibrations are considered to be independent of the loading
effects of the surrounding air, while the acoustic vibrations of the surrounding air are forced by the
motion of the muffler. This allows the muffler vibration and acoustic radiation problems to be solved
in sequence, using the submodeling procedure in ABAQUS. The results for the sequentially coupled
model are verified by comparing them to results using the fully coupled procedure.
Full modeling vs. submodeling

The fully coupled model includes the effect of the acoustic pressure in the surrounding air loading the
muffler body during vibration of the system. When modeling the acoustics of metal structures in air,
such as in this case, such acoustic pressure loading is often negligible in comparison with other forces
in the structure. The submodeling capability (*SUBMODEL) can be used in this situation. The part of
the interacting system that is unaffected by the other is treated as the "global" model, while the part
whose solution depends strongly on the solution of the other is treated as the "submodel." In the case
of an acoustic analysis, of course, this nomenclature refers to the hierarchy of the solutions, not the
geometric sizes of the models.
When sequential coupling is physically appropriate, its use offers an advantage over a fully coupled
solution. Two problems, each smaller than the fully coupled problem, are less computationally
expensive. If the applicability of the sequentially coupled solution method is uncertain, the user should
make characteristic test computations in the frequency range of interest. If these computations show
little difference between the fully and sequentially coupled solutions, the less expensive sequentially
coupled method can be used.
Geometry and model

The system considered here consists of a cylindrical muffler and the interacting air. The muffler is a
simple tube 180 mm in diameter and 1 m in length, with inlet and outlet pipes 70 mm in diameter and
100 mm in length. The muffler structure is made from stainless steel sheeting, 0.75 mm in thickness. A
porous packing material, which dampens the acoustic field, surrounds the inner pipe.
Although this problem is in essence axisymmetric, a narrow three-dimensional wedge (subtending an
angle of 10°) of the coupled system is modeled because ABAQUS has a limitation on the use of
submodeling with axisymmetric shells. Appropriate boundary conditions are applied to the
three-dimensional model so that the axisymmetric solution is captured. The meshes of the surrounding
air, the exterior muffler shell, and the air inside the muffler are shown in Figure 7.1.2-1, Figure 7.1.2-2,
and Figure 7.1.2-3, respectively.
The air inside the muffler is meshed with AC3D10 elements (second-order tetrahedra). The innermost
column of fluid elements models the undamped air. The adjacent annulus models the air in the region
of the packing material. These two regions are highlighted in Figure 7.1.2-3, where the annulus is
shown as the darker region. The effect of the packing material is modeled using the *ACOUSTIC
MEDIUM, VOLUMETRIC DRAG option. The muffler is meshed with S4R shell elements.
The exterior fluid is shown in Figure 7.1.2-1. Its outer boundary is made up of spherical and cylindrical
segments, on which spherical and cylindrical absorbing boundary conditions are imposed using
7-1079
Acoustic Analyses
*SIMPEDANCE, TYPE=SPHERE and *SIMPEDANCE, TYPE=CIRCULAR, respectively. The

cylindrical and spherical absorbing boundary conditions can be combined in ABAQUS, allowing the
external mesh to conform to the geometry of the radiating object more closely. Combinations of
different boundary condition types are most effective when the boundaries are continuous in slope as
well as displacement. Second-order hexahedral acoustic elements ( AC3D20) are used to fill in the
volume of the exterior fluid region. In the submodeling procedure the interface between the
surrounding air and the muffler is meshed with 8-node acoustic interface elements ( ASI8). The choice
of mesh density (element size) is discussed in `Àcoustic and coupled acoustic-structural analysis, ''
Section 6.9.1 of the ABAQUS/Standard User's Manual. In both cases the inner boundary of the
exterior air mesh conforms to the muffler shell and to rigid baffles, which isolate the exterior field
from the exhaust and inlet noise. These baffle pipes are the same diameter as the inlet and exhaust
pipes but are modeled simply by imposing no boundary condition on the acoustic elements in this
region. This is equivalent to imposing the condition that the acceleration on this boundary is zero,
which is correct for a rigid baffle.
We are most interested in performing a frequency sweep about the first resonant frequency of the fully
coupled system. However, a direct determination of the eigenfrequencies using the *FREQUENCY
procedure yields modes and frequencies associated with the decoupled fluid and solid systems (see
`Àcoustic and coupled acoustic-structural analysis, '' Section 6.9.1 of the ABAQUS/Standard User's
Manual, for more details). For problems involving air and metal structures, the structure usually
dominates the behavior of the system. Therefore, an estimate of the first important resonance of the
coupled system is found by performing a frequency sweep in the vicinity of the first eigenfrequency of
the muffler shell, computed without any interaction with the interior or exterior air. This occurs at
f = 172 Hz . Although the resonant frequencies of the fully coupled system do not coincide with the
resonant frequencies of the muffler shell alone, they are close, especially at lower frequencies.
Using the *STEADY STATE DYNAMICS, DIRECT procedure to search around 172 Hz, we find that
the first resonant frequency for the fully coupled system occurs at approximately 178 Hz. A frequency
sweep of both the fully coupled and the sequentially coupled models from 177.2 Hz to 179.4 Hz at 0.2
Hz increments is performed. A pressure wave of unit magnitude is applied to the muffler inlet at each
frequency, and a plane wave absorbing boundary condition is applied at the muffler outlet.
The material properties for the air are a bulk modulus Kf of 0:142 MPa and a density ½f of 1.2
3
kg=m , yielding a characteristic sound speed of 344 m=s. The volumetric drag, r, specified for the air
in the packing material region is 1:2 N s = m. Volumetric drag values are considered "small" if they are
small compared to 2 ¼ ½f f , a condition satisfied by r = 1:2 N s = m for the frequency range of interest.
The muffler is made of stainless steel with Young's modulus E of 190 MPa, Poisson's ratio º of 0.3,
3
and density ½s of 7920 kg=m .
Material properties affect the mesh parameters appropriate for wave problems. The characteristic
wavelength of air at f = 180Hz , − = 2¼ £ 180 ¼ 1131 rad=sec , is
q
Kf
¸a = ½f −2
¼ 1:9m
, which is long compared to the overall system geometry. The internodal spacing of roughly 40 mm
7-1080
Acoustic Analyses
used in the surrounding acoustic mesh and 30 mm in the interior acoustic mesh is adequate for this
frequency. The acoustic wavelength must also be considered in selecting the overall size of the exterior
domain. Accuracy of the solution requires placement of the radiating boundary at least one-quarter
wavelength from the acoustic sources; in this problem a standoff distance of approximately 700 mm is
selected. The characteristic flexural wavelength ¸p of the steel plating can be computed using the
thickness h and the formula
2¼ Eh2 1=4
¸p = p (
− 12½s
) ¼ 190mm
. The discretization requirements of the finite element method in wave problems require at least six
nodes per wavelength; here, we use an internodal distance of approximately 30 mm for the shells.
The fully coupled model consists of all three meshes shown in Figure 7.1.2-1, Figure 7.1.2-2, and
Figure 7.1.2-3, constrained at their abutting surfaces using the *TIE option.
The sequentially coupled analysis is performed in two jobs. The "global" model job consists of the
meshes shown in Figure 7.1.2-2 and Figure 7.1.2-3. Displacements and displacement phases from the
shell elements are saved from this analysis and drive the second "submodel" analysis through the use
of the *BOUNDARY, SUBMODEL option. The second model consists of the exterior air mesh
(Figure 7.1.2-1) used in the fully coupled case, with ASI8 elements placed on the boundary that abuts
the shell surface. These elements convert the displacements from the "global" analysis to the
appropriate boundary conditions for acoustic elements. In this analysis the ASI8 elements conform to
the acoustic submodel mesh but not to the shell mesh of the global model. The nodes of the ASI8
elements are placed in a node set, specified in the model data by the *SUBMODEL option. The
GLOBAL ELSET parameter must be used in this case to ensure that only the displacements of the
ASI8 elements are driven by the shell elements. Without the GLOBAL ELSET parameter, ABAQUS
may attempt to drive the acoustic pressure of the ASI8 elements by the interior acoustic elements, since
those elements share the shell nodes in the "global" model.

It is good practice to check the absorbing boundary conditions used on a particular mesh at a desired
frequency by analyzing only the exterior fluid mesh with some test forcing on the boundary where
acoustic excitations are expected. If the forcing is at a single point, the pressure phase angles should
show a pattern of concentric circles, minimally distorted by the radiating boundary. While not a
rigorous numerical test, such a result usually coincides with a properly offset radiating boundary. As
shown in Figure 7.1.2-4, this criterion is met by the mesh used in this analysis.
Figure 7.1.2-5 is a plot of the radial displacement of the muffler inlet as a function of frequency for
both the fully coupled and the global models. The resonant peak for the fully coupled model at 178.1
Hz is clearly illustrated. In contrast, the resonant peak for the "global" model (without the acoustic
medium) occurs at approximately 178.2 Hz. The difference in the two peaks can be accounted for by
the fact that the exterior air on the fully coupled model adds a small amount of damping, due to
radiation, as well as mass to the system, which result in a lowered natural frequency, as well as a
slightly lower peak response. It is clear from Figure 7.1.2-5 that for the frequency range of interest the
7-1081
Acoustic Analyses
coupling between the exterior air and the muffler is most important at 178.1 Hz.
Figure 7.1.2-6 and Figure 7.1.2-7 contain contour plots of the pressure magnitude and phase for the
muffler interior at 179.4 Hz for both the "global" model and the fully coupled model. In both cases the
results indicate that the modeling assumptions of the sequentially coupled analysis appear to be valid
for the solutions in the muffler interior.
Contour plots of the pressure magnitude and phase for the muffler exterior at 179.4 Hz are shown in
Figure 7.1.2-8 and Figure 7.1.2-9. The resulting pressure magnitude in the exterior air is small in both
cases. The differences in the pressure amplitudes and phase as computed by the two analyses are not
considered to be significant. Two factors that account for the small differences are the different
modeling methods (fully coupled vs. sequentially coupled) and the different techniques used to couple
the muffler to the exterior air (*TIE vs. acoustic interface elements).
Figure 7.1.2-10 and Figure 7.1.2-11 contain contour plots of the pressure magnitude and phase for the
muffler interior at 178.1 Hz for both the "global" model and the fully coupled model. It is clear that at
179.4 Hz, the modeling assumptions of the sequentially coupled analysis are less valid than they are at
178.1 Hz for the solutions in the muffler interior. This result is anticipated by Figure 7.1.2-5. However,
the solutions are still reasonably close to one another, indicating that the sequentially coupled analysis
is still a reasonable approximation for this system even at a resonant peak.
Contour plots of the pressure magnitude and phase for the muffler exterior at 178.1 Hz are shown in
Figure 7.1.2-12 and Figure 7.1.2-13. Again, the resulting pressure magnitude in the exterior air is small
in both cases. The differences in the pressure amplitudes and phase as computed by the two analyses is
less evident in the exterior than they were in the interior.
The pressure magnitudes along the muffler centerline at both 178.1 Hz and 179.4 Hz are shown in
decibels in Figure 7.1.2-14. The reference pressure is chosen at one unit for convenience. The plot
illustrates the variation of acoustic pressure in the muffler near resonance.
Table 7.1.2-1 shows comparative solution times and memory requirements for the fully and
sequentially coupled analyses. The total computational time for the sequentially coupled case is lower,
and the peak memory requirements are significantly lower. These differences will be greater for larger
models. Optimal speed increases occur when global and submodels have nearly equal numbers of
degrees of freedom. Here, solving the fully coupled system does not impose as much of a speed penalty
as might be expected, because the sparse solver used by ABAQUS exploits the extreme sparsity of the
fluid-solid coupling term. When the number of system nodes involving fluid-solid coupling is a large
percentage of the total number of nodes, the sparsity of the coupling term decreases, favoring the
sequentially coupled procedure. Sequentially coupled analyses are even more advantageous than fully
coupled analyses when many different submodels need to be analyzed, driven by a single set of global
results.
ABAQUS issues a series of warning messages in this example, because the narrow wedge domain
results in some three-dimensional acoustic elements with bad aspect ratios. These messages can be
ignored in this study, since the solutions are essentially axisymmetric and the gradient of the solution
in the circumferential direction is nearly zero. Moreover, elements with scalar degrees of freedom,
such as the acoustic elements used in this example, are much less sensitive to geometric distortion than
7-1082
Acoustic Analyses
elements with vector degrees of freedom, such as continuum stress/displacement elements.
Input files
muffler_full.inp
Three-dimensional, fully coupled model.
muffler_globl.inp
Muffler and internal air global model.
muffler_submo.inp
Exterior air submodel.
muffler_shell_nodes.inp
Nodal coordinates for muffler shell mesh.
muffler_intair_nodes.inp
Nodal coordinates for interior air mesh.
muffler_extair_nodes.inp
Nodal coordinates for surrounding air mesh.
muffler_shell_elem.inp
Element definitions for muffler shell mesh.
muffler_intair_elem.inp
Element definitions for interior air mesh.
muffler_extair_elem.inp
Element definitions for surrounding air mesh.
muffler_freq.inp
Natural frequency extraction for shell mesh.
muffler_bctest.inp
Element definitions for surrounding air mesh.
Table
Table 7.1.2-1 Comparison of relative CPU times (normalized with respect to the CPU time for the
sequential analysis) and approximate problem size for the frequency sweep excluding preprocessing.
Memory DOF Relative CPU
Time
Global model 10 Mb 10030 0.325
Submodel 15 Mb 19030 0.675
7-1083
Acoustic Analyses
Fully coupled 29 Mb 29060 1.086

model
Sequential analysis 1.000
Figures
Figure 7.1.2-1 Mesh of surrounding air.
Figure 7.1.2-2 Mesh of muffler.
Figure 7.1.2-3 Mesh of interior air.
7-1084
Acoustic Analyses
Figure 7.1.2-4 Radiating boundary condition test at 165 Hz.
Figure 7.1.2-5 Radial displacement of the muffler inlet as a function of frequency.
Figure 7.1.2-6 Muffler internal pressure magnitudes at 179.4 Hz, muffler inlet at top: fully coupled
solution on left, "global" model (without the exterior acoustic medium) on right.
7-1085
Acoustic Analyses
7-1086
Acoustic Analyses
Figure 7.1.2-7 Muffler internal pressure phase at 179.4 Hz, muffler inlet at top: fully coupled solution
on left, "global" model (without the exterior acoustic medium) on right.
7-1087
Acoustic Analyses
7-1088
Acoustic Analyses
Figure 7.1.2-8 Muffler external pressure magnitudes at 179.4 Hz, muffler inlet at top: fully coupled
Figure 7.1.2-9 Muffler external pressure phase at 179.4 Hz, muffler inlet at top: fully coupled solution
7-1089
Acoustic Analyses
on left, "global" model (without the exterior acoustic medium) on right.
Figure 7.1.2-10 Muffler internal pressure magnitudes at 178.1 Hz, muffler inlet at top: fully coupled
7-1090
Acoustic Analyses
7-1091
Acoustic Analyses
Figure 7.1.2-11 Muffler internal pressure phase at 178.1 Hz, muffler inlet at top: fully coupled
7-1092
Acoustic Analyses
7-1093
Acoustic Analyses
Figure 7.1.2-12 Muffler external pressure magnitudes at 178.1 Hz, muffler inlet at top: fully coupled
Figure 7.1.2-13 Muffler external pressure phase at 178.1 Hz, muffler inlet at top: fully coupled
7-1094
Acoustic Analyses
Figure 7.1.2-14 Muffler internal pressure magnitude at 178.1 and 179.4 Hz: dB along muffler
centerline.
7-1095
Acoustic Analyses
Sample listings
7-1096
Acoustic Analyses
Listing 7.1.2-1
*HEADING
FULLY COUPLED ACOUSTIC ANALYSIS OF A MUFFLER
M, KG, S
** -------------------------------------------
**
** PART INSTANCE: INTERIORAIR-1
**
*NODE, INPUT=muffler_intair_nodes.inp
*ELEMENT, TYPE=AC3D10,
INPUT=muffler_intair_elem.inp
** -------------------------------------------
*ELSET, ELSET=INTERIORAIR-1_I1, GENERATE
549, 1457, 1
*SOLID SECTION, ELSET=INTERIORAIR-1_I1,
MATERIAL=AIRABSORB
1.,
** -------------------------------------------
*ELSET, ELSET=INTERIORAIR-1_I2, GENERATE
1, 548, 1
*SOLID SECTION, ELSET=INTERIORAIR-1_I2,
MATERIAL=AIR
1.,
** -------------------------------------------
**
** PART INSTANCE: MUFFLERSHELL-1
**
*NODE,INPUT=muffler_shell_nodes.inp
*ELEMENT, TYPE=S4R, INPUT=muffler_shell_elem.inp
** -------------------------------------------
*ELSET, ELSET=MUFFLERSHELL-1_I1, GENERATE
1458, 1500, 1
*SHELL SECTION, ELSET=MUFFLERSHELL-1_I1,
MATERIAL=STEEL
0.00075, 5
** -------------------------------------------
**
** PART INSTANCE: OUTERAIR-1
**
*NODE,INPUT=muffler_extair_nodes.inp
*ELEMENT, TYPE=AC3D20,
INPUT=muffler_extair_elem.inp
7-1097
Acoustic Analyses
** -------------------------------------------
*ELSET, ELSET=OUTERAIR-1_I1, GENERATE
1501, 2734, 1
*SOLID SECTION, ELSET=OUTERAIR-1_I1,
MATERIAL=AIR
1.,
** -------------------------------------------
*NSET, NSET=SHELLSYM,UNSORTED
3301,3376,3375,3299,3373,3297,3340,3339,3338,
3337,3336,3335,3334,3333,3332,3331,3330,3329,
3328,3327,3326,3325,3324,3323,3322,3321,3320,
3319,3318,3317,3316,3315,3314,3313,3312,3311,
3310,3309,3295,3307,3292,3304,3303,3291
** -------------------------------------------
*NSET, NSET=SHELLCONSTRAINT,UNSORTED
3302,3377,3378,3300,3374,3298,3341,3342,3343,
3344,3345,3346,3347,3348,3349,3350,3351,3352,
3353,3354,3355,3356,3357,3358,3359,3360,3361,
3362,3363,3364,3365,3366,3367,3368,3369,3370,
3371,3372,3296,3308,3293,3305,3306,3294
** -------------------------------------------
*NSET, NSET=SHELLDISP
3304,3325,3375
** -------------------------------------------
*NSET, NSET=CENTERLINE
10, 11, 118, 119, 120, 121, 122, 123, 124, 125,
126, 127, 128, 129, 130, 131, 132, 133, 134, 135,
136, 137, 138, 139, 140, 141, 142, 143, 144, 145,
146, 147, 148, 149, 150, 151, 152, 153, 154, 155,
156, 157, 158, 159, 160, 161, 162, 163, 164, 165,
166, 167, 168, 169, 170, 171, 172, 173, 174, 175,
176, 831, 832, 833, 834, 835, 836, 837, 838, 839,
840, 841, 842, 843, 844, 845, 846, 847, 848, 849,
850, 851, 852, 853, 854, 855, 856, 857, 858, 859,
860, 861, 862, 863, 864, 865, 866, 867, 868, 869,
870, 871, 872, 873, 874, 875, 876, 877, 878, 879,
880, 881, 882, 883, 884, 885, 886, 887, 888, 889,
890
** -------------------------------------------
*NSET, NSET=INLETPRESSURE
11, 12, 13, 177, 186, 618, 891, 892, 893, 913,
914, 915
** -------------------------------------------
7-1098
Acoustic Analyses
*ELSET, ELSET=INNERAIROUTLET_S4
288, 299, 367
*SURFACE, NAME=INNERAIROUTLET
INNERAIROUTLET_S4, S4
** -------------------------------------------
*ELSET, ELSET=INNERAIR2SHELL_S1
367, 374, 384, 385, 386, 387, 388, 392,
393, 894, 895, 896, 1192, 1195, 1196, 1197,
1198, 1201, 1202, 1203, 1210, 1213, 1214, 1216,
1237, 1238, 1239, 1240, 1243, 1248, 1250, 1330,
1331, 1334, 1337, 1341, 1359
370, 373, 375, 380, 383, 391, 559, 634,
894, 895, 1193, 1194, 1200, 1204, 1206, 1208,
1211, 1212, 1215, 1217, 1235, 1242, 1244, 1246,
1247, 1249, 1251, 1276, 1298, 1320, 1321, 1322,
1323, 1326, 1329, 1335, 1336, 1338, 1340, 1343,
1347, 1349
379, 592, 635, 858, 917, 1199, 1205, 1209,
1236, 1241, 1252, 1253, 1254, 1325, 1339, 1344,
1345, 1346, 1348, 1350, 1351, 1352, 1353, 1354,
1355, 1356, 1357, 1358, 1360, 1361, 1362
377, 378, 389, 390, 557, 560, 561, 568,
683, 888, 897, 1207, 1245, 1277, 1324, 1327,
1328, 1332, 1333, 1342, 1387, 1388
*SURFACE, NAME=INNERAIR2SHELL
INNERAIR2SHELL_S1, S1
** -------------------------------------------
*ELSET, ELSET=SHELL2INNERAIR_SNEG, GENERATE
1458, 1500, 1
*SURFACE, NAME=SHELL2INNERAIR
SHELL2INNERAIR_SNEG, SNEG
*ELSET, ELSET=SHELL2OUTERAIR_SPOS, GENERATE
1458, 1500, 1
*SURFACE, NAME=SHELL2OUTERAIR
SHELL2OUTERAIR_SPOS, SPOS
** -------------------------------------------
*ELSET, ELSET=OUTERAIR2SHELL_S6
7-1099
Acoustic Analyses
2100,2116,2420
*ELSET, ELSET=OUTERAIR2SHELL_S3
2100,2420,2421,2422
*ELSET, ELSET=OUTERAIR2SHELL_S4, GENERATE
1519, 1975, 19
*SURFACE, NAME=OUTERAIR2SHELL
OUTERAIR2SHELL_S6, S6
** -------------------------------------------
*ELSET, ELSET=SPHERE1_S3, GENERATE
2340, 2355, 1
2664, 2734, 5
*SURFACE, NAME=SPHERE1
SPHERE1_S3, S3
SPHERE1_S4, S4
** -------------------------------------------
2036, 2051, 1
1979, 2035, 4
*SURFACE, NAME=SPHERE2
SPHERE2_S3, S3
SPHERE2_S4, S4
** -------------------------------------------
*ELSET, ELSET=CYLINDER_S6, GENERATE
1501, 1957, 19
*SURFACE, NAME=CYLINDER
CYLINDER_S6, S6
** -------------------------------------------
*ELSET, ELSET=INNERAIRABSORB_S3
382, 394, 403, 404, 405, 408, 411, 415, 417, 418,
419, 423, 429, 430, 431, 437, 438, 439, 448, 451,
452, 456, 457, 458, 459, 461, 462, 463, 464, 467,
468, 473, 477, 478, 482, 483, 484, 498, 499, 502,
503, 505, 506, 507
407, 428, 436, 443, 471, 476, 504
402, 410, 414, 416, 422, 426, 434, 435, 440, 441,
444, 450, 453, 455, 460, 466, 472, 475, 479, 481,
485, 486, 487, 488, 490, 491, 492, 495, 496, 497,
7-1100
Acoustic Analyses
500
427, 432, 433, 442, 445, 446, 447, 449, 454, 465,
469, 470, 474, 480, 489, 493, 494, 501
*SURFACE, NAME=INNERAIRABSORB
INNERAIRABSORB_S3, S3
** -------------------------------------------
**
** MATERIALS
**
*MATERIAL, NAME=AIR
*ACOUSTICMEDIUM,BULKMODULUS
1.42E5,
*DENSITY
1.2,
*MATERIAL, NAME=AIRABSORB
*ACOUSTICMEDIUM,BULKMODULUS
1.42E5,
*ACOUSTICMEDIUM,VOLUMETRICDRAG
1.2,
*DENSITY
1.2,
*DENSITY
7920.,
*ELASTIC
1.9E+11, 0.3
*IMPEDANCEPROPERTY,TYPE=SPH,NAME=SPHERE
0.7
*IMPEDANCEPROPERTY,TYPE=C,NAME=CY
0.7
** -------------------------------------------
*TIE,NAME=ACOST1
INNERAIR2SHELL, SHELL2INNERAIR
OUTERAIR2SHELL, SHELL2OUTERAIR
**
** BOUNDARY CONDITIONS AND CONSTRAINTS
**
*BOUNDARY
SHELLSYM,ZSYMM
7-1101
Acoustic Analyses
*EQUATION
2
SHELLCONSTRAINT,1,0.176326981,
SHELLCONSTRAINT,3,-1.
2
SHELLCONSTRAINT,2,1., SHELLSYM,2,-1.
** -------------------------------------------
**
** STEP: STEP-1
**
*STEP
STEP-1: FREQUENCY RESPONSE
*STEADY STATE DYNAMICS, DIRECT,
FREQUENCY SCALE=LINEAR
177.21,179.41,12
*BOUNDARY
INLETPRESSURE,8,8,1.0
*SIMPEDANCE
CYLINDER,CY
*SIMPEDANCE
SPHERE1,SPHERE
*SIMPEDANCE
SPHERE2,SPHERE
*SIMPEDANCE
INNERAIROUTLET,
*OUTPUT,FIELD,OP=NEW,FREQUENCY=6
*NODE OUTPUT
U,PU,POR,PPOR
*OUTPUT,HISTORY,OP=NEW,FREQUENCY=1
*NODE OUTPUT,NSET=CENTERLINE
POR,
*NODE OUTPUT,NSET=SHELLDISP
U,
*NODE FILE,NSET=CENTERLINE,FREQUENCY=1
COORD,POR,PPOR
*NODE FILE,NSET=SHELLDISP,FREQUENCY=1
U,PU
*END STEP
7-1102
Soils Analyses
8. Soils Analyses
8.1 Soils analyses
8.1.1 Plane strain consolidation
Most consolidation problems of practical interest are two- or three-dimensional, so that the
one-dimensional solutions provided by Terzaghi consolidation theory (see ``The Terzaghi
consolidation problem,'' Section 1.14.1 of the ABAQUS Benchmarks Manual) are useful only as
indicators of settlement magnitudes and rates. This problem examines a linear, two-dimensional
consolidation case: the settlement history of a partially loaded strip of soil. This particular case is
chosen to illustrate two-dimensional consolidation because an exact solution is available (Gibson et
al., 1970), thus providing verification of this capability in ABAQUS.
Geometry and model

The discretization of the semi-infinite, partially loaded strip of soil is shown in Figure 8.1.1-1. The
loaded region is half as wide as the depth of the sample. The reduced-integration plane strain element
with pore pressure, CPE8RP, is used in this analysis. Reduced integration is almost always
recommended when second-order elements are used because it usually gives more accurate results and
is less expensive than full integration. No mesh convergence studies have been done, although the
reasonable agreement between the numerical results provided by this model and the solution of Gibson
et al. (1970) suggests that the model used is adequate--at least for the overall displacement response
examined. In an effort to reduce analysis cost while at the same time preserve accuracy, the mesh is
graded from six elements through the height, under the load, to one element through the height at the
outer boundary of the model, where a single infinite element (type CINPE5R) is used to model the
infinite domain. This requires the use of two kinematic constraint features provided by ABAQUS.
Consider first the displacement degrees of freedom along line AC in Figure 8.1.1-1. The 8-node
isoparametric elements used for the analysis allow quadratic variation of displacement along their
sides, so the displacements of nodes a and b in elements x and y may be incompatible with the
displacement variation along side AC of element z. To avoid this, nodes a and b must be constrained
to lie on the parabola defined by the displacements of nodes A, B, and C: The QUADRATIC MPC
("multi-point constraint") is used to enforce this kinematic constraint: it must be used at each node
where this constraint is required (see planestrainconsolidation.inp ). Pore pressure values are obtained
by linear interpolation of values at the corner nodes of an element. When mesh gradation is used, as
along line AC in this example, an incompatibility in pore pressure values may result for the same
reason given for the displacement incompatibility discussed above. To avoid this, the pore pressure at
node B must be constrained to be interpolated linearly from the pore pressure values at A and C: This
is done by using the P LINEAR MPC.
The material properties assumed for this analysis are as follows: the Young's modulus is chosen as 690
GPa (108 lb/in2); the Poisson's ratio is 0; the material's permeability is 5.08 ´ 10-7 m/day (2.0 ´ 10-5
in/day); and the specific weight of pore fluid is chosen as 272.9 kN/m 3 (1.0 lb/in 3).
8-1103
Soils Analyses
The applied load has a magnitude of 3.45 MPa (500 lb/in 2). The strip of soil is assumed to lie on a
smooth, impervious base, so the vertical component of displacement is prescribed to be zero on that
surface. The left-hand side of the mesh is a symmetry line (no horizontal displacement). The infinite
element models the other boundary.
Time stepping
As in the one-dimensional Terzaghi consolidation solution (see ``The Terzaghi consolidation
problem,'' Section 1.14.1 of the ABAQUS Benchmarks Manual), the problem is run in two steps. In the
first *SOILS, CONSOLIDATION step, the load is applied and no drainage is allowed across the top
surface of the mesh. This one increment step establishes the initial distribution of pore pressures which
will be dissipated during the second *SOILS, CONSOLIDATION step.
During the second step drainage is allowed to occur through the entire surface of the strip. This is
specified by prescribing the pore pressure (degree of freedom 8) at all nodes on this surface (node set
TOP) to be zero. By default in a *SOILS, CONSOLIDATION step such boundary conditions are
applied immediately at the start of the step and then held fixed. Thus, the pore pressures at the surface
change suddenly at the start of the second step from their values with no drainage (defined by the first
step) to 0.0.
Consolidation is a typical diffusion process: initially the solution variables change rapidly with time,
while at the later times more gradual changes in stress and pore pressure are seen. Therefore, an
automatic time stepping scheme is needed for any practical analysis, since the total time of interest in
consolidation is typically orders of magnitude larger than the time increments that must be used to
obtain reasonable solutions during the early part of the transient. ABAQUS uses a tolerance on the
maximum change in pore pressure allowed in an increment, UTOL, to control the time stepping. When
the maximum change of pore pressure in the soil is consistently less than UTOL the time increment is
allowed to increase. If the pore pressure changes exceed UTOL, the time increment is reduced and the
increment is repeated. In this way the early part of the consolidation can be captured accurately and the
later stages are analyzed with much larger time steps, thereby permitting efficient solution of the
problem. For this case UTOL is chosen as 0.344 MPa (50 lb/in 2), which is 10% of the applied load.
This is a fairly coarse tolerance but results in an economical and reasonable solution.
The choice of initial time step is important in consolidation analysis. As discussed in ``The Terzaghi
consolidation problem,'' Section 1.14.1 of the ABAQUS Benchmarks Manual, the initial solution
(immediately following a change in boundary conditions) is a local, "skin effect" solution. Due to the
coupling of spatial and temporal scales, it follows that no useful information is provided by solutions
generated with time steps smaller than the mesh and material-dependent characteristic time. Time steps
very much smaller than this characteristic time provide spurious oscillatory results (see Figure
3.1.5-2). This issue is discussed by Vermeer and Verruijt (1981), who propose the criterion
°!
¢t ¸ (¢h)2 ;
6Ek
where ¢h is the distance between nodes of the finite element mesh near the boundary condition
change, E is the elastic modulus of the soil skeleton, k is the soil permeability, and °! is the specific
8-1104
Soils Analyses
weight of the pore fluid. In this problem ¢h is 8.5 mm (0.33 in), so--using the material properties
shown in Figure 8.1.1-1--
¢tinitial = 1 £ 10¡5 days.
We actually use an initial time step of 2 ´ 10-5 days, since the immediate transient just after drainage
begins is not considered important in the solution.

The prediction of the time history of the vertical deflection of the central point under the load (point P
in Figure 8.1.1-1) is plotted in Figure 8.1.1-2, where it is compared with the exact solution of Gibson
et al. (1970). There is generally good agreement between the theoretical and finite element solutions,
even though the mesh used in this analysis is rather coarse.
Figure 8.1.1-2 also shows the time increments selected by the automatic scheme, based on the UTOL
tolerance discussed above. The figure shows the effectiveness of the scheme: the time increment
changes by two orders of magnitude over the analysis.
Input file
planestrainconsolidation.inp
Input data for this example.
References
· Gibson, R. E., R. L. Schiffman, and S. L. Pu, "Plane Strain and Axially Symmetric Consolidation
of a Clay Layer on a Smooth Impervious Base," Quarterly Journal of Mechanics and Applied
Mathematics, vol. 23, pt. 4, pp. 505-520, 1970.
· Vermeer, P. A., and A. Verruijt, "An Accuracy Condition for Consolidation by Finite Elements,"
International Journal for Numerical and Analytical Methods in Geomechanics, vol. 5, pp. 1-14,
1981.
Figures
Figure 8.1.1-1 Plane strain consolidation example: geometry and properties.
8-1105
Soils Analyses
Figure 8.1.1-2 Consolidation history and time step variation history.
8-1106
Soils Analyses
Sample listings
8-1107
Soils Analyses
Listing 8.1.1-1
*HEADING
PLANE STRAIN CONSOLIDATION
*NODE
1,
13,,2.
801,2.
813,2.,2.
1401,4.
1413,4.,2.
1506,4.5,.8333
1601,5.
1603,5.,.5
1607,5.,1.
1610,5.,1.5
1613,5.,2.
2001,9.
2007,9.,1.
2013,9.,2.
2201,17.
2213,17.,2.
*NGEN,NSET=LHS
1,13
*NGEN,NSET=TOP
13,813,100
813,1413,100
1413,1613,100
1613,2013,100
13,813,100
*NGEN,NSET=BOT
1,801,100
801,1401,100
1401,1601,100
1601,2001,100
1,1401,100
*NSET,NSET=TOP1,GENERATE
13,1413,100
*NFILL
BOT1,TOP1,12,1
*NGEN
8-1108
Soils Analyses
1607,2007,100
*NSET,NSET=ALLN,GENERATE
1,9999
**
*ELEMENT,TYPE=CPE8RP,ELSET=SOIL
1,1,201,203,3,101,202,103,2
101,801,1001,1005,805,901,1003,905,803
201,1401,1601,1607,1405,1501,1603,1506,1403
202,1405,1607,1613,1413,1506,1610,1513,1409
301,1601,1801,1813,1613,1701,1807,1713,1607
*ELGEN,ELSET=SOIL
1,6,2,1,4,200,10
101,3,4,1,3,200,10
301,2,200
*ELEMENT,TYPE=CINPE5R,ELSET=SOILIN
401,2013,2001,2201,2213,2007
*SOLID SECTION,ELSET=SOIL,MATERIAL=A1
*SOLID SECTION,ELSET=SOILIN,MATERIAL=A1IN
**
*MATERIAL,NAME=A1
*ELASTIC
1.E8,
*PERMEABILITY,SPECIFIC=1.0
2.E-5,
*MATERIAL,NAME=A1IN
*ELASTIC
1.E8,
ALLN,1.5
*MPC
QUADRATIC,802,801,803,805
QUADRATIC,804,801,803,805
QUADRATIC,806,805,807,809
QUADRATIC,808,805,807,809
QUADRATIC,810,809,811,813
QUADRATIC,812,809,811,813
QUADRATIC,1407,1405,1409,1413
QUADRATIC,1411,1405,1409,1413
QUADRATIC,1603,1601,1607,1613
QUADRATIC,1610,1601,1607,1613
P LINEAR,803,801,805
P LINEAR,807,805,809
P LINEAR,811,809,813
8-1109
Soils Analyses
P LINEAR,1409,1405,1413
P LINEAR,1607,1601,1613
**
*STEP
SET UP INITIAL PORE PRESSURES
*SOILS,CONSOLIDATION
1.E-7,1.E-7
*BOUNDARY
LHS,1
BOT,2
*DLOAD
6,P3,500.
16,P3,500.
*NODE PRINT,NSET=N1
U,POR,RVT
*NODE FILE,FREQUENCY=2,NSET=N1
U,
POR,
*NODE OUTPUT,NSET=N1
U,
POR,
*END STEP
**
*STEP,INC=300
*SOILS,CONSOLIDATION,UTOL=50.,END=SS
2.E-5, 1.E-1,,, 1.E-2
*BOUNDARY
TOP,8
*END STEP
8.1.2 Calculation of phreatic surface in an earth dam

This example illustrates the use of ABAQUS to solve for the flow through a porous medium in which
fluid flow is occurring in a gravity field and only part of the region is fully saturated, so the location of
the phreatic surface is a part of the solution. Such problems are common in hydrology (an example is
the well draw-down problem, where the phreatic surface of an aquifer must be located, based on
pumping rates at particular well locations) and in some problems of dam design, as in this example.
The basic approach takes advantage of the ABAQUS capability to perform partially and fully saturated
analysis: the phreatic surface is located at the boundary of the fully saturated part of the model. This
8-1110
Soils Analyses
approach has the advantage that the capillary zone, just above the phreatic surface, is also
identified.
Boundary conditions
A typical dam is shown in Figure 8.1.2-1. We consider fluid flow only: deformation of the dam is
ignored. Thus, although we use the fully coupled pore fluid flow-deformation elements, all
displacement degrees of freedom are prescribed to be zero. A more general analysis would include
stress and deformation of the dam.
The upstream face of the dam (surface S1 in Figure 8.1.2-1) is exposed to water in the reservoir behind
the dam. Since ABAQUS uses a total pore pressure formulation, the pore pressure on this face must be
prescribed to be uw = (H1 ¡ z )g½w , where H1 is the elevation of the water surface, z is elevation, g
is the gravitational acceleration, and ½w is the mass density of the water. (g½w , the weight density of
the water, must be given as the value of the SPECIFIC parameter on the *PERMEABILITY option.)
Likewise, on the downstream face of the dam (surface S2 in Figure 8.1.2-1), uw = (H2 ¡ z )g½w :
The bottom of the dam (surface S3 ) is assumed to rest on an impermeable foundation. Since the natural
boundary condition in the pore fluid flow formulation provides no flow of fluid across a surface of the
model, no further specification is needed on this surface.
The phreatic surface in the dam, S4 , is found as the locus of points at which the pore fluid pressure,
uw , is zero. Above this surface the pore fluid pressure is negative, representing capillary tension
causing the fluid to rise against the gravitational force and creating a capillary zone. The saturation
associated with particular values of capillary pressure for absorption and exsorption of fluid from the
porous medium is a physical property of the material and is defined in the *SORPTION option.
A special boundary condition is needed if the phreatic surface reaches an open, freely draining surface,
as indicated on surface S5 in Figure 8.1.2-1. In such a case the pore fluid can drain freely down the
face of the dam, so uw =0 at all points on this surface below its intersection with the phreatic surface.
Above this point uw <0, with its particular value depending on the solution. This example is
specifically chosen to include this effect to illustrate the use of the ABAQUS drainage-only flow
boundary condition.
This drainage-only flow condition consists of prescribing the flow velocity on the freely draining
surface in a way that approximately satisfies the requirement of zero pore pressure on the completely
saturated portion of this surface (Pagano, 1997). The flow velocity is defined as a function of pore
pressure, as shown in Figure 8.1.2-2. For negative pore pressures (those above the phreatic surface) the
flow velocity is zero--the proper natural boundary condition. For positive pore pressures (those below
the phreatic surface) the flow velocity is proportional to the pore pressure value. When this
proportionality coefficient, ks , is large compared to k=°w c--where k is the permeability of the
medium, °w is the specific weight of the fluid, and c is a characteristic length scale--the requirement of
zero pore pressure on the free-drainage surface below the phreatic surface will be satisfied
approximately. The drainage-only seepage coefficient in this model is specified as ks =10-1 m3/Nsec.
This value is roughly 10 5 times larger than the characteristic value, k=°w c, based on the material
properties listed below and an element length scale ¼10-1 m. This condition is prescribed using the
*FLOW option with the drainage-only flow type label (QnD) as shown in phreaticsurf_cpe8rp.inp.
8-1111
Soils Analyses
Geometry and model

The geometry of the particular earth dam considered is shown in Figure 8.1.2-3. This case is chosen
because an analytical solution is available for comparison (Harr, 1962). The dam is filled to two-thirds
of its height. Only a part of its base is impermeable. Since the dam is assumed to be long, we use
CPE8RP coupled pore pressure/displacement plane strain elements (the mesh is shown in Figure
8.1.2-4). In addition, an input file containing element type CPE4P is included for verification purposes.
Additional input files are included to demonstrate the use of *CONTACT PAIR and *TIE in coupled
pore pressure-displacement analyses.
Material
The permeability of the fully saturated earth of which the dam is made is 0.2117 ´ 10-3 m/sec. The
default assumption is used for the partially saturated permeability: that it varies as a cubic function of
saturation, decreasing from the fully saturated value to a value of zero at zero saturation. The specific
weight of the water is 10 kN/m 3. The capillary action in the dam is defined by a single
absorption/exsorption curve that varies linearly between a negative pore pressure of 10 kN/m 2 at a
saturation of 0.05 and zero pore pressure at fully saturated conditions. This is not a very realistic model
of physical absorption/exsorption behavior, but this will not affect the results of the steady-state
analysis significantly insofar as the location of the phreatic surface is concerned. Accurate definition of
this behavior would be required if definition of the capillary zone created by filling and emptying the
dam at given rates is needed.
The initial void ratio of the earth material is 1.0. The initial conditions for pore pressure and saturation
are assumed to be those corresponding to the dam being fully saturated to the upstream water level: the
initial saturation is, therefore, 1.0; and the initial pore pressures vary between zero at the water level
and a maximum value of 12.19 kN/m 2 at the base of the dam.

The weight of the water is applied by GRAV loading, and the upstream and downstream pore pressures
are prescribed as discussed above. A steady-state *SOILS analysis is performed in five increments to
allow ABAQUS to resolve the high degree of nonlinearity in the problem.

The steady-state contours of pore pressure are shown in Figure 8.1.2-5. The upper-right part of the dam
shows negative pore pressures, indicating that it is partly saturated or dry. The phreatic surface is best
shown in Figure 8.1.2-6, where we have chosen to draw the contours in the vicinity of zero pore
pressure. This phreatic surface compares well with the analytical phreatic surface calculated by Harr
(1962), shown in Figure 8.1.2-2. Figure 8.1.2-7 shows contours of saturation that indicate a region of
fully saturated material under the phreatic zone and decreasing saturation in and above the phreatic
zone.
Input files
8-1112
Soils Analyses
phreaticsurf_cpe8rp.inp
Phreatic surface calculation (element type CPE8RP).
phreaticsurf_cpe4p.inp
Element type CPE4P.
phreaticsurf_cpe4p_contactpair.inp
Element type CPE4P using the *CONTACT PAIR option.
phreaticsurf_cpe4p_tie.inp
Element type CPE4P using the *TIE option.
References
· Harr, M. E., Groundwater and Seepage, McGraw-Hill, New York, 1962.
· Pagano, L., "Steady State and Transient Unconfined Seepage Analyses for Earthfill Dams,"
ABAQUS Users' Conference, Milan, pp. 577-585, 1997.
Figures
Figure 8.1.2-1 Phreatic surface problem.
Figure 8.1.2-2 Pore pressure-flow velocity relationship defined on the drainage-only surface.
8-1113
Soils Analyses
Figure 8.1.2-3 Configuration of earth dam and analytical phreatic surface.
Figure 8.1.2-4 Finite element mesh.
Figure 8.1.2-5 Pore pressure contours at steady state.
8-1114
Soils Analyses
Figure 8.1.2-6 Pore pressure contours showing phreatic surface (displayed by setting CMIN= -10,
CMAX= 10).
Figure 8.1.2-7 Saturation contours at steady state (displayed by setting CMIN= -0.6, CMAX= 0.9).
Sample listings
8-1115
Soils Analyses
Listing 8.1.2-1
*HEADING
EARTH DAM
STEADY STATE FREE SURFACE SEEPAGE
*** UNITS: M, KG, SEC, NEWTON
*NODE,NSET=ALLN
1,0.,0.
39,4.8768,0.
601,1.8288,1.8288
639,3.048,1.8288
*NGEN,NSET=BOT
1,39,1
*NGEN,NSET=TOP
601,639,1
*NFILL,NSET=ALLN
BOT,TOP,12,50
*NSET,NSET=PORN0,GENERATE
1,39,2
101,139,2
201,239,2
301,339,2
401,439,2
501,539,2
601,639,2
*NSET,NSET=PORN
PORN0, PORN1, PORN2, PORN3
*NSET,NSET=OUTN,GENERATE
1,601,100
21,621,100
*ELEMENT,TYPE=CPE8RP,ELSET=DAM
1,1,3,103,101,2,53,102,51
*ELGEN,ELSET=DAM
1,19,2,1,6,100,20
*ELSET,ELSET=FSIDE,GENERATE
19,119,20
*ELSET,ELSET=FBOT,GENERATE
8-1116
Soils Analyses
16,19
*ELSET,ELSET=OUTE,GENERATE
11,111,20
*SOLID SECTION,ELSET=DAM,MATERIAL=FILL
*MATERIAL,NAME=FILL
*ELASTIC
1000.,
*DENSITY
2000.,
*PERMEABILITY,SPECIFIC=10000.
2.1167E-4,
*SORPTION
-100000.,.04
-10000.,.05
0.,1.
*INITIAL CONDITIONS,TYPE=SATURATION
ALLN,1.
*INITIAL CONDITIONS,TYPE=PORE PRESSURE
PORN, 12192.0, 0.0, 0.0, 1.2192
PORN4,0.
PORN5,0.
PORN6,0.
ALLN,1.
*BOUNDARY
ALLN,1
ALLN,2
*STEP,INC=5
*SOILS
.2,1.
*DLOAD
DAM,GRAV,10.,0.,-1.,0.
*BOUNDARY
1,8,,12192.
101,8,,9144.
201,8,,6096.
301,8,,3048.
401,8,,0.
*FLOW
FSIDE,Q2D,0.1
FBOT,Q1D,0.1
8-1117
Soils Analyses
*NODE PRINT,FREQUENCY=10,NSET=OUTN
POR,
*EL PRINT,FREQUENCY=10,ELSET=OUTE
SAT,POR
*NODE FILE,FREQUENCY=10,NSET=OUTN
POR,
*NODE OUTPUT,NSET=OUTN
POR,
*EL FILE,FREQUENCY=10,ELSET=OUTE
SAT,POR
*ELEMENT OUTPUT,ELSET=OUTE
SAT,POR
*END STEP
8.1.3 Axisymmetric simulation of an oil well

This example simulates the settlement of soil near an oil well. It is assumed that the oil in question is
too thick for normal pumping. Therefore, steam is injected in the soil in the vicinity of the well to
increase the temperature and decrease the oil's viscosity. As a result creep becomes an important
component of the soil inelastic deformation and in the prediction of the effects of the oil pumping. Five
years of oil pumping are simulated. This coupled displacement/diffusion analysis illustrates the use of
ABAQUS to solve problems involving fluid flow through a saturated porous medium, inelastic
material properties with time-dependent creep behavior, and thermal loading. No experimental data
exist to compare with the numerical results of this example.
Geometry and model

The example considers an axisymmetric model of an oil well and the surrounding soil, as shown in
Figure 8.1.3-1. The radius of the well is 81 m (265 ft), and the well extends from a depth of 335 m
(1100 ft) to 732 m (2500 ft). A depth of 1463 m (4800 ft) is modeled with 11 different soil layers.
Reduced-integration axisymmetric elements with pore pressure, CAX8RP, are used to model the soil
in the vicinity of the well. The far-field region is modeled with axisymmetric infinite elements,
CINAX5R, to provide lateral stiffness. Reduced integration is almost always recommended when
second-order elements are used, because it usually gives more accurate results and is less expensive
than full integration. A coarse mesh is selected for the illustrative purpose of this example. No mesh
convergence study has been performed.
Soil layers designated by S1, T1, U1, and L1 are modeled using the Drucker-Prager plasticity model
and are specified on the *DRUCKER PRAGER option. Both the elastic and inelastic material
properties are tabulated in Table 8.1.3-1. The linear form of the Drucker-Prager model with no
intermediate principal stress effect (K = 1.0) is used. The model assumes nonassociated flow;
8-1118
Soils Analyses
consequently, the material stiffness matrix is not symmetric. The use of UNSYMM=YES on the
*STEP option improves the convergence of the nonlinear solution significantly. The
hardening/softening behavior is specified by the *DRUCKER PRAGER HARDENING option, and the
data are listed in Table 8.1.3-1. No creep data are provided for these layers since these are far removed
from the loading. These layers are assumed to be saturated with water. A high permeability is assumed
for the two top soil layers S1 and T1, while a low permeability is assigned to layers U1 and L1.
Layers D1 through D7 are modeled with the modified Drucker-Prager Cap plasticity model. The
material property data are tabulated in Table 8.1.3-2and are specified by the *CAP PLASTICITY
option. As required by the creep model, no intermediate principal stress effect is included (i.e., K =
1.0), and no transition region on the yield surface is defined (i.e., ® = 0.0). The material's volumetric
strain-driven hardening/softening behavior is specified with the *CAP HARDENING option, and the
data are listed in Table 8.1.3-2. The initial cap yield surface position, "in
vol (0) , is set to 0.02. ABAQUS
automatically adjusts the position of the cap yield surface if the stress lies outside the cap surface.
Consolidation creep is modeled with a Singh-Mitchell type creep model. The creep material data are
specified with the *CAP CREEP option and are dependent on temperature. The following creep data
are specified:
A=2.2E¡7 1/day, ®=3.05 1/MPa (0.021 1/psi), t1 =1.0 day, n=1.0 at 10°C (50°F)
A=3.5E¡4 1/day, ®=3.05 1/MPa (0.021 1/psi), t1 =1.0 day, n=1.0, at 100°C (212°F)
These layers consist of rich organic matter and are saturated with oil. The temperature-dependent
permeability data are specified by the *PERMEABILITY option.
A uniform thermal expansion coefficient of 5.76E-6 1/°C (3.2E-6 1/°F) and a constant weight density
1.0 metric ton/m3 (64.6 lbs/ft 3) are assumed for all layers.
For a coupled diffusion/displacement analysis care must be taken when choosing the units of the
problem. The coupled equations may be numerically ill-conditioned if the choice of the units is such
that the numbers generated by the equations of the two different fields differ by many orders of
magnitude. The units chosen for this example are inches, pounds, and days.
Initial conditions
An initial geostatic stress field is defined through the *INITIAL CONDITIONS option and is based on
the soil weight density integrated over the depth. A coefficient of lateral stress of 0.85 is assumed. An
initial void ratio of 1.5 is used throughout all soil layers with an initial uniform temperature field of
10°C (50°F).
Loading
The problem is run in five steps. The first step of the analysis is a *GEOSTATIC step to equilibrate
geostatic loading of the finite element model. This step also establishes the initial distribution of pore
pressure. Since gravity loading is defined with distributed load type BZ and not with gravity load type
GRAV, the pore fluid pressure reported by ABAQUS is defined as the pore pressure in excess of the
hydrostatic pressure required to support the weight of pore fluid above the elevation of the material
point.
The second step is a *SOILS, CONSOLIDATION step to equilibrate any creep effects induced from
8-1119
Soils Analyses
the initial geostatic loading step. The choice of the initial time step is important in a consolidation
analysis. Because of the coupling of spatial and temporal scales, no useful information is provided by
solutions generated with time steps that are smaller than the mesh and material-dependent
characteristic time. Time steps that are very much smaller than this characteristic time provide spurious
oscillatory results. For further discussion on calculating the minimum time step, refer to ``Coupled
pore fluid diffusion and stress analysis,'' Section 6.7.1 of the ABAQUS/Standard User's Manual. For
this example a minimum initial time step of one day was selected.
The third step of the analysis models the injection of steam into the well region between a depth of 366
m to 732 m (1200 ft to 2400 ft). The region is indicated by the shaded area in Figure 8.1.3-1. The
nodes in this region are heated to 100°C (212°F) during a *SOILS, CONSOLIDATION analysis. The
NO CREEP parameter is included; therefore, creep effects are not considered. The injection of the
steam increases the permeability of the oil and increases the soil creep behavior.
The fourth step simulates the pumping of oil by prescribing an excess pore pressure of -1.2 MPa (-170
psi) at nodes located at the depth of 427 m to 550 m (1400 ft to 1800 ft) below the surface. The
pressure produces a pumping rate of approximately 172.5 thousand barrels per day at the end of the
fifth year.
The final step consists of a consolidation analysis performed over a five-year period to investigate the
settlement that results from pumping and creep effects in the vicinity of the well.

The two initial steps show negligible deformations, indicating that the model is in geostatic
equilibrium. Figure 8.1.3-2shows a contour plot of the soil settlement resulting from consolidation
after the five-year period. A settlement of 0.13 m (0.4 ft) is expected at the surface. A maximum soil
dislocation of 0.24 m (0.78 ft) occurs above the pump intake. Figure 8.1.3-3shows a contour plot of the
excess pore pressure. The negative pore pressure represents the suction of the pump. During the
five-year period, a total of 313.5 million barrels of oil are pumped (as determined from nodal output
variable RVT). Figure 8.1.3-4 through Figure 8.1.3-6 show contour plots of the vertical stress
components, plastic strains, and creep strains, respectively. Plastification occurs in soil layers D3
through D5. Significant creep occurs in the area in which steam is injected.
Input files
axisymoilwell.inp
Finite element analysis.
axisymoilwell_thermalexp.inp
Same as axisymoilwell.inp except that the thermal expansion of the pore fluid is also included.
Tables
Table 8.1.3-1 Soil data using Drucker-Prager model.
8-1120
Soils Analyses
Soil layer Elastic Inelastic Hardening behavior

propertie
properties
s
S1 E = 124 MPa ¯ = 42.0° 0.075 MPa, 0.0
º = 0.3 K=1.0 0.083 MPa, 0.058
' = 0.0° 0.075 MPa, 0.116
T1 E = 2068 MPa ¯ = 36.0° 0.48 MPa, 0.0
º = 0.25 K=1.0 0.62 MPa, 0.058
' = 0.0° 0.48 MPa., 0.116
U1 E = 468.8
¯ = 38.0° 1.97 MPa, 0.0
MPa
º = 0.22 K=1.0 3.17 MPa, 0.0037
' = 0.0° 2.47 MPa, 0.04
L1 E = 2482 MPa ¯ = 38.0° 1.97 MPa, 0.0
º = 0.29 K=1.0 3.17 MPa, 0.0037
' = 0.0° 2.47 MPa, 0.04
Table 8.1.3-2 Soil data using modified Drucker-Prager cap model.

Soil layer Elastic Inelastic Hardening
properties properties behavior
D1 E = 328 MPa d = 1.38 MPa ® = 0.0 2.75 MPa, 0.0
º = 0.17 ¯ = 36.9° K=1.0 4.14 MPa, 0.02
R = 0.33 vol (0) = 0.02
"in 5.51 MPa, 0.05
6.20 MPa, 0.09
D2 E = 434 MPa d = 1.38 MPa ® = 0.0 1.38 MPa, 0.0
º = 0.17 ¯ = 39.4° K=1.0 4.14 MPa, 0.02
R = 0.33 vol (0) = 0.02
"in 6.89 MPa, 0.04
55.1 MPa, 0.1
D3 E = 546 MPa d = 1.38 MPa ® = 0.0 1.38 MPa, 0.0
º = 0.19 ¯ = 42.0° K=1.0 3.45 MPa, 0.02
R = 0.34 vol (0) = 0.02
"in 13.8 MPa, 0.04
62.0 MPa, 0.06
D4 E = 411 MPa d = 1.2 MPa ® = 0.0 1.38 MPa, 0.0
º = 0.2 ¯ = 40.1° K=1.0 5.03 MPa, 0.02
R = 0.3 vol (0) = 0.02
"in 6.90 MPa, 0.10
62.0 MPa, 0.3
D5 E = 494 MPa d = 1.38 MPa ® = 0.0 2.75 MPa, 0.0
º = 0.17 ¯ = 40.4° K=1.0 4.83 MPa, 0.02
R = 0.3 vol (0) = 0.02
"in 5.15 MPa, 0.04
62.0 MPa, 0.08
D6 E = 775 MPa d = 17 MPa ® = 0.0 2.76 MPa, 0.0
º = 0.17 ¯ = 50.2° K=1.0 4.14 MPa,
0.005
R = 0.23 vol (0) = 0.02
"in 7.58 MPa, 0.02
62.0 MPa, 0.05
D7 E = 1,121 MPa d = 1.7 MPa ® = 0.0 3.44 MPa, 0.0
º = 0.17 ¯ = 58.5° K=1.0 4.14 MPa,
0.006
8-1121
Soils Analyses
R = 0.23 vol (0) = 0.02

"in 7.58 MPa,
0.012
67.6 MPa, 0.03
Figures
Figure 8.1.3-1 Axisymmetric model of oil well and surrounding soil.
Figure 8.1.3-2 Soil settlement after five-year period.
Figure 8.1.3-3 Contour plot of the pore pressure.
8-1122
Soils Analyses
Figure 8.1.3-4 Contour plot of the vertical stress components.
Figure 8.1.3-5 Contour plot of the vertical plastic strain components.
8-1123
Soils Analyses
Figure 8.1.3-6 Contour plot of the vertical creep strain components.
Sample listings
8-1124
Soils Analyses
Listing 8.1.3-1
*HEADING
EXAMPLE-AXISYMMETRIC SIMULATION OF OIL WELL
UNITS: F = lbs, L = in., T = days
*NODE
1, 0.0, 0.0, 0.0
9, 3180.0, 0.0, 0.0
11, 9216.0, 0.0, 0.0
13, 21504.0, 0.0, 0.0
15, 46080.0, 0.0, 0.0
17, 95232.0, 0.0, 0.0
*NGEN
1, 9, 1
9, 11, 1
11, 13, 1
13, 15, 1
15, 17, 1
*NSET,NSET=CORNODES, GENERATE
1, 17, 2
*NSET,NSET=MIDNODES, GENERATE
2, 16, 2
*NSET,NSET=BASE
CORNODES,
MIDNODES,
*NCOPY, CHANGE NUMBER=100, OLD SET=CORNODES,
NEW SET=MID, SHIFT
0,7200.

NEW SET=CORNER, SHIFT
0,14400.

NEW SET=MID, SHIFT
0,18000.

0,21600.

NEW SET=MID, SHIFT
8-1125
Soils Analyses
0,25200.

0,28800.

NEW SET=MID, SHIFT
0,31200.

0,33600.

NEW SET=MID, SHIFT
0,34800.

0,36000.

NEW SET=MID, SHIFT
0,38400.

0,40800.

NEW SET=MID, SHIFT
0,42000.

0,43200.

NEW SET=MID, SHIFT
0,43800.
8-1126
Soils Analyses

0,44400.

NEW SET=MID, SHIFT
0,46800.

0,49200.

NEW SET=MID, SHIFT
0,52500.

0,55800.

NEW SET=MID, SHIFT
0,56700.

0,57600.
*NCOPY, CHANGE NUMBER=200, OLD SET=MIDNODES,

NEW SET=MID, SHIFT
0,14400.

NEW SET=MID, SHIFT
0,21600.

NEW SET=MID, SHIFT
0,28800.

NEW SET=MID, SHIFT
0,33600.
8-1127
Soils Analyses

NEW SET=MID, SHIFT
0,36000.

NEW SET=MID, SHIFT
0,40800.

NEW SET=MID, SHIFT
0,43200.

NEW SET=MID, SHIFT
0,44400.

NEW SET=MID, SHIFT
0,49200.

NEW SET=MID, SHIFT
0,55800.

NEW SET=MID, SHIFT
0,57600.
1, 2201, 100
*NSET, NSET=NALL
MID,
MIDNODES,
CORNER,
CORNODES,
*NSET,NSET=FREESURF,GENERATE
2201, 2215, 2
**
** REGION WHERE STEAM IS INJECTED
**
*NSET, NSET=TARGET
601,602,603,604,605,606,607,608,609
701,702,703,704,705,706,707,708,709
8-1128
Soils Analyses
801,802,803,804,805,806,807,808,809
901,902,903,904,905,906,907,908,909
1001,1002,1003,1004,1005,1006,1007,1008,1009
1101,1102,1103,1104,1105,1106,1107,1108,1109
1201,1202,1203,1204,1205,1206,1207,1208,1209
1301,1302,1303,1304,1305,1306,1307,1308,1309
1401,1402,1403,1404,1405,1406,1407,1408,1409
1501,1502,1503,1504,1505,1506,1507,1508,1509
1601,1602,1603,1604,1605,1606,1607,1608,1609
**
** ELEMENTS GENERATION
**
*ELEMENT,TYPE=CAX8RP
1, 1,3,203,201,2,103,202,101
*ELGEN, ELSET=ESOIL
1, 7,2,1,11,200,100
*ELEMENT, TYPE=CINAX5R, ELSET=FAR
8, 215, 15, 17, 217, 115
108, 415, 215, 217, 417, 315
208, 615, 415, 417, 617, 515
308, 815, 615, 617, 817, 715
408, 1015, 815, 817, 1017, 915
508, 1215, 1015, 1017, 1217, 1115
608, 1415, 1215, 1217, 1417, 1315
708, 1615, 1415, 1417, 1617, 1515
808, 1815, 1615, 1617, 1817, 1715
908, 2015, 1815, 1817, 2017, 1915
1008, 2215, 2015, 2017, 2217, 2115
*ELSET, ELSET=L1, GENERATE
1,7,1
*ELSET, ELSET=U1, GENERATE
101,107,1
*ELSET, ELSET=D7, GENERATE
201,207,1
301,307,1
401,407,1
501,507,1
601,607,1
8-1129
Soils Analyses
701,707,1
801,807,1
*ELSET, ELSET=T1, GENERATE
901,907,1
*ELSET, ELSET=S1, GENERATE
1001,1007,1
**
** SOIL STRATA L1 (3600 ft - 4800 ft)
** ----------------------------------
*MATERIAL,NAME=L1
*ELASTIC
360000.,0.29
*DRUCKER PRAGER, SHEAR CRITERION=LINEAR
38.,1.,0.
*DRUCKER PRAGER HARDENING
286.,0.
460.,0.0037
358.,0.04
*PERMEABILITY,SPECIFIC=3.61e-2
15.2,
*EXPANSION
0.32E-05,
**
** SOIL STRATA U1 (3000 ft - 3600 ft)
** ----------------------------------
*MATERIAL,NAME=U1
*ELASTIC
68000.,0.22
*DRUCKER PRAGER
38.,1.,0.
286.,0.
460.,0.0037
358.,0.04
15.2,
*EXPANSION
0.32E-05,
**
** SOIL STRATA D7 (2400 ft - 3000 ft)
** ----------------------------------
*MATERIAL,NAME=D7
8-1130
Soils Analyses
*ELASTIC
162600.,.17
*CAP PLASTICITY
247.,58.5,0.23,0.02,0.,1.
*CAP HARDENING
500.,.0
600.,.006
1100.,.012
9800.,.03
*CAP CREEP, LAW=SINGHM, MECHANISM=CONSOLIDATION
2.2e-7, 2.1e-2, 1.0, 1.0, 50.0
3.5e-4, 2.1e-2, 1.0, 1.0, 212.0
16.8,,0
32.0,,250
*EXPANSION
0.32E-05,
**
** -----------------------------------
*MATERIAL,NAME=D6
*ELASTIC
112400.,.17
*CAP PLASTICITY
247.,50.2,0.23,0.02,0.,1.
*CAP HARDENING
400.,.0
600.,.005
1100.,.02
9000.,.05
2.2e-7, 2.1e-2, 1.0, 1.0, 50.0
3.5e-4, 2.1e-2, 1.0, 1.0, 212.0
16.8,,0
32.0,,250
*EXPANSION
0.32E-05,
**
** ----------------------------------
*MATERIAL,NAME=D5
*ELASTIC
8-1131
Soils Analyses
71600.,.17
*CAP PLASTICITY
200.,40.4,0.30,0.02,0.,1.
*CAP HARDENING
400.,.0
700.,.02
2000.,.04
9000.,.08
2.2e-7, 2.1e-2, 1.0, 1.0, 50.0
3.5e-4, 2.1e-2, 1.0, 1.0, 212.0
16.8,,0
32.0,,250
*EXPANSION
0.32E-05,
**
** ----------------------------------
*MATERIAL,NAME=D4
*ELASTIC
59600.,.2
*CAP PLASTICITY
174.,40.1,0.30,0.02,0.,1.
*CAP HARDENING
200.,.0
730.,.02
1000.,.04
9000.,.1
2.2e-7, 2.1e-2, 1.0, 1.0, 50.0
3.5e-4, 2.1e-2, 1.0, 1.0, 212.0
16.8,,0
32.0,,250
*EXPANSION
0.32E-05,
**
** ----------------------------------
*MATERIAL,NAME=D3
*ELASTIC
79200.,.19
8-1132
Soils Analyses
*CAP PLASTICITY
200.,42.0,0.34,0.02,0.,1.
*CAP HARDENING
200.,.0
500.,.02
2000.,.04
9000.,.06
2.2e-7, 2.1e-2, 1.0, 1.0, 50.0
3.5e-4, 2.1e-2, 1.0, 1.0, 212.0
16.8,,0
32.0,,250
*EXPANSION
0.32E-05,
**
** ----------------------------------
*MATERIAL,NAME=D2
*ELASTIC
63000.,.17
*CAP PLASTICITY
200.,39.4,0.33,0.02,0.,1.
*CAP HARDENING
200.,.0
600.,.02
1000.,.04
8000.,.1
2.2e-7, 2.1e-2, 1.0, 1.0, 50.0
3.5e-5, 2.1e-2, 1.0, 1.0, 212.0
16.8,,0
32.0,,250
*EXPANSION
0.32E-05,
**
** ---------------------------------
*MATERIAL,NAME=D1
*ELASTIC
47700.,.17
*CAP PLASTICITY
8-1133
Soils Analyses
200.,36.9,0.33,0.02,0.,1.
*CAP HARDENING
400.,.0
600.,.02
800.,.05
900.,.09
2.2e-7, 2.1e-2, 1.0, 1.0, 50.0
3.5e-5, 2.1e-2, 1.0, 1.0, 212.0
16.8,,0
32.0,,250
*EXPANSION
0.32E-05,
**
** SOIL STRATA T1 (150 ft - 700 ft)
** --------------------------------
*MATERIAL,NAME=T1
*ELASTIC
300000.,0.25
*DRUCKER PRAGER
36.,1.,0.
70.,0.
90.,.058
70.,.116
40,
*EXPANSION
0.32E-05,
**
** SOIL STRATA S1 (0 ft - 150 ft)
** ------------------------------
*MATERIAL,NAME=S1
*ELASTIC
18000.,0.3
*DRUCKER PRAGER
42.,1.,0.
11.,0.
12.,.058
11.,.116
8-1134
Soils Analyses
40.,
*EXPANSION
0.32E-05,
** FAR FIELD MATERIAL DATA
** -----------------------
*MATERIAL,NAME=FAR
*ELASTIC
47700.,.17
**
*SOLID SECTION, ELSET=L1, MATERIAL=L1
*SOLID SECTION, ELSET=U1, MATERIAL=U1
*SOLID SECTION, ELSET=D7, MATERIAL=D7
*SOLID SECTION, ELSET=T1, MATERIAL=T1
*SOLID SECTION, ELSET=S1, MATERIAL=S1
*SOLID SECTION, ELSET=FAR, MATERIAL=FAR
**
*BOUNDARY
FREESURF, 8,,0.
**
*INITIAL CONDITIONS, TYPE=RATIO
NALL, 1.5
**
NALL, 50
L1, -2151.2, 0.,-1615.7, 14400., .85 , .85
U1, -1615.7, 14400.,-1346.4, 21600., .85 , .85
D7, -1346.4, 21600.,-1077.1, 28800., .85 , .85
D6, -1077.1, 28800., -897.6, 33600., .85 , .85
D5, -897.6, 33600., -807.8, 36000., .85 , .85
D4, -807.8, 36000., -628.3, 40800., .85 , .85
D3, -628.3, 40800., -538.6, 43200., .85 , .85
D2, -538.6, 43200., -493.7, 44400., .85 , .85
D1, -493.7, 44400., -314.2, 49200., .85 , .85
T1, -314.2, 49200., -67.3, 55800., .85 , .85
S1, -67.3, 55800. , 0.0, 57600., .85 , .85
FAR,-2151.2, 0., 0.0, 57600., .85 , .85
8-1135
Soils Analyses
**
**
** GEOSTATIC STEP TO BALANCE GRAVITY LOADING
**
*STEP, UNSYMM=YES, NLGEOM
*GEOSTATIC
*BOUNDARY
BASE, 2
AXIS, 1
*DLOAD
S1, BZ, -0.0374
T1, BZ, -0.0374
D1, BZ, -0.0374
D2, BZ, -0.0374
D3, BZ, -0.0374
D4, BZ, -0.0374
D5, BZ, -0.0374
D6, BZ, -0.0374
D7, BZ, -0.0374
U1, BZ, -0.0374
L1, BZ, -0.0374
*END STEP
**
** CONSOLIDATION ANALYSIS (1 MONTH-BALANCE CREEP)
**
*STEP,NLGEOM,INC=100,UNSYMM=YES
*SOILS,CONSOLIDATION,CETOL=0.0001,UTOL=10
1.,31.,
*EL FILE, FREQ=0, ELSET=ESOIL
S,E,PE,POR,VOIDR,CE
*END STEP
**
** THERMAL LOADING - INJECTING STEAM
**
*STEP, NLGEOM, INC=100, UNSYMM=YES
*SOILS, CONSOLIDATION, UTOL=25, NO CREEP
1.,1.
*TEMPERATURE,OP=MOD
TARGET, 212.
S,E,PE,POR,VOIDR,CE
*END STEP
8-1136
Soils Analyses
**
** INSTALL PUMP
**
*STEP, NLGEOM, INC=100, UNSYMM=YES, AMP=RAMP
*SOILS, CONSOLIDATION, UTOL=25
1.,1.
*BOUNDARY, OP=MOD
1001,8,8,-170
1201,8,8,-170
S,E,PE,POR,VOIDR,CE
*END STEP
**
** PUMPING OIL OVER FIVE YEAR PERIOD
**
*STEP,NLGEOM,INC=200,UNSYMM=YES
*SOILS,CONSOLIDATION,CETOL=0.0001,UTOL=20
1.,1825.,
*CONTROLS, PARAMETER=FIELD, FIELD=DISPLACEMENT
,1.
*CONTROLS, PARAMETER=FIELD,
FIELD=PORE FLUID PRESSURE
,1.
*EL FILE,FREQ=200, ELSET=ESOIL
S,E,PE,CE,POR,VOIDR
*EL PRINT,FREQ=0, ELSET=ESOIL
S,
CE,
PE,
*NODE PRINT,FREQ=0
U,RVF,RVT
*END STEP
8.1.4 Analysis of a pipeline buried in soil

Oil and gas pipelines are usually buried in the ground to provide protection and support. Buried
pipelines may experience significant loading as a result of relative displacements of the ground along
their length. Such large ground movement can be caused by faulting, landslides, slope failures, and
seismic activity.
ABAQUS provides a library of pipe-soil interaction ( PSI) elements to model the interaction between a
buried pipeline and the surrounding soil. The pipeline itself is modeled with any of the beam, pipe, or
elbow elements in the ABAQUS/Standard element library. The ground behavior and soil-pipe
8-1137
Soils Analyses
interaction are modeled with the pipe-soil interaction elements. These elements have only
displacement degrees of freedom at their nodes. One side or edge of the element shares nodes with the
underlying beam, pipe, or elbow element that models the pipeline. The nodes on the other edge
represent a far-field surface, such as the ground surface, and are used to prescribe the far-field ground
motion. The elements are described in detail in ``Pipe-soil interaction elements,'' Section 18.7.1 of the
The purpose of this example is to determine the stress state along the length of a infinitely long buried
pipeline subjected to large fault movement of 1.52 m (5.0 ft), as shown in Figure 8.1.4-1. The pipeline
intersects the fault at 90.0°. The results are compared with results from an independent analysis, as
described below.
Problem description
The problem consists of an infinitely long pipeline buried at a depth of 6.1 m (20.0 ft.) below the
ground surface. Only a 610.0 m (2000.0 ft.) long section of the pipeline is modeled. The outside
diameter of the pipe is 0.61 m (24.0 in), and the wall thickness is 0.0254 m (1.0 in). The pipeline is
modeled with 50 first-order PIPE21 elements. A nonuniform mesh, with smaller elements focused near
the fault, is used.
The pipe-soil interaction behavior is model with PSI24 elements. The PSI elements are defined so that
one edge of the element shares nodes with the underlying pipe element, and the nodes on the other
edge represent a far-field surface where ground motion is prescribed. The far-field side and the side
that shares nodes with the pipeline are defined by the element connectivity.
A three-dimensional model that uses PIPE31 and PSI34 elements is also included for verification
purposes.
Material
The pipeline is made of an elastic-perfectly plastic metal, with a Young's modulus of 206.8 GPa (30 ´
106 lb/in2), a Poisson's ratio of 0.3, and a yield stress of 413.7 MPa (60000 lb/in 2).
The pipe-soil interaction behavior is elastic-perfectly plastic. The *PIPE-SOIL STIFFNESS,
TYPE=NONLINEAR option is used to define the interaction model. The behavior in the vertical
direction is assumed to be different from the behavior along the axial direction. It is further assumed
that the pipeline is buried deep below the ground surface so that the response is symmetric about the
origin. ABAQUS also allows a nonsymmetric behavior to be defined in any of the directions (this is
usually the case in the vertical direction when the pipeline is not buried too deeply). The ultimate force
per unit length in the axial direction is 730.0 N/m (50.0 lb/ft), and in the vertical direction it is 1460.0
N/m (100.0 lb/ft). The ultimate force is reached at 0.0304 m (0.1 ft) in both the horizontal and vertical
directions.
The loading occurs in a plane (axial-vertical), so the properties for the pipe-soil interaction behavior in
the transverse horizontal direction are not important.
Loading
8-1138
Soils Analyses
The loading on the pipeline is caused by a relative vertical displacement 1.52 m (5.0 ft) along the fault
line. It is assumed that the effect of the vertical ground motion decreases linearly over a distance of
91.4 m (300.0 ft.) from the origin of fault, as shown in Figure 8.1.4-1.
This linear distribution of ground motion is prescribed as follows. Rigid ( R2D2) elements are
connected to the far-field edges of the PSI to create two rigid surfaces, one on each side of the fault
line. These surfaces extend a distance of 91.4 m (300.0 ft.) from the origin of the fault. The rigid body
reference nodes are also placed a distance of 91.4 m (300.0 ft.) from the fault on the ground surface.
The fault movement is modeled by prescribing a rotation to each of the rigid body reference nodes so
that a positive vertical displacement of 0.76 m (2.5 ft) is obtained on one side of the fault and a
negative vertical displacement of 0.76 m (2.5 ft) is obtained on the other side of the fault, as shown in
Figure 8.1.4-2. All degrees of freedom on the remaining far-field nodes are fully fixed. In addition, the
two end points of pipeline are fully fixed. Figure 8.1.4-2 does not show the PSI elements or any of the
remaining nodes on the ground surface.
Reference solution
The reference solution is obtained by using JOINTC elements between the pipeline and ground nodes
to model the pipe-soil interaction. These elements provide an internal stiffness, which is modeled with
linear or nonlinear springs; nonlinear springs are used in this example. The behavior of the nonlinear
spring is elastic in the sense that reversed loading does not result in permanent deformation. This
behavior is different from the behavior provided by the nonlinear PSI elements. However, this is not a
limitation in this example since the loading is monotonic.
Another distinct difference between JOINTC elements and PSI elements is that the spring behavior
associated with JOINTC elements is defined in terms of total force, whereas the constitutive behavior
for PSI elements is defined as a force/unit length. This difference requires us to define a separate
stiffness for each JOINTC element or to use a uniform mesh with JOINTC elements spaced at unit
length intervals along the pipeline. A unit length mesh is used in this example.

Figure 8.1.4-3 and Figure 8.1.4-4 show the axial and vertical forces per unit length applied to the
pipeline due to relative ground motion. The figures show that permanent deformation occurs in the
pipe-soil interaction model near the fault along the axial and horizontal directions, with purely elastic
behavior further from the fault.
Figure 8.1.4-5 compares the axial stress in the bottom wall of the pipeline with the reference solution.
The figure shows that the pipeline behavior is purely elastic. The figure also shows close agreement
with the reference solution. The small differences between the solutions can be accounted for by the
different mesh densities. The reaction forces at the pipeline edges and the maximum pipeline
displacements are also in close agreement with the reference solution.
Input files
buriedpipeline_2d.inp
Two-dimensional model using PSI24 elements.
8-1139
Soils Analyses
buriedpipeline_3d.inp
Three-dimensional model using PSI34 elements.
buriedpipeline_ref.inp
Reference solution using JOINTC elements.
Reference
· Audibert, J. M. E., D. J. Nyman, and T. D. O'Rourke, "Differential Ground Movement Effects on
Buried Pipelines," Guidelines for the Seismic Design of Oil and Gas Pipeline Systems, ASCE
publication, pp. 151-180, 1984.
Figures
Figure 8.1.4-1 Pipe with fault motion.
Figure 8.1.4-2 Displaced shape (magnification factor=10.0).
Figure 8.1.4-3 Axial force/unit length applied along the pipeline.
8-1140
Soils Analyses
Figure 8.1.4-4 Vertical force/unit length applied along the pipeline.
Figure 8.1.4-5 Axial stress along the bottom of the pipeline.
8-1141
Soils Analyses
Sample listings
8-1142
Soils Analyses
Listing 8.1.4-1
*HEADING
RELATIVE FAULT MOTION ACCROSS BURIED PIPELINE
PSI24 elements Units: N, m
*PRE PRINT,ECHO=NO,MODEL=YES,CONTACT=YES,
HISTORY=YES
*NODE
1, -91.4, 6.1
2, 91.4, 6.1
101, -305.0, 0.0
151, 305.0, 0.0
501, -305.0, 6.1
551, 305.0, 6.1
*NODE,NSET=NL
111, -91.4, 0.0
511, -91.4, 6.1
*NODE,NSET=NC
126, 0.0, 0.0
526, 0.0, 6.1
*NODE,NSET=NR
141, 91.4, 0.0
541, 91.4, 6.1
*NGEN,NSET=ALL
101, 111
141, 151
501, 511
541, 551
*NFILL,NSET=ALL,BIAS=1.111111
NL, NC, 15
*NFILL,NSET=ALL,BIAS=0.900000
NC, NR, 15
*NSET,NSET=PIPE,GEN
101, 151
*NSET,NSET=SURF,GEN
501, 510
526, 526
542, 551
*NSET,NSET=ENDS
101, 151
*ELEMENT,TYPE=PIPE21
101, 101, 102
*ELGEN,ELSET=PIPELINE
8-1143
Soils Analyses
101, 50
*ELSET,ELSET=PIPEPLOT,GEN
111,140
*ELEMENT,TYPE=PSI24
501, 101, 102, 502, 501
*ELGEN,ELSET=SOIL
501, 50
*ELEMENT,TYPE=R2D2
1111, 511, 512
1127, 527, 528
*ELGEN,ELSET=SURFL
1111, 14
*ELGEN,ELSET=SURFR
1127, 14
*BEAM SECTION,SECTION=PIPE,ELSET=PIPELINE,
MATERIAL=STEEL
0.61, 0.0254
*ELASTIC
206.8E6, 0.3
*PLASTIC
413.7E6
*PIPE-SOIL INTERACTION,ELSET=SOIL
*PIPE-SOIL STIFFNESS,TYPE=NONLINEAR,DIR=AXIAL
-730.0, -0.0304
0.0, 0.0
730.0, 0.0304
*PIPE-SOIL STIFFNESS,TYPE=NONLINEAR,DIR=VERTICAL
-1460.0, -0.0304
0.0, 0.0
1460.0, 0.0304
*PIPE-SOIL STIFFNESS,TYPE=NONLINEAR,DIR=HORIZONTAL
-1460.0, -0.0304
0.0, 0.0
1460.0, 0.0304
*RIGID BODY,ELSET=SURFL,REFNOD=1
*RIGID BODY,ELSET=SURFR,REFNOD=2
1: APPLY FAULT MOTION
*STATIC
*BOUNDARY
SURF, 1, 2
ENDS, 1, 2
8-1144
Soils Analyses
ENDS, 6
1, 1, 2
2, 1, 2
1, 6, , 0.0083368
2, 6, , 0.0083368
*NODE PRINT,FREQ=10
U,
RF,
*ELPRINT,FREQ=10,ELSET=SOIL
S, PE
COORD,
*ELPRINT,FREQ=10,ELSET=PIPEPLOT
1
S, PE
COORD,
*EL FILE,ELSET=PIPEPLOT
S
*EL FILE,ELSET=SOIL
S, PE, E
NFORC,
*ELEMENT OUTPUT,ELSET=PIPELINE
S,E,EE,EP
*NODE OUTPUT,NSET=ALL
U,
*NODE OUTPUT,NSET=ENDS
U,RF
*END STEP
8-1145
ABAQUS/Aqua Analyses
9. ABAQUS/Aqua Analyses
9.1 ABAQUS/Aqua analyses
9.1.1 Jack-up foundation analyses
Products: ABAQUS/Standard ABAQUS/Aqua
This example simulates a jack-up rig on a sand foundation subjected to alternating wind loading.
Geometry and model

The model--a simplified planar model for the analysis of a multiple leg, portal frame-type structure--is
intended for the analysis of 3-leg jack-up rigs with shallow foundation supports. Figure 9.1.1-1is a
schematic of a 3-leg jack-up, as represented by the model. The jack-up hull is assumed to be rigid and
triangular, and the connection between the hull and the legs is also taken to be rigid. The jack-up has
two windward legs and one leeward leg; the model is projected onto the vertical symmetry plane that
passes through the leeward leg and between the windward legs. Elastic beam columns are used to
model both the upper and lower segment of each leg. The soil model is chosen to be macro-yield sand.
Three degrees of freedom--vertical, horizontal, and rotational--are assumed at each spud can at the
base of each leg. Mass is assumed to be concentrated at the center of the hull. The horizontal degree of
freedom at the center is assumed to represent the motion of the rig for analysis purposes. Wind loading
on the rig is applied as a horizontal force below the center of gravity of the hull.
The leg segments are modeled using B21 elements, and the *BEAM GENERAL SECTION option is
used to define the structural properties of the beam. The interaction between the spud can and the soil
is modeled through JOINT2D elements and the *JOINT ELASTICITY and *JOINT PLASTICITY
options. Rigid beam elements, RB2D2, are used to model the rigid hull.
The dimensions of the rig and the material properties of the sand and the spud can are as follows (force
units are in kN, and length units are in meters):
Leg length upper segment 49.4
Leg length lower segment 13.5
Leg EI upper segment 2.7 ´ 108
Leg EI lower segment 2.7 ´ 109
Leg AE upper segment 2.2 ´ 108
Leg AE lower segment 2.2 ´ 109
Leg GA upper segment 8.1 ´ 107
Leg GA lower segment 8.1 ´ 108
Horizontal distance from platform
center of gravity to leeward leg 23.4
Horizontal distance from platform
center of gravity to windward leg 11.7
Spud can diameter 10.9
Spud can cone angle 180°
Foundation preload per spud can 50600
Foundation tensile capacity 0
Operational vertical load (weight) 62700
9-1146
Vertical distance from center of

gravity
to load application point 5.7
Soil submerged unit weight 10.0
Soil friction angle 33°
Soil Poisson's ratio 0.2
Foundation elastic shear moduli, 5.14 ´ 104
Gºº
Ghh 3.87 ´ 103
Grr 2.04 ´ 103
Constant coefficient, ¤1 1.0
Constant coefficient, ¤2 0.5

The base nodes of the JOINT2D elements are always fixed. The required preload is applied to each
spud can using the *INITIAL CONDITIONS option. In the first step the weight loading is applied at
the center of gravity of the hull. The rig is then subjected to an alternating horizontal wind loading
applied at the specified location below the center of gravity of the hull. The load is applied by using
the *CLOAD option.
The rig is loaded from zero to 5370 kN, unloaded to zero and then to 6440 kN in the opposite
direction, reloaded to 9130 kN in the initial direction, unloaded and reloaded to 9770 kN in the
opposite direction, and unloaded to zero again. Each of these loadings is done in a separate step and is
ramped from zero to the specified magnitude at the end of the step.

The estimated load path for the leeward spud can foundation is plotted in a graph of equivalent
p
horizontal load, R = (M=D )2 + ¤1 H 2 , versus V =Vc : The plot is shown in Figure 9.1.1-2 and is in
good agreement with the load path predicted by an independent analysis, as detailed in the reference
below. The moment-horizontal load response (i.e., M=D versus H) for the leeward spud can
foundation, shown in Figure 9.1.1-3, compares well with the independent analysis.
Input file
jackup.inp
Input data for this example.
Reference
· Wong, P. C. and J. D. Murff, "Dynamic Analysis of Jack-Up Rigs Using Advanced Foundation
Models," Proceedings, 13th International Conference on Offshore Mechanics and Arctic
Engineering (OMAE), vol. 2 - Safety and Reliability, Houston, pp. 93-109, February 1994.
Figures
9-1147
Figure 9.1.1-1 Schematic representation of jack-up rig.
Figure 9.1.1-2 Load path for leeward spud can.
Figure 9.1.1-3 Moment versus horizontal load for leeward spud can.
9-1148
Sample listings
9-1149
Listing 9.1.1-1
*HEADING
3-LEG JACK UP ON SAND --- CYLINDRICAL
**
**NODE DEFINITIONS
**
*NODE,NSET=BEAMS
11,0.
12,0.,13.5,0.
13,0.,62.9,0.
14,35.1,0,0.
15,35.1,13.5,0.
16,35.1,62.9,0.
17,11.7,62.9,0.
18,11.7,70.0,0.
999,11.7,75,0.
*NODE,NSET=BASE
1, 0.
2, 0.
3,35.1
*NSET,NSET=EPJ
BASE,11,14
**
**ELEMENT DEFINITIONS
**
*ELEMENT,ELSET=BEAML,TYPE=B21
1,11,12
3,11,12
5,14,15
*ELEMENT,ELSET=BEAMU,TYPE=B21
2,12,13
4,12,13
6,15,16
*ELEMENT,ELSET=RBEAMS,TYPE=RB2D2
11,13,17
12,17,16
13,17,18
*ELEMENT,TYPE=JOINT2D, ELSET=JOINT2D
101,1,11
102,2,11
103,3,14
**
9-1150
**ORIENTATIONS AND TRANSFORMATIONS

**
*ORIENTATION,NAME=ORI1
0,1,0,1,0,0
*TRANSFORM,NSET=EPJ
0,1,0,1,0,0
**
**PROPERTY DEFINITIONS
**
*BEAM GENERAL SECTION,ELSET=BEAMU,SECTION=GENERAL
1.0,1.2273,0.,1.2273,2.4545,
2.2E8,8.1E7
*BEAM GENERAL SECTION,ELSET=BEAML,SECTION=GENERAL
1.0,1.2273,0.,1.2273,2.4545,
2.2E9,8.1E8
*RIGID BODY,ELSET=RBEAMS,REF NODE=999
**
*EPJOINT,ELSET=JOINT2D,ORIEN=ORI1,SECTION=SPUD
10.9,0.
*JOINT ELASTIC,NDIM=2,MODULI=SPUD CAN
5.14E4, 3.87E3, 2.04E4, 0.2
*JOINT PLASTIC,TYPE=SAND
0.,1,0.5,33.0,10.0
**
*INITIAL CONDITIONS, TYPE=SPUD PRELOAD
JOINT2D,50600
**
**
*STEP,INC=10000
APPLY WEIGHT
*STATIC
1.0,1.0
*BOUNDARY
1,1,6,0
2,1,6,0
3,1,6,0
**
** Apply weight
**
*CLOAD
17,2,-62700
9-1151
*EL PRINT,ELSET=JOINT2D,FREQ=0
S,
PE,
*EL FILE,ELSET=JOINT2D,FREQ=20
S,
E,
PE,
*NODE FILE,FREQ=20
U,
RF,
*NODE OUTPUT, NSET=EPJ
U,RF
*ELEMENT OUTPUT, ELSET=JOINT2D
S,E,PE
*OUTPUT, HISTORY, FREQUENCY=10
*NODE OUTPUT, NSET=EPJ
U,RF
*ELEMENT OUTPUT, ELSET=JOINT2D
S,E,PE
*END STEP
*STEP,INC=10000
ALTERNATING LOADING ANALYSIS FIRST STEP
*STATIC
0.001,1.0,,0.005
*BOUNDARY
1,1,6,0
2,1,6,0
3,1,6,0
*CLOAD
18,1,5370
*END STEP
**
**
**
*STEP,INC=10000
UNLOADING TO ZERO
*STATIC
0.001,1.0,,0.005
*BOUNDARY
1,1,6,0
2,1,6,0
3,1,6,0
9-1152
*CLOAD
18,1,0
*END STEP
**
**
**
*STEP,INC=10000
LOADING IN THE OPPOSITE DIRECTION
*STATIC
0.001,1.0,,0.005
*BOUNDARY
1,1,6,0
2,1,6,0
3,1,6,0
*CLOAD
18,1,-6440
*END STEP
**
**
**
*STEP,INC=10000
RELOADING IN THE INITIAL DIRECTION
*STATIC
0.001,1.0,,0.005
*BOUNDARY
1,1,6,0
2,1,6,0
3,1,6,0
*CLOAD
18,1,9130
*END STEP
**
**
**
*STEP,INC=10000
UNLOADING TO ZERO (2)
*STATIC
0.001,1.0,,0.005
*BOUNDARY
1,1,6,0
2,1,6,0
3,1,6,0
*CLOAD
9-1153
18,1,0
*END STEP
**
**
**
*STEP,INC=10000
REVERSE LOADING
*STATIC
0.001,1.0,,0.005
*BOUNDARY
1,1,6,0
2,1,6,0
3,1,6,0
*CLOAD
18,1,-9770
*END STEP
**
**
**
*STEP,INC=10000
UNLOADING BACK TO ZERO
*STATIC
0.001,1.0,,0.005
*BOUNDARY
1,1,6,0
2,1,6,0
3,1,6,0
*CLOAD
18,1,0
*END STEP
9.1.2 Riser dynamics

Products: ABAQUS/Standard ABAQUS/Aqua
Pipelines extending from the sea floor to the ocean surface (risers) are subject to many types of load:
self-weight, buoyancy, internal and external pressure, tensile forces arising from surface moorings,
current drag, and oscillatory loads resulting from wave motion. The response of a riser to these loads is
complex, and the difficulty of such analysis is heightened by the relative length of such pipelines (deep
water risers). In this example a riser is analyzed under conditions specified by the American Petroleum
Institute for comparison of drilling riser analyses (API BULLETIN 2J, 1977), and the results are
compared with the results shown in that publication.
Geometry and model
9-1154
The riser is shown in Figure 9.1.2-1. Its length is 463.3 m (1520 ft), and it stands in 448.1 m (1470 ft)
of water. The outer diameter of the riser is 405 mm (1.33 ft), and it has a wall thickness of 15.88 mm
(0.0521 ft). The pipeline is made of steel, with a Young's modulus of 206.8 GPa (4.32 ´ 109 lb/ft2) and
a density of 11508.685 kg/m 3 (22.332 lb-s 2/ft4). The riser is modeled with 10 beam elements of type
B21. No mesh convergence studies have been performed; hence, more elements may be required for
accurate prediction of the stress in the riser.
Loading
The riser has a weight of 2575 N/m (176.36 lb/ft) and is loaded by a top tension of 2.224 MN (5 ´ 105
lb). Drag loading is applied by a steady current flowing by the riser with a velocity distribution varying
linearly from 0.257 m/s (0.844 ft/s) at the mean water level to zero at the base of the riser. The
coefficients in Morison's equation are transverse drag coefficient ( CD ) 0.7, tangential drag coefficient
(CT ) 0.0, and transverse inertia coefficient (CM ) 1.5.
The effective outer diameter for the drag calculations is 0.66 m (2.167 ft). Waves of peak to trough
height 6.1 m (20 ft) travel across the water surface with a period of 9 seconds; these are modeled with
the Airy wave theory provided in the *AQUA option (`ÀBAQUS/Aqua analysis,'' Section 6.10.1 of
the ABAQUS/Standard User's Manual). The density of the fluid is taken to be 1021 kg/m 3 (1.982
lb-s2/ft4). In ABAQUS/Aqua, user subroutine UWAVE can be used to specify user-defined wave
kinematics. We illustrate this capability by repeating this analysis with a user-specified Airy wave
theory that is identical to the built-in Airy wave option in ABAQUS/Aqua.
Boundary conditions
The base of the riser is "gimballed," supporting no moments. The top of the riser has two motions
prescribed: an initial offset of 13.716 m (45 ft) from the vertical position of the riser and a sinusoidal
motion about this static configuration, representing the surge of a vessel attached to the riser, with
peak-to-peak amplitude of 1.22 m (4 ft) and a period of 9 seconds. The vessel surge is 15° out of phase
with the surface waves.
Analysis
The analysis is done in two steps. The first is the static step, in which the top tension is applied and the
riser is moved from the vertical to its offset position by specifying the necessary horizontal
displacement at the top of the pipeline. The top tension is 2.224 MN (5 ´ 105 lb).
In the second step, which is a dynamic step, the time increment is chosen as a fixed value of 0.125
second. The prescribed displacement at the top of the riser has a 9-second period, so this time step
should provide reasonably accurate time integration once the higher modes are damped out by the fluid
drag. The "half-step residual" values calculated by ABAQUS provide a measure of accuracy of the
solution, and these values are typically of order 4.4 kN (1000 lb). Since these values are smaller than
typical actual forces, they suggest that the time integration is reasonably accurate.

The initial static step, which moves the riser to its offset position and applies the static loads, is
9-1155
completed in four increments. The first increment requires more iterations than subsequent increments,
which is typical of this class of problem: the riser is initially unstressed and, therefore, is highly
flexible. After some loading is applied, the axial tension stabilizes the system, and convergence is
more rapid.
At the end of the static step the top of the riser makes an angle of 1.17° with the vertical. This value
agrees well with the value of 1.20° presented in API BULLETIN 2J (1977). The angle predicted at the
base of the riser is 2.48°, which compares to 2.55° reported in the API bulletin. The slight
discrepancies are attributed to the relative coarseness of the model.
The dynamic solution is carried out for 18 seconds of response. Typically one equilibrium iteration is
required in each of the time increments. Half-step residual values for the first few increments are of
order 178 MN (4.0 ´ 107 lb), and at the end of the run they are of order 4.4 kN (1000 lb). This result is
typical: initially there is much high frequency content in the solution, which is reflected in the larger
half-step residual values. As the analysis proceeds, the fluid drag dissipates this "noise," the solution
becomes smoother, and the half-step residual values drop accordingly.
The envelope of pipeline excursions during the dynamic analysis is plotted in Figure 9.1.2-2, and the
envelope of bending stress is shown in Figure 9.1.2-3. These results are in basic agreement with those
given in the API bulletin.
As expected, the results obtained by the model with the Airy wave theory implemented in user
subroutine UWAVE are identical to those due to the built-in Airy wave option.
Input files
riserdynamics_airy_disp.inp
Analysis with the Airy wave theory. User subroutine DISP is used to prescribe the sinusoidal
surge motion. This motion could be prescribed instead through the use of the *AMPLITUDE
option. User subroutine DISP is used to illustrate the use of this routine to prescribe a nonzero
boundary condition value.
riserdynamics_airy_disp.f
User subroutine DISP used in riserdynamics_airy_disp.inp.
riserdynamics_wavedata.inp
Wave data for use in riserdynamics_airy_disp.inp.
riserdynamics_stokes_disp.inp
Analysis with the Stokes wave theory.
riserdynamics_stokes_disp.f
User subroutine DISP used in riserdynamics_stokes_disp.inp.
riserdynamics_airy_disp_uwave.inp
Analysis with the Airy wave theory implemented in user subroutine UWAVE.
9-1156
riserdynamics_airy_disp_uwave.f
User subroutines UWAVE and DISP used in riserdynamics_airy_disp_uwave.inp.
Reference
· American Petroleum Institute, "Comparison of Marine Drilling Riser Analyses," API Bulletin 2J,
Washington, D. C., January 1977.
Figures
Figure 9.1.2-1 Riser problem definition.
Figure 9.1.2-2 Horizontal displacement envelope during dynamic response.
9-1157
Figure 9.1.2-3 Bending stress envelope during dynamic response.
9-1158
Sample listings
9-1159
Listing 9.1.2-1
*HEADING
RISER DYNAMIC CHECK OUT - AIRY THEORY
NEEDS FILE riserdynamics_wavedata.inp FOR
WAVE DATA
*NODE ,NSET=ENDS
1,,1520.
11,0.
*NGEN
1,11
*ELEMENT,TYPE=B21
1,1,2
*ELGEN,ELSET=PIPE
1,10
*ELSET,ELSET=PIPE1,GEN
1,5,1
*ELSET,ELSET=PIPE2,GEN
6,10,1
*BEAM GENERAL SECTION,SECTION=PIPE,DENSITY=22.332,
ELSET=PIPE
.6667,0.0521
,,-1.
4.32E9,2.16E9
*AQUA
0.,1470.,32.2,1.982
0.,0.,0.,0.
0.844,0.,0.,1470.
*WAVE,TYPE=AIRY,INPUT=riserdynamics_wavedata.inp
*STEP,NLGEOM
INITIAL OFFSET
*STATIC
.25E-5,1.E-5
*DLOAD
PIPE,PY,-176.36
*****
** Aqua Normal drag load
PIPE,FDD, 1., 2.167, 0.7, 1.
**
** Aqua Inertial drag load
PIPE, FI, 1., 2.167, 1.5, 0.5
**
9-1160
*CLOAD
1,2,5.E5
*BOUNDARY
11,1,2
*BOUNDARY,USER
1,1
*NODE FILE,NSET=ENDS,FREQUENCY=1
U,RF
*PRINT,RESIDUAL=NO
*ENERGY PRINT
*ENERGY FILE
*ENERGY PRINT,ELSET=PIPE1
*ENERGY FILE,ELSET=PIPE1
*EL PRINT,ELSET=PIPE1,TOTALS=YES
ELEN,
SF,
SF,
U,RF
U,RF
*END STEP
DYNAMICS
*DYNAMIC
.125,18.
SF,
SF,
*EL PRINT,ELSET=PIPE1,TOTALS=YES
ELEN,
*END STEP
9-1161
Underwater Shock Analyses
10. Underwater Shock Analyses

10.1 Underwater shock analyses
10.1.1 The cylinder whip problem
Products: ABAQUS/Standard ABAQUS/USA
An important phase of response for underwater structures is the late-time response excited by the flow
field created from an explosion. The response of long underwater structures to these late-time
excitations is often referred to as whipping. The capability exists within USA to analyze these late-time
responses due to the flow field created by an explosion. The explosion creates a gas bubble that
pulsates and (possibly) migrates toward the surface. USA will calculate the velocity potential due to
the bubble, but the user must provide certain information about the explosive event. The user selects a
bubble-pulse excitation by using the BUBBLE parameter on the *USA INCIDENT PRESSURE
option. Once this has been selected, the user must then provide the TNT equivalent charge weight of
the explosion, a bubble drag coefficient, a length scale factor ( USA bubble calculations have some
embedded constants that use units of feet), a bubble cutoff time, and the elevation of the free surface.
In addition, the user can indicate that the bubble is allowed to migrate toward the free surface by using
the BUBBLE=MIGRATE parameter on the *USA INCIDENT PRESSURE option. Free surface
corrections will not be performed unless the fluid surface information is given on the *USA FLUID
PROPERTIES option. If the fluid surface data are given on the *USA FLUID PROPERTIES option,
USA will use a pulsating and migrating bubble excitation regardless of whether the user specifies
BUBBLE or BUBBLE=MIGRATE. For more information on bubble whip and the bubble-pulse
calculations, see DeRuntz (1989) and DeRuntz et al. (1986).
Geometry and model

The model consists of an open-ended cylindrical shell submerged 30.48 m (1200 in) below the surface.
The origin of the coordinate system is at the cylinder's center, and only a quarter-model of the cylinder
is required. For this three-dimensional analysis the fluid elevation direction must be oriented along the
positive z-axis. The positive x-axis is along the axis of the cylinder. The explosive charge lies 15.24 m
(600 in) directly below the cylinder's center. The charge standoff point (structural point nearest the
charge) is at the lower surface of the cylinder.
Shell finite element model

The cylindrical shell is modeled with S4R5 elements, using eight elements around the
half-circumference and 10 along the half-length. Eighty USI4 interface elements are used to represent
the fluid surface. Since both XSYMM and YSYMM are specified on the *USA FLUID PROPERTIES
option, the quarter-model must be defined in the positive x-half-plane and the positive y-half-plane.
The overall geometry of the problem is shown in Figure 10.1.1-1, while the quarter-shell mesh is
shown in Figure 10.1.1-2. The input file for the S4R5/USI4 model is shown in ucw_usi4_p.inp.
In the modeling approach with USI4 elements there is a one-to-one correspondence between USI
interface and structural elements. For a given structural (or solid model) it is also possible to define a
10-1162
coarser fluid model using fluid element "overlay" interface elements. This approach is effective for
reducing the size of the fluid element model in regions of small fluid pressure gradient and relatively
large curvature. ABAQUS/USA supports two types of fluid element overlay elements: USI6O is a
6-node interface element for overlaying two structural elements, and USI9O is a 9-node interface
element for overlaying four structural elements. Example fluid element overlay models are included
with the ABAQUS release. Refer to the corresponding input files listed in the "Input files" section.
Beam element model

The same problem is modeled using beam elements to discretize the structure and surface of revolution
USI2SOR elements to model the fluid surface. In this case the total number of nodes is reduced to 11.
Ten B31 elements are used along the half-length of the cylinder. A pipe cross-section is used for the
beams, with the pipe radius and thickness given as 4.572 m (180.0 in) and 50.8 mm (2.0 in). Ten
USI2SOR elements are used to model the interface to the surrounding fluid. Cosine and sine terms are
selected for the SOR Fourier coefficients. Since this is a half-model (as opposed to the shell
quarter-model), only XSYMM is used for the *USA FLUID PROPERTIES and *BOUNDARY
specifications. All other aspects of the geometry and fluid model remain the same as above.
This model requires comparatively fewer degrees of freedom and will run much faster. However, it is
also limited in scope since it cannot model any local deformations (cross-sectional deformations). The
input file for this model is shown in ucw_b31_p.inp.
A slightly more realistic model can be generated using elbow elements in place of the beams. This
allows some cross-sectional deformations of the structure to be represented. However, only nodal
forces (from the fluid) are applied to the structure, and ovalization of the structure will not affect the
fluid model (or its calculation of fluid pressure). Files ucw_s4_p.inp and ucw_s4_pm.inp contain
elbow/SOR models.

The results for both models are discussed below.
Shell models
The velocity history and strain history results obtained by ABAQUS/USA with the S4R5/USI4 model
are compared to the USA-STAGS solution. The velocity histories of two points at the cylinder
midspan (XSYMM plane) are shown in Figure 10.1.1-3. The upper plot shows the radial velocity
history for the point closest to the charge, node 9 in Figure 10.1.1-2. The lower plot shows the radial
velocity history for the cylinder midspan location away from the charge, node 1. The ABAQUS/USA
and USA-STAGS velocity results compare very well for the low-frequency bending behavior. The
difference in the high-frequency behavior is probably the result of differences in the formulations of
the ABAQUS shell element (reduced integration) and the STAGS shell element (fully integrated).
Experimentation with this example problem has shown that the velocity results (especially the high
frequency content) change significantly with a change in the time increment size. The problem is
sensitive enough to slight numerical changes that the changing word length (precision) from one
machine to another may cause small differences in the velocity response.
10-1163
Strain history results at six locations along the length are shown in Figure 10.1.1-4. Strain histories (a)
- (b) are taken from the "back" side of the cylinder, from the centroids of elements 1, 11, 21, 31, 41,
and 51, respectively. The strain histories show very good agreement in the low-frequency range and
moderate agreement in the high-frequency range.
All of the USA-STAGS data shown for comparison were obtained from De Runtz (1991) and are the
same data used in DeRuntz et al. (1986).
In addition, results obtained for the model with 80 USI4 interface elements are compared against
results obtained using a model with 20 USI9O fluid overlay interface elements and a model with 40
USI6O elements. All models employ the same structural model. For these models the velocity histories
at the node closest to the charge are shown in Figure 10.1.1-5to be in good agreement, with any
differences being caused by the different fluid models.
Beam/USI2SOR model
The velocity history at selected nodal points from the ABAQUS/USA solution is compared against the
USA-STAGS solution. The z-direction nodal velocity history of the cylinder symmetry plane (node 1)
is shown in Figure 10.1.1-6. The ABAQUS/USA results compare well to the USA-STAGS results.
The USA-STAGS data shown for comparison were obtained from DeRuntz (1994) and are the same
data used in DeRuntz et al. (1986).
Input files
Pulsating bubble excitation:
ucw_b31_p.inp
10-element B31/USI2SOR model.
ucw_10b31_p.inp
10-element multibranch B31/USI2SOR model.
ucw_13b31_p.inp
13-element multibranch B31/USI2SOR model.
ucw_usi2sor_0.inp
10-element ELBOW31/USI2SOR model, 0 ovalization modes.
ucw_usi2sor_4.inp
10-element ELBOW31/USI2SOR model, 4 ovalization modes.
ucw_s3r_p.inp
160-element S3R/USI3 element model.
ucw_s4_p.inp
80-element S4/USI4 element model.
10-1164
ucw_s4r_p.inp
ucw_usi4_p.inp
80-element S4R5/USI4 model.
ucw_usi90_p.inp
80 S4R5 and 20 USI9O elements.
ucw_usi60_p.inp
80 S4R5 and 40 USI6O elements.
ucw_stri3_p.inp
160-element STRI3/USI3 element model.
ucw_mixed_ele.inp
Mixed element model that uses shells and beams to model the structure and uses USI3, USI4, and
USI2SOR elements to represent the fluid interface.
Pulsating and migrating bubble excitation:
ucw_s3r_pm.inp
ucw_s4_pm.inp
ucw_s4r_pm.inp
ucw_usi4_pm.inp
ucw_stri3_pm.inp
Pulsating and migrating bubble excitation with free surface corrections:
ucw_s3r_pf.inp
ucw_s4_pf.inp
ucw_s4r_pf.inp
10-1165
ucw_usi4_pf.inp
ucw_stri3_pf.inp
References
· DeRuntz, J. A., Jr., Private Communication, 1991.
· DeRuntz, J. A., Jr., Private Communication, 1994.
· DeRuntz, J. A., Jr., "The Underwater Shock Analysis Code and its Applications, " 60th Shock and
Vibration Symposium Proceedings, vol. 1, pp. 89-107, 1989.
· DeRuntz, J. A., Jr., C. C. Rankin, F. A. Brogan, and M. E. Reglebrugge, "Enhanced Analysis

Capability in USA-STAGS," LMSC-D062084, Lockheed Palo Alto Research Laboratory, 1986.
Figures
Figure 10.1.1-1 Cylinder whip overall geometry.
Figure 10.1.1-2 Cylinder whip quarter-shell model.
10-1166
Figure 10.1.1-3 Shell model velocity time histories. Solid line: ABAQUS/USA (S4R5) results.
Dashed line: USA-STAGS results.
10-1167
Figure 10.1.1-4 Shell model strain time histories. Solid line: ABAQUS/USA (S4R5) results. Dashed
line: USA-STAGS results.
10-1168
Figure 10.1.1-5 General and fluid overlay interface element model velocity time histories. Solid line:
Reference USI4 results. Dashed line: USI9O results. Dotted line: USI6O results.
10-1169
Figure 10.1.1-6 SOR model velocity time history. Vertical velocity of the X-symmetry plane. Solid
line: ABAQUS/USA results. Dashed line: USA-STAGS results.
Sample listings
10-1170
Listing 10.1.1-1
*HEADING
NSWC BUBBLE-PULSE CYLINDER WHIP
CYLINDER QUARTER MODEL
80 element S4R5 model
all dimensions in inches
Cylindrical Pipe;length=3600.0,radius=180.0,
thickness = 2.0
Cylindrical Pipe Properties;E=30E6,nu=0.3,
rho=1.6274
Fluid properties;rho=9.346E-5,c=60000.0
Infinite Fluid, Pulsating (Non-Migrating)
Bubble Excitation
*NODE, NSET=ENDS
1, 0.0, 0.0, 180.0
9, 0.0, 0.0, -180.0
1001, 0.0, 0.0, 0.0
101, 1800.0, 0.0, 180.0
109, 1800.0, 0.0, -180.0
2001, 1800.0, 0.0, 0.0
*NGEN, LINE=C, NSET=MID
1, 9,, 1001,,,, -1.0, 0.0, 0.0
*NGEN, LINE=C, NSET=END
101, 109,, 2001,,,, -1.0, 0.0, 0.0
*NFILL, NSET=ALL
MID, END, 10, 10
*NSET, NSET=EDGES, GENERATE
1, 101, 10
9, 109, 10
**
** Structural Shell Elements
**
*ELEMENT, TYPE=S4R5
1, 1, 11, 12, 2
*ELGEN, ELSET=CYL
1, 8, 1, 1, 10, 10, 10
*SHELL SECTION, ELSET=CYL, MATERIAL=STEEL
2.0,
*ELASTIC
30.0E6, 0.3
10-1171
*DENSITY
4.2117E-3,
**
** USA structural interface element definitions
**
*ELEMENT, TYPE=USI4
1001, 1, 11, 12, 2
*ELGEN, ELSET=FLUID
1001, 8, 1, 1, 10, 10, 10
*INTERFACE, ELSET=FLUID
*USA FLUID PROPERTIES, DAA2=0.0, XSYMM, YSYMM
9.346E-5, 60000.0
**
*BOUNDARY
MID, XSYMM
EDGES, YSYMM
**
** Step 1
**
*STEP
*USA ADDED MASS GENERATION
*END STEP
**
** Step 2
**
*STEP, INC=500
*DYNAMIC, ALPHA=0.0
0.002, 1.0
** bubble is pulsating only (no migration,
** no free surface corrections)
*USA INCIDENT PRESSURE, BUBBLE
0.0, 0.0, -600.0, 0.0, 0.0, -180.0
1000.0, 2.25, 0.083333, 0.57
1800.0,
*NSET, NSET=TEMP
1, 9
*NODE FILE, NSET=TEMP, FREQUENCY=1
V,
*NODE OUTPUT,NSET=TEMP
V,
10-1172
V,
*ELSET, ELSET=TEMP
1, 11, 21, 31, 41, 51
*EL FILE, ELSET=TEMP, FREQUENCY=1
3,
E,
*MONITOR, NODE=9, DOF=3, FREQUENCY=999
*END STEP
10-1173
Listing 10.1.1-2
*HEADING
NSWC BUBBLE-PULSE BEAM WHIP
CYLINDER HALF MODEL
10 element B31/USI2SOR model
thickness=2.0
rho=1.6274
Bubble Excitation
*NODE, NSET=MID
1, 0.0, 0.0, 0.0
*NODE, NSET=END
11, 1800.0, 0.0, 0.0
*NGEN, NSET=ALL
1, 11
**
** Structural Beam Elements
**
*ELEMENT, TYPE=B31, ELSET=CYL
1, 1, 2
*ELGEN, ELSET=CYL
1, 10, 1, 1
*BEAM SECTION, SECTION=PIPE, ELSET=CYL,
MATERIAL=STEEL
180.0, 2.0
*ELASTIC
30.0E6, 0.3
*DENSITY
4.2117E-3,
**
**
*ELEMENT, TYPE=USI2SOR, ELSET=FLUID
1001, 1, 2
1002, 2, 3
1003, 3, 4
10-1174
1004, 4, 5
1005, 5, 6
1006, 6, 7
1007, 7, 8
1008, 8, 9
1009, 9,10
1010, 10,11
*SOR PROPERTIES, ELSET=FLUID, COSINE, SINE
0,
0.0, 180.0
*USA FLUID PROPERTIES, XSYMM, DAA2=0.0
9.346E-5, 60000.0
*BOUNDARY
MID, XSYMM
**
** Step 1
**
*STEP
*END STEP
**
** Step 2
**
*STEP, INC=500
*DYNAMIC, ALPHA=0.0
0.002, 1.0
** bubble is pulsating only (no migration, no
** free surface corrections)
0.0, 0.0, -600.0, 0.0, 0.0, -180.0
1000.0, 2.25, 0.083333, 0.57
1800.0,
*NSET, NSET=TEMP
1,
*NODE FILE, NSET=TEMP, FREQUENCY=1
V,
V,
10-1175
V,
*ELSET, ELSET=TEMP
1,
E,
*ELSET, ELSET=TEMP
1001,
*ELEMENT OUTPUT,ELSET=TEMP
PTOT,
*ELEMENT OUTPUT,ELSET=TEMP
PTOT,
PTOT,
*END STEP
10-1176
Listing 10.1.1-3
*HEADING
NSWC BUBBLE-PULSE CYLINDER WHIP
CYLINDER QUARTER MODEL
20 (5x4) USI9O fluid element, 80 S4R5 structural
element model
thickness=2.0
rho=1.6274
Bubble Excitation
*NODE, NSET=ENDS
1, 0.0, 0.0, 180.0
9, 0.0, 0.0, -180.0
1001, 0.0, 0.0, 0.0
101, 1800.0, 0.0, 180.0
109, 1800.0, 0.0, -180.0
2001, 1800.0, 0.0, 0.0
*NGEN, LINE=C, NSET=MID
1, 9,, 1001,,,, -1.0, 0.0, 0.0
*NGEN, LINE=C, NSET=END
101, 109,, 2001,,,, -1.0, 0.0, 0.0
*NFILL, NSET=ALL
MID, END, 10, 10
*NSET, NSET=EDGES, GENERATE
1, 101, 10
9, 109, 10
**
** Structural Shell Elements
**
*ELEMENT, TYPE=S4R5
1, 1, 11, 12, 2
*ELGEN, ELSET=CYL
1, 8, 1, 1, 10, 10, 10
*SHELL SECTION, ELSET=CYL, MATERIAL=STEEL
2.0,
*ELASTIC
10-1177
30.0E6, 0.3
*DENSITY
4.2117E-3,
**
**
*ELEMENT, TYPE=USI9O
1001, 1, 21, 23, 3, 11, 22, 13, 2, 12
*ELGEN, ELSET=FLUID
1001, 4, 2, 1, 5, 20, 10
*INTERFACE, ELSET=FLUID
*USA FLUID PROPERTIES, DAA2=0.0, XSYMM, YSYMM
9.346E-5, 60000.0
*BOUNDARY
MID, XSYMM
EDGES, YSYMM
**
**
** Step 1
**
*STEP
*END STEP
**
** Step 2
**
*STEP, INC=500
*DYNAMIC, ALPHA=0.0
0.002, 1.0
** bubble is pulsating only (no migration, no
** free surface corrections)
0.0, 0.0, -600.0, 0.0, 0.0, -180.0
1000.0, 2.25, 0.083333, 0.57
1800.0,
*NSET, NSET=NOUT
1, 9
V,
10-1178
V,
*NODE FILE, NSET=NOUT, FREQUENCY=1
V,
*ELSET, ELSET=ELF
1001, 1004
*EL FILE, ELSET=ELF, FREQUENCY=1
PTOT,
*ELEMENT OUTPUT,ELSET=ELF
PTOT,
*ELEMENT OUTPUT,ELSET=ELF
PTOT,
*FILE FORMAT, ZERO INCREMENT
*END STEP
10-1179
Postprocessing of ABAQUS Results Files
11. Postprocessing of ABAQUS Results Files

11.1 Postprocessing examples
11.1.1 User postprocessing of ABAQUS results files: overview
This chapter illustrates how to write and utilize postprocessing programs to manipulate data stored on
the ABAQUS results file. The results file, which is identified by the file extension .fil, is created by
the *CONTACT FILE, *EL FILE, *NODE FILE, *ENERGY FILE, *MODAL FILE, and
*RADIATION FILE options and contains results based on user-specified output requests. The standard
file format is binary, but it can be changed by user request for each run with the *FILE FORMAT,
ASCII option. Alternatively, it can be set to a default ASCII format during site installation. ABAQUS
uses FORTRAN unit 8 to communicate with the results file.
Sample postprocessing programs that perform commonly exercised tasks are presented in separate
sections in this chapter. These include merging multiple results files and converting the resulting
results file from binary format to ASCII, or vice-versa; computing principal values and directions of
stress and strain; and computing a perturbed mesh for a collapse analysis by incorporating a
user-specified geometric imperfection in the form of the critical buckling mode shape.
Each postprocessing program must be linked using the make parameter when running the ABAQUS
execution procedure (see `Èxecution procedure for ABAQUS/Make,'' Section 3.2.10 of the
ABAQUS/Standard User's Manual). To link properly, the postprocessing program cannot contain a
FORTRAN PROGRAM statement. Instead, the program must begin with a FORTRAN SUBROUTINE
with the name HKSMAIN.
General programming concepts, ABAQUS FORTRAN interfaces, and data processing concepts are
described below. Refer to Chapter 5, "File Output Format," of the ABAQUS/Standard User's Manual
for additional information. The program listings in each section provide details on the program flow,
how to interface with various computer platforms that use different operating systems and FORTRAN
compilers, and how to interface with ABAQUS subroutines to handle data files and records.
Initialization
Details about the variables that are used in the postprocessing programs are discussed in `Àccessing
the results file information,'' Section 5.1.3 of the ABAQUS/Standard User's Manual. ABAQUS uses a
512-word buffer named ARRAY for the reading and writing of data on the results file. This is
dimensioned as ARRAY(513). The integer equivalent is JRRAY(513) for a 64-bit computer or
JRRAY(2,513) for a 32-bit computer. The EQUIVALENCE statement is used to equivalence ARRAY
and JRRAY to simplify manipulation of real and integer numbers in the data record stored in the
buffer.
The information concerning the FORTRAN unit number and format of the results file that is read is
defined in LRUNIT(2,NRU), where NRU is the number of files to be processed. The FORTRAN unit
number for the nth file is stored in LRUNIT(1,n). The information about the file format is stored in
LRUNIT(2,n), which is initialized to 1 for ASCII format and to 2 for binary format. If a new results
11-1180
file is to be created by the postprocessing program, the file format of the output file is defined similarly
via the variable LOUTF, which is also initialized to 1 for ASCII format and 2 for binary format. The
root file name for both input and output results files is defined through the character variable FNAME.
The root file name case will be the same as the case in which FNAME is defined; ABAQUS defines the
file extensions to be lowercase letters. See `Àccessing the results file information,'' Section 5.1.3 of
the ABAQUS/Standard User's Manual, for a discussion of the naming convention for the file
extensions.
The final initialization phase is done internally by calling the ABAQUS subroutines INITPF and
DBRNU. The FORTRAN interfaces are
CALL INITPF(FNAME, NRU, LRUNIT, LOUTF)
CALL DBRNU (JUNIT)
where the arguments in the call to INITPF are as described above, and JUNIT is the FORTRAN unit
number connecting the file.
These integer variables must be defined before the subroutines are called.
Data processing
Data manipulation requires knowledge of each data record. Details of these records are found in
``Results file output format,'' Section 5.1.2 of the ABAQUS/Standard User's Manual.
The data organization in the results file uses a sequential format. Each record must, therefore, be
retrieved in a sequential manner via a call to DBFILE using the interface
CALL DBFILE(0, ARRAY, JRCD)
This call can be placed inside a DO-loop, and the loop count should exceed the number of records
stored in the file. Alternatively, DBFILE can be called as long as JRCD is equal to 0. The first
argument, 0, indicates that a record is to be read. Each record that is read is stored in the buffer ARRAY
and returned to the calling program for manipulation. The last argument, JRCD, is a return code that is
set to 0 unless an end-of-file condition or an incomplete record is processed, in which case JRCD is set
to 1.
If it is desirable to extract or modify certain records and save them in a new results file with the same
data organization as an ABAQUS-generated results file, then the subroutine DBFILW should be called
with the interface
CALL DBFILW(1, ARRAY, JRCD)
The new results file will be written with the file extension .fin. Refer to `Ùtility routines for
accessing the results file,'' Section 5.1.4 of the ABAQUS/Standard User's Manual.
11.1.2 Joining data from multiple results files and converting file
format: FJOIN
This example illustrates how to use a FORTRAN program to extract specific data from different
11-1181
ABAQUS results files and to join the data into a single results file. This program can also be used to
convert the format of results files.
Postprocessing
Sometimes it is desirable to combine a number of results files into a single file or to create a new
results file by retrieving selected data from different results files. The ABAQUS/Append procedure
joins two results files by stripping the header information from the second results file and appending
the step information to the end of the first results file. See `Èxecution procedure for joining results
(.fil) files,'' Section 3.2.7 of the ABAQUS/Standard User's Manual, for more information on this
utility.
This example postprocessing program demonstrates how a FORTRAN program can be used to extract
specific information from results files created by separate analyses of the same model. In this example
the stress and strain records in three analyses will be merged to create a new results file.
Programming details
The general discussion on programming concepts and ABAQUS FORTRAN interfaces in `Ùser
postprocessing of ABAQUS results files: overview,'' Section 11.1.1, should be reviewed before
running or modifying this program. Review of the results file format in Chapter 5, "File Output
Format," of the ABAQUS/Standard User's Manual is also recommended.
The program FJOIN (named fjoin.f on the ABAQUS release media), prompts for the values of
NRU, LRUNIT(1,NRU), LRUNIT(2,NRU), and FNAME. Then subroutines INITPF and DBNRU are
called to complete the necessary initializations and file connections. Data processing starts with a
double DO-loop looping over all of the records to be read, one-by-one, via a call to DBFILE. A record
can be skipped or written to the new results file with or without any modifications. Each record is
identified by its record key, which is stored in the second entry of the record (see ``Results file output
format,'' Section 5.1.2 of the ABAQUS/Standard User's Manual).
Each file contains a number of header records, ( 1900-series). These records contain general
information about the model. Different analyses using the same model place essentially the same
information in these records. Hence, when combining results files from different analyses of the same
model, all 1900-series records from the first file that is processed should be kept. Similar records in
subsequent files should be skipped to avoid duplication and confusion. However, the 1910, 1911,
1922, and 1980 records should be kept. They are useful for processing superelement results, output
requests, and natural frequency extraction results.
The data for each increment of an analysis begins with the increment start record, which is identified
by record key 2000. Record 2000 is followed by the records that correspond to the data requested
through file output options specified in the ABAQUS input file. Record 1, the element header record,
is automatically written to the results file when the *EL FILE option is used in the input file. It is of
interest when postprocessing since it contains important information about the element data, including
the location of data within an element (i.e., whether data are written at the element integration points,
the centroid, nodes, etc.). For this example, records 11 and 21, the stress and strain records,
respectively, are written to the results file since stress and strain were requested through the *EL FILE
11-1182
option. The increment end record is identified by record key 2001. When an end-of-file condition is
encountered and the previously processed record is a 2001 record, a FORTRAN CLOSE is executed
on the current FORTRAN unit number so that the processing of the next file can begin.
Program compilation and linking

The ABAQUS/Make procedure is designed to compile and link this type of postprocessing program. It
will also make the aba_param.inc file available during compilation. The ABAQUS/Make
command to compile and link the FJOIN program is as follows:
abaqus make job=fjoin
This command will have to be repeated if FORTRAN errors are discovered during the compilation or
link. The commands used by the ABAQUS/Make procedure can be changed if necessary. The
ABAQUS Site Guide lists the typical compile and link commands for each computer type.
Program execution
Before program execution, the analysis jobs must be run to generate results files to be read by the
program. In this example three jobs are run. The input files for these analyses are fjoin002.inp,
fjoin003.inp, and fjoin004.inp. The results files from these analyses are output in binary
format and are called fjoin002.fil, fjoin003.fil, and fjoin004.fil.
The FJOIN program will read these files via FORTRAN units 2, 3, and 4. The name of the new file
will be fjoinxxx. Before running the program, the results files must be renamed to
fjoinxxx.002, fjoinxxx.003, and fjoinxxx.004. Note that the root file names are the same
(defined using FNAME), and that the extensions are set to the FORTRAN unit numbers used to open
the files.
When the program is executed using the command abaqus fjoin, the first prompt will be
Enter the number of files to be joined:
Enter 3 to set NRU=3. The second prompt will be
Enter the unit number of input file # 1:
Enter 2 to define LRUNIT(1,1)=2. At the third prompt,
Enter the format of input file # 1 (1-ASCII, 2-binary):
enter 2. This sets LRUNIT(2,1)=2 and means that the file being read is binary. The second and third
prompts are repeated for each additional file to be processed. The program will then ask whether the
new results file should be written in ASCII or binary format,
Enter the format of the output file (1-ASCII, 2-binary):
Enter 2 to set LOUTF=2, which specifies that binary format has been chosen for the new results file.
The format of the output file may be different from the format of the input files, so this program can
also be used to convert the format of results files. Finally, when the program issues the prompt
Enter the name of the input files (w/o extension):
enter fjoinxxx to define FNAME (the input files must have been given the root file name
11-1183
fjoinxxx; the output file will be created as fjoinxxx.fin).

As soon as the nth file has been processed, the message END OF FILE # n is written to the terminal.
After all files have been processed, the program stops and the new results file is created. The new
results file created by this program contains stress and strain records at all integration points in each
element and at all nodal points.
Analysis description
The structure is a 10 ´ 10 square plate with unit thickness. The plate lies in the X-Y plane such that its
bottom edge coincides with the x-axis and the left edge coincides with the y-axis. The finite element
model employs a 2 ´ 2 mesh of CPS8R elements. The material is linear elastic with Young's modulus
= 30 ´ 106 and Poisson's ratio = 0.3. Three separate analyses are performed with displacement
controlled load steps.
In the first analysis (fjoin002.inp), the plate is subjected to biaxial tension by prescribing a vertical
displacement of 0.25 along the top edge, a horizontal displacement of 0.25 along the right edge and
symmetry boundary conditions on the left and bottom edges.
In the second analysis (fjoin003.inp), the structure is forced to deform in simple shear by applying a
horizontal displacement of 0.25 to the top edge while holding the bottom edge fixed and allowing the
horizontal displacement to vary linearly with y along the left and right edges. The vertical
displacement is zero everywhere.
In the third analysis (fjoin004.inp), the plate is subjected to uniaxial tension by applying a
displacement of 0.25 in the y-direction to the nodes along the top edge and symmetry boundary
conditions to the nodes along the x- and y-axes.

Since the state of stress and strain is homogeneous, the integration point and nodal averaged values of
stress and strain are identical everywhere. A typical record obtained at the end of each step is included
below:
Analysis ¾xx ¾yy ¾xy
fjoin002.inp 1.07 ´ 1.07 ´ 0.0
106 106
fjoin003.inp 0.0 0.0 2.88 ´
105
fjoin004.inp 0.0 7.50 ´ 0.0
10 5
Analysis "xx "yy "xy

fjoin002.inp 2.50 ´ 10 -2 2.50 ´ 10 -2 0.0
fjoin003.inp 0.0 0.0 2.50 ´ 10-2
fjoin004.inp -7.50 ´ 10-3 2.50 ´ 10-2 0.0
Input files
fjoin002.inp
11-1184
First analysis file.

fjoin003.inp
Second analysis file.
fjoin004.inp
Third analysis file.
fjoin.f
Postprocessing program.
Sample listings
11-1185
Listing 11.1.2-1
SUBROUTINE HKSMAIN
C====================================================================
C This program must be compiled and linked with the command:
C abaqus make job=fjoin
C Run the program using the command:
C abaqus fjoin
C====================================================================
C
C Purpose (this program performs two functions):
C
C 1. It can be used to join together a number of ABAQUS results files.
C The program will prompt the user for the number of files to be
C joined, the FORTRAN unit numbers associated with each file and
C the file format, ASCII or binary. The user will also be prompted
C for the format of the output file and the root name of the files.
C
C 2. It can be used to convert the format of a file from binary to
C ASCII or vice-versa. This can be accomplished by reading one
C file as input and giving the opposite format for the output file.
C
C Input File names:
C
C The results file to be processed should be named 'FNAME.0xx',
C where xx is a 2-digit FORTRAN unit number less than 31.
C Certain units within this range are used by ABAQUS internally and
C by this program and cannot be used by the user. These are 01,
C 05, 06, 07, 09, 11, 12, 13, 20 and 28.
C
C Output File name:
C
C 'FNAME'.fin
C
C====================================================================
C
C Variables used by this program:
C
C ARRAY -- Real array containing values read from results file
C (.fil). Equivalenced to JRRAY.
C JRRAY -- Integer array containing values read from results file
C (.fil). Equivalenced to ARRAY.
C FNAME -- Root file name of input file(s) and output file.
11-1186
C NRU -- Number of results files (.fil) to be read.

C LRUNIT -- Array containing unit number and format of results files:
C LRUNIT(1,*) --> Unit number of input file.
C LRUNIT(2,*) --> Format of input file.
C LOUTF -- Format of output file:
C 1 --> ABAQUS results file ASCII format.
C 2 --> ABAQUS results file binary format.
C JUNIT -- Unit number of file to be opened.
C JRCD -- Error check return code:
C .EQ. 0 --> No errors.
C .NE. 0 --> Errors detected.
C KEY -- Current record key identifier.
C
C====================================================================
C
C The use of ABA_PARAM.INC eliminates the need to have different
C versions of the code for single and double precision.
C ABA_PARAM.INC defines an appropriate IMPLICIT REAL statement and
C sets the value of NPRECD to 1 or 2, depending on whether the
C machine uses single or double precision.
C
C====================================================================
C
DIMENSION ARRAY(513), JRRAY(NPRECD,513)
C
C====================================================================
C Set the dimensions of LRUNIT to be the maximum number of results
C files to be joined.
C
C====================================================================
PARAMETER (MXUNIT=21)
INTEGER LRUNIT(2,MXUNIT),LUNIT(10)
CHARACTER FNAME*80
DATA LUNIT/1,5,6,7,9,11,12,13,20,28/
C
C====================================================================
C Input the number of files to be joined and then the unit number and
C format of each of the files.
C
C====================================================================
5 CONTINUE
11-1187
WRITE(6,10) MXUNIT
10 FORMAT(1X,'Enter the number of files to be joined (MAX:',I3,'):')
READ(5,'(I3)') NRU
IF (NRU .GT. MXUNIT) GOTO 5
C
C
DO 40 INRU = 1, NRU
15 CONTINUE
WRITE(6,20) INRU
20 FORMAT(1X,'Enter the unit number of input file #',I3,':')
READ(5,*) LRUNIT(1,INRU)
DO 41 K1=1,9
IF (LRUNIT(1,INRU) .EQ. LUNIT(K1)) THEN
WRITE(6,*) 'ERROR! Unit number cannot be ',LUNIT(K1)
GOTO 15
ENDIF
41 CONTINUE
42 CONTINUE
WRITE(6,30) INRU
30 FORMAT(1X,'Enter the format of input file #',I3,
1 ' (1-ASCII, 2-binary):')
READ(5,*) LRUNIT(2,INRU)
IF (LRUNIT(2,INRU).NE. 1 .AND. LRUNIT(2,INRU) .NE. 2) THEN
WRITE(6,*) 'ERROR! This number must be 1 or 2'
GOTO 42
ENDIF
40 CONTINUE
C
C====================================================================
C Set LOUTF equal to the format of the output file. If this program
C is to be used only to convert the file format from one type to
C another, set NRU=1 (to read only one file) and specify a value of
C LOUTF which is opposite to the value specified for LRUNIT(2,1).
C
C====================================================================
45 CONTINUE
WRITE(6,50)
50 FORMAT(1X,'Enter the format of the output file ',
1 '(1-ASCII, 2-binary):')
READ(5,*) LOUTF
IF (LOUTF .NE. 1 .AND. LOUTF .NE. 2) THEN
WRITE(6,*) 'ERROR! This number must be 1 or 2'
GOTO 45
11-1188
ENDIF
C
WRITE(6,60)
60 FORMAT(1X,'Enter the name of the input file(s) (w/o extension):')
READ(5,'(A)') FNAME
C
C
KEYPRV = 0
C
C====================================================================
C Loop through NRU input files...
C
C====================================================================
DO 100 INRU = 1, NRU
JUNIT = LRUNIT(1,INRU)
CALL DBRNU (JUNIT)
I2001 = 0
C====================================================================
C ...and cover a maximum of 10 million records in each file.
C
C====================================================================
DO 80 IXX2 = 1, 100
DO 80 IXX = 1, 99999
C WRITE(6,*) 'KEY/RECORD LENGTH = ', JRRAY(1,2),JRRAY(1,1)
IF (JRCD .NE. 0 .AND. KEYPRV .EQ. 2001) THEN
WRITE(6,*) 'END OF FILE #', INRU
CLOSE (JUNIT)
GOTO 100
ELSE IF (JRCD .NE. 0) THEN
WRITE(6,*) 'ERROR READING FILE #', INRU
CLOSE (JUNIT)
GOTO 110
ENDIF
C
C====================================================================
C Initialize the flag to write a record to the file:
C LWRITE=0 -- write disabled
C LWRITE=1 -- write enabled
C
C====================================================================
LWRITE=1
11-1189
C
C====================================================================
C For files other than the first, skip the 1900-series header records
C except for the superelement path (1910; for superelement analyses),
C output request (1911), heading (1922), and modal (1980; for natural
C frequency extraction) records. In a merged file, the heading
C record serves as a file delimiter.
C
C====================================================================
IF (INRU.GT.1) THEN
IF (JRRAY(1,2).GE.1900 .AND. JRRAY(1,2).LE.1909) LWRITE=0
IF (JRRAY(1,2).GE.1912 .AND. JRRAY(1,2).LT.1922) LWRITE=0
C
C====================================================================
C Skip the first 2001 record (this indicates the end of the header
C records).
C
C====================================================================
IF (JRRAY(1,2) .EQ. 2001 .AND. I2001 .EQ. 0) THEN
I2001 = 1
LWRITE = 0
ENDIF
ENDIF
C
C====================================================================
C If this is the first input file, or if the write flag has not been
C disabled for records in subsequent files, then write the data to
C the output file. We are interested in retrieving the header
C records (relevant 1900-series records), the increment start and
C end records (2000 and 2001), the element header record, (1) and
C the stress and strain records (11 and 21).
C
C====================================================================
IF (INRU .EQ. 1 .OR. LWRITE .EQ. 1) THEN
KEY=JRRAY(1,2)
IF((KEY.EQ.1900).OR.(KEY.EQ.1901).OR.(KEY.EQ.1902).OR.
1 (KEY.EQ.1910).OR.(KEY.EQ.1911).OR.(KEY.EQ.1921).OR.
4 (KEY.EQ.21)) THEN
CALL DBFILW(1,ARRAY,JRCD)
IF (JRCD .NE. 0) THEN
WRITE(6,*) 'ERROR WRITING FILE'
11-1190
CLOSE (JUNIT)
GOTO 110
ENDIF
ENDIF
ENDIF
KEYPRV = JRRAY(1,2)
80 CONTINUE
100 CONTINUE
110 CONTINUE
C
RETURN
END
11.1.3 Calculation of principal stresses and strains and their

directions: FPRIN
This example illustrates the use of a FORTRAN program to read stress and strain records from an
ABAQUS results file and to calculate principal stress and strain values and their directions.
General description
This program shows how to retrieve integration point and nodal averaged stress and strain components
from an ABAQUS results file and then compute principal values and directions using the ABAQUS
subroutine SPRIND. Usage of this subroutine is documented in the program listing provided below,
and further details about the interface to this subroutine are discussed in `ÙMAT,'' Section 23.2.29 of
the ABAQUS/Standard User's Manual. The results file created by the FJOIN program in ``Joining data
from multiple results files and converting file format: FJOIN,'' Section 11.1.2, is used here to verify
that the records that have been put together are retrievable. The previously generated results file was
named fjoinxxx.fin. To use it as an input file for postprocessing program FPRIN, the file
extension must be changed. This program will assume that the results file has the default .fil
extension, which corresponds to FORTRAN unit 8.
Programming details
The user should first review the general discussion on programming concepts and ABAQUS
FORTRAN interfaces in `Ùser postprocessing of ABAQUS results files: overview,'' Section 11.1.1,
and the detailed discussion of postprocessing given in Chapter 5, "File Output Format," of the
When running program FPRIN (this program is named fprin.f on the ABAQUS release media), the
user will be prompted for the file name that initializes FNAME. Other variables, such as LOUTF, NRU,
LRUNIT(1,NRU), and LRUNIT(2,NRU), are initialized inside the program. INITPF and DBNRU
are then called to complete the neccesary initializations and file connections. Data processing starts
with a double DO-loop over all the records to be read, one-by-one, via a call to DBFILE. Each record
11-1191
is identified by its record key, which is stored in the second entry of the record. When records 1922
and 2000 are processed by program FPRIN, the heading and the current step and increment numbers
are written out so as to provide a way to recognize the beginning of data in each analysis. Record type
1 is then examined to determine the output location of stress and strain, the number of direct and shear
stress and strain components, and either the element number or the node number for which the records
are written. The stress and strain records ( 11 and 21, respectively) will be filtered out for processing
by the ABAQUS subroutine SPRIND. When a stress or strain record is passed into SPRIND, principal
stresses or strains and the corresponding principal directions are calculated and returned in an unsorted
order.

Before program execution, the FORTRAN program has to be compiled and linked. Both operations, as
well as the inclusion of the aba_param.inc file, are performed by a single execution of the
ABAQUS/Make procedure:
abaqus make job=fprin
This may have to be repeated until all FORTRAN errors are corrected. After successful compilation,
the program's object code is automatically linked with the ABAQUS object codes stored in the shared
program library and interface library in order to build the executable program. Refer to Chapter 3,
"Environment file," of the ABAQUS Site Guide to see which compile and link commands are used for
a particular computer.
Program execution
Before the program is executed, a results file must have been created. In this example the results file
fjoinxxx.fin created by the FJOIN program discussed in ``Joining data from multiple results files
and converting file format: FJOIN,'' Section 11.1.2, is used. This file must be renamed to
fjoinxxx.fil since FORTRAN unit 8 (which is associated with the .fil file extension) is used
in the program to read the file. When the program is executed using the command abaqus fprin,
the prompt
Enter the name of the input file (w/o .fil):
will appear. Enter fjoinxxx to define FNAME. The program processes the data and produces a file
named pvalue.dat, which contains information about principal stresses and strains and their
directions.

The computed principal stress and strain values and their directions are tabulated below.
Analysis Principal Stress Strain Dir-1 Dir-2 Dir-3
Componen
File ´ 105 ´ 10-3
t
fjoin002.i 1 10.714 25.0 1.0 0.0 0.0
np 2 10.714 25.0 0.0 1.0 0.0
3 0.0 0.0 0.0 0.0 1.0
fjoin003.i 1 -2.8846 -12.5 0.707 -0.707 0.0
11-1192
np 2 2.8846 12.5 0.707 0.707 0.0

3 0.0 0.0 0.0 0.0 1.0
fjoin004.i 1 0.0 -7.5 1.0 0.0 0.0
np 2 7.5 25.0 0.0 1.0 0.0
3 0.0 0.0 0.0 0.0 1.0
Input file
fprin.f
Sample listings
11-1193
Listing 11.1.3-1
SUBROUTINE HKSMAIN
C====================================================================
C abaqus make job=fprin
C abaqus fprin
C====================================================================
C
C Purpose:
C
C This program computes the principal stresses and strains and their
C directions from stress and strain values stored in an ABAQUS
C results file (.fil).
C
C Input File names: `FNAME.fil', where FNAME is the root file name of
C the input file.
C
C Output File name: pvalue.dat
C
C====================================================================
C
C Variables used by this program and ABAQUS subroutine SPRIND :
C
C NDI -- Number of direct components in stress/strain tensor.
C NSHR -- Number of shear components in stress/strain tensor.
C NDIP1 -- NDI + 1
C ARRAY -- Real array containing values read from results file
C (.fil). Equivalenced to JRRAY.
C JRRAY -- Integer array containing values read from results file
C (.fil). Equivalenced to ARRAY.
C FNAME -- Root file name of input file (w/o .fil extension).
C NRU -- Number of results files (.fil) to be read.
C LRUNIT -- Array containing unit number and format of results files:
C LRUNIT(1,*) --> Unit number of input file.
C LRUNIT(2,*) --> Format of input file.
C LOUTF -- Format of output file:
C 0 --> Standard ASCII format.
C JUNIT -- Unit number of file to be opened.
C JRCD -- Error check return code.
11-1194

C KEY -- Current record key identifier.
C JELNUM -- Current element number.
C INTPN -- Integration point number.
C LSTR -- Indicates type of principal value (stress/strain) and
C ordering used:
C For calculation of principal value (stress/strain):
C 1 --> stress.
C 2 --> strain.
C For calculation of directions:
C 1 --> stress.
C 2 --> strain.
C S -- Array containing stress tensor.
C PS -- Array containing principal stresses.
C ANPS -- Array containing directions of principal stresses.
C E -- Array containing strain tensor.
C PE -- Array containing principal strains.
C ANPE -- Array containing directions of principal strains.
C
C====================================================================
C
C ABA_PARAM.INC defines an appropriate IMPLICIT REAL statement
C and sets the value of NPRECD to 1 or 2, depending on whether
C the machine uses single or double precision.
C
C====================================================================
C
DIMENSION ARRAY(513), JRRAY(NPRECD,513), LRUNIT(2,1)
C
C====================================================================
DIMENSION S(6), E(6), PS(3), PE(3), ANPS(3,3), ANPE(3,3)
CHARACTER FNAME*80
C
C====================================================================
C Get the name of the results file.
C
C====================================================================
WRITE(6,*) 'Enter the name of the input file (w/o .fil):'
11-1195
READ(5,'(A)') FNAME
C
C====================================================================
C Open the output file.
C
C====================================================================
OPEN(UNIT=9,FILE='pvalue.dat',STATUS='NEW')
C
NRU = 1
LOUTF = 0
LRUNIT(1,1) = 8
LRUNIT(2,1) = 2
C
CALL INITPF(FNAME,NRU,LRUNIT,LOUTF)
C
JUNIT = 8
C
CALL DBRNU(JUNIT)
C
C====================================================================
C Read records from the results (.fil) file and process the data.
C Cover a maximum of 10 million records in the file.
C
C====================================================================
DO 1000 K100 = 1, 100
DO 1000 K1 = 1, 99999
IF (JRCD .NE. 0) GO TO 1001
KEY = JRRAY(1,2)
C
C====================================================================
C Get the heading (title) record.
C
C====================================================================
IF (KEY .EQ. 1922) THEN
WRITE(9,1100) (ARRAY(IXX),IXX=3,12)
1100 FORMAT(1X,10A8)
C
C====================================================================
C Get the current step and increment number.
C
C====================================================================
ELSE IF (KEY .EQ. 2000) THEN
11-1196
WRITE(9,1200) JRRAY(1,8), JRRAY(1,9)

1200 FORMAT(1X,'** STEP ',I2,' INCREMENT ',I3)
C
C====================================================================
C Get the element and integration point numbers, JELNUM and INTPN,
C and the location of INTPN (0--at int.pt., 1--at centroid,
C 4--nodal average) and the number of direct and shear components
C in the analysis.
C
C====================================================================
JELNUM = JRRAY(1,3)
INTPN = JRRAY(1,4)
LOCATE = JRRAY(1,6)
NDI = JRRAY(1,8)
NSHR = JRRAY(1,9)
NDIP1 = NDI + 1
IF(LOCATE.LE.1) THEN
WRITE(9,1201) JELNUM, INTPN ,NDI,NSHR
1201 FORMAT(2X,'ELEMENT NUMBER = ',I8,5X,
1 'INT. PT. NUMBER = ',I2,5X,
2 'NDI/HSHR = ',2I2)
ELSEIF(LOCATE.EQ.4) THEN
WRITE(9,1191) JELNUM, NDI,NSHR
1191 FORMAT(2X,'NODE NUMBER = ',I8,5X,
1 'NDI/HSHR = ',2I2)
END IF
C
C====================================================================
C Get the stress tensor.
C
C====================================================================
WRITE(9,1202)
1202 FORMAT(3X,'STRESSES:')
C
DO 10 IXX = 1, NDI
S(IXX) = ARRAY(IXX+2)
10 CONTINUE
WRITE(9,1203) (S(IZZ), IZZ = 1, NDI)
1203 FORMAT(4X,'S11 = ',E12.5,' S22 = ',E12.5,' S33 = ',E12.5)
DO 20 IYY = NDI + 1, NSHR + NDI
S(IYY) = ARRAY(IYY+2)
11-1197
20 CONTINUE
WRITE(9,1204) (S(IZZ), IZZ = NDI + 1, NSHR + NDI)
1204 FORMAT(4X,'S12 = ',E12.5,' S13 = ',E12.5,' S23 = ',E12.5)
C
C
C====================================================================
C Calculate the principal stresses and corresponding principal
C directions in unsorted order.
C====================================================================
LSTR = 1
CALL SPRIND(S,PS,ANPS,LSTR,NDI,NSHR)
WRITE(9,1205) PS(1), ANPS(1,1), ANPS(1,2), ANPS(1,3)
1205 FORMAT(4X,'PS1 = ',E12.5,/,
1 5X,'PD11 =',F8.3,2X,'PD12 =',F8.3,2X,'PD13 =',F8.3)
1206 FORMAT(4X,'PS2 = ',E12.5,/,
1 5X,'PD21 =',F8.3,2X,'PD22 =',F8.3,2X,'PD23 =',F8.3)
1207 FORMAT(4X,'PS3 = ',E12.5,/,
1 5X,'PD31 =',F8.3,2X,'PD32 =',F8.3,2X,'PD33 =',F8.3)
C
C
C====================================================================
C Get the strain tensor.
C
C====================================================================
WRITE(9,2202)
2202 FORMAT(3X,'STRAINS:')
C
DO 30 IXX = 1, NDI
E(IXX) = ARRAY(IXX+2)
30 CONTINUE
WRITE(9,2203) (E(IZZ), IZZ = 1, NDI)
2203 FORMAT(4X,'E11 = ',E12.5,' E22 = ',E12.5,' E33 = ',E12.5)
DO 40 IYY = NDI + 1, NSHR + NDI
E(IYY) = ARRAY(IYY+2)
40 CONTINUE
WRITE(9,2204) (E(IZZ), IZZ = NDI + 1, NSHR + NDI)
2204 FORMAT(4X,'E12 = ',E12.5,' E13 = ',E12.5,' E23 = ',E12.5)
C
C
C====================================================================
11-1198
C Calculate the principal strains and corresponding principal

C directions in unsorted order.
C====================================================================
LSTR = 2
CALL SPRIND(E,PE,ANPE,LSTR,NDI,NSHR)
WRITE(9,2205) PE(1), ANPE(1,1), ANPE(1,2), ANPE(1,3)
2205 FORMAT(4X,'PE1 = ',E12.5,/,
1 5X,'PD11 =',F8.3,2X,'PD12 =',F8.3,2X,'PD13 =',F8.3)
2206 FORMAT(4X,'PE2 = ',E12.5,/,
1 5X,'PD21 =',F8.3,2X,'PD22 =',F8.3,2X,'PD23 =',F8.3)
2207 FORMAT(4X,'PE3 = ',E12.5,/,
1 5X,'PD31 =',F8.3,2X,'PD32 =',F8.3,2X,'PD33 =',F8.3)
C
END IF
C
1000 CONTINUE
1001 CONTINUE
C
CLOSE (UNIT=9)
C
RETURN
END
11.1.4 Creation of a perturbed mesh from original coordinate

data and eigenvectors: FPERT
This example illustrates the use of a FORTRAN program to create a perturbed mesh by superimposing
a small imperfection in the form of the weighted sum of several buckling modes on the initial
geometry. The program retrieves the original nodal coordinates and the desired eigenvectors from an
ABAQUS results file, then calculates new nodal coordinates for the perturbed mesh.
General description
Collapse studies of a structure's postbuckling load-displacement (Riks) behavior are often conducted to
verify that the critical buckling load and mode predicted by an eigenvalue buckling analysis are
accurate. They are also done to investigate the effect of an initial geometric imperfection on the
load-displacement response. A typical assumption is that an imperfection made up of a combination of
the eigenmodes associated with the lowest eigenvalues will be the most critical. One method of
PM
introducing an imperfection of this type into the model is by adding i=1 ®i ui to the original mesh
coordinates. In this case ui is the ith eigenmode, ®i is a scaling factor of the ith eigenmode, and M is
11-1199
the total number of eigenmodes extracted in the buckling analysis. Since the eigenvector is typically
normalized to a maximum absolute value of one, ®i is usually some fraction of a geometric parameter,
such as the shell thickness. The postprocessing program described below can be used to introduce an
imperfection of this type into a model.
The perturbation procedure is illustrated in ``Buckling of a cylindrical shell under uniform axial
pressure,'' Section 1.2.3 of the ABAQUS Benchmarks Manual. An eigenvalue buckling analysis,
fpert001, is run first. This analysis creates the results file, fpert001.fil, which contains the
original nodal coordinates and the eigenvectors for the buckling modes. This results file is then used to
generate a perturbed mesh for the postbuckling load-displacement analysis. The postprocessing
program perturbs the original mesh using the relation
M
X
0
X =X+ ®i ui ;
i=1
where X0 is the vector containing the new global coordinates; X is the vector of original coordinates;
M is the number of buckling modes; and ®i is the imperfection factor for the ith eigenvector, ui . The
new coordinates are written to the file fpert002.015, which is read by the load-displacement
analysis fpert002.
Programming details
The general discussion on programming concepts and ABAQUS FORTRAN interfaces in `Ùser
postprocessing of ABAQUS results files: overview,'' Section 11.1.1, should be reviewed before
running or modifying this program. Review of the results file format in Chapter 5, "File Output
Format," of the ABAQUS/Standard User's Manual is also recommended.
The FPERT program (this program is named fpert.f on the ABAQUS release media) makes some
assumptions concerning the type of results file it will be reading. Variables NRU, LRUNIT(1,NRU),
and LRUNIT(2,NRU) are initialized within the program to 1, 8, and 2. These values indicate that
one file will be read, the FORTRAN unit used will be 8, and the file type will be binary. See
`Àccessing the results file information,'' Section 5.1.3 of the ABAQUS/Standard User's Manual, for
more information on opening and initializing postprocessing files.
Once the file specification parameters are set, the INITPF and DBNRU subroutines are called to open
and ready the file, whose name is stored in FNAME, for reading. The file to which the perturbed
coordinates are to be written can be directly opened using a FORTRAN OPEN statement. The
ABAQUS file utilities are not necessary since the file is a plain text file.
The records with the original nodal coordinates are read using the DBFILE routine and stored in the
local array COORDS(3,8000). The first index of the COORDS array indicates the x-, y-, and
z-coordinate of the node. The second index indicates the node number. The second dimension should
be increased if there are more than 8000 nodes in a model.
Components of the eigenvector are stored in the local array DISP(6,8000). This array holds up to 6
displacement terms for each node. The second dimension should be increased if there are more than
8000 nodes in a model. Subroutine NODEGEN, a subroutine local to this postprocessing program, is
11-1200
then called to compute the new nodal coordinates. Once all the requested mode shapes are computed,
the new nodal coordinates are written to the plain text file opened earlier.

The ABAQUS/Make procedure is designed to compile and link this type of postprocessing program. It
will also make the aba_param.inc file available during compilation. The ABAQUS/Make
command to compile and link the FPERT program is as follows:
abaqus make job=fpert
This command will have to be repeated if FORTRAN errors are discovered during the compilation or
link. The commands used by the ABAQUS/Make procedure can be changed if necessary. The
ABAQUS Site Guide lists the typical compile and link commands for each computer type.
Program execution
Before the program is executed, an eigenvalue buckling job must have been run with ABAQUS. In this
example the input file fpert001.inp is used to generate the results file fpert001.fil. When
the FPERT program is executed using the command abaqus fpert, the first prompt will be
Enter the name of the results file (w/o .fil):
Enter fpert001 to define FNAME. The second prompt will be
Enter the mode shape(s) to be used in calculating the perturbed
mesh (zero when finished):
Enter 1 followed by 0, since this is the only eigenvector available in the results file for this example.
At the third prompt,
Enter the imperfection factor to be introduced into the geometry
for this eigenmode:
enter 0.25. This sets ® = 0.25, the shell thickness for this model. The program then processes the data
and writes the nodal coordinates for the new mesh to fpert002.015.
Analysis description
For a full discussion of the analysis, refer to ``Buckling of a cylindrical shell under uniform axial
pressure,'' Section 1.2.3 of the ABAQUS Benchmarks Manual. The input file fpert001.inp (same
file as bucklecylshell_s9r5_n3.inp) contains a 2 ´ 20 mesh of S9R5 elements and data lines
for a buckling analysis. The input file fpert002.inp contains data lines for a Riks analysis using a
perturbed mesh. The source code for the FPERT program is in fpert.f.

Plots produced by these analyses are shown in Figure 11.1.4-1 and Figure 11.1.4-2. Figure 11.1.4-1 is
obtained from the eigenvalue buckling analysis and shows the original (cylindrical) mesh and the
critical buckling mode. Figure 11.1.4-2 is generated when the load level has reached a local maximum
(increment 8) in the Riks analysis using the perturbed mesh.
11-1201
Input files
fpert001.inp
Eigenvalue buckling analysis.
fpert002.inp
Riks analysis using a perturbed mesh.
fpert.f
Figures
Figure 11.1.4-1 Undeformed shape and eigenvalue buckling mode.
Figure 11.1.4-2 Deformed shape at first peak load in Riks analysis.
11-1202
Sample listings
11-1203
Listing 11.1.4-1
SUBROUTINE HKSMAIN
C====================================================================
C abaqus make job=fpert
C abaqus fpert
C====================================================================
C
C PURPOSE:
C This program computes a perturbed mesh based on a user-specified
C perturbation factor. The original coordinate data and
C eigenvectors are read from an ABAQUS results (.fil) file.
C
C PROMPTS:
C 1. Ènter the name of the results file (w/o .fil):'
C 2. Ènter the mode shape(s) to be used in calculating the
C perturbed mesh (zero when finished):'
C 3. Ènter the imperfection factor to be introduced into the
C geometry for this eigenmode:'
C
C====================================================================
C
C INPUT FILE -- `FNAME'.fil
C
C OUTPUT FILE -- fpert002.015
C
C====================================================================
C
C ABA_PARAM.INC defines an appropriate IMPLICIT REAL statement and
C sets the value of NPRECD to 1 or 2, depending on whether the
C machine uses single or double precision.
C
C====================================================================
C ARRAY = Described in Section 7.0.0 of the Verification manual
C JRRAY = Described in Section 7.0.0 of the Verification manual
C LRUNIT = Described in Section 7.0.0 of the Verification manual
C DISP = Contains the eigenvector for a particular eigenmode
C COORD = Original coordinate data
C INODE = Original node label
11-1204
C IDOF = DOF for the element

C JEIGNO = Array of mode shapes used for calculating the perturbed
C mesh
C FNAME = Name of the results file
C NODEMAX = Number of nodes in the model
C IELMAX = Number of elements in the model
C====================================================================
DIMENSION ARRAY(513), JRRAY(NPRECD,513), LRUNIT(2,1)
C===================================================================
C ITOTAL must be greater than or equal to the number of nodes in the
C model
C===================================================================
PARAMETER (ITOTAL = 8000)
C
C====================================================================
C
DIMENSION DISP(6,ITOTAL), COORD(3,ITOTAL)
DIMENSION INODE(ITOTAL), IDOF(30), JEIGNO(10)
CHARACTER FNAME*80,OUTFILE*(*)
PARAMETER (OUTFILE = 'fpert002.015')
C
C====================================================================
C Define flags and counters.
C
C====================================================================
ICYCLE = 0
I1901 = 0
I101 = 0
I = 1
K = 1
C
C====================================================================
C Define file access variables.
C
C====================================================================
NRU = 1
LRUNIT(1,NRU) = 8
LRUNIT(2,NRU) = 2
LOUTF = 0
C
C====================================================================
11-1205
C Open output file.

C
C====================================================================
OPEN(UNIT=15,FILE=OUTFILE,STATUS='UNKNOWN',IOSTAT = J)
IF (J .NE. 0) THEN
WRITE(*,900) OUTFILE
GOTO 950
ENDIF
C
C====================================================================
C Get the name of the results (.fil) file.
C
C====================================================================
WRITE(*,2000)
WRITE(6,*) ' Enter the name of the results file (w/o .fil):'
READ(5,'(A)', IOSTAT = J ) FNAME
IF (J .NE. 0) GOTO 950
C
C====================================================================
C Access ABAQUS libraries to set up input file.
C
C====================================================================
C
JUNIT = LRUNIT(1,NRU)
C
CALL DBRNU (JUNIT)
C
C====================================================================
C Read a record from the input file.
C On the first pass through the file obtain the number of nodes for
C a diagnostic check.
C====================================================================
CALL DBFILE (0, ARRAY, JRCD)
DO WHILE (JRCD .EQ. 0)
IF (JRRAY(1,2) .EQ. 1980) IEIGNO = JRRAY(1,3)
IF (JRRAY(1,2) .EQ. 1921 ) THEN
NODEMAX = JRRAY(1,8)
IELMAX = JRRAY(1,7)
ICYCLE = ICYCLE +1
ENDIF
ENDDO
11-1206
C
C===================================================================
C User is given a choice of eigenmodes for calculating the perturbed
C mesh.
C
C===================================================================
WRITE(*,2010) NODEMAX, IELMAX
WRITE(*,2015) IEIGNO
5 READ(5,*,ERR = 950) JEIGNO(I)
IF (JEIGNO(I) .EQ. 0) GOTO 10
I=I+1
GOTO 5
C
10 CONTINUE
C
C
C====================================================================
C If this is the first pass through the file and the current record
C is the nodal coordinate record, then read the original nodal
C coordinates and the node numbers. Make sure that the third
C coordinate exists before saving it.
C
C====================================================================
IF (JRRAY(1,2) .EQ. 1901 .AND. ICYCLE .LE. 1) THEN
I1901 = I1901 + 1
INODE(I1901) = JRRAY(1,3)
COORD(1,I1901) = ARRAY(4)
COORD(2,I1901) = ARRAY(5)
COORD(3,I1901) = 0.0D0
IF (JRRAY(1,1) .GE. 6) COORD(3,I1901) = ARRAY(6)
C
C====================================================================
C If this is the first pass through the file and the current record
C is the active degree of freedom record, save the active d.o.f.
C If the d.o.f. is active in the model, IDOF(XX) equals the
C position of d.o.f. XX in the output arrays. If the d.o.f. is not
C active, IDOF(XX) is zero for d.o.f. XX (i.e., for planar models
C IDOF(1) = 1, IDOF(2) = 2, IDOF(3) = 0, IDOF(4) = 0, IDOF(5) = 0,
C IDOF(6) = 3, etc.). ITRANS equals the number of translational
C d.o.f.'s in the model.
11-1207
C
C====================================================================
ELSE IF (JRRAY(1,2) .EQ. 1902) THEN
DO 15 IXX = 1, JRRAY(1,1)-2
IDOF(IXX) = JRRAY(1,IXX+2)
15 CONTINUE
ITRANS = 3
IF (IDOF(3) .EQ. 0) ITRANS = 2
C
C====================================================================
C If the current record is the modal record, save the current
C eigenvalue number.
C
C====================================================================
C
ELSE IF (JRRAY(1,2) .EQ. 1980) THEN
IEIGNO = JRRAY(1,3)
DO J = 1, I-1
IF (JEIGNO(J) .EQ. IEIGNO) K = J
ENDDO
C
C====================================================================
C If the current record is the displacement record and the current
C eigenvalue was requested, read the displacement data. The data
C will be in the coordinate system specified in the
C `*NODE FILE,GLOBAL=' option. If nodal transformations were
C performed and GLOBAL=NO was used, the displacements will be in
C the local system. If nodal transformations were used and
C GLOBAL=YES, the results will be in the global system. In all
C other cases the results will be in the global system. Also,
C make sure that degrees of freedom are active in the model before
C saving them in the appropriate array location.
C
C====================================================================
C
ELSE IF (JRRAY(1,2) .EQ. 101 .AND. IEIGNO .EQ. JEIGNO(K)) THEN
I101 = I101 + 1
DISP(1,I101) = ARRAY(4)
DISP(2,I101) = ARRAY(5)
IF (IDOF(3) .NE. 0) DISP(3,I101) = ARRAY(IDOF(3)+3)
11-1208
IF (INODE(I101) .EQ. 0) INODE(I101) = JRRAY(1,3)

IF( I101 .EQ. NODEMAX ) THEN
WRITE(6,16) JEIGNO(K)
16 FORMAT(/,2X,'Nodal coordinate data being computed for',
1 ' eigenvalue . . .',I5)
C
C====================================================================
C FACTOR should be entered as a perturbation factor in terms of a
C percentage value multiplied by a geometric parameter
C (e.g., shell thickness)
C====================================================================
C
WRITE(6,2020)
READ(5,*,IOSTAT = J) FACTOR
IF (J .NE. 0) GOTO 950
ICYCLE = ICYCLE + 1
I101 = 0
C
C====================================================================
C Compute the perturbed mesh. This section assumes that nodal
C displacements are in the GLOBAL coordinate system. If they are
C not, the correct transformations should be applied prior to
C perturbing the mesh. The user should supply this coding. Also,
C only the translational degrees of freedom should be used for the
C perturbation (X = Xo + u).
C====================================================================
C====================================================================
C Subroutine NODEGEN is the actual mesh generator.
C====================================================================
CALL NODEGEN(COORD,DISP,I1901,FACTOR)
ENDIF
ENDIF
C
ENDDO
C
C====================================================================
C Next line is added for diagnostics
C
C====================================================================
IF (NODEMAX .NE. I1901) THEN
WRITE(*,2025) NODEMAX,I1901
GOTO 950
11-1209
ENDIF
C
IF (ICYCLE .LE. 1) THEN
WRITE(*,30)
30 FORMAT(2X,'. . . NO EIGENVECTORS WERE FOUND . . .',/
1 2X,'The input file for the buckling analysis must contain:',
2 /,2X,'--> *NODE FILE <--',
3 /,2X,'--> U <--')
GOTO 950
ENDIF
C
C===================================================================
C Output the coordinates of the perturbed mesh and close the file.
C===================================================================
WRITE(*,100) OUTFILE
100 FORMAT(//,2X,'The perturbed mesh data are being written to:',
1 1X,A,//)
C
DO K = 1, NODEMAX
WRITE(15,110) INODE(K), (COORD(J,K),J = 1, ITRANS)
110 FORMAT(I6,3(',',1PE14.6))
ENDDO
CLOSE (15)
C
C====================================================================
C
900 FORMAT(//,
1 /,2X,'TROUBLE OPENING FILE',1X,A)
950 WRITE(*,1000)
1000 FORMAT(//,
1 /,2X,' . . . TROUBLE READING DATA . . . ',
2 /,2X,' . . . PROGRAM STOPPED . . . ',/)
2000 FORMAT(//,
1 /,' +----------------------------------------------+',
2 /,' | |',
3 /,' | P R O G R A M --- F P E R T |',
4 /,' | P E R T U R B E D M E S H |',
5 /,' | G E N E R A T O R |',
6 /,' | |',
7 /,' +----------------------------------------------+',//)
2010 FORMAT(//,
1 /,2X,'Number of nodes in model . . . . . . . ',I5,
2 /,2X,'Number of elements in model . . . . . . . ',I5)
11-1210
2015 FORMAT(//,
1 /,2X,'Number of mode shapes available . . . . . . .',I5,
2 //,2X,'Enter the mode shape(s) to be used in calculating',
3 /,2x,'the perturbed mesh (zero when finished):')
2020 FORMAT(//,
1 /,2X,'Enter the imperfection factor to be introduced ',
2 /,2X,'into the geometry for this eigenmode: ')
2025 FORMAT(//,
1 /,2X,'. . . TROUBLE READING COORDINATE DATA . . . ',
2 /,2X,'Number of coordinates in model . . . . . .',I5,
3 /,2X,'Number of coordinates read . . . . . . . .',I5)
RETURN
C====================================================================
C
SUBROUTINE NODEGEN(COORD,DISP,I1901,FACTOR)
C
C====================================================================
C PURPOSE: Defines new coordinate data based upon a fraction of the
C eigenvector obtained in a buckling analysis
C
C INPUT:
C
C COORD = Original coordinate data
C DISP = Displacement data (eigenvector)
C I1901 = Total number of nodes
C FACTOR = Imperfection factor (e.g., percentage of shell
C thickness)
C====================================================================
C
DIMENSION COORD(3,*),DISP(6,*)
C
DO I = 1, I1901
COORD(1,I) = COORD(1,I) + FACTOR*DISP(1,I)
ENDDO
RETURN
END
11.1.5 Output radiation viewfactors and facet areas: FRAD

11-1211
This example illustrates the use of a FORTRAN program to read the radiation viewfactors and the
facet areas from the results file.
General description
The program shows how to retrieve the viewfactors and the facet areas from the results file. The results
file created from the benchmark problem detailed in `Àxisymmetric elemental cavity radiation
viewfactor calculations,'' Section 1.6.6 of the ABAQUS Benchmarks Manual, is used to verify that the
output records have been read and output correctly. This program will assume that the results file has
the default file extension, .fil, which corresponds to FORTRAN unit 8.
Programming details
Before proceeding, review the general discussion on programming concepts and ABAQUS FORTRAN
interfaces in `Ùser postprocessing of ABAQUS results files: overview,'' Section 11.1.1, and the
detailed discussion of postprocessing given in Chapter 5, "File Output Format," of the
When running the program FRAD (this program is named frad.f on the ABAQUS release media),
the user will be prompted for the file name that initializes FNAME. Other variables, such as LOUTF,
NRU, LRUNIT(1,NRU), and LRUNIT(2,NRU), are initialized inside the program. INITPF and
DBNRU are then called to complete the necessary initializations and file connections. By default, the
results file is processed for all steps and increments in the results file. The user can restrict the output
by setting LSTEPA and LINCA to the required step and increment and uncommenting the simple IF
- END IF block. Data processing starts with a DO-loop over all the records to be read, one-by-one,
by means of a call to DBFILE. Each record is identified by its record key, which is stored in the
second entry of the record. When records 1922 and 2000 are processed by FRAD, the heading and the
current step and increment numbers are written out so as to provide a way to recognize the beginning
of data in each analysis. Record types 1605, 1606, 1607, and 1609 are then read; and the desired
output is written to the output file vfout.

Before it can be executed, the FORTRAN program must be compiled and linked. Both operations, as
well as the inclusion of the aba_param.inc file, are performed by a single execution of the
ABAQUS/Make procedure:
abaqus make job=frad
This procedure may have to be repeated until all FORTRAN errors are corrected. After successful
compilation, the program's object code is linked automatically with the ABAQUS object codes stored
in the shared program library and the interface library to build the executable program. Refer to the
ABAQUS Site Guide for information about the compile and link commands for a particular computer.
Program execution
Before the program is executed, a results file must have been created with the desired output being
written to that file. In this example the results file xrvda4n1.fil created by running the input file
11-1212
xrvda4n1.inp discussed in `Àxisymmetric elemental cavity radiation viewfactor calculations,''

Section 1.6.6 of the ABAQUS Benchmarks Manual, is used. When the program is executed using the
command abaqus frad, the prompt
will appear. Enter xrvda4n1 to define FNAME. The program processes the data and produces a file
named vfout, which contains the required information.

The radiation viewfactors and facet areas are read and output to vfout. The output agrees with the
expected results.
Input file
frad.f
Sample listings
11-1213
Listing 11.1.5-1
SUBROUTINE HKSMAIN
C====================================================================
C abaqus make job=frad
C abaqus frad
C====================================================================
C
C PURPOSE:
C This program reads the results file and outputs the radiation
C viewfactors, and facet areas associated with different facets
C in a cavity.
C
C PROMPTS:
C 1. 'Enter the name of the results file (w/o .fil):'
C
C====================================================================
C
C INPUT FILE ---- 'FNAME'.fil
C OUTPUT FILE ---- vfout
C
C====================================================================
C
DIMENSION ARRAY(513),JRRAY(NPRECD,513),LRUNIT(2,1)
EQUIVALENCE (ARRAY(1),JRRAY(1,1))
DIMENSION COORD(3),TRACT(3),CLEAR(3)
C
CHARACTER*80 FNAME,OUTFILE,TEMP
CHARACTER*8 TEMP1
CHARACTER*3 NULL
LOGICAL STRNCMP
PARAMETER (ZERO=0.0D0,ONE=1.0D0,TWO=2.0D0,NULL=' ')
C------------------------------------------------------------
C NRU Number of results files (*.fil) to be read
C LRUNIT(1,*) Unit number of results file
C LRUNIT(2,*) Format of input file (1 = ASCII, 2 = BINARY)
C LOUTF Format of output file (not needed here)
C COORD(*) Stores the nodal coordinates
C IXX Record length
C KEY Record key
11-1214
C LSTEPA Step Number

C LINCA Increment Number
C------------------------------------------------------------
C
C--- Initialize File ---------------------------------------
NRU = 1
LRUNIT(1,NRU) = 8
LRUNIT(2,NRU) = 2
LOUTF = 0
OUTFILE = 'vfout'
FNAME = NULL
C--- Change below to set to desired step and increment when
C--- requesting restricted output; Default is for all steps
C--- and increments available in the results file
LSTEPA = 1
LINCA = 1
C---
C
C Get the name of the results (.fil) file
C
C------------------------------------------------------------
WRITE(6,*) ' Enter the name of the results file (w/o .fil): '
READ (5,'(A)', IOSTAT=J) FNAME
IF (J .NE. 0) WRITE(*,*) 'ERROR IN READING INPUT DATA'
C
C
C Access ABAQUS libaries and open input file
C------------------------------------------------------------
CALL INITPF(FNAME, NRU, LRUNIT, LOUTF)

OPEN(UNIT=15,FILE=OUTFILE,STATUS='UNKNOWN')
JUNIT = LRUNIT(1,NRU)
CALL DBRNU(JUNIT)
C------------------------------------------------------------
C End access
C
if (jrcd .ne. 0) write(15,*) 'ERROR IN FILE-ACCESS'
C
C IXX = Record length
C
11-1215
IXX = JRRAY(1,1)
C
C KEY = Record type key
C
KEY = JRRAY(1,2)
C
C Output Request Definition
C
IF(KEY .EQ. 2000) THEN

LINC = JRRAY(1,9)
LSTEP= JRRAY(1,8)
END IF
C Output desired quantities to the output file

C
C-----------------------------------------------------------
C Uncomment below to get output for specific increment/step.
C Default output is for all increments available in the file.
C-----------------------------------------------------------
C IF((LINC .EQ. LINCA) .AND. (LSTEP .EQ. LSTEPA)) THEN
IF (KEY .EQ. 2000) THEN

WRITE(15,'(/,1X,A,1X,I6,1X,A,I6)')'STEP NUMBER', LSTEP,
1 'INCREMENT',LINC
WRITE(*,'(/,1X,A,1X,I6,1X,A,I6)') 'WRITING OUTPUT FOR STEP',
1 LSTEP, 'INCREMENT', LINC
END IF
IF(KEY .EQ. 1605)THEN

NUMFACETS = JRRAY(1,3)
WRITE(TEMP,'(A)') ARRAY(4)
WRITE(15,30)TEMP,NUMFACETS
30 FORMAT(//,1X,'CAVITY NAME:',1X,A,/,
1 1X,'NUMBER OF FACETS:',1X,I6)
ELSEIF(KEY .EQ. 1609) THEN
MVFLEN = JRRAY(1,3)
MATRIXSIZE = INT(SQRT(REAL(MVFLEN)))
ELSEIF( KEY .EQ. 1606 ) THEN
WRITE(15,'(/,1X,A)')'VIEWFACTOR MATRIX (ROW - COLUMN)'
WRITE(15,40) (ARRAY(I),I=3,JRRAY(1,1))
40 FORMAT(1X,1PE14.6,', ')
ELSEIF(KEY .EQ. 1607) THEN
11-1216
WRITE(15,'(/,1X,A)') 'FACET AREA'

WRITE(15,50)(ARRAY(I),I=3,JRRAY(1,1))
50 FORMAT(1X,1PE14.6,', ')
ENDIF
C ENDIF
ENDDO
RETURN
END
11.1.6 Creation of a data file to facilitate the postprocessing of

elbow element results: FELBOW
This example illustrates the use of a FORTRAN program to read selected element integration point
records from an ABAQUS results file to facilitate the postprocessing of elbow element results. X-Y
data are created that are suitable for use with the X-Y plotting capability in ABAQUS/Viewer.
General description
This program shows how to retrieve integration point data for elbow elements from an ABAQUS
results file to visualize one of the following:
1. Variation of a variable along a line of elbow elements,
2. Variation of a variable around the circumference of a given elbow element, or
3. Ovalization of a given elbow element.
An ASCII file containing X-Y data is created that can be read into ABAQUS/Viewer for visualization
purposes.
To execute option 1, the elbow elements must be numbered such that they increase monotonically
within the range of elements considered; all elements in the desired range must be elbow elements. X-Y
data will be created with the X-data being the distance along the line of elbow elements, measured
along the elbow centerline and the Y -data being the variable value. The user must ensure that the
integration point coordinates (COORD) are written to the results file if either option 2 or 3 is needed.
For option 2 X-data are the distance around the circumference of the elbow element, measured along
the middle surface, and Y -data are the variable value. For option 3 the X-Y data are the current
11-1217
coordinates of the middle-surface integration points around the circumference of the elbow element,
projected to a local coordinate system in the plane of the deformed cross-section. The origin of the
local system coincides with the center of the cross-section; the plane of the deformed cross-section is
defined as the plane that contains the center of the cross-section and integration points 1 and 2.
Programming details
The user is prompted for the name of the results file (assumed to be binary) and the postprocessing
option (1, 2, or 3). The user is then prompted for additional information depending on the option that
was chosen; this information includes
· The range of element numbers (options 2 and 3 require only a single element number),
· The section point number (options 1 and 2 only),
· The integration point number (option 1 only),
· The element variable (options 1 and 2 only),
· The component of the variable (as defined in ``Results file output format,'' Section 5.1.2 of the
ABAQUS/Standard User's Manual, options 1 and 2 only),
· The step number, and
· The increment number.
The data are processed in a double DO-loop over all records, via a call to DBFILE. The desired data
are stored in variable VAR; the integration point coordinates are stored in COORDS. The program
checks to make sure the requested data are available in the results file. An error is issued if the user
tries to process data that are not found in the results file.

Before program execution, compile and link the FORTRAN program by using the ABAQUS/Make
procedure:
abaqus make job=felbow
Repeat this command until all FORTRAN errors are corrected. After successful compilation, the
program's object code is linked automatically with the ABAQUS object codes stored in the shared
program library and interface library to build the executable program. Refer to Chapter 3,
"Environment file," of the ABAQUS Site Guide to see which compile and link commands are used for
a particular computer.
Program execution
Before executing the program, run an analysis that creates a results file containing the appropriate
output. This analysis includes, for example, output for the elements in a given range and the integration
point coordinates of the elements. When the program is executed using the command abaqus
felbow, the prompt
11-1218

will appear. Enter the name of the results file to define FNAME. The user is then prompted for other
information, such as the desired postprocessing option, element number, etc. The program processes
the data and produces a file named output.dat that contains the information required to visualize
the elbow element results.

`Èlastic-plastic collapse of a thin-walled elbow under in-plane bending and internal pressure, '' Section
1.1.2, contains several figures created with the aid of this program. The output agrees with the
expected results.
Input file
felbow.f
Sample listings
11-1219
Listing 11.1.6-1
SUBROUTINE HKSMAIN
C====================================================================
C abaqus make job=felbow
C abaqus felbow
C==================================================================
C
C PURPOSE:
C
C This program reads results stored in an ABAQUS results (.fil)
C file and creates an input file for ABAQUS/Viewer that permits
C postprocessing and visualization of elbow element results.
C
C
C Input file names: `fname.fil', where fname is the root file name
C of the input file.
C
C Output file name: output.dat
C
C==================================================================
C
C VARIABLES USED BY THIS PROGRAM:
C
C ndi -- Number of direct components in stress/strain tensor.
C nshr -- Number of shear components in stress/strain tensor.
C ndip1 -- ndi + 1
C array -- Real array containing values read from results file.
C Equivalenced to jrray.
C jrray -- Integer array containing values read from results file.
C Equivalenced to array.
C fname -- Root file name of input file (w/o .fil extension).
C nru -- Number of results files (.fil) to be read.
C lrunit -- Array containing unit number and format of results files:
C lrunit(1,*) --> Unit number of input file.
C lrunit(2,*) --> Format of input file.
C loutf -- Format of output file:
C 0 --> Standard ASCII format.
C junit -- Unit number of file to be opened.
11-1220
C jrcd -- Error check return code.

C key -- Current record key identifier.
C jelnum -- Current element number.
C intpn -- Integration point number.
C
C==================================================================
C
C ABA_PARAM.INC defines an appropriate IMPLICIT REAL statement
C and sets the value of NPRECD to 1 or 2, depending on whether
C the machine uses single or double precision. ABA_PARAM.INC is
C referenced from the \site (for NT) or /site (for Unix)
C ABAQUS release subdirectory when ABAQUS/Make is executed.
C
C==================================================================
C
include 'aba_param.inc'
parameter (melem=10000,mnode=10000,mkey=3000,mintpt=360)
dimension array(513),jrray(nprecd,513),lrunit(2,1)
equivalence (array(1),jrray(1,1))
C
C==================================================================
C
character fname*80
C
C SOME DIMENSION STATEMENTS MAY BE NEEDED IF YOU ARE DOING ADDITIONAL
C CALCULATIONS ON THE DATA
C
dimension var(513,melem),coords(3,mintpt)
dimension iconn(3,melem),loc_el(melem)
dimension naxipt(melem),iglb_nod(mnode),loc_nod(mnode)
dimension islct(mkey),nintpt(melem),nodel(melem),iglb_el(melem)
dimension xyz(3,mnode),xpos(mnode),tang1(3),tang2(3)
dimension rot(3,3),xc(3),xpr(3,mintpt),scord(3,mintpt),xnorm(3)
dimension isctchk(melem),icrdchk(melem),ivarchk(melem)
C
C==================================================================
C
C GET THE NAME OF THE RESULTS FILE.
11-1221
C
C==================================================================
C
isect1 = 0
intpt1 = 0
icomp = 1
C
write(6,*) 'Enter the name of the input file (w/o .fil):'
read(5,'(A)') fname
C
100 write(6,*) 'Enter the postprocessing option:'
write(6,*) ' 1 - variation along the riser '
write(6,*) ' 2 - variation around the circumference'
write(6,*) ' of the elbow'
write(6,*) ' 3 - ovalization of elbow cross-section'
read(5,*) ipost
if (ipost .ne. 1 .and. ipost .ne. 2 .and.

& ipost .ne. 3) then
write(6,*)'Invalid entry - try again'
go to 100
endif
C
write(6,*)'Enter the first element number'
read(5,*) jel_1
C
if (ipost .ne. 1) then
jel_2 = jel_1
else
write(6,*)'Enter the last element number'
read(5,*) jel_2
endif
if (jel_2 .lt. jel_1) then
write(6,*)'last elem number must be greater than first'
return
endif
jdiff = jel_2 - jel_1 + 1
if (jdiff .gt. melem) then
write(6,*)'Too many elements - max is',melem
return
endif
C
if (ipost .lt. 3) then
11-1222
write(6,*)'Enter the section point number'

read(5,*) isect1
endif
C
if (ipost .eq. 1) then
write(6,*)'Enter the integration point number'
read(5,*) intpt1
endif
C
if (ipost .lt. 3) then
998 write(6,999)
999 format('Enter the key for the variable you wish to process:')
C
write(6,500)
500 format(2x,'TEMP ...... 2',/,
& 2x,'COORD...... 8',/,
& 2x,'S.......... 11',/,
& 2x,'SINV....... 12',/,
& 2x,'SF......... 13',/,
& 2x,'E.......... 21',/,
& 2x,'PE......... 22',/,
& 2x,'CE......... 23',/,
& 2x,'IE......... 24',/,
& 2x,'EE......... 25',/,
& 2x,'THE........ 88',/,
& 2x,'LE......... 89',/,
& 2x,'NE......... 90',/,
& 2x,'SP.........401',/,
& 2x,'EP.........403',/,
& 2x,'NEP........404',/,
& 2x,'LEP........405',/,
& 2x,'EEP........408',/,
& 2x,'IEP........409',/,
& 2x,'THEP.......410',/,
& 2x,'PEP........411',/,
& 2x,'CEP........412')
C
read(5,*) key
C
if (key .ne. 2 .and. key .ne. 8 .and.
& key .ne. 11 .and. key .ne. 12 .and.
11-1223

& key .ne. 411 .and. key .ne. 412) then
C
write(6,*)'Invalid key - try again'
go to 998
C
endif
C
C Integration point coordinates available only at middle surface
C Overwrite the section point and set to the default
C
if (key .eq. 8) isect1 = 0
C
if (key .ge. 8) then
write(6,*)'Enter the attribute number'
read(5,*) icomp
endif
C
if (key .eq. 13) then
write(6,*)'Section force data written only once per element'
write(6,*)'Will use default integration & section point'
write(6,*)' '
intpt1 = 1
isect1 = 0
elseif (ipost .gt. 1) then
write(6,*)'Only use with option 1: Processing terminated'
return
endif
endif
C
endif
C
write(6,*)' Enter the step number'
read(5,*) istep1
write(6,*)' Enter the increment number'
read(5,*) iinc1
11-1224
C
C SELECT THE RECORDS TO BE PROCESSED
C
C IF THE ISLCT ARRAY IS SET TO 1 THE DATA WILL BE PROCESSED
C IF THE ISLCT ARRAY IS SET TO 0 THE DATA WILL NOT BE PROCESSED
C
do ii = 1, mkey
islct(ii) = 0
end do
C
C ELEMENT DEFINITION
islct(1900)=0
C NODE DEFINITION
islct(1901)=0
C INCREMENT START
islct(2000)=0
C ELEMENT HEADER
islct(1)=0
C COORD
islct(8)=0
C
if (ipost .lt. 3) islct(key) = 1
keyprv=0
C
nels = 0
numnp = 0
C
itimchk = 0
ielchk = 0
do i = 1, melem
loc_el(i) = 0
isctchk(i) = 0
icrdchk(i) = 0
ivarchk(i) = 0
do j = 1, 513
var(j,i) = 0.0
end do
end do
C
C==================================================================
C
C OPEN THE OUTPUT FILE.
C
11-1225
C==================================================================
open(unit=9,file='output.dat',status='unknown')
C
nru = 1
loutf = 0
lrunit(1,1) = 8
lrunit(2,1) = 2
C
call initpf(fname,nru,lrunit,loutf)
C
junit = 8
C
call dbrnu(junit)
C
C REWIND FILE
C
call dbfile(2,array,jrcd)
C
C
C==================================================================
C
C Read records from the results (.fil) file and process the data.
C Cover a maximum of 10 million records in the file.
C
C==================================================================
C
do 1000 k100 = 1, 100
do 1000 k1 = 1, 99999
C
C READ RECORD FROM .FIL FILE AND PROCESS DATA.
C
call dbfile(0, array, jrcd)
if (jrcd .ne. 0 .and. keyprv .eq. 2001) then

write(6,*)'End of file'
close(junit)
go to 1001
else if (jrcd .ne. 0) then
write(6,*)'Error reading input file'
return
endif
C
C LENGTH AND KEY RECORD NUMBER ARE ALWAYS NEEDED
11-1226
C
lenf= jrray(1,1)
key = jrray(1,2)
C
C SPECIFIC RECORD NUMBERS FOLLOW...
C
C==================================================================
C
C ELEMENT DEFINITIONS
C
C==================================================================
if ( key .eq. 1900 ) then
jelnum=jrray(1,3)
if (jelnum .gt. melem) then
write(6,*)'Maximum global elem number too large'
write(6,*)'Increase melem'
return
endif
ctype= array(4)
if (jelnum .ge. jel_1 .and.
& jelnum .le. jel_2) then
ielchk = 1
nels = nels + 1
jel = nels
iglb_el(jel) = jelnum
loc_el(jelnum) = jel
nodel(jel) = lenf - 4
naxipt(jel) = 1
if (ctype .ne. 'ELBOW31' .and.
& ctype .ne. 'ELBOW31B' .and.
& ctype .ne. 'ELBOW31C' .and.
& ctype .ne. 'ELBOW32') then
write(6,*)'Invalid element type encountered'
return
endif
if (ctype .eq. 'ELBOW32')naxipt(jel) = 2
do kk=5,lenf
ii=kk-4
iconn(ii,jel)=jrray(1,kk)
end do
endif
keyprv = key
goto 9099
11-1227
C
C==================================================================
C
C NODE DEFINITIONS
C
C==================================================================
else if ( key .eq. 1901 ) then
jnod = jrray(1,3)
if (jnod .gt. mnode) then
write(6,*)'Maximum global node number exceeded'
write(6,*)'Increase mnode'
return
endif
numnp = numnp + 1
iglb_nod(numnp) = jnod
loc_nod(jnod) = numnp
xyz(1,numnp) = array(4)
keyprv = key
goto 9099
C
C==================================================================
C
C CURRENT STEP AND INCREMENT NUMBER.
C
C==================================================================
ttime=array(3)
stime=array(4)
sfreq=array(12)
istep=jrray(1,8)
iinc =jrray(1,9)
C
C SET THE FLAG ITIME IF THIS IS THE REQUESTED STEP/INC
C
if (istep .eq. istep1 .and.
& iinc .eq. iinc1) then
itime = 1
itimchk = 1
else
itime = 0
11-1228
endif
c
keyprv = key
goto 9099
C
C==================================================================
C
C ELEMENT VARIABLES AT INTEGRATION POINTS WITHIN THE ELEMENT
C
C==================================================================
C
C ELEMENT DATA
C
C
C PROCESS THIS RECORD IF THIS IS THE CORRECT STEP/TIME
C
if (itime .eq. 1) then
jelnum = jrray(1,3)
jel = loc_el(jelnum)
ielem = 0
C
C PROCESS THIS ELEMENT IF IT IS IN THE LIST OF DESIRED ELEMS
C
if (jel .gt. 0) then
ielem = 1
intpn = jrray(1,4)
if (intpn .gt. mintpt) then
write(6,*)'Max number of int points exceeded'
write(6,*)'Increase mintpt'
return
endif
isect = jrray(1,5)
ilocn = jrray(1,6)
if (ilocn .ne. 0) then
write(6,*)'Element data must be at the int points'
return
endif
crbar = array(7)
ndi = jrray(1,8)
nshr = jrray(1,9)
ndir = jrray(1,10)
11-1229
nsfc = jrray(1,11)
if (isect .eq. isect1)then
isctflg = 1
isctchk(jel) = 1
else
isctflg = 0
endif
if ( (ipost .eq.1 .and. intpn .eq. intpt1) .or.
& ipost .gt. 1) then
iptflg = 1
else
iptflg = 0
endif
endif
endif
keyprv = key
goto 9099
C
C==================================================================
C
C COORDINATES
C
C==================================================================
else if ( key .eq. 8 .and. ipost .gt. 1) then
C
C PROCESS IF THIS IS THE CORRECT STEP/INC, ELEMENT, SECTION
C POINT, AND INTEGRATION POINT
C
icheck = itime*ielem*iptflg
if (icheck .eq. 1 ) then
c if (ipost .eq. 3) nintpt(jel) = intpn
icrdchk(jel) = 1
if (ipost .gt. 1) nintpt(jel) = intpn
do kk=3,lenf
ii=kk-2
coords(ii,intpn)=array(kk)
end do
C
if (islct(key) .eq. 1 .and. ipost .lt. 3) then
ivarchk(jel) = 1
jvar = intpn
nintpt(jel) = intpn
do kk=3,lenf
11-1230
ii=kk-2
var(ii,jvar)=array(kk)
end do
if (icomp .gt. ii) then
write(6,*)'Invalid component'
return
endif
endif
C
endif
keyprv = key
goto 9099
C
C==================================================================
C
C VARIABLE RECORD
C
C==================================================================
C PROCESS IF THIS IS THE DESIRED VARIABLE
else if (islct(key) .eq. 1 .and. ipost .lt. 3) then

C
C CONTINUE PROCESSING IF THIS IS THE CORRECT STEP/INC, ELEMENT,
C SECTION POINT, AND INTEGRATION POINT
C
icheck = itime*ielem*isctflg*iptflg
if (icheck .eq. 1 ) then
ivarchk(jel) = 1
jvar = jel
else
jvar = intpn
endif
nintpt(jel) = intpn
do kk=3,lenf
ii=kk-2
var(ii,jvar)=array(kk)
end do
if (icomp .gt. ii) then
write(6,*)'Invalid component'
return
endif
endif
11-1231
keyprv = key
goto 9099
C
C==================================================================
C
C ANYTHING ELSE
C
C==================================================================
else
c write (6,9000) key
9000 format(I5,' *** NO CODING FOR THIS RECORD ***')
keyprv = key
9099 end if
C
C==================================================================
C
C END LOOPS
C
C==================================================================
1000 continue
1001 continue
C
C==================================================================
C POSTPROCESS DATA HERE
C==================================================================
do ii = 1, nels
nintpt(ii) = nintpt(ii)/naxipt(ii)
end do
C
C==================================================================
C VERIFY THAT APPROPRIATE DATA WERE WRITTEN TO RESULTS FILE
C==================================================================
c iquit = 0
if (ielchk .eq. 0) then
write(6,*)'Desired element not found'
return
endif
C
if (itimchk .eq. 0) then
write(6,*)'Desired step/increment not found'
return
endif
C
11-1232
do iel = 1, nels
if (isctchk(iel) .eq. 0 .and. ipost .lt. 3) then
write(6,*)'Desired element and/or section point not found'
return
endif
if (icrdchk(iel) .eq. 0 .and. ipost .gt. 1) then
write(6,*)'Integration point coords not found'
return
endif
if (ivarchk(iel) .eq. 0 .and. ipost .lt. 3) then
write(6,*)'Desired element variable not found'
return
endif
end do
C
C****************************************
C
C PLOT VARIABLE ALONG ELBOW LENGTH
C
C****************************************
C
C
nn_old = iconn(nodel(1),1)
do iel = 1, nels
nn_new = iconn(1,iel)
if (iel .gt. 1 .and. nn_new .ne. nn_old) then
write(6,*)'Error in element connectivity'
return
endif
nnum_b = iconn(1,iel)
nnum_b = loc_nod(nnum_b)
nnum_e = iconn(nodel(iel),iel)
nnum_e = loc_nod(nnum_e)
if (iel .eq. 1) xpos(nnum_b) = 0.0
xlength = 0.d+0
do jj = 1, 3
delx = xyz(jj,nnum_e) - xyz(jj,nnum_b)
xlength = xlength + delx*delx
end do
xlength = sqrt(xlength)
xpos(nnum_e) = xpos(nnum_b) + xlength
x_mid = 0.5*(xpos(nnum_e) + xpos(nnum_b))
11-1233
C
C EXTRACT VARIABLE INFORMATION AND WRITE TO FILE
C
data = var(icomp,iel)
write(9,2000)x_mid,data
2000 format(5x,e17.9,5x,e17.9)
C
nn_old=iconn(nodel(iel),iel)
C
end do
return
C
C
C****************************************
C
C PLOT VARIABLE AROUND CIRCUMFERENCE OF ELBOW
C
C****************************************
C
else if (ipost .eq. 2) then
C
xpos(1) = 0.0
C
locel = loc_el(jel_1)
iaxial = 1
C
if (naxipt(locel) .eq. 2) then
3000 write(6,*)'Enter the axial station'
read(5,*)iaxial
if (iaxial .ne. 1 .and. iaxial .ne. 2) then
write(6,*)'Try again - station must be 1 or 2'
go to 3000
endif
endif
C
do ipt = 1 , nintpt(locel)
jpt = ipt + (iaxial - 1)*nintpt(locel)
if (ipt .gt. 1) then
xlength = 0.0
do jj = 1, 3
delx = coords(jj,jpt) - coords(jj,jpt-1)
xlength = xlength + delx*delx
end do
11-1234
xlength = sqrt(xlength)
xpos(ipt) = xpos(ipt-1) + xlength
endif
C
C EXTRACT VARIABLE INFORMATION AND WRITE TO FILE
C
data = var(icomp,jpt)
write(9,2000)xpos(ipt),data
C
end do
ipt = 1
data = var(icomp,jpt)
xfinal = xpos(nintpt(locel)) + (xpos(2)- xpos(1))
write(9,2000)xfinal,data
return
C
C****************************************
C
C OVALIZATION PLOT
C
C****************************************
C
else if (ipost .eq. 3) then
C
locel = loc_el(jel_1)
iaxial = 1
C
if (naxipt(locel) .eq. 2) then
3500 write(6,*)'Enter the axial station'
read(5,*)iaxial
if (iaxial .ne. 1 .and. iaxial .ne. 2) then
write(6,*)'Try again - station must be 1 or 2'
go to 3500
endif
endif
C
C AVERAGE THE COORDS OF THE INTEGRATION POINTS TO GET THE CENTER
C
do i = 1, 3
xc(i) = 0.0
11-1235
end do
C
xnint = real(nintpt(locel))
C
C
do i = 1, 3
xc(i) = xc(i) + coords(i,jpt)/xnint
end do
C
end do
C
C SHIFT THE COORDINATES SO THAT CENTER OF SECTION IS AT ORIGIN
C
do i = 1, 3
scord(i,jpt) = coords(i,jpt) - xc(i)
end do
C
end do
C
C DETERMINE TANGENT VECTOR 1 (FROM CENTER OF CROSS-SECTION
C TO INT PT 1)
C
jpt = 1 + (iaxial - 1)*nintpt(locel)
xmag = 0.0
do i = 1, 3
tang1(i) = coords(i,jpt) - xc(i)
xmag = xmag + tang1(i)*tang1(i)
end do
xmag = sqrt(xmag)
do i = 1, 3
tang1(i) = tang1(i)/xmag
end do
C
C GUESS TANGENT VECTOR 2 (FROM CENTER OF CROSS-SECTION
C TO INT PT 2)
C
jpt = 2 + (iaxial - 1)*nintpt(locel)
xmag = 0.0
do i = 1, 3
11-1236
tang2(i) = coords(i,jpt) - xc(i)

xmag = xmag + tang2(i)*tang2(i)
end do
xmag = sqrt(xmag)
do i = 1, 3
tang2(i) = tang2(i)/xmag
end do
C
C DETERMINE THE NORMAL VECTOR, TANG1 X TANG2
C
xnorm(1) = tang1(2)*tang2(3) - tang1(3)*tang2(2)
xnorm(2) = - tang1(1)*tang2(3) + tang1(3)*tang2(1)
xnorm(3) = tang1(1)*tang2(2) - tang1(2)*tang2(1)
xmag = 0.0
do i = 1, 3
xmag = xmag + xnorm(i)*xnorm(i)
end do
xmag = sqrt(xmag)
do i = 1, 3
xnorm(i) = xnorm(i)/xmag
end do
C
C REDEFINE TANGENT VECTOR 2
C
tang2(1) = - tang1(2)*xnorm(3) + tang1(3)*xnorm(2)
tang2(2) = tang1(1)*xnorm(3) - tang1(3)*xnorm(1)
tang2(3) = - tang1(1)*xnorm(2) + tang1(2)*xnorm(1)
C
C DETERMINE THE ROTATION MATRIX
C
do i = 1, 3
rot(1,i) = tang1(i)
rot(2,i) = tang2(i)
rot(3,i) = xnorm(i)
end do
C
C TRANSFORM COORDINATES INTO LOCAL SYSTEM
C
do i = 1, 3
xpr(i,ipt) = 0.0
11-1237
do j = 1, 3
xpr(i,ipt) = xpr(i,ipt) + rot(i,j)*scord(j,jpt)
end do
end do
end do
C
C OUTPUT COORDS TO FILE
C
do ipt = 1, nintpt(locel)
write(9,4000)(xpr(i,ipt),i=1,2)
4000 format(5x,e17.9,5x,e17.9,5x,e17.9)
end do
C
ipt = 1
write(9,4000)(xpr(i,ipt),i=1,2)
endif
C
close (unit=9)
C
return
end
11-1238
Product Index
ABAQUS/Standard
Section 1.1.1 Axisymmetric analysis of bolted pipe flange connections
Section 1.1.2 Elastic-plastic collapse of a thin-walled elbow under in-plane bending and internal
pressure
Section 1.1.3 Parametric study of a linear elastic pipeline under in-plane bending
Section 1.1.4 Indentation of an elastomeric foam specimen with a hemispherical punch
Section 1.1.5 Collapse of a concrete slab
Section 1.1.6 Jointed rock slope stability
Section 1.1.7 Indentation of a crushable foam specimen with a hemispherical punch
Section 1.1.8 Notched beam under cyclic loading
Section 1.1.9 Hydrostatic fluid elements: modeling an airspring
Section 1.1.10 Shell-to-solid submodeling of a pipe joint
Section 1.1.11 Stress-free element reactivation
Section 1.1.12 Symmetric results transfer for a rubber bushing
Section 1.1.13 Transient loading of a viscoelastic bushing
Section 1.1.15 Damage and failure of a laminated composite plate
Section 1.1.16 Analysis of an automotive boot seal
Section 1.1.17 Pressure penetration analysis of an air duct kiss seal
Section 1.1.18 Self-contact in rubber/foam components: jounce bumper
Section 1.1.19 Self-contact in rubber/foam components: rubber gasket
Section 1.1.20 Submodeling of a stacked sheet metal assembly
Section 1.2.1 Snap-through buckling analysis of circular arches
Section 1.2.2 Laminated composite shells: buckling of a cylindrical panel with a circular hole
Section 1.2.4 Elastic-plastic K-frame structure
Section 1.2.5 Unstable static problem: reinforced plate under compressive loads
Section 1.2.6 Buckling of an imperfection sensitive cylindrical shell
Section 1.3.1 Upsetting of a cylindrical billet in ABAQUS/Standard: quasi-static analysis with
rezoning
Section 1.3.3 Superplastic forming of a rectangular box
Section 1.3.4 Stretching of a thin sheet with a hemispherical punch
Section 1.3.5 Deep drawing of a cylindrical cup
Section 1.3.6 Extrusion of a cylindrical metal bar with frictional heat generation
Section 1.3.8 Axisymmetric forming of a circular cup
Section 1.3.17 Upsetting of a cylindrical billet: coupled temperature-displacement and adiabatic
analysis
Section 1.3.18 Unstable static problem: thermal forming of a metal sheet
0-1239
Section 1.4.1 A plate with a part-through crack: elastic line spring modeling
Section 1.4.2 Conical crack in a half-space with and without submodeling
Section 1.4.3 Elastic-plastic line spring modeling of a finite length cylinder with a part-through
axial flaw
Section 1.4.4 Crack growth in a three-point bend specimen
Section 1.5.1 Springback of two-dimensional draw bending
Section 1.5.2 Deep drawing of a square box
Section 2.1.1 Nonlinear dynamic analysis of a structure with local inelastic collapse
Section 2.1.2 Detroit Edison pipe whip experiment
Section 2.1.5 Pressurized fuel tank with variable shell thickness
Section 2.1.6 Modeling of an automobile suspension
Section 2.1.12 Rigid multi-body mechanism
Section 2.2.1 Analysis of a rotating fan using superelements and cyclic symmetry model
Section 2.2.2 Linear analysis of the Indian Point reactor feedwater line
Section 2.2.3 Response spectra of a three-dimensional frame building
Section 3.1.1 Symmetric results transfer for a static tire analysis
Section 3.1.2 Steady-state rolling analysis of a tire
Section 3.1.3 Subspace-based steady-state dynamic tire analysis
Section 4.1.1 Thermally coupled analysis of a disc brake
Section 4.1.2 Exhaust manifold assemblage
Section 4.1.3 Coolant manifold cover gasketed joint
Section 4.1.4 Radiation analysis of a plane finned surface
Section 5.1.1 Eigenvalue analysis of a piezoelectric transducer
Section 5.2.1 Thermal-electrical modeling of an automotive fuse
Section 6.1.1 Hydrogen diffusion in a vessel wall section
Section 6.1.2 Diffusion toward an elastic crack tip
Section 7.1.1 Coupled acoustic-structural analysis of a car
Section 7.1.2 Fully and sequentially coupled structural acoustics of a muffler
Section 8.1.1 Plane strain consolidation
Section 8.1.2 Calculation of phreatic surface in an earth dam
Section 8.1.3 Axisymmetric simulation of an oil well
Section 8.1.4 Analysis of a pipeline buried in soil
Section 9.1.1 Jack-up foundation analyses
Section 9.1.2 Riser dynamics
Section 10.1.1 The cylinder whip problem
Section 11.1.2 Joining data from multiple results files and converting file format: FJOIN
Section 11.1.3 Calculation of principal stresses and strains and their directions: FPRIN
Section 11.1.4 Creation of a perturbed mesh from original coordinate data and eigenvectors:
FPERT
0-1240
Section 11.1.5 Output radiation viewfactors and facet areas: FRAD
Section 11.1.6 Creation of a data file to facilitate the postprocessing of elbow element results:
FELBOW
ABAQUS/Explicit
Section 1.1.5 Collapse of a concrete slab
Section 1.1.7 Indentation of a crushable foam specimen with a hemispherical punch
Section 1.1.9 Hydrostatic fluid elements: modeling an airspring
Section 1.1.14 Indentation of a thick plate
Section 1.2.3 Buckling of a column with spot welds
Section 1.3.2 Upsetting of a cylindrical billet in ABAQUS/Explicit
Section 1.3.4 Stretching of a thin sheet with a hemispherical punch
Section 1.3.6 Extrusion of a cylindrical metal bar with frictional heat generation
Section 1.3.7 Rolling of thick plates
Section 1.3.8 Axisymmetric forming of a circular cup
Section 1.3.9 Cup/trough forming
Section 1.3.10 Forging with sinusoidal dies
Section 1.3.11 Forging with multiple complex dies
Section 1.3.12 Flat rolling: transient and steady-state
Section 1.3.13 Section rolling
Section 1.3.14 Ring rolling
Section 1.3.15 Axisymmetric extrusion: transient and steady-state
Section 1.3.16 Two-step forming simulation
Section 1.3.17 Upsetting of a cylindrical billet: coupled temperature-displacement and adiabatic
analysis
Section 1.5.1 Springback of two-dimensional draw bending
Section 1.5.2 Deep drawing of a square box
Section 2.1.3 Plate impact simulation
Section 2.1.4 Tennis racket and ball
Section 2.1.7 Explosive pipe closure
Section 2.1.8 Knee bolster impact with double-sided surface contact
Section 2.1.9 Cask drop with foam impact limiter
Section 2.1.10 Oblique impact of a copper rod
Section 2.1.11 Water sloshing in a baffled tank
Section 2.1.12 Rigid multi-body mechanism
Section 4.1.1 Thermally coupled analysis of a disc brake
ABAQUS/Design
0-1241
Section 1.1.12 Symmetric results transfer for a rubber bushing
ABAQUS/Aqua
Section 9.1.1 Jack-up foundation analyses
Section 9.1.2 Riser dynamics
ABAQUS/USA
Section 10.1.1 The cylinder whip problem
0-1242

Abaqus Example Problems Manual

Uploaded by

Abaqus Example Problems Manual

Uploaded by

ABAQUS Example Problems Manual

· The version of ABAQUS that are you using.

· The type of computer on which you are running ABAQUS.

· Workarounds or tests that you have already tried.

ABAQUS Web server

· Frequently asked questions

· ABAQUS systems information and machine requirements

· Error status reports

· ABAQUS documentation price list

· Training seminar schedule

Anonymous ftp site

Writing to technical support

Support for academic institutions

Getting Started with ABAQUS/Standard: This document is a self-paced tutorial designed to

ABAQUS/Standard User's Manual: This volume contains a complete description of the

Quality Assurance Plan: This document describes HKS's QA procedures. It is a controlled

1. Static Stress/Displacement Analyses

Geometry and model

A3 = (AC + AD )=4; A5 = (AD + AE )=4; A7 = (AE + AF )=4;

Loading and boundary conditions

superelement the face µ =90° is constrained with the boundary condition ux = 0.

Results and discussion

Third-level superelement generation data referenced by boltpipeflange_3d_super3_1.inp and

Closures,'' Transactions of ASME, vol. 113, pp. 398-401, 1991.

Figure 1.1.1-1 Schematic of the bolted joint. All dimensions in mm.

Figure 1.1.1-2 Axisymmetric model of the bolted joint.

Figure 1.1.1-3 22.5° segment three-dimensional model of the bolted joint.

** Note: Areas associated with contact nodes are determined

QUADR, 1934, 202, 218, 228

524, 2116, 2128, 5009, 2118, 2121, 2129, 2122, 2117

547, 2196, 2210, 2212, 2198, 2203, 2211, 2204, 2197

1.1.2 Elastic-plastic collapse of a thin-walled elbow under

Geometry and model

Kinematic boundary conditions

Results and discussion

· Goodall, I. W., "Large Deformations in Plastically Deforming Curved Tubes Subjected to

· Sobel, L. H. and S. Z. Newman, "Elastic-Plastic In-Plane Bending and Buckling of an Elbow:

Figure 1.1.2-1 MLTF elbow: geometry and measured material response.

Figure 1.1.2-2 Models for elbow/pipe interaction study.

Figure 1.1.2-3 Moment-rotation response: mesh convergence studies.

Figure 1.1.2-4 Moment-rotation response: pressure dependence.

Figure 1.1.2-6 In-plane bending of an elbow, elastic-plastic collapse moment results.

Figure 1.1.2-9 Ovalization of elements 451 and 751.

*KINEMATIC COUPLING, REF NODE = 100

1.1.3 Parametric study of a linear elastic pipeline under in-plane

Geometry and model

Results and discussion

values ranging from 0.03175 cm to .2032 cm (0.0125 inches to 0.08 inches).

mean_rad = outer_pipe_radius - wall_thick/2.0

outputstring = str(k) + ', ' + str(rf) + '\n'

ELBOW31 model with 6 Fourier modes.

· Lutz, M., and D. Ascher, Learning Python, O'Reilly, 1999.

Figure 1.1.3-1 Pipeline geometry with inward prescribed tip displacements.

Figure 1.1.3-2 Normalized tip reaction force: closing displacement case.

Figure 1.1.3-3 Normalized tip reaction force: opening displacement case.

# geometrical properties of the problem:

1.1.4 Indentation of an elastomeric foam specimen with a

Geometry and model

Procedure and loading definitions in ABAQUS/Standard

Procedure and loading definitions in ABAQUS/Explicit

Design sensitivity analysis

are given under the *PARAMETER SHAPE VARIATION option.

Results and discussion

Figure 1.1.4-1 Model for foam indentation by a spherical punch.

Figure 1.1.4-2 Elastomeric foam stress-strain curves.

Figure 1.1.4-3 Elastic modulus relaxation curve.

Figure 1.1.4-4 Deformation and contour plots of viscoelastic foam, ABAQUS/Standard.

Figure 1.1.4-9 Initial (undeformed) configuration, ABAQUS/Explicit.

Figure 1.1.4-10 Stress S22 contour plot at 0.06 s, ABAQUS/Explicit.

Figure 1.1.4-11 Logarithmic strain in the z-direction at 0.06 s, ABAQUS/Explicit.

Figure 1.1.4-12 Punch reaction force, ABAQUS/Explicit.

Figure 1.1.4-13 Punch reaction force versus indentation, ABAQUS/Explicit.

.0107, .08, .0030

*HYPERFOAM,N=2,TEST DATA INPUT,POISSON=0.0,

1, 0.000, 0.00000, 0.000

1.1.5 Collapse of a concrete slab