LS-DYNA Introductory Oct 30 Nov 1-2012

1
LS-DYNA Introductory Course
Oasys LS-DYNA Environment

TRAINING COURSES
INTRODUCTORY TRAINING
COURSE
The Arup Campus, Blythe Gate, Blythe Valley Park,

Solihull, West Midlands, B90 8AE
tel: +44 (0) 121 213 3399
email: [email protected]
October 30th – November 1st, 2012
LS-DYNA ENVIRONMENT
2
What does this course aim at? LS-DYNA Introductory Course
This course aims at providing the trainee with:
– An overview of LS-DYNA software
– An ability to prepare and run LS-DYNA input files
– An ability to debug LS-DYNA data
– An overview about post-processing the results
Please note that this training course is designed with an emphasis on Explicit
solution capabilities of LS-DYNA, hence topics related to Implicit capabilities of
LS-DYNA will not be dealt with here.
back to contents
LS-DYNA ENVIRONMENT
3
Assumptions LS-DYNA Introductory Course
This course includes a number of hands-on

workshops. For the purpose of Pre-/ Post processing
of LS-DYNA data, the trainer will be using OASYS
Software Suite (PRIMER/ D3PLOT/ T-HIS).
back to contents
LS-DYNA ENVIRONMENT
4
Outline Of Training Course LS-DYNA Introductory Course
• BACKGROUND p6 • THEORY AND PRACTICE

• Time Integration p42
• Element Formulation p53
• INTRODUCTION • Hourglassing p77
• Key principles p13
• Materials p82
• Running p21
• Contact Surfaces p160
• Output files p30
• Rigids p197
• Post Processing p33
• Boundary Conditions p206
• Connections p214
• Control & Output p240
• Other Techniques p274
• ANALYSIS
• Checking p294
back to contents
LS-DYNA ENVIRONMENT
5
• WORKSHOP EXAMPLES
• 1 – Keyword structure p38
• 2 – Time step p52
• 3 – Meshing – crush tube p76
• 4 – Hourglass p81
• 5 – Material – Dogbone p111
• 6 – Material – Foam block p126
• 7 – Contact p191
• 8 – Tied contact p196
• 9 to 13 – Rigid wheel p205
• 14 – Model set-up p328
back to contents
LS-DYNA ENVIRONMENT
6
LS-DYNA Introductory Course B

A
C
K
G
R
O
U
N
D
BACKGROUND
LS-DYNA ENVIRONMENT
7
LS-DYNA Development LS-DYNA Introductory Course B

A
1970’s C
• John Hallquist develops DYNA3D at Lawrence Livermore K
National Laboratory (LLNL). Public domain code. G
Explicit finite element code – solid (brick) elements only. R
O
1980’s
U
• Shell elements added. N
• 1987 Hallquist forms Livermore Software Technology D
Corporation (LSTC).
1990’s
• 1990 1st version of LS-DYNA3D released.
• 2D Capability added to code – renamed LS-DYNA.
• ALE brick elements added.
• Implicit capabilities added.
2000’s
• MPP version adopted as main production version.
• Main production code for vehicle crashworthiness worldwide.
back to contents• Latest Version LS-DYNA 971.
LS-DYNA ENVIRONMENT
8
LS-DYNA Capabilities LS-DYNA Introductory Course B

A
C
• Transient dynamics (linear and non-linear) K
• Non-linear material behaviour (>200 models) G
• Contact surfaces R
O
• Flexible & Rigid Bodies U
• Quasi-static simulations N
• Thermal analysis D
• Fluid / Structure interaction (ALE)

• Implicit solver
back to contents
LS-DYNA ENVIRONMENT
9
LS-DYNA Capabilities LS-DYNA Introductory Course B

A
C
There are two types of time integration schemes available: Explicit and Implicit. K
G
This course concentrates on the EXPLICIT solver.
R
Explicit solution O
U
• Most common solution type, used for impact / crash analyses, short duration
N
events, blast.
D
• Nodal accelerations are calculated using a force balance at every timestep. The
timesteps are very small. The timestep size is set for stability.
Implicit solution
• Used for static or long duration analyses.
• A stiffness matrix is assembled and inverted at each timestep.
• Small number of timesteps that are longer and more expensive compared with the
explicit solution. The timestep size is set for accuracy.
back to contents
LS-DYNA ENVIRONMENT
10
LS-DYNA Applications LS-DYNA Introductory Course B

A
C
• Aerospace
K
• Automotive crashworthiness G
R
• Ballistics
O
• Biomechanics U
N
• Blast
D
• Civil Engineering
• Heat Transfer
• Impact and drop testing
• Manufacturing processes
• Metal forming
• Occupant Safety
• Seismic Engineering
• Rail Engineering
back to contents
LS-DYNA ENVIRONMENT
11
LS-DYNA Introductory Course I

N
T
R
O
D
U
C
T
I
INTRODUCTION O
N
LS-DYNA ENVIRONMENT
12
Outline - INTRODUCTION LS-DYNA Introductory Course I

N
T
1. Key Principles of LS-DYNA R
• Analysis process overview O
D
• Keyword Format U
2. Preparing a Model for LS-DYNA C
• Key Concepts T
I
• Element Hierarchy O
• Units N
3. Submitting DYNA jobs
• Submitting
• Output files
4. Post Processing jobs
5. Workshop
back to contents
LS-DYNA ENVIRONMENT
13
Analysis Process Overview LS-DYNA Introductory Course K

E
Y
• Model preparation (meshing)
Prepare the model – Ansa, Medina P
(Pre-Processing) R
I
• Model preparation (LS-DYNA specific) N
– Oasys Primer, LS PrePost C
I
P
Solve Model • Solution
L
– LS-DYNA E
S
• Post Processing
Review the results
– Oasys D3Plot, T/his, LS PrePost
(Post-Processing)
• Reporting
– Oasys Reporter
Produce a report
back to contents
LS-DYNA ENVIRONMENT
14
Keyword Format LS-DYNA Introductory Course K

E
Y
• The Keyword file (input deck) defines all data needed for LS-DYNA.
• The input file starts with *KEYWORD P
R
• LS-DYNA terminates reading the input file after *END I
• Each Keyword starts with an asterisk * N
C
• The order of the Keywords in the input deck is arbitrary (a few exceptions). I
• Format is fixed (specific columns) or free (comma separated). P
• The cards (lines) after each keyword define any data associated with that keyword. L
E
• There is frequently more than one card for a given keyword. S
• There is a maximum 80 characters to each card (long format under
development).
• Each card is usually 8 fields of 10 characters with some exceptions, e.g.:
– *ELEMENT (typically 10 fields of 8 characters)
– *NODE (uses fields of 16 characters for coordinates, for numerical precision)
– specified in the manual (above the card) if it isn‟t 8 fields of 10 characters –
see also slide 16.
back to contents
LS-DYNA ENVIRONMENT
15
Keyword Format - Example keyword file LS-DYNA Introductory Course K

E
• Comment lines begin with a $ symbol. Y
• Any data after the *END is typically created/used by pre-processors to store P

additional information about the model. (assemblies, connections, dummy positioning etc) R
I
N
C
*KEYWORD
$
I
$ ==================== P
$ MAT (Material) cards L
$ ==================== E
$2345678901234567890123456789012345678901234567890123456789012345678901234567890 S
$
*MAT_PIECEWISE_LINEAR_PLASTICITY
1000000 7.85E-9 206000.0 0.292 300.0 2000.0 0.0 0.0
0.0 0.0 0 0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
$
$
*END
back to contents
LS-DYNA ENVIRONMENT
16

E
Y
• Most lines of data (cards) are 8 fields with a length of 10 characters each.
• Below is an example of a card as it appears in the Keyword manual – the fields are P
numbered at the top 1-8: R
I
N
C
I
P
L
E
S
• Type is generally Floating Point (F) or Integer (I).
• Can also be Character (C) or Alphanumeric (A).
• The default is used if field is blank, zero is entered or if the card is omitted.
• If default is „none‟, a value must be specified.
• Remarks refer to notes in the manual.
• Comma separated variable (free format keyword ) can also be used.
• Fixed and free formats can be mixed throughout the deck but NOT within a card.
back to contents
LS-DYNA ENVIRONMENT
17

E
Y
Some keywords have different options that can be selected
P
e.g. *LOAD_BODY_OPTION R
*ELEMENT_SHELL_{OPTION} I
N
C
• If it is shown as OPTION then one of the possible options must be selected.
I
• If it is shown as {OPTION} then it can be left blank. P
• The available options are specified in the manual: L
E
S
back to contents
LS-DYNA ENVIRONMENT
18
Keyword Hierarchy LS-DYNA Introductory Course K

E
Y
N1
P
R
I
N4 N2
EID N
C
I
P
N3
L
*ELEMENT_SHELL (eid,pid,nid1,nid2,nid3,nid4) E
S
*PART (pid,sid,mid)
*NODE (nid,x,y,z)
*NODE (nid,x,y,z)
*SECTION_SHELL *MAT_ELASTIC *NODE (nid,x,y,z)
(Thickness, formulation) (Material properties) *NODE (nid,x,y,z)
back to contents
LS-DYNA ENVIRONMENT
19
Keyword Hierarchy LS-DYNA Introductory Course K

E
Y
Example showing how a *PART card references the *SECTION and *MAT cards
using their unique id numbers (SECID, MID). P
R
I
N
C
I
P
L
*PART
E
Mounting Plate S
1 10 20 0 0 0 0 0
$
$
*SECTION_SHELL
10 2 0.0 0 0.0 0.0 0 0
1.0 1.0 1.0 1.0 0.0 0.0
$
$
*MAT_ELASTIC
20 7.0E-9 210000.0 0.3 0.0 0.0 0.0
$
back to contents
LS-DYNA ENVIRONMENT
20
Units in LS-DYNA LS-DYNA Introductory Course K

E
Y
LS-DYNA doesn‟t have a fixed set of units. Any set can be used, but they must be
consistent. This can be checked: P
1 (force unit) = 1 (mass unit) x 1 (acceleration unit) R
I
1 (acceleration unit) = 1 (length unit) / [ 1 (time unit) ] 2 N
C
Examples of sets of consistent units are: I
Measurement S1 (SI) S2 S3 Blast P
L
Length metre millimetre millimetre centimetre
E
Time second second millisecond microsecond S
Mass kilogram tonne kilogram gram
Force Newton Newton kiloNewton Mkg f
Pressure Pa MPa GPa 100GPa
Young‟s modulus of steel 210E+09 210E+03 210 2.10
Density of steel 7.85E+03 7.85E-09 7.85E-06 7.85
Gravitation 9.81 9.81E+03 9.81E-03 9.81E-10
back to contents
LS-DYNA ENVIRONMENT
21
Preparing a model LS-DYNA Introductory Course R

U
N
• It is recommended that LS-DYNA models are built using a commercial N
pre-processor. I
N
G
• Most pre-processors (ANSA, Medina etc) will write a LS-DYNA Keyword
Deck, but may not include all LS-DYNA options.
• LS-DYNA specific pre-processors (Oasys PRIMER, LS-PrePost) or text

editors may need to be used to set up more advanced options.
back to contents
LS-DYNA ENVIRONMENT
22
Oasys Primer LS-DYNA Introductory Course R

U
N
Oasys PRIMER N
I
• Oasys PRIMER has been written as a pre-processing tool to manipulate N
keyword decks (default <name>.key). G
• It aids the user to prepare and check keyword decks prior to LS-DYNA
submission.
• Oasys PRIMER can read
NASTRAN bulk data files and
translate to LS-DYNA keyword
files.
back to contents
LS-DYNA ENVIRONMENT
23
Single precision vs. Double precision LS-DYNA Introductory Course R

U
• Single Precision N
Note: In the LS-DYNA version 971
• 32 bit storage R4.2.1 and newer, the single N
precision version stores all the nodal I
• Used for most applications of LS-DYNA history variables in double precision. N
• Double precision G
• 64 bit storage
• Mandatory for implicit
• Useful for applications involving extremely small deflections, longer event
times, or larger structures
• Useful for buckling problems which are sensitive to small imperfections
• More computationally expensive (30-40%) than single precision
• The executables for the various options for version 971R6.0 for Windows 32 bit are
listed below as an example:
• ls971_s_R6.0_win32_p.exe → Single Precision SMP version
• ls971_d_R6.0_win32_p.exe → Double Precision SMP version
• mpp971_s_R6.0_win32_p.exe → Single Precision MPP version
• mpp971_d_R6.0_win32_p.exe → Double Precision MPP version
back to contents
LS-DYNA ENVIRONMENT
24
SMP vs. MPP LS-DYNA Introductory Course R

U
The explicit and implicit LS-DYNA analyses can be run either SMP (Shared N
Memory Parallel) or MPP (Massively Parallel Processing). N
I
N
SMP – Shared Memory Parallel G
• Run the analysis on a single computer.
• Parts of the calculation are split up into parallel threads, e.g. element processing.
• All processors have access to one “large” global memory.
• Communication between processors is done internally within the machine across the
data bus.
• Not very efficient when running on more than 4 - 8 parallel threads.
MPP – Massively Parallel Processing (or Distributed Memory Parallel)

• Run on one or more computers, each with one or more cores.
• The model is decomposed into a number of sub models via a domain decomposition.
• Each Core runs a sub model as an independent process.
• Model information is passed between the processes.
• Communication is maintained between the processors using the MPI (Message
Passing Interface), both within the computer and across the network.
• More efficient than SMP, can achieve good parallelisation on 64, 128+ cores.
back to contents
LS-DYNA ENVIRONMENT
25
MPP – Domain Decomposition LS-DYNA Introductory Course R

U
N
Element Domains for 2, 4, 8, 16, 32 and 64 way MPP analyses: N
I
N
G
back to contents
LS-DYNA ENVIRONMENT
26
Submitting jobs LS-DYNA Introductory Course R

U
N
LS-DYNA can either be submitted from the command line, or using one of various N
GUIs (e.g. Oasys Shell , LS Program Manager). I
N
G
Command Line Submission:
ls-dyna_executable i={input_file}
- e.g. for model1.key:
/prg/LS-DYNA/ls-dyna.exe i=model1.key
Additional options can be specified after i={input_file}.
See the Keyword Manual → Getting Started → Execution Syntax for more details.
back to contents
LS-DYNA ENVIRONMENT
27

U
GUI Submission – Oasys Shell N
N
I
LS-DYNA N
Version G
Input
File
Optional
input/output
files
Number
of CPU
(cores)
Submit
back to contents
LS-DYNA ENVIRONMENT
28

U
GUI Submission – LS Program Manager N
N
I
N
LS-DYNA G
Version
Submit
Input
File
Number
of CPU
(cores) Optional
input/output
files
back to contents
LS-DYNA ENVIRONMENT
29
Sense Switches LS-DYNA Introductory Course R

U
N
Sense switches allow you to control the behaviour of an analysis while it is running. N
To activate one of the sense switch options a text file called jobname.kil needs to be I
created, containing one line of text, which is the sense switch you want to use: N
G
e.g. sw1. (include the dot)
Some of the options are:
sw1. – A restart file is written and LS-DYNA terminates
sw2. – LS-DYNA responds with time & cycle info
sw3. – A restart file is written and LS-DYNA continues
sw4. – A plot state is written and LS-DYNA continues
swa. – Flush ASCII file buffers
The jobname.kil file

can also be created
using the Oasys Shell:
back to contents
LS-DYNA ENVIRONMENT
30
File naming convention LS-DYNA Introductory Course O

U
T
• When submitting a job, by default the result files will be written as: P
D3HSP, D3PLOT, D3THDT, D3DUMP, RUNRSF, etc. U
T
• However, as it is difficult to track which results belong to which model with
this naming convention, a lot of users prefer instead the following naming
convention for the results files:
<name>.otf, <name>.ptf, <name>.thf, <name>.dpf, <name>.rrd
• This is automatically done if the jobs are submitted via the Oasys Shell, but
needs to be setup manually when using the LS Program Manager.
• The next slide summarizes the various files and their purpose.
back to contents
LS-DYNA ENVIRONMENT
31
LS-DYNA File Summary LS-DYNA Introductory Course O

U
T
P
Input Files
U
<name>.key keyword input file ASCII T
<name>.dpf d3dump restart file BINARY
<name>.adf runrsf restart file BINARY
Output Files
<name>.otf d3hsp analysis status ASCII
<name>.ptf d3plot plot state data BINARY
<name>.thf d3thdt time history data BINARY
binout analysis results BINARY
messag analysis status ASCII
<name>.log log analysis status ASCII
<name>.dpf d3dump restart file BINARY
<name>.adf runrsf restart file BINARY
back to contents
LS-DYNA ENVIRONMENT
32
Data Output LS-DYNA Introductory Course O

U
T
Controlling Disk Space Usage P
LS-DYNA can produce a large amount of data, so care must be taken to control U
disk space usage. T
Binary File Families

• Ease handling and manipulation, new file on reaching size limit:
e.g. model.ptf, model.ptf01, model.ptf02, etc.
• Family member size can be controlled on submission.
• Small models will write more than one state to a file.
• Allows selective deletion of results if there is too much data.
• Always keep the parent file, e.g. model.ptf.
back to contents
LS-DYNA ENVIRONMENT
33
Post-Processing LS-DYNA Introductory Course P

O
S
There are two main purposes to post-processing a model: T
• To understand the behaviour of the model P

• To extract & document the results from the simulation R
O
C
There are also various levels of post-processing: E
S
• Standard post-analysis checks S
• Detailed analysis checks I
N
• Examination of model behaviour G
back to contents
LS-DYNA ENVIRONMENT
34

O
Standard post-analysis checks S
T
• Did the job reach the termination time?
P
• Are there any errors (check .otf or messag files)? R
• Is the overall motion of the model believable ? Check the animations. O
C
• Is the total energy of the system sensible? E
• For impact cases the total energy should be reasonably constant S
S
• Check hourglass energy & Sliding interface energy I
• Are there sections of the model that should be connected but connections have N
G
been missed out?
• Do joints hold together?
• Do the contact surfaces penetrate?
• Do contact surfaces need extending?
• Do model kinematics look sensible?
back to contents
LS-DYNA ENVIRONMENT
35

O
S
Detailed post-analysis checks T
• Is the deformed shape (both overall and localized) realistic? P

• Energy breakdown of the system. R
• Which parts of the model absorb the energy? O
C
• Do any of the energy traces have „spikes‟ or „steps‟? E
• Is the mesh density sufficient to capture the deformation modes? S
S
• Is the added mass due to mass-scaling acceptable during the whole analysis?
I
• In which part maximum % of mass is added? N
• Do any of the elements suffer extreme distortion? G
• Would any of the areas/connections have failed?
back to contents
LS-DYNA ENVIRONMENT
36

O
S
Model behaviour T
• View animations for whole model and parts of model to see interactions in more P
detail. R
• Look at stresses – von Mises stress typical for metals. O
C
• Look at principal stresses, both magnitude and directions to examine load paths.
E
• Look at plastic strains and their development over the course of the analysis. S
• Use cut sections through the model to gain better understanding of behaviour S
within the model. I
N
G
back to contents
LS-DYNA ENVIRONMENT
37
Data Evaluation - Model Post-Processing LS-DYNA Introductory Course P

O
S
Post processing software: T/HIS & D3PLOT used in this course T
P
R
O
C
E
S
S
I
N
G
back to contents
LS-DYNA ENVIRONMENT
38
Workshop Example 1 LS-DYNA Introductory Course
1. Edit the keyword file „example1.key‟ (using a text editor e.g. Wordpad, GVim etc.).
2. Open up the online pdf LS-DYNA keyword manuals (Vol I & II) for reference.
3. Create 4 nodes which will represent the corners of a quad shell (*NODE) .
4. Create a simple elastic material for the shell (*MAT_ELASTIC).
5. Create a section with the shell section properties (*SECTION_SHELL).
6. Create a part for the shell - this links the material and section together.
7. Create a quad shell with these 4 nodes (*ELEMENT_SHELL) and put it into the part that
was created in step 6.
8. Submit the job and check there are no errors.
*PART *ELEMENT_SHELL *NODE (coordinates)
*NODE
*NODE
*MAT_ELASTIC
(Material properties) *NODE
*SECTION_SHELL
(Thickness, formulation)
back to contents
LS-DYNA ENVIRONMENT
39
LS-DYNA Introductory Course T

H
E
O
R
Y
&
THEORY AND PRACTICE P

R
A
C
T
I
C
E
LS-DYNA ENVIRONMENT
40
• BACKGROUND • THEORY AND PRACTICE

• Time Integration
• Element Formulation
• INTRODUCTION • Hourglassing
• Key principles
• Materials
• Running
• Contact Surfaces
• Output files
• Rigids
• Post Processing
• Boundary Conditions
• Connections
• Control & Output
• Other
• ANALYSIS
• Checking
back to contents
LS-DYNA ENVIRONMENT
41
Outline – THEORY AND PRACTICE LS-DYNA Introductory Course T

H
E
1. Time Integration 4. Materials O
• Description of explicit time 5. Contact Surfaces R
integration Y
• Types of contact
• Description of implicit time
integration • Contact method (depth, stiffness) &
• Time-step calculation • Initial penetrations

P
• Summary of cost considerations • Other contact issues R
• Contact guidelines A
2. Element Formulation C
6. Rigid Bodies T
• Description of element types
I
– Solid – Beam 7. Boundary Conditions
C
– Shell – Discrete 8. Connections E
– Thick Shell – Other 9. Control and Output
• Meshing Notes
10. Other Techniques
3. Hourglassing
back to contents
LS-DYNA ENVIRONMENT
42
Outline – THEORY AND PRACTICE LS-DYNA Introductory Course T

I
M
1. Time Integration 4. Materials E
• Description of explicit time 5. Contact Surfaces
integration I
• Types of contact N
integration • Contact method (depth, stiffness) T
E
G
• Contact guidelines A
2. Element Formulation T
6. Rigid Bodies I
O
N
– Shell – Discrete 8. Connections
• Meshing Notes
3. Hourglassing
back to contents
LS-DYNA ENVIRONMENT
43
Time Integration LS-DYNA Introductory Course T

I
M
Explicit Time Integration: Acceleration assumed to be constant over timestep.
E
I
N
Equation of motion for mass: T
Fo = applied force, dt = timestep, a=acceleration, x=displacement, v=velocity E
G
F0 - k xt = m at Know all values at time=t, want to solve for t+dt R
F0 - k xt+dt = m at+dt A
T
I
xt+dt = xt + vt+dt/2 dt Central Difference Method O
vt+dt/2 = vt-dt/2 + at dt N
m at+dt = F0 - k ( xt + vt-dt/2 dt + at dt2 )
Solving a multi-degree of freedom system requires inversion of the mass matrix which is
trivial.
back to contents
LS-DYNA ENVIRONMENT
44

I
M
Implicit Time Integration: Acceleration assumed to vary linearly over timestep.
E
I
N
Equation of motion for mass: T
Fo = applied force, dt = timestep, a=acceleration, x=displacement, v=velocity E
G
F0 - k xt = m at Know all values at time=t, want to solve for t+dt R
F0 - k xt+dt = m at+dt A
T
I
xt+dt = xt + 0.5 ( vt + vt+dt ) dt Average values used. O
vt+dt = vt + 0.5 ( at + at+dt ) dt Extra „at+dt‟ term to deal with…. N
m at+dt = F0 - k (xt + dt ( vt + 0.25 ( at + at+dt ) dt))

(m + 0.25 k dt2)at+dt = F0 - k (xt + vt dt + 0.25 at dt2 )
Solving a multi-degree of freedom system requires inversion of the STIFFNESS matrix which
is computationally expensive.
back to contents
LS-DYNA ENVIRONMENT
45

I
M
Explicit Time Integration E
In LS-DYNA we solve for dynamic equilibrium, as follows:
I
N
[M] a + [C] v + [K] x - F = 0 T
E
G
F are the applied loads R
[K] is the stiffness matrix A
[C] is the damping matrix T
[M] is a mass matrix I
O
x are the nodal deflections N
v are the nodal velocities
a are the nodal accelerations
back to contents
LS-DYNA ENVIRONMENT
46

I
M
Considering the usual case where damping is negligible:
I
[M] a + [K] x - F = 0 or: [M] a = F - [K] x N
T
We define the Internal Forces I as: E
G
I = [K] x so that: [M] a = F - I R
A
This set of equations is solved at the beginning of a time step. The mass matrix is known T
(from the elements), and the applied forces are also known (defined in the model). I
Only the internal forces need to be determined in order to solve for a, and these are O
calculated based on the displacements from the previous time step. N
We can then quickly solve for the accelerations by inverting the diagonal mass matrix:
a = [M]-1 ( F - I )
Once the accelerations are known, the central difference method can be used to determine
the new velocities and the nodal displacements for this time step. The new stresses and
strains can then be calculated from these nodal displacements.
back to contents
LS-DYNA ENVIRONMENT
47
Time Integration – Process Loop LS-DYNA Introductory Course T

I
Start M
E
Write databases Apply force I

and update and other N
geometry boundary T
conditions E
G
R
A
T
Calculate nodal Process elements I
vt+dt/2 velocities and and calculate Material O
xt+dt displacements out-of-balance Model N
forces
F = ma
Use out-of-balance
nodal forces to
calculate nodal
accelerations
back to contents
LS-DYNA ENVIRONMENT
48
Time Integration – Timestep LS-DYNA Introductory Course T

I
Explicit Time Integration M
• The central difference method does not require any iterations, and hence it does not E
have to converge.
I
• However, it is only conditionally stable, i.e. the time step must be kept small enough, or
the solution will not have any physical meaning. N
• The time step selection is governed by the Courant Condition. T
E
G
Hammer Ideal Elastic bar
Speed of acoustic wave: R
Pressure
Pressure
A
l E T
Time Time c 
t  I
(Time for stress
O
 wave to cover N
l t l
E one element)
 l
For an accurate solution → t  l   where ∆t = time step for numerical time integration.
E c
 where K = TSSFAC on *CONTROL_ TIMESTEP
LS-DYNA will set → t  K  l 
E card, by default = 0.9)
back to contents
LS-DYNA ENVIRONMENT
49
Time Integration – Timestep LS-DYNA Introductory Course T

I
M
The Courant Condition limits the time step to be less than the time taken for a
I
sound wave to traverse the smallest element in the model (excluding rigid
N
bodies). In this way, we guarantee that a stress wave which travels through an T
element will cause stresses to develop in the element, as it would occur in E
reality. If the time step were too large, the motion of the shock wave could G
"jump" across an element undetected. (Note: the schematics are for illustration R
only) A
Stress (pressure) wave propagation down a beam T
I
Timestep > Courant Condition O
N
F
Step 1 Step 2 Step 3 Step 4 Step 5
Timestep < Courant Condition
Time
Step 1 Step 2 Step 3 Step 4 Step 5
back to contents
LS-DYNA ENVIRONMENT
50
Timestep calculation & Characteristic length LS-DYNA Introductory Course T

I
LS-DYNA calculates the timestep for each individual element in the model and M
E
chooses the lowest as the analysis timestep. The idea is that if the stability is
guaranteed to the element with the lowest timestep, all others are stable. I
N
T
Beam elements: E
G
R
A
Shell elements: T
I
O
N
Solid elements:
Note: For discrete elements,

Spring elements: the length of the element does
not matter.
back to contents
LS-DYNA ENVIRONMENT
51
Explicit vs. Implicit LS-DYNA Introductory Course T

I
Explicit Implicit M
E
Small time step Large time step
Inversion of diagonal mass matrix only Inversion of mass & stiffness matrix I
N
Dynamic and Quasi-static problems Static and Dynamic problems T
Low memory High memory + High I/O E
G
Stability - controls the timestep used Accuracy – controls the timestep used R
A
Cost per time step – small Cost per timestep - large T
Element processing – controls the cost Matrix inversion – controls the cost I
O
Simple element formulation Complex element formulation N
Explicit code CPU time
Element Process. Matrix inversion Time integration Others
Implicit code CPU time

back to contents
LS-DYNA ENVIRONMENT
52
Workshop Example 2 LS-DYNA Introductory Course T

I
M
1. Submit the job „example2.key‟.
E
2. Open the output file example2.otf (whilst it is running).
3. Find the timestep controlling element by searching for “smallest” I
and note the element id and the timestep DYNA gives. N
4. Create a new folder, copy example2.key into example2_new.key T
and edit this file. E
5. Find the part then material that the timestep controlling element G
references. R
6. Change the material Young‟s modulus to 2000, and reduce the A
density by a factor of 10. T
I
7. Try to predict (by hand) the new timestep that DYNA will use. You
O
will need to use the current values of E and ρ to help you calculate
N
the new timestep.
8. Submit this new job and see if the prediction is correct.
9. Have a look at the results of either run using T/HIS and D3PLOT.
Answer: Timestep proportional to √(density/modulus)

Change is proportional to √(0.1/0.025) = √(4)
New timestep = Old timestep x 2.0 Units: mm, s, t, N, MPa
back to contents
LS-DYNA ENVIRONMENT
53
Outline – THEORY AND PRACTICE LS-DYNA Introductory Course E

L
E
1. Time Integration 4. Materials M
• Description of explicit time 5. Contact Surfaces E
integration N
• Types of contact T
integration • Contact method (depth, stiffness)
F
O
• Contact guidelines M
2. Element Formulation U
6. Rigid Bodies L
A
T
– Shell – Discrete 8. Connections I
– Thick Shell – Other O
9. Control and Output N
• Meshing Notes
3. Hourglassing
back to contents
LS-DYNA ENVIRONMENT
54
Elements LS-DYNA Introductory Course E

L
• Elements in LS-Dyna are linear and generally have E
simple formulations. Shells M
E
• Accelerations, velocities and displacements are N
evaluated at the nodes. T
Solids
F
• Stresses and strains are calculated at the O
integration points. R
Beams Trusses M
• Element quality is not as critical as for some implicit U
codes (e.g. Nastran) but highly deformed elements L
Springs Lumped Dampers A
are less accurate. mass
T
I
• Some element formulations are more costly than N4 N3
acc4 vel4 acc3 vel3 O
others; there is a trade-off between accuracy and disp4
σ, ε
disp3 N
CPU. Use higher cost elements selectively.
N1
• Single point integration is the default for solids, acc1 vel1
acc2 vel2
N2 disp2
shells and beams. disp1
back to contents
LS-DYNA ENVIRONMENT
55
Elements LS-DYNA Introductory Course E

L
E
Element size is a balance between capturing geometry vs. run time (number of M
elements and timestep). E
Remember: N
T
The *ELEMENT_ ... keyword references
• Parts F
O
• Nodes R
M
U
The *SECTION_ ... keyword references L
A
• Element type (formulation)
T
• Number of integration points I
• Thickness (for shell elements) O
N
back to contents
LS-DYNA ENVIRONMENT
56
Elements – Element formulation LS-DYNA Introductory Course E

L
E
• LS-DYNA provides the following controls to user with regard to element M
properties selection: E
• Element formulation (or element type) N
T
• Number of integration points – both in the plane and through thickness
• Choice of integration rule (the user can also define their own integration F
rule). O
R
M
U
L
A
T
I
O
N
back to contents
LS-DYNA ENVIRONMENT
57
Elements – Element formulation LS-DYNA Introductory Course E

L
E
• LS-DYNA elements use various M
EQ.0: 1 point corotational for *MAT_MODIFIED_HONEYCOMB. See remark 3.
mathematical formulations for EQ.1: constant stress solid element (default), E
calculating the stress and strain EQ.2: fully integrated S/R solid. See remark 4 below,
N
EQ.3: fully integrated quadratic 8 node element with nodal rotations,
in the element; the choice of EQ.4: S/R quadratic tetrahedron element with nodal rotations, T
element type on the *SECTION EQ.5: 1 point ALE,
EQ.6: 1 point Eulerian,
card is based on the specifics of EQ.7: 1 point Eulerian ambient,
F
the model. EQ.8: acoustic, O
EQ.9: 1 point corotational for *MAT_MODIFIED_HONEYCOMB. See remark 3. R
• For example, the available EQ.10: 1 point tetrahedron.
M
EQ.11: 1 point ALE multi-material element
element types for EQ.12: 1 point integration with single material and void. U
*SECTION_SOLID are: EQ.13: 1 point nodal pressure tetrahedron. See remark 14 below. L
EQ.14: 8 point acoustic
A
• Standard (Lagrange) EQ.15: 2 point pentahedron element.
T
EQ.16: 4 or 5 point 10-noded tetrahedron (See Remark 13)
• Eulerian (ALE, fluid)
• Acoustic
EQ.17: 10-noded composite tetrahedron I
EQ.18: 8 point enhanced strain solid element for linear statics only
• Tet element EQ.19: 4 point cohesive element
O
• Penta element EQ.20: 4 point cohesive elements with offsets for use with shells N
• Linear Static (Implicit) EQ.41: Mesh-free solid formulation
• Cohesive element EQ.99: simplified linear element for time-domain vibration studies. See remarks.
GT.100.and.LT.106: User defined solid
• Other
back to contents
LS-DYNA ENVIRONMENT
58
Solid Elements LS-DYNA Introductory Course E

L
E
*ELEMENT _SOLID, *SECTION_SOLID M
E
• 8 node brick (hex) elements N
Wedge T
• 6 node wedge (penta) elements Brick
• 4 node tetrahedral (tet) elements F
O
R
• 3 d.o.f.s at each node (in general) M
Tetrahedron
U
• Generally just take the defaults, except for a tetrahedral mesh, when L
A
element Type 10 is recommended, since they are more stable than the
T
degenerate solid elements (by default a tet is created through collapsing I
nodes on a brick element, thus obtaining a degenerate element). O
N
• Alternatively set ESORT=1 on the *CONTROL_SOLID card and this will
automatically assign Type 10 for all tet elements and Type 15 for all penta
elements in the model.
back to contents
LS-DYNA ENVIRONMENT
59
Solid Elements LS-DYNA Introductory Course E

L
E
Constant Stress Solid (Type 1 – default) M
• Single integration point at the centre. E
• No bending stiffness (need multiple elements through N
T
thickness).
F
Fully Integrated S/R Solid (Type 2) O
R
• 8 integration points (one for each node).
M
• Does have bending stiffness, so one element through U
thickness can be used, however multiple elements will L
still be more accurate. A
T
• Does not suffer from hourglassing problem (discussed
I
later). O
N
1 Point Tetrahedron (Type 10)
• Special element formulation for use with tet elements
(more accurate mass distribution).
back to contents
LS-DYNA ENVIRONMENT
60
Shell Elements LS-DYNA Introductory Course E

L
E
*ELEMENT_SHELL , *SECTION_SHELL M
E
• 4 noded quad shell elements and 3 noded N
T
tria shell elements.
Quadrilateral Triangular F
• 5 d.o.f‟s at each node shell shell O
R
(no drilling resistance). M
U
• Use of trias is necessary because of meshing requirements, but you should L
A
aim for less than 5% of total number of shells as they are over stiff.
T
I
• Setting ESORT=1 on the *CONTROL_SHELL card is recommended, as O
this will automatically assign the improved element formulation (Type 4) to N
any tria elements in the model (default will use degenerate quad elements).
back to contents
LS-DYNA ENVIRONMENT
61

L
Belytshcko-Tsay (Type 2 – default) E
M
• Single integration location in plane.
E
• Default two through-thickness integration points (can change N
with NIP option *SECTION_SHELL). T
• If a single layer of shells is used for bending, then several
integration points needed. F
• Cheapest and the most robust shell element in LS-Dyna. O
R
M
Fully Integrated Shell (Type 16)
U
• 4 integration locations in plane. L
• Default two through-thickness integration points at each A
location. T
• More accurate if the mesh becomes badly distorted. I
• Does not suffer from hourglassing problem (discussed later). O
N
C0 Triangular Shell (Type 4)

• Special element formulation for use with tria elements (more
accurate than the degenerate elements).
back to contents
LS-DYNA ENVIRONMENT
62

L
E
Shell elements type 2 vs. type 16 – general guidelines for typical M
crush structures: E
N
• For large element sizes (> approx. 5mm), shell type 2 should be T
used, as it gives more accurate results. Type 16 can give an
F
overstiff response for this case.
O
• For small element sizes (<approx. 5mm), shell type 16 is R
recommended. M
U
L
A
T
I
O
N
back to contents
LS-DYNA ENVIRONMENT
63

L
E
• Each shell element is designated a local coordinate system, which moves and M
deforms with the element. E
• The direction of the shell normal is important for example when applying a N
pressure load, or when looking at results, such as stress/strain reported at the T
middle, top and bottom layers for shells with several through-thickness
F
integration points (more info in section *DATABASE_EXTENT_BINARY).
O
Z
R
Y
Node4 Node4
Y
M
Node3 U
Node3 L
A
X T
Z X I
O
Node1 Node1 N
Node2 Node2
• Z-axis set by cross-product of diagonals
• X-axis set by Node1 → Node2 direction
View from above
• Y-axis set by cross product of X and Z
back to contents
LS-DYNA ENVIRONMENT
64
Thick Shell Elements LS-DYNA Introductory Course E

L
E
*ELEMENT_TSHELL, *SECTION_TSHELL t
M
• Plane stress 8 node thick shell elements (or s E
brick shells). n8 n7 N
T
• 3 d.o.f.‟s at each node. n4 n3
• One Point Reduced Integration (Type 1 – F
r
default). n5
n6 O
• Single integration location in plane. R
n1 n2 M
• Two through thickness integration points by U
default. L
• Care must be taken when meshing thick A
t-axis
shells, as node order implies the direction of T
shell thickness. I
O
N
back to contents
LS-DYNA ENVIRONMENT
65
Beam Elements LS-DYNA Introductory Course E

L
E
*ELEMENT_BEAM, *SECTION_BEAM M
• 2 node beam elements, mass is lumped at nodes, 6 d.o.f.‟s at each node. E
• 3rd node used to define orientation of the section (avoid using a structural node). N
T
• Solid or tubular circular or rectangular sections can be defined, or user specified
cross sectional properties used. F
O
n3 s Parameter CST in *SECTION_BEAM defines R
r
rectangular, circular or arbitrary cross section M
type U
n2 s s s
L
t A
t t
t T
n1 I
O
N
• A multitude of pre-determined section types can also be specified using the
STYPE option on *SECTION_BEAM card, or using the OR/IRID option to
reference a *INTEGRATION_BEAM card depending on beam element type.
back to contents
LS-DYNA ENVIRONMENT
66

L
E
Hughes-Liu Integration Beam (Type 1 – default): M
• Single integration location along the length of the beam. E
• Default 2x2 grid of integration points (QR/IRID option *SECTION_BEAM) N
T
QR/IRID=2 (default for integrated beams) QR/IRID=1 F

(2x2 Gauss quadrature) (one integration point) O
R
M
U
L
• QR/IRID can also be used to reference the *INTEGRATION_BEAM keyword, A
where more complicated cross sections profiles can be specified: T
I
e.g. I-Section, Channel, T-Section, etc O
N
back to contents
LS-DYNA ENVIRONMENT
67

L
E
Hughes-Liu Integration Beam (Type 1 – default): M
• The bending moment is constant along the length due to single integration E
point. N
T
• So, in order to model a cantilever, several elements are required along the
length for accuracy: F
F O
R
M
U
L
One Element Three Elements A
M M
T
I
O
N
d d
back to contents
LS-DYNA ENVIRONMENT
68

L
E
Belytschko-Schwer Resultant Beam (Type 2): M
• Uses an explicit linear beam formulation with plastic conditions introduced at E
the ends (F=δAE/L). N
T
• Area (A) and 2 Moment of Area (Iss, Itt) are defined on the
nd
*SECTION_BEAM card, or the STYPE option can be used to specify a F

particular cross section using a set of dimensions. O
• Values are calculated at the nodes. R
• The bending moment varies linearly along the length. M
U
• So in order to model a cantilever only one element is required along the L
length for accuracy (small deflection only). However typically several A
elements are used to capture the deformed shape along the whole length. T
M
I
O
F N
d
back to contents
LS-DYNA ENVIRONMENT
69

L
E
Discrete beams (Type 6): M
*ELEMENT_BEAM, *SECTION_BEAM E
• Simple force deflection characteristic (linear, non-linear, elastic-plastic) N
T
depending on material selected.
• Use materials *MAT_066 → *MAT_071 in conjunction with the discrete beam. F
• CID and SCOOR options on the *SECTION_BEAM card are used to define O
element local axis system and how it changes during the analysis. R
M
• More details in later section „Connections → Discrete Beams‟. U
L
A
T
I
O
N
back to contents
LS-DYNA ENVIRONMENT
70
Discrete Elements (Springs, Dampers) LS-DYNA Introductory Course E

L
E
*ELEMENT_DISCRETE, *SECTION_DISCRETE M
• 2 node elements apply force to end nodes depending on user specified E
characteristics; can also define between one node and ground (N2 = 0). N
T
• The elements have no mass and therefore can cause timestep problems (extra
*ELEMENT_MASS may need to be added to the nodes). F
• Translational or rotational (angles in radians) specified using the DRO option on O
*SECTION_DISCRETE card. R
• Linear, non-linear and general springs/dampers can be defined depending on M
U
the material chosen.
L
e.g. *MAT_SPRING_ELASTIC, *MAT_SPRING_ELASTOPLASTIC, A
*MAT_DAMPER_VISCOUS, etc. T
• Non-linear properties input as curves of force vs displacement. I
O
• More details about spring and damper materials in later section „Connections → N
Discrete Elements‟.
back to contents
LS-DYNA ENVIRONMENT
71
Discrete Elements (Springs, Dampers) LS-DYNA Introductory Course E

L
• By default, discrete elements act in the direction between nodes (and E
“displacement” means change of length). M
E
• Optionally, an orientation vector can be defined for each discrete element using N
*DEFINE_SD_ORIENTATION so that discrete element acts along vector – need T
to consider carefully if vector definition should vary during the event.
F
• An orientation vector must be defined for rotational and zero-length elements.
O
• Define using card *ELEMENT_DISCRETE: R
N2
M
z
VID
U
y L
N1 A
x
T
I
• VID orientation option is used to reference a *DEFINE_SD_ORIENTATION card O
for the vector. N
• IOP option on *DEFINE_SD_ORIENTATION sets the orientation:
• along or normal to a vector = fixed during the analysis
• along or normal to vector defined using two nodes = varies with the motion of the nodes
back to contents
LS-DYNA ENVIRONMENT
72
Other Elements LS-DYNA Introductory Course E

L
E
Lumped Mass *ELEMENT_MASS M
• Adds user specified mass to a node, for example for a discrete element E
(picture below). N
T
• Can also be used to add structural mass to components.
F
Mass element – added at the free
node of discrete element O
R
M
Discrete element – no mass
U
L
A
Shell elements – T
mass at nodes based I
on density & volume
O
N
Inertial *ELEMENT_INERTIA
• Defines a lumped inertia element assigned to a nodal point.
back to contents
LS-DYNA ENVIRONMENT
73
Element Meshing LS-DYNA Introductory Course E

L
E
• Typical size 5-20mm in automotive. M
• Need enough elements to capture expected deformation mode: E
• min 3-4 on buckling half-wavelength. N
T
• min 6-8 elements across a crushing section.
• 3 or more elements through a bending section. F
• 2 or 3 element wide flanges in critical areas. O
R
M
U
L
A
T
I
O
N
back to contents
LS-DYNA ENVIRONMENT
74

L
E
• Mesh sensitivity in buckling should be tested. M
E
N
Buckling Load
Optimum
T
element
size
F
O
R
M
U
L
Mesh Density A
T
• Elements should be as square as possible in critical areas. I
O
• Generally, model radii, holes, joggles and other features only if bigger than N
pre-selected minimum element size and in an area likely to deform.
• Always model crush initiators.
back to contents
LS-DYNA ENVIRONMENT
75

L
E
• Most companies have internal mesh quality requirements for models e.g.: M
– Minimum length 5 mm E
– Maximum warpage 20 deg N
T
– Maximum aspect ratio 5:1
– Maximum skew 60 deg F
– Minimum tria internal angle 30 deg O
R
– Maximum tria internal angle 120 deg M
– Minimum quad internal angle 40 deg U
– Maximum quad internal angle 140 deg L
A
T
• Check and fix in Pre-processor. I
O
• Oasys PRIMER can check mesh quality.
N
back to contents
LS-DYNA ENVIRONMENT
76
Workshop Example 3 LS-DYNA Introductory Course E

L
E
1. Animate the results of „crush_tube‟. M
2. Compare the crushed shape of each tube (use cut-sections). E
3. Compare crush force for each tube (X Sections). N
T
4. Compare the amount of absorbed energy by each tube (Parts → IE Energy).
F
O
R
M
U
L
A
Tube 3 T
I
O
Tube 2 N
Tube 1
Units: mm, s, t, N, MPa
back to contents
LS-DYNA ENVIRONMENT
77
Outline – THEORY AND PRACTICE LS-DYNA Introductory Course H

O
U
1. Time Integration 4. Materials R
• Description of explicit time 5. Contact Surfaces G
integration L
• Types of contact A
integration • Contact method (depth, stiffness) S
S
I
• Summary of cost considerations • Other contact issues N
• Contact guidelines G
2. Element Formulation
6. Rigid Bodies
• Meshing Notes
3. Hourglassing
back to contents
LS-DYNA ENVIRONMENT
78
Hourglassing LS-DYNA Introductory Course H

O
• Hourglassing (zero energy deformation modes) is most commonly a problem in single U
integration point solid and shell elements (should typically be <5% of total energy). R
• For example, for a solid: G
L
– 3 d.o.f.‟s at each node = 24 d.o.f.‟s for an 8 node A
solid element. S
S
– single integration point generates 6 stresses (3
I
direct, 3 shear) + 6 rigid body modes.
N
– Therefore 12 modes of deformation have no G
stiffness and have to be suppressed.
• As the mesh gets distorted the element accuracy is
reduced and the timestep gets smaller.
Hourglass modes for

y y
underintegrated solid
elements
x x
Under integrated shell element Fully integrated shell element

back to contents
LS-DYNA ENVIRONMENT
79

O
U
Some tips for avoiding hourglassing: R
• Avoid subjecting single point integration G
solid elements to bending. L
A
• Avoid point loads - use pressures instead.
S
• Avoid single node or line contacts. S
• Use fully integrated elements. I
N
• Monitor the hourglass energy in your result
G
files and make sure it is low compared with
the internal energy (<5%).
• Re-mesh with a finer mesh if possible.
Bad Good
back to contents
LS-DYNA ENVIRONMENT
80

O
If the hourglassing cannot be avoided with the methods suggested previously, it can be U
somewhat controlled using the *HOURGLASS card. R
The hourglass control can be applied at part level *HOURGLASS card: G
L
• For solids:
A
• IHQ =1,2,3 = viscous hourglassing S
• IHQ = 4,5,6 = stiffness hourglassing S
• For shells the hourglass control is based on the formulation of Belytschko and Tsay, i.e., I
options 1-3 are identical, and options 4-5 are identical: N
• IHQ = 1,2,3 = viscous hourglassing G
• IHQ = 4,5 = stiffness hourglassing
• Shells type #16 should use IHQ =8 for warping stiffness; for shells type #2 it is
recommended IHQ=4, QM=0.05
• The hourglass coefficient QM = 0.1 by default.
• Viscous hourglass control – acts to resist further hourglass deformation, recommended for
high velocity deformations.
• Stiffness control – acts to restore node to the original position, recommended for lower
velocities.
• Also consider *CONTROL_BULK_VISCOSITY
back to contents
LS-DYNA ENVIRONMENT
81
Workshop Example 4 LS-DYNA Introductory Course H

O
U
In this example, an incomplete input deck is given,
R
block0.key. The task steps are: G
L
1. Complete the model, following the extra A
instructions given in „Example4_instructions.docx‟: S
start with „block0.key‟ and set up the material, S
section and element type, boundary constraints I
and loading; save the updated model. N
G
2. Submit the analysis and view results - your model
should give similar results to the model in folder
„BLOCK1‟.
3. Modify the model to eliminate hourglass.
4. Review the models in „BLOCK_2‟ and „BLOCK_3‟
HINT: in Primer, use DISPLAY-ENTITIES (or press „e‟) to

make constraints visible. Units: mm, s, t, N, MPa
back to contents
LS-DYNA ENVIRONMENT
82
Outline – THEORY AND PRACTICE LS-DYNA Introductory Course M

A
T
integration I
• Types of contact A
integration • Contact method (depth, stiffness) L
S
• Summary of cost considerations • Other contact issues
• Contact guidelines
6. Rigid Bodies
• Meshing Notes
3. Hourglassing
back to contents
LS-DYNA ENVIRONMENT
83
Materials LS-DYNA Introductory Course M

A
T
• There are over 200 materials available E
Key to numbers in the square brackets:
in LS-DYNA. R
I
A
• Materials usually have both a numerical
L
and a descriptive equivalent name that S
can be used.
e.g. *MAT_001 = *MAT_ELASTIC
• The Keyword manual lists, in the first

pages in the material section, all the
materials with information about the Key to numbers in the curly brackets:
element formulations available for the
respective material and the number of {n} = the default number of history variables per
default history variables. element integration point that are stored in
addition to the 7 history variables stored by
e.g. *MAT_007: *MAT_BLATZ- default.
KO_RUBBER [0,2,8B]{9}
back to contents
LS-DYNA ENVIRONMENT
84

A
T
The Keyword manual also includes a table with all the materials, some of their E
attributes and the general classes of physical material to which the numerical R
models might be applied: I
A
L
S
For example, *MAT_003 includes Strain-rate effects, has Failure option and
can be applied to Composites, Metals and Plastic materials.
back to contents
LS-DYNA ENVIRONMENT
85

A
General concepts used in definition of materials: T
E
• Define a curve: R
I
• This is needed for example to input material data, load curves, etc.
A
• Use *DEFINE_CURVE card. L
• For constitutive material models the curves are extrapolated if the values on the S
abscissa go off scale.
Material Curve Note: For load curves, LS-
DYNA does not extrapolate
σ the curve past the last input
value on the x-axis.
ε
• LS-DYNA divides the curve into 100 (default) equally spaced points, this can be
changed on *CONTROL_SOLUTION, parameter LCINT.
• Define a table of curves:
• Needed for example to define a strain rate dependency, see slides on strain rate.
• Use *DEFINE_TABLE card.
back to contents
LS-DYNA ENVIRONMENT
86

A
General concepts used in definition of materials: T
E
R
• Failure (see later slides for more detail): I
• Material failure – for example based on plastic strain or time step size. A
• If the material has the failure option then the element will be deleted when L
S
the failure value is reached.
• If the material does not allow for failure then *MAT_ADD_EROSION can
be used in conjunction with your material.
• Strain Rate Sensitivity (see later slides for more detail):
• Some materials have different properties depending on the speed they
are loaded at (e.g. steel, plastics).
• Certain material models include options to model this.
• Also referred to as strain rate effect or strain rate stiffening.
• Orthotropic and Anisotropic Materials (see later slides for more detail):
• Some materials allow the user to specify different properties in different
directions (e.g. composites, wood etc).
back to contents
LS-DYNA ENVIRONMENT
87
Materials – Modelling Failure LS-DYNA Introductory Course M

A
T
• Elements are deleted when the failure criterion is reached. E
• Contacts on the failed elements are also eliminated. R
• Failure can be based on a range of criteria depending on the material model, two I
A
of the main ones are:
L
• Plastic strain (based on material data) S
• Time step reduction (for model stability)
• When using tensile test data to estimate failure strain, the reduction in cross
sectional area has to be taken into account.
True failure strain

Ao
area
= loge ( A failed )
Afailed
back to contents
LS-DYNA ENVIRONMENT
88

A
T
• The mode of failure can be important - failure due to bending strains can be E
very different to failure by axial strains. R
• For some materials the parameter NUMINT can be set for the number of I
through-thickness integration points which must fail before the element is A
deleted. L
S
back to contents
LS-DYNA ENVIRONMENT
89

A
T
The element size has an influence on the load at which failure starts to occur. E
• For a given plastic strain at failure, the smaller the element, the lower the R
failure load. (local stress concentration and gauge length) I
A
• So, to achieve failure at a particular load, the plastic strain at failure from the
L
test has to be adjusted to take into account the elements used in the model. S
• This is typically done by modelling the tensile test and changing the failure
strain until the failure load matches the test.
Results for three tension samples with different
element sizes but the same failure strain
Total Load
X Stress
back to contents
LS-DYNA ENVIRONMENT
90
Materials – Strain Rate Sensitivity LS-DYNA Introductory Course M

A
T
A large number of engineering materials exhibit strain rate sensitivity, where the stress
E
strain characteristics of the material are dependent on the rate they are loaded at.
R
I
 Fast Loading A
(high strain rate)
L
S
Slow Loading
(low strain rate)

Strain is calculated as: Strain rate is:
L d
   [units: 1/s]
L0 dt
F
L0 L
back to contents
LS-DYNA ENVIRONMENT
91

A
T
• A number of material models in LS-DYNA can include strain rate sensitivity (e.g. E
Mat_24). R
I
• The next few slides cover the main three methods available for modelling strain A
L
rate sensitivity: S
1. Curve defining the strain rate scaling effect on the yield stress.
2. Input actual stress-strain curves at various strain rates via a *DEFINE_TABLE.
3. Using Cowper and Symonds model.
• In combination with any of the three methods above, a fully Viscoplastic

formulation is also available.
back to contents
LS-DYNA ENVIRONMENT
92

A
T
1. Curve defining the strain rate scaling effect on the yield stress.
E
This method uses a curve to define a factor to be applied to the basic stress R
depending on the strain rate. I
A
L
• Input a curve (LCSR) which sets the S
scaling factor for the yield stress vs.
strain rate:
Effective stress
Scale factor
Effective plastic strain

Strain rate (log10)
back to contents
LS-DYNA ENVIRONMENT
93

A
T
2. Input actual stress-strain curves at various strain rates via a *DEFINE_TABLE.
E
If different stress-strain curves for a range of strain rates are available, these can be input R
using *DEFINE_CURVE and then referenced using a *DEFINE_TABLE. I
• The *DEFINE_TABLE card contains a list *DEFINE_TABLE_TITLE
A
of the strains rates associated with each My_Table L
55021
curve. $ S
• The *DEFINE_CURVE cards then follow $ Table 55021 references 3 load-curves.
$ List of factors:
the table in the order they are listed. $
1.00000E-3
• All the curves should finish at the same x 1.00000000
100.000000
value for efficiency. $
$ Table 55021 load-curves follow.
$
*DEFINE_CURVE
55027 0 1.0 1.0 0.0 0.0 0
0.0 16.505850
9.2266803E-4 35.299141
....
*DEFINE_CURVE
55028 0 1.0 1.0 0.0 0.0 0
0.0 17.337440
8.4261497E-4 36.563351
...
*DEFINE_CURVE
55030 0 1.0 1.0 0.0 0.0 0
Note: Also see *DEFINE_TABLE_2D, which 0.0 19.279110
allows for each value in the table to specify a 5.9243903E-4 36.797409
...
curve – from version 971 R5.1.1
back to contents
LS-DYNA ENVIRONMENT
94

A
T
3. Cowper and Symonds model for strain rate
E
This is a power law function which scales the stress according to the following equation: R
where:  = strain rate
I
Stress Factor  1  ( / C )
1/ p
C = material constant A
Typically for Mild steel: C = 1300, p = 5 p = material constant L
S
Input material curve Resulting curves for some
different strain rates
Cowper-Symonds can give

insufficient scaling at low strains
back to contents
LS-DYNA ENVIRONMENT
95

A
Viscoplastic formulation: T
E
• A fully viscoplastic formulation can also be used in combination with any of the R
three different methods (set using the VP option on the material card). I
• This formulation can help to reduce the noisy response that sometimes occurs A
when modelling strain rate sensitivity. L
S
• Recommend use VP = 1 → Visco-plastic formulation – uses the second
invariant of equivalent plastic strain rate, consequently no oscillations in the
stress that is computed, although it will add some CPU cost.
Example of *MAT_024 with strain rate of 0.1(1/s)
No strain rate treatment;
Cowper and Symonds VP=0
Cowper and Symonds VP=1
back to contents
LS-DYNA ENVIRONMENT
96
Materials – Orthotropic and Anisotropic Materials LS-DYNA Introductory Course M

A
T
There are a number of orthotropic and anisotropic materials in LS-DYNA that allow the user
E
to specify different properties for each of the local material axis (typically referred to as a,b,c
R
instead of x,y,z).
I
A
The AOPT option on the material card allows the user to control how the material axis is L
calculated in a number of ways. The main ones are: S
AOPT = 0: Based on the element local axis including BETA angle offset.
AOPT < 0: The value given is the id of a *DEFINE_COORDINATE_SYSTEM or

*DEFINE_COORDINATE_VECTOR to be used for the material local axis.
back to contents
LS-DYNA ENVIRONMENT
97

A
Y-element T
• For each shell element the *ELEMENT_SHELL_BETA
E
option allows an offset (beta angle) to be specified between
R
the local element axis and the local material axis system.
I
• Most pre-processors can automatically set the beta angle X-material A
for shells elements so that the material axes for those L
elements all line up with a given vector. BETA
S
X-element
Shell Element Local X-Axis Shell Element Material X-Axis

BETA angle set to align with global x axis
(cosine vector 1,0,0 in PRIMER & AOPT=0)
back to contents
LS-DYNA ENVIRONMENT
98

A
T
• For each solid element the *ELEMENT_SOLID_ORTHO option allows the local material
E
axis to be set using two additional optional cards.
R
• Generally for solid elements the AOPT<0 option on the material card for referencing a I
*DEFINE_COORDINATE_... is used instead to define a vector for the material axis. A
L
S
Solid Element Local X-Axis Solid Element Material X-Axis

AOPT referencing a CID
(*DEFINE_COORDINATE_VECTOR as 1,0,0)
back to contents
LS-DYNA ENVIRONMENT
99
Material Models LS-DYNA Introductory Course M

A
T
The following material categories are discussed in the next section: E
R
• General I
A
• Viscoelastic L
• Foam and rubber S
• Special
• Johnson Cook
• Thermal
• Composite
• Fabric
• Discrete beam
• Concrete
back to contents
LS-DYNA ENVIRONMENT
100

A
GENERAL MATERIAL TYPES T
E
*MAT_ELASTIC [type 1] R
I
Straightforward elastic material model.
A
*MAT_PLASTIC_KINEMATIC [type 3] (*) L
Simple bilinear stress-strain relationship. S
*MAT_PIECEWISE_LINEAR_PLASTICITY [type 24] (*)
Stress-strain curve, strain rate effects and failure.
*MAT_MODIFIED_PIECEWISE_LINEAR_PLASTICITY [type 123]
Similar to type 24, has extra failure options – can have failure in tension only, i.e. no
failure in compression, suitable for metals; also allows for specifying number of
integration points that can fail before the shell is deleted.
*MAT_PLASTICITY_COMPRESSION_TENSION [type 124] (*)
Isotropic elastic-plastic material with unique yield stress vs. plastic strain defined for
tension and compression.
*MAT_PLASTICITY_POLYMER [type 89] (*)
Elasto-plastic material suitable for polymers.
back to contents
LS-DYNA ENVIRONMENT
101

A
T
*MAT_PLASTIC_KINEMATIC [type 3] E
A simple bilinear elastic plastic model R
I
MID RO E PR SIGY ETAN BETA A
L
S
SRC SRP FS VP
RO Density
E Young‟s modulus
PR Poisson‟s ratio ζ
ETAN
SIGY Yield stress
ETAN Tangent modulus SIGY
BETA Hardening parameter
SRC Cowper Symonds C parameter E
SRP Cowper Symonds P parameter FS
FS Plastic strain at failure ε
VP Viscoplastic formulation flag
back to contents
LS-DYNA ENVIRONMENT
102

A
T
*MAT_PLASTIC_KINEMATIC [type 3] E
R
• Von Mises yield criteria.
ζ I
• BETA=0 Kinematic hardening keeps yield A
ζmax
surface diameter constant whilst moving the L
centre of the surface. S
ζyield
• BETA=1 Isotropic hardening keeps the yield
surface centre fixed in space whilst allowing the 2ζyield
diameter to increase.
• Typical material properties for mild steel: ε
β=0 Kinematic
Young‟s modulus (E) 205 GN/m2
Poisson‟s ratio () 0.3
Beta 1 (isotropic)
-ζmax
yield stress (y) 200 MN/m2 β=1 Isotropic
hardening modulus(Eh) 1000 MN/m2
density () 7890 kg/m3
back to contents
LS-DYNA ENVIRONMENT
103

A
T
*MAT_PIECEWISE_LINEAR_PLASTICITY [type 24] E
An elastic-plastic model using a stress vs plastic strain curve after yield. R
I
MID RO E PR SIGY ETAN FAIL TDEL A
L
S
C P LCSS LCSR VP
RO Density
E Young‟s modulus
PR Poisson‟s ratio ζ LCSS
SIGY Yield stress (if no curve defined)
ETAN Tangent modulus (if no curve defined)
SIGY
FAIL Plastic strain at failure
C Cowper Symonds C parameter E
P Cowper Symonds P parameter
FAIL
LCSS Stress vs plastic strain curve (or table)
ε
LCSR Stress scale factor vs strain rate curve
VP Viscoplastic formulation flag
back to contents
LS-DYNA ENVIRONMENT
104

A
T
*MAT_PIECEWISE_LINEAR_PLASTICITY [type 24]
E
Used for: R
• Metallic parts (where good data is available) I
• Glass A
• Engineering plastics such as ABS and polypropylene L
(care needed with stress-strain curve to allow stable S
yielding)
A stress vs plastic strain curve is used for the post yield
behaviour. Typically this is obtained from tensile test data
or a bending test. ζ LCSS
The stress strain curve used is True Stress vs True
Plastic Strain. SIGY
Typically test data is given in engineering stress and strain,
E
so it may need converting.
FAIL
True stress = engineering stress x (1 + engineering strain); ε
True strain = loge (1 + engineering strain);
back to contents
LS-DYNA ENVIRONMENT
105

A
*MAT_PIECEWISE_LINEAR_PLASTICITY [type 24] T
Once necking occurs in the tensile test, the measured curve no longer applies to the whole E
gauge length of the sample, so its meaning is therefore questionable. R
I
It is better to just extrapolate the curve beyond the point where necking starts.
A
L
ζ Necking ζ Necking ζ Necking
S
Yield Yield Yield
ε εp εp
True Stress vs Strain True Stress vs Plastic Strain True Stress vs Plastic Strain
(extrapolated after necking)
• No CPU penalty for lots of points.
• LS-DYNA divides the curve into 100 (default) equally spaced points:
• This can lead to a loss of accuracy if the maximum strain value is very large
• It is recommended that the last point is at a strain of about 2.0
• The spacing can be changed on *CONTROL_SOLUTION, parameter LCINT.
back to contents
LS-DYNA ENVIRONMENT
106

A
T
*MAT_PLASTICITY_COMPRESSION_TENSION [type 124] E
Elastic plastic model with separate stress vs. plastic strain curves for tension and R
compression. I
A
MID RO E PR C P FAIL TDEL L
S
LCIDC LCIDT LCSRC LCSRT SRFLAG LCFAIL
PC PT PCUTC PCUTT PCUTF
RO Density LCIDC Compressive stress vs plastic strain

E Young‟s modulus LCIDT Tensile stress vs plastic strain
PR Poisson‟s ratio LCSRC Stress scale factor vs strain rate (comp)
C Cowper Symonds C parameter LCSRT Stress scale factor vs strain rate (ten)
P Cowper Symonds P parameter PC Compressive yield stress
FAIL Plastic strain at failure PT Tensile yield stress
back to contents
LS-DYNA ENVIRONMENT
107

A
T
*MAT_PLASTICITY_POLYMER [type 89] E
An elastic plastic model using a stress vs strain curve and a strain rate dependent R
failure strain. I
A
MID RO E PR L
S
C P LCSS LCSR
EFTX DAMP RATEFAC LCFAIL
RO Density LCSS True stress vs true total strain curve

E Young‟s modulus LCSR Stress scale factor vs strain rate curve
PR Poisson‟s ratio EFTX Failure type flag
C Cowper Symonds C parameter LCFAIL Failure strain vs strain rate curve
P Cowper Symonds P parameter
back to contents
LS-DYNA ENVIRONMENT
108

A
T
*MAT_PLASTICITY_POLYMER [type 89] E
Suitable for materials where the elastic to plastic transition is not as distinct as it is R
for metals, such as Polypropylene, ABS, etc. I
A
• The stress strain curve used is True Total Stress vs True Total Strain
L
• Strain rate sensitivity can be modelled using the Cowper Symonds rule or a yield S
stress scaling curve.
• Many plastics become more brittle at higher strain rates so this model allows the
plastic strain at failure (tension only) to be defined using a load curve of Failure
Strain vs. Strain Rate instead of a single value.
Strain Rate Effects Strain Rate Effects

Stress Failure Strain
ζ
Slow Loading
Medium Loading
Fast Loading
ε
back to contents
LS-DYNA ENVIRONMENT
109
Materials - *Mat_SAMP-1 (187) LS-DYNA Introductory Course M

A
*MAT_SAMP-1 T
• Semi-Analytical Model for Polymers. E
R
• This material model uses an isotropic C-1 smooth yield surface for the description of I
non-reinforced plastics. A
• Thermoplastics are currently modelled in crash simulations using von Mises plasticity L
laws – this is generally ok for large plastic deformations and failure, when the unloading S
behaviour is not relevant.
• However, for applications of pedestrian protection, the elastic recovery is of importance,
and the traditional models are not very accurate, as yield of thermoplastics in tension,
compression and shear will not fit a von Mises type yield surface.
• With *MAT_SAMP-1 and the modified Drucker-Prager yield surface, the bending
stiffness of the material can be better predicted.
• The unloading behaviour is approximated by a simple elastic damage model, which
represents an important improvement concerning problems where the elastic rebound
plays a dominant role, for example leg impact in pedestrian protection.
• Many tests needed to fully categorize the material.
• More information about this material can be found in the paper by S. Kolling et al, „A
Constitutive Formulation for Polymers Subjected to High Strain rates‟, presented at the
9th International LS-DYNA Users Conference.
back to contents
LS-DYNA ENVIRONMENT
110
Materials - *Mat_SAMP-1 (187) LS-DYNA Introductory Course M

A
*MAT_SAMP-1 T
For this material a series of stress-strain curves are required which are used to calculate a E
yield surface. Depending on what data is supplied a different shape yield surface is R
calculated. I
A
Tension        L
S
Compression    
Shear   
Biaxial Tension   
q q q q
Resultant
Yield Surface
p p p p
Von Mises Drucker-Prager General Isotropic General Isotropic
Quadratic Quadratic
(least squares fit)
back to contents
LS-DYNA ENVIRONMENT
111
Workshop Example 5 – Dogbone Sample LS-DYNA Introductory Course M

A
T
A uniaxial tension test was done on a dogbone sample (mm, s, t, N, MPa.) The dimensions
E
of the sample are given below:
R
I
A
127mm
L
S
680mm
1. Set up a *MAT_003 model for the steel sample. Use Young‟s modulus=3333MPa,
density=7.8E-9t/mm3, Poisson‟s ratio=0.3, Yield stress=500MPa and Etan=58MPa.
2. Run the analysis and asses if the simulation results are acceptable:
• Normal or error termination?
• Does the animation look sensible?
• Plot stress-strain for the history shell 1323 – how does this compare to the material
parameters entered?
back to contents
LS-DYNA ENVIRONMENT
112

A
T
MATERIAL TYPES FOR VISCOELASTIC E
R
*MAT_VISCOELASTIC [type 6] I
Dense rubbery foams. A
*MAT_GENERAL_VISCOELASTIC [type 76] L
Isotropic viscoelastic for modelling dense continuum rubbers and solid explosives. S
*MAT_ORTHOTROPIC_VISCOELASTIC [type 86]
Allows definition of an orthotropic material with a viscoelastic part for shell elements.
back to contents
LS-DYNA ENVIRONMENT
113

A
T
MATERIAL TYPES FOR FOAMS AND RUBBERS E
R
*MAT_BLATZ-KO_RUBBER [type 7] I
Foams with hysteresis. A
*MAT_MOONEY-RIVLIN_RUBBER [type 27] L
Simple rubber model. S
*MAT_LOW_DENSITY_FOAM [type 57] (*)
Foams with hysteresis.
*MAT_VISCOUS_FOAM [type 62]
E.g. Confor foam (rate dependent).
*MAT_CRUSHABLE_FOAM [type 63] (*)
Foams with no or little hysteresis.
*MAT_LOW_DENSITY_VISCOUS_FOAM [type 73]
Suitable for modelling low density urethane foam with high compressibility and rate
sensitivity – applications in seat cushion, Side Impact Dummies.
back to contents
LS-DYNA ENVIRONMENT
114

A
T
MATERIAL TYPES FOR FOAMS AND RUBBERS E
R
*MAT_HYPERELASTIC_RUBBER [type 77] I
General hyperhelastic rubber model combined optionally with linear viscoelasticity. A
*MAT_FU_CHANG_FOAM [type 83] (*) L
Model low and medium density foams. S
*MAT_ARRUDA_BOYCE_RUBBER [type 127]
Hyperelastic rubber model combined optionally with linear viscoelasticity.
*MAT_SIMPLIFIED_RUBBER/FOAM [type 181] (*)
Model for rubber and foam material based on simple uniaxial test data.
*MAT_SIMPLIFIED_RUBBER_WITH_DAMAGE [type 183]
Model incompressible rubber with different loading and unloading curves, similar to
*MAT_181 with the addition of damage option.
back to contents
LS-DYNA ENVIRONMENT
115

A
T
*MAT_LOW_DENSITY_FOAM [type 57] E
General purpose model for foams that recover their original shape on unloading. R
I
MID RO E LCID TC HU BETA DAMP A
L
S
SHAPE FAIL BVFLAG ED BETA1 KCON REF
RO Density
E Young‟s modulus for tension
LCID Stress vs volumetric strain curve
TC Tension cut off stress
HU Hysteresis unloading factor
DAMP Viscous damping coefficient
SHAPE Unloading shape factor
FAIL Post tension cut off behaviour – tensile stress remains at cut-off value or is reset to zero
KCON Stiffness coefficient for contact interface stiffness
back to contents
LS-DYNA ENVIRONMENT
116

A
T
• The main applications are for modelling seat cushions and the padding inside R
side impact dummies. I
A
• Compression is controlled by a Engineering Stress vs Volumetric Strain
L
curve. S
• Tensile and shear behaviour is elastic and governed by Young‟s modulus.
• An upper limit for the tension can be set using the TC parameter.
• The DAMP parameter can be used to add in some basic strain rate sensitivity to
the material.
back to contents
LS-DYNA ENVIRONMENT
117

A
T
• The HU and SHAPE parameters control the amount of hysteresis during R
unloading (energy dissipation). I
• It is recommended to set HU=0.01and use the SHAPE parameter to achieve the A
L
desired amount of hysteresis. S
ζ
Loading
Unloading
(SHAPE=0.1)
Unloading
(SHAPE=10)
ε
back to contents
LS-DYNA ENVIRONMENT
118

A
T
Name Variable Typical Value Notes R
I
If real density is very low, using
RO Density 4 e-11 T/mm3 artificially high density will help timestep
A
and contact instability L
S
E Young‟s Modulus 5 – 50 N/mm2 Governs tensile and shear behaviour
Define compressive stress vs. nominal
LCID Load Curve for compression compressive strain.
First point should be (0,0)
TC Tension cut-off stress 10 N/mm2
Default 1; generally, set to 0.01 and use
HU Hysteretic unloading factor
SHAPE to control hysteresis instead
0.001 – 0.1 Can be used to model simple strain rate
DAMP Damping factor
effects
High values increase energy dissipation
SHAPE Shape Factor 0.1 – 10 N/mm2
on unloading
If contact robustness problems occur, try KCON = 30N/mm2 and change the stress-strain load curve after
75% to make it stiffer.
back to contents
LS-DYNA ENVIRONMENT
119

A
T
*MAT_CRUSHABLE_FOAM [type 63] E
Used for isotropic materials that show permanent crush (e.g. expanded R
polystyrene, urea formaldehyde foam, etc.). I
A
MID RO E PR LCID TSC DAMP L
S
RO Density
E Young‟s modulus
PR Poisson‟s ratio
LCID Stress vs volumetric strain curve
TSC Tension cut off stress
back to contents
LS-DYNA ENVIRONMENT
120

A
T
• Compression is controlled by a Engineering Stress vs Volumetric Strain R
curve. I
A
• Unloading is elastic based on the Young‟s modulus and Poisson's ratio.
L
• Tensile and shear behaviour is elastic and governed by Young‟s modulus and S
Poisson's ratio.
• An upper limit for the tension can be set using the TSC parameter.
• The DAMP parameter can be used to add in some basic strain rate sensitivity to
the material.
ζ Loading
Unloading
back to contents ε
LS-DYNA ENVIRONMENT
121

A
T
Name Variable Typical Value Notes R
I
If real density is very low, using
RO Density 4 e-11 T/mm3 artificially high density will help timestep
A
and contact instability L
S
E Young‟s Modulus 5 – 50 N/mm2 Governs tensile and shear behaviour
PR Poisson‟s Ratio 0
Define compressive stress vs.

LCID Load curve for compression volumetric strain.
First point should be (0,0)
TSC Tension cut-off stress 10 N/mm2
Can be used to model simple strain rate
DAMP Damping factor 0.001 – 0.1
effects
back to contents
LS-DYNA ENVIRONMENT
122

A
T
*MAT_FU_CHANG_FOAM [type 83] E
A more detailed foam model that allows for strain rate sensitivity and also R
non-linear behaviour in tension. I
A
MID RO E ED TC FAIL DAMP TBID L
S
BVFLAG SFLAG RFLAG TFLAG PVID SRAF REF HU
RO Density
E Young‟s modulus
TC Tension cut off stress
FAIL Post tension cut-off behaviour
TBID Table of stress vs strain curves for a range of strain rates
TFLAG Tension behaviour flag (linear or non-linear)
back to contents
LS-DYNA ENVIRONMENT
123

A
T
*MAT_FU_CHANG_FOAM [type 83] E
• Based on the unified constitutive equations for foams by Chang. R
• Drop-tower test curves for engineering stress vs. strain can be input directly. I
A
• Tension behaviour can be linear or based on the stress vs strain curves (TFLAG). L
• A range of different unloading behaviours are available. S
ζ
Table of stress vs strain

curves for a range of
strain rates
Compression
Tension ε
Tension based
on curve data
TFLAG=1
Tension based on
Young’s modulus
TFLAG=0 (default)
back to contents
LS-DYNA ENVIRONMENT
124

A
T
*MAT_SIMPLIFIED_RUBBER/FOAM [type 181] E
This simplified model allows for a fast generation of input data based on uniaxial R
static and dynamic test data → Ogden parameters are computed from uniaxial I
tensile and compressive data only. A
L
MID RO KM MU G SIGF REF PRTEN S
SGL SW ST LC/TBID TENSION RTYPE AVGOPT PR/BETA
RO Density
KM Bulk modulus
MU Damping Coefficient
SGL Specimen gauge length
SW Specimen width
ST Specimen thickness
LC/TBID Stress vs strain curve or table
back to contents
LS-DYNA ENVIRONMENT
125

A
T
*MAT_SIMPLIFIED_RUBBER/FOAM [type 181] E
• If a Poisson‟s ratio is defined the foam formulation is activated. R
• Otherwise an incompressible rubber is assumed (Poisson‟s ratio = 0.495). I
A
• The stress strain behaviour is defined using the results from a uniaxial test or a L
series of tests at a range of strain rates. S
• A force vs displacement curve is defined along with the test specimen
dimensions, or a stress vs strain curve can be used and the dimensions set to 1.
ζ
Table of stress vs strain
curves for a range of
strain rates
Tension
Compression ε
back to contents
LS-DYNA ENVIRONMENT
126
Workshop Example 6 – Foam block LS-DYNA Introductory Course M

A dynamic test has been done on a 100x100x50mm block of foam. The foam was placed on a A
flat surface and a flat impactor of mass 10kg was dropped onto the block. T
The acceleration and displacement of the impactor are provided in the curve file E
„example5_mat_test.cur‟. Units: mm, s, t, N, MPa. R
I
1. Use T/HIS to calculate the stress on the block using the acceleration data, the impactor A
mass and the foam block cross-sectional area. L
2. Use T/HIS to calculate the strain in the block using the displacement data. S
3. Use the T/HIS COMBINE function to derive a stress-strain curve for the foam.
4. An input file containing everything except the material definition for the foam has already
been created. Create a material definition using your stress strain curve and run the model.
5. Compare the acceleration of the impactor in your simulation to the original test data.
6. Compare z-stress vs. z-strain results for a
history solid with the input material curve – how
does the hysteresis compare?
7. Try changing the parameters that control the
amount of hysteresis in your foam model:
HU=0.01 & SHAPE=5 and look at the stress-
strain results again.
(Hints: Use *MAT_057 for the foam, density=1.0e-11,
E=100.0)
back to contents
LS-DYNA ENVIRONMENT
127

A
SPECIAL MATERIAL TYPES T
E
*MAT_NULL [type 9] (*) R
This material has no stiffness or strength. I
A
*MAT_RIGID [type 20]
L
Non-deforming structure - SEE SECTION ON RIGID BODIES. S
*MAT_HONEYCOMB [type 26]
Orthotropic crushing characteristics - e.g. aluminium honeycomb structures.
*MAT_MODIFIED_HONEYCOMB [type 126] (*)
Based on type 26 with additional yield options.
*MAT_LAMINATED_GLASS [type 32]
Layered glass, including polymeric layers and glass failure option.
*MAT_SPOTWELD [type 100]
Spotweld beams - see section on Spotwelds.
*MAT_ARUP_ADHESIVE [type 169]
Model adhesive bonding in aluminium structures.
back to contents
LS-DYNA ENVIRONMENT
128

A
T
*MAT_NULL [type 9] E
This material is used for shell elements that have been created on solid element R
faces (coating) to help with contact modelling (particularly with soft materials). I
A
L
MID RO PC MU TEROD CEROD YM PR
S
RO Density
MU Dynamic viscosity coefficient (for fluid modelling)
YM Young‟s modulus only used for contact stiffness calculation
PR Poisson‟s ratio only used for contact stiffness calculation
back to contents
LS-DYNA ENVIRONMENT
129

A
T
*MAT_NULL [type 9] E
• This material has no stiffness or strength. R
• The elements will still have a mass however (density*area*thickness). I
A
• The Young‟s modulus and Poisson‟s ratio are only used for calculating the L
contact stiffness (see later section). S
• Can also be used to model a fluid in combination with an equation of state *EOS
to describe the pressure/volume relationship.
back to contents
LS-DYNA ENVIRONMENT
130

A
T
*MAT_MODIFIED_HONEYCOMB [type 126] E
Used for modelling orthotropic crushable materials/structures such as aluminium R
honeycomb. I
A
MID RO E PR SIGY VF MU BULK L
S
LCA LCB LCC LCS LCAB LCBC LCCA LCSR
EAAU EBBU ECCU GABU GBCU GCAU AOPT MACF
RO Density LCS Shear stress vs shear strain curve

E Young‟s modulus (compacted) LCAB, LCBC, Shear Stress strain curves (ab,bc,ca)
PR Poisson‟s ratio (compacted) LCCA
SIGY Yield stress (compacted) EAAU, EBBU, Young‟s modulus (un-compacted)
VF Relative volume for compaction ECCU
LCA, LCB, Stress strain curves (aa,bb,cc) GABU, GBCU, Shear modulus (un-compacted)
LCC GCAU
back to contents
LS-DYNA ENVIRONMENT
131

A
T
This material exists in two phases: at the start of the analysis it is Un-Compacted; R
when the relative volume (Vr = V/Vo= 1 at start of analysis) reaches the value set I
for VF, it switches to Compacted. A
L
S
Un-compacted
Orthotropic elastic plastic behaviour controlled by Young‟s modulus, Shear
modulus, and Stress-Strain curves for each axis (Direct: aa, bb, cc, Shear: ab,bc,ca):
(Elastic: EAAU, EBBU, ECCU, GABU, GBCU, GCAU)
(Plastic: LCA, LCB, LCC, LCS, LCAB, LCBC, LCCA)
Compacted (when relative volume < VF)

Isotropic elastic perfectly plastic controlled by Young‟s modulus, Poisson‟s ratio
and the Yield stress:
(Elastic: E, PR)
(Plastic: SIGY)
back to contents
LS-DYNA ENVIRONMENT
132

A
T
• The load curves must start at zero (x-value) or LS-DYNA will assume that x- R
values are relative volumes = V/V0 (1→0) instead of volumetric strain =1-V/V0 I
(0→1). A
L
• All load curves should have points at the same x-values. Otherwise the CPU S
costs greatly increase due to the required interpolation.
• Use as few points in load curves as possible for maximum efficiency.
• The elastic modulus should be greater than the steepest part of the stress
strain curves to prevent negative hysteresis.
ζ
Negative
Hysteresis
ε
back to contents
LS-DYNA ENVIRONMENT
133

A
*MAT_MODIFIED_HONEYCOMB [type 126] T
E
Name Variable Typical Value Notes
R
1e-11 to 1e-10 If real density is very low, use a high I
RO Density 3
T/mm density to give a reasonable timestep A
Compacted Young's Lower than real material to give a L
E 100 N/mm2
modulus reasonable timestep S
PR Compacted Poisson's ratio 0.0 Unimportant
SIGY Compacted Yield stress 1000 N/mm2 Normally use a high value
Relative volume at full Setting this too low increases the
VF 0.1
compaction chance of elements turning inside out
MU Damping factor 0.001 – 0.1 Can be used to model strain rate effects
LCA, LCB, LCC Stress-Strain curves
LCS or LCAB, Shear Stress-Strain If in doubt use curves identical to main
LCBC, LCCA curves. crush curves with values halved
EAAU, EBBU, Uncompacted Young's A bit less than the fully crushed
10 N/mm2
ECCU modulus modulus
GABU, GBCU, Uncompacted shear 2 Half the uncompressed Young's
5 N/mm
GCAU modulus modulus to give a Poisons ratio of 0.0
back to contents
LS-DYNA ENVIRONMENT
134

A
T
JOHNSON-COOK MATERIALS E
R
*MAT_JOHNSON_COOK [type 15]
I
Model strain and temperature sensitive plasticity – use if adiabatic temperature A
increase due to plastic heating cause material softening. L
*MAT_SIMPLIFIED_JOHNSON_COOK [type 98] S
Thermal effects and damage are ignored in this model, 50% faster than *MAT_15;
to compensate for the lack of thermal softening, limiting stress values are used.
*MAT_SIMPLIFIED_JOHNSON_COOK_OTHOTROPIC_DAMAGE [type 99]
Extension of mat 098 with multiple through thickness integration points;
implemented for aluminium panels, to allow orthotropic damage.
*MAT_MODIFIED_JOHNSON_COOK [type 107]
Complex model based on *MAT_015 with additional options for damage,
constitutive laws used, fracture criteria.
back to contents
LS-DYNA ENVIRONMENT
135
Material models LS-DYNA Introductory Course M

A
T
*MAT_SPRING_... or *MAT_DAMPER_... must be used in conjunction with the
E
*SECTION_DISCRETE , i.e. for springs and dampers.
R
I
Many different behaviours can be modelled: A
L
• *MAT_SPRING_ELASTIC [S01] – simple elastic spring
S
• *MAT_SPRING_ELASTOPLASTIC [S03] – bi-linear plasticity spring
• *MAT_SPRING_NONLINEAR_ELASTIC [S04] – non-linear elastic spring
• *MAT_SPRING_GENERAL_NONLINEAR [S06] – general non-linear plasticity spring
with loading and unloading curves
• *MAT_SPRING_INELASTIC [S08] – non-linear plasticity tension/compression only
spring with loading curve and unloading stiffness
• *MAT_DAMPER_VISCOUS [S02] – simple linear viscous damper
• *MAT_DAMPER_NONLINEAR_VISCOUS [S05] – non-linear viscous damper
back to contents
LS-DYNA ENVIRONMENT
136

A
T
THERMAL MATERIALS – for inputting temperature dependent structural material E
properties – not for thermal heat flow problems: R
I
*MAT_ELASTIC_PLASTIC_THERMAL [type 04] A
Temperature dependent material using at least two (and up to eight) temperatures L
and their corresponding material properties. S
*MAT_ORTHOTROPIC_THERMAL [type 21]
Linear elastic orthotropic material with orthotropic thermal expansion.
*MAT_TEMPERATURE_DEPENDENT_ORTHOTROPIC [type 23]
Orthotropic elastic material with temperature dependency; similar to *MAT_021 but
allows to input the constitutive parameters (Young‟s modulus, Poisson‟s ratio,
Shear modulus) for three directions and in addition, for various temperatures.
For thermal heat flow problems, a different set of materials are used:
*MAT_THERMAL_{OPTION} – see more about this in the „Other Techniques‟ section.
back to contents
LS-DYNA ENVIRONMENT
137

A
COMPOSITE MATERIALS T
• The method of modelling composites depends on the scope of the simulation. E
• Main techniques include: R
I
One Shell Element Layers of Solid Elements Layers of Shell Elements A
L
S
Mat3
Mat1
Mat4 Mat3 Mat3
Mat1 Mat1
Mat5 Mat4
Mat4
Mat2 Mat5 Mat5
Mat1
Mat2 Mat2
• With *PART_COMPOSITE all the layers Mat1 Mat1
can be defined using a single shell
element. • Each layer can be modelled • Each layer can be modelled
• One integration point available for each separately. separately.
layer. • The connection between layers can • The connection between layers
• Can control with MAXINT amount of be node-to-node, or using can be done using a _TIEBREAK
output in thickness. _TIEBREAK contact, or by contact, or by physically modelling
• Cannot model delamination. physically modelling adhesive. adhesive.
back to contents • Can model delamination. • Can model delamination.
LS-DYNA ENVIRONMENT
138
Composite Modelling – Shell Orientation LS-DYNA Introductory Course M

A
T
For shells of orthotropic/anisotropic materials, there are 3 coordinate systems E
to consider: R
I
• Global A
• Element local L
• Material Orientation S
N3 Z
Y
N4 z c
b a X
y x
N2 Global coordinates
N1
z-axis = Shell Normal
a-axis = Material Orientation – coincides with x-shell if AOPT=0
back to contents
LS-DYNA ENVIRONMENT
139

A
T
Two ways of changing the orientation of the material: E
R
1. Via parameter BETA added to the shell local x, which can be input in I
the following ways: A
o *ELEMENT_SHELL_BETA (base angle-offset) L
o *PART_COMPOSITE (layer angle-offset) S
o Also on some *MAT_...
2. Via *PART_COMPOSITE, using parameters B1, B2, ...Bn
back to contents
LS-DYNA ENVIRONMENT
140

A
1. BETA Example: *ELEMENT_SHELL_BETA T
E
R
*ELEMENT_SHELL_BETA
1 1 1 2 13 12
I
0.0 0.0 0.0 0.0 45.000000 A
L
S
Element Shell Thickness (If zero, BETA – Used to define the
then *SECTION_SHELL card is used) Material Orientation offset
from the basic element x-axis
Example BETA = 45deg

N4 N3
Material a Note that the BETA parameter is available on:
Orientation
1. *ELEMENT_SHELL_BETA (base angle-offset)
2. *PART_COMPOSITE (layer angle-offset)
45deg
X
3. Also on some *MAT
N1 N2
back to contents
LS-DYNA ENVIRONMENT
141
PART_COMPOSITE LS-DYNA Introductory Course M

A
2. *PART_COMPOSITE T
*PART_COMPOSITE is a simplified method of defining a layered material, and has replaced E
the older method of using a *INTEGRATION_SHELL & *SECTION_SHELL with a separate R
part definition for each layer. I
A
Card2 PID ELFORM SHRF NLOC MAREA HGID ADOPT ITHELFM L
S
Card3 MID1 THICK1 B1 TMD1 MID2 THICK2 B2 TMD2
Card4 MID3 THICK3 B3 TMD3 MID4 THICK4 B4 TMD4
*PART_COMPOSITE
Card1 Example Part
Card2 1 2 0.0 0.0 0.0 0 0 0
Card3 1 1.0 0.0 0 2 2.0 90.0 0
Card4 2 2.0 90.0 0 1 1.0 0.0 0
$
*MAT_ELASTIC
1 7.85E-9 210000.0 0.3 0.0 0.0 0.0
*MAT_ELASTIC
2 2.7E-9 60000.0 0.3 0.0 0.0 0.0
back to contents
LS-DYNA ENVIRONMENT
142
PART_COMPOSITE LS-DYNA Introductory Course M

A
T
Resultant Part Layup for the card in the previous slide: E
R
I
A
0° L
Layer 4
Mat 1, Thick 1.0mm S
Total Thickness = 6.0mm

Shell Normal
Layer 3 90°
Mat 2, Thick 2.0mm
90°
Layer 2
Mat 2, Thick 2.0mm
Layer 1 0°
Mat 1, Thick 1.0mm
back to contents
LS-DYNA ENVIRONMENT
143

A
T
E
• The total offset of one of the layers in the composite is the sum of the: R
„base angle-offset‟ + „layer angle-offset‟: I
A
L
S
B1 = 30° („layer offset‟, on
*PART_COMPOSITE)
N4 N3
Material a
orientation for 30°
this particular BETA = 25° („base offset‟, on
layer – total 25° *ELEMENT_SHELL_BETA,
angle of 55° X overrides BETA on *MAT)
N1 N2
back to contents
LS-DYNA ENVIRONMENT
144

A
T
COMPOSITE MATERIALS E
R
The main difference between the LS-DYNA orthotropic and composite materials is that for
I
the composite materials an option for failure of fibres and matrix is available.
A
*MAT_COMPOSITE_DAMAGE [type 22] L
Orthotropic material with optional brittle failure. S
*MAT_ENHANCED_COMPOSITE_DAMAGE [type 54-55] (*)
Enhanced type 22, allows different stiffness in mutually perpendicular directions
and failure based on fibre tensile, compressive strain, maximum matrix strain,
maximum shear strain and time step size.
*MAT_LAMINATED_COMPOSITE_FABRIC [type 58]
Model composite materials with unidirectional layers, complete laminates and
woven fabrics – for shells and thick shells only.
*MAT_COMPOSITE_FAILURE_SHELL_MODEL [type 59]
Basic composite model for use with shells.
*MAT_COMPOSITE_FAILURE_SOLID_MODEL [type 59]
Basic composite model for use with solids.
back to contents
LS-DYNA ENVIRONMENT
145
Material Models MATERIALS LS-DYNA Introductory Course M

A
T
*MAT_ENHANCED_COMPOSITE_DAMAGE [type 54-55] E
Arbitrary orthotropic material for modelling unidirectional (uniaxial) layers of R
composite material that allows for different failure criteria in the fibre and matrix I
directions. A
L
S
MID RO EA EB EC PRBA PRCA PRCB
GAB GBC GCA KF AOPT
RO Density
EA Young‟s modulus in a direction
EB Young‟s modulus in b direction
PRBA Poisson‟s ration in ba plane
GAB Shear modulus in the ba plane
GBC Shear modulus in the bc plane
GCA Shear modulus in the ca plane
back to contents
LS-DYNA ENVIRONMENT
146

A
T
*MAT_ENHANCED_COMPOSITE_DAMAGE [type 54-55] E
• An enhanced version of *MAT_022 that is only valid for thin shells. R
• Two different matrix failure models available. I
A
• Chang/Chang – *MAT_54 L
• Tsai-Wu – *MAT_55 S
• Lamination theory can be activated in *CONTROL_SHELL (LAMSHT) to correct
the assumption of constant shear strain through the thickness.
• Damage option available to model interlaminar shear failure.
back to contents
LS-DYNA ENVIRONMENT
147

A
T
FABRIC MATERIAL E
R
*MAT_FABRIC [type 34]
I
• Only used with shell elements – it invokes a special membrane element. A
• Applications: airbags, fabric roofs and seatbelt elements. L
• It is a composite layered orthotropic material with a liner built in together with the fabric. S
tliner
t
tliner
LRATIO 
t
E, ν
Ea, Eb, νab, Gab
back to contents
LS-DYNA ENVIRONMENT
148

A
T
FABRIC MATERIAL E
R
*MAT_FABRIC [type 34]
I
• Since in reality buckling of the tension-only fibres can easily occur, a special flag (CSE) A
can be set to prevent compressive stresses. In addition a linear elastic liner is used to L
support compressive stresses and to prevent these elements from crushing. S
• Compression in liner (LNRC) can be switched off until reference geometry is reached.
• Several special membrane element formulations are available:
• FORM =0,1,2, 12 – orthogonal material angles, defined through AOPT = 0, 2, 3
• FORM = 3, 4, 13, 14 – non-orthogonal material angles, a and b vectors defined
through ICOMP=1 in keyword *SECTION_SHELL
• FORM = 4 and 14, definition of loading and unloading curves is permitted.
back to contents
LS-DYNA ENVIRONMENT
149

A
T
DISCRETE BEAM MATERIAL E
R
*MAT_LINEAR_ELASTIC_DISCRETE_BEAM [type 66]
I
*MAT_NONLINEAR_ELASTIC_DISCRETE_BEAM [type 67] A
*MAT_NONLINEAR_PLASTIC_DISCRETE_BEAM [type 68] L
Connections, bolts, etc. - see section on Connections for more details on discrete beam. S
back to contents
LS-DYNA ENVIRONMENT
150
Materials – Discrete
Connections – Discrete
Beams
Beams LS-DYNA Introductory Course M
s
A
T
• With the discrete beam (Type 6) element use
E
one of the following materials: r
n2 R
• *MAT_LINEAR_ELASTIC_DISCRETE_ I
BEAM [66] – simple linear elastic response. A
t
• *MAT_NONLINEAR_ELASTIC_DISCRETE_ n1 L
BEAM [67] – non-linear elastic response S
Transl. stiff. Rot. stiff.
defined by load curves.
• *MAT_NONLINEAR_PLASTIC_DISCRETE_
BEAM [68] – non-linear elastic-plastic
response defined by load curves with optional
failure by load or displacement limit. Transl. damp. Rot. damp.
• These materials use 6 linear or nonlinear
springs to define stiffness in 6 degrees-of-
freedom for each beam.
• The input is force vs. deflection or moment vs. Preload force Preload moment
rotation (in radians); damping and preload can
also be defined.
• Cards for *MAT_066 shown to the right:
back to contents
LS-DYNA ENVIRONMENT
151
Materials – Discrete
Beams
Beams LS-DYNA Introductory Course M
A
T
• The diagrams below show examples of input for materials *MAT_066, _067 and _068.
E
• For material 067, the compression behaviour can be different from the tension behaviour. R
• For material 068 the initial yield point corresponds to a plastic displacement of zero. Can I
still have an elastic behaviour in any desired degree of freedom, by setting the curve ID to A
zero and non-zero stiffness coefficient (e.g. TKR). Failure can also be defined. L
S
Force/Moment Force/Moment Force/Moment
Elastic Elastic Plastic

Displacement Displacement Displacement
MAT_066 MAT_067 MAT_068
back to contents
LS-DYNA ENVIRONMENT
152
Materials - *MAT_ARUP_ADHESIVE LS-DYNA Introductory Course M

A
T
*MAT_ARUP_ADHESIVE E
• This material model was written for adhesive bonding. R
I
• The plasticity model is not volume conserving and hence avoids the A
spuriously high tensile stresses that can develop if adhesive is modelled L
using traditional elasto-plastic material models. S
• It is available only for solid elements of formulations 1 & 2 and it is
assumed that the smallest dimension of the element is the through-
thickness dimension of the bond.
• In-plane stresses are set to zero: it is assumed that the stiffness and
strength of the substrate is large compared with that of the adhesive, given
the relative thicknesses.
back to contents
LS-DYNA ENVIRONMENT
153
Materials - *MAT_ARUP_ADHESIVE LS-DYNA Introductory Course M

A
T
*MAT_ARUP_ADHESIVE E
• Generally with adhesives, if the substrate is modelled using shells, then the shells are R
expected to lie at the mid-plane of the substrate, and so the adhesive solid elements will I
be thicker than the original bond, as shown below. A
• This means that if the elastic stiffness of the bond is important for the simulation, then L
the Young‟s modulus needs to be scaled accordingly. S
• However, with the new *MAT_ARUP_ADHESIVE there is an option EXTRA=2 &
BTHCK=„real thickness‟, which will do the scaling of the elastic stiffness automatically.
Real
components
FE model FE bond -
Solid element
Panels
Real Shells at
Bond mid-plane
back to contents
LS-DYNA ENVIRONMENT
154
*MAT_120/ *MAT_GURSON LS-DYNA Introductory Course M
A
• *MAT_GURSON – Gurson formulation is a phenomenological constitutive T
model for a progressively cavitating ductile material, which comprises a E
macroscopic yield condition and an evolution law for the void volume fraction, R
assumed to fully describe the material microstructure. I
A
• This model is available for both shell and solid elements. L
S
• The material model introduces adjusting parameters in the constitutive model,
in order to better simulate the loss of material stress carrying capacity during
void growth and coalescence.
• Strain Rate effects can be included by using a table (*DEFINE_TABLE) to

define a series of tension stress-strain curves at different strain rates, instead
of a single curve. F0 and FN in the unstressed condition can be determined
from scanning electron microscope images of the virgin material.
• A detailed description of this model can be found in the following references:

Gurson [1975, 1977], Chu and Needleman [1980] and Tvergaard and
Needleman [1984]. The implementation in LS-DYNA is based on the
implementation of Feucht [1998] and Faßnacht [1999], which was recoded at
LSTC.
back to contents
LS-DYNA ENVIRONMENT
155
*MAT_ADD_EROSION LS-DYNA Introductory Course M

A
T
*MAT_ADD_EROSION
E
• Many of the material models in LS-DYNA do not allow failure and erosion.
R
I
• This option provides a way of including failure in these models.
A
L
• Define *MAT_ADD_EROSION for each material requiring failure.
S
• From ls971_R4, this option applies to 2D continuum, 3D solid elements, 3D shell
elements and thick shell elements type 1 & 2.
• Various failure criteria available:

• Pressure (positive in compression)
• Maximum principal stress
• Equivalent stress (Von Mises)
• Maximum principal strain
• Shear strain
• Stress impulse – calculated by Tuler-Butcher criterion (a measure of amount that
the maximum principal stress exceeds a „threshold‟ stress)
back to contents
LS-DYNA ENVIRONMENT
156

A
T
CONCRETE MATERIALS E
Concrete is a commonly used material but its properties and behaviour are R
I
relatively complicated. The main properties that need to be considered are:
A
L
• Young‟s Modulus S
• Poisson‟s Ratio
• Compressive Strength
• Tensile Strength
• Tri-axial behaviour
• Volumetric crushing (pressure vs. volumetric strain response)
• Cracking and crack orientation
• Compressive and Tensile strain softening
• Strain rate behaviour
• Shear-dilation
• Reinforcement
back to contents
LS-DYNA ENVIRONMENT
157

A
T
R
*MAT_SOIL_AND_FOAM [type 5] or *MAT_SOIL_AND_FOAM_FAILURE [type 14] I
Relatively simple model for concrete. Should only be used where confinement is A
present. L
S
*MAT_SOIL_CONCRETE [type 78]
Reasonably good model for concrete under large compression loads.
*MAT_CONCRETE_DAMAGE [type 72]

Relatively complex model for concrete.
*MAT_WINFRITH_CONCRETE [type 84/85]

Relatively simple to define. Oriented tensile cracking. Most suitable to tensile failure
situations.
back to contents
LS-DYNA ENVIRONMENT
158

A
T
R
Typical values for the basic properties of standard concrete can be determined I
from the following equations (taken from ACI 318-95) based on the compressive A
strength fc‟. L
S
• For units in psi
Young‟s Modulus E = 57000  fc‟
Tensile Strength ft = 6.7  fc‟
• For units in MPa

Young‟s Modulus E = 4733  fc‟
Tensile Strength ft = 0.556  fc‟
Typical values for Poisson‟s Ratio are between 0.15 and 0.22, with 0.19
commonly used.
back to contents
LS-DYNA ENVIRONMENT
159

A
Summary of Concrete Material Model Capabilities T
E
Material Model Types R
5 /14 72 78 84 (85)
I
A
Soil/Foam Concrete_damage Soil_concrete Winfrith
L
Compressive Initial Yield Surface     S
Compressive Meridian Surface    ()
Compressive Residual Surface   () 
Compressive Strain Softening    
Volumetric Crushing    
Strain Rate Sensitivity    
Separate Tensile Meridian    
Tensile Failure  / ()  () 
Tensile Strain Softening    
Oriented Tensile Cracks    
Reinforcement  ()  
back to contents
LS-DYNA ENVIRONMENT
160
Outline – THEORY AND PRACTICE LS-DYNA Introductory Course C

O
N
1. Time Integration 4. Materials T
• Description of explicit time 5. Contact Surfaces A
integration C
integration • Contact method S

• Summary of cost considerations • Other contact issues
6. Rigid Bodies
• Meshing Notes
3. Hourglassing
back to contents
LS-DYNA ENVIRONMENT
161
Contact LS-DYNA Introductory Course C

O
N
General characteristics T
A
• Contacts are surfaces defined from existing nodes and elements to prevent C
parts of the model from passing through (or separating from) each other. T
S
– Penalty based – Use numerical springs to:
• prevent penetration of nodes of one surface into segments of the other,
• transfer of loads between two parts.
– Constraint based – Create a kinematic constraint between nodes of one
part and segments of the other.
back to contents
LS-DYNA ENVIRONMENT
162
Contact LS-DYNA Introductory Course C

O
N
General characteristics T
A
• LS-DYNA offers a large number of contact types and a number of parameters C
that can be used to improve the contact performance for a specific case. T
S
• Contacts which have „AUTOMATIC‟ in the name are an improved type and are
non-oriented, i.e. can detect penetration coming from either side of the shell
element – recommended as more robust.
back to contents
LS-DYNA ENVIRONMENT
163
Contact Types LS-DYNA Introductory Course C

O
N
Penalty based: T
– *CONTACT_..._SINGLE_SURFACE types used when the exact position A
of contact is unknown e.g. a component folding on itself or to cover large C
areas of groups of components. T
S
– *CONTACT_..._SURFACE_TO_SURFACE or
*CONTACT_..._NODE_TO_SURFACE types used when the direction of
contact between two or more components is predictable.
...
Constraint-based:
– *CONTACT_TIED types that constrain parts together.
– *CONTACT_TIEBREAK types that tie and allow failure.
...
Force transducers:
– *CONTACT_FORCE_TRANSDUCER only measures contact forces
generated by other penalty based contact between one area and the rest
of the model.
back to contents
LS-DYNA ENVIRONMENT
164
Contact Theory – Terminology LS-DYNA Introductory Course C

O
N
• Segment - face of an element included in the contact – e.g. 3 or 4-noded shell or T
one face of a solid. A
• Slave/master - names for two opposing sides of a contact surface. C
• Contact can use one-way or two-way definition. For two-way masters and slaves T
are reversed and contact checked both ways. S
• Typically, if the name of the contact does not specify it, the contact is two-way;
single surface contacts are two way.
Slave
Slave ONE-WAY
Master
Slave
Master
TWO-WAY
back to contents Master
LS-DYNA ENVIRONMENT
165
Contact Theory – Penalty Method LS-DYNA Introductory Course C

O
How do contacts work? N
T
• The penetrating slave node is projected on the master segment.
A
• A force is calculated by the “penalty” method and applied to the slave node, to prevent C
penetration. T
• A reaction force is applied to the master segment nodes, the magnitude of the force on S
each node depends on how close it is to the projected slave node, such that the total force
on the master nodes equals the applied slave node force (n1-n4 are weighting factors).
• Both a normal (FN ) and a friction (Ff = µ *FN) penalty force are applied to the slave node.
Normal forces Friction forces
n2*FN FN n1*FN
n2*Ff Ff
n3*FN n1*Ff
n4*FN
+ n3*Ff
Direction of
Projection of slave node n4*Ff relative velocity
to master segment
back to contents
LS-DYNA ENVIRONMENT
166
Contact Theory – Contact Stiffness LS-DYNA Introductory Course C

O
N
• The slave node is projected on to the master T
segment surface. F F = k * DP A
• The penetration depth DP is calculated. C
T
• The slave node force is calculated by the k
S
“penalty” method and a penalty force is applied
on the slave node (see also next slide):
Force = Stiffness * DP d
• The force is then reacted on the master
DP
segment nodes.
Slave node
t/2 0 t/2
back to contents
LS-DYNA ENVIRONMENT
167
Contact Theory – Contact Stiffness LS-DYNA Introductory Course C

O
N
The basic calculation of the stiffness of a contact segment is as follows: T
A
fs  Area  K
2
fs  Area  K
k k C
Volume Minimum _ diagonal T
S
for segments on solid elements for segments on shell elements
where:
• Area – the area of the contact segment,
• K – the bulk modulus of the contacted element and
• fs – the penalty factor (same as the SLSFAC parameter in the
*CONTROL_CONTACT)
• Note that a rigid material still requires E and ν values, to calculate the bulk modulus,
K, for the contact stiffness:
E
K
3(1  2 )
For shells on solids, the shells determine the contact stiffness.

back to contents
LS-DYNA ENVIRONMENT
168
Contact Theory – Contact Thickness LS-DYNA Introductory Course C

O
N
• The contact thickness TC (for single surface and automatic contacts) is half the T
element thickness (or half the sum of slave and master thickness), but with a A
small no-man‟s-land (PENMAX) in the centre. C
T
• The contact is released and the node will pass through once D = T –PENMAX. S
P C
TC = (TS+TM)/2
Slave mid-plane
TS
TC
DP
TM
Master mid-plane DP= 0 DP> 0

back to contents
LS-DYNA ENVIRONMENT
169
Contact Theory – Contact Search LS-DYNA Introductory Course C

O
Bucket Search - which nodes are contacting which segments? N
T
A
• The bucket search was introduced for efficiency reasons, such that a contact penetration C
check is not done each cycle for each slave node against each segment in the model.
T
• In a bucket search the model is divided into a grid of boxes or buckets. This can be S
visualised in two dimensions as shown below:
• Segments in bucket 9 will be checked for penetration by nodes in buckets 1 to 9 to find

closest segment. This is based on the current spatial location of the nodes and
segments.
• Default bucket sort frequency is every 10 - 100 cycles for SMP, depending on contact
type, and 200 for MPP – avoid reducing if possible, as the sort is expensive.
• During every cycle, an incremental search is performed for all the pairs of slave nodes
and master segments which were identified during the bucket search. Then, the buckets
are reassessed at the bucket sort frequency.
back to contents
LS-DYNA ENVIRONMENT
170
Contact Theory LS-DYNA Introductory Course C

O
N
• Parameters are set in a hierarchy: T
1. *CONTROL_CONTACT sets overall defaults, A
2. *CONTACT_… sets parameters for that contact – overrides 1 above, C
T
3. *PART_CONTACT sets parameters for that part in the contact – overrides S
1 & 2 above.
• Important parameters are:

– Contact thickness and scaling factors
– Penalty stiffness scale factors or SOFT options
– Friction values
– Treatment of initial penetrations with IGNORE option
• Note: in all AUTOMATIC contacts, all segments that are connected to the
neighbour master node are checked for penetration by the slave node – this
allows to treat cavities correctly.
back to contents
LS-DYNA ENVIRONMENT
171
Contact – Parameters definition LS-DYNA Introductory Course C

O
N
Creating a contact – mandatory cards: T
Card 1 A
C
T
S
• Define Slaves (SSTYP) and Masters using (MSTYP):

• Set of Segments (0), set of Shells (1), set of Parts (2), Part (3), Node Set (4) - slave
nodes for node to surface contact.
• All (5) - for single surface definitions (no master defined).
• Set of exempted parts (6) - slave side only.
back to contents
LS-DYNA ENVIRONMENT
172

O
N
Card 2 A
C
T
S
• Define friction FS (static) and FD (dynamic):

• Fstatic = Fdynamic = 0.2 is typical
• Contact birth time (BT) and death time (DT).
back to contents
LS-DYNA ENVIRONMENT
173

O
N
Card 3 A
C
T
S
• Penalty stiffness scale factors for slave (SFS) and master (SFM) – increase this if contact
not stiff enough, but must take care as model can become unstable.
• Optional thickness for slave (SST) and master (MST) – overrides the true shell thickness
for contact calculations only.
• Scale factor for slave (SFST) and master (SFMT) surface thickness – scales the true shell
thickness for contact calculation only.
back to contents
LS-DYNA ENVIRONMENT
174

O
N
Creating a contact – optional cards: T
Card A A
C
T
S
• SOFT constraint option can be changed from the default penalty formulation (=0) to:
• (=1 & =2) – the contact stiffness will be calculated based on stability using time-step
and nodal masses, could be used for contact between dissimilar material stiffnesses
or mesh densities.
• (=2) – turns the node-to-segment formulation into segment-to-segment formulation.
back to contents
LS-DYNA ENVIRONMENT
175

O
N
Card B DP A
C
T
S
PENMAX*t
t/2 0 t/2
t
• PENMAX – defines maximum penetration allowed – different use
for different contacts:
- for example for *CONTACT_AUTOMATIC_SINGLE_SURFACE
the default is 0.4, which means 40% of segment thickness.
• THKOPT and SHLTHK – for old type contacts.
back to contents
LS-DYNA ENVIRONMENT
176

O
N
Card C A
C
T
S
• IGNORE – ignore initial penetration – more about this later.
back to contents
LS-DYNA ENVIRONMENT
177
Contact Problems – Contact not stiff enough LS-DYNA Introductory Course C

O
F N
When a contact is not stiff enough and parts F2 T
start to penetrate through each other, there are A
F1
two main approaches to fixing the problem: C
F
k F = k * DP T
1. Penalty Stiffness change: S
• Increase penalty stiffness factors k1 k

SFS/SFM (F1) d
• Or, if dealing with a rigid or null material,
increase Young‟s modulus (F1) DP
2. Contact Thickness change: F

• New thickness which overrides true
thickness SST/MST (F2)
• Scale true thickness SFST/SFMT (F2) DP
t/2 t/2 0 t/2 t/2

back to contents
LS-DYNA ENVIRONMENT
178
Contact Problems – Edge contact LS-DYNA Introductory Course C

O
If no slave nodes are detected when they are projected on to the master surface, N
as in the examples below, penetrations will occur: T
A
C
Master T
Edge contact S
Slave
Solve by:
Using SOFT = 2 on the *CONTACT card, which forces segment-based searching. This type
of search picks up the slave surface passing through the master surface, using segment-vs-
segment rather than node-vs-segment searches. Typically SBOPT=3 and DEPTH=5 are
also set, so that both surface penetrations and edge-to-edge penetrations are checked.
Only works with *AUTOMATIC_SURFACE_TO_SURFACE and *AUTOMATIC_SINGLE_
SURFACE.
back to contents
LS-DYNA ENVIRONMENT
179
Contact – Initial Penetrations and Crossed Edges LS-DYNA Introductory Course C

O
N
• The quality of the mesh in terms of initial penetrations and crossed edges is T
important for the purpose of how the contacts will work in the model. A
• The difference between an initial penetration and a crossed edge is shown below. C
T
• There is a work around for initial penetrations, as shown in the next slide;
S
however, crossed edges need to be corrected before the model is run.
Initial
Penetration Crossed Edge
back to contents
LS-DYNA ENVIRONMENT
180
Contact – Initial Penetrations LS-DYNA Introductory Course C

O
• If the model contains initial penetrations, these will affect the contacts. N
• By default LS-DYNA moves the nodes in order to remove the initial penetrations. T
• An alternative is to set parameter IGNORE=1 (available for Automatic Contact types). A
C
• This can be set on *CONTROL_CONTACT (card 4) for all contacts in the model T
or selectively on each contact (Optional Card C). S
Default Contact IGNORE=1

The penetrating node is The node is not moved, the contact
projected back to outer edge of thickness is reduced for the penetrating
the contact surface node, such that is doesn‟t penetrate.
(other nodes not initially penetrating still see
the full thickness)
If later on the node moves away from the surface, the full contact depth is reinstated.
Note: IGNORE=2 gives the same result as 1 but outputs more info – see manual
back to contents
LS-DYNA ENVIRONMENT
181
Contact – Monitoring contact forces LS-DYNA Introductory Course C

O
N
• The contact force time histories can be output in the RCFORC file.
T
(Surface-to-surface and nodes-to-surface contacts only) A
• The forces on particular parts in single surface contacts can be measured using C
the *CONTACT_FORCE_TRANSDUCER_PENALTY card. T
S
– Does not generate any forces,

only measures the forces from
other contacts.
– A *DEFINE_BOX can be used to
identify clearly which contact
event is being measured. Ffront Fback
– Example will measure forces on
front of blue component only.
selected segments are green – if
back segments were included as
well, total force might sum to
zero.
back to contents
LS-DYNA ENVIRONMENT
182
Contact – Extent of contact surface LS-DYNA Introductory Course C

O
• No problem to put whole model in one contact - recommend using N
T
*CONTACT_AUTOMATIC_SINGLE_SURFACE
A
• A *DEFINE_BOX can be used to reduce contact size and save CPU: C
T
S
• Can supplement one large contact with local contacts with different properties,
but avoid having two or more contacts generating forces due to the same
penetration.
back to contents
LS-DYNA ENVIRONMENT
183
Contact – Extent of contact surface LS-DYNA Introductory Course C

O
N
• Important to exclude certain parts: T
A
– Spotweld beams, bushes, etc - most beam elements, springs. C
T
S
– Any parts made of solid elements that can fail (e.g. aluminium castings).
These need to use *CONTACT_ERODING… as this will recalculate the
contact surface after element failure.
– Deformable barriers, dummies, airbags – these are generally set up with

their own internal contacts with special settings.
• Can define by Excluded Part Set (SSTYP=6).

• A node can be in more than one contact, but should not be in slave and master
of the same contact.
back to contents
LS-DYNA ENVIRONMENT
184
Contact – Checking contact definitions LS-DYNA Introductory Course C

O
N
• In Pre-processor: T
– Try to check for penetrations and crossed elements while A
creating the mesh, e.g.: C
T
S
• In Oasys PRIMER:
– Sketch contact, check that the right parts are covered.
– Initial penetration and crossed element check.
– Automatic correction.
back to contents
LS-DYNA ENVIRONMENT
185

O
N
• After initialising in LS-DYNA: T
A
– Check .otf file for messages for initial penetrations. C
T
– Check contact timestep (not needed for SOFT=1 & 2): timestep S
reported should be greater than model timestep or instability could
occur during the run.
– Magnify displacements x100 or x1000, plot model after 5 or 10
cycles, look for deformation caused by initial penetrations.
• Fixing initial penetrations:
– Best option – fix mesh to eliminate.

– If in a hurry use IGNORE=1.
back to contents
LS-DYNA ENVIRONMENT
186

O
N
• Contact energies are written to the following files: T
– GLSTAT - Sliding interface energy A
– SLEOUT - Slave energy, Master energy, Frictional energy C
T
• Contact energy in a contact spring = ½*(contact force)*(penetration) S
= ½*fcon*Dp = ½*kcon*Dp2 =½*(contact stiffness)*(penetration)2
• Total contact energy = sum of contact energies in all the contact springs.
• Contact forces are written to the RCFORC file (need to request with
*DATABASE_RCFORC).
• Contact forces & energies are calculated incrementally every timestep.
• It is advisable to always switch FRCENG=1 on *CONTROL_CONTACT.
This will help in understanding the magnitude of frictional contact energy &
contact energy due to contact forces.
• Contact energies are non-physical. For the purpose of stability, it is
advisable that ratio of the total contact energy to the deformation energy
should be small (<5%). Reducing the magnitude of penetration reduces the
contact energy.
back to contents
LS-DYNA ENVIRONMENT
187
Contact Problems LS-DYNA Introductory Course C

O
N
Contacts are now reasonably reliable, but in case of problems... T
A
• Identify the time at which the problem occurs - look for steps in contact energy
C
in GLSTAT (all contacts) and SLEOUT (individual contacts) files. T
• Request energy output using: S
– *CONTROL_ENERGY
– *DATABASE_GLSTAT and *DATABASE_SLEOUT
• This is non-physical energy stored in the contact

“springs” and should be small compared with
deformation energy.
• Also plot internal energy of all parts by adding
*DATABASE_MATSUM and identify parts
showing biggest steps.
back to contents
LS-DYNA ENVIRONMENT
188
Contacts types and usage LS-DYNA Introductory Course C

O
N
• The following table presents a summary of commonly used sliding T
contacts and their application for various contact situations. A
– Beam to Beam C
– Beam to Shell Edge T
– Beam to Shell Surface S
– Shell Edge to Shell Edge (External edges)
– Shell Segment to Segment
– Shell Edge to Shell Edge (Internal edges)
• The examples were run with SOFT values of 0, 1 and 2
• The following results are reported:
 PASS - Contact worked
 FAIL - Contact failed
 Reverted to SOFT=1 - Contact switched to SOFT=1 internally by LS-DYNA
• The examples were run in R6.0, revision 70232, MPP, single precision.
back to contents
LS-DYNA ENVIRONMENT
189
Contacts types and usage LS-DYNA Introductory Course
Contact SOFT Beam to beam Beam to shell Beam to surf Edge to edge 1 Segm to segm Edge to edge 2
AUTOMATIC_ 0 PASS PASS PASS PASS PASS FAIL

GENERAL
1 PASS PASS PASS PASS PASS FAIL
2 Reverted to SOFT=1 Reverted to SOFT=1 Reverted to SOFT=1 Reverted to SOFT=1 Reverted to SOFT=1 FAIL
AUTOMATIC_ 0 PASS PASS PASS PASS PASS FAIL

GENERAL_
1 PASS PASS PASS PASS PASS FAIL
INTERIOR
2 Reverted to SOFT=1 Reverted to SOFT=1 Reverted to SOFT=1 Reverted to SOFT=1 Reverted to SOFT=1 FAIL
AUTOMATIC_ 0 FAIL FAIL FAIL FAIL FAIL FAIL

SURFACE_TO_
SURFACE 1 FAIL FAIL FAIL FAIL FAIL FAIL
2 FAIL FAIL FAIL PASS PASS PASS
AUTOMATIC_ 0 FAIL FAIL PASS FAIL FAIL FAIL

SINGLE_SURFACE
1 FAIL FAIL PASS FAIL FAIL FAIL
back to contents
LS-DYNA ENVIRONMENT
190
Contacts types and usage LS-DYNA Introductory Course
Contact SOFT Beam to beam Beam to shell Beam to surf Edge to edge 1 Segm to segm Edge to edge 2
AUTOMATIC_ 0 FAIL PASS PASS FAIL FAIL FAIL
BEAMS_TO_
SURFACE 1 FAIL PASS PASS FAIL FAIL FAIL
2 FAIL PASS PASS FAIL FAIL FAIL
0 FAIL FAIL FAIL PASS FAIL FAIL

SINGLE_EDGE
1 FAIL FAIL FAIL PASS FAIL FAIL
2 FAIL FAIL FAIL Reverted to SOFT=1 FAIL FAIL

AIRBAG_SINGLE_
SURFACE
back to contents
LS-DYNA ENVIRONMENT
191
Workshop Example 7 LS-DYNA Introductory Course C

O
N
1. Animate the results of analysis „contact‟ and plot the contact surface force T
magnitude in T/HIS. A
2. Edit the keyword file and examine the contact surface data - how would you C
make the contact work properly in this example? Look at the material data T
for each part. S
3. Copy contact.key into contact1.key and try to fix the contact.

4. Run the fixed model, if it is still bad then try other methods (for example
SOFT=1, null shells coating, penalty stiffness factor, contact thickness).
5. Animate the results of analysis „contact1‟ and plot the new contact surface
force magnitude in T/HIS.

back to contents
LS-DYNA ENVIRONMENT
192
Tied Contacts LS-DYNA Introductory Course C

O
N
• Tied contact – ties the slave side of the contact to the master side. T
• Tied contacts can use either a constraint or a penalty formulation. A
• Useful for solid mesh transitions, modelling welds etc. C
T
• Slave nodes must be “close” to be TIED – close defined as: S
D < 0.6 * (master thickness + slave thickness)
• If the node is too far away it will not be tied and a warning will be printed in the
d3hsp file (.otf file) – example below.
*** Warning Slave node is not constrained
since it is too far from segment.
tied interface # = 1
slave node # = 220
segment # = 126
distance from segment = 0.2009796E+01
segment nodes:
118 165 164 113
back to contents
LS-DYNA ENVIRONMENT
193

O
N
• Several types of TIED contact available:
T
– *CONTACT_TIED_NODES_TO_SURFACE – constrains translational DOF only. A
– *CONTACT_TIED_SURFACE_TO_SURFACE – constrains translational DOF only. C
– *CONTACT_TIED_SHELL_EDGE_TO_SURFACE – (TSE2S) constrains T
translational AND rotational DOF. S
– *CONTACT_AUTOMATIC_..._TIEBREAK - allows defined failure forces/stresses.
– *CONTACT_SPOTWELD (alternative name for TSE2S).
• Important considerations when using the default definition:
– Basic tied contacts use a constraint formulation and cannot be used with rigid parts
or other constraints.
– Slave nodes are projected onto the master segment at initialisation altering geometry
if D < 0.6 * (master thickness + slave thickness):
Slave
D < 0.6*(TS + TM)
Master
back to contents
LS-DYNA ENVIRONMENT
194

O
N
Additional options are available for TIED contact types, which maintain the offset of T
slave nodes (i.e. do not alter geometry): A
C
– _OFFSET – maintains offset and switches to penalty formulation. T
S
– _BEAM_OFFSET – maintains offset and switches to penalty formulation

(TSE2S only); the name „BEAM‟ is due to the fact that it uses „artificial
beams‟ in the contact calculation, which allows for moment transfer, i.e. it is
more accurate for the offset option (and so it is not limited to beams).
– _CONSTRAINED_OFFSET – constraint formulation (use with TSE2S only),

moments are computed.
back to contents
LS-DYNA ENVIRONMENT
195
Recommended contact types & settings LS-DYNA Introductory Course C

O
Application Contact type Special Parameters
N
T
A
General model *CONTACT_AUTOMATIC_SINGLE_SURFACE SOFT=1
C
Soft foams
*CONTACT_AUTOMATIC_SINGLE_SURFACE + *MAT_NULL properties: T
Cover solids with null shells 2
E=3000N/mm , thk=2mm S
Model with many
*CONTACT_AUTOMATIC_SINGLE_SURFACE IGNORE=1
initial penetrations
SOFT=2, SBOPT=3,
*CONTACT_AUTOMATIC_SINGLE_SURFACE or DEPTH=5
Edge contact
*CONTACT_AUTOMATIC_GENERAL (no SOFT
problems Apply selectively in problem
option required)
areas
Deformable
*CONTACT_AUTOMATIC_SURFACE_TO_SURFACE As above
barrier to model
Beams or other *CONTACT_AUTOMATIC_NODES_TO_SURFACE or
non-continuous *CONTACT_AUTOMATIC_GENERAL (considers cross
mesh sectional area of beams along its length)
SOFT=2, DEPTH=3
Folded airbag *CONTACT_AIRBAG_SINGLE_SURFACE BSORT = 4 or 1 in extreme
cases
back to contents
LS-DYNA ENVIRONMENT
196

O
N
This example demonstrates the usage of the tied contact. T
1. The model has been already run, visualize the results in directory 01 and A
find out if the contact of the T-junction is working ok. C
T
2. How many of the nodes are failing the tied contact? Use .otf file and search
S
for „warning‟.
3. Make the necessary modifications to correct the tied contact into a new
folder and re-run the model.
4. Consider alternative ways of fixing the problem.

back to contents
LS-DYNA ENVIRONMENT
197
Outline – THEORY AND PRACTICE LS-DYNA Introductory Course R

I
G
1. Time Integration 4. Materials I
• Description of explicit time 5. Contact Surfaces D
integration
• Types of contact B
integration • Contact method (depth, stiffness) O
D
I
• Summary of cost considerations • Other contact issues E
• Contact guidelines S
6. Rigid Bodies
• Meshing Notes
3. Hourglassing
back to contents
LS-DYNA ENVIRONMENT
198
Rigid Bodies LS-DYNA Introductory Course R

I
G
• An LS-DYNA part becomes a rigid body when the material model is defined as: I
*MAT_RIGID (Dyna material type 20). D
B
• Elastic modulus and Poisson ratio are only used for contact stiffness. O
• Restraint conditions: use parameters CMO, CON1, CON2. D
• Restraints are applied at the centre of mass. I
E
• Mass, CofG and inertia tensor are automatically calculated from element volumes S
and material density - these can be overridden by using *PART_INERTIA
• Another type of rigid: *CONSTRAINED_NODAL_RIGID_BODY
• Applications:
• non-deforming objects such as engines and gearboxes,
• connections between components,
• spreading load over several nodes to avoid hourglassing.
back to contents
LS-DYNA ENVIRONMENT
199
Rigid Bodies - Theory LS-DYNA Introductory Course R

I
The forces on the rigid body are summed and applied at the CofG. G
I
Note: giving the rigid bodies a realistic mass is important for the equations of D
motion – if the mass and inertia of rigid parts are too small it can cause problems.
F1 F6 B
O
F5 Fg
M1
Which is The geometry D
F2 equivalent to: can be updated: I
Mg E
Fg  Fi S
M2
F3
F4 M g  M i  (ri  Fi )
v3 v4
a  Fg / m The motion v5
v imposed on v
vt dt / 2  vt dt / 2  a  dt
the nodes:
θ   M g / I θ
v2
t dt / 2  t dt / 2   dt v1 vi  v    ri
back to contents
LS-DYNA ENVIRONMENT
200
Rigid Bodies - Incompatibilities LS-DYNA Introductory Course R

I
G
As rigid bodies work by imposing motion on nodes, the following features I
which also impose motion should not be applied to nodes on rigid bodies: D
B
– node restraints *BOUNDARY_SPC_NODE (use *MAT_RIGID
O
restraints instead). D
– multiple rigid bodies on same node, e.g. two rigid parts meshed I
together. E
– constraint types of contact, i.e.: S
– *CONTACT_TIED_SURFACE_TO_SURFACE
– *CONTACT_TIED_NODES_TO_SURFACE
– *CONTACT_TIED_SHELL_EDGE_TO_SURFACE
– *BOUNDARY_PRESCRIBED_MOTION_NODE (use instead
*BOUNDARY_PRESCRIBED_MOTION_RIGID).
back to contents
LS-DYNA ENVIRONMENT
201
Rigid Bodies - Contact Surface Stiffness LS-DYNA Introductory Course R

I
G
Material properties of rigid bodies (E and ν) are used to calculate the contact I
stiffness. In some cases these values may need to be altered. A general rule is: D
B
– Start with rigid body Young‟s Modulus equal to that of the deformable
O
elements contacting the body. D
I
– If the rigid elements are bigger (orders of magnitude) than the E
deformable elements scale down the rigid body Young‟s modulus in S
proportion to the element size to avoid contact instability.
(SEE FIX FOR EXAMPLE 7)
back to contents
LS-DYNA ENVIRONMENT
202
Rigid Bodies - Applications LS-DYNA Introductory Course R

I
G
Rigid Parts (i.e. Parts that use *MAT_RIGID) can for example be used as connections: I
D
Make patches of elements A, B, C, D rigid (each has its own PID) and use rigid body
mergers (*CONSTRAINED_RIGID_BODIES) A-C and B-D. The advantage is that the seat B
can be replaced and the connections will work straight away, if the new seat has rigid O
patches A and B defined. D
I
E
S
back to contents
LS-DYNA ENVIRONMENT
203
Rigid Bodies - Common Pitfalls LS-DYNA Introductory Course R

I
G
There are some common pitfalls I
that need to be avoided: D
PID1
B
• Two Rigid Parts, which are meant to
O
be separate, have the same Part ID D
and so are constrained together. PID1
I
E
S
MAT1
• Different parts reference the same
rigid material - if material has
restraints, can lead to unwanted
restraint conditions on some of the
parts. MAT1
back to contents
LS-DYNA ENVIRONMENT
204
Rigid Bodies - Common Pitfalls LS-DYNA Introductory Course R

I
G
There are some common pitfalls that need to be avoided: I
D
• Attempt to make a rigid body contact a rigidwall. Rigid bodies can interact
B
with fixed rigid walls if RWPNAL is set on the *CONTROL_CONTACT.
O
D
• The TIED type contacts do not work in conjunction with rigid bodies, since I
they are constraint based – use the _OFFSET option on these contacts E
(penalty based). S
• When modelling linkages/joints, failure to define sensible cross sectional

area, leading to very low mass so that joints are not stiff enough.
back to contents
LS-DYNA ENVIRONMENT
205
Workshop Example 9 LS-DYNA Introductory Course R

I
G
The model shown on the right I
needs completing before the upper fork D
wheel will roll up the ramp,
compressing the strut, as it B
does so.
lower fork O
D
wheel I
Firstly, a contact surface needs E
to be defined so that the wheel S
is aware of the ground – use
*CONTACT_AUTOMATIC_
SURFACE_TO_SURFACE.
pivot ground
(Units: mm, s, t, N, Mpa)
Refer to example10.key to
check the answer
back to contents
LS-DYNA ENVIRONMENT
206
Outline – THEORY AND PRACTICE LS-DYNA Introductory Course B

O
U
1. Time Integration 4. Materials N
• Description of explicit time 5. Contact Surfaces D
integration A
• Types of contact R
integration • Contact method (depth, stiffness) Y
C
• Summary of cost considerations • Other contact issues O
• Contact guidelines N
2. Element Formulation D
6. Rigid Bodies I
T
I
– Shell – Discrete 8. Connections O
– Thick Shell – Other N
9. Control and Output S
• Meshing Notes
3. Hourglassing
back to contents
LS-DYNA ENVIRONMENT
207
Boundary Conditions – Initial Velocity LS-DYNA Introductory Course B

O
U
Most impact models contain an initial velocity (i.e. the velocity just before impact).
N
There are two main ways of setting the initial velocity:
D
1. *INITIAL_VELOCITY – will specify a translation & rotation on nodes. A
2. *INITIAL_VELOCITY_GENERATION – specify an initial rotation and R
translation of a part Y
C
*INITIAL_VELOCITY *INITIAL_VELOCITY_GENERATION
O
N
D
I
T
I
O
N
S
back to contents
LS-DYNA ENVIRONMENT
208
Boundary Conditions – Initial Velocity LS-DYNA Introductory Course B

O
U
*INITIAL_VELOCITY N
• Sections of the model defined by node sets or boxes identify nodes D
that have the velocity applied to them. A
R
Y
C
O
N
D
I
• When changing the velocity of a whole model, a very common mistake T
is to change *INITIAL_VELOCITY and forget that initial velocities on I
rigid parts can be defined with *PART_INERTIA. O
N
• IRIGID allows velocities applied to rigid body nodes by
S
*INITIAL_VELOCITY to overwrite those on *PART_INERTIA.
• Use IRIGID option set to -2.
back to contents
LS-DYNA ENVIRONMENT
209
Boundary Conditions - Loading LS-DYNA Introductory Course B

O
Nodal Loads Pressure Loads U
N
D
A
R
Y
C
O
• Loads can vary arbitrarily with time and can be applied as:
N
– Concentrated nodal loads: *LOAD_NODE_OPTION D
– Surface pressure loads: *LOAD_SHELL_OPTION I
*LOAD_SEGMENT_OPTION T
– Gravity is applied using: *LOAD_BODY_OPTION I
O
Force
Load Curve N
• Note that isolated nodal loads tend to promote hourglassing.
S
• Also note that when *DEFINE_CURVE is used for defining the
loading, the curves are not extrapolated (as they are when used
in the definition of materials) – instead, function values are set to
zero if the abscissa goes off scale. time
back to contents
LS-DYNA ENVIRONMENT
210
Boundary Conditions - Prescribed Motion LS-DYNA Introductory Course B

O
U
Define an imposed nodal motion using keyword: N
*BOUNDARY_PRESCRIBED_MOTION_OPTION1 – where the option can be D
prescribed on a node, node set or a rigid body: A
_NODE R
_SET Y
_RIGID
C
O
Prescribed motion is defined via a load curve (*DEFINE_CURVE), and can
N
prescribe one of the following: D
- Velocity I
- Acceleration T
- Displacement I
O
N
S
back to contents
LS-DYNA ENVIRONMENT
211
Boundary Conditions – Restraints LS-DYNA Introductory Course B

O
U
For deformable parts, define nodal single point constraints using keyword:
N
*BOUNDARY_SPC_OPTION1 – where the option can be prescribed on a node or a node D
set. A
To output the reaction forces, there are two options: R
1. Add *DATABASE_SPCFORC card and the individual nodal reaction forces will be Y
output in the SPCFORC file.
2. Use *DATABASE_NODAL_FORCE_GROUP and define the node groups for which to C
output forces; add *DATABASE_NODFOR card and resultant for each group will be O
reported in the NODFOR file. N
D
F I
T
The resultant forces are reported for each of I
the groups defined, as well as the individual O
nodes in the group. N
R
R R S
Note: The restraints on rigid parts (parts using rigid materials) are applied on the
*MAT_RIGID using CMO, CON1 and CON2 parameters.
back to contents
LS-DYNA ENVIRONMENT
212
Boundary Conditions - Rigidwalls LS-DYNA Introductory Course B

O
U
Simple way of treating contact between a rigid surface and nodes in the model: N
D
• Keyword *RIGIDWALL A
R
• Various shapes of rigidwall available, but Planar and Flat versions are Y
most common;
• Shaped walls can be built up from several rigidwalls; C
O
• A geometric (constrained) definition is used to determine if the node
N
„contacts‟ the wall – this is not the generic penalty stiffness method of the D
conventional contacts *CONTACT_ but it is a purely geometric method I
and can sometimes introduce noise. T
I
O
N
Planar S
rigidwall
Geometric
rigidwall
back to contents
LS-DYNA ENVIRONMENT
213
Boundary Conditions - Rigidwalls LS-DYNA Introductory Course B

O
U
• Rigidwalls are NOT the same as contact surfaces – they use no contact N
stiffness. D
• Penetrating slave nodes are simply moved back to the rigidwall surface – can A
R
give very „rigid‟ and abrupt response. Y
• Rigid parts can make contact only if RWPNAL parameter is set on
*CONTROL_CONTACT . C
O
• Can use *DEFINE_BOX to reduce number of slave nodes. N
• Planar rigidwall can be given mass and initial velocity. D
I
• Geometric rigidwall can be given prescribed motion. T
• Rigidwalls allow frictionless sliding, a friction coefficient or a „stick‟ condition. I
O
• Sometimes easier/better to use a contact surface on a plane of rigid elements N
rather than a rigidwall. S
back to contents
LS-DYNA ENVIRONMENT
214
Workshop Example 10 LS-DYNA Introductory Course B

O
U
1. To drive the wheel and strut Units: mm, s, t, N, MPa N
assembly we will force the upper fork D
upper fork to move with the A
following velocity. Use the R
*BOUNDARY_PRESCRIB lower fork Y
ED_MOTION_RIGID:
C
wheel O
Time Velocity
(sec) (mm/s) N
D
0 0
I
0.005 -2000 T
I
1 -2000
O
pivot N
2. Restrain the ground so that ground S
it cannot move. The
ground is rigid, so apply the
restraints to the
*MAT_RIGID card.
LS-DYNA ENVIRONMENT
215

O
N
1. Time Integration 4. Materials N
• Description of explicit time 5. Contact Surfaces E
integration C
integration • Contact method (depth, stiffness) I
O
N
• Summary of cost considerations • Other contact issues S
6. Rigid Bodies
• Meshing Notes
3. Hourglassing
back to contents
LS-DYNA ENVIRONMENT
216
Connections LS-DYNA Introductory Course C

O
N
• Many different types of connections: N
E
– Spotwelds/Rivets C
T
– Welds
I
– Bolts and Mounts O
– Adhesives N
S
– Joints (e.g. Universal joints)
• Many different methods of modelling these connections:

– Spotwelds
– Mesh connectivity
– Discrete Beams
– Springs and Dampers
– Contact surfaces
– Constraints
back to contents
LS-DYNA ENVIRONMENT
217
Rigid Connections LS-DYNA Introductory Course C

O
1. Rigid body to Rigid Body: Slave rigid part is merged with the master rigid part. Inertial N
properties of merged rigid body is the combination of all of the merged rigid parts, N
unless master rigid part has its inertial properties explicitly defined (*PART_INERTIA). E
All restraints, initial velocities, etc. of master rigid part are applied to whole merged C
rigid body. T
*CONSTRAINED_RIGID_BODIES
A I
PIDM PIDS IFLAG O
B N
S
2. Rigid body to Deformable body:
*CONSTRAINED_EXTRA_NODES_OPTION
A
PID NID/NSID IFLAG
3. Deformable mesh to Deformable mesh:

*CONSTRAINED_NODAL_RIGID_BODY_{OPTION}_{OPTION}_{OPTION}
PID CID NSID PNODE IPRT DRFLAG
CMO CON1 CON2

back to contents
LS-DYNA ENVIRONMENT
218
Connections – Spotwelds LS-DYNA Introductory Course C

O
N
N
• Spotwelding is one of the oldest welding E
processes and it is used widely for C
connecting steel sheet body panels. T
• The metal is locally melted and allowed to I
fuse, thus creating a connection between the O
N
panels.
S
• In order to accurately model the spotweld
connections, the following ideally should be
captured:
– Grain structure and zones in weld
region
– Failure modes: Peel, Shear, Bending
– Material behaviour in strain rates
– Damage consideration
back to contents
LS-DYNA ENVIRONMENT
219

O
N
Tests on spotwelds show a very ductile failure, due to local deformation of the N
panels around the spotweld. Explicitly modelling this localised deformation in E
the panels would require very small elements, so ductility is represented in the C
spotweld element (beam or solid element – see next slide). T
I
Test O
N
S
CAE
back to contents
LS-DYNA ENVIRONMENT
220

O
N
N
• Detailed FE models of spotwelds can be E
created by using a very small mesh size. Modeling of different grains & C
temperature zones
• However, this is not practical for spotwelds T
in a full crash vehicle model. I
O
• Some of the most used simplified modelling N
techniques are: S
– Beams elements
– Single solid element
– Solid nugget (cluster of multiple solids)
Beam element ELFORM #9 Solid element ELFORM #1 Solid nugget ELFORM #1
back to contents
LS-DYNA ENVIRONMENT
221

O
N
Typical Spotweld Shear and Peel Test Results N
E
C
Failure at 5500N
T
I
O
Yield at 3500N N
S
NB peel failure force < shear failure force
Shear Test
Peel Test
back to contents
LS-DYNA ENVIRONMENT
222

O
Beam spotwelds N
N
• By tradition beam elements have been the most used option for modelling spotwelds. E
However, they can pose issues, as for example in a shear situation: C
T
F F Spotweld test data
I
Force
Rotating beam: Shear O
forces are more N
tensile as load S
F F increases (not Tension
pure shear any
Displacement
more)
• Tensile failure input is a problem - could cause premature failure in shear.

• Not able to model tensile failure correctly without causing premature failure in shear.
• It has been generally found to be best if the tensile failure force is set to the same value
as the shear failure force.
• Additional problem with beam spotwelds: they are dependent of the location of the beam
on the shell element – see also *CONTROL_SPOTWELD_BEAM
back to contents
LS-DYNA ENVIRONMENT
223
Spotwelds – Beam elements LS-DYNA Introductory Course C

O
Beam spotwelds N
N
E
Spotweld
Created using a preprocessor C
Element
T
Spotweld I
Beam: ELFORM=9
Section O
N
Spotweld
*MAT_SPOTWELD (*MAT_100) S
Material
*CONTACT_TIED_SHELL_EDGE_TO_SURFACE or
Spotweld
*CONTACT_SPOTWELD or
Contact
*CONTACT_TIED_SHELL_EDGE_TO_SURFACE_OFFSET
• Beam spotwelds are easier to set up, but are not as accurate as solid spotwelds.
• Use *ELEMENT_BEAM_PID - independent of panel mesh, give 2 parts ID‟s connected
by the weld; the _PID option only works with type 9 beams.
• In the contact definition, the nodes on the spotweld beam are the slave nodes, while the
panels to be welded are the master segments. Also, it is advisable to define one single
contact to include all spotwelds and panels to be welded.
back to contents
LS-DYNA ENVIRONMENT
224
Spotwelds – Solid elements LS-DYNA Introductory Course C

O
Solid spotwelds N
N
E
Spotweld Created using a preprocessor:
C
Element one solid or nugget
T
Spotweld I
Solid: ELFORM=1
Section O
N
Spotweld
*MAT_SPOTWELD (*MAT_100) S
Material
Spotweld *CONTACT_TIED_SURFACE_TO_SURFACE or
Contact *CONTACT_TIED_SURFACE_TO_SURFACE_OFFSET
• The solid spotwelds are more accurate than beams since they have 4 nodes that can be
tied.
• Use brick solids ELFORM=1 on the *SECTION_SOLID card with hourglass control type 6
QM=1.
• In the contact definition, the nodes on the spotweld solid are the slave nodes, while the
panels to be welded are the master segments. Also, it is advisable to define one single
contact to include all spotwelds and panels to be welded.
back to contents
LS-DYNA ENVIRONMENT
225

O
N
• *MAT_SPOTWELD_{OPTION} [100] – the option can be <BLANK> or set to DAMAGE-
N
FAILURE
E
• This material applies to beam element type 9 and solid element type 1 with type 6 C
hourglass control. T
• Allows input of basic parameters such as density, Young‟s modulus, Poisson‟s ratio, initial I
yield stress and hardening modulus; use low Young‟s modulus (typically 10x lower than O
steel to reduce added mass from mass-scaling) N
• Can set the timestep of spotweld elements by mass-scaling (DT) – if this is used, the S
added mass is reported after the first cycle.
• Failure can be incorporated based on:
– plastic failure strain, or
– force/moment resultants or stress using the option DAMAGE-FAILURE - Nonzero „RS‟
(rupture strain) on card 3, invokes damage mechanics combined with the plasticity
model to achieve a smooth drop off of the resultant forces prior to removal of the
spotweld.
• The „OPT‟ parameter determines the method used in computing resultant based failure,
which is unrelated to damage.
back to contents
LS-DYNA ENVIRONMENT
226
Connections – Welds LS-DYNA Introductory Course C

O
N
There are several methods of modelling weld areas – here are some: N
E
• *CONSTRAINED_GENERALIZED_WELD_OPTION. C
• Allows definition of spotwelds, fillet and butt welds by applying a constraint T
between a set of nodes. Weld failure treated as ductile (plastic strain limit at I
nodes) or brittle (stress levels). Not generally used since it is difficult to O
N
define and post-process.
S
back to contents
LS-DYNA ENVIRONMENT
227
Connections – Welds LS-DYNA Introductory Course C

O
N
Other methods of modelling weld areas: N
• Model explicitly with solid elements; necessary only if accurate and detailed E
behaviour of connection is required; generally needs very fine mesh with many C
small elements. T
I
• Continuous mesh – rather than specifically modelling the weld, simply connect O
the components together via the mesh and monitor the stresses/strains in the N
weld area. S
• Using nodal rigid bodies.
• If edge nodes not aligned, can use *TIED_SHELL_EDGE_TO_SURFACE.
back to contents
LS-DYNA ENVIRONMENT
228
Connections – Bolts and Mounts LS-DYNA Introductory Course C

O
There are a number of different methods for modelling bolts and mounts – here are some: N
N
• Explicitly with solid elements – necessary only if accurate and detailed behaviour of
E
connection is required; generally needs many small elements, with fine mesh on both
C
sides of the connection;
T
• Deformable beam I
• Discrete beam – short beam element connecting two nodal rigid body NRB spiders; O
elastic, plastic, viscous and failure properties can be defined for all 6 dofs; N
• Spring and damper elements – discrete element connecting two nodes. Elastic, S
plastic, viscous and failure properties can be defined; can be oriented to work in a
particular vector/plane, but interaction between springs is not possible;
• Nodal rigid bodies (NRB);
Beam Discrete NRB

beam
back to contents
LS-DYNA ENVIRONMENT
229
Connections – Discrete Beams *ELEMENT_BEAM LS-DYNA Introductory Course C

O
N
Discrete beams can be used when translational and rotational stiffness information for bolts
N
is available, in conjunction with materials 66-67-68; it uses 6 linear or nonlinear springs to
E
define stiffness in 6 degrees-of-freedom.
C
T
Force/Moment
T
I
O
N
T *MAT_066
Displacement S
Force/Moment
T
+
*MAT_067
Displacement
R
Force/Moment
R
*MAT_068
R Displacement with failure
back to contents
LS-DYNA ENVIRONMENT
230

O
N
• The discrete beam element can have zero or finite length.
N
• In section card *SECTION_BEAM must define the following: E
• ELFORM – element Type 6 = discrete beam. C
• CID – refer to a coordinate system ID to orient the beam element (define this T
beforehand using *DEFINE_COORDINATE. I
• If finite length beam element, set SCOOR to 2 for correct beam orientation – the O
coordinate system will follow the beam tracking of its rotations. N
S
• Define beam volume on VOL to give mass – important for timestep.
• Define beam mass moment of inertia – required for rotational stability of discrete beam.
• It is recommended that Discrete beam be used whenever possible, especially if
orientation is specified. The discrete beam tends to be more accurate and cost
effective. *ELEMENT_DISCRETE option is no longer being developed.
back to contents
LS-DYNA ENVIRONMENT
231

O
N
For a non zero length beam the SCOOR = -3, -2, 2, or 3 options should be used. N
SCOOR = -2 (2 is the same but centred on N2 not N1) below: N2
E
C
The angular velocity on N1 rotates triad, but T
the r-axis follows the line between N1 and N2. I
O
N2
N
S
N1
r
N2
s
N1
r
Initial State s
N1
Deformed State
back to contents
LS-DYNA ENVIRONMENT
232
*Section_Beam - SCOORBeams
options
*ELEMENT_BEAM LS-DYNA Introductory Course C
O
N
SCOOR = -3 (3 is the same but centred on N2 not N1) bellow: N
N2
E
C
The angular velocity on N1 rotates triad. r T
I
N2 O
s N
N1 S
r
N2
s
N1
Initial State
N1
Deformed State
back to contents
LS-DYNA ENVIRONMENT
233
Connections – Discrete Beams *ELEMENT_BEAM

Simple Example LS-DYNA Introductory Course C
O
N
In the case of the y or z shear loading it can be seen that as the beam rotates with N
SCOOR=-3 no axial force is generated whereas with SCOOR=-2 a force is E
generated. C
SCOOR = -3 SCOOR = -2 T
I
O
N
S
r
s
back to contents
LS-DYNA ENVIRONMENT
234
Connections – Springs and Dampers LS-DYNA Introductory Course C
*ELEMENT_DISCRETE O
N
• Define spring and damper translation and rotation elements using *SECTION_DISCRETE.
N
• The parameter DRO switches between translation and rotation. E
• Additional options on *SECTION_DISCRETE: C
– Clearance (CL) – for modelling gaps between components; compressive displacement T
that must be closed before spring acts, therefore compression-only. I
– Failure deflection in tension, compression or twist (FD, CDL, TDL). O
N
Force Force Deflection limit S
- tension
Clearance
Deflection
Deflection limit Deflection

- compression
Clearance Failure deflection

Once the *SECTION_DISCRETE is used, a material such as *MAT_SPRING_... or
*MAT_DAMPER_... must be chosen.
back to contents
LS-DYNA ENVIRONMENT
235
Constraints – Nodal Rigid Bodies NRB LS-DYNA Introductory Course C

O
N
Nodal Rigid Body – Simple way of rigidly linking N
nodes together to form a rigid „spider‟ E
• Keyword C
*CONSTRAINED_NODAL_RIGID_BODY T
I
• Use node sets to identify which slave nodes O
are constrained together. N
• Inertia of nodal rigid body can be defined, but S
taken from the nodal masses by default.
• Occasionally, default inertia is too small and
causes numerical instability – can either
define inertia or add mass to nodes.
• Can be used to connect parts in a
deformable mesh.
• NRB‟s have part ID‟s defined – keep this
unique to avoid clashing with other
component‟s ID‟s.
back to contents
LS-DYNA ENVIRONMENT
236
Constraints - Joints LS-DYNA Introductory Course C

O
Joints allow relative rotation between rigid parts. N
N
• Keyword: *CONSTRAINED_JOINT_OPTION_{OPTION}
E
• Input is simple - define nodes connecting the joint C
together. T
• Based on a penalty stiffness method. Stiffness of the I
joint is dependent on time step of the simulation. Check O
the stiffness in the ~.otf file. N
S
Joint types (to connect rigid bodies)

• SPHERICAL Commonly • TRANSLATIONAL_MOTOR
• REVOLUTE used joint • ROTATIONAL_MOTOR
types
• CYLINDRICAL • GEARS
• PLANAR • RACK_AND_PINION
• UNIVERSAL • CONSTANT_VELOCITY
• PULLEY
• TRANSLATIONAL
• SCREW
• LOCKING
back to contents
LS-DYNA ENVIRONMENT
237
Constraints – Joints LS-DYNA Introductory Course C

O
N
Cylindrical Joint Definition (e.g. Tube over a shaft) N
*CONSTRAINED_JOINT_CYLINDRICAL E
C
Node N1 Node N2 Node N3 Node N4 T
I
Body A O
N
Body B S
Coincidental nodal pairs
• Joints are generally defined with pairs of nodes – different joints need different
numbers of pairs defined.
• Cylindrical joint uses nodes in pair order: N1, N2,N3, N4.
• Nodal pairs should be coincidental.
• No resistance to sliding – if resistance is required change to _REVOLUTE type
or, add a discrete (spring) element.
back to contents
LS-DYNA ENVIRONMENT
238
Constraints - Joints LS-DYNA Introductory Course C

O
N
Checking Joint Stiffness N
Check the joint stiffness in the .otf file. Depends on rigid body mass and timestep. E
Stiffness is reported as a function of timestep: C
T
J o i n t s t i f f n e s s d a t a I
O
joint rigid body rigid body stiffness
no. no. 1 no. 2 * dt**2 scale
N
S
1 1 2 1.6224E-8
8
The .otf file gives: stiffness  dt  1.622 10
2
1.622 108
So, assuming dt  110 5
, the joint stiffness is: stiffness  5 2
 162 N / mm
(110 )
• Check stiffness against expected force, e.g. in this case 10kN would cause
10000/162 = 61.7mm displacement. The joint stiffness can be modified using
RPS (relative penalty stiffness) on *CONSTRAINED_JOINT, but this value
should not exceed 10. Better to increase the mass of the rigid bodies.
back to contents
LS-DYNA ENVIRONMENT
239
Workshop Example 11
9 LS-DYNA Introductory Course C
O
1. The pivot should be connected to N
the spindle with: N
*CONSTRAINED_JOINT_REVOL E
UTE spindle C
T
2. Two discrete elements have
I
already been defined between the
O
upper and lower forks. Add
N
section and material definitions so
S
that one element becomes a
spring, the other a damper.
*SECTION_DISCRETE
*MAT_SPRING_ELASTIC
*MAT_DAMPER_VISCOUS
For the spring, K=200N/mm
For the damper, use a constant of
10N/mm/s.
3. Also, the wheel needs attaching to
the pivot with a constrained rigid
HINT: temporarily translate the spindle a small distance so
body. that the 4 nodes required for the joint can be picked. Use
Units: mm, s, t, N, MPa DISPLAY-ENTITIES to visualise the nodes in the model.
back to contents
LS-DYNA ENVIRONMENT
240

O
N
1. Time Integration 4. Materials T
integration O
• Types of contact L
integration • Contact method (depth, stiffness)
&
• Summary of cost considerations • Other contact issues O
• Contact guidelines U
2. Element Formulation T
6. Rigid Bodies P
U
T
• Meshing Notes
3. Hourglassing
back to contents
LS-DYNA ENVIRONMENT
241
Control & Output LS-DYNA Introductory Course C

O
N
There is a large number of control and output cards available in LS-DYNA. The T
most important ones are summarized in this section. R
O
Control data: Output data: L
*CONTROL_CONTACT *DATABASE_BINARY_D3PLOT
*CONTROL_ENERGY *DATABASE_BINARY_D3THDT &
*CONTROL_OUTPUT *DATABASE_BINARY_D3DUMP O
*CONTROL_SHELL *DATABASE_BINARY_RUNRSF U
*DATABASE_... ASCII files T
*CONTROL_SOLID
P
*CONTROL_ACCURACY *DATABASE_CROSS_SECTION U
*CONTROL_TERMINATION *DATABASE_HISTORY T
*CONTROL_TIMESTEP *DATABASE_EXTENT_BINARY
If a particular card is not defined in the input deck, LS-DYNA will assume the
defaults for that particular card.
back to contents
LS-DYNA ENVIRONMENT
242
Control Data LS-DYNA Introductory Course C

O
N
*CONTROL_CONTACT
T
R
Card 1 O
L
&
O
U
T
Card 2 P
U
T
RWPNAL – Set to 1.0 if there are any *RIGIDWALL card in the model.
Everything else – leave as default.
back to contents
LS-DYNA ENVIRONMENT
243

O
N
*CONTROL_CONTACT
T
R
Card 3 O
L
&
O
U
T
Card 4 P
U
T
IGNORE – Set to 1.0 to ignore any initial penetrations for all contacts.
Everything else – leave as default.
back to contents
LS-DYNA ENVIRONMENT
244

O
N
*CONTROL_ENERGY T
R
O
L
&
O
U
T
P
U
HGEN – Set to 2 to compute hourglass energy. Check that hourglass
T
energy is less than about 5% of total energy. Use MATSUM file
to check hourglass energy for each individual part.
SLNTEN – Set to 2 to compute contact energy – helps when debugging.
back to contents
LS-DYNA ENVIRONMENT
245

O
N
*CONTROL_OUTPUT T
R
O
L
&
O
U
NPOPT – Set to 1 to stop the .otf file getting too big (suppresses T
printing of node and element data). P
U
T
NREFUP – Set to 1 if you want to see how beam elements‟ orientations change
during the analysis, this will move the third node with the beam.
N.B. In this case, the beam 3rd nodes must NOT be part of the
structure and each beam should have a unique 3rd node.
back to contents
LS-DYNA ENVIRONMENT
246

O
N
*CONTROL_SHELL T
R
O
Card 1 L
&
O
U
T
P
U
ESORT – set to 1, sorts degenerate quads into type 4 triangular shells.
T
ISTUPD – if highly strained elements present, recommended to set to 1, as
membrane strain causes a thickness change; leave as default otherwise.
Important - these are not the LS-DYNA defaults!

back to contents
LS-DYNA ENVIRONMENT
247

O
N
*CONTROL_SHELL T
R
O
Card 2 L
Optional
&
O
U
T
P
U
NFAIL1 – flag to delete highly distorted under-integrated shell elements. T
NFAIL4 – flag to delete highly distorted fully-integrated shell elements.
back to contents
LS-DYNA ENVIRONMENT
248

O
N
*CONTROL_SOLID T
R
O
Card 1 L
&
O
U
T
P
U
ESORT – set to 1, sorts degenerate tetrahedron and pentahedron into type 10
T
and type 15, respectively, which are more stable formulations
back to contents
LS-DYNA ENVIRONMENT
249

O
N
*CONTROL_ACCURACY T
R
O
L
&
O
U
T
P
U
OSU – set to 1, objective stress updates helps with large rotations, for
T
example when using *INITIAL_VELO_GEN.
INN – set to 2 for shells or 3 for solids, makes the element behaviour
independent of node numbering – can help to reduce hourglassing – see
next slide for details.
back to contents
LS-DYNA ENVIRONMENT
250

O
*CONTROL_ACCURACY N
T
INN For shells, the default local X-axis is based on the side N1-N2 of the element, and
R
this can cause non-physical rotations of the stress tensor, under large in-plane shear O
deformations. L
For distorted elements, this can lead to different results if the node numbering changes.
&
With the INN parameter switched on, a permutation of the nodes is done, which minimizes
the effect of the nodes sequence, while adding a <5% cost. O
U
Initial Shape Deformed Shape (INN=0/1) Deformed Shape (INN=2) T
N4 N3 N4 N3
N4 N3 P
U
y T
N1 N1
N1 N2
Note: Schematic diagram only
back to contents N2 N2
LS-DYNA ENVIRONMENT
251

O
N
*CONTROL_TERMINATION T
R
O
L
&
O
U
T
P
U
T
ENDTIM – Termination time: always need this!
ENDMAS – Prevents CPU wastage due to model getting too heavy because
of mass scaling. Set to 20 to cause termination if mass increases
by 20%;
back to contents
LS-DYNA ENVIRONMENT
252

O
N
*CONTROL_TIMESTEP T
R
O
L
&
O
U
T
DT2MS – When negative (positive value not often used): P
Elements are made heavier to stop timestep U
Analysis
getting below absolute value of DT2MS*TSSFAC Time Step T
(use for dynamic analysis). A typical value: -1.2e-6
(This gives timestep of 0.9*1.2e-6 sec = approx. Max limit: LCTM
1.08e-6 sec.) Check in .otf file that mass increase
Time Step is kept
is not unacceptable (should be <1-2%). between these 2 values
LCTM – Load curve ID that limits the maximum time step size Min limit
|DT2MS|*TSSFAC
back to contents
LS-DYNA ENVIRONMENT
253
Output Data LS-DYNA Introductory Course C

O
N
*DATABASE_BINARY_D3PLOT T
Sets output interval for .ptf (or d3plot) files. R
O
L
&
O
U
T
P
U
T
DT – If for example 100 plot states need to be output (this depends on the
analysis), the DT/CYCL in this case should be set to
= (Termination Time)/100.
back to contents
LS-DYNA ENVIRONMENT
254

O
N
*DATABASE_BINARY_D3THDT T
• Output interval for .thf file (or d3thdt). Contains time history data for nodal, R
element data output, etc in T/HIS, card below. O
• Most of this data must be specifically requested using L
*DATABASE_HISTORY_.. cards (examples to follow).
&
O
U
T
P
U
T
back to contents
LS-DYNA ENVIRONMENT
255
Output
OutputData
Data– DATABASE ASCII LS-DYNA Introductory Course C
O
N
*DATABASE_...ASCII files – these files have been traditionally called ASCII files, T
but they can now be written to a binary database (LSDA or binout). R
Example of useful ASCII files: O
L
*DATABASE_GLSTAT – Global model data
*DATABASE_MATSUM – Material energies/velocities by part &
*DATABASE_SLEOUT – Contact energies
O
*DATABASE_JNTFORC – Joint forces U
*DATABASE_SECFORC – Asks for output of forces into file SECFORC, see T
P
also next slides and *DATABASE_CROSS_SECTION U
*DATABASE_NODOUT – See also next slides and *DATABASE_HISTORY T
*DATABASE_RCFORC – Resultant interface forces
*DATABASE_ELOUT – Element data
*DATABASE_DFORC – Discrete elements
*DATABASE_SPCFORC – Reaction forces
*DATABASE_RWFORC – Rigid wall forces
back to contents
LS-DYNA ENVIRONMENT
256
Output
OutputData
Data– DATABASE ASCII LS-DYNA Introductory Course C
O
N
*DATABASE_...ASCII files T
LS-DYNA v970 and onwards creates an LSDA file named “binout”. This is a binary R
file containing all the ASCII file outputs in a directory structure within the file. It is a O
much more compact version of the ASCII files. L
&
O
U
T
P
U
T
DT – Time interval for outputs.

BINARY – if =1: only ASCII files written, default on SMP.
if =2: only binout files written, default on MPP – recommended.
if =3: both ASCII and binout written on SMP – option only valid for
back to contents
GLSTAT on MPP.
LS-DYNA ENVIRONMENT
257

O
N
*DATABASE_CROSS_SECTION T
In order to get output of forces and moments in a specific cross section of the R
model (output files SECFORC and binout) the following two steps need to be O
performed: L
&
1. Add card *DATABASE_SECFORC below:
O
U
T
P
U
T
2. Define the cross-section, using *DATABASE_CROSS_SECTION, as

explained on next slide.
back to contents
LS-DYNA ENVIRONMENT
258

O
*DATABASE_CROSS_SECTION_{OPTION}_{OPTION2} N
T
R
Example – Define a cross-section, using
O
*DATABASE_CROSS_SECTION_PLANE L
&
O
LENM U
T
P
(XHEV,
YHEV,
U
ZHEV) T
L
(XCT,
YCT, z
ZCT) (XCH,
YCH, y
back to contents
N ZCH)
x
LS-DYNA ENVIRONMENT
259

O
N
*DATABASE_HISTORY_{OPTION} T
• Allows to selectively request outputs for nodes and elements. R
• In order to get output of, for example nodes displacement, velocity and O
L
acceleration, need to:
1. Add card *DATABASE_NODOUT below and data will be output in the &
NODOUT file (and .thf or d3thdt):
O
U
T
P
U
T
2. Define the history nodes, using *DATABASE_HISTORY_NODE, as

explained on next slide.
back to contents
LS-DYNA ENVIRONMENT
260

O
N
*DATABASE_HISTORY_{OPTION} T
• Define the history nodes, using *DATABASE_HISTORY_NODE or R
_NODE_SET, where ID1… are the nodes or node sets ID‟s to be included in O
the output: L
&
O
U
T
P
• Or, use a description of the node/node set, using *DATABASE_HISTORY_ U
NODE_ID or NODE_SET_ID: T
back to contents
LS-DYNA ENVIRONMENT
261
Output Data – *DATABASE_EXTENT_BINARY LS-DYNA Introductory Course C

O
N
*DATABASE_EXTENT_BINARY T
Controls output of element stresses and strains to .ptf, .thf and ASCII files: R
O
Card 1 L
&
O
U
T
MAXINT – number of SHELL integration points written, default = 3; see next slide for details. P
STRFLG – set =1 to include strain tensors for solids & shells; default = 0 exclude. U
T
SIGFLG – set =1 to include stress tensor for shells, default; set =2 to exclude.
EPSFLG – set =1 to include effective plastic strains for shells, default; set =2 to exclude.
RLTFLG – set =1 to include stress resultant for shells, default; set =2 to exclude.
ENGFLG – set =1 to include internal energy density & thickness for shells, default; set =2 to
exclude.
Card 2 – take defaults (advanced options).
back to contents
LS-DYNA ENVIRONMENT
262
Data Output – Integration Points LS-DYNA Introductory Course C

O
N
On the *DATABASE_EXTENT_BINARY card you need to specify the number of shell
T
integration points for which you want data to be output, using MAXINT. If a part has less
R
integration points than specified by MAXINT, then the extra points will be given a value of
O
zero.
L
If you don‟t specify anything on the *DATABASE_EXTENT_BINARY card or set MAXINT=3 &
then Dyna will calculate the values for the middle, bottom, and top surfaces of the shell
element, and will output these values for integration points 1, 2 and 3. O
U
Example – Shell with 8 integration points and with MAXINT=3: T
P
U
T
Data Output
back to contents
LS-DYNA ENVIRONMENT
263
Data Output – Integration Points LS-DYNA Introductory Course C

O
N
If you specify another value than 3 on the *DATABASE_EXTENT_BINARY card, then Dyna
T
will just report the data for those integration points.
R
O
Example – Shell with 8 integration points and with MAXINT=4: L
&
O
U
T
P
Data Output U
T
back to contents
LS-DYNA ENVIRONMENT
264

O
N
• Stresses and strains for elements are by default output in the following way: T
R
• Shells: O
• binout (LSDA) → Local coordinate system L
• .thf (D3THDT) → Global coordinate system
• .ptf (D3PLOT) → Global coordinate system &
O
• Solids:
U
• binout (LSDA) → Global coordinate system T
• .thf (D3THDT) → Global coordinate system P
• .ptf (D3PLOT) → Global coordinate system U
T
• For certain materials and elements this can be altered (example solid elements
type with *MAT_SPOTWELD)
back to contents
LS-DYNA ENVIRONMENT
265

O
N
*DATABASE_BINARY_D3DUMP T
• Output interval (in cycles) for output of .dpf (or d3dump) – which are R
permanent restart dumps. O
• Outputting too many .dpf files uses up lots of disk space – typically set L
DT/CYCL to something large to avoid this.
• If this keyword is omitted, then only one set of .dpf files (one for each core if &
run MPP) will automatically be written at termination time.
O
U
T
P
U
T
back to contents
LS-DYNA ENVIRONMENT
266

O
N
*DATABASE_BINARY_RUNRSF T
• Output interval (in cycles) for output of .adf (or runrsf) –which are R
overwritten restart dumps. O
• Output every 10000 cycles recommended. L
• If the analysis crashes (and no .dpf files are available), the .adf files can be
used for restart. &
O
U
T
P
U
T
back to contents
LS-DYNA ENVIRONMENT
267
Output data – Accelerometers LS-DYNA Introductory Course C

O
N
• Accelerometers are useful for replicating test z
x T
measurements and should be used instead of looking
R
at „raw‟ nodal accelerations.
O
• Define using L
*ELEMENT_SEATBELT_ACCELEROMETER card;
• The accelerometer is fixed to a rigid body containing 3 &
y
nodes NID1, NID2 and NID3.
• Include Node 1 of the accelerometer in O
*DATABASE_HISTORY_NODE to output U
accelerometer data. T
• The velocity and acceleration of Node 1 will then be P
output in the local coordinates (X from node 1 to node U
2, Y in the 1-2-3 plane). T
x
z
• In the example shown, a rigid solid brick element is ND2
using *CONSTRAINED_EXTRA_ NODES to attach to
the three accelerometer nodes and the rest of the ND3
structure.
y ND1
back to contents
LS-DYNA ENVIRONMENT
268
Output data – Accelerometers LS-DYNA Introductory Course C

O
N
*ELEMENT_SEATBELT_ACCELEROMETER
T
• It is recommended that a rigid patch or nodal rigid body be created around the R
accelerometer and attach a suitable mass to Node 1 to represent the accelerometer; this O
reduces noise and gives more reliable acceleration output. L
• If acceleration outputs are still very noisy before filtering, results could be misleading; in this
case use a post-processor to differentiate the velocity instead of taking acceleration directly &
– note that if the accelerometer undergoes rotations, the INTOPT=1 needs to be used for
the differentiated velocity to give the correct acceleration. O
• An automatic filter of nodal accelerations can be set in the *CONTROL_OUTPUT using the U
IACCOP=1 parameter. T
• A more detailed example can be supplied on request. P
U
T
back to contents
LS-DYNA ENVIRONMENT
269
*DAMPING LS-DYNA Introductory Course C

O
• Damping can be specified globally or for part sets using *DAMPING_GLOBAL or N
T
*DAMPING_PART_MASS.
R
• Mass weighted and Rayleigh‟s damping options available. O
L
• Define mass weighted nodal damping that applies globally to the nodes of deformable
bodies and to the mass centre of the rigid bodies. For mass weighted damping, the &
acceleration is computed,
O
where M is mass matrix, Pn is the external load vector, Fn is the internal load vector, U
T
Fndamp is the force vector due to system damping:
P
U
T
ωmin is lowest frequency of interest (radians per unit time)
• Damping Energy dissipated through mass weighted damping is reported as system
Damping Energy in the GLSTAT file.
back to contents
LS-DYNA ENVIRONMENT
270
Control Cards - Recommendations LS-DYNA Introductory Course C

O
N
The following slides show a summary of the recommended control card:
T
*CONTROL_BULK_VISCOSITY R
• Q1 = Q2 = 0 O
• TYPE= -2 – Must always have this for shells type #16 L
*CONTROL_CONTACT
&
• IGNORE=1 or 2 – Ignore initial penetrations
• SPOTHIN=0.5 if spotwelds exist in the model O
*CONTROL_ENERGY U
• HGEN=2 – Calculate the hourglass energy and add it to the total energy calc T
• RWEN=2 – Calculate rigidwall energy P
U
• SLNTEN=2 – Calculate sliding interface energy
T
• RYLEN=2 – Damping energy calculated
*CONTROL_OUTPUT
• NOPT=1 – Suppress nodal coords, element connectivity, rigid wall def. and
initial velocities to be printed to the .otf
• NEECHO=3 – Both element and node printing are suppressed
back to contents
LS-DYNA ENVIRONMENT
271

O
N
*CONTROL_MPP_IO_NOFULL
T
• To stop the d3full restart files R
*CONTROL_MPP_IO_LSTC_REDUCE O
• Gives more consistent summation of floating point data among processors L
*CONTROL_SHELL
&
• ESORT=1 – Auto sorting of 3-noded shells
• ISTUPD= 0 – Unless large membrane stretching in shells, for which case O
recommend 1 U
• BWC=1 – Add warping stiffness T
• Set NFAIL1=1 & NFAIL4=1 – Check for highly distorted shells and delete them P
*CONTROL_SOLID U
T
• ESORT=1 – Auto sorting of tetrahedral and pentahedral solids
back to contents
LS-DYNA ENVIRONMENT
272

O
N
*CONTROL_ACCURACY
T
• OSU=1 – Objective stress update R
• INN=4 – Invariant node numbering for shells and solids O
*DATABASE_FORMAT L
• IBINARY=1 – 32 bit output for D3PLOT files if running double precision
&
O
U
T
P
U
T
back to contents
LS-DYNA ENVIRONMENT
273

O
N
1. Request output of .ptf (every
T
0.001s) and .thf files (every upper fork R
0.0001s). Also request ASCII
O
output for contact forces and
L
discrete element forces. lower fork
2. Request HISTORY_NODE output &
for node 513.
3. Run the model. The spring is the wheel O
only element with a calculated U
timestep which leads to an unstable T
run. We need to force the timestep P
to be a smaller value. U
4. Create a loadcurve of timestep vs T
time and add to
*CONTROL_TIMESTEP. Use a
pivot
timestep of 1e-6 sec.
5. Run the new model and plot the ground
spring, damper and contact surface
forces and node output in T/HIS
back to contents
LS-DYNA ENVIRONMENT
274
Outline – THEORY AND PRACTICE LS-DYNA Introductory Course O

T
H
integration
integration • Contact method (depth, stiffness) E
C
H
• Summary of cost considerations • Other contact issues N
• Contact guidelines I
2. Element Formulation Q
6. Rigid Bodies U
E
S
• Meshing Notes
3. Hourglassing
back to contents
LS-DYNA ENVIRONMENT
275
LS-DYNA – Other Techniques LS-DYNA Introductory Course O

T
H
With the LS-DYNA code, there are other analysis methods available, such as the E
following: R
• Thermal T
E
• Implicit C
• ALE – Arbitrary Lagrangian-Eulerian H
• SPH – Smoothed Particle Hydrodynamics N
I
Q
U
E
S
back to contents
LS-DYNA ENVIRONMENT
276
LS-DYNA – Thermal analysis LS-DYNA Introductory Course O

T
H
With the LS-DYNA code the following Thermal problems can be solved: E
R
• Steady State Thermal
T
• Transient Heat Transfer
E
• Coupled Thermal-structural C
H
N
I
Q
U
E
S
back to contents
LS-DYNA ENVIRONMENT
277

T
H
E
• Steady State Thermal – solves for equilibrium thermal state, boundary R
conditions do not vary with time.
T
E
C
H
Tconst=40 Tconst= 0 N
I
Q
U
E
S
Temperature at an
internal point is
calculated
back to contents
LS-DYNA ENVIRONMENT
278

T
H
E
• Transient Heat Transfer – boundary conditions can vary with time. R
T
E
C
H
T=40 T= 0 N
I
Q
U
E
S
The analysis solves for
the temperature at an
internal point
back to contents
Time
LS-DYNA ENVIRONMENT
279

T
H
• In order to run a thermal LS-DYNA analysis the following control keywords E
are needed: R
• *CONTROL_SOLUTION – to specify the analysis type, i.e. structural only, T

thermal only or coupled structural-thermal. E
C
H
• *CONTROL_THERMAL_SOLVER – to specify options, such as steady-state N
or transient analysis, linear or non linear analysis, solver type, etc. I
Q
U
• *CONTROL_THERMAL_TIMESTEP – to specify the thermal timestep; E
implicit time integration is used for the thermal calculations, while explicit time S
integration is used for the structural calculations; implicit time integration is
unconditionally stable and so a larger timestep can be taken. The control
cards *CONTROL_TIMESTEP sets the timestep for the structural part of the
analysis.
back to contents
LS-DYNA ENVIRONMENT
280

T
H
• For a Thermal analysis, the thermal material is referred to via TMID in the E
*PART card: R
T
E
C
H
N
I
Q
• Below are a few thermal materials; the thermal materials allow for Heat U
Capacity and Thermal Conductivity to be specified: E
S
*MAT_THERMAL_ISOTROPIC (*MAT_T01)
*MAT_THERMAL_ORTHOTROPIC (*MAT_T02)
*MAT_THERMAL_ISOTROPIC_TD (*MAT_T03)
*MAT_THERMAL_ORTHOTROPIC_TD (*MAT_T04)
…
back to contents
LS-DYNA ENVIRONMENT
281

T
H
• For a coupled Thermal-structural analysis, both a thermal and a structural E
material need to be defined; R
T
E
C
H
N
I
Q
• In order for the material properties to vary due to temperature changes, a U
structural material with temperature dependent material properties should be E
used (E1, PR1, T1, E2, PR2, T2, ...), such as S
*MAT_ELASTIC_PLASTIC_THERMAL (*MAT_004). Also the Thermal
Coefficient of Expansion is input here.
• Thermal expansion can alternatively be included by using a
*MAT_ADD_THERMAL_EXPANSION in conjunction with a non-thermal
material.
back to contents
LS-DYNA ENVIRONMENT
282

T
H
• A coupled thermal-structural analysis can be set up using for example: E
*BOUNDARY_TEMPERATURE_{OPTION} R
*BOUNDARY_FLUX
T
*INITIAL_TEMPERATURE_{OPTION} E
*LOAD_HEAT_GENERATION C
H
N
• For a structural only analysis with predefined nodal temperatures use for I
example: Q
*LOAD_THERMAL_LOAD_CURVE U
E
*LOAD_THERMAL_VARIABLE
S
• For cases where different components come in contact, the heat transition
between surfaces can be modelled using:
*CONTACT_SURFACE_TO_SURFACE_THERMAL
back to contents
LS-DYNA ENVIRONMENT
283
LS-DYNA Implicit – set up analysis LS-DYNA Introductory Course O

T
LS-DYNA Implicit analysis H
• This short introduction to the implicit analysis in LS-DYNA focuses on linear E
and non-linear static analysis for a range of different analysis types including: R
• Simple static analysis T
• Eigenvalue analysis E
• Buckling analysis C
H
• Frequency Response analysis N
• In an implicit static analysis a termination time and a timestep are specified I
by the user. However, for implicit static, time doesn‟t have any actual Q
meaning as the analysis is trying to achieve static equilibrium at each U
timestep. E
S
• The terminology time is used however as Implicit and Explicit share a large
number of the same keywords within LS-DYNA.
• Setting up an implicit analysis from an existing explicit model is quite straight
forward and the cards needed are summarized in the next slide.
• There are some restrictions on materials, contacts, etc which work with the
implicit solver.
back to contents
LS-DYNA ENVIRONMENT
284

T
• In order to run an implicit LS-DYNA analysis the following keywords are H
E
needed:
R
• *CONTROL_IMPLICIT_GENERAL – allows specification of the time step; T

also allows an automatic switch between implicit and explicit. E
C
H
• *CONTROL_IMPLICIT_SOLUTION – to specify if the implicit analysis is
N
either linear or non-linear. I
Q
• *CONTROL_IMPLICIT_SOLVER – to specify the linear equation solver used U
to perform the stiffness matrix inversion (use defaults). E
S
• *CONTROL_IMPLICIT_AUTO – to specify how LS-DYNA can automatically

adjust the timestep to achieve a solution.
• The termination time is specified in the *CONTROL_TERMINATION – for a

single step (static) analysis the step size is set to the termination time.
back to contents
LS-DYNA ENVIRONMENT
285
Example 1: Simple Cantilever Beam Analysis LS-DYNA Introductory Course O

T
H
A simple cantilever beam modelled in shell mm
E
elements is fixed at one end and a load applied
R
at the other.
T
Load (F) = 0.778N E
Length (L) = 300mm C
Width (b) = 40mm H
Thickness (d) = 1mm N
I
I (2nd Moment of Area) Q
= b d3 / 12 U
= 40 * 13 / 12 E
= 3.333 S
End deflection
= F L3 / 3 E I
= 0.778 * 3003 / 3 * 210e3 * 3.333
= 10.00mm
back to contents
LS-DYNA ENVIRONMENT
286

T
H
• LS-DYNA implicit can also be used to perform Eigenvalue analyses, in which E
case the following cards are required: R
*CONTROL_IMPLICIT_GENERAL T
*CONTROL_IMPLICIT_EIGENVALUE E
C
H
• Shell element types 18-21 and solid element type 18 should be used in N
eigenvalue analyses (ELFORM on the *SECTION card) to get the most I
accurate results. Q
U
E
• A .ptf (d3plot) file is output from the analysis, however the actual eigenvalue S
results are stored in the „eigout‟ and „d3eigv‟ files.
back to contents
LS-DYNA ENVIRONMENT
287
Example 6: Eigenvalue Analysis LS-DYNA Introductory Course O

T
H
Modelling Points E
R
• *CONTROL_IMPLICIT_EIGENVALUE card used to define and eigenvalue
T
analysis:
E
C
*CONTROL_IMPLICIT_EIGENVALUE
H
8 0.0 0 0.0 0 0.0 0 0.0
N
0 0 0 0 0 I
Q
• Shell element type 18 specified on *SECTION_SHELL card: U
E
*SECTION_SHELL S
1 18 0.0 0 0.0 0.0 0 0
1.0 1.0 1.0 1.0 0.0 0.0 0.0 0
back to contents
LS-DYNA ENVIRONMENT
288
Eigenvalue Analysis LS-DYNA Introductory Course O

T
H
The d3eigv file is a plot state file similar to the ptf file that allows the various mode shapes
E
to be visualized. Each state represents a particular mode and can be animated in Oasys
R
D3PLOT
T
E
C
H
The Modeshape
N
animation style is
I
recommend for viewing
Q
these results
U
Anim → Style → Modeshape
E
S
back to contents
LS-DYNA ENVIRONMENT
289

T
H
A tutorial for LS-DYNA Implicit can be found at: E
R
http://www.oasys-software.com/dyna/en/training/tutorials.shtml
T
E
C
H
N
I
Q
U
E
S
back to contents
LS-DYNA ENVIRONMENT
290
LS-DYNA – Hydrocode LS-DYNA Introductory Course O

T
A hydrocode refers to a computer program used for the study of the dynamic H
E
response of materials and structures to impulse, such as blast and impact, R
involving everything from car and aircraft collisions to impact of space structures
by assorted debris. T
E
Explicit FE – structure C
H
Smoothed Particle Hydrodynamics N
Lagrangian (SPH) – structure, fluid, fluid-structure I
interaction (FSI) Q
U
Hydrocode Element Free Galerkin (EFG) E
– structure (meshfree) S
Eulerian – fluid
mechanics
Arbitrary Lagrangian-
Eulerian (ALE) – fluid-
structure interaction (FSI)
back to contents
LS-DYNA ENVIRONMENT
291
LS-DYNA – FSI capabilities overview LS-DYNA Introductory Course O

T
H
Fluid Structure Interaction (FSI and ALE - Arbitrary Lagrangian-Eulerian ) E
Within LS-DYNA a block of solid elements can be defined as a Eulerian control R
volume allowing fluids and fluid flow to be modelled.
T
This Eulerian mesh can also then be coupled to a structural (Lagrangian) mesh
E
using *CONSTRAINED_LAGRANGE_IN_SOLID such that the two interact. C
H
N
I
Q
U
E
S
back to contents
LS-DYNA ENVIRONMENT
292
LS-DYNA – SPH capabilities overview LS-DYNA Introductory Course O

T
H
Smoothed Particle Hydrodynamics (SPH) E
SPH is a Lagrangian explicit method where the structure is modelled using mesh R
of particles which aren‟t directly connected together, but rather have a sphere of
influence within which surrounding particles can affect each other. This allows for T
very high deformations to be modelled. E
C
H
N
I
Q
U
E
S
back to contents
LS-DYNA ENVIRONMENT
293
• BACKGROUND • THEORY AND PRACTICE

• Time Integration
• Element Formulation
• INTRODUCTION • Hourglassing
• Key principles
• Materials
• Running
• Contact Surfaces
• Output files
• Rigids
• Post Processing
• Boundary Conditions
• Connections
• Control & Output
• Other Techniques
• ANALYSIS
• Checking
back to contents
LS-DYNA ENVIRONMENT
294
Outline – ANALYSIS LS-DYNA Introductory Course A

N
A
1.0 Checking L
2.0 Restarting the Model Y
S
I
S
back to contents
LS-DYNA ENVIRONMENT
295
Model Checking LS-DYNA Introductory Course A

N
A
There are many checks on the model that should be performed before and L
after the analysis: Y
S
I
– Standard pre-processor checks
S
– Primer checks
– Initialization checks
– Post-analysis checks
back to contents
LS-DYNA ENVIRONMENT
296
Model Checking - Pre-processor LS-DYNA Introductory Course A

N
A
• Element Quality (check against in-house targets):
L
– Warpage Y
– Aspect ratio S
– Internal angles I
S
• Allowable element quality is dependent on element location and model purpose;
• Mesh Quality:
– Duplicate elements
– Free edges/faces
– Coincident nodes
– Small sized elements (low timesteps + mass scaling issues)
– Element normals (required for specific features only)
back to contents
LS-DYNA ENVIRONMENT
297
Model Checking - Pre-processor LS-DYNA Introductory Course A

N
• Model Geometry and Structure: A
– Visual check for inter-penetrating parts L
Y
– Panel gauges/thicknesses
S
– Connections and attachments I
S
• Material Properties:
– Material types and parameters
– Appropriate material types for element types
• Boundary Conditions, Initial Conditions and Interactions:

– Restraints and symmetry conditions
– Initial velocities
– Contact surface definitions
• Output Definitions:
– Node and element time-histories
– Reaction forces, cross-section definitions, etc.
back to contents
LS-DYNA ENVIRONMENT
298
Model Checking – Primer LS-DYNA Introductory Course A

N
Initial velocity vector plot A
L
Y
S
I
S
Vector plot shows error:

engine has different velocity
back to contents
LS-DYNA ENVIRONMENT
299

N
A
L
Y
Visual check of panel gauges S
I
S
back to contents
LS-DYNA ENVIRONMENT
300

N
A
L
Y
Mild steel and high strength S
steel - check that correct I
materials have been used S
back to contents
LS-DYNA ENVIRONMENT
301

Model Checking - Primer
N
Oasys PRIMER A
L
Can change settings to show Y
only small timestep S
elements. I
S
Especially check spring
elements - may need to add
lumped mass.
back to contents
LS-DYNA ENVIRONMENT
302

Model Checking - Primer
N
Oasys PRIMER A
L
Y
Shows where LS-DYNA will S
add mass I
S
back to contents
LS-DYNA ENVIRONMENT
303
Model Checking – Primer - Find attached LS-DYNA Introductory Course A

N
A
Primer finds all LS-DYNA data attached to VISIBLE L
entities: display in Line mode to avoid missing entities Y
attached to hidden nodes/elements S
I
S
Use these buttons to

control which types of
entities will be displayed
back to contents
LS-DYNA ENVIRONMENT
304

N
A
L
Y
In this case, only extra S
nodes on rigid bodies are I
directly connected... S
back to contents
LS-DYNA ENVIRONMENT
305

N
A
L
Press Apply again, now we can see Y
some elements and springs that are S
attached to the extra nodes ... I
S
back to contents
LS-DYNA ENVIRONMENT
306

N
A
L
Press Apply again, now we can see Y
some beams that are attached to the S
springs ... I
S
back to contents
LS-DYNA ENVIRONMENT
307

N
A
L
Y
S
I
S
Press Apply again...
back to contents
LS-DYNA ENVIRONMENT
308
Model Checking – Primer – Model Checker LS-DYNA Introductory Course A

N
A
• Run Model Checker in L
Primer. Y
• Model Checker will only S
I
ensure that the model is S
consistent within itself.
back to contents
LS-DYNA ENVIRONMENT
309
Model Checking - Initialisation LS-DYNA Introductory Course A

N
A
• The model can be initialized by setting
L
CPU = 1 sec on the submission shell Y
• Then check <jobname>.otf S
– Did it complete initialization stage? Normal I
termination? S
– Are there any errors or warning messages? Search
for „Error‟ or „Warning‟.
– Are the total masses of the model and each part
correct? Search for „t o t a l‟.
– Are the rigid body restraint conditions correct?
– Which element is controlling the timestep? A spring?
Search for „smallest‟.
– Is the timestep reasonable?
– Are the contact surface timesteps less than model
time step?
– Is the initial energy correct?
– How much mass has been added in mass scaling?
Is this acceptable?
back to contents
LS-DYNA ENVIRONMENT
310
Model Checking - Initialization LS-DYNA Introductory Course A

N
A
• Using D3PLOT: L
Y
– Magnify displacements x100 or x1000 and compare model at „zero‟ state S
I
(undeformed geometry) and at „first‟ state (time zero). Do any parts of the
S
model “jump”?
– Plot velocity vectors at first state. Are the initial velocities correct?
– Are contact surfaces correctly defined?
– Are parts joined correctly – check for discrepancies in velocity?
back to contents
LS-DYNA ENVIRONMENT
311
Model Checking – Post analysis – .otf file LS-DYNA Introductory Course A

N
T i m i n g i n f o r m a t i o n
CPU(seconds) %CPU Clock(seconds) %Clock A
---------------------------------------------------------------- L
Initialization ....... 4.7000E-01 0.01 5.0935E-01 0.01
Element processing ... 4.0769E+03 47.79 2.0127E+03 47.58 Y
Binary databases ..... 3.3946E+00 0.04 1.7572E+00 0.04 In the .otf file, S
ASCII database ....... 2.0024E+00 0.02 1.1212E+00 0.03 check the time
Contact algorithm .... 4.3704E+03 51.23 2.1756E+03 51.43 I
Interface ID 1 3.4023E+03 39.88 1.6947E+03 40.06 necessary for the S
Interface ID 2 9.4749E+02 11.11 4.7645E+02 11.26
Contact entities ..... 0.0000E+00 0.00 0.0000E+00 0.00 various steps in
Rigid bodies ......... 7.7279E+01 0.91 3.8627E+01 0.91 the analysis.
Implicit Nonlinear ... 0.0000E+00 0.00 0.0000E+00 0.00
Implicit Lin. Alg. ... 0.0000E+00 0.00 0.0000E+00 0.00
----------------------------------------------------------------
T o t a l s 8.5305E+03 100.00 4.2303E+03 100.00
Problem time = 5.0000E-01

Problem cycle = 1171653
Total CPU time = 8531 seconds ( 2 hours 22 minutes 11 seconds)
CPU time per zone cycle = 810 nanoseconds
Clock time per zone cycle= 402 nanoseconds
Number of CPU's 2
NLQ used/max 272/ 272
Start time 09/13/2010 16:53:41
End time 09/13/2010 18:04:50
Elapsed time 4269 seconds( 1 hours 11 min. 9 sec.) for 1171653 cycles
N o r m a l t e r m i n a t i o n
back to contents
LS-DYNA ENVIRONMENT
312
Model Checking - Post analysis – .otf file LS-DYNA Introductory Course A

N
100 smallest timesteps A
---------------------- L
element timestep Y
shell 27156 0.35129E-06
shell 12452 0.35129E-06
Check timesteps in S
shell 38534 0.41072E-06 structure... I
shell 38533 0.41072E-06 S
shell 1101 0.41072E-06
shell 1096 0.41072E-06
… and the stable timestep for contacts – the model

timestep should be smaller than the timestep used for
contacts, or instability could occur during the run.
slave surface of interface # 1 type= 13
surface timestep= 0.761E-06 current minimum= 0.761E-06
master surface of interface # 3 type= 7
The LS-DYNA time step size should not exceed 0.498E-06
back to contents
LS-DYNA ENVIRONMENT
313
Model Checking - Post analysis – .otf file LS-DYNA Introductory Course A

N
A
L
Y
S
1170000 t 4.9930E-01 dt 4.24E-07 flush i/o buffers I
1171652 t 5.0000E-01 dt 4.24E-07 write d3plot file Check the final time S
step used in the
*** termination time reached *** analysis.
1171653 t 5.0000E-01 dt 4.24E-07 write crush_tube_smp.dpf01 file
1171653 t 5.0000E-01 dt 4.24E-07 write d3plot file
Final cycle Final analysis Time step used

number time reached in final cycle
back to contents
LS-DYNA ENVIRONMENT
314
Model checking – Post analysis – .otf file LS-DYNA Introductory Course A

N
A
L
problem cycle = 1200 Y
time = 1.2949E-03 Check added mass S
added mass = 4.3696E-04 in structure... I
percentage increase = 2.7420E-02 S
Deformable Spotwelds *****

total added spotweld mass = 1.3139E-05 … and in spotwelds
percentage mass increase = 8.2451E-04
back to contents
LS-DYNA ENVIRONMENT
315
Model Checking - Post analysis LS-DYNA Introductory Course A

N
A
• Using T/HIS L
Y
– Check global energy balance (GLSTAT file) S
I
– Check contact forces and energies S
– Check individual part energies
back to contents
LS-DYNA ENVIRONMENT
316
Model Checking - Post analysis LS-DYNA Introductory Course A

N
A
• Using T/HIS L
– Check global energy balance Y
(GLSTAT file) S
I
– Total Energy = Internal + Kinetic +
S
Contact (sliding) + Hourglass +
System damping + Rigidwall Energies
back to contents
LS-DYNA ENVIRONMENT
317
Running LS-DYNA Analyses LS-DYNA Introductory Course A

N
A
• Submitting LS-DYNA jobs L
• Runtime of an analysis Y
S
• Speed-up via mass-scaling I
S
• Restarting the analysis
back to contents
LS-DYNA ENVIRONMENT
318
Running LS-DYNA - Controlling CPU Time LS-DYNA Introductory Course A

N
A
• The CPU time is proportional to the model size and the user should always check L
the CPU required. Y
• The CPU time can be estimated by following the next steps: S
– Submit the analysis and let it run past initialisation. I
S
– Open the .otf file and scroll to the end, where you will se lines similar to the ones below:
…
shell 4739 0.64040E-06
shell 8555 0.64040E-06
shell 5675 0.64040E-06
1 t 0.0000E+00 dt 6.19E-07 flush i/o buffers
1 t 0.0000E+00 dt 6.19E-07 write d3plot file
5000 t 3.0883E-03 dt 6.16E-07 flush i/o buffers
8136 t 4.9996E-03 dt 6.02E-07 write d3plot file
10000 t 6.1120E-03 dt 5.91E-07 flush i/o buffers
– Identify the time reached by the analysis (TA), as marked above.

– Identify the time elapsed since the analysis started (TE) – use the last time the .otf
file was updated (this will actually include the initialization time).
– The termination time, specified in the analysis (TT).
– The estimated CPU time = (TT)/(TE)*(TA).
– Note: for a large model the initialization takes longer…
back to contents
LS-DYNA ENVIRONMENT
319
Running LS-DYNA - Controlling CPU Time LS-DYNA Introductory Course A

N
A
CPU requirements can be reduced by: L
– reducing the event time Y
– avoiding complicated or fully integrated elements S
I
– avoid small elements - beware of automeshers and meshes imported S
from implicit codes
– avoid sharp changes in spring characteristics
– use simple material formulations wherever possible
– use rigid bodies, shells, beams, and springs where possible
– reduce size of contacts and contact search frequencies
– use timestep optimisation by mass scaling - always check the % added
mass
back to contents
LS-DYNA ENVIRONMENT
320
LS-DYNA: Optimising the timestep LS-DYNA Introductory Course A

N
A
• Tune mass-scaling parameter DT2MS on *CONTROL_TIMESTEP to achieve L
acceptable added mass: Y
– Acceptable added mass depends on circumstances but typically 1%-2% S
of total model mass I
S
– Also check that the locations of added mass are acceptable e.g. not all
added in one place
• If added mass is unacceptably large:

– either reduce DT2MS and accept that the model will take longer to run
– consider re-meshing areas where much mass is added
back to contents
LS-DYNA ENVIRONMENT
321
Running LS-DYNA: SMP Parallel processing LS-DYNA Introductory Course A

N
A
3 L
Typical for crash models: 2.2 to 2.7 on 4 CPUs Y
2.5 S
Speed increase
I
2 S
1.5 Typical
Parallel
1 Speed-up
0.5
1 2 3 4 5 6 7 8
Number of CPU’s
The elapsed time of the analysis can be reduced by using parallel processing.
However, there is not much gain when using more than 4 processors for a single
job in the SMP version of LS-DYNA.
back to contents
LS-DYNA ENVIRONMENT
322
MPP Parallel processing speed up LS-DYNA Introductory Course A

N
A
MPP Performance of 1,090,000 Element Metal-stamping Model L
Y
S
40000 I
S
35000
30000
elapsed time (sec)
25000
20000
15000
10000
5000
0
1 2 4 8 16 32 64
number of processors
back to contents
LS-DYNA ENVIRONMENT
323
Outline – ANALYSIS LS-DYNA Introductory Course A

N
A
1.0 Checking L
2.0 Restarting the Model Y
S
I
S
back to contents
LS-DYNA ENVIRONMENT
324
Restarting LS-DYNA analyses LS-DYNA Introductory Course A

N
A
The Restart Process L
Y
• LS-DYNA, a transient code, can be stopped and restarted any time. This is S
I
useful if a job should stop due to a machine or disk crash or if a particular
S
section of an analysis needs more detailed results.
• Restarts are achieved using a binary restart “dump” file <name>.dpf

(d3dump) or <name>.adf (d3runsf) and an ASCII file <name>.key01 to
describe model changes (optional) at the restart time.
• User creates ASCII input file; binary dump file selected from File Selector
Dialogue.
back to contents
LS-DYNA ENVIRONMENT
325
Restarting LS-DYNA Analyses LS-DYNA Introductory Course A

N
A
Restart dump files - d3dump/*.dpf, or runrsf/*.adf L
– Family of dump files d3dump01, d3dump02 etc - limit these to avoid Y
excessive disk space usage S
I
– Single auxiliary dump file runrsf/.adf overwritten each time
S
– Output frequency controlled from within keyword file
(*DATABASE_BINARY_D3DUMP and
*DATABASE_BINARY_RUNRSF)
Types of Restart:
– Simple restart with no changes (no keyword input file needed)
– Small restart with some small changes (small restart keyword .key01 file
needed) - e.g. change termination time (see next slide)
– Full deck restarts allow users to make significant changes. Expert users
only! Major changes to model requires new keyword file .key01 with all
model data and additional changes. Add new keyword into restart file
*STRESS_INITIALIZATION
back to contents
LS-DYNA ENVIRONMENT
326
Restarting LS-DYNA Analyses LS-DYNA Introductory Course A

N
A
Simple restarts: keyword input file L
<name>.key01 file usually created by hand using a text editor. Example: Y
S
I
S
*KEYWORD 20000000 Memory - should be same as
*CONTROL_TERMINATION initial run
0.150 New termination time
*END
Select the restart time in the Oasys shell, then turn on the “input deck” button
to ask the shell to look for the .key01 input file.
Keywords allowed in simple restart keyword file are given in last section of the
Users Manual (Restart Input Data).
back to contents
LS-DYNA ENVIRONMENT
327
Workshop Example 13 LS-DYNA Introductory Course A

N
Units: mm, s, t, N, MPa A
1. Change the termination
L
time to 0.05 seconds and upper fork Y
run the model.
S
2. When the job has finished, I
create a restart file and lower fork S
change the termination
time back to 0.1 seconds in
a .key01 file. wheel
3. Restart the analysis and
check the ptf output in
D3PLOT
pivot
ground
back to contents
LS-DYNA ENVIRONMENT
328
Workshop Example 14 LS-DYNA Introductory Course
In this example a model will be set up and run.

The meshes of a plate and a ball are supplied. The following steps are necessary to set up
the analysis (mm, s, tone, N):
1. Define materials, using *MAT_PLASTIC_KINEMATIC (003):
a) ball: Density=7.89E-9t/mm3, Young‟s modulus=210,000MPa, Poisson‟s ratio=0.3,
Yield stress=200MPa, Tangent modulus=1000MPa;
b) plate: Density=7.89E-9t/mm3, Young‟s modulus=100MPa, Poisson‟s ratio=0.3, Yield
stress=1MPa, Tangent modulus=1MPa;
2. Define boundary conditions – constrain all edges of the plate.
3. Define initial velocity for the ball (2000mm/s) towards the plate.
4. Define contact between ball and plate.
5. Set control termination to 0.02s.
6. Set the output for the ASCII files GLSTAT, MATSUM and RCFORC to 1.0E-5.
7. Set the output for binary files D3PLOT to 1.0E-3 and D3THDT to 1.0E-5.
8. Run analysis and review results.
back to contents
LS-DYNA ENVIRONMENT

LS-DYNA Introductory Oct 30 Nov 1-2012

Uploaded by

Copyright:

Available Formats

LS-DYNA Introductory Oct 30 Nov 1-2012

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

LS-DYNA Introductory Oct 30 Nov 1-2012

Uploaded by

Copyright:

Available Formats

1

LS-DYNA Introductory Course

Oasys LS-DYNA Environment

The Arup Campus, Blythe Gate, Blythe Valley Park,

October 30th – November 1st, 2012

What does this course aim at? LS-DYNA Introductory Course

This course aims at providing the trainee with:

– An overview of LS-DYNA software

– An ability to prepare and run LS-DYNA input files

– An ability to debug LS-DYNA data

– An overview about post-processing the results

Assumptions LS-DYNA Introductory Course

This course includes a number of hands-on

Outline Of Training Course LS-DYNA Introductory Course

• BACKGROUND p6 • THEORY AND PRACTICE

Outline Of Training Course LS-DYNA Introductory Course

LS-DYNA Introductory Course B

LS-DYNA Development LS-DYNA Introductory Course B

LS-DYNA Capabilities LS-DYNA Introductory Course B

• Fluid / Structure interaction (ALE)

LS-DYNA Capabilities LS-DYNA Introductory Course B

LS-DYNA Applications LS-DYNA Introductory Course B

LS-DYNA Introductory Course I

Outline - INTRODUCTION LS-DYNA Introductory Course I

Analysis Process Overview LS-DYNA Introductory Course K

Keyword Format LS-DYNA Introductory Course K

Keyword Format - Example keyword file LS-DYNA Introductory Course K

• Any data after the *END is typically created/used by pre-processors to store P

Keyword Format LS-DYNA Introductory Course K

Keyword Format LS-DYNA Introductory Course K

Keyword Hierarchy LS-DYNA Introductory Course K

Keyword Hierarchy LS-DYNA Introductory Course K

Units in LS-DYNA LS-DYNA Introductory Course K

Preparing a model LS-DYNA Introductory Course R

• LS-DYNA specific pre-processors (Oasys PRIMER, LS-PrePost) or text

Oasys Primer LS-DYNA Introductory Course R

Single precision vs. Double precision LS-DYNA Introductory Course R

SMP vs. MPP LS-DYNA Introductory Course R

MPP – Massively Parallel Processing (or Distributed Memory Parallel)

MPP – Domain Decomposition LS-DYNA Introductory Course R

Submitting jobs LS-DYNA Introductory Course R

- e.g. for model1.key:

Additional options can be specified after i={input_file}.

Submitting jobs LS-DYNA Introductory Course R

Submitting jobs LS-DYNA Introductory Course R

Sense Switches LS-DYNA Introductory Course R

The jobname.kil file

File naming convention LS-DYNA Introductory Course O

LS-DYNA File Summary LS-DYNA Introductory Course O

Data Output LS-DYNA Introductory Course O

Binary File Families

Post-Processing LS-DYNA Introductory Course P

• To understand the behaviour of the model P

Post-Processing LS-DYNA Introductory Course P

Post-Processing LS-DYNA Introductory Course P

• Is the deformed shape (both overall and localized) realistic? P

Post-Processing LS-DYNA Introductory Course P

Data Evaluation - Model Post-Processing LS-DYNA Introductory Course P

Workshop Example 1 LS-DYNA Introductory Course

*PART *ELEMENT_SHELL *NODE (coordinates)

PART ELEMENT_SHELL *NODE (coordinates)