Transfer, Development, and Splice Length For Strand/Reinforcement in High-Strength Concrete (2008)

This PDF is available at http://nap.nationalacademies.
org/13916
Transfer, Development, and Splice Length for

Strand/Reinforcement in High-Strength
Concrete (2008)
DETAILS
122 pages | 8.5 x 11 | PAPERBACK
ISBN 978-0-309-11747-0 | DOI 10.17226/13916
CONTRIBUTORS
BUY THIS BOOK
FIND RELATED TITLES SUGGESTED CITATION

National Academies of Sciences, Engineering, and Medicine. 2008. Transfer,
Development, and Splice Length for Strand/Reinforcement in High-Strength
Concrete. Washington, DC: The National Academies Press.
https://doi.org/10.17226/13916.
Visit the National Academies Press at nap.edu and login or register to get:
– Access to free PDF downloads of thousands of publications
– 10% off the price of print publications
– Email or social media notiﬁcations of new titles related to your interests
– Special offers and discounts
All downloadable National Academies titles are free to be used for personal and/or non-commercial
academic use. Users may also freely post links to our titles on this website; non-commercial academic
users are encouraged to link to the version on this website rather than distribute a downloaded PDF
to ensure that all users are accessing the latest authoritative version of the work. All other uses require
written permission. (Request Permission)
This PDF is protected by copyright and owned by the National Academy of Sciences; unless otherwise
indicated, the National Academy of Sciences retains copyright to all materials in this PDF with all rights
reserved.
Transfer, Development, and Splice Length for Strand/Reinforcement in High-Strength Concrete
NATIONAL COOPERATIVE HIGHWAY RESEARCH PROGRAM
NCHRP REPORT 603

Transfer, Development, and
Splice Length for
Strand/Reinforcement in
High-Strength Concrete
Julio A. Ramirez
PURDUE UNIVERSITY
West Lafayette, IN
Bruce W. Russell
OKLAHOMA STATE UNIVERSITY
Stillwater, OK
Subject Areas
Bridges, Other Structures, Hydraulics and Hydrology
Research sponsored by the American Association of State Highway and Transportation Officials
in cooperation with the Federal Highway Administration
TRANSPORTATION RESEARCH BOARD

WASHINGTON, D.C.
2008
www.TRB.org
Copyright National Academy of Sciences. All rights reserved.

NATIONAL COOPERATIVE HIGHWAY NCHRP REPORT 603

RESEARCH PROGRAM
Systematic, well-designed research provides the most effective Project 12-60
approach to the solution of many problems facing highway ISSN 0077-5614
administrators and engineers. Often, highway problems are of local ISBN: 978-0-309-11747-0
interest and can best be studied by highway departments individually Library of Congress Control Number 2008907268
or in cooperation with their state universities and others. However, the © 2008 Transportation Research Board
accelerating growth of highway transportation develops increasingly
complex problems of wide interest to highway authorities. These
problems are best studied through a coordinated program of COPYRIGHT PERMISSION
cooperative research.
Authors herein are responsible for the authenticity of their materials and for obtaining
In recognition of these needs, the highway administrators of the written permissions from publishers or persons who own the copyright to any previously
American Association of State Highway and Transportation Officials published or copyrighted material used herein.
initiated in 1962 an objective national highway research program Cooperative Research Programs (CRP) grants permission to reproduce material in this
employing modern scientific techniques. This program is supported on publication for classroom and not-for-profit purposes. Permission is given with the
understanding that none of the material will be used to imply TRB, AASHTO, FAA, FHWA,
a continuing basis by funds from participating member states of the
FMCSA, FTA, or Transit Development Corporation endorsement of a particular product,
Association and it receives the full cooperation and support of the method, or practice. It is expected that those reproducing the material in this document for
Federal Highway Administration, United States Department of educational and not-for-profit uses will give appropriate acknowledgment of the source of
any reprinted or reproduced material. For other uses of the material, request permission
Transportation.
from CRP.
The Transportation Research Board of the National Academies was
requested by the Association to administer the research program
because of the Board’s recognized objectivity and understanding of
NOTICE
modern research practices. The Board is uniquely suited for this
purpose as it maintains an extensive committee structure from which The project that is the subject of this report was a part of the National Cooperative Highway
Research Program conducted by the Transportation Research Board with the approval of
authorities on any highway transportation subject may be drawn; it the Governing Board of the National Research Council. Such approval reflects the
possesses avenues of communications and cooperation with federal, Governing Board’s judgment that the program concerned is of national importance and
state and local governmental agencies, universities, and industry; its appropriate with respect to both the purposes and resources of the National Research
Council.
relationship to the National Research Council is an insurance of
The members of the technical committee selected to monitor this project and to review this
objectivity; it maintains a full-time research correlation staff of
report were chosen for recognized scholarly competence and with due consideration for the
specialists in highway transportation matters to bring the findings of balance of disciplines appropriate to the project. The opinions and conclusions expressed
research directly to those who are in a position to use them. or implied are those of the research agency that performed the research, and, while they have
been accepted as appropriate by the technical committee, they are not necessarily those of
The program is developed on the basis of research needs identified
the Transportation Research Board, the National Research Council, the American
by chief administrators of the highway and transportation departments Association of State Highway and Transportation Officials, or the Federal Highway
and by committees of AASHTO. Each year, specific areas of research Administration, U.S. Department of Transportation.
needs to be included in the program are proposed to the National Each report is reviewed and accepted for publication by the technical committee according
Research Council and the Board by the American Association of State to procedures established and monitored by the Transportation Research Board Executive
Committee and the Governing Board of the National Research Council.
Highway and Transportation Officials. Research projects to fulfill these
needs are defined by the Board, and qualified research agencies are The Transportation Research Board of the National Academies, the National Research
Council, the Federal Highway Administration, the American Association of State Highway
selected from those that have submitted proposals. Administration and and Transportation Officials, and the individual states participating in the National
surveillance of research contracts are the responsibilities of the National Cooperative Highway Research Program do not endorse products or manufacturers. Trade
Research Council and the Transportation Research Board. or manufacturers’ names appear herein solely because they are considered essential to the
object of this report.
The needs for highway research are many, and the National
Cooperative Highway Research Program can make significant
contributions to the solution of highway transportation problems of
mutual concern to many responsible groups. The program, however, is
intended to complement rather than to substitute for or duplicate other
highway research programs.
Published reports of the
NATIONAL COOPERATIVE HIGHWAY RESEARCH PROGRAM

are available from:
Transportation Research Board

Business Office
500 Fifth Street, NW
Washington, DC 20001
and can be ordered through the Internet at:

http://www.national-academies.org/trb/bookstore
Printed in the United States of America


COOPERATIVE RESEARCH PROGRAMS
CRP STAFF FOR NCHRP REPORT 603

Christopher W. Jenks, Director, Cooperative Research Programs
Crawford F. Jencks, Deputy Director, Cooperative Research Programs
David B. Beal, Senior Program Officer
Eileen P. Delaney, Director of Publications
Ellen M. Chafee, Assistant Editor
NCHRP PROJECT 12-60 PANEL

Field of Design—Area of Bridges
R. Scott Christie, Pennsylvania DOT (Chair)
Theresa Ahlborn, Michigan Technological University
Thomas Beitelman, Sound Structures Engineering, Inc., Tallahassee, FL
Vijay Chandra, PB Americas, Inc., Herndon, VA
Allan W. Frank, Kentucky Transportation Cabinet
Gregg A. Freeby, Texas DOT
Bijan Khaleghi, Washington State DOT
David H. Sanders, University of Nevada—Reno
Joey Hartmann, FHWA Liaison
Eric P. Munley, FHWA Liaison
Stephen F. Maher, TRB Liaison
AUTHOR ACKNOWLEDGMENTS
The research reported herein was performed under NCHRP Project 12-60 by Purdue Research Foun-
dation through the School of Civil Engineering at Purdue University with subcontracting services being
provided by the School of Civil and Environmental Engineering of Oklahoma State University (OSU).
The Principal Investigator on this project was Julio A. Ramirez of Purdue University. The other author
of this report was Bruce W. Russell of Oklahoma State University. The work was done under the general
supervision of Julio A. Ramirez. The work at OSU was under the supervision of Bruce W. Russell.

FOREWORD
By David B. Beal
Staff Officer
Transportation Research Board
This report documents research performed to develop recommended revisions to the

AASHTO LRFD Bridge Design Specifications to extend the applicability of the transfer, devel-
opment, and splice length provisions for prestressed and non-prestressed concrete mem-
bers to concrete strengths greater than 10 ksi. The report details the research performed and
includes recommended revisions to the AASHTO LRFD Bridge Design Specifications. The
material in this report will be of immediate interest to bridge designers.
The AASHTO LRFD Bridge Design Specifications contain barriers to the use of high-
strength concrete. These barriers restrict the application of existing and new technology
to bridges. The AASHTO LRFD Bridge Design Specifications state that design concrete
compressive strengths above 10 ksi shall be used only when allowed by specific articles or
when physical tests are made to establish the relationships between the concrete strength
and other properties. When the AASHTO LRFD Bridge Design Specifications were written,
there was a lack of data to demonstrate that the provisions were applicable to concrete
compressive strengths above 10 ksi. Recent research has started to address design issues
with higher strength concretes. FHWA Showcase Projects encourage the use of high-
performance concretes—including high-strength concrete—in bridge structures. As the
industry moves toward the use of high-strength concrete, the need to revise the AASHTO
LRFD Bridge Design Specifications is more urgent. There is, therefore, a need to expand the
AASHTO LRFD Bridge Design Specifications to allow greater use of high-strength concrete.
Two recent NCHRP reports complement the work accomplished in NCHRP Project 12-60
in removing barriers to the use of high-strength concrete. NCHRP Report 579: Application of
LRFD Bridge Design Specifications to High-Strength Structural Concrete: Shear Provisions iden-
tifies all barriers in the AASHTO LRFD Bridge Design Specifications to the use of high-strength
concrete and provides research findings to remove the barriers related to shear. NCHRP Report
595: Application of the LRFD Bridge Design Specifications to High-Strength Structural Concrete:
Flexure and Compression Provisions addresses flexural and compression issues. Recommenda-
tions from these reports have already been adopted into the AASHTO LRFD Bridge Design
Specifications.
The objective of NCHRP Project 12-60 was to develop recommended revisions to the
AASHTO LRFD Bridge Design Specifications for normal-weight concrete having compres-
sive strengths up to 15 ksi, relating to transfer and development length of prestressing strand
with diameters up to 0.62 in. and development and splice length in tension and compres-
sion of individual bars, bundled bars, and welded wire reinforcement and development
length of standard hooks. This research was performed by Purdue University and Oklahoma
State University. The report fully documents the research leading to the recommended revi-
sions to Section 5 of the AASHTO LRFD Bridge Design Specifications. AASHTO is expected
to consider these recommendations for adoption in 2008.

CONTENTS
1 Summary
4 Chapter 1 Introduction and Research Approach
4 1.1 Problem Statement and Research Objective
4 1.2 Research Approach
6 Chapter 2 Literature Review
6 2.1 Introduction
7 2.2 Literature Review
18 2.3 Identification of Issues and Needs
21 2.4 Issues Related to Testing Protocols
25 2.5 Summary
26 Chapter 3 Experimental Program and Results
26 3.1 Introduction to the Experimental Program
26 3.2 The Standard Test Method for the Bond of Prestressing Strands
31 3.3 The NASP Bond Test in Concrete
38 3.4 Measured Transfer Lengths versus Varying Concrete Strengths
and Varying NASP Bond Test Values
54 3.5 Development Length Tests
72 3.6 Discussion of Design Recommendations
77 3.7 The Effect of Concrete Strength on Bond Performance—
Summary, Conclusions, and Recommendations
78 3.8 Experimental Program—Mild Steel Anchorage of Uncoated Bars
in Tension
89 3.9 Anchorage of Epoxy-Coated Bars in Tension
99 3.10 Anchorage of Bars Terminated with Standard Hooks in Tension
110 Chapter 4 Design Recommendations
110 4.1 Introduction
110 4.2 Design Recommendations
111 4.3 Details of the Design Recommendations
119 References
122 Appendices

SUMMARY
Transfer, Development, and Splice Length

for Strand/Reinforcement in
High-Strength Concrete
Article 5.4.2.1 of the 3rd edition of the AASHTO LRFD Bridge Design Specifications limits
the applicability of the specifications for concrete compressive strengths of 10,000 psi or less
unless physical tests are made to establish the relationships between concrete strength and
other properties (AASHTO 2004). A comprehensive, article-by-article review of Section 5 of
the AASHTO LRFD Bridge Design Specifications pertaining to transfer, development, and splice
length for strand/reinforcement was performed under NCHRP Project 12-60 to identify all the
provisions that have to be revised to extend their use to high-strength, normal-weight concrete
up to 15 ksi. Upon completion of the experimental work under NCHRP Project 12-60, draft
specifications and accompanying commentary for provisions to extend the application of the
LRFD bridge design specifications to high-strength concrete were developed. The provisions
cover the transfer and development length of prestressing strand and the development and
splice length of reinforcement in normal-weight concrete with compressive strengths up to
15 ksi. Researchers from Purdue University and Oklahoma State University have jointly pre-
pared this report.
Transfer Length and Development Length for Strand

Recommendations include new transfer length and development length equations for
incorporation into Articles 5.11.4.1 and 5.11.4.2 of the AASHTO LRFD Bridge Design
Specifications. Also, a new requirement is introduced for addition to Article 5.4.4.1 for the
purpose of qualifying the basic bonding properties of prestressing strand.
Article 5.4.4.1 addresses the material properties of prestressing strand. Heretofore, Article
5.4.4.1 addressed the mechanical properties of strand only, i.e., breaking strength, yield
strength, and strand size. Based on research described in this report, a “Standard Test
Method for the Bond of Prestressing Strands” (also called the “Standard Test for Strand
Bond”) is recommended for inclusion by reference in Article 5.4.4.1. Details for testing
procedures and material acceptance are included in Appendix H. The Standard Test Method
for the Bond of Prestressing Strand requires that prestressing strands obtain an average
minimum pull-out value of 10,500 lb for 0.5-in. strands and 12,600 lb for 0.6-in. strands.
Further, the research supports, and this report recommends, that transfer length and
development length equations include a parameter for concrete strength. The research
shows a clear correlation between shortening of transfer and development lengths and in-
creasing concrete strength. Therefore, a new transfer length expression is recommended
for inclusion into Article 5.11.4.2 of the AASHTO LRFD Bridge Design Specifications:
⎡ 120db ⎤
lt = ⎢ ⎥ ≤ 40db
⎣ f'ci ⎦

where
db = strand diameter,
f c′i = concrete strength at release, and
lt = transfer length.
At a concrete release strength of 4 ksi, the recommended expression provides for a transfer
length of 60 strand diameters, which matches historic design procedures. The recommended
expression also provides for a transfer length of at least 40 strand diameters, effectively lim-
iting the benefits from release strength to about 9 ksi.
Increases in concrete strength also result in shorter development lengths. Therefore a new
development length expression is recommended for inclusion in Article 5.11.4.3 of the
AASHTO LRFD Bridge Design Specifications:
⎡ 120 225 ⎤
ld = ⎢ + ⎥ db ≥ 100db
⎣ f ci′ f c′ ⎦
This expression provides a development length of about 150 strand diameters for con-
crete with release strength of 4 ksi and design strength of 6 ksi. The expression is different in
form than the current expression, but more “user friendly” to the designer. In the develop-
ment length equation, ld is the development length (in.), db is the strand diameter (in.), f c′i is
the concrete strength at release (ksi), and f c′i is the concrete design strength (ksi). The
expression provides for a development length of at least 100 strand diameters.
Recommendations are also made to revise the part of Article 5.11.4.3 of the AASHTO
LRFD Bridge Design Specifications dealing with debonded, or shielded, strands. In brief, rec-
ommendations contained in this report would remove the 2.0 multiplier applied to
debonded strands, but add some restrictions to the use of debonded strands.
Development Length and Splice Length

for Reinforcement
The proposed recommendations stemming from the work conducted under NCHRP
Project 12-60 cover two aspects for mild steel:
1. Development length of black and epoxy-coated reinforcing bars anchored by means of

straight embedment length and splices and
2. Development length of black and epoxy-coated bars terminated with a standard hook.
Based on observations from tests conducted during NCHRP Project 12-60 on 18 top cast
beam-splice specimens and the examination of an extensive database of previous tests
compiled by ACI Committee 408, it is proposed that extension of the AASHTO LRFD Bridge
Design Specifications to concrete strengths up to 15 ksi follow a format similar to the one used
in ACI: 318-05: Building Code Requirements for Structural Concrete and Commentary
(ACI 2005), with the following exceptions:
• Removal of the bar size factor for #6 bars and smaller bars (thus γ = 1.0 in all cases).
• Use of a single factor for epoxy-coated bars of 1.5 regardless of the ratio of cover to bar diameter.
• Exclusion of evaluations of beam splice specimens with bottom cast bars in this study. ACI
Committee 408 has indicated that the current approach in the 318 Code (ACI 2005) overes-
timates the bar force at failure in many specimens with bottom bars available in the ACI
Committee 408 database, especially for specimens with concrete compressive strengths greater
that 10 ksi (ACI Committee 408 2003). ACI Committee 408 proposed a modified expression

for development and splice length in addition to removal of the bar size factor to address this
issue. In the evaluation of test data conducted under NCHRP Project 12-60, the researchers
found that the use of a bottom cast modification factor of 1.2 for uncoated bars anchored in
concrete with compressive strengths greater than 10 ksi appeared to address the safety
concerns raised by ACI Committee 408. This factor would not be needed for bottom cast
epoxy-coated bars (because of the single modification factor of 1.5) or for uncoated top bars.
This approach could be used as an alternative to the approach suggested by ACI Committee
408. The researchers note that additional testing of bottom cast uncoated splices is justified
with higher strength concretes.
Article 5.11.2.4 of the AASHTO LRFD Bridge Design Specifications (AASHTO 2004) was
verified for high-strength concrete in the experimental work plan for NCHRP Project 12-60
with the exception of the lightweight aggregate factor. Based on the analysis of tests
conducted during NCHRP Project 12-60 (21 full-scale tests of hooked bar anchorages) and
the analysis of tests of additional specimens in the literature, it is possible to support the
extension of the approach in the 318 Code (ACI 2005) provision for anchorage of bars
terminated with standard hooks, black and epoxy-coated, to normal-weight concrete with
concrete compressive strength of up to 15 ksi, with these two modifications:
1. A minimum amount of transverse reinforcement (at least #3 U bars at 3db spacing) needs to
be provided to improve the bond strength of both epoxy-coated and black #11 bars and larger
bars in tension anchored by means of standard hooks.
2. A modification factor of 0.8 instead of the current factor of 0.7 for #11 and smaller hooks with
side cover (normal to plane of hook) not less than 2.5 in. and for 90-deg hooks with cover on
bar extension beyond hook not less than 2 in.
Tension lap splices were also evaluated under NCHRP Project 12-60. Splices of bars in
compression were not part of the experimental program. Class C splices were eliminated
based on the modifications to development length provisions. The proposed modifications
to Article 5.11.2.1 of the AASHTO LRFD Bridge Design Specifications contain several changes
that eliminated many of the concerns regarding tension splices due to closely spaced bars
with minimal cover; however, development lengths, on which splice lengths are based, have
in some cases increased. A two-level splice length was retained primarily to encourage
designers to splice bars at points of minimum stress and to stagger splices to improve
behavior of critical details; however, such provisions are not intended to reflect the strength
of the splice.

CHAPTER 1
Introduction and Research Approach
1.1 Problem Statement The knowledge gained and the resulting improvements to
and Research Objective the 4th edition of the AASHTO LRFD Bridge Design Specifi-
cations will assist engineers in the safe design of high-strength
This final report documents research findings regarding concrete bridge members (AASHTO 2007).
the transfer length and development length of prestressing The 4th edition of the AASHTO LRFD Bridge Design Spec-
strand and the development length and splice length of rein- ifications provides a good starting point for higher strength
forcement in normal-weight concrete with compressive concrete bridge members (AASHTO 2007). However, the
strengths up to 15 ksi. Recommended revisions to the 4th edi- existing provisions were optimized for concrete strengths be-
tion of the AASHTO LRFD Bridge Design Specifications were tween 5 and 10 ksi and require modifications if they are to be
also developed in the research (AASHTO 2007). extended to higher strength concretes. Because of its brittle
Structural engineers have continually tried to optimize nature, high-strength concrete must be properly utilized in
building materials by improving their durability and effec- both design and construction. The more brittle nature of
tiveness. An example of such efforts is the use of high-strength high-strength concrete means that if cracks form, they may
concrete in bridge members. High-strength concrete, defined propagate more extensively than they would in lower
for this report as concrete having a compressive strength in ex- strength concretes. This may result in the loss of effectiveness
cess of 10 ksi, is more brittle than normal-strength concrete; of the concrete cover and raise safety concerns regarding the
consequently, the designer has to be cautious in extending em- bond strength of strand and deformed bars embedded in
pirically based rules to this new material. The 2004 AASHTO high-strength concrete.
LRFD Bridge Design Specifications (with 2005 and 2006 in-
terim revisions) states that concrete with compressive
1.2 Research Approach
strengths above 10 ksi should be used only when physical tests
are made to establish the relationships between concrete Chapter 2 of this report provides a review of relevant prac-
strength and other properties (AASHTO 2004). This require- tice, performance data, research findings, and other relevant
ment reflects the lack of data to demonstrate that the provi- information related to the transfer and development length
sions were applicable to concrete strengths above 10 ksi at the of prestressing strand and the development and splice length
time the specification was written. of mild reinforcement. This chapter presents an information
The objective of this study was to develop recommended summary on the factors that affect the behavior of transfer
revisions as appropriate to the 4th edition of the AASHTO and development length of prestressing strand and the devel-
LRFD Bridge Design Specifications for normal-weight concrete opment and splice length of mild reinforcement. Chapter 2
having compressive strengths up to 15 ksi, relating to the includes a critical review of existing testing protocols for
following (AASHTO 2007): determining bond characteristics and presents the testing
protocols used in NCHRP Project 12-60.
• Transfer and development length of prestressing strand Article 5.4.2.1 of the 3rd edition of the AASHTO LRFD
with diameters up to 0.62 inches; and Bridge Design Specifications limits the applicability of the spec-
• Development and splice length in tension and compres- ifications to concrete compressive strengths of 10,000 psi or
sion of individual bars, bundled bars, and welded wire less unless physical tests are made to establish the relationships
reinforcement and development length of standard hooks. between concrete strength and other properties (AASHTO

2004). A comprehensive, article-by-article review of Section 5 by both of these reports as having the potential for preventing
of the 2nd edition of the AASHTO LRFD Bridge Design Speci- the extension of development and splice length specifications
fications (with Update 2000) was performed under NCHRP to high-strength concrete were reviewed and considered dur-
Project 12-56 to identify all the provisions that directly or ing the work conducted under NCHRP Project 12-60.
indirectly have the potential to prevent the extension of the The results of the expanded plan of research are included in
specifications in their current form to high-strength concrete Chapter 3. The findings of the extensive literature review and
(AASHTO 1998; Hawkins and Kuchma 2002). In addition, a the experimental program conducted under NCHRP Project
comprehensive, article-by-article review of Section 5 of the 12-60 were used to recommend revisions as appropriate to
2nd edition of the AASHTO LRFD Bridge Design Specifications the AASHTO LRFD Bridge Design Specifications for transfer,
with the 1999, 2000, and 2001 interim revisions (AASHTO development, and splice lengths for strand/reinforcement in
1998) was performed as part of the FHWA study, Compilation normal-weight concrete having compressive strengths up to
and Evaluation of Results from High Performance Concrete 15 ksi. Chapter 4 contains design recommendations supported
Bridge Projects (Russell et al. 2006). The provisions identified by the research conducted under NCHRP Project 12-60.

CHAPTER 2
Literature Review
2.1 Introduction issues that may influence the transfer and development of pre-
stressing strands include confining reinforcement and strand
The comprehensive and critical literature review under- spacing. The research reported in the literature indicates that
taken during NCHRP Project 12-60 is described in this chap- 0.6-in. strand can be spaced at 2.0 in. center to center (c/c) or
ter. In this report, important findings from prior research are that 0.5-in. strands may be spaced at 1.75 in. without penalty
reviewed, with particular attention to the impact of these to the transfer and development of strands. The reported
findings on the work plan for NCHRP Project 12-60. research also indicates that confining reinforcement has little
The objective of the work related to prestressing strand was or no effect on transfer length of strands, but it can be quite
to gather and synthesize existing data and information on the beneficial to strand development. Standardized confinement
transfer length and development length of strand with diam- details were employed in beam testing where warranted.
eters up to 0.6 in. In the area of mild reinforcement, the effort In the area of bond of mild reinforcement, the single most
concentrated on development and splice length in tension of important issue not currently accounted for in the AASHTO
individual bars and development length of bars in tension an- LRFD Bridge Design Specifications is the effect of confining
chored with standard hooks. The database constructed from reinforcement on the bond strength of tension reinforcement
this effort includes 71 tension development and splice tests of in the case of splitting type failures. This parameter is espe-
specimens with top cast uncoated reinforcing bars, 493 speci- cially important as bars are being developed in higher strength
mens with bottom cast uncoated reinforcing bars, 27 speci- concretes. The review also revealed that for epoxy-coated bar
mens with top cast epoxy-coated bars, and 48 with bottom development/splice length and development length of bars
cast epoxy-coated bars. In addition, 33 specimens with un- terminated with standard hooks there is a paucity of data on
coated bars terminated with standard hooks and 13 specimens concretes with cylinder strengths above 10 ksi.
reinforced with epoxy-coated bars have been reviewed. A significant effort during the initial 6 months of the
A comprehensive analysis of the data collected was con- NCHRP Project 12-60 study was focused on identifying and
ducted to identify issues and needs related to bond of strand evaluating testing protocols related to the experimental work
and mild steel in high-strength concrete. This analysis assisted to be conducted. In the area of bond in concrete of prestress-
in the identification of several key variables that are likely to ing strand, particular attention was given to the surface char-
affect the transfer and development of prestressing strands, acterization tests to evaluate strand “bond-ability.” There are
development/splice length of bars in tension, and develop- three tests that have been offered in recent years as possible
ment length of bars in tension terminated with standard tests to standardize acceptance of strand based on its “bond-
hooks. Some of these variables are currently included in the ability”: (1) the Moustafa Test, where untensioned strands
AASHTO LRFD Bridge Design Specifications while some are are pulled from large concrete blocks; (2) the PTI Bond Test,
not. In the area of transfer length and development length of where untensioned strands are pulled from a neat cement
prestressing strand, specifications do not account for variables mortar; and (3) the NASP Bond Test, where untensioned
such as concrete strength, strand size, “top bar” effects, epoxy strands are pulled from a sand cement mortar. In testing per-
coating, bond quality of individual strand samples, and struc- formed by the North American Strand Producers (NASP),
tural behavior issues (e.g., the interaction of shear and bond). the NASP Bond Test has proven to be the most reliable test of
The work plan for NCHRP Project 12-60 included procedures the three. It has produced test results from “blind trials” with
and testing to evaluate some, but not all, of these effects. Other the best repeatability and reproducibility.

In three separate rounds of testing, the Moustafa Test (now to directly measure development and splice strength in full-
called the Large Block Pull-Out Test) was performed at differ- scale members.
ent sites to determine its reproducibility across sites. In NASP The experimental work supporting the current require-
Round I testing, the Moustafa Test was performed at Coreslab ments in the AASHTO LRFD Bridge Design Specifications
Structures in Colorado and at Florida Wire and Cable Co. and ACI 318-05: Building Code Requirements for Structural
(FWC). For the purpose of carrying out Moustafa Tests, FWC Concrete and Commentary (2005) for development of stan-
built a completely automated testing machine so that the dard hooks in tension was conducted using a test setup rep-
Moustafa procedures could be precisely followed. Round I test- resenting an exterior beam column joint. Because of the
ing showed widely dissimilar results from the two testing sites. paucity of data on concrete strengths above 10 ksi, the evalu-
In NASP Round II, the Moustafa Test and the PTI Bond Tests ation of uncoated and epoxy-coated bars terminated with
were performed at three testing sites: Coreslab Structures, FWC, standard hooks in tension to normal-weight concrete with
and the University of Oklahoma. Additionally, the NASP Bond compressive strength up to 15 ksi was performed using a sim-
Test was introduced in an early form as a test very similar to the ilar exterior beam column joint test setup.
PTI Bond Test except that a sand-cement mortar was used. The results of the initial work of NCHRP Project 12-60
Seven different strand samples were shipped to the different confirmed the basic premises stated in the original project
testing sites. The trials were blind. Again, the Moustafa Test proposal. Thus, the efforts of the experimental program and
failed to produce reproducible results across testing sites. Of the the order of priority of these efforts remained as originally
three tests, the NASP Bond Test showed the highest statistical stated. The experimental program focused on the following
correlation across testing sites. In the NASP Round III testing, major efforts listed in priority order:
a more refined version of the NASP Bond Test again outper-
formed the Moustafa Test in blind trials at the three testing sites. 1. Determining design equations for transfer and develop-
In all three rounds of testing, when the Moustafa Test was used, ment length of strand in prestressed concrete bridge mem-
it failed to produce results that were consistent across sites. The bers. Variables included concrete strength at release, con-
NASP Bond Test proved more reliable at providing the same or crete strength at time of development length testing, use of
similar results across testing sites in Rounds II, III, and IV. Be- air-entraining admixtures, “top bar effects,” and strand size.
cause of the NASP Bond Test’s more consistent results, the 2. Development and splice length in tension of reinforcing
NCHRP Project 12-60 testing program was built upon the bars. Variables included concrete strength, bar size, con-
NASP Bond Test. crete cover/bar spacing, amount of transverse reinforce-
The review conducted on testing for development/splice ment, epoxy coating, and casting position.
length of deformed bars in tension showed that the generally 3. Development length in tension of bars terminated with
recommended testing protocol for full-scale specimens because standard hooks. Variables included concrete strength, bar
of both the relative ease of fabrication and the realistic state of size, concrete cover/bar spacing, amount of transverse
stress achieved during testing is the beam-splice specimen. reinforcement, and epoxy coating.
Thus, beam splice specimens were used in the development of
experimental data related to development/splice length of mild A comprehensive article-by-article review of Section 5 of
reinforcement during the course of the NCHRP Project 12-60 the 2nd edition of the AASHTO LRFD Bridge Design Specifi-
study. It is well established that testing protocols to evaluate de- cations with the 1999, 2000, and 2001 interim revisions
velopment and splice length requirements for deformed bars (AASHTO 1998) was conducted during the initial 6 months
and wire in tension must be of an appropriate scale, containing of the NCHRP Project 12-60 study. In this review, the provi-
more than one bar or wire, with due regard for a realistic trans- sions of Section 5 that directly or indirectly affect transfer and
fer of force between concrete and steel reinforcement and development length of prestressing strand and develop-
cover/bar spacing effects. Splice tests have in the past been ac- ment/splice length of mild reinforcement by the use of high-
curate simulations of real conditions in structures; however, de- strength concrete were extracted and critically reviewed.
velopment length tests have been largely conducted using pull-
out tests, in which splitting failures are purposely avoided. As a
result, the bond stresses developed along splices are low com- 2.2 Literature Review
pared with the bond stresses along a bar in a pull-out test. This 2.2.1 Strand Transfer and Development
difference in test methods is responsible for large differences in Length
code-required anchorage lengths for splices and development
of single bars. Pull-out failures occur in cases of high confine- A number of experimental investigations related to high-
ment and short bonded lengths. In most structural applications, strength concrete have been conducted in North America and
however, splitting failures tend to control. Beam-splice speci- overseas. Hence, a significant body of knowledge currently ex-
mens are deemed to represent larger-scale specimens designed ists with respect to the performance of high-strength concrete

members. Amongst the experimental data are various studies large number of research programs intent on measuring the
dealing with transfer length and development length of pre- transfer and development of prestressing strands. Research
stressing strand and splice length and development length of was performed at the University of Texas (Russell and Burns
black and epoxy-coated reinforcement. In this study, a com- 1996, 1997), Florida DOT (Shahawy, Issa, and Batchelor
prehensive and critical literature review was undertaken to 1992), McGill University (Mitchell et al. 1993), and Auburn
gather and synthesize existing data and information related to University (Cousins et al. 1993). The arbitrary 1.6 multiplier
the transfer length and development length of prestressing from the original FHWA moratorium is now incorporated
strand with diameters up to 0.6 in., and development and into the AASHTO LFRD Bridge Design Specifications.
splice length in tension and compression of individual bars, By the mid-1990s, it became apparent that the studies ex-
bundled bars, and welded wire reinforcement and develop- amining the transfer and development of prestressing strands
ment length of standard. had not resulted in a consensus on design standards. As a
The literature review centered on collecting information whole, the research displayed a large scatter of the test results,
on testing protocols for determining surface bond character- with measured transfer lengths for 0.5-in. strand ranging
istics of strand, performance of members containing trans- from a low of less than 20 in. to a high of more than 60 in.
verse reinforcement, bond and transfer length, and tests Thus, it became apparent that other variables were in play
addressing deformation capacity. Information available from and that such variables were not properly accounted for in
the field—including FHWA showcase projects and the either design equations or specifications.
unpublished experiences of engineers, bridge owners, and Since the mid-1990s, research work has concentrated on
producers—was reviewed and used to supplement other developing a standardized test to assess the bond characteris-
work conducted in this study. tics of individual prestressing strands. It was suspected that
The development of reliable code expressions for transfer different strand manufacturers produced strand with quite
and development of prestressing strand is made more diffi- dissimilar bonding characteristics. Hence, it was important to
cult by the large experimental scatter reported by researchers quantify the bonding characteristics of an individual strand
over the past 40 years. The original code expressions for before the transfer length and development length data
transfer and development length of pretensioned strands would be meaningful. To that end, three or four different
were developed from testing performed in the late 1950s and testing programs were undertaken to assess the viability of
early 1960s on Grade 250, stress-relieved strand (Hanson and various “standardized tests” and the suitability of such tests
Kaar 1959; Kaar, LaFraugh, and Mass 1963; Tabatabai and for predicting the “bond-ability” of prestressing strand.
Dickson 1993). Based on these early tests, the ACI Building The first such testing program was developed by Rose and
Code (ACI 2005) and the AASHTO LRFD Bridge Design Russell (1997). The various testing programs found little
Specifications adopted provisions governing the design for correlation between a “simple pull-out” test and measured
strand transfer and development. Manufacturing innovation transfer lengths. From these research programs, the precast
has brought about Grade 270 low-relaxation strand as the in- concrete industry adopted a set of standard test procedures
dustry standard, while the code expressions for transfer and that were to be employed in performing “pull-out” tests. The
development length have changed very little. set adopted is known as the “Moustafa Test.” Early results
Furthermore, contemporary strand production employs using the Moustafa Test indicated that the test could be used
induction heating to stress relieve strand, whereas convection to compare the bonding characteristics of strand on a relative
heating was used in the late 1950s and early 1960s. Convec- basis. Logan (1997) demonstrated that the Moustafa Test, at
tion heating created hotter surface temperatures on strand the recommended threshold value, would provide strand
that may have burned off much of the surface residues with bonding capability more than adequate to meet current
remaining from the wire drawing process. Today’s processes, design assumptions.
using induction heating, may have created surface tempera- In the meantime, the Post-Tensioning Institute commis-
tures lower than those created by convection heating and sioned a study at Queen’s University in Ontario (Hyett,
thereby may have effectively changed the bonding character- Dube, and Bawden 1994). The study produced yet another
istics of the surface of prestressing strands (Rose and Russell bond test, the “PTI Bond Test.” The PTI Bond Test’s primary
1997). purpose was to assess the bond characteristics of 0.6 in. di-
In the mid-1980s, Cousins, Johnston, and Zia (1990) ameter strand and show the strand’s suitability for use as a
measured transfer lengths that exceeded the standard design rock anchor. In an appendix to ASTM A 416, the ASTM has
predictions by a wide margin. Their findings led FHWA to adopted the PTI Bond Test on a provisional basis for 0.6 in.
adopt a moratorium on the use of 0.6 in. diameter strands diameter strand that is to be used as rock anchors.
and to increase the development length for other sizes of pre- Subsequent testing sponsored by the North American
stressing strands. The FHWA action led to the creation of a Strand Producers Association (NASP) led to the development

of a third bond test, now called the “NASP Bond Test” portation agencies of the experimental results on transfer
(Russell and Paulsgrove 1999b). In “blind trial” testing, the length and development length of strand in concrete.
Moustafa Test, the PTI Bond Test, and the NASP Bond Test Round II of the NASP tests examined the proposed stan-
were performed at multiple sites. The results of the blind trial dardized tests for repeatability and reproducibility. The re-
testing indicated that the NASP Bond Test provided the best sults clearly indicated that the NASP Bond Test was the most
repeatability. Based on these results and on as yet unpub- reliable test of the three tests examined. Results from the
lished results from NASP Round III testing, the NASP Moustafa Test are shown in Figures 2.1 and 2.2. Note that in
recommended the use of the NASP Bond Test as the stan- the Moustafa Test, results from a majority of strands tended
dardized test to assess the bond characteristics of prestressing to cluster near the threshold level, and a more poorly
strands. Overall, experimental results clearly show that performing strand was inconsistently rated. In a similar plot,
inherent quality differences exist in the bond of prestressing Figure 2.3 compares results from two different test series per-
strands from various manufacturers. Accordingly, it is im- formed at the University of Oklahoma (OU) featuring the
perative in a testing program to evaluate the bonding charac- NASP Bond Test. Finally, Figure 2.4 compares the NASP
teristics of the prestressing strands used. The standardization Bond Test results at two different test sites. The repro-
process will make possible nationwide adoption by trans- ducibility of test results proved to be quite remarkable and
50
r2 = 0.92
STRESSCON=(0.597)OU+14.0 C
STRESSCON DATA (kips)
40
M K B PW
A
30 Z
20 J
"PERFECT" TEST
10
0
0 10 20 30 40 50
OU DATA (kips)
Figure 2.1. Comparison of Moustafa pull-out values from

Stresscon and OU.
50
40
"PERFECT" TEST
FWC DATA (kips)
C
30
K B PW
A M
20 Z
J r2 = 0.88
10 FWC=(0.625)OU+5.25
0
0 10 20 30 40 50
OU DATA (kips)
Figure 2.2. Comparison of Moustafa pull-out values from

FWC and OU.

10
25
r2 = 0.97
OU SERIES TWO DATA (kips)

20 SERIES TWO = (0.83)SERIES ONE + 2.34
P C
A
K M
15
W B
10
Z
5
J
"PERFECT" TEST
0
0 5 10 15 20 25
OU SERIES ONE DATA (kips)
Figure 2.3. Comparison of NASP Bond Test results at OU in

separate test series.
can be seen in the figures. The test has received unanimous also cancel out the most compelling reasons to use high-
endorsement by the NASP as its testing standard. strength concrete in pretensioned girder applications. Russell
(1994) showed that 0.6-in. strands must be placed at a spac-
ing of about 2.0 in. c/c to enable designs to take advantage of
2.2.1.1 Effects of Strand Spacing
high-strength concrete.
Historically, AASHTO limited the strand clear spacing to a The Auburn report (Cousins et al. 1993) was one of the
minimum of three times the strand diameter (3 db). In bridge more recent works dedicated to investigating the effects of
codes prior to the AASHTO LRFD Bridge Design Specifica- strand spacing on transfer and development lengths of pre-
tions, this provision was made an explicit part of the design tensioned strands. In the Auburn study, 0.5-in. pretensioned
code. It is likely that this code provision mirrored the stan- strands were fully stressed and placed at 1.75 in. c/c in some
dard of placing 0.5-in. strands at 2.0 in. c/c. If this provision beams and 2.0 in. c/c in others. The research demonstrated
were extended to the larger diameter 0.6-in. strands, then the that there was no substantive difference in transfer lengths
0.6-in. strands would have to be placed at 2.4 in. c/c. Never- measured on beams. For beams with strands spaced at 2.0 in.
theless, using this strand spacing would cancel out the eco- c/c, the measured transfer lengths averaged 44 in. For beams
nomic value inherent in the use of 0.6-in. strand and would with strands spaced at 1.75 in. c/c, the measured transfer
25
FWC SERIES TWO DATA (kips)
20
"PERFECT" TEST
C
P
A
15
K M
B
W
10
Z r2 = 0.97
FWC TWO = (0.658) OU ONE + 3.21
J
5
0
0 5 10 15 20 25
OU SERIES ONE DATA (kips)
Figure 2.4. Comparison of NASP Bond Test results at two

different test sites.

11
lengths averaged 47 in. The researchers concluded that the cle) had transfer lengths that varied between 18 and 36 in.
strand spacing had no effect on the measured transfer lengths. Other strand manufacturers provided strand that varied
In the same study, beams were also tested for strand devel- between 18 in. and 21 in. In NASP Round II testing, nine
opment. As with the transfer length measurements, the data different strand samples were tested. The NASP Bond Test
demonstrated that beams performed similarly regardless of demonstrated significant and measurable differences
whether strands were spaced at 1.75 in. or 2.0 in. c/c. The re- between strands. In the NASP Round III testing, 10 different
searchers concluded that spacing 0.5-in. strand at 1.75 in. c/c strand samples were tested. In these tests, the differences in
did not adversely affect transfer or development length of the pull-out test results were demonstrated to correlate directly
strands. The researchers also concluded that the research results with strand transfer and development lengths.
could be extended to the use of 0.6-in. strands at 2.0 in. c/c.
From their research, Cousins et al. (1993) drew two con-
2.2.1.3 Influence of Concrete Strength
clusions. First, “decreasing the strand spacing in preten-
sioned, prestressed members from 2.0 inches to 1.75 inches Cousins et al. (1993) also tested for transfer and develop-
has no significant effect on transfer length and does not result ment lengths in two different strength classes of concrete. The
in splitting of members at transfer of prestressing force.” normal-strength concrete mixture resulted in concrete
Second, “decreasing the strand spacing in pretensioned, strengths between 6,000 and 8,000 psi. The high-strength
prestressed members from 2.0 inches to 1.75 inches has no concrete mixture resulted in concrete strengths between
significant effect on development length or nominal moment 10,000 and 12,000 psi. Transfer lengths measured in the high-
capacity.” With regard to 0.6-in. strand, Cousins et al. (1993) strength concrete were, on average, 37 in.; the transfer lengths
make the following statement, “. . . for the results reported measured in the normal-strength concrete were, on average,
herein for specimens prestressed with 0.5 inch diameter 51 in. The higher concrete strength resulted in transfer
strand, the use of 0.6 inch diameter strand at a spacing of lengths that were about 25 percent shorter. The researchers
2.0 inches does appear reasonable.” concluded that “increasing the concrete strength . . . reduces
Deatherage, Burdette, and Chew (1994) also reported on the transfer length and development length.”
research performed to determine the effect that strand spac- Two other significant research programs examined the ef-
ing had on transfer and development lengths. In their study, fects of concrete strength on transfer and development
0.5 in. diameter strand was placed in pretensioned beams length. The first, undertaken by Zia and Moustafa (1977),
with 1.75-in. and 2.0-in. spacing. Also, strands of three dif- recommended code expressions for transfer and develop-
ferent diameters (0.5 in., 0.525 in., and 9/16 in.) were placed ment length that included the concrete strength parameter.
in beams with 2.0-in. spacing. In their studies, Deatherage, Nearly 20 years later, Abrishami and Mitchell (1993) also per-
Burdette, and Chew (1994) concluded that a c/c spacing of formed transfer and development length tests. They also rec-
1.75 in. should be permitted for 0.5 in. diameter strands. Also, ommended that concrete strength be incorporated into the
the researchers stated that their data indicated that the bond code provisions. However, as noted above, the lack of data
strength of pretensioned strand was roughly proportional to that are consistent from one research program to another has
its strand diameter, indicating that strand spacing did not prevented the development of a consensus for code expres-
influence the bond characteristics of strand appreciably. sions related to transfer and development length of preten-
Accordingly, the authors recommended that the spacing sioned strands.
requirements for 0.5-in. strand be reduced from 4.0 strand
diameters to 3.5 diameters. If this principle is applied to
2.2.1.4 Tests of Strands Pretensioned
0.6-in. strands, the authors would effectively recommend a
in High-Performance Concrete
2.1-in. spacing for 0.6 in. diameter strands.
In the 1990s, several research programs were undertaken
by various states to design and build bridges using high-
2.2.1.2 Strand from Different Manufacturers
performance concrete (HPC). Most, if not all, of these proj-
Deatherage, Burdette, and Chew (1994) included 0.5 in. di- ects incorporated high-strength concrete as part of the HPC.
ameter strands from various manufacturers. The researchers In several of the projects, strand transfer length was measured,
provide strand transfer and development length test data, but and development length tests were conducted to ensure ade-
provide little comment on differences between manufactur- quate bonding properties from the pretensioned strands and
ers. The data indicate that differences in measured transfer to add to the body of knowledge regarding the transfer and de-
lengths exist among strands made by different manufactur- velopment of pretensioned strands in high-strength concrete.
ers. In the Deatherage, Burdette, and Chew study (1994), the Perhaps the first of these tests was performed in Texas by
0.5-in. strand provided by FWC (as designated in their arti- Gross and Burns (1995). In this research, two rectangular

12
beams, 42 in. deep, were fabricated. Each beam employed pre- transfer length was 23.4 in. The concrete strength at release
tensioned 0.6 in. diameter strands with 2.0 in. spacing. Trans- was 7,800 psi.
fer lengths were measured and development length tested at The box beams were also tested for development length.
each of the four ends. Concrete strengths were 7,040 psi at re- Concrete strength at the time of development length testing
lease and 13,160 psi at the time of development length testing. was 11,000 psi. For embedment lengths in excess of 60 in., the
From the four beam ends, an average transfer length of 14.3 strands demonstrated the ability to develop adequate tension
in. was measured. This value is significantly less than the cur- force to support the flexural capacity of the beams. Subse-
rent transfer length provision of 60 db found in the AASHTO quent failures were labeled as flexural failures. However,
LRFD Bridge Design Specifications. Similarly, the development when the strand embedment length was set at 60 in. and 59
length for these 0.6-in. strands was found to be less than 78 in., in., web shear cracking formed in the webs of the box beams,
which roughly corresponds to the development length given and strand anchorage failures ensued. The researchers re-
by current AASHTO provisions. The history of these beams ported that the development length for the strand was 60 in.
is also interesting. They were dubbed the “Hoblitzell-Buckner” Additionally, several research projects were undertaken in
beams. Hoblitzell was employed by FHWA and was instru- the 1990s in part to investigate the transfer and development
mental in developing the federal programs encouraging the length of 0.6-in. strands. Uniformly, these projects featured
use of HPC. Buckner authored a report for FHWA titled, pretensioned 0.6 in. diameter strands and spaced at 2 in. c/c.
An Analysis of Transfer and Development Lengths for Preten- The projects were sponsored by Texas (Barnes and Burns
sioned Concrete Structures (Buckner 1994). Buckner reviewed 2000), Virginia (Roberts-Wollmann et al. 2000; Ozyildirim
transfer length and development length data prior to 1992/ and Gomez 1999), and Georgia (Khan, Dill, and Reutlinger
1993 and developed some design recommendations based on 2002). Uniformly, these researchers concluded that 0.6 in. di-
that earlier data. In his report, Buckner recommended that the ameter strands could be deployed safely using 2-in. spacing.
design provision for transfer lengths be changed to reflect the The state of Virginia has also supported transfer length
stress in the pretensioned strand prior to release (fpi) as testing of 0.6 in. diameter strains in HPC. Results reported by
opposed to using the “effective prestress” after all losses, Ozyildirim and Gomez (1999) and Roberts-Wollmann et al.
which is still found in the 318-02 Code (ACI 2002). Effectively, (2000) indicate that transfer lengths measured in HPC were
Buckner’s recommendation would have increased the re- substantially less than the transfer length predicted by the
quirement for transfer length by about 25 percent. current code expressions.
More interesting was Buckner’s design equation for devel- Barnes and Burns (2000) reported on transfer lengths that
opment length. In reviewing the data, Buckner concluded were measured on 36 AASHTO Type I beams pretensioned
that the strain experienced by the prestressing steel at flexural with 0.6-in. strands. Strand spacing was 2 in. c/c. Concrete
strength level was an important component in the develop- compressive strengths at release ranged from 3,950 to 11,000
ment of strand. His design equation required the design en- psi. Altogether, transfer lengths from 192 independent meas-
gineer to increase development length requirements as the urements are discussed, and the report includes data on
steel strain at flexural strength level increased. The Hoblitzell- strands that are fully bonded to the ends of the member and
Buckner beams were designed, therefore, to develop ex- strands that are shielded, or debonded, at the ends of the
tremely large strains in the prestressing steel at flexural member. The results of the Barnes and Burns study (2000)
strength and test Buckner’s proposal. In the subsequent de- indicate a definite trend in which transfer lengths tend to de-
velopment length tests reported by Gross and Burns (1995), crease in inverse proportion to the square root of the concrete
the strands were able to achieve their ultimate tensile capac- strength at release. A “best fit” line reported by the authors in-
ity, undergo very large elongation strains, and adequately cludes the square root of the concrete strength at release in
develop their tension capacities within the current AASHTO the denominator. This relationship is shown in Figure 2.5.
design provision. The results of these tests suggested that However, the data demonstrate wide variation, and the sta-
strand strain did not play an important role in strand devel- tistical correlation is relatively weak. Nonetheless, it appears
opment, and therefore it would not be necessary to recom- that concrete strength is an important factor that may affect
mend that the AASHTO LRFD Bridge Design Specifications the bond of pretensioned strand.
should contain a development length provision based on Barnes and Burns (2000) also reported results on transfer
predicted strand strain at flexural strength levels. lengths of strand from various strand manufacturers. Their
The state of Colorado sponsored a research program results are illustrated in Figure 2.6. The data illustrated in Fig-
specifically designed to assess the transfer length and devel- ure 2.6 demonstrate that wide variations in measured transfer
opment length of 0.6-in. strands pretensioned in HPC box length may be the result of differences among strand manu-
beams (Cooke, Shing, and Frangopol 1998). In these beams, facturers. This finding highlights the need to establish an in-
0.6-in. strands were spaced at 2 in. c/c. The average measured dustry standard for the “bond-ability” of prestressing strand.

13
1 MPa0.5 = 0.381 ksi0.5; 1 MPa-0.5 = 2.63 ksi0.5

70
fpt
AASHTO LRFD lt = 0.22 MPa-0.5 db
60
√ f'ci
ACI/AASHTO Standard
Transfer Length (db)

50
40 fpt
lt = 0.13 MPa-0.5 db
√ f'ci
30 R = 0.37
20
10
0
0 50 100 150 200 250 300
fpt
(MPa0.5)
√ f'ci
Figure 2.5. Data comparing transfer lengths to concrete

strength at release (Barnes and Burns 2000).
1 MPa0.5 = 0.381 ksi0.5; 1 MPa-0.5 = 2.63 ksi-0.5

150
Manufacturer A fpt
Manufacturer B lt = 0.5 MPa-0.5 db
120 √ f'ci
Manufacturer C
Transfer Length (db)
Manufacturer D
90 Unknown
AASHTO LRFD
60
ACI/AASHTO Standard
30
0
0 50 100 150 200 250 300
fpt 0.5
(MPa )
√ f'ci
Figure 2.6. Data highlighting differences among strand

manufacturers (Barnes and Burns 2000).
In addition to the research projects explicitly discussed examine the effects of air entrainment on pretensioned bond.
herein, there have been other projects across the United The research reported herein incorporates the use of air en-
States that have incorporated the use of 0.6 in. diameter trainment; however, it should be noted that air entrainment
strands and spaced at 2.0 in. c/c. Many of those projects have is not usually specified in combination with high-strength
measured transfer lengths. One of the projects was performed concrete/HPC because air entrainment directly causes a
by Kahn, Dill, and Reutlinger (2002). In some cases, the decline in concrete strengths.
research reports are still in a preliminary format and use of
the data is being reserved by the authors and the research
2.2.1.6 Water Reducers and High Range
sponsors. However, it is safe to say that, uniformly, these
Water Reducers
projects are employing 0.6 in. diameter strands at 2.0-in.
spacing without adverse effects. There is no evidence of a systematic testing program ex-
amining the effects of water reducers (WRs) or high range
water reducers (HRWRs) on the transfer and development of
2.2.1.5 Effects of Air Entrainment
prestressing strands. Since WRs and HRWRs are used in
There is no evidence that a systematic testing program more than 95 percent of the pretensioned prestressing plants
examining the effects of air entrainment on the transfer and throughout North America, this is an important variable that
development of prestressing strands exists. There is a need to warrants investigation.

14
2.2.2 Development and Splice Length is present in the development length. Tests have also shown
for Mild Reinforcement that the development length of deformed welded wire rein-
forcement is not affected by epoxy coating, and thus the
To identify needed experimental research, the literature re- epoxy coating factor in the current ACI Code is 1.0 for epoxy-
view focused on the analysis of test results from bond tests on coated deformed wire fabric. In recent years, welded wire
development and splice length in tension of coated and un- fabric (WWF) has been used widely as shear reinforcement in
coated bars and development length of coated and uncoated thin-webbed girders because of the ease of construction over
bars terminated with standard hooks in tension. Based on the the use of conventional stirrups. Research conducted to date,
reported bond performance of individual and bundled bars with concrete compressive strengths up to 12 ksi, indicates
in compression, it was determined that no additional exper- that this reinforcement can be used effectively to resist shear
imental work was required in this area. Compression devel- (Mansur, Lee, and Lee 1987; Xuan, Rizkalla, and Maruyama
opment lengths are considerably shorter than tension 1988; Pincheira, Rizkalla, and Attiogbe 1989; and Zhongguo,
development lengths because there are no transverse cracks Tadros, and Baishya 2000). It was shown that two cross wires
in compression zones; the harmful effect of such cracks in ini- welded at a spacing of 2 in. at the open ends (top and bottom)
tiating splitting is absent. However, the major difference be- of WWF cages provide satisfactory anchorage. Such anchor-
tween tension and compression development and splice age was found to be more effective for deformed WWF than
lengths is the ability of the bars in compression to transfer smooth WWF. The increase in concrete compressive strength
load to the concrete directly by bearing. In tests conducted by has been shown to further improve the anchorage of this
Pfister and Mattock (1963), bearing stresses equal to five reinforcement.
times the cylinder strength of the concrete were attained at
the square-cut ends of bars in compression splices. Addi-
tional experimental work conducted at the Otto-Graf-
2.2.2.1 Databases
Institute of the University of Stuttgart by Leonhardt and
Teichen (1972) conclusively showed the following: There are two databases. One consists of 71 tension devel-
opment and splice tests of specimens with top cast uncoated
• End bearing is responsible for the majority of splice failures reinforcing bars, 493 specimens with bottom cast uncoated
in compression irrespective of the splice length tested. The reinforcing bars, 27 specimens with top cast epoxy-coated
splice lengths varied between 9 and 38 bar diameters. bars, and 48 specimens with bottom cast epoxy-coated bars,
• The bearing capacity of the concrete at the bearing ends of for a total of 639 specimens. The other database consists of 33
the bars was increased by the presence of confining rein- specimens with uncoated bars terminated with standard
forcement. Under such conditions, concrete bearing hooks and 13 specimens with epoxy-coated bars, for a total of
stresses of 17 ksi were measured (for concrete with a uni- 46 specimens.
axial compressive strength around 4 ksi). The provisions for development length of reinforcement in
• An increase in the thickness of the concrete cover over the Section 5 of the AASHTO LRFD Bridge Design Specifications
compression splice resulted only in very minor improve- are based on the provisions of ACI 318-89 (ACI 1989). The
ments in bond performance. 1989 provisions in the ACI Code were extensively modified
• Under long-term loading, the bearing pressure under the in the 1995 version of the ACI Code (ACI 1995) with a view
ends of the compression bars diminishes because of creep; to formulating a more “user-friendly” format while main-
hence, the splice performance improves. taining the same general agreement with professional judg-
ment and research results. Tests conducted by Azizinamini et
The available information on the anchorage in tension of al. (1993, 1999a) have indicated that in the case of high-
welded wire reinforcement indicated that a significant exper- strength concrete, some minimum amount of transverse re-
imental effort was not required as part of NCHRP Project inforcement is needed to ensure adequate ductility from the
12-60 (Furlong, Fenves, and Kasl 1991; Griezic, Cook, and splice at failure. A proposed modification to ACI 318-99 (ACI
Mitchell 1994; and Guimaraes, Kreger, and Jirsa 1992). In the 1999), based on these tests, called for the determination of a
case of plain wire fabric, the development in tension depends basic straight development length for bars in tension without
on the mechanical anchorage from at least two cross wires. including the presence of transverse reinforcement, together
Deformed welded wire reinforcement derives anchorage with a minimum area of transverse steel in the form of stir-
from bond stresses along the deformed wires and from rups, Asp, crossing potential splitting planes. In these studies,
mechanical anchorage from the cross wires. Current code ex- over 70 specimens with concrete compressive strengths rang-
pressions for development length in tension of deformed ing between 5 ksi and 16 ksi were tested (Azizinamini et al.
welded wire reinforcement assume that at least one cross wire 1993, 1999a).

15
Although the modification proposed in another paper by above 8 ksi. In order to assess whether the limit on f c′ can be
Azizinamini and colleagues (1999b) was not adopted in the removed by examining the existing data for development and
2002 version of the 318 Code (ACI 2002), it was deemed an splice length of uncoated and epoxy-coated bars in tension,
improvement over the current AASHTO LRFD provisions. the ratio of test to calculated bond strength is plotted versus
Therefore, the 2005 318 Code (ACI 2005) provisions were the concrete compressive strength (f c′) evaluated throughout
used in NCHRP Project 12-60 as the basis for further exten- the range of concrete cylinder strengths. The bond strength
sion of the AASHTO provisions to higher strength concrete. ratio is determined in terms of bar stresses at failure versus
The experimental work conducted in the mild steel phase of calculated bar stress, using Equations 2.2 through 2.5:
NCHRP Project 12-60 was focused on filling the gaps identi-
ls 3 fs
fied in order to extend the applicability of the present = (2.2)
db 40 ⎛ c + K tr ⎞
AASHTO LRFD Bridge Design Specifications to normal- fc′ ⎜
weight concrete with compressive strengths up to 15 ksi. ⎝ db ⎟⎠
The 639-specimen database is shown in Figures 2.7 through c = ( c min + 0.5 * db ) (2.3)
2.10 by plotting the bond strength, utest, versus the concrete
compressive strength, f c′. The bond strength is defined as Atr * f yt
K tr = (2.4)
1500 * s * n
Ab f su
utest = (2.1) To limit the probability of a pull-out failure, 318 Code (ACI
πdb ls
2005) requires that
In Equation 2.1, Ab is the area of bar being developed or
c + K tr
spliced, fsu is the stress in the bar estimated at failure using ≤ 2.5 (2.5)
db
moment-curvature type analysis and compatibility of defor-
mations, db is the diameter of the bar, and ls is anchorage/ The additional parameters in the equations are the follow-
splice length. As can be seen from Figures 2.7 and 2.8, there ing: fs is the stress in the reinforcing bar; cmin is the smaller of
is a lack of data for development and splice lengths of un- minimum cover or one-half of the clear spacing between
coated bars in tension above 16 ksi. Figures 2.9 and 2.10 show bars; Atr represents the area of each stirrup or tie crossing the
that there are limited data for epoxy-coated bars in tension potential plane of splitting adjacent to the reinforcement
1.60
1.40
1.20
1.00
Utest (ksi)
0.80
0.60
0.40
0.20
0.00
0 2000 4000 6000 8000 10000 12000 14000 16000 18000
f c' (psi)
Bottom Cast Uncoated
Figure 2.7. Bond stress at failure (utest) versus the concrete compressive strength ( fc′ ) of bottom cast uncoated
specimens.

16
1.60
1.40
1.20
1.00
utest (ksi)
0.80
0.60
0.40
0.20
0.00
0 2000 4000 6000 8000 10000 12000 14000 16000 18000
f 'c (psi)
Top Cast Uncoated
Figure 2.8. Bond stress at failure (utest) versus the concrete compressive strength ( fc′ ) of top cast uncoated
specimens.
1.00
0.90
0.80
0.70
0.60
utest (psi)
0.50
0.40
0.30
0.20
0.10
0.00
0 2000 4000 6000 8000 10000 12000
f 'c (psi)
Epoxy-Coated Bottom Bars
Figure 2.9. Bond stress at failure (utest) versus the concrete compressive strength ( fc′ ) of bottom cast epoxy-
coated specimens.

17
0.80
0.70
0.60
0.50
utest (ksi)
0.40
0.30
0.20
0.10
0.00
0 2000 4000 6000 8000 10000 12000 14000
f 'c (psi)
Epoxy-Coated Top Cast
Figure 2.10. Bond stress at failure (utest) versus the concrete compressive strength ( fc′ ) of top cast
epoxy-coated specimens.
being developed, spliced, or anchored; fyt is the yield strength Tests have shown that the bar force is transferred rapidly
of the stirrup reinforcement; s is the spacing of stirrups; and into the concrete, and the portion following a hook is gener-
n is the number of bars being developed or spliced. The re- ally ineffective and can potentially be limited by the tensile
sults of the evaluation indicated that the average of the ratio strength of the concrete. Marques and Jirsa (1975) reported
for bars not confined by stirrups is 1.23 with a standard devi- on the results of 22 tests conducted using two #7 or two #11
ation of 0.28 for all f c′ values, and 1.23 with a standard devia- uncoated bars. Standard 90- or 180-deg hooks conforming to
tion of 0.23 for concrete compressive strengths below 10 ksi. the 318 Code were used (ACI 2005). The concrete compres-
In the case of bars confined by stirrups, the average is 1.23, sive strength was around 5 ksi. The specimens simulated
and the standard deviation is 0.3 for all f c′ values. For f c′ values exterior beam column joints. Hamad, Jirsa, and D’Abreu de
below 10 ksi, the average is 1.24 and the standard deviation is Paulo (1993) reported on the results of 24 tests to evaluate the
0.30. In members with confined bars, the stirrups are as- anchorage performance of epoxy-coated hooked bars. Based
sumed to be uniformly spaced throughout the splice/devel- on these results, a 20-percent increase on the basic develop-
opment length. The value for the members in the database, ment length was recommended for epoxy-coated hooked
calculated by the ACI provisions, gives approximately the bars. It was shown that the relative anchorage strength of un-
same scatter throughout the range of concrete compressive coated and epoxy-coated hooked bars was independent of bar
strengths up to a maximum of 16 ksi for members with and size, concrete strength, side concrete cover, or hook geome-
without stirrups. This conclusion supports the extension of try. The maximum concrete strength of the specimens was 7
these provisions to higher concrete compressive strengths ksi. These tests serve as the basis of the 318 Code anchorage
with a few verification tests of uncoated bars at the upper provisions for bars anchored by means of standard hooks
limit, mainly to establish the role of the minimum amount of (ACI 2005). The specimen and the test setup used in NCHRP
transverse reinforcement on the mode of failure of splices in Project 12-60 was similar to the one used in the Marques and
tension recommended in the Azizinamini et al. studies (1993, Jirsa (1975) and Hamad, Jirsa, and D’Abreu de Paulo (1993)
1999a). On the other hand, it is recognized that there is a studies. However, only 90-deg hooks were evaluated, since
paucity of data on the performance of epoxy-coated bars in Hamad, Jirsa, and D’Abreu de Paulo found little difference in
concretes with compressive strengths above 10 ksi. Therefore, the performance of 90- and 180-deg hooks. It should be
a more intense verification testing effort was carried out in noted that sections of the AASHTO LRFD Bridge Design
this study to close this gap. Specifications dealing with the anchorage of bars in tension

18
terminated with a standard hook are similar to those in the

2005 version of the ACI Code.
2.3 Identification of Issues C1 C2

and Needs
The work described in the previous section was used to as- Cc Cs = min{C1, C2/2}
semble a comprehensive list of issues pertaining to transfer
(a) (b) (c)
length, development length, and splice length of strand/
reinforcement to normal-weight concrete with compressive Figure 2.12. Anchorage failure modes:
strengths in excess of 10 ksi and up to 15 ksi. In this section, a (a) vertical splitting, (b) splitting in the
discussion of the main issues related to bond performance of horizontal plane of the bars, and (c) pull-out
without splitting (ACI 408 2003).
reinforcement is presented, and gaps found in the existing data-
base are addressed. The experimental program described in
Chapter 3 of this report was directed at addressing the identi- perimeter of the bar, resulting in a pull-out type failure. Tests
fied needs in order to extend the AASHTO LFRD Bridge Design have shown that these two types of failures can take place at
Specifications to allow greater use of high-strength concrete. stresses close to the tensile strength of the reinforcement. Pull-
In reinforced and prestressed concrete structures, suffi- out failures occur in cases of high confinement and low
cient transfer of forces between concrete and reinforcement bonded lengths. However, splitting failures are more common
is required for a satisfactory design. The transfer of forces in structural applications. For this reason, it is recommended
occurs through a combination of chemical surface adhesion, that experimental data considered for development of design
friction, and bearing of bar deformations against the sur- equations should have a minimum embedment length.
rounding concrete. Initially, the transfer of forces occurs Another important observation is that transverse reinforce-
mainly by chemical adhesion; after initial slip, most of the ment has been observed to rarely yield during splitting failures
force is transferred by bearing and friction. In the case of plain (Maeda, Otani, and Aoyama 1991; Sakaruda, Morohashi, and
bars or wires, slip-induced friction—resulting from shape Tanaka 1993; and Azizinamini et al. 1999a). Therefore, it is
and surface roughness—plays an important role in the force important to limit in design provisions the level of confine-
transfer. In the case of deformed reinforcement, as slip in- ment provided by transverse reinforcement. The many factors
creases, bearing of the ribs against the surrounding concrete affecting bond performance are presented in two main cate-
becomes the principal mechanism of force transfer between gories: member properties and material properties. Some of
concrete and steel. The forces on the bar surface are balanced the factors are common to both strand and mild reinforce-
by compressive and shearing stresses in the concrete (see Fig- ment while others are unique to one or the other.
ure 2.11). The concrete stresses result in tensile stresses that, Initially, in the testing of prestressing strand for bond to con-
if high enough, can lead to cracking in planes both parallel crete, the simple pull-out tests were criticized because they did
and perpendicular to the reinforcement, as shown in Figure not include the wedging action, or Hoyer’s effect, associated
2.12. These transverse cracks can lead to splitting failure. with pretensioned strands in real beams. However, subsequent
If the concrete cover, bar spacing, or amount of transverse testing with both the Moustafa Test and the NASP Bond Test
reinforcement is sufficient to prevent or delay the splitting fail- have demonstrated that a direct correlation exists between re-
ure, then failure can occur along a surface surrounding the sults from these simple pull-out tests and strand performance
in pretensioned beams. Therefore, in this testing program the
NASP Bond Test was employed as an assessment tool to quan-
bearing and friction
forces on bar
tify the “bond-ability” of prestressing strands that will be
employed. Testing sponsored by the NASP has demonstrated
that the NASP Bond Test has superior repeatability and repro-
ducibility when compared with the Moustafa Test.
2.3.1 Member Properties

adhesion and friction forces
along the surface of the bar
2.3.1.1 Transfer Length of Prestressing Strand
Figure 2.11. Mechanisms of force transfer In the specific case where prestressing strands are bonded
between concrete and reinforcement- to concrete, bond stresses are derived through a combination
deformed bars. of adhesion, friction, and mechanical interlocking (Hanson

19
and Kaar 1959). It has been widely believed that a wedging length. In the case of anchored reinforcement by means of
effect, called Hoyer’s Effect, unique to pretensioned strands, straight embedment, longitudinal splitting will initiate at either
creates significant bond stresses in the transfer zone where the a free surface or at a flexural crack location. In the case of spliced
effective prestressing force is transferred from the preten- bars, splitting will start at the ends of the splice and move toward
sioned strand to the concrete. In those same regions, slip the center. The mode of failure explains the fact that the non-
occurs between strand and concrete due to the difference in loaded end of a developed/spliced bar is less effective than the
strain condition. Research has indicated that the strand end loaded end in transferring forces between concrete and rein-
slip can be used as a quality control measure for the bond of forcement. It can be concluded that there is a non-proportional
prestressing strands (Rose and Russell 1997). Furthermore, relationship between development/splice length and bond
the relative slip between strand and concrete virtually ensures strength. Thus, even though bond strengths have been meas-
that adhesion plays little or no role in the transfer of pre- ured for very short embedment lengths, it is not appropriate to
stressing forces to concrete (Russell and Burns 1996). linearly extrapolate such findings to code development lengths.
This observation suggests the need for testing at appropriate
scale for development of design provisions.
2.3.1.2 Development and Splice Length In beams tested for strand development, it is equally appar-
Bond forces are not uniformly distributed over the length of ent that cracking causes the mobilization of the strand relative
anchorage (see Figure 2.13). Thus, bond failures are incremen- to concrete. Commonly, bond stresses that develop strand
tal, initiating in the region of highest bond force per unit of tension from the transfer zone to the point where flexural
capacity is required are called “flexural bond stresses.” Flex-
ural bond results primarily from a combination of mechani-
cal interlocking and friction. The mechanical interlocking
bond stresses are derived by the helical windings of the 7-wire
prestressing strand, which act similarly to the mechanical de-
formations found on rolled, mild reinforcement.
Development length testing of pretensioned beams indi-
cates that splitting occurs less frequently than in convention-
ally reinforced beams (although splitting cracks have been
observed in pretensioned bond failures). Issues for strand
development are more related to the cracking patterns that
occur as the pretensioned beams approach their ultimate
strength. In testing on beams with debonded strands, it is
clear that cracks that propagate through or near the transfer
zone of pretensioned strands cause strands to slip. In many of
those tests, cracking in the transfer zones of pretensioned
strands caused bond failure of pretensioned strands (Russell,
Burns, and ZumBrunnen 1994; Russell and Burns 1994).
Additionally, in pretensioned strands with fully bonded
beams, it is important to note that sections with narrow webs,
specifically I-shaped beams, have failed in bond in concert
with web shear cracking that occurs near or through the
transfer zones of pretensioned strands (Jacob 1998; Kaufman
and Ramirez 1988; Russell and Burns 1993). In contrast, tests
on rectangular prestressed beams will not produce web shear
cracks, so the behavior of rectangular cross sections can be
Figure 2.13. Variation of steel significantly different than cross sections with narrow webs.
and bond forces in a reinforced
For that reason, the testing program includes testing of both
concrete member subjected to
pure bending: (a) cracked con-
rectangular and I-shaped sections.
crete region, (b) bond stresses
acting on a reinforcing bar, (c) 2.3.1.3 Transverse Reinforcement
variation of tensile force in steel,
and (d) variation of bond force Orangun, Jirsa, and Breen (1977) indicated that transverse
along the bar (Nilson 1997). reinforcement confines the concrete around anchored bars

20
and limits the progression of splitting cracks. An additional strength of vertical bars seems to be reduced only by 25 per-
beneficial effect of transverse reinforcement is that increases cent with respect to the bond strength of horizontal bars. A
in transverse reinforcement lead eventually to pull-out failures single factor of 1.3 is recommended for all vertical bars where
rather than splitting-type failures. However, the Orangun, the center of the splice or the development length has more
Jirsa, and Breen study also noted that transverse reinforce- than 24 in. of concrete cast below.
ment in excess of the amount required to cause the change in
mode of failure is not as effective and eventually leads to no
further increase in bond strength. These observations and the 2.3.1.5 Concrete Cover and Spacing
observations by Maeda, Otani, and Aoyama 1991; Sakaruda, of Reinforcement
Morohashi, and Tanaka 1993; and Azizinamini et al. 1995 that As shown in Figure 2.12, splitting failure is expected to con-
the transverse reinforcement confining the anchored bar sel- trol in the majority of structural applications. In this type of fail-
dom yields in splitting failure indicates the need for an upper ure, the actual location of the splitting cracks in the case of bot-
limit on the improvement in bond strength provided by the tom cast reinforcement depends on the relative values of the
presence of transverse reinforcement. concrete bottom cover, concrete side cover, and one-half of the
clear spacing between bars. If the bottom cover is less than the
2.3.1.4 Casting Position side cover and one-half the spacing between bars, splitting
occurs through the cover to the bottom free surface. If either the
It has been observed by various researchers that top cast side cover or one-half the bar spacing is smaller than the bot-
bars have lower bond strengths than bottom cast bars. Clark tom cover, then splitting of the concrete occurs either through
(1946), using pull-out type specimens cast in a horizontal the side cover or between the reinforcement. This observation
position, noted that in the top position, bars were two-thirds supports the need to modify the current AASHTO LRFD Bridge
as effective in bond as in the bottom position. The depth of Design Specifications for bond and development length of mild
the concrete under the bar in the top position was 15 in., and reinforcement to incorporate the effects of cover, bar spacing,
the depth of the concrete under the bar in the bottom posi- and transverse reinforcement.
tion was 2 in. The concrete slump was 4.25 in., and the com-
pressive strength averaged 5.6 ksi. Ferguson and Thompson
(1965) noted that with 12 in. of concrete below the bar, the 2.3.2 Material Properties
strength dropped from 3 to 13 percent as the slump was in- 2.3.2.1 Reinforcement Properties
creased. They noted that for the beam depths tested, from 13
to 22 in., the 1.4 factor used in the specifications was conser- For a given bonded length required to achieve a given steel
vative. This observation is currently recognized in the 318 stress level, reinforcement of different areas will achieve dif-
Code where a 1.3 factor is used to increase the development ferent levels of force at the onset of splitting failure, with the
length or splice of bars cast horizontally with more than 12 in. larger area reinforcement achieving higher forces. Therefore
of fresh concrete cast in the member below the bar (ACI larger area reinforcement will require longer development/
2005). A 1.4 factor is currently prescribed in the AASHTO splice length than smaller area reinforcement for the same
LRFD Bridge Design Specifications to cover this case, and it is degree of confinement. The size of the reinforcement being
thus conservative if the effects of cover and transverse rein- developed also plays a role in the contribution of the confin-
forcement are included in the specifications. ing reinforcement for the case of deformed bars. As large bars
Additional research (Jirsa and Breen 1981) indicates that slip, higher strains are mobilized in the transverse reinforce-
the concrete slump plays an important role in determining ment, thus the beneficial effect of transverse reinforcement
the effects of casting position, and this is most significant on the bond strength of deformed bars increases as the area
when very large depths of concrete are cast below the bars or of the bar increases.
splices. The 1981 study by Jirsa and Breen further indicated It is now customary to relate bond performance to bar
that the so-called top bar factor should vary with the depth of geometry by means of the relative rib area factor, Rr, defined as:
concrete cast below the reinforcement and recommended a projected rib area normal to bar axis
maximum factor of 1.3 for slumps of less than 4 in. For Rr =
(nom. bar peerimeter) × (center − to − center rib spacing)
slumps between 4 and 6 in., a maximum factor of 1.35 is rec-
(2.6)
ommended for depths below 24 in., and a maximum factor
of up to 1.6 is recommended for depths greater than 48 in. Typical bars currently used in the United States have relative
For slumps greater than 6 in., a factor of 1.8 for depths below rib area factors ranging between 0.057 and 0.087 (Choi et al.
24 in. is recommended, and a factor of 2.2 for depths below 1990). Darwin and Graham (1993a, 1993b) concluded that the
48 in. is recommended. It is further stated that the basic bond bond strength is independent of deformation pattern if the bar

21
is under small cover conditions and there is no transverse rein- In unconfined bar bond tests, the use of basalt aggregates has
forcement. Darwin and Graham observed that under large been shown to increase bond strength by almost 13 percent
cover or with transverse reinforcement, bond strength in- over the bond strength of concretes with weaker aggregate such
creased with an increase in relative rib area. They also found that as limestone (Zuo and Darwin 1998, 2000). Tests on bars con-
deformations parallel to the splitting cracks were more effective. fined by transverse reinforcement (Darwin et al. 1996; Zuo and
The bond strength of epoxy-coated bars has been found to Darwin 1998) also indicate an increase in bond strength in the
increase under all conditions of confinement as the relative presence of stronger aggregates showing a significant effect on
rib area is increased. Zuo and Darwin (1998) recommended the contribution from the transverse reinforcement.
that for epoxy-coated bars with relative rib areas greater than Lower strength aggregates, on the other hand, have a
or equal to 0.1 and concrete with compressive strength below detrimental effect on the bond strength. Reports by ACI
10 ksi, development and splice length should be increased by Committee 408 (1966, 1970) have emphasized the paucity of
20 percent instead of the 50-percent increase for cover less experimental data on the bond strength of reinforced con-
than 3db or clear spacing less than 6db. For concrete strengths crete elements made with lightweight aggregate concrete. The
greater than 10 ksi, a 50-percent increase appeared warranted AASHTO LRFD Bridge Design Specifications includes a fac-
regardless of the value of Rr. tor of 1.3 for development length to reflect the lower ten-
The surface condition is important from the standpoint of sile strength of lightweight aggregate concrete and allows that
bond strength because it affects adhesion, friction, and bear- factor to be taken as 0.22 fc′ / fct ≥ 1.0 if the average splitting
ing in the transfer of forces between steel and surrounding strength, fct, of the lightweight aggregate concrete is specified.
concrete. Items such as cleanliness, rust, and coatings affect For lightweight sand, where fct is not specified, a factor of 1.2
the surface condition of the reinforcement. Specifications is specified. Although design provisions, in general, require
require that the reinforcement be free of mud and other longer development lengths for lightweight aggregate con-
substances capable of reducing bond strength. It is well crete, test results from previous research are contradictory, in
established that the presence of epoxy coatings reduces the part, because of the different characteristics associated with
bond strength of reinforcement (Mathey and Clifton 1976; the particular type of aggregate and mix design. The use of
Johnston and Zia 1982; Treece and Jirsa 1989; Choi et al. lightweight aggregate concrete is outside the scope of NCHRP
1990, 1991; Cleary and Ramirez 1993). Project 12-60.
It has been widely observed that as the concrete compres-
sive strength increases, the bond strength of the same con-
2.3.2.2 Concrete Properties
crete also increases—albeit at a slower rate—leading to po-
Compressive strength and lightweight aggregate are ac- tentially more brittle failures (Azizinamini et al. 1993, 1999a).
knowledged in codes and specifications as influencing bond On the other hand, the tensile strength of the concrete is not
strength. In addition, tensile strength and fracture energy, the only factor controlling bond strength, as it has been noted
mineral admixtures, and consolidation and vibration are also by Zuo and Darwin (1998, 2000) for deformed bars. The Zuo
factors affecting bond strength of reinforcement. and Darwin studies recommended the use of f c′1/4 instead of
Azizinamini et al. (1993, 1999a) noted that for higher the traditional f c′1/2 to represent the effect of concrete com-
strength concretes, the higher bearing capacity prevents pressive strength on bond strength for unconfined bars. They
crushing of the concrete in front of the ribs, thus reducing the also noted that the presence of confinement influenced the
local slip. These researchers further noted that the reduced power of the compressive strength and recommended the use
slip also limited the number of ribs participating in the load of f c′3/4 as a good representation of the influence of compres-
transfer between concrete and reinforcement. The reduced sive strength on bond strength.
participation of the ribs increases the local tension stresses Most of the work related to bond has focused on the effect
and further leads to a non-uniform distribution of bond of silica fume. The studies have shown increases of less than
force. Although traditionally fc′ has been used to reflect the 10 percent on bond strength in the presence of the mineral
concrete compressive strength in bond calculations, Zuo and admixture (DeVries, Moehle, and Hester 1991; Hamad and
Darwin (1998, 2000) have postulated that f c′1/4 for members Itani 1998).
without stirrups and f c′3/4 for members with stirrups better re-
flect the effect of concrete strength on bond. These findings
2.4 Issues Related to Testing
indicate that if bond strengths are normalized with respect to
Protocols
f c′1/2, the effect of concrete strength on the bond strength is se-
verely overestimated. High-strength concrete has been shown A review of testing protocols for determining bond charac-
to improve anchorage of prestressing strand, thus reducing teristics was presented. From our review of available research,
the required transfer length and development length. we recommended that the NASP Bond Test be employed

22
throughout the experimental program to quantify the bond transfer and development of prestressing strands. Through
characteristics of individual strand samples. The testing the years, several researchers have included concrete strength
program includes “round robin” testing at both Purdue Uni- as a variable in transfer and development length equations.
versity and Oklahoma State University (OSU) to validate the However, the lack of consistency in the strand products
repeatability of the test procedure. The NASP Bond Test themselves has worked against developing a consensus re-
procedure has been refined through this research and is now garding the effect of concrete strength. By quantifying strand
recommended for adoption into the AASHTO LRFD Bridge bond characteristics through the NASP Bond Test, this re-
Code as the Standard Test Method for the Bond of Prestress- search has been able to assess the effects of concrete strength
ing Strands. The testing protocols for splice/development on transfer and development of pretensioned strands.
lengths and bars terminated with standard hooks are also pre- Recommendations to include concrete strength in the design
sented. The beam-splice test for splice/development length equations have been made.
and the Marques and Jirsa (1975) and Hamad, Jirsa, and
D’Abreu de Paulo (1993) exterior beam column joint setup
2.4.1.3 Influence of Water-Reducing Admixtures
for bars in tension anchored by means of standard hooks are
recommended for use in this study. There is a lack of data available to assess what effect, if any,
water-reducing admixtures have on the bond of pretensioned
strands. This contrasts directly with the fact that more than
2.4.1 Testing Protocols for
99 percent of the prestressing plants in North America use
Prestressing Strand
HRWRs (the source for this information is an informal,
Since 1994, three new test procedures or protocols have unpublished committee report on a survey of precast/
been developed for assessing the bonding characteristics of prestressing plants done by the Prestressing Steel Committee
prestressing strand: the Moustafa Bond Test, the Post- of PCI circa 1998). For historical perspective, it is noted that
Tensioning Institute (PTI) Bond Test, and the NASP Bond the bulk of development regarding the Moustafa Test em-
Test. Testing has demonstrated that the NASP Bond Test deployed concrete that did not contain HRWRs. A majority of
livers the greatest degree of repeatability and reproducibility the Moustafa testing has been performed at Stresscon
of the three tests. Therefore, the testing program for NCHRP Corporation in Colorado Springs, where HRWRs are not
Project 12-60 employed the NASP Bond Test as the standard commonly employed. Yet, others that have participated in
test to assess the relative “bond-ability” of prestressing Moustafa Testing have employed HRWRs as part of the stan-
strands. Previous experience with research on strand bond dard casting procedures used in the local prestressing plants.
demonstrates the importance of quantifying the strand bond- Variations that result from the use of HRWR have not been
ing properties prior to or concurrent with testing programs measured or quantified. These data were compiled informally
for transfer and development length of strands. through the work of the Prestressing Steel Committee of the
Precast/Prestressed Concrete Institute and are not available
for publication.
2.4.1.1 Prestressing Strand up to 0.6 in. in Diameter
Engineers and contractors concerned with the bond of pre-
2.4.1.4 Influence of Air Entrainment
stressing strand used for rock anchors developed the PTI Bond
Test. In conformance with standard practice for rock anchors, There is a lack of data available to assess what effect, if any,
the test protocol indicates explicitly that testing should be con- air entrainment has on the bond of pretensioned strands.
ducted on 0.6-in. strand. The test protocol has been modified The experiences of the states are mixed with regard to
to accept 0.5-in. strand, but the acceptance value has not been whether air entrainment is required in pretensioned beams.
adjusted or evaluated using 0.5-in. strand. Both the Moustafa The NASP Bond Test was employed to examine what effects,
Test and the NASP Bond Test were developed using 0.5-in. if any, air entrainment has on bond. Results indicate that
strand. In the experimental program, the NASP Bond Test was concrete strength is more important to bond strength than
performed using 0.6-in. strand. The testing demonstrated that air entrainment.
the Standard Test Method for the Bond of Prestressing
Strands is suitable for 0.6-in. strands as well as 0.5-in. strands.
2.4.1.5 Top Bar Effects
There is a small database in existence available to examine
2.4.1.2 Influence of Concrete Strength
the “top bar effect” on transfer and development of pre-
As noted in the literature review, concrete strength has stressing strands. This information is primarily available from
long been described as an important variable affecting the testing programs on prestressed concrete piling. The top bar

23
effect is expected to be assessed during casting of the scale situations encountered in structures. Transverse compression
model and full-sized specimens by including pretensioned has a beneficial effect on bond strength and yields an overly
strands in the top half of the cross section. optimistic assessment of the actual performance of struc-
tures. For this reason, various testing schemes have been pro-
posed to eliminate transverse compression (see Figure 2.14[f]
2.4.2 Testing Protocols for Mild
and [g]). In the case of semi-beam specimens, such as those
Reinforcement
shown in Figure 2.14(f), it is critical to properly account for
A review of testing protocols was conducted to determine the increase in the length over which splitting resistance tends
the appropriate testing protocol(s) for addressing gaps in the to be mobilized due to the confining pressure at the end of the
experimental data. It is well established that testing protocols bar (if the bar end is not shielded). ACI Committee 408
to evaluate development and splice length requirements for (1964) prepared a detailed guide for the determination of
deformed bars and wire in tension must be of an appropriate bond strength in beam specimens. The more popular varia-
scale, containing more than one bar or wire; testing protocols tion, the so-called beam splice test with the splice located in
should also show due regard for a realistic transfer of force the constant moment region (the most critical condition is
between concrete and steel reinforcement, as well as cover one where both bars in the splice are subjected to high
and bar spacing effects. The more commonly used testing stresses), can be seen in Figure 2.14(i).
configurations are shown in Figure 2.14. Although they Splice tests have been realistic simulations of real conditions
are economically appealing, pull-out tests used by earlier in structures, but development length tests have been con-
researchers to evaluate bond performance of various ducted largely using pull-out tests in which splitting failures
reinforcing bars embedded in concrete of different strengths are purposely avoided. As a result, the bond stresses developed
(Figures 2.14[a] through [e]) present the problem of intro- along splices are low compared with the bond along a bar in a
ducing transverse compression, a compression not typical of pull-out test. This difference in test methods is responsible for
(h)
(i)
(a) (b)
(j)
(k)
(l)
(g)
(e)
(d)
(c)
(f)
(m)
P = applied load.
Figure 2.14. Testing methods to evaluate bond strength.

24
large differences in code-required anchorage lengths for

splices and development of single bars. Pull-out failures occur
in cases of high confinement and short bonded lengths. In
most structural applications, however, splitting failures tend
to control. On this basis, the data developed for extending the
current AASHTO LRFD specification for splice/development
length had a minimum bonded length to bar diameter ratio of
15 in order to avoid unrealistically high values of bond
strength. A similar concept of minimum embedment length
should be included in any proposed specifications.
Splice specimens such as those shown in Figure 2.14(i) are
deemed to represent larger-scale specimens designed to directly
measure development and splice strength in full-scale mem-
bers. Because of the relative ease of fabrication and the realistic
state of stress achieved during testing, splice specimens were
used in the development of experimental data on development/
splice length of mild reinforcement in this research.
Review of experimental data on anchorage of bars termi-
nated using standard hooks indicates the need for additional
testing to extend the current AASHTO LRFD specifications to
concrete strengths up to 15 ksi (see Section 2.2.2). Tests have Figure 2.15. Exterior beam-to-column joint setup to
shown that the bar force is transferred rapidly into the con- evaluate bond performance of bars developed
crete, and the portion following a hook is generally ineffective using standard hooks.
and can potentially be limited by the tensile strength of the
concrete. Further study of failures of hooked bars indicates that – mineral admixtures;
splitting of the concrete cover is the primary cause of failure – chemical admixtures, including specific gravity and
and that splitting originates at the inside of the hook, where the percent solids; and
local stress concentrations are higher. Thus, it has been deter- – fine and coarse aggregates and their properties (e.g.,
mined that hook development is a direct function of bar di- specific gravity [SSD] and absorption).
ameter, db, which governs the magnitude of compressive • the concrete compressive strength, as obtained from a
stresses on the inside of the hook. The experimental work sup- standard concrete cylinder (which should be cured side-
porting the current requirements for development of standard by-side with, and in the same manner as, the bond/splice
hooks in tension was conducted using the test setup shown in specimens), and including:
Figure 2.15. In NCHRP Project 12-60, a similar specimen and – size of the compressive strength specimens,
test setup was used in the evaluation of uncoated and epoxy- – type and thickness of the cylinder caps used on the spec-
coated bars terminated with standard hooks in tension to imens, and
normal-weight concrete with compressive strength up to 15 ksi. – age of the specimen at testing.
A useful test protocol to help understanding the bond • the concrete flexural strength, including:
strength of mild reinforcement in concrete members must – size of the flexural strength specimens,
define a minimum level of information to be provided. The – age of specimen at testing, and
recommended level of information is described in the – flexural test method used.
following subsections.
2.4.2.2 Reinforcement Properties
2.4.2.1 Concrete Properties
The properties of the reinforcing steel are required for
The following information on concrete properties should basic identification and, in most cases, are needed to fully
be provided: characterize the steel used in the tests. The following infor-
mation should be provided for each heat or production run
• the source of the concrete. of reinforcing steel: the standard (ASTM) under which the
• the mix proportions, including identification of the bars were manufactured, the nominal diameter, bar designa-
components: tion, yield strength, tensile strength, proof strength (if appli-
– cement type; cable), elongation at failure, weight (mass) per unit length,

25
rib spacing and rib height (according to the standard under strain gages on the developed/spliced reinforcing bars; loca-
which the bar was manufactured as well as the average value), tion, type, and number of displacement transducers; the full
relative rib area, rib angle (the included angle between the rib load-deflection curves for the tests; and the number and
and the bar axis), rib-face angle, and type of coating and coat- location of load measurement devices. More information on
ing thickness (if applicable). the local behavior of the splice or development length can be
obtained by placing strain gages on the bar itself. The strain
2.4.2.3 Specimen Characteristics gage instrumentation provides information on the changes in
bar force along its length. In order to avoid excessive distur-
The following specimen characteristics should be recorded: bances due to the presence of the gages, they are installed by
exterior dimensions, the location of structural/tensile rein- splitting the bar in half, forming a channel along the center-
forcement (including the effective depth), bottom/top cover, line. Strain gages and wires are then placed in the channel,
side cover, the clear spacing between bars, and the length of the and the bar is welded together. An acceptable alternative is to
specimen. These additional specimen characteristics should be place the wires and gages in grooves cut along the longitudi-
recorded: the length of the developed/spliced bars, the number
nal ribs of the bar. Transverse reinforcement should be
of developed/spliced bars, the number and average spacing of
instrumented as well. Surface concrete strains should be
stirrups/ties used as transverse reinforcement in the region of
monitored in the test region using strain gages and/or Zurich-
the developed/spliced bars, the tensile strength of this rein-
type gages by means of a reference grid attached to the surface
forcement, and the load (tensile/bending) on the specimen at
the time of failure (including the specimen self-weight and the of the specimen.
weight of the test system). Specimen dimensions should be
measured after casting. It is also recommended that the cover 2.5 Summary
be measured after casting and/or testing.
Based on the results of the literature review conducted in
this chapter, the experimental plan for NCHRP Project 12-60
2.4.2.4 Test Information
was refined and carried out with the approval from the proj-
The following basic information should be recorded for ect panel. The results of the entire experimental program for
each test: a description of the test system; the weight of the both strand and mild steel are presented and discussed in
loading system; the rate of loading; the presence or absence of Chapter 3 of this report.

26
CHAPTER 3
Experimental Program and Results
3.1 Introduction to the Experimental to support the recommendation for minimum threshold
Program values from the Standard Test Method for the Bond of
Prestressing Strands. Based on the results, a design
Each section of this chapter focuses on a phase of the re- expression for development length is recommended that
search program. Each section begins with a discussion of test- includes a factor for concrete strengths up to 15 ksi.
ing procedures for the phase of the research program covered • The object of the mild steel phase of the experimental pro-
and then discusses the test results. Prestressed sections are gram was to evaluate the bond strength under monotonic
discussed in Sections 3.2 through 3.7. The mild steel phase is loading of lap-spliced and hooked uncoated and coated
discussed in Section 3.8. The experimental thrust areas con- bars in tension embedded in normal weight higher
sist of the following: strength concrete. The 318 Code (ACI 2005) places an
upper limit on the fc′ of 100 psi in the calculation of
• Refinement of the NASP Bond Test, culminating in the re- splice length/development length of bars as well as on the
sults from round robin testing by OSU and Purdue. Based development length of standard hooks in tension in higher
on its repeatability and the reproducibility, the NASP Bond strength concretes. This limitation was first introduced
Test is presented as the Standard Test Method for the Bond in the 1989 edition of 318 Code (ACI 1989). Section 5.4.2.1
of Prestressing Strands (Standard Test for Strand Bond). of the 2004 edition of the AASHTO LRFD Bridge Design
The Standard Test Method for the Bond of Prestressing Specifications states that design concrete strengths above 10
Strands is recommended for adoption by AASHTO. The ksi shall be used only when allowed by specific articles or
Standard Test Method for the Bond of Prestressing Strands when physical tests are made to establish the relationships
was modified by testing strand in concrete of varying between the concrete strength and other properties. The
strengths. The tests demonstrate that the bond strength be- object of this experimental program was to provide the
tween strand and concrete is improved by increases in con- information necessary to determine whether these limita-
crete strength. The results indicate that bond performance tions can be removed for concrete compressive strengths
improves in proportion to the square root of the concrete up to 15 ksi.
strength. This relationship is subsequently used in the rec-
ommended code expressions for transfer length and devel-
opment length.
• Transfer length measurements made on pretensioned con-
3.2 The Standard Test Method
for the Bond of Prestressing
crete beams, both rectangular beams and I-shaped beams.
Strands
The results show a direct correlation between decreasing
transfer length and increasing concrete strength. Based on In the past, testing programs intended to measure transfer
the results, a design expression for transfer length is rec- and/or development length have instead highlighted the vari-
ommended that includes a factor for concrete strength. ation in bond-ability that resulted from the varying bonding
• Development length tests on the pretensioned concrete properties of prestressing strands. So, rather than addressing
beams. The results show that strand development length the primary research focus, which was often to develop code
requirements shorten with increased concrete strength. equations for strand transfer and development lengths, the
Additionally, the development length tests provide the data results of these testing programs were muddled and confus-

27
ing to transportation agencies and others. Therefore, as a first represents a different stage in the development and refine-
step in this research program, a Standard Test Method for the ment of the NASP Bond Test.
Bond of Prestressing Strands was refined from prior testing.
The research reported herein continues and expands research
3.2.1 Refinement of the NASP Bond Test
begun by NASP. The focus of NASP’s research has been to de-
velop a standardized test for bond that would be repeatable at The NASP Bond Test was originally developed in Round II
a testing site, reproducible among sites, and provide a reliable and Round III research sponsored by NASP. The NASP
prediction of the performance of a pretensioned concrete research investigated the repeatability and reproducibility of
product. With the development of a repeatable, reproducible the test method together and also compared the NASP Bond
standard test, design expressions for transfer and develop- Test with other test methods. In Rounds II and III, the
ment length can be developed. research showed that the NASP Bond Test was a better pre-
Figure 3.1 shows a NASP specimen mounted in the load- dictor for bond than the Moustafa Test or the PTI Bond Test.
ing frame at OSU. Each test specimen is prepared by casting The NASP Bond Test also showed convincing results when
a single prestressing strand in a sand-cement mortar within compared with transfer lengths measured on prestressed con-
a cylindrical steel casing. The sand-cement mortar is pro- crete beams. Additionally, Round III testing showed evidence
portioned to produce a strength of 4750 ± 250 psi at 24 hr, that the NASP Bond Test could be used to ensure adequate
after standard curing. Additionally, the sand-cement mortar strand development. The early versions of the NASP Bond
is required to produce a flow in the range of 100 to 125 as Test protocols are included in Appendix I. Appendix I con-
measured by ASTM C 1437. The strand is pulled from the tains two versions of the NASP Bond Test, the first dated
concrete mortar at a displacement rate of 0.10 in./min, 24 hr August 2001 and the second dated May 2004. The earliest ver-
after casting. Pull-out force is measured in relation to the sion of the NASP Bond Test was employed for Rounds II and
movement of the free end of the strand to the hardened III of the NASP-sponsored research. The May 2004 protocols
mortar. The NASP Bond Test records the pull-out force that were used for NCHRP Project 12-60 for the purpose of
corresponds to 0.10 in. of free strand end slip. One single further refining the NASP Bond Test. Some refinements in
NASP Bond Test consists of six or more individual test spec- protocol were made to develop the final version found in
imens; the average value from the set of six becomes the Appendix H and titled, “Standard Test Method for the Bond
“NASP Bond Test Value.” The appendices to this report of Prestressing Strands.”
contain three separate bond test protocols; each protocol For this research, minor changes were made to the NASP
Test procedures that were used in NASP Round III research.
Although the underlying methodology in the procedure was
not changed significantly, changes in the sample preparation
were made and test procedures were refined. The NASP pro-
tocols in 2001 specified a sample preparation in which the
cement mortar had a sand-cement-water ratio of 2:1:0.45 and
a target 1-day mortar cube strength of 3,500 to 5,000 psi. The
wide range in the mortar cube strength proved to adversely
affect the NASP Bond Test values. Weaker mortar produced
lower pull-out strengths, whereas stronger mortar produced
higher pull-out strengths. The May 2004 protocols used in
the NCHRP research targeted a smaller range (4,750 ± 250
psi) for mortar cube strength. Later, through refinement, the
mortar proportions were not specified so that consistent
mortar strengths could be produced despite possible varia-
tions in the constituent materials from site to site. Therefore,
the August 2006 protocol for the Standard Test Method for
the Bond of Prestressing Strands required mortar strength in
the range of 4,500 and 5,000 psi, but did not specify the mix-
ture proportions.
Additionally, the test methodology adopted a mortar flow
requirement in the range of 100 to 125, whereas flow mea-
Figure 3.1. NASP specimen on the surements were not made during the NASP Round III tests of
loading frame at OSU. the August 2001 protocols. The standardized flow rates help

28
ensure workability of the mortar and consistent consolida- Strands requires a loading rate of 0.1 in./min, as before, and
tion of the mortar. The strand is centered in a steel casing the NASP value is reported as the load at which the free strand
with an outer diameter of 5 in. and a bond length of 16 in. The end slip is 0.1 in. The average of six or more NASP specimens
cement mortar is cast and consolidated in the steel casing. is reported as the NASP value for the strand. Studies con-
The NASP Bond Test protocols in 2001 did not specify the ducted earlier in the NASP Round II concluded that the least
frame used for loading the NASP specimen. The loading variation in the NASP values is exhibited for the 0.1 in. of
frames used in the Round III trials were more “flexible” when strand end slip. The largest variation in the NASP values was
compared with the frame used in the current NCHRP re- reported in the 0.01 in. of free strand end slip.
search, which is more “rigid.” Because the NASP Bond Test The Moustafa Test and the PTI Bond Test, which are used
protocols require a displacement rate, the rigidity of the test by some to identify the bonding properties of prestressing
apparatus affects the loading rate. Therefore, the Standard strands with concrete, were neither repeatable nor repro-
Test Method for the Bond of Prestressing Strands limits the ducible. The NASP Bond Test was convincingly superior to
loading rate to 8,000 lb/min for 0.5 in. diameter strands and the others in its ability to reproduce results among sites.
9,600 lb/min for 0.6 in. diameter strands. In its recommended Table 3.1 provides the results of NASP Bond Tests that
form, the Standard Test Method for the Bond of Prestressing were performed at OSU. Ten different 0.5 in. diameter
Table 3.1. Results of NASP Bond Tests at OSU.
Diameter (in.)
Mortar NASP Test Results
STRAND ID
NCHRP ID
Water-
NASP IV
Strength
Batch #
Strand
to- Pull-Out
f ci Force at S LC/DC
Cement N
Ratio 0.1" slip (lbs.)
(psi) (lbs.)
8N 0.45 4765 C D 0.5 6,870 12 861 DC
11N 0.45 4730 G A 0.5 20,710 11 1604 DC
14N 0.45 4953 G A 0.5 20,010 12 3088 LC
15N 0.45 4815 G A 0.5 21,930 6 1106 LC
15N 0.45 4815 G A 0.5 21,190 6 1333 DC
17N 0.45 4484 C D 0.5 8,710 5 432 LC
17N 0.45 4484 C D 0.5 6,910 5 338 DC
21N 0.5 4043 G A 0.5 20,060 12 1129 LC
22N 0.5 4117 C D 0.5 6,110 12 421 DC
23N 0.5 3981 G A 0.5 16,360 12 1629 DC
24N 0.4 5763 C D 0.5 8,420 12 415 DC
27N 0.45 4933 K6 0.6 19,010 5 4311 DC
27N 0.45 4933 L6 A 0.6 17,960 6 1292 DC
28N 0.45 4843 K6 0.6 22,420 5 1964 DC
28N 0.45 4843 L6 A 0.6 18,610 6 717 DC
29N 0.45 4723 A C 0.5 14,130 6 1144 DC
29N 0.45 4723 E 0.5 15,950 6 1266 DC
30N 0.45 4723 J B 0.5 19,330 5 808 DC
30N 0.45 4723 E 0.5 17,210 6 823 DC
31N 0.45 4927 J B 0.5 21,090 6 733 DC
31N 0.45 4927 A C 0.5 13,300 6 1763 DC
34N 0.45 4659 H 0.5 15,940 6 1153 DC
34N 0.45 4659 F 0.5 13,570 6 968 DC
35N 0.45 4659 H 0.5 18,080 6 1202 DC
35N 0.45 4659 F 0.5 16,540 6 684 DC
36N 0.45 4451 I 0.5 12,100 6 1455 DC
36N 0.45 4451 B 0.5 13,440 6 1243 DC
37N 0.45 4724 I 0.5 14,710 6 1181 DC
37N 0.45 4724 B 0.5 15,600 6 1044 DC
38N 0.45 4153 K6 0.6 19,510 12 2079 DC
39N 0.45 4303 D E 0.5 5,240 6 635 DC

29
strands were tested along with two different 0.6 in. strands. In Test Method for the Bond of Prestressing Strands requires
Table 3.1, LC/DC refers to whether the test was conducted displacement control instead of load control.
using load control (LC) or displacement control (DC). These The Standard Test Method for the Bond of Prestressing
tests were critical to refining the test protocols and also to Strands (see Appendix H) includes specific dimensions for
determining which strand samples would provide high and the test specimens and the procedures for the test. Figure 3.2
low targets for NCHRP Project 12-60 transfer length and de- shows a schematic of the Standard Test Method for the Bond
velopment length tests. Testing also included variations in of Prestressing Strands. Additional details for the NASP Bond
water-to-cement ratio (w/c), which resulted in variations in Test are shown in Figure 3.3. Figure 3.4 shows detail for the
mortar strength. W/c ratios of 0.40, 0.45, and 0.50 were methodology employed to measure the strand end slip on its
tested. Additionally, some tests were performed using load- “free” end, i.e., the end of the strand that is not loaded in ten-
controlled protocols instead of displacement control. From sion. The photograph in Figure 3.5 shows the strand end slip
these tests, it was determined that displacement control pro- measurement device. Finally, in Figure 3.6, the photograph
vides more data that can be valuable in evaluating strand shows an entire Strand Bond Test specimen placed within the
bond performance. Therefore, the recommended Standard loading frame and ready for testing.
2'-6"
8"
5.5" LOAD TRANSDUCER WITH 22,000

LB CAPACITY
4.5" 2.5"
12 12 " x 9" x 1 14 " THK STEEL PLATE
1" BOLT
2 14" x 8" CHANNEL SECTION
8" CLEAR DISTANCE
2'-8"
Vertical Axis line
12 12 " x 9" x 1 14 " THK STEEL PLATE WITH 43" WIDE

SLOT AT THE CENTER THROUGH 4 12"
9 12 " x 9" x 1 14 " THK STEEL PLATE WITH 43" WIDE

SLOT AT THE CENTER THROUGH 4 12"
5 14" CLEAR DISTANCE
2 14" x 8" CHANNEL SECTION

7"
1" BOLT
MTS CONSOLE
Figure 3.2. Schematic diagram of NASP Test setup.

30
Consolidated mortar mix
1 6
2" or 10 " diameter strand
2" long styrofoam bond breaker

firmly attached to the strand
Base plate 6" x 6" x 14" with 58" diameter hole
welded to the specimen cylinder
Neoprene pad 6" x6"
Base plate 6"x6"x 43" steel plate
Steel plate 3 14" x 3 14" x 12" with 85" diameter hole
1
2" chuck
Figure 3.3. Details of the NASP Bond Test specimen.
3.2.2 Reproducibility of the NASP Bond Test Strand Bond Test data from OSU was developed from tests
Between Sites with 12 samples. At Purdue, all of the tests had a sample size
of six. Table 3.2 lists both the NASP identifiers (Round III and
The NASP Bond Test was performed on specific strand Round IV) and the NCHRP strand ID. Purdue performed the
samples at Purdue and OSU. Round robin trials were per- bond tests as completely blind trials—even the Purdue strand
formed on five 0.5 in. diameter and two 0.6 in. diameter, identifiers were changed from those used at OSU.
Grade 270, low-relaxation strands. The strands included in The results from the round robin testing are reported in
the round robin trials are shown in Table 3.2. Some of the Table 3.3. The five 0.5 in. diameter strand samples are Strand
A, Strand B, Strand C, Strand D and Strand E. The two 0.6 in.
13.0"
7.50"
1 43 " steel block bolted to the
aluminum plate for weight
Aluminum plate 13" x 1" x 43 in
9.0" 3" clear spring loaded LVDT (DCT

1000A)
Magnetic base with control

switch
NASP specimen casing

Mortar mix cured for 24 hours
1
2" or 106 " diameter strand
2" long styrofoam bond breaker
firmly attached to the strand
Figure 3.4. NASP Test specimen strand end slip

measurement. Figure 3.5. LVDT on NASP Test specimen.

31
match the “perfect fit” line very closely, with an R2 value of

0.92.
3.2.3 Recommendation for the Standard

Test Method for the Bond
of Prestressing Strands
The NASP Bond Test performed in this research program
was conducted on ten 0.5 in. diameter and two 0.6 in. diam-
eter strands. In the NCHRP testing, round robin tests were
performed at Purdue and OSU. As shown in Figure 3.7, the
results from the two testing sites closely match.
This research builds upon earlier work done by NASP to
develop a standard test for bond. The NCHRP research fur-
ther refined the testing protocols to the point where the test
results are now demonstrably reproducible between testing
sites. The refined test is recommended as a standard test
method to evaluate the ability of a prestressing strand to bond
with concrete.
Figure 3.6. NASP Test specimen 3.3 The NASP Bond Test
inside loading frame. in Concrete
The NASP Bond Test protocol was modified to test the
diameter samples are labeled Strand A6 and Strand B6. For strand in concrete in place of mortar. This is important to the
example, Strand E had the lowest reported results at both test- overall research because the NASP Bond Test modified for
ing sites, 5240 lb at OSU and 6070 lb at Purdue. For Strand C, concrete demonstrates the relationship between bond strength
OSU reported an average of 13,715 lb, whereas Purdue and concrete strength. The overarching conclusion from this
reported an average of 14,710 lb. Note that the table reports segment of the testing was that bond strength improves in
results of testing with both 0.5 in. and 0.6 in. strands. proportion to the square root of the concrete strength. This
The results in Table 3.3 are illustrated in Figure 3.7. Figure conclusion stems from an examination of data from three of
3.7 plots the average NASP values from OSU against the av- the 0.5 in. diameter strands and one of the 0.6 in. diameter
erage NASP values from Purdue. A linear regression line and strands. The concrete used for the modified NASP Bond Test
a “perfect fit” line are plotted in the figure. The test results had 1-day strengths varying from 4 ksi and to 10 ksi.
Table 3.2. Round-robin testing at OSU and Purdue.
NASP NASP
STRAND NCHRP
ROUND III ROUND IV
DIAMETER (IN) OSU ID OSU Purdue
ID ID
0.5 A C x
0.5 B x
0.5 FF C D x x
0.5 II D E x x
0.5 E x
0.5 F x
0.5 AA G A x x
0.5 H x
0.5 I x
0.5 J B x x
0.6 K6 B6 x x
0.6 L6 A6 x x

32
Table 3.3. Results from round-robin testing—Standard Test for the

Bond of Prestressing Strands in Concrete.
NASP Test Results NASP Test Results
Diameter (in)
STRAND ID
at OSU at Purdue
NCHRP
Strand
Mortar Strength Pull-Out Force Mortar Strength Pull-Out Force
f ci at 0.1" slip f ci at 0.1" slip
(psi) (lb) (psi) (lb)
C 0.50 4,723 14,130 4,498 14,270

C 0.50 4,927 13,300 4,810 15,150
Avg. 13,715 14,710
D 0.50 4,765 6,870 4,665 7,280
D 0.50 4,484 6,910 4,365 9,770 1
D 0.50 4,767 9,970
Avg. 6,890 8,625
E 0.50 4,303 5,240 4,000 6,070
A 0.50 4,730 20,710 4,847 2,0880
A 0.50 4,815 21,190 4,318 16,470 1
A 0.50 4,638 18,880
Avg. 20,950 19,880
B 0.50 4,723 19,330 4,893 22,700
B 0.50 4,927 21,090 4,798 22,280
Avg. 20,210 22,490
B6 0.60 4,843 22,420 4,356 19,130
B6 0.60 4,933 19,010
B6 0.60 4,153 19,510 1
Avg. 20,715 19,130
A6 0.60 4,933 17,960 4,628 15,450
A6 0.60 4,843 18,610
Avg. 18,285 15,450
1
Value omitted from average because the mortar strength was out of range.
The modified NASP Bond Test was conducted in con- NASP tests in concrete were conducted on three 0.5 in. di-
crete to understand the effects of varying concrete strengths ameter strands with NCHRP strand designations A, B, and
on the bond of prestressing strands. The test procedure was D and on one 0.6 in. diameter strand with an NCHRP
identical to the NASP Bond Test protocols discussed in strand designation of A6. The number of NASP tests con-
Section 3.2 except that concrete with varying strengths was ducted on concrete for varying concrete strengths is re-
used instead of the standard cement-sand mortar. The ported in Table 3.4. Each test listed in Table 3.4 contains six
25000
B
Average NASP Pull-out Values (lb)
20000 A
B6
Purdue University
15000 C
Linear Regression
A6 2
R = 0.92
10000 D
"Perfect" Test
5000
E
0
0 5,000 10,000 15,000 20,000 25,000
Average NASP Pull-out Values (lb)
Oklahoma State University
Figure 3.7. Comparison of NASP Bond Tests at OSU and Purdue.

33
Table 3.4. Number of NASP Bond Tests modified was performed (Tessema 2006) to arrive at the concrete mix
for concrete with varying concrete target proportions and the target fresh and hardened properties.
strengths. The results and discussion on the concrete mix proportions
are beyond the scope of this report. The mixture propor-
NASP Strand Target 1d Concrete Strengths
NCHRP tions reported in Table 3.5 were also employed to make the
Round IV Diameter (ksi)
ID (inches)
ID 4 6 8 10 transfer length and development length beams.
G A 0.5 1 1 1 1
The NASP Bond Test, modified to be tested in concrete,
J B 0.5 1 1 1 1 conforms to the same protocols for NASP Bond Testing that
C D 0.5 2 4 2 1 are found in Appendix I. The only variation is that concrete
L A6 0.6 1 1 1 1 is used in place of the sand-cement mortar. Also, concrete
slumps of 2 to 3 in. were achieved instead of the mortar flow
rates of 100 to 125. The handling and preparation of the
or more NASP specimens. The target concrete strengths for strands, the steel casing, and the bond breakers were identi-
each of the tests were 4, 6, 8, and 10 ksi. The concrete cal to the NASP Bond Tests conducted in sand-cement mor-
mixtures used for making the NASP specimens in concrete tar. The mixing procedures used for the NASP Bond Test
included Type III cement from Lafarge North America, conformed to ASTM C 192. The fresh concrete is placed in
coarse and fine aggregate from Dolese Bros. Co., cement two layers; each layer is consolidated using a handheld elec-
slag from Lafarge North America, and admixtures from De- tric vibrator. The slump, unit weight, and air content are
gussa Admixtures, Inc. Admixtures used included HRWRs, measured per ASTM C 143, ASTM C 138, and ASTM C 231,
normal range water reducers (NRWRs), and air entraining respectively. The NASP specimens and the test cylinders were
admixtures (AEAs). Table 3.5 gives the mix proportions and cured in conformance with ASTM C 192. The compressive
the target fresh and hardened properties for the concrete strength testing was conducted during the time of the NASP
cast in the modified NASP specimens. The mix proportions Bond Test in concrete, in conformance with ASTM C 39. The
were named based on the target 1-day strength. The mix NASP specimens are then kept in a laboratory curing room
C-0 targets a concrete strength of 4 ksi at release. Similarly, for 22 to 24 hr from the time of hydration. Curing conditions
C-I, C-II, and C-III target strengths of 6, 8, and 10 ksi at near 73.4 °F and 100-percent relative humidity were main-
release, respectively. The concrete mix C-IA has a target tained. The modified NASP Bond Test is performed at 24 ± 2
release strength of 6 ksi with AEA. Detailed trial batching hr after the hydration of the cement. The NASP specimen in
Table 3.5. Concrete mixture proportions for transfer and devel-

opment length testing and for the NASP Bond Test in concrete.
Concrete Mixture Designations C-0 C-I C-IA C-II C-III
Cement (PCY) 650 800 800 800 900

Cement Slag (PCY) 100
Coarse Aggregates (PCY) 1,800 1,703 1,800 1,805 1,747
Fine Aggregates (PCY) 1,243 1,203 922 1,219 1,183
Water (PCY) 298 303 272 277 251
Glenium 3200 (fl oz/cm. wt) 10 14 7
Glenium 3400 (fl oz/cm. wt) 8 5 5.5
Polyheed 997 (fl oz/cm. wt) 3
MB-AE 90 (fl oz/cm. wt) 1.88
Target Properties for Fresh and Hardened Concrete
1-Day Strength (ksi) 4 6 6 8 10
28-Day Strength (ksi) 6 8 8 10 14
56-Day Strength (ksi) n/a 10 10 14 15
Slump (in) 8 8 8 8 9
Unit Weight (pcf) 145 148 148 150 157
Air Content (%) 2 2 6 2 2

34
concrete is mounted on a rigid steel frame in the same man- (S) for each set of tests are reported for the modified NASP
ner described for the NASP Bond Test (in mortar). Bond Test in concrete.
3.3.1 Results from the NASP Bond Tests 3.3.2 Discussion of the Results from the
in Concrete NASP Bond Tests in Concrete
The NASP Bond Test standardized in mortar was con- Figure 3.8 shows the pull-out values from the Modified
ducted in concrete to understand the effect of concrete NASP Bond Test for Strands A and B plotted versus the
strengths on the NASP Bond Test. The results from this ex- concrete strength. There are a total of 8 data points, also re-
perimental testing are reported in Table 3.6. Table 3.6 reports ported in Table 3.6, found in Figure 3.8. Both linear regres-
the NCHRP Strand ID, the NASP Strand ID (for comparison sion and the power regression curves are plotted on the
purposes), the 1-day concrete strength (f c′i), the NASP Bond figure. The coefficient of determination (R2) value for both
Test result (from the Standard Test for Strand Bond in mor- the regressions is 0.82. The linear and the power best-fit
tar), and the NASP Bond Test when modified and performed equations are reported in Figure 3.8. Figure 3.8 clearly shows
in concrete. The table reports the w/cm (water–cementitious that increases in concrete strength result in a higher NASP
materials) ratio because there were pozzolanic materials pull-out value for NCHRP Strands A and B. Note that the
added for some of the concrete mixtures reported in Table NASP Bond Test pull-out value for the standardized test in
3.5. The concrete strengths reported in Table 3.6 are averages mortar is 20.95 kips for NCHRP Strand A and 20.21 kips for
of three or more concrete specimens tested during the NASP Strand B. Also, note that the regression plots cross the 4 ksi
test. The number of NASP Bond Test specimens (N) that concrete strength at a corresponding NASP Bond Test
were included as part of the test and the standard deviation (modified) value of about 23 kips.
Table 3.6. Results of NASP Bond Tests in concrete.
Strand ID Concrete NASP Test Results

NCHRP ID
in Mortar
NASP
STRAND
Diameter
NASP IV
S
Strand
NASP
(kips)
f ci Value
(in.)
f ci
ID
w/cm N (ksi)
(ksi) (kips)
ksi
A G 0.5 0.425 4.52 2.13 23.58 6 0.66
A G 0.5 0.38 7.02 2.65 26.35 6 1.44
20.95
A G 0.5 0.36 8.05 2.84 30.68 6 1.77

A G 0.5 0.235 11.79 3.43 35.29 6 2.33
B J 0.5 0.46 3.56 1.89 22.55 6 5.57
B J 0.5 0.4 5.58 2.36 30.8 6 1.04
20.21
B J 0.5 0.32 7.11 2.67 28.78 6 4.55

B J 0.5 0.24 10.06 3.17 34.33 6 4.17
D C 0.5 0.45 4.71 2.17 7.48 6 2.76
D C 0.5 0.46 4.56 2.13 6.66 6 2.52
D C 0.5 0.36 6.99 2.64 8.96 6 2.23
D C 0.5 0.38 7.34 2.71 9.51 6 2.64
6.89
D C 0.5 0.4 6.13 2.48 6.74 6 0.25

D C 0.5 0.3 8.67 2.94 10.26 6 0.26
D C 0.5 0.32 8.34 2.89 9.97 6 1.06
D C 0.5 0.26 9.95 3.15 11.56 6 0.84
A6 L 0.6 0.46 2.23 1.49 11.6 6 0.61
A6 L 0.6 0.38 5.02 2.24 23.13 6 1.24
18.29
A6 L 0.6 0.28 8.79 2.96 24.84 6 0.82

A6 L 0.6 0.235 10.42 3.23 28.74 6 1.39

35
45.00
NASP Pull-out Value in Concrete (kip)

40.00
35.00
30.00
y = 1.5293x + 18.02
25.00 R2 = 0.8153
20.00
15.00
y = 14.078x0.3734
10.00 R2 = 0.8222
5.00
0.00
0 2 4 6 8 10 12 14
Concrete Strength (ksi)
Figure 3.8. Pull-out values from the modified NASP Bond Test for
Strands A/B versus concrete strength.
Figure 3.9 illustrates the pull-out values from the Modified that the regression plots cross the 4 ksi concrete strength at
NASP Bond Test for NCHRP Strand D plotted against con- a corresponding NASP Bond Test (modified) value of about
crete strength. NCHRP Strand D had a NASP Bond Test 6 kips.
value of 6.89 kips in the standardized test, which was lower Figure 3.10 makes the same comparison as Figures 3.8 and
than the standardized NASP Bond Test values of Strands A 3.9, but for 0.6 in. diameter strand, NCHRP A6. The data
and B. There are a total of 8 data points for Strand D, and the shown in Figure 3.10 are also reported in Table 3.6. Both
data shown in Figure 3.9 correspond to data reported in Table linear regression and the power regression curves are plotted
3.6. Linear regression and the power regression curves are on Figure 3.10 for NCHRP Strand A6. Please note that the ex-
plotted on Figure 3.9 for Strand D. The coefficient of deter- ponent in the best-fit power curve is approximately 0.56. As
mination (R2) values are 0.89 for the linear regression and in Figures 3.8 and 3.9, Figure 3.10 clearly shows that increases
0.84 for the power regression. The linear and the power best- in concrete strength result in higher NASP pull-out values for
fit equations are also reported in the figure. Figure 3.9 clearly NCHRP Strand A6. Please note that the NASP Bond Test
shows that increases in concrete strength result in a higher pull-out value for the standardized test in mortar is 18.29 kips
NASP pull-out value for NCHRP Strand D. Please note that for the NCHRP Strand A6, and that the regression plots cross
the NASP Bond Test pull-out value for the standardized test the 4 ksi concrete strength at a corresponding NASP Bond
in mortar is 6.89 kips for the NCHRP Strand D. Also, note Test (modified) value of about 17.5 kips.
14.00
NASP Pull-out Value in Concrete (kip)
12.00
10.00
y = 0.8833x + 2.6335
8.00
R2 = 0.8917
6.00
4.00
y = 2.4161x0.6677
2.00
R2 = 0.8431
0.00
0 2 4 6 8 10 12 14
Strand D versus concrete strength.

36
45.00
NASP Pull-out Value in Concrete (kips)

40.00
35.00
30.00
25.00
y = 1.8439x + 9.8805
20.00 R2 = 0.8566
15.00
10.00 y = 7.9751x0.5565
5.00
R2 = 0.9203
0.00
0 2 4 6 8 10 12 14
Strand A6 versus concrete strength.
Table 3.6 also reports values for the square of the 1-day The data presented in Table 3.6 are normalized and pre-
concrete strength. The NASP pull-out values and the square sented all together in Figure 3.12. The NASP Bond Test val-
root of the concrete strength are presented for all the strands ues were normalized by dividing by the NASP Bond Values in
tested in concrete in Figure 3.11. The linear best fit-lines are concrete by the Standard NASP Bond Test (in mortar) values.
plotted in the figure with the corresponding R2 values for the Figure 3.12 includes data from all three 0.5 in. diameter
four strands tested in the modified NASP test in concrete. In strands and the one 0.6 in. diameter strand. Concrete
Figure 3.11, the best-fit curves tend to have a steeper slope for strength at 24 hr from the modified NASP Bond Test is plot-
strands with higher NASP values in the same range of con- ted against normalized NASP values. The data are plotted
crete strengths. The NASP value increases with increases in against a best-fit power regression curve, also shown in Fig-
concrete strength, and the high-performing strands have a ure 3.12. The R2 value for the test data is 0.80, indicating that
steeper best-fit line. Thus, for a given change in the concrete the power regression equation closely agrees with the test
strength, the NASP results can have a higher variation for the data. The best-fit equation is given in Equation 3.1.
high-performing strands (strands with higher NASP values)
when compared with the moderately performing strands ( NASPconcrete )
= 0.49139 f ci′0.51702 (3.1)
(strands with lower NASP values). NASP
40.00
35.00
Strand B
2
R = 0.7003 Strand A
R2 = 0.9351
30.00
NASP Pull-out Value (kip)
25.00
20.00 Strand A6
2
R = 0.9053
15.00
10.00 Strand D
R2 = 0.8166
5.00
0.00
0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00
fci′
0.5
(ksi)
Figure 3.11. NASP pull-out values versus fci′ for all strands.

37
2.00
1.80
NASP in Concrete Normalized by

NASP in Mortar (kips/kips)
1.60
1.40
1.20
1.00
y = 0.49139x0.51702
0.80 R2 = 0.79826
0.60
0.40
0.20
0.00
0 2 4 6 8 10 12 14
Concrete Strength at 24 hours (ksi)
NCHRP A NCHRP B NCHRP D NCHRP A6 (0.6")
Figure 3.12. Normalized NASP pull-out values versus concrete

strength for all strands.
where The result of the regression is remarkable for two reasons.

NASPconcrete = the value obtained from the NASP Bond One, the data’s best fit regression demonstrates a coefficient
Test in concrete, and of determination of 0.79, illustrating that the data set is fairly
NASP = the value obtained from the Standard NASP well predicted by the regression; two, the data demonstrate
Bond Test (NASP Bond Test in mortar). that bond improvements are directly proportional to the con-
crete strength at 1 day of age. Furthermore, the normalized
The equation is modified to fit the NASP values as a func-
value of 1.0 is achieved at an f c′i of 4 ksi. These significant re-
tion of the square root of concrete strengths. In Figure 3.13,
sults are used later in the recommendation for transfer and
the normalized NASP pull-out values are plotted against the
development length code expressions. Also note that the
square root of the concrete strength. The linear regression
modified NASP Bond Test in concrete nearly matches the
results in the following equation:
Standard NASP Bond Test if the concrete strength is only
( NASPconcrete ) 4 ksi, as compared to the requirement for mortar strength of
= 0.51 fci′ (3.2)
NASP 4,500 to 5,000 psi.
2.00
1.80
NASP in Concrete Normalized by
1.60
NASP in Mortar (kips/kips)
1.40
1.20 y = 0.5096x
R2 = 0.7889
1.00
0.80 y = 0.51 f ci
0.60
0.40
0.20
0.00
0 1 2 3 4 5
Square Root of Concrete Strength at 24 hours ( ksi)
NCHRP A NCHRP B NCHRP D NCHRP A6 (0.6") Linear (all)
Figure 3.13. Normalized NASP pull-out values versus fci′ .

38
3.4 Measured Transfer Lengths on transfer lengths. Three different sources for 0.5 in. diam-
versus Varying Concrete eter strand and one source for 0.6 in. diameter strand were
Strengths and Varying NASP employed in this research program.
Bond Test Values Figure 3.18 shows the details of the I-shaped cross section.
In the I-shaped beams made with 0.5 in. diameter strands,
The research aims at assessing the effects that varying con- four strands were located within the bottom bulb of the cross
crete strength can have on strand bond. This section deals section with a fifth strand located 2 in. from the top of the
primarily with transfer lengths measured on pretensioned cross section. In the I-beams made with 0.6 in. diameter
beams. Variables included strand with varying bond quality strands, three strands were located within the bottom bulb of
and concrete strengths varying between 4 ksi at release and 10 the cross section with a fourth strand located 2 in. from the
ksi at release. Beams were either rectangular in shape or top of the cross section.
I-shaped. A total of 43 rectangular-shaped beams and 8
I-shaped beams were cast using 4 different strand sources.
3.4.1 Fabrication of Beams
The number of beams made and the corresponding research
variables are reported in Table 3.7. Two-strand rectangular Transfer lengths were measured at release on all the beams
beams included two strands placed near the bottom of the using strand end slips. On some of the beams, transfer lengths
cross section. The four-strand rectangular beams had two were measured using a detachable mechanical strain gage
strands placed near the bottom and two strands placed near (DEMEC gage), which effectively measures changes in con-
the top. Beams were cast using both 0.5 in. and 0.6 in. diam- crete surface strains. The transfer lengths measured from
eter strands. Figure 3.14 illustrates the beam numbering strand end slips are compared with those measured using the
system that describes the variables that are contained within DEMEC gage.
each beam specimen. Cross section details for the rectangu- The rectangular beams were 17 ft in length with a cross
lar beams are found in Figures 3.15 and 3.16. section that was 6.5 in. wide by 12 in. high. Two #6 bars
Figure 3.17 depicts some of the rectangular beams during were placed within 1 in. of the top of the cross section in all
fabrication, prior to release of the prestressing strands. Beams rectangular beams to ensure ductile flexural failures. The
were made with a target 1-day concrete strength of 4,000, cross section for the I-shaped beams is shown in Figure
6,000, 8,000 and 10,000 psi. The 6,000-psi release strength 3.18. The beams were fabricated 24 ft in length. The beams
concrete beams were made using both air-entrained and non- cast had 0.5 in. diameter strands and 0.6 in. diameter
air-entrained concrete to study the effects of air entrainment strands. All of the I-beams contained horizontal web rein-
Table 3.7. Number of transfer length beams and research variables

employed.
0.6 in
Target Concrete 0.5 in Diameter Strands Diameter
Target Air
Release Design Strands
Content
Strength Strength
(%)
(ksi) (ksi) Strand A Strand B Strand D Strand A6
Two-Strand Rectangular Beams

4 6 2 0 2 2 2
6 10 2 2 0 2 3
6 10 6 2 0 2 0
8 14 2 2 0 2 3
10 15 2 2 0 2 3
Four-Strand Rectangular Beams
6 10 2 2 0 2 0
8 14 2 2 0 2 0
10 16 2 2 0 2 0
I-Shaped Beams
6 10 2 1 0 1 2
10 15 2 1 0 1 2

39
Beam Shape Internal hoop reinforcements were placed in the form of

R: Rectangular 6.5” x 12” triangular cages at both ends of the beam.
I: “I” Shaped beams Strands were tensioned to 75 percent of fpu (the guaran-
teed breaking strength) or 202.5 ksi. The expected elonga-
Strand Source
A, B, or D for 0.5 in. strand tions were calculated and compared with the measured
A for 0.6 in. strand elongations to ensure proper stressing. The strands were
stressed to an initial level of 2,000 lb. Once the force on the
Nominal Concrete Strength at Release strand reached 2,000 lb, the strand was marked with a per-
4, 6, 8, or 10 ksi
A6: for 6 ksi with Air Entrainment manent marker coinciding with the datum level marking
on the prestressing bed. The strand was then stressed to
202.5 ksi. The elongation was then measured as the distance
the mark on the strand moved from the datum marking on
RA6-5-1-T the prestressing bed.

Concrete was batched onsite at Coreslab’s batch plant.
Fresh properties of concrete, slump, unit weight, and air con-
tent were checked before casting the concrete. Extensive trial
batching was performed (Tessema 2006) to determine the
Strand Size
5 for 0.5 in. diameter fresh and hardened properties of the concrete mix designs. If
6 for 0.6 in. diameter the fresh properties of unit weight, slump, or air content did
not meet with the design expectations, the concrete was not
used. Concrete cylinders were made at the site and placed in
Specimen Number the same prestressing bed as the test beams until transfer.
1, 2, or 3 is the number in a series Steam curing was used if the ambient temperatures were low.
of companion beams
The test beams together with the concrete cylinders were kept
under cover if steam curing was used.
Top Strand 3.4.2 Measuring Transfer Lengths

If the rectangular beam contains top
strands, T is used. Not applicable
for “I” shaped beams
Transfer lengths were measured on all strands by measur-
ing the distance each strand slipped into the concrete after
prestress release. A depth micrometer was used in combina-
Figure 3.14. Beam number identification.
tion with specially made clamps to measure the strand end
slip. Figure 3.19 shows the depth micrometer measuring
forcement consisting of four or two #4 bars, 96 in. long, lo- strand end slips immediately after prestress release.
cated near the ends of the beams and anchored with stan- Strand end slips are directly related to measured transfer
dard hooks. Two horizontal #4 bars were placed at the lengths, as shown in Figure 3.20. In Figure 3.20, stresses are
south end of every beam, and four horizontal #4 bars were used to indicate the loss of prestress caused by elastic short-
placed at the north end. The deck slab contained two #3 ening (ES). After release, ES is the primary prestress loss. The
straight bars in the longitudinal direction in the deck slab. transfer length of the strand is directly related to the area of
17'
6½
2
2 – ½ in. Ø Strands
2- #6 Bars 16'-8"
12
#3 Tie at 6" c/c
2 2- ½ in. Ø Strands
Figure 3.15. Details of four-strand beams.

40
17'
6½
2
2 – #6 Bars 16'-8"
12
#3 Tie at 6" c/c
2 2- ½ in. Ø Strands
Figure 3.16. Details of two-strand beams.
the shaded triangle shown in Figure 3.20. The shaded area di-
vided by the elastic modulus of the strand gives the strand end
slip measurement. Thus, by measuring the strand end slip,
the transfer length can be calculated directly. Over time, the
beam experiences additional losses and a lengthening of the
transfer length. The transfer length over time is illustrated in
Figure 3.20 by the larger, unshaded triangle. In Figure 3.20, fsi
is the stress in the prestressing strand just prior to release, and
fse is the strand stress after all losses. ES is the elastic shorten-
ing loss that occurs immediately upon release of the pre-
stressing force.
Changes in concrete surface strains were measured on some
of the specimens using a DEMEC gage. The DEMEC gage
is pictured in Figure 3.21. DEMEC target points were set at
100-mm spacings. The DEMEC gage spans 200 mm, so read-
Figure 3.17. Fabrication of rectangular beams. ings were taken over a 200-mm gage length. The procedure
23
# 3 bars on deck at 9" c/c and 2 bars

throughout the length
1.5
2
# 3 stirrups at 7" c/c
3
24 # 4 bars with standard hooks 2" c/c for 96"
23
20.5 20 from ends
4 bars at north end and 2 bars at south end
# 3 bars 4" c/c shape for internal hoop

6.5 reinforcement for 72" from end
3
Prestressing strand
10 Mild steel reinforcement
Figure 3.18. Details of I-shaped beams.

41
Figure 3.21. Concrete surface strain measurements

with DEMEC gage.
requires initial readings to be made prior to strand cutting.

After release, the measurements are repeated, and the differ-
ences can be plotted as a strain profile, such as the one shown
Figure 3.19. Strand end slip measurement using a in Figure 3.22. As shown, concrete strains on the north end
micrometer. and the south end are plotted along the length of the beam.
The strain profile is “smoothed” by averaging three measure-
ment points. The Average Mean Strain (AMS) is found out by
averaging the points on the strain plateau on the north and the
Initial Losses south sides independently. The measured transfer length
obtained from the DEMEC readings is the location where the
95-percent AMS line intersects the Smoothed Strain profile.
f si (f si – ES) f se
3.4.3 Results of the Transfer Length
Measurements
Results of the transfer length measurements are reported
Lt
in several tables, generally organized by strand type. Table
3.8 reports the transfer lengths computed from measured
Figure 3.20. Variation in strand stress variations strand end slips on Strands A and B. Table 3.8 reports trans-
with length and relation to strand end slip fer lengths only on strands located at the bottom of the
measurements. cross sections. Table 3.8 reports a transfer length for each
Concrete Strains (10-6 in/in)

450
400
350
300
250
200
150
100
50
0
0 2 4 6 8 10 12 14 16
Length of the Beam (north to south)
Unsmoothed Profile Smoothed Profile
Figure 3.22. Concrete strain profile highlighting strand transfer

lengths.

42
Table 3.8. Summary of transfer lengths at release for

bottom Strands A/B.
X S f ci f c (56d )
Beam Number Location North South
(kips) (kips) (psi) (psi)
RB4-5-1 East 17.06 18.31
20.12
West 19.78 18.66
2.56
4,033 7,050
RB4-5-2 East 24.13 22.47
West 18.1 22.45
RA6-5-1 East 20.66 20.24
West 17.68 16.16
RA6-5-2 East 15.94 11.78
18.02
West 17.12 18.23
2.23
6,183 8,500
RA6-5-1T East 19.39 18.7
West 20.62 18.93
RA6-5-2T East 18.7 18.84
West 19.07 16.27
RA8-5-1 East 12.01 13.09
West 14.58 13.9
RA8-5-2 East 13.9 11.74
13.63
West 15.93 12.42
1.32
8,570 13,490
RA8-5-1T East (a) 12.51
West (b) 14.71
RA8-5-2T East 14.52 15.6
West 12.55 13.36
RA10-5-1 East (c) (d)
West (e) 13.57
RA10-5-2 East 12.75 15.25
13.72
West 12.75 14.8

2.27
9,711 14,470
RA10-5-1T East 17.74 12.06
West 18.16 11.32
RA10-5-2T East 12.2 11.78
West 11.46 14.53
(a) Lt of 1.48 in. not included (c) Lt of 25.65 in. not included
(b) Lt of 5.26 in. not included (d) Lt of 5.82 in. not included
(e) Lt of 22.89 in. not included
strand, two at each end of the beam, with each end of the Table 3.10 reports the measured transfer lengths for Strand
beam designated as either north or south; thus, all together, A, placed near the tops of cross sections in the respective
four transfer length measurements are reported for each beams. Again, no clear pattern emerges of the top strands
beam. having longer transfer lengths than the bottom strands.
Table 3.8 also reports the average transfer length for all of Table 3.11 reports the measured transfer lengths for 0.5 in
the transfer length measurements on beams for a particular diameter Strand D. Table 3.11 includes the beam number, the
concrete strength, X . The standard deviation, S is reported in measured transfer length for each strand, the average transfer
inches for the data set. The release strength, fci′ (psi), is the length for Strand D by concrete strength, the standard devia-
average of at least three 4 in. by 8 in. cylinders. The 56-day tion of the transfer lengths, and 1-day and 56-day concrete
strength, fci′ (56d), is the average of three cylinders placed in strengths.
laboratory curing conditions. Table 3.12 reports transfer lengths measured on Strand D
Table 3.9 reports the measured transfer lengths on beams placed in top locations of four-strand beams. Again, there is
that contained air-entrained concrete. Only Strand A was used no clear pattern of top strands having longer transfer lengths
for this set of beams. While the transfer lengths measured in than bottom strands.
air-entrained concrete appear to be longer than the transfer Table 3.13 reports transfer lengths measured on I-shaped
lengths measured in the companion beams without air en- beams, including data from both 0.5 in. diameter strands—
trainment, no clear pattern emerges with the limited data. Strand A and Strand D—and data from the 0.6 in. diameter

43
Table 3.9. Summary of transfer lengths at release of

bottom Strands A/B in air-entrained concrete.
Beam Number Location North South X S f ci f c (56d )

RA6A-5-1 East 19.26 17.47
West 16.22 17.88
20.49
3.39
7,960 11,420
RA6A-5-2 East 26.41 22.63
West 22.6 21.42
RD6A-5-1 East 36.25 30.04
West 34.55 28.15
26.26
6.93
7,960 11,420
RD6A-5-2 East 21.16 21.79
West 19.79 18.36
strand, Strand A6. As in Tables 3.8 through 3.12, measured transfer length measured using both methods. Figure 3.23
transfer lengths are reported for each strand, the average and presents the data from Tables 3.15 through 3.17 graphically
standard deviation are reported for each beam, along with and shows that generally the transfer lengths measured by the
concrete strengths at release and at 56 days. Table 3.13 DEMEC gage are approximately the same as the transfer
includes data collected from strands located in the bottom lengths obtained from strand end slip measurements.
bulbs on the I-shaped beams only. Tables 3.18 through 3.22 provide the transfer length meas-
Table 3.14 reports the measured transfer lengths on top urements over time, from release through 240 days after
strands from the I-shaped beams. The data are erratic, so release. Strand end slips can be measured individually for
no conclusions can be drawn from these measurements. each strand. In the tables reporting measured transfer
All of the transfer lengths reported in Tables 3.8 through lengths from strand end slips, the east strand is represented
3.14 report transfer lengths measured immediately after in the column headed by “E” whereas the west strand is re-
release. ported in the columns headed by “W.” Transfer lengths were
Tables 3.15, 3.16 and 3.17 all include both transfer length not measured beyond release for beams RB4-5-1 and RB4-5-2.
measurements made with the DEMEC gage and transfer As the data indicate, transfer lengths grow over time, and the
length measurements made from strand end slips for com- 240-day transfer lengths are considerably longer than
parison. Approximately 43 percent of the beam ends had the transfer lengths measured at release. All of the transfer
Table 3.10. Summary of transfer lengths at release of

Strand A in top locations.
X S f ci f c (56d )
RA6-5-1T East 21.03 19.11
19.04
West 19.47 20.58

2.07
6,183 8,500
RA6-5-2T East 17.07 16.52
West 21.82 16.71
RA8-5-1T East 13.38 14.88
14.55
West 10.74 14.42

1.96
8,570 13,490
RA8-5-2T East 17.61 15.7
West 15.12 14.56
RA10-5-1T East 14.93 13.5
West 14.53 11.32
12.90
1.65
9,711 14,470
RA10-5-2T East 10.63 11.2
West 14.11 12.99

44
Table 3.11. Summary of transfer lengths at release for bottom

Strand D.
X S f ci f c (56d )
RD4-5-1 East 31.69 32.11
32.90
West 33.88 29.93
2.64
4,033 7,050
RD4-5-2 East 36.9 (a)
West (b) (c)
RD6-5-1 East 29.88 30.42
West 30.6 25.71
RD6-5-2 East 25.35 30.15
26.19
West 25.84 28.29
2.99
6,183 8,500
RD6-5-1T East 23.89 25.12
West 23.43 26.59
RD6-5-2T East 25.53 19.93
West 24.67 23.71
RD8-5-1 East 21.16 20.89
West 19.13 19.41
RD8-5-2 East 16.79 21.43
West 10.54 13.17 20.94
6.05
8,570 13,490
RD8-5-1T East 35.63 29.78
West 15.94 26.34
RD8-5-2T East 20.87 21.99
West 18.99 23.01
RD10-5-1 East 23.48 16.16
West 28.59 17.54
RD10-5-2 East 13.95 19.33
18.36
West 15.74 17.12

3.72
9,711 14,470
RD10-5-1T East 21.76 16.22
West 21.1 17.4
RD10-5-2T East 16.36 15.25
West 17.13 16.58
(a) Excessive movement of the beams during flame cutting, Lt observed as 50.43 in.
(b) Excessive movement of the beams during flame cutting, Lt observed as 47.48 in.
(c) Excessive movement of the beams during flame cutting, Lt observed as 48.98 in.
length measurements over time were made using the strand Figure 3.24 illustrates the transfer length measurements at
end slip method. release plotted against the concrete strengths at 1 day of age
for Strands A/B. (Although Strand A and Strand B represent
two different sources of strand, their NASP Bond Test values
3.4.4 Discussion of Transfer Length
were very similar; therefore, the data from the two strands are
Measurements
treated as part of one data set.) Two regression curves are
The discussion on transfer lengths focuses on two essential shown in Figure 3.24; one shows the best fit for data derived
elements: (1) what effects, if any, concrete strength has on from the DEMEC gage, and the other shows the best fit for
transfer length and (2) whether the NASP Bond Test provides the data derived from strand end slip measurements. Both re-
an indicator regarding transfer length. Another objective of gression curves in Figure 3.24 show that transfer lengths
this discussion is to present to the industry a reasonable code shorten as concrete strengths increase.
equation to adequately predict the transfer lengths of preten- Figure 3.25 shows the transfer length measurements at re-
sioned strands. lease plotted against the concrete strengths at 1-day of age for

45
Table 3.12. Summary of transfer lengths at release for

Strand D in top locations.
X S f ci f c (56d )
Beam Number Location North South (kips) (kips) (psi) (psi)
RD6-5-1T East 27.91 21.46
23.76
West 27.52 20.26
3.03
6,183 8,500
RD6-5-2T East 23.7 20.2
West 23.61 25.43
RD8-5-1T East 22.06 16.45
22.64
West 17.61 18.84
4.68
8,570 13,490
RD8-5-2T East 27.82 23.71
West 28.67 25.94
RD10-5-1T East 16.79 15.56
15.93
West 17.27 16.32
2.22
9,711 14,470
RD10-5-2T East 18.98 15.02
West 16.19 11.29
Table 3.13. Summary of transfer lengths at release for I-shaped

beams—bottom Strands B and D (0.5 in.) and Strand A6 (0.6 in.).
X S f ci f c (56d )
IB6-5-1 East 16.12 6.42

10.77
West 17.82 2.9 5.43 5,810 9,350

Cent. 10.93 9.45
Midd. 16 6.48
IB10-5-1 East 11.14 12.45
10.59
West 10.03 5.8

2.15
7,615 13,490
Cent. 11.6 12.45
Midd. 11.31 9.9
ID6-5-1 East 24.47 12.23
18.49
West 23.47 2.56

9.88
5,492 9,840
Cent. 26.69 (a)
Midd. 28.96 11.04
ID10-5-1 East 19.03 19.03
20.82
West 20.34 23.61

2.8
8,225 14,160
Cent. 15.99 21.13
Midd 23.51 23.94
IA6-6-1 East 18.36 16.33
21.17
4.68
West 29.83 22.21

Cent. 20.15 20.15 4,381 8,990
IA6-6-2 East 9.62 14.18
16.04
4.49
West 15.48 19.47

Cent. 22.58 14.92
IA10-6-1 East 9.4 21.15
13.29
5.91
West 14.35 5.81 10,480 14,990

Cent. 10.19 18.85
IA10-6-2 East 17.94 10.64
14.72
3.46
West 13.85 10.76 10,590 14,930

Cent. 17.83 17.32
(a) Spalling of concrete surface during flame cutting

46
Table 3.14. Summary of transfer lengths at release for top

strands in I-shaped beams.
X S f ci f c (56d )
IA6-6-1 Top 22.84
9.36 4,381 8,990
18.57 6.23
IA6-6-2 Top 20.22
21.84
IA10-6-1 Top 3.82 2.87 1.35 10,480 14,990
1.91
IA10-6-2 Top 9.3 9.17 0.18 10,590 14,930
9.04
IB6-5-1 Top 21.43 13.80 10.80 5,810 9,350
6.16
ID6-5-1 Top 36.25 33.12 4.4 5,492 9,840
29.99
ID10-5-1 Top (a) 16.86 - 8,225 14,160
16.86
(a) End clamp loosened during detensioning
Table 3.15. Transfer length at release measured by DEMEC gage

and strand end slip for 0.5-in. Strands A/B.
Strand End Slips DEMEC

Beam North (in.) South (in.) North (in.) South (in.)
RB4-5-1 18.4 18.5 24.2 27.1
RB4-5-2 21.1 22.5 – –
RA6A-5-1 17.7 17.7 16.0 17.5
RA6A-5-2 24.5 22.0 – –
RA6-5-1 19.2 18.2 – –
RA6-5-2 16.5 15.0 – –
RA6-5-1-T 20.3 19.8 – –
RA6-5-2-T 19.4 16.6 – –
RA8-5-1 13.3 13.5 14.3 12.0
RA8-5-2 14.9 12.1 – –
RA8-5-1-T 12.1 14.7 12.0 15.6
RA8-5-2-T 16.4 15.1 – –
RA10-5-1 24.3 9.7 24.3 14.4
RA10-5-2 12.8 15.0 – –
RA10-5-1-T 14.7 12.4 12.5 11.7
RA10-5-2-T 12.4 12.1 – –
IB6-5-1 12.2 Not available 15.2 –
IB10-5-1 11.1 Not available 11.0 –
– measurements were not taken.

47
Table 3.16. Transfer length at release measured by DEMEC gage

and strand end slip for 0.5-in. Strand D.

RD4-5-1 32.8 31.0 25.6 24.8
RD4-5-2 36.9 Not available – –
RD6A-5-1 35.4 29.1 39.0 26.4
RD6A-5-2 20.5 20.1 – –
RD6-5-1 30.2 28.1 – –
RD6-5-2 25.6 29.2 – –
RD6-5-1-T 27.7 20.9 – –
RD6-5-2-T 23.7 22.8 – –
RD8-5-1 20.2 20.2 11.3 18.5
RD8-5-2 13.7 17.3 – –
RD8-5-1-T 19.8 17.6 12.4 12.0
RD8-5-2-T 28.2 24.8 – –
RD10-5-1 26.0 16.9 23.4 19.4
RD10-5-2 14.8 18.2 – –
RD10-5-1-T 17.0 15.9 16.1 15.7
RD10-5-2-T 17.6 13.2 – –
ID6-5-1 25.2 Not available 25.9 –
ID10-5-1 17.5 Not available 19.7 –
Table 3.17. Transfer Length at release measured by DEMEC gage

and strand end slip for 0.6-in. Strand A6.

RA4-6-1 33.4 25.0 31.4 30.3
RA4-6-2 30.2 29.3 – –
RA6-6-1 29.7 28.2 22.4 21.1
RA6-6-2 31.7 30.1 – –
RA6-6-3 25.8 33.6 – –
RA8-6-1 28.2 29.2 19.5 22.0
RA8-6-2 28.2 25.7 – –
RA8-6-3 22.8 28.3 – –
RA10-6-1 20.0 21.9 16.6 15.0
RA10-6-2 15.6 21.8 – –
RA10-6-3 16.3 22.7 – –
IA6-6-2 24.3 26.1 15.9 16.2
IA10-6-1 18.0 Not available 11.3 –
IA10-6-2 16.0 Not available 16.5 –

48
40.00
35.00
Transfer Length from DEMEC (in)
30.00
R2 = 0.5768
25.00 0.5" A Strand
0.5" B Strand
20.00
0.5" D Strand
15.00 0.6" A6 Strand
Perfect Fit
10.00
Linear (All Strands)
5.00
0.00
0.00 5.00 10.00 15.00 20.00 25.00 30.00 35.00 40.00
Transfer Length from End Slips (in)
Figure 3.23. Transfer lengths measured by DEMEC gage versus transfer lengths
measured by strand end slip.
Table 3.18. Change in transfer lengths over time for bottom 0.5 in. diameter Strands A/B in
two-strand rectangular beams.
Transfer Length at
Release from Strand End Transfer Length after 60 Transfer Length after 90 Transfer Length from 240
Slips Days from Strand End Slips Days from Strand End Slips Days from Strand End Slips
Beam (in.) (in.) (in.) (in.)
Number
North South North South North South North South
Average W & E Average W & E Average W & E Average W & E
RB4-5-1 18.42 18.48
RB4-5-2 21.11 22.46
RA6-5-1 19.17 18.20 33.07 28.86 33.07 29.55 33.69 30.03
RA6-5-2 16.53 15.01 26.57 20.82 26.56 23.38 27.95 23.52
RA6A-5-1 17.74 17.68 25.23 26.62 26.33 28.14 26.54 28.55
RA6A-5-2 24.50 22.02 28.92 27.72 31.41 29.03 31.75 29.38
RA8-5-1 13.30 13.50 15.59 21.13 17.68 21.46 24.91 22.54
RA8-5-2 14.92 12.08 22.07 19.24 23.69 19.71 35.23 19.98
RA10-5-1 24.27 9.69 23.92 9.83 24.13 12.11 24.34 13.14
RA10-5-2 12.75 15.02 16.47 16.67 18.05 17.23 19.15 17.30

49
Table 3.19. Change in transfer lengths over time for bottom 0.5-in. diameter Strand D in
two-strand rectangular beams.
Transfer Length at
Number
RD4-5-1 32.78 31.02

RD4-5-2 42.19 49.70
RD6-5-1 30.24 28.07 43.00 38.57 46.82 44.37 49.75 45.26
RD6-5-2 25.60 29.22 36.79 39.87 41.99 44.72 44.24 48.27
RD6A-5-1 35.40 29.10 37.39 34.47 39.40 36.41 39.94 37.16
RD6A-5-2 20.48 20.08 26.26 35.24 30.73 39.37 32.39 40.07
RD8-5-1 20.15 20.15 28.34 26.55 32.66 30.33 39.08 34.54
RD8-5-2 13.66 17.30 34.14 46.08 36.82 47.73 37.38 50.41
RD10-5-1 26.03 16.85 26.31 25.27 26.45 26.51 30.24 27.14
RD10-5-2 14.85 18.23 17.47 20.16 18.71 22.30 22.30 22.03
Table 3.20. Change in transfer lengths over time for 0.6-in. Strand A6 in two-strand rectangular
beams.
Transfer Length at
Number

RA4-6-1 33.42 24.98
RA4-6-2 30.24 29.35
RA6-6-1 29.73 28.19
36.87 41.73 39.00 44.45 40.85 55.13
RA6-6-2 31.65 30.10
47.03 46.36 49.24 48.20 52.18 49.37
RA6-6-3 25.83 33.63
39.73 44.60 44.08 44.82 44.96 45.93
RA8-6-1 28.21 29.17
42.46 41.87 43.87 43.26 45.48 43.41
RA8-6-2 28.20 25.70
42.68 38.55 46.28 42.35 46.35 42.37
RA8-6-3 22.80 28.26
36.85 44.00 41.17 46.93 43.00 49.22
RA10-6-1 20.03 21.92
25.69 25.77 28.08 28.82 29.98 32.15
RA10-6-2 15.62 21.78
20.99 25.99 26.14 29.47 26.79 30.70
RA10-6-3 16.34 22.73
24.46 28.82 26.13 32.30 27.73 33.32

50
Table 3.21. Change in transfer length over time for 0.5-in. Strand A in four-strand rectangular beams.
Transfer Length at Transfer Length after 60 Transfer Length after 90 Transfer Length after 240
Release from Strand Days from Strand End Days from Strand End Days from Strand End
Beam Number and End Slips (in.) Slips (in.) Slips (in.) Slips (in.)
Location
RA8-5-1-T (Top) 12.06 14.65 24.27 24.67 24.96 26.18 25.16 27.21
(Bottom) 3.37 13.61 11.59 21.29 11.66 25.11 12.80 27.27
RA8-5-2-T (Top) 16.37 15.13 27.30 27.30 28.20 28.68 28.96 29.44
(Bottom) 13.54 14.48 22.03 24.25 23.31 25.40 24.05 25.33
RA6-5-1-T (Top) 20.25 19.84 33.01 32.69 34.46 34.96 34.60 34.89
(Bottom) 20.00 18.82 31.92 26.29 34.06 28.36 34.06 28.56
RA6-5-2-T (Top) 19.44 16.61 37.07 35.15 39.33 37.28 40.64 37.49
(Bottom) 18.89 17.55 45.33 35.77 47.82 42.71 49.55 43.34
RA10-5-1-T (Top) 14.73 12.41 21.59 18.79 22.16 19.43 22.16 19.70
(Bottom) 17.95 11.69 19.00 14.40 19.07 15.30 19.62 15.93
RA10-5-2-T (Top) 12.37 12.10 14.29 22.15 15.42 22.22 15.63 22.36
(Bottom) 11.83 13.16 16.28 16.01 16.56 16.29 17.33 16.43
Table 3.22. Change in transfer length over time for 0.5-in. Strand D in four-strand rectangular beams.
Transfer Length at Transfer Length after 60 Transfer Length after 90 Transfer Length from 240
Release from Strand Days from Strand End Days from Strand End Days from Strand End
Beam Number and End Slips (in.) Slips (in.) Slips (in.) Slips (in.)
Location North South North South North South North South
RD8-5-1-T (Top) 19.84 17.64 35.57 35.90 40.15 39.59 41.18 42.47
(Bottom) 25.78 28.06 23.98 38.41 30.63 41.05 27.73 44.59
RD8-5-2-T (Top) 28.25 24.82 65.51 67.04 67.56 68.62 68.52 68.62
(Bottom) 19.93 22.50 49.52 32.36 50.91 33.26 52.86 35.00
RD6-5-1-T (Top) 27.71 20.86 53.89 56.65 57.32 59.03 58.79 60.29
(Bottom) 23.66 25.85 38.10 38.09 40.76 42.20 42.89 45.27
RD6-5-2-T (Top) 23.66 22.81 49.07 48.64 57.90 53.33 63.27 54.49
(Bottom) 25.10 21.82 65.45 39.67 69.56 44.05
RD10-5-1-T (Top) 17.03 15.94 26.10 24.12 27.87 26.36 30.19 27.11
(Bottom) 21.43 16.81 23.51 19.77 23.51 21.63 23.77
RD10-5-2-T (Top) 17.58 13.15 24.81 23.58 26.30 24.95 26.58 26.99
(Bottom) 16.74 15.92 24.05 23.01 25.98 23.01 28.18 23.01

51
40
35
30 From DEMEC
Transfer Lengths (in.)

2
R = 0.4215
25
20
From Strand End Slips
2
15 R = 0.6429
10
0
0 2000 4000 6000 8000 10000 12000
Concrete Strengths (psi)
Strand End Slip DEMEC Omitted Data
Figure 3.24. Transfer length versus fci′ for Strands A/B in rectangular beams.
40
35
R2 = 0.6556
30
25
From DEMEC
20
R2 = 0.3195
15
10
0
0 2000 4000 6000 8000 10000 12000
Strand End Slip DEMEC Omitted Data
Figure 3.25. Transfer length versus fci′ for Strand D in rectangular beams.
Strand D. Again, it is clear that the transfer length decreases Bond Test value and the square root of concrete strength. The
with increasing concrete strength. coefficient of determination in those comparisons is a very
Finally, Figure 3.26 illustrates the transfer length measure- robust 0.8. If the NASP Bond Test value, which is a direct
ments taken on beams made with the 0.6 in. diameter strand, measure of bond between the strand and concrete, varies with
Strand A6. Again, the data clearly show the inverse relation- the square root of concrete strength, then it is logical that
ship between transfer lengths and concrete strength. The data the transfer length would also vary with the square root of
from all three of the strand sources are illustrated in Figure concrete strength.
3.27, where the transfer lengths for each strand are plotted Figure 3.28 plots the same data as Figure 3.27, but does a
against the concrete strengths at release. best-fit curve from power regressions. The coefficients of
Figures 3.24 through 3.27 show the relation between trans- determination for these power curves are nearly as good as
fer length data and linear regression models. Linear regression the coefficients of determination for the linear regressions.
is often used because the methodology is less abstract than Furthermore, the best-fit regressions provide an exponent in
others and perhaps more easily understood. However, there is the equation of −0.56, −0.83 and −0.56. As a reminder, the
a direct relationship between the NASP Bond Test values in inverse of the square root would be an exponent of −0.50.
concrete and the square root of concrete strengths. Figures Figure 3.29 plots the transfer lengths for Strands A/B at re-
3.12 and 3.13 show a strong correlation between the NASP lease and at 240 days after release. The data are fitted to a

52
40
35

30
R2 = 0.6399
25
20
15 From DEMEC
10
R2 = 0.687
0
0 2000 4000 6000 8000 10000 12000
Strand End Slip DEMEC
Figure 3.26. Transfer length versus fci′ for Strand A6 (0.6 in) in rectangular
beams.
45
40 Strand D
R2 = 0.6556
Transfer Length (in.)
35
30 Strand A6 (0.6 in.)

R2 = 0.6399
25
20
15
Strands A/B
10
R2 = 0.6429
5
0
0 2 4 6 8 10 12 14

Strands A/B Strand D Strand A6 (0.6 in.)
Figure 3.27. Linear regression for transfer lengths and f′ci.
70
Strand D
60
R2 = 0.5815
Transfer Length (in.)
50 Strand A6 (0.6 in.)

R2 = 0.6526
40
30
20 Strand A/B
R2 = 0.6112
10
0
0 2 4 6 8 10 12 14

Strand A/B Strand D Strand A6 (0.6 in.)
Figure 3.28. Power regression for transfer lengths and f′ci.

53
60
Transfer Length at Release (in.)

50
40
y = 149.93x-0.8913
30
R2 = 0.3886
20
y = 44.79x-0.56
10 R2= 0.72
0
0 2 4 6 8 10 12 14
Concrete Strength at Release (ksi)
Figure 3.29. Transfer lengths versus concrete strengths

for 0.5-in. Strands A/B at release and at 240 days.
power regression curve. The best-fit equations are also shown for Strand D are considerably longer than those for Strands
in Figure 3.29. In Figures 3.29 through 3.32, transfer length A/B. Recall that Strand D had a NASP Bond Test value of
data obtained immediately after release are represented by 6,890 lb, whereas both Strands A and B had NASP Bond Test
diamond-shaped data points and the solid regression curve. values in excess of 20,000 lb (see Table 3.3). These data would
Transfer lengths measured at 240 days are represented by support the idea that higher NASP Bond Test values will
triangular-shaped data points and the dashed regression curve. result in shorter transfer lengths.
Figure 3.30 plots the transfer lengths for Strand D at both Figure 3.31 plots the same data but for the 0.6 in. diameter
release and at 240 days after release. Again, these data are fit- strand, Strand A6. Again, the data clearly show that transfer
ted to a power regression curve. Note that the transfer lengths lengths decrease with increasing concrete strength.
60
50
y = 432.02x-1.19
40 R2 = 0.63
30
y = 115.53x-0.83
20
R2 = 0.71
10
0
0 2 4 6 8 10 12 14
Figure 3.30. Transfer lengths versus concrete strengths for

0.5-in. Strand D at release and at 240 days.
60
50
y = 156.73x-0.75
40 R2 = 0.87
30
y = 68.78x-0.56
R2 = 0.68
20
10
0
0 2 4 6 8 10 12
Figure 3.31. Transfer lengths versus concrete strengths for

0.6-in. Strand A6 at release and at 240 days.

54
60
Transfer Lengths at Release vs. Normalized NASP Test
Transfer Lengths at 240 d
50 Design Curve for Transfer Length
Power (Transfer Lengths at Release vs. Normalized NASP Test)
Power (Transfer Lengths at 240 d)
Power (Design Curve for Transfer Length)

40
y = 103.17x-0.45
R2 = 0.49
30
20
y = 97.2x-0.5
y = 67.78x-0.46
10 R2 = 0.58
0
0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 45.0
Normalized NASP Value (kips)
Figure 3.32. Transfer lengths versus normalized NASP bond values for Strands A/B and
plotted together at release and at 240 days.
Finally, in Figure 3.32, the transfer lengths at release and at improving bond, as do the increasing NASP Bond Test val-
240 days after release are plotted against the normalized ues. The proposed design equation shown indicates that
NASP Bond Test value. The normalized NASP Bond Test transfer length can be obtained by dividing 97.2 in. by the
value is obtained from Equation 3.1. In Equation 3.1, the nor- square root of the normalized NASP Bond Test value. The
malized value can be obtained because the ratio of the NASP normalized NASP Bond Test value factors in the same factor
Bond Test Value in concrete to the standard NASP Bond Test for the square root of the concrete strength. The proposed
value (in mortar) is essentially equal to one-half of the square design equation will provide a transfer length of 60 strand
root of the concrete strength at 1 day. In this manner, data diameters for concrete strength of 4 ksi. Increasing concrete
from strands with widely dissimilar NASP Bond Test values strengths will reduce the proposed transfer length in propor-
can be plotted on the same chart and the results compared. In tion to the square root of the concrete strength.
Figure 3.32, we see that the power regression curve fits
through both sets of data. The data set shown with lower
3.5 Development Length Tests
NASP Bond Test values, toward the left side of the chart, are
the data derived from Strand D; data with higher NASP Bond Measured transfer and development lengths of prestress-
Test values, toward the right side of the chart, are obtained ing strands are indications of the quality of bond between the
from Strands A/B. The power regression curve shows a best strand and concrete. The research conducted as part of
fit with an exponent of −0.46. NCHRP Project 12-60 and described earlier resulted in five
Also plotted on Figure 3.32 is the curve that corresponds to overarching conclusions:
the proposed equation for transfer length. The normalized
NASP Bond Test value is obtained from Equation 3.1. For ex- 1. The Standard NASP Bond Test method provides a reliable
ample, for a concrete strength of 4 ksi, the transfer length and repeatable method to test for the bond performance
should be 60 strand diameters. For 0.5 in. diameter strand, of prestressing strands. Results were found to be repeat-
the transfer length would be 30 in. The data illustrated in able at different testing sites.
Figure 3.32 show that transfer lengths are shortened with 2. The Standard NASP Bond Test is able to determine ac-
increasing NASP Bond Test values. However, it is not pro- ceptable quality levels for bond of prestressing strands.
posed to shorten the transfer length equation as a function of This ability is demonstrated by the correlation of NASP
NASP Bond Test values. It is worth noting, however, that the Bond Test results with measured transfer lengths. In-
data illustrated in Figures 3.12, 3.13, and 3.32 clearly show creases in measured transfer lengths correlate directly with
that increases in concrete strength have a similar effect in decreases in bond performance, as measured by the NASP

55
Bond Test. The development length tests reported in this measured on each beam end, either directly by measuring
chapter supplement the findings from the NASP Bond concrete surface strain or indirectly by measuring strand
Tests and transfer length measurements. end slip before release.
3. The NASP Bond Test can be modified by testing strand • I-shaped beam specimens. In all, eight I-shaped beam
bond performance in concrete instead of mortar. The specimens were fabricated with target concrete release
modified NASP Bond Tests in concrete demonstrate that strengths of 6 ksi and 10 ksi. These beams were 24 ft in
increases in concrete strength result in improving bond length and designed to be tested at each end. Transfer
performance. The results develop a strong statistical cor- lengths were also measured on these beams prior to devel-
relation, and the best fit indicates that bond strength opment length testing.
improves in proportion to the square root of concrete
strength. 3.5.1.1 Terminology
4. Concrete strength influences the bond of prestressing steel
with concrete. In the NASP Bond Tests (modified), in- The testing program terminology was as follows.
creasing concrete strengths resulted in increasing bond
strength between strand and concrete. In beams where Embedment length, le. For the purposes of this research
transfer lengths were measured, increasing concrete and generally in the broader literature, the embedment length
strength correlated to shortening transfer lengths. The is the length of bond provided from the beginning of bond
measured pull-out forces from the NASP Bond Tests (usually at the end of the beam) to the critical section of the
established that bond strength improves in proportion to beam. The critical section in these tests is generally under-
the square root of concrete strength at release. stood to be the section where maximum moment occurs. In
5. The transfer length data further establish that measured this testing program, the embedment length is the distance
transfer lengths decrease in inverse proportion (approxi- from the end of the beam to the point of loading, which
mately) to the square root of concrete strength. The influ- corresponds to the point of maximum moment.
ence of concrete strength and NASP Bond Test value Development length, ld. Development length of prestress-
correlates with the inverse of strand end slip measure- ing strands is the minimum distance from the free end of the
ment, which is a direct indicator of transfer lengths. strand over which the strand should be bonded to concrete so
that the section under consideration achieves its full nominal
The development length tests are necessary to determine the capacity.
following:
Flexural bond length. The flexural bond length is meas-
• Whether the NASP Bond Test can be used as a predictor of ured from the section where the prestressed force is fully
strand bond performance in flexural applications, effective (at the end of the transfer length) to the critical
• The minimum acceptable level of bond performance as section. In these tests, the flexural bond length is equivalent
measured by the NASP Bond Test, and to the embedment length minus the transfer length. Often,
• What modifications are necessary to the LRFD develop- the flexural bond length is used in conjunction with the
ment length equation to account for variations in concrete development length, ld . In that case, the embedment length is
strength. the development length minus the transfer length.
3.5.1 Testing Program 3.5.1.2 Beam Identification System and

Section Properties
The experimental program consisted of the flexural tests
on two types of beam specimens: Each beam carries a unique identifying name. The system
for identification is described in Figure 3.14. The identifica-
• Rectangular beam specimens. In all, 43 rectangular beam tion system indicates the following beam characteristics:
specimens were fabricated with target release concrete shape (rectangular [R] or I-shaped [I]), strand source, strand
strengths varying from 4 ksi to 10 ksi. Rectangular beams size, nominal concrete strength at release, and specimen
were cast with two prestressing strands at a depth of 10 in. number in a series. The section properties and materials are
in a beam 12-in. deep. Both 0.5 in and 0.6 in diameter described in the sections under transfer length.
strands were used. The rectangular beams were 7 ft in Rectangular beams 17 ft in length were fabricated with
length and designed to be tested independently at each end two strands in each beam. Longitudinal top steel was in-
to assess the development length of embedded strands. cluded in the cross section to provide additional compres-
Prior to development length testing, transfer lengths were sion reinforcement and to ensure under-reinforced flexural

56
conditions at capacity. As shown in Figures 3.15 and 3.16, 3.5.1.3 Loading Geometry
#3 stirrups, or “ties,” were provided on 6-in. centers. The
Both rectangular and I-shaped beams were designed to be
nominal flexural capacity, Mn, of the rectangular beams var-
tested on both ends, enabling a distinct development length
ied from about 700 k-in. for the lower strength concrete
test at each beam end. The loading geometry varied from end
(nominal 4 ksi at release) to approximately 754 k-in. for the
to end so that a different embedment length was tested at each
10 ksi (release) concrete.
end. Embedment lengths varied for each test and were cho-
Four-strand beams were cast for transfer length measure-
sen depending upon results from prior tests.
ments with two strands in the bottom of the cross section and The typical loading geometries for rectangular beams with
two strands at the top of the cross section. Four-strand beams 0.5-in. strands are shown in Figure 3.33. The geometry shown
were not tested for development length and are not discussed for the south end corresponds with an embedment length of 58
in this chapter. in., which is approximately 80 percent of the computed devel-
Figure 3.18 shows the cross-section of the I-shaped beam opment length requirement. The geometry shown for the north
with the reinforcement details. Each I-shaped beam was cast end corresponds with an embedment length of 73 in., which is
with a length of 24 ft. Top flanges were reinforced longitudi- approximately equivalent to the AASHTO LRFD and ACI re-
nally with two #3 bars that ran the length of the beam. Trans- quirements for development length. Rectangular beams with
verse reinforcement in the top flange consisted of #3 bars at 0.6-in. strands required longer embedment lengths than those
9-in. centers over the beam length. Stirrups were made from shown in Figure 3.33. The two testing lengths, 73 in. and 58 in.,
#3 bars with standard 90° hooks and spaced at 7-in. centers. were established through testing programs conducted by Rose
Stirrups were arranged so that the legs alternated directions. and Russell (1997) and Logan (1997).
Horizontal reinforcement was placed in the webs of each end The typical loading geometries for the I-shaped beams are
of each I-shaped beam. shown in Figure 3.34. The geometry illustrated is typical for
Figure 3.33. Typical loading geometry for rectangular beams.

57
Figure 3.34. Typical loading geometry for I-shaped beams.
beams made with 0.5 in. strands. The test on the south end to the pretensioned concrete test beam. The loading geometry
shows an embedment length of 58 in. The north end shows was arranged so that constant bending moment is applied
development length test geometry for an embedment length between the two load points. In this picture, the beam that is
approximately equal to the LRFD and ACI requirements, 72 being tested is a rectangular beam. It is supported by a pin on
in. As with the rectangular beam series, tests on I-shaped the near end and a roller at the far end.
beams with 0.6 in. diameter strands required longer embed-
ment lengths than those shown in Figure 3.34.
3.5.1.5 Instrumentation
3.5.1.4 Test Frame The following instrumentation was used.
The test frame was designed to perform flexural tests on both Electronic data acquisition. Load, hydraulic pressure,
rectangular beams and I-beams. The photograph in Figure 3.35 beam deflection, and strand end slips were measured and
shows the test frame with a beam in position for testing. The test recorded by an electronic data acquisition system. Data were
frame has four sides that form a rectangular “frame.” Load is sampled and recorded at regular intervals without manual
applied through a hydraulic actuator (attached to the top hori- prompting. The rate of sampling was fixed at 1.0 Hz, which
zontal beam in the frame) to a spreader beam (attached to the provided smooth transition of load, displacement, and strand
bottom of the actuator). The spreader beam distributes loading end slip values. The data were stored on a laptop computer

58
and were then available for analysis. During each development

length test, data were also recorded manually in the event that
electronic data were corrupted by unforeseen circumstance.
Load. Load was measured electronically with a load cell

placed between a spherical head under the hydraulic actuator
and the spreader beam. The load cell can be seen in Figure
3.35 just above the steel loading beam. Load was applied
hydraulically, and the hydraulic pressure was also monitored
electronically by a pressure transducer. The pressure trans-
ducer also sent electronic signals to the data acquisition sys-
tem for monitoring and recording. A hydraulic pressure gage
was employed during the test for visual observations and
manual recording.
Deflection. Wire transducers with a range of 30 in. and

accuracy ±0.005 in. were used to determine the vertical de-
flection. Deflection was measured below the center of the
loading point. Two wire transducers were used to measure de-
flection, one on each side of the beam, so that any twisting of
the beam would be taken out when computing the average be-
tween the two sides. Data from the wire transducers were Figure 3.36. Wire transducers
recorded and stored electronically. In addition to the elec- (foreground) and a dial gage.
tronic data, a dial gage with a precision of one one-thousandth
of an inch was used to manually record deflection readings.
The dial gage was also used to monitor displacements when
Strand end slip. Linear voltage displacement transducers
the testing switched from load-controlled testing to displace-
(LVDTs) were used to measure strand movement relative to the
ment controls. The wire transducers and the dial gage are
concrete. The LVDTs had a stroke limit of 1.0 in. and recorded
shown in Figure 3.36.
strand end slips to one one-thousandth of an inch. Clamps were
attached to the strands, and LVDTs were mounted on these
clamps at a location providing an initial reading of approxi-
mately 0.9 in. with an error of (±0.003 in. Strand end slips were
measured and recorded for each strand on the “test” end. The
photograph in Figure 3.37 shows the LVDTs clamped to strands
to measure strand end slip relative to concrete.
At the far end of the beam, or at the end of the beam
opposite the end being tested, strand end slips were measured
by a mechanical deflection gage with an electronic readout.
The device and arrangement are shown in Figure 3.38.
Measurements with a precision of ±0.005 in. were possible
using this technique.
3.5.1.6 Testing Procedure

For each test, the instrument readings were initialized prior
to the application of external load. Load was then applied to
beams in regular load increments. Load was applied manually
by an hydraulic pump. At all load increments, values of load,
displacement, and strand end slips, as well as DEMEC read-
ings (wherever applicable) were noted and recorded manually.
Figure 3.35. Test frame with a In addition to electronic data being stored at the 1-Hz refresh
rectangular beam readied for testing rate on the data acquisition system, data were recorded man-
rupture. ually. Once cracking began, cracks were marked with perma-

59
until failure. Failure was defined by the beam’s inability to

sustain or maintain load with increasing deflections or by
abrupt failures of the concrete or strand.
Throughout the test, manual readings at every load
increment were noted along with any significant develop-
ment such as first flexural crack, first shear crack, appear-
ance of flexure-shear crack, first strand end slip, concrete
spalling, concrete crushing, and any audible developments.
Written summaries of each development length test appear
in the appendices. Detailed progress of each test was docu-
mented and is included along with significant photographs
and data plots in Appendices C through G. Also, plots of
moment versus deflection, strand end slip versus deflec-
tion, and shear versus average shear strain were plotted
from the acquired data. Shear strains were measured from
Figure 3.37. LVDTs clamped to strands to measure DEMEC target points attached to the webs of I-shaped
strand end slip relative to concrete. beams. Shear stress was determined by dividing the shear
force applied by the product of the web width and the beam
depth.
nent markers as soon as they were observed. The loads at
which the cracks first appeared were noted alongside of the
cracks. Photographs were taken at regular intervals to record 3.5.2 Experimental Results
cracking patterns. from Development Length Testing
As displacements became larger with smaller increments in
All together, 50 flexural tests were performed on rectangu-
load, the manual system of loading switched from regular
lar beams and 14 tests on I-shaped beams. All of these tests
load increments to regular displacement increments. This
were carried out at the Civil Engineering Laboratory at OSU.
was done arbitrarily by the researchers conducting the test.
Most of the beams were tested on both ends. For each beam
At each load or displacement increment, manual readings
test, the embedment length was determined on the basis of
of hydraulic pressure and beam displacements were made.
various factors, including the AASHTO development length
Additionally, manual and electronic instruments were
equation with changes to account for prior results, concrete
checked to determine whether strand end slip had occurred
strength, or strand bond strength. In this section, Tables 3.23
during the prior loading increment. Loading was continued
through 3.27 report the results from development length test-
ing. These tables report on the following parameters:
• Concrete strength at release;

• Concrete strength at 56 days;
• Average NASP Bond Test value for the strands contained
in the beams;
• Embedment length for each individual test;
• Test span;
• Failure Moment, which is the maximum applied moment
measured during the test;
• Percentage of the Failure Moment to the nominal flexural
capacity, Mn, as determined by strain compatibility. The
calculation for Mn assumes that the strands are fully devel-
oped; no reduction in flexural capacity was assumed for
embedment lengths provided that are less than the calcu-
lated development length;
Figure 3.38. Mechanical deflection gage arrangement • Maximum beam deflection;
for measuring strand end slips at the beam end • Maximum strand end slip; and
opposite to the test end. • Classification for each type of failure.

60
Table 3.23. Development length test results on rectangular beams containing Strand D.
fc fc Avg. NASP Avg. Lt Max.

Avg. Lt @ Actual Failure Deflection
@ 56 Pull-Out (56-Day or Span %Mn End Failure
Beam End Release Le Moment @ Failure
Release Days Value @ Test) (in) Slip Mode
(in) (in) (kip-in) (in)
(psi) (psi) (lb) (in) (in)
RD-4-5-1-N 4,033 7,050 6,890 32.79 38.54 73 162 804 115 3.4 0.00 Flexure
RD-4-5-1-S 4,033 7,050 6,890 31.02 42.28 58 132 759 108 1.6 0.35 Bond
RD-4-5-2-N 4,033 7050 6,890 42.19 63.05 73 162 831 119 2.7 0.40 Flexure
RD-4-5-2-S 4,033 7,050 6,890 49.71 51.81 58 132 513 73 2.5 0.57 Bond
RD-6-5-1-N 6,183 8,500 6,890 30.24 49.75 73 162 797 111 2.5 0.06 Flexure
RD-6-5-1-S 6,183 8,500 6,890 28.07 45.26 58 132 788 109 2.0 0.18 Flexure
RD-6-5-2-N 6,183 8,500 6,890 25.60 44.24 73 162 735 102 2.0 0.01 Flexure
RD-6-5-2-S 6,183 8,500 6,890 29.22 48.27 58 132 724 100 2.0 0.25 Bond
RD-6A-5-1-N 7,960 11,420 6,890 35.4 39.94 73 162 794 106 2.3 0.00 Flexure
RD-6A-5-1-S 7,960 11,420 6,890 29.1 37.16 58 132 805 108 2.5 0.08 Flexure
RD-6A-5-2-S 7,960 11,420 6,890 20.08 40.07 58 132 778 104 1.9 0.02 Flexure
RD-8-5-1-N 8,570 13,490 6,890 20.15 39.08 73 162 811 107 2.6 0.00 Flexure
RD-8-5-1-S 8,570 13,490 6,890 20.15 34.54 58 132 805 106 2.6 0.08 Flexure
RD-8-5-2-N 8,570 13,490 6,890 13.67 37.38 58 132 775 102 2.2 0.08 Flexure
RD-8-5-2-S 8,570 13,490 6,890 17.30 50.41 58 132 813 107 2.0 0.00 Flexure
RD-10-5-1-N 9,711 14,470 6,890 26 30.24 58 132 821 108 2.1 0.00 Flexure
RD-10-5-1-S 9,711 14,470 6,890 13.57 27.14 46 120 819 107 2.6 0.00 Flexure
RD-10-5-2-N 9,711 14,470 6,890 14.85 22.30 58 132 788 103 1.9 0.00 Flexure
RD-10-5-2-S 9,711 14,470 6,890 18.23 22.03 46 120 794 104 1.9 0.01 Flexure
3.5.2.1 Tabulated Beam Test Results—Rectangular Table 3.23 also reports the maximum strand end slip that
Beams occurred during testing, which corresponds to the maximum
strand end slip measured at the time the beam failed, whether
Table 3.23 reports the results from development length tests a flexural failure or a bond failure. Note that it is not uncom-
on rectangular beams made with Strand D. Strand D was the mon for strand end slips to be measured even though a beam
0.5 in. strand with the lower NASP pull-out value, 6,890 lb. fails in flexure. For example, RD-4-5-2-N failed in flexural at
Concrete strengths at release varied from a target of 4 ksi to a a load that exceeded its nominal capacity by 19 percent.
target of 10 ksi. 56-day concrete strengths ranged from 7.05 ksi Further, the beam achieved adequate ductility as demon-
to 14.47 ksi. Table 3.23 reports only three bond failures, all strated by 2.7 in. of overall deflection while sustaining capac-
occurring with lower strength concretes. Also, all of the bond ity. However, the measured strand end slip was 0.40 in. This
failures occurred at an embedment length of only 58 in., which finding is consistent with other research that has been
is approximately 80 percent of the ACI- and AASHTO- performed to date. More notably, the results in Table 3.23
prescribed development lengths. Of the three bond failures, demonstrate that the measured strand end slips decrease
two occurred at an applied moment that matched or exceeded measurably with increasing concrete strengths. At higher
Mn, the nominal flexural capacity for the beams. Table 3.23 concrete strengths, strand end slips did not occur. Overall,
also shows that at higher strengths, in general, flexural failures the results support a conclusion that higher concrete
were observed in all tests. For example, two ends of the beams strengths result in increasing bond strength and reducing the
with 14.47 ksi concrete were tested with an embedment length required development lengths. Detailed testing summaries
of only 46 in., or approximately 63 percent of ld. In these cases, on each development length test are found in Appendix C.
the development length test resulted in flexural failures with- Table 3.24 reports the results from development length
out bond slip (beams RD-10-5-1-S and RD-10-5-2-S). tests on rectangular beams made with Strands A/B. Strands A

61
Table 3.24. Development length test results on rectangular beams containing Strands A/B.

@ 56 Pull-Out (56-Day or Span End- Failure
Beam End Release Le Moment %Mn @ Failure
RA-6-5-1-N 6,183 8,500 20,950 19.2 33.70 73 162 790 110 2.1 0.00 Flexure
RA-6-5-1-S 6,183 8,500 20,950 18.2 30.03 58 132 800 111 2.1 0.00 Flexure
RA-6-5-2-N 6,183 8,500 20,950 16.5 28.00 58 120 772 107 1.5 0.00 Flexure
RA-6-5-2-S 6,183 8,500 20,950 15.01 23.50 46 120 777 108 1.5 0.00 Flexure
RA-6A-5-1-N 7960 11,420 20,950 17.74 26.54 73 162 769 103 2.4 0.00 Flexure
RA-6A-5-1-S 7,960 11,420 20,950 17.68 28.55 58 132 770 103 1.7 0.00 Flexure
RA-6A-5-2-N 7,960 11,420 20,950 24.51 31.75 58 132 788 105 1.9 0.00 Flexure
RA-6A-5-2-S 7,960 11,420 20,950 22.03 29.38 46 120 788 105 1.7 0.01 Flexure
RA-8-5-1-N 8,570 13,490 20,950 13.3 24.91 58 132 829 109 1.7 0.01 Flexure
RA-8-5-1-S 8,570 13,490 20,950 13.5 22.54 46 120 832 110 1.9 0.00 Flexure
RA-10-5-1-N 9,711 14,470 20,950 24.27 24.34 58 132 788 103 1.7 0.00 Flexure
RA-10-5-1-S 9,711 14,470 20,950 9.69 13.14 46 120 796 104 1.7 0.00 Flexure
RB-4-5-1-N 4,033 7,050 20,210 18.42 22.10 73 162 776 111 1.9 0.00 Flexure
RB-4-5-1-S 4,033 7,050 20,210 18.49 20.51 58 132 802 114 2.0 0.00 Flexure
RB-4-5-2-N 4,033 7,050 20,210 21.12 22.52 73 162 721 103 2.4 0.00 Flexure
RB-4-5-2-S 4,033 7,050 20,210 22.46 23.75 58 132 748 107 1.7 0.00 Flexure
Table 3.25. Development length test results on rectangular beams containing Strand A6 (0.6-in. diameter).

@ 56 Pull-Out (56-Day or Span End- Failure
Beam End Release Le Moment %Mn @ Failure
RA-4-6-1-N 4,033 7,050 18,290 33.42 41.82 88 192 1084 114 3.0 0.00 Flexure
RA-4-6-1-S 4,033 7,050 18,290 24.96 28.87 70 156 964 102 2.7 0.00 Flexure
RA-4-6-2-N 4,033 7,050 18,290 30.24 37.66 73 162 1011 107 2.4 0.13 Flexure
RA-4-6-2-S 4,033 7,050 18,290 29.35 33.19 58 148 921 97 3.0 0.33 Bond
RA-6-6-1-N 4,855 8,040 18,290 29.73 40.85 88 192 1012 104 2.5 0.00 Flexure
RA-6-6-2-N 4,855 8,040 18,290 31.65 52.18 73 162 1001 103 2.1 0.02 Flexure
RA-6-6-2-S 4,855 8,040 18,290 30.1 49.37 58 148 913 94 2.7 0.41 Bond
RA-6-6-3-N 4,855 8,040 18,290 25.83 44.96 88 192 1046 108 2.6 0.00 Flexure
RA-8-6-1-N 5,413 8,220 18,290 28.21 45.48 88 192 1008 103 2.4 0.00 Flexure
RA-8-6-2-N 5,413 8,220 18,290 28.2 46.35 73 162 1007 103 2.0 0.01 Flexure
RA-8-6-2-S 5,413 8,220 18290 25.7 42.37 58 132 988 ~101 2.5 0.14 Bond
RA-10-6-1-N 9,150 14,610 18,290 20.03 29.98 88 192 1084 102 2.8 0.00 Flexure
RA-10-6-2-N 9,150 14,610 18290 15.62 26.79 73 162 1070 101 2.5 0.00 Flexure
RA-10-6-2-S 9,150 14,610 18,290 21.78 30.70 58 148 1083 102 2.4 0.00 Flexure

62
Table 3.26. Development length test results on I-shaped beams containing 0.5 in. diameter strands.
Measured fc fc Avg. NASP Max.

Maximum Maximum
Overall @ 56 Pull-Out Le Span End- Failure
Beam End Moment %Mn Deflection
Depth (h) Release Days Value (in) (in) Slip Mode
(kip-in) (in)
(in) (psi) (psi) (lb) (in)
IB-6-5-1-N 24 5,810 9,350 20,210 58 166 3,526 82 1.1 0.04 Shear

IB-6-5-1-S 24 5,810 9,350 20,210 72 222 3,980 98 3.1 0.03 Flexure
IB-10-5-1-N 24 7,615 13,490 20,210 54 168 4,282 102 2.0 0.03 Flexure
IB-10-5-1-S 24 7,615 13,490 20,210 58 180 4,196 100 1.6 0.02 Flexure
ID-6-5-1-N 24 5,492 9,840 6,890 72 222 3,538 82 2.5 0.80 Bond
ID-6-5-1-S 24 5,492 9,840 6,890 88 270 3,280 81 3.5 0.75 Bond
ID-10-5-1-N 24 8,225 14,160 6,890 88 270 4,026 92 5.2 0.08 Flexure
ID-10-5-1-S 24 8,225 14,160 6,890 72 222 4,039 92 3.7 0.75 Bond
and B were used interchangeably in this beam series as the ment length test are found in Appendix D, for Rectangular
two strand samples tested with approximately the same NASP Beams Made with Strands A and B.
Bond Test value. Concrete strengths at release varied from a Both Tables 3.23 and 3.24 report results on beams made
target of 4 ksi to a target of 10 ksi. 56-day concrete strengths with air-entrained concrete. The development length test
ranged from 7.05 ksi to 14.47 ksi. Table 3.24 reports no bond results on the air-entrained beams closely match from beams
failures. These results demonstrate that the NASP Bond Test made with 6-ksi concrete without air entrainment. In other
is a good predictor of the ability of strands to perform in pre- words, all of the ends tested with air-entrained concrete failed
tensioned applications. At concrete strengths above 4 ksi, em- in flexure, with strand end slip in only a few cases. These
bedment lengths as short as 46 in. were tested. All of these results mirrored the results of the development length tests
tests also resulted in flexural failures without any strand end without air entrainment.
slip. All of the flexural failures occurred at an applied moment Table 3.25 reports the results from development length
that matched or exceeded Mn, the nominal flexural capacity tests on rectangular beams made with 0.6 in. diameter strand.
for the beams. Detailed testing summaries on each develop- The strand is called Strand A6. Strand A6 had an NASP Bond
Table 3.27. Development length test results on I-shaped beams containing 0.6 in. diameter Strand A6.
Measured fc fc Avg. NASP Max.

Maximum Maximum
Overall @ 56 Pull-Out Le Span End-
Beam End Moment %Mn Deflection Failure Mode
Depth (h) Release Days Value (in) (in) Slip
(kip-in) (in)
(in) (psi) (psi) (lb) (in)
IA-6-6-1-N 24.125 4,381 8,990 18,290 75 156 3,267 81 1.7 0.05 Shear @ opposite end
IA-6-6-1-S 24.125 4,381 8,990 18,290 91 188 4,387 109 2.8 0.12 Flexure
IA-6-6-2-N 24.125 4,381 8,990 18,290 88 270 4,125 102 3.2 0.13 Shear
Flexure w/ Strand
IA-10-6-1-S 24.25 10,480 14,990 18,290 72 222 4,620 112 2.5 0.03
Rupture
IA-10-6-2-S 24.375 10,590 14,930 18,290 88 270 4,559 111 5.7 0.00 Flexure

63
Test value of 18,290 lb, which is interesting as it falls between Strand D. Table 3.26 reports three bond failures out of
the higher and lower NASP Bond Test values for the 0.5 in. four tests on I-shaped beams made with the lower bond per-
strands tested. Concrete strengths at release varied from a tar- former, Strand D. The fourth flexural test resulted in a
get of 4 ksi to a target of 10 ksi. The range for 56-day concrete flexural failure; this beam was made with the higher strength
strengths was 7.05 ksi to 14.61 ksi. concrete. Bond failures occurred at both the lower concrete
For 0.6 in. strands, the ACI and AASHTO development strength and the higher concrete strength. Unlike the rectan-
length provision would require a development length of ap- gular beams, bond failures of Strand D occurred at lengths
proximately 88 in. From Table 3.25, it can be seen that several equal to and exceeding the ACI and AASHTO development
of the tests were performed at an embedment length of 88 in., length design equation. At the lower concrete strength, 9.48
which roughly corresponds to 100 percent of the AASHTO re- ksi, bond failures occurred at embedment lengths of 72 and
quired development length. At the embedment length equal 88 in. At the higher concrete strength, 14.16 ksi, one bond
to the required development length of 88 in., all of the beam failure occurred at an embedment length of 72 in. The flex-
specimens failed in flexure, regardless of concrete strength. ural failure had an embedment length of 88 in. These results
This would indicate that the strand performance was adequate support two primary conclusions:
and suitable for making pretensioned concrete beams.
Other tests on beams made with Strand A6 were conducted 1. The strand with an NASP Bond Test value of 6,890 lb is
at an embedment length of 72 in., which roughly corresponds inadequate to develop the tension necessary to support
to 80 percent of ld. This was done intentionally to mirror the flexural failures as intended, and
80 percent of ld that was tested for 0.5 in. strands. Addition- 2. Higher concrete strength can improve the bond between
ally, note that some tests were conducted at an embedment prestressing steel and concrete.
length of 58 in., which is about 55 percent of ld.
Three bond failures occurred in the tests on rectangular Strand B. Table 3.26 reports results of four tests done on
beams made with Strand A6. Notably, all three bond failures beams made with Strand B. In the four tests, none of the
occurred at embedment lengths of 58 in., which is consider- beams failed in bond. The highest strand end slip measured
ably shorter than the required development length. The three was 0.04 in. Of the four failures, one was a shear failure and
bond failures occurred in beams made with the three lower the other three were flexural failures. Three of the four tests
concrete strengths, with nominal release strengths of 4 ksi, were conducted with embedment lengths of 52, 54, and 58
6 ksi, and 8 ksi. In contrast, the fourth beam, made from con- in., lengths which are significantly less than the development
crete with a nominal release strength of 10 ksi, failed in flex- length prescribed by ACI and AASHTO. These results sup-
ure when tested at an embedment length of 58 in. The results port one of the primary conclusions, i.e., that strand with a
of these tests would support the conclusion that increasing high NASP Bond Test value, in this case 20,210 lb, will
concrete strength improves the bond performance of pre- provide bond that exceeds the implicit requirement of the
stressing strands. Detailed testing summaries on each devel- development length design equations.
opment length test are found in Appendix N, for Rectangu-
lar Beams Made with 0.6 in. Strands A, or Strand A6. Strand A6. Table 3.27 reports the results from develop-
ment length tests on I-shaped beams made with 0.6 in.
strands. Strand A6 was the only 0.6 in. strand cast in beams.
3.5.2.2 Tabulated Beam Test Results—I-Shaped
It has an NASP pull-out value of 18,290 lb. Four beams were
Beams
made, two with a target release strength of 6 ksi and two with
Table 3.26 reports the results from development length a target release strength of 10 ksi. These casts achieved the tar-
tests on I-shaped beams made with 0.5 in. strands. Strand D get release strength of 10 ksi, and 1-day strengths measured
was the 0.5 in. strand with the lower NASP pull-out value, 10,590 lb. The range for 56-day strengths was 8,990 and
6,890 lb, and Strand B possessed the higher NASP Test value 14,910 lb. Of the seven beam ends tested, three ends failed in
of 20,210 lb. Two different concrete strengths were employed, shear at the end opposite the “test” end. The larger diameter
concrete with a target release strength of 6 ksi and concrete strands required longer testing spans, and the beams were not
with a target release strength of 10 ksi. The beams were made able to overcome the damage sustained during tests on the
in pairs, and the release strength of 10 ksi was not achieved. south end when tests were performed on the north end.
56-day concrete strengths ranged from 9.35 ksi to 14.16 ksi, Of the four tests that would qualify as development length
which is very near the target design strengths of 10 and 15 ksi. tests, one resulted in shear failure whereas the other three
Detailed testing summaries on each development length test tests resulted in a flexural failure. None of the failures resulted
are found in Appendix F, for I-Shaped Beams Made with 0.5 from bond failure. At the lower strength, some strand end
in. strands, including both Strand B and Strand D. slips were measured and observed; however, these strand end

64
slips were consistent with behavior that was noted in previ- Flexural failures of rectangular beams. A typical flexural
ous testing and did not prevent the strands from develop- failure is observed from the test on the south end of Beam RB-
ment tension adequate to support flexural failures at, or 4-5-1. The rectangular beam contained two 0.5 in. strands,
exceeding, the nominal flexural capacity. Detailed testing with a Strand B designation and a 56-day concrete strength of
summaries on each development length test are found in 7.05 ksi. The embedment length for this test was 58 in.,
Appendix G, for I-Shaped Beams Made with 0.6 in. strand, or about 80 percent of the AASHTO design requirement for
Strand A6. 0.5 in. strands. Strand B had a relatively high NASP Bond Test
value of 20,120 lb.
The moment versus deflection curve is found in Figure
3.5.3 Discussion of Development Length
3.39. Note that the beam achieves its nominal flexural ca-
Test Results
pacity, Mn, and that it also displays the ability to sustain the
Development length tests must be conducted to failure, moment under large deflections. Additionally, for this
and the type of failure observed determines whether the em- beam, strand end slips remained small or the strand did not
bedment length provided was adequate to ensure proper slip at all. The beam failed in flexure as the concrete in the
strand development. Three distinct types of beam failures compression zone crushed. A photograph of the beam at
were observed in the conduct of the development length tests: failure is shown in Figure 3.40.
(1) flexural failure, (2) bond failure, or (3) shear failure.
Flexural failures of I-shaped beams. The test on the south
end of I-shaped beam IA-10-6-1 provides a good example of
3.5.3.1 Types of Failure—Flexure
a flexural failure. In this test, one of the strands ruptured in
Flexural failures are characterized by two primary criteria: tension, an obvious indicator that the strand was able to fully
develop the tension necessary to resist the flexural capacity.
1. The beam is able to resist a flexural moment that ap- The embedment length for this test was 72 in., which is
proaches and often exceeds the nominal flexural capacity approximately 80 percent of the ACI and AASHTO required
(strength), and development length for 0.6 in. strands. The NASP Bond Test
2. The beam is able to undergo large deformations while sus- value for Strand A6 was 18,290 lb.
taining its capacity for resistance (ductility). The moment versus deflection curve is found in Figure
3.41. Note that the beam achieves its nominal flexural capac-
Flexural failures of the beam specimens were typically ity, Mn, and that it also displays the ability to sustain the load
characterized by the crushing of concrete at the top of the under large deflections. The beam failed at a moment of 4,620
cross section where the compression zone exists. The beams kip-in., which exceeded the calculated Mn by about 12 per-
were designed to be under-reinforced, which ensures that the cent. In this beam, the strands slipped a small amount as
strands themselves will experience large strains at flexural loads increased to capacity; the maximum strand end slip
failure. Even so, crushing of the concrete is the most common measured was 0.03 in. This small amount of strand end slip is
failure mode. In one or two specimens of this test series, also consistent with many of the flexural failures that occur
strands ruptured in tension. It should be noted that some during development length testing.
strand end slip can be observed even during a flexural failure. The beam failed when one of the strands ruptured in ten-
The strands consistently exhibit an ability to develop strand sion. Strand rupture was accompanied by a loud noise. The
tension even with small amounts of slip. However, when cracking pattern at failure is shown in the photograph shown
larger amounts of slip are observed, often the result is a bond in Figure 3.42. The cracking pattern is typical for I-shaped
failure. When strand end slips are observed, the determina- beams. There are two distinct regions of cracking. Flexural
tion of whether the failure is a flexural failure with adequate cracking is predominant in the regions of maximum mo-
strand bond or a bond failure is based on whether the beam ment. These cracks are distinguished by a vertical propaga-
meets the two criteria listed above for a flexural failure. tion near the bottom fibers of the beam. Web shear cracking
Beams that failed in flexure also showed considerable duc- occurs in the webs within the shear span of the tested end.
tility, with deflection increasing dramatically with sustained These cracks are distinguished by their diagonal nature. It was
loads or with some incremental load increases. In some fl- uniformly observed that web crack propagation was limited
exural failures, strand fractures occurred. Typically, strand to the webs of the I-shaped beams until loads approaching
fractures occurred in beams made with higher strength flexural capacity were applied. As loading increased, the web
concrete. In these cases, failures did not cause crushing of cracks would propagate into the bottom “bulb” of the
concrete at the top surface; rather, the applied moments were I-shaped beam. Additionally, the photograph in Figure 3.42
large enough to cause the strands to rupture in tension. shows inclined flexural cracks that propagate vertically from

65
1200 1.50
1.35
1000
1.20
1.05
800
Mn = 705.3 kip-in
Moment (kip-in)
0.90
End-slip (in.)
600 0.75
Moment vs Deflection 0.60
400
0.45
0.30
200
0.15
Strand End Slip vs Deflection
0 0.00
0 0.5 1 1.5 2 2.5 3 3.5 4
Deflection (in.)
Figure 3.39. Moment versus deflection and strand end slip for Beam RB-4-5-1-S.
the bottom of the beam, but then incline as the crack Oftentimes, although not always, bond failures can be
approaches and enters the webs. abrupt and occur without warning. However, it is generally
noted that test beams failing in bond demonstrate some
measure of gradual failure; that is, they possess an ability to
3.5.3.2 Types of Failure—Bond
sustain some load through large deformations. However,
Failures of pretensioned bond are characterized by the bond failures nearly always occur at loads less than the calcu-
following two primary markers: (1) an inability to develop lated nominal flexural capacity, Mn.
resistance to meet its design capacity and (2) excessive strand
end slip. Bond failures in rectangular beams. A typical bond failure
is observed from the test on the south end of Beam RD-4-5-2.
The rectangular beam contained two 0.5 in. strands. The con-
crete strength at 56 days was 7.05 ksi. The embedment length
for this test was 58 in., or about 80 percent of the AASHTO
design requirement for 0.5 in. strands. The beam contained
strands from the sample Strand D, which possessed a rela-
tively low NASP Bond Test value of 6,890 lb.
The moment versus deflection curve is shown in Figure
3.43. The moment versus deflection curve illustrates that the
beam was unable to reach its nominal flexural capacity, Mn.
Mn for this beam was 705 kip-in., and the beam’s actual ca-
pacity was 513 kip-in., as measured during the test. In re-
viewing the load versus deflection curve and the strand end
slip curve, it is apparent that the strand started slipping very
soon after flexural cracking first occurred. The beam was
RB-4-5-1-S unable to resist loads that were much larger than the cracking
moment, and strand end slips continued to increase with
Figure 3.40. Concrete crushing in the compression additional beam deflections. At a total deflection of about
zone of Beam RB-4-5-1-S. 3 in., the compression block at the top of the beam exhibited

66
5000 1
4500 0.9
Mn = 4110 kip-in
4000 0.8
3500 0.7
Strand End Slip (in.)

Moment vs Deflection
Moment (kip-in)
3000 0.6
2500 0.5
2000 0.4
1500 0.3
1000 0.2
500 Strand End Slip vs Deflection 0.1
0 0
0 0.5 1 1.5 2 2.5 3 3.5 4
Deflection (in.)
Figure 3.41. Load versus deflection and strand end slip for IA-10-6-1 South.
crushing failure. The cracking pattern and the crushing fail- four ends tested, bond failures occurred on the beams where
ure of the beam can be viewed in Figure 3.44. Note the one the embedment length was only 58 in. The companion beam
wide flexural crack, which is often a characteristic of bond to Beam RD-4-5-2 (south end), described above, was Beam
failures. Because the beam was unable to achieve its nominal RD-4-5-1 (south end). It also failed in bond but at a load
flexural capacity and because the beam exhibited excessive equal to the nominal flexural capacity. Still, the beam exhib-
strand end slips, this test was classified as a bond failure. ited excessive strand end slip during the test, and the failure
It should be noted that two rectangular beams were con- was not particularly ductile in that the beam was unable to
structed with Strand D and a targeted release strength of 4 ksi. sustain its resistance through large deformations. A descrip-
These beams are the RD-4-5-1 and RD-4-5-2 beams. Of the tion of that test and all other development length tests can be
found in the appendices to this report.
Bond failures in I-shaped beams. In development length

tests on I-shaped beams made with 0.5 in. strands, three bond
failures occurred. All of the bond failures occurred in beams
made with Strand D, the strand with the lower NASP Bond
Test value of 6,890 lb. Of the three tests that failed in bond, two
ends failed at embedment lengths of 72 in. and 88 in. These
were two ends of the same beam that had a release strength of
5,490 psi and a 56-day strength of 9,840 psi. On the higher
strength beam, with a 56-day concrete strength of 14.16 ksi, a
bond failure occurred at an embedment length of 72 in., and
a flexural failure occurred at an embedment length of 88 in.
These tests demonstrated that Strand D, with an NASP Bond
Test value of 6,890 lb, was inadequate in its ability to bond
with concrete and satisfy the design requirements implied in
the ACI and AASHTO expressions for development length.
Figure 3.42. Cracking pattern for Beam Test IA-10-6-1 Beam ID-6-5-1 (south end) shows a typical bond failure.
South, at strand. This I-shaped beam contained five 0.5 in. strands; the con-

67
1200 1.50
1.35
1000
1.20
1.05
800

Mn = 705.3 kip-in
Moment (kip-in)
0.90
600 0.75
0.60
400
0.45
0.30
200
Strand End Slip vs Deflection 0.15
0 0.00
0 0.5 1 1.5 2 2.5 3 3.5 4
Deflection (in.)
Figure 3.43. Applied moment versus deflection and strand end slip for Beam RD-4-5-2-South.
crete strength at 56 days was 9.84 ksi. The embedment length kip-in. was only about 81 percent of its calculated nominal
for this test was 88 in., or about 120 percent of the AASHTO flexural capacity. In reviewing the results from the test, it is
design requirement for 0.5 in. strands. The beam contained apparent that the incidence of web shear cracking coincided
strands from the sample Strand D, which possessed a rela- with the initial strand end slips. Strand end slips continued to
tively low NASP Bond Test value of 6,890 lb. increase with increased beam loadings and increased beam
The moment versus deflection curve is found in Figure deflections. The test was concluded at a total deflection of
3.45. The moment versus deflection curve illustrates that the about 3.5 in., when it was apparent that deflections were in-
beam was unable to reach its nominal flexural capacity, Mn. creasing without further increase in beam capacity. The
The results indicate that the beam’s flexural capacity of 3,280 cracking pattern and the crushing failure of the beam can be
viewed in Figure 3.46. The photograph shows one flexural
crack under the loading point that became very wide under
load. The excessive width of the crack is further evidence of
bond failure in the prestressing strand. Because the beam was
unable to achieve its nominal flexural capacity and because
the beam exhibited excessive strand end slips, this test was
classified as a bond failure.
3.5.3.3 Types of Failure—Shear Failure

Two shear failures occurred in I-shaped beams; no shear
failures occurred in the rectangular beams. Prior research has
shown that significant interaction can exist between shear
and bond behaviors, especially in I-shaped beams with nar-
row webs (Kaufman and Ramirez 1988). In these beams,
shear behavior is improved considerably by the inclusion of
Figure 3.44. Cracking pattern of bond failure for horizontal mild reinforcement within the webs and extend-
Beam RD-4-5-2 (South). ing for the first 96 in. from each end of the I-shaped beam.

68
5000 1
4500 0.9
Mn = 4051.6 kip-in
4000 0.8
3500 0.7

Moment (kip-in)
3000 0.6
2500 0.5
2000 0.4
1500 0.3
1000 0.2

500 0.1
0 0
0 0.5 1 1.5 2 2.5 3 3.5 4
Deflection (in.)
Figure 3.45. Applied moment versus deflection and strand end slip for Beam ID-6-5-1 (South).
An example shear failure is observed from the test on Beam moment versus deflection curve follows a pattern indicative
IA-6-6-2 (north end). This I-shaped beam contained four 0.6 of a flexural failure. The curve also shows that the beam was
in. strands; the concrete strength at 56 days was 8.99 ksi. The unloaded and then reloaded a second time. Web shear
embedment length for this test was 88 in., or approximately cracking and flexural cracking occurred at the same load
equal to the AASHTO design requirement development increment, corresponding to a moment of about 2,400 kip-
length for 0.6 in. strands. The beam contained strands from in. Strand end slips did not occur with the initial web crack,
the sample Strand A6, which possessed a NASP Bond Test but soon followed.
value of 18,290 lb. Also, this beam was dropped and damaged One of the interesting things about this test is that the shear
during handling at the prestressing plant. Several cracks re- failure occurred as the beam had reached its nominal flexural
sulted from the dropping of the beam. capacity. The large deformations also suggest that strand
The moment versus deflection curve and the strand end yielding was probably occurring, and, as the test on the beam
slip versus deflection curve are shown in Figure 3.47. The was being conducted, a flexural failure was indicated. How-
ever, as one can view in the photograph shown in Figure 3.48,
the beam failed suddenly and violently with a diagonal com-
pression failure of the web. The shear failure shows that even
though the beam is failing in shear, the strand possesses bond
adequate to develop the beam’s capacity.
3.5.3.4 Summary of Development Length Tests

There are three key issues:
1. Whether the NASP Bond Test can be used as a predictor

of strand bond performance in flexural applications,
2. What the minimum acceptable level of bond performance
is as measured by the NASP Bond Test, and
3. What modifications are necessary to the LRFD develop-
Figure 3.46. Cracking patterns at the maximum load ment length equation to account for variations in concrete
(failure) of Beam ID-6-5-1 (South). strength.

69
5000 1
4500 0.9
Mn = 4040 k-in
4000 0.8
3500 0.7
Shear Failure

Moment (kip-in)
3000 0.6
2500 0.5
2000 Moment vs Deflection 0.4
1500 0.3
1000 0.2

500 0.1
0 0
0 0.5 1 1.5 2 2.5 3 3.5 4
Deflection (in.)
Figure 3.47. Applied moment versus deflection and strand end slip for Beam Test IA-6-6-2
(North).
The NASP Bond Test as a predictor of strand bond per- bond failures at shorter embedment lengths. In I-shaped
formance in flexural applications. Two 0.5 in. strands were beams made with Strand D, the strand failed in bond even at
tested in beams. Strand D had an NASP Bond Test value of lengths in excess of the ACI and AASHTO design require-
6,890 lb and Strands A and B had an NASP Bond Test value ments for development length. In other words, Strands A and
exceeding 20,000 lb. In the rectangular beams made with B, which have relatively high NASP Bond Test values, demon-
Strands A or B, no bond failures were experienced, even at rel- strated excellent bond characteristics. In contrast, Strand D,
atively short embedment lengths. In I-shaped beams made with a relatively low NASP Bond Test value, demonstrated
with Strand B, no bond failures were experienced, even at em- poor bond characteristics. The results clearly show that the
bedment lengths shorter than the AASHTO design require- NASP Bond Test can distinguish between strands with good
ment for development length. In contrast, both rectangular bonding behavior and strands with poor bonding behavior.
beams and I-shaped beams made with Strand D experienced
The minimum acceptable level of bond performance as
measured by the NASP Bond Test. To determine a mini-
mum level of bond performance as measured by the NASP
Bond Test, results from the testing program conducted in the
NASP Round III testing program are required. However, the
results from the testing described in this chapter clearly indi-
cate that the minimum value for the NASP Bond Test should
be greater than the value measured on Strand D, 6,890 lb, but
need not be as strong as the bond value measured on Strands
A and B, which exceeded 20,000 lb.
Modifications necessary to the LRFD development

length equation to account for variations in concrete
strength. The results clearly show that increases in concrete
strength bring about improvements in strand development.
Strand D, which failed in bond at lower concrete strengths,
was still able to fully develop adequate tension at the higher
Figure 3.48. IA-6-6-2 (North) at shear failure. concrete strengths.

70
3.5.4 Discussion of Test Results The variables for development length tests in this research
were embedment length, concrete strength, and the type of
This section includes analysis in three primary areas: strand. These parameters were changed for flexural tests on
both rectangular and I-shaped beam specimens.
1. What influence does concrete strength have on the devel- The current ACI/AASHTO equation does not include the
opment length for pretensioned prestressing strands? concrete strength parameter for calculating transfer and de-
2. What is the proper expression for development length? velopment length. However, results obtained during the flex-
3. What should be the minimum NASP Bond Test Value of ural tests strongly suggest that the anchorage ability of the
the prestressing strand for achieving adequate anchorage? strands is improved as concrete strength increases. The next
section reports on the effects of increasing concrete strength
The NASP Bond Tests in concrete clearly demonstrate that on the results obtained during the flexural tests.
concrete strength can exert great influence over the bond of
strand with concrete. This trend was also demonstrated in
measured transfer lengths as the transfer length for a given 3.5.4.2 Direct Tabular Method
strand was shortened as concrete strength increased. In this
Table 3.28 summarizes the results from development
section, the results from development length tests are ana-
length tests performed on Strand D cast in rectangular beams.
lyzed to determine the influence of concrete strength. Based
In the Tables 3.28 through Table 3.30, the letter “F”’ denotes
on the analysis, certain modifications to the current
a flexural failure, and the letter “B” denotes a bond failure. In
AASHTO equation for development length are recom-
Table 3.28, the results indicate that for embedment lengths of
mended. Comparisons among flexural test results are used to
73 in. and concrete release strengths of about 4 ksi (56-day
assess the validity of such recommendations.
strength of 7 ksi), Strand D was able to develop the necessary
tension to achieve a flexural failure in the beam. However, at
3.5.4.1 Evaluating Development Length an embedment length of 58 in. and tested at the opposite ends
from the Flexural Tests of the same beams, Strand D failed in bond.
The embedment length of 73 in. corresponds to 100 per-
The development length is the length for which the strand cent of the development length prescribed in the AASHTO
must be fully bonded to ensure strand anchorage adequate to
develop the tension stress necessary to support the nominal Table 3.28. Development length tests on rectangular
flexural capacity of the cross section. The development length beams with 0.5-in. Strand D (average NASP pull-out
is distinguished from the embedment length, which is the value = 6,870 lb).
length of bond that is actually provided. In the course of test-
ing, a specific embedment length may be longer or shorter fc fc
than the strand’s development length. If a beam test results in @ 56
Beam No. Release Days Embedment Length (in)
a bond failure, then one must conclude that the embedment (psi) (psi)
length provided was shorter than the required development 46 58 73
length. Conversely, if a beam test results in a flexural failure, RD-4-5-1 4,033 7,050 B F
then one can conclude that the embedment length provided RD-4-5-2 4,033 7,050 B F
was longer than the required development length. Each inde-
RD-6-5-1 6,183 8,500 F F
pendent beam test therefore becomes a single data point that
can indicate whether the embedment was sufficient. In most RD-6-5-2 6,183 8,500 B F
cases, it is difficult to discern from a single test what the “true” RD-6A-5-1 7,960 11,420 F F
development length must be.
RD-6A-5-2 7,960 11,420 F F
Ideally, the “true” value of development would be when the
flexural test results in simultaneous flexural, shear, and bond RD-8-5-1 8,570 13,490 F F
failures (Meyer 2002). Research that varies the embedment RD-8-5-2 8,570 13,490 F, F* -
length between the values corresponding to complete flexural
RD-10-5-1 9,711 14,470 F F -
failure and the values corresponding to complete bond failure
can get closer to identifying the “true” development length. RD-10-5-2 9,711 14,470 F F -
Based on prior test results, the embedment length can be F = Flexural failures
systematically lengthened or shortened for the purpose of brack- B = Bond failures
eting the test results. In this manner, an accurate picture for de- * Both ends were tested at an embedment length of 58 in. Both ends failed in
flexure.
velopment length may be obtained through multiple beam tests.

71
Table 3.29. Development length tests on rectangular The tests demonstrate that, had the concrete strength been
beams with 0.5-in.Strands A/B (average NASP 7 ksi, the development length required for Stand D would be
pull-out value for A = 20,210 lb and for B = 20,950 lb). less than 73 in. but greater than 58 in. The dark line in the
table separates the zone of bond failures from the zone of flex-
fc fc
Beam No.
ural failures. The test results clearly show that the strand bond
RLS 56 Days Embedment Length (in)
(psi) (psi) improves in development length applications with increases
46 58 73
in concrete strength.
RB-4-5-1 4,033 7,050 F F Table 3.29 shows the results from development length
RB-4-5-2 4,033 7,050 F F tests performed on beams made with Strands A/B. The re-
RA-6-5-1 6,183 8,500 F F sults show that (1) Strands A/B bonded better with concrete
RA-6-5-2 6,183 8,500 F F than Strand D, and (2) the bond of Strands A/B improved
RA-6A-5-1 7,960 11,420 F F
as concrete strength increased. The dark line in the table
separates the zone of bond failures from the zone of flexural
RA-6A-5-2 7,960 11,420 F F
failures.
RA-8-5-1 8,570 13,490 F F
Table 3.30 summarizes the results of beam tests on rectan-
RA-10-5-1 9,711 14,470 F F
gular beams made with 0.6 in. strands. The current AASHTO
F = Flexural failures
B = Bond failures
expression gives a development length requirement equal to
88 in. for 0.6 inch diameter strands. Test results show that
flexural failures occurred at lengths of 88 in. and 73 in. for all
LRFD Bridge Design Specifications, while the embedment concrete strengths. The results also show that bond failures
length of 58 in. corresponds to 80 percent of the code- occurred for the three concrete strengths when an embed-
specified value. Important to the purposes of this research, ment length of 58 in. was tested. However, when Strand A6
the bond of Strand D demonstrates marked improvement as was cast in concrete with a release strength of 10 ksi and a
concrete strengths increase. At a concrete strength of 11 ksi, 56-day strength of over 14 ksi, the strand was able to develop
Strand D was able to develop the necessary tension at the required tension force at an embedment length of 58 in.
embedment lengths of either 58 in. or 73 in. The test results The dark line in the table separates the zone of bond failures
indicate that for Strand D, cast in 11 ksi concrete, the devel- from the zone of flexural failures. These results show clear
opment length required is equal to or less than 58 in. Further, improvements in strand bond behavior with increasing con-
in Beams RD-10-5-1 and RD-10-5-2, Strand D was able to crete strength.
develop its tensile force in only 46 in. of bonded length. These The current ACI/AASHTO equation does not include the
tests indicate that for Strand D cast in 14 ksi concrete, the concrete strength parameter for calculating transfer and
development length required is equal to or less than 46 in. development length. However, results obtained during the
Table 3-30. Development length tests on rectangular

beams with 0.6-in. Strand A6 (average NASP pull-out
value = 18,920).
fc fc
Beam End RLS 56 Days Embedment Length (in)
(psi) (psi)
58 70 73 88
RA-4-6-1 4,033 7050 F F
RA-4-6-2 4,033 7,050 B F
RA-6-6-1 4,855 8,040 F
RA-6-6-2 4,855 8,040 B F
RA-6-6-3 4,855 8,040 F
RA-8-6-1 5,413 8,220 F
RA-8-6-2 5,413 8,220 B F
RA-10-6-1 9,150 14,610 F

RA-10-6-2 9,150 14,610 F F
F = Flexural failures
B = Bond failures

72
flexural tests demonstrate that the anchorage ability of the following relationship between NASP values in concrete and
strands is improved as concrete strength increases. NASP values in mortar (standard NASP values):
( NASPconcrete )
3.6 Discussion of Design = 0.49139 f ci′0.51702 (3.3)
NASP
Recommendations
The equation was further modified to fit the NASP values
The current AASHTO code provisions do not include as a function of square root of concrete strengths. Figure 3.13
the effects of concrete strength when calculating the is a plot of normalized NASP values against the square roots
required development length of prestressing strands. As a of corresponding concrete strengths. Following is the rela-
result, the development length for strands is the same tionship shown in Figure 3.13:
regardless of concrete strength. However, the results of this
research clearly demonstrate that the required transfer ( NASPconcrete )
= 0.51 fci′ (3.4)
and development lengths are shortened as concrete NASP
strength increases.
With the help of this relationship, it was possible to use the
The approach develops from the findings of the research:
Standardized NASP Bond Test, conducted in mortar, to
estimate the bond strength as if the test were conducted in con-
1. The current AASHTO transfer length of 60db is adequate
crete with various strengths. The graphs in Figures 3.9 through
to predict the transfer length of prestressing strands in
3.13 demonstrate that the NASP Bond Test pull-out value in
“normal strength concrete” (4-ksi release strength).
concrete is inversely proportional to the square root of the con-
2. The data support modification of the AASHTO transfer
crete strength. From these data, one can further assert that the
length to account for variations in concrete release strength
average bond stress, taken as the pull-out force divided by the
and in recognition of the finding that bond strength
bonded length, is also inversely proportional to the concrete
improves in proportion to the square root of the concrete
strength at release. Further evidence for this same relationship
strength.
between bond strength and pull-out force is found in Figures
3. The current AASHTO development length equation can
3.24 through 3.32, which chart measured transfer lengths ver-
be used to adequately predict required development
sus concrete strengths. The transfer length data demonstrate
lengths for “normal strength concrete” with a release
that transfer lengths change inversely with concrete release
strength in the range of 4 ksi and a design strength of 6 ksi.
strength. Figure 3.32, which charts transfer length measured on
4. The data demonstrate that shorter development lengths
three different strand samples, shows that transfer lengths are
are required as concrete strength increases.
approximately inversely proportionate to the square root of the
concrete strength. The best fit power regression indicates an ex-
3.6.1 Discussion of Transfer Length ponent of −0.46 for measured concrete strengths at release.
Recommendations This is approximately equal to the inverse of the square root. It
can therefore be concluded that transfer length is inversely pro-
The standard NASP Bond Test is a test where a prestress-
portional to the square root of concrete strength. Therefore, a
ing strand is pulled from sand-cement mortar. The mortar is
transfer length expression is recommended that is equivalent
made from sand, cement, and water and possesses a 1-day
to the current design expression of 60 strand diameters at a
compressive strength of 4,500 to 5,000 psi. The NASP Bond
release strength of 4 ksi, but that shortens in proportion to the
Test can be modified to perform the test in concretes with
square root of the concrete strength at release. The recom-
varying concrete strengths. However, the NASP Bond Test
mended code provision also provides a minimum transfer
values used in the discussions regarding minimum Bond Val-
length of 40 db. The 40 db value corresponds to 10-ksi concrete,
ues are pull-out strengths obtained from the standardized
which was the highest 1-day strength tested.
NASP Bond Test performed in mortar.
The transfer length equation is modified by the square root
The results from NASP pull-out tests in concrete are pre-
of the concrete release strength, as follows:
sented and compared in this section. Figure 3.12 presents
normalized NASP values (obtained by dividing the NASP 120db
lt = (3.5)
pull-out values in concrete by the NASP standardized test val- fci′
ues [from tests conducted in mortar]) versus the concrete
strengths for the NASP tests in concrete. The tests demon- where
strate remarkable correlation between the bond-ability of lt = transfer length (in.),
prestressing strand and the concrete strength. Compared f c′i = release concrete strength (ksi), and
with a power regression, the chart in Figure 3.12 shows the db = diameter or prestressing strand (in.).

73
Using concrete release strength of 4 ksi, this equation re- where

sults in a transfer length equal to 60 db. The recommendation ld = development length,
for transfer length is only modified so that a minimum length lt = transfer length,
for transfer length is used, regardless of concrete strength. db = diameter of the prestressing strand, and
The recommendation effectively limits improvements in f c′ = design concrete strength.
transfer length based on a concrete release strength of 9 ksi,
Using concrete design strength of 6 ksi, which roughly cor-
which is less than the maximum release strength obtained in
responds to a “normal” concrete strength within the industry
the beams cast for this research (9.7 ksi on rectangular
and forms the base case from the experimental results, the co-
beams). Therefore, the final recommended expression for
efficient of 225 corresponds to flexural bond length of 90
transfer length is the following:
strand diameters.
120 Like the transfer length expression, the development
lt = db ≥ 40db (3.6)
fci′ length expression is limited by a minimum value. The rec-
ommended expression for development length, therefore, is
based on a limiting concrete strength of approximately 14
3.6.2 Development Length ksi, which is slightly less than the maximum concrete
Recommendations strength attained in beams tested in the research program
(14.9 ksi). Thus, the recommended development length
Since the inception of the pretensioned, prestressed concrete equation is as follows:
industry in the United States, the development length equation
has been made from the sum of two components: (1) transfer ⎡ 120 225 ⎤
ld = ⎢ + ⎥ db ≥ 100db (3.8)
length and (2) “flexural bond length,” which is the additional ⎣ f ci′ f c′ ⎦
length of bond beyond the transfer length required for devel-
opment. This approach has been utilized in the industry for
decades. Research continues to demonstrate that the approach 3.6.3 Distribution of Failure Types in Beams
is adequate to explain observed behavior and predict results. Tested
Thus, the same approach is followed, but with modifications
to include the effects of varying concrete strengths: This section presents the development length test results in
graphical fashion. The result of each beam test, whether flex-
• The results demonstrate that for all types of 0.5 in. ural failure or bond failure, is plotted on a chart showing con-
strands—Strands A/B and Strand D—flexural failures oc- crete strength versus embedment length. The recommended
curred at embedment lengths of 73 in. The embedment design equation for development length is also shown on
length of 73 in. corresponds to 100 percent of the current each of the charts. Note that the development length varies
code provision for development length for these speci- with concrete strength. For the purpose of plotting the values
mens. The results included tests on beams made with while using the equation, release strength is taken as 66.7 per-
concrete strength of approximately 4 ksi at release and cent of the design strength. This is a reasonable ratio of release
approximately 6 ksi at the time of the beam test. strength to design strength, borne out by years of experience
• The results uniformly indicate that the development length in prestressed concrete.
requirements diminish with increasing concrete strength. Figure 3.49 shows the results of development length tests
• The required development length calculated from the cur- on Strand D. Strand D demonstrated below average to poor
rent code provisions is approximately 150 db, although bond performance with a relatively low NASP Bond Test
some variations will exist due to variations in strand stress- result (6,890 lb), longer transfer lengths, and longer devel-
ing, beam geometry and subsequent variations in com- opment length requirements than Strands A/B. Figure 3.49
puted prestress losses. shows that bond failures occurred in rectangular beams
• If the transfer length is approximately 60 db, and the devel- with embedment lengths of 58 in. at the lower concrete
opment length is approximately 150 db, then the flexural strengths. More importantly, the figure shows improve-
bond length must be approximately 90 db. ment in strand bond behavior as concrete strengths in-
creased.
The development length expression can then be written as Note, however, that bond failures occurred in I-shaped
follows: beams cast with Strand D. Results of the tests demonstrate that
the Strand D, with an NASP Bond Test value of only 6,890 lb,
225db does not provide adequate bond-ability with concrete. Figure
ld = lt + (3.7)
fc′ 3.50 shows the results of development length tests on Strands

74
16000
14000
12000 Proposed Design Equation

Concrete Strength, f'c (psi)
10000
8000
6000
4000
2000
0
0 20 40 60 80 100
Embedment length (in.)
Flexural Failures - R-Beams Bond Failures - R-Beams Flexural Failures - I-Beams Bond Failures - I-Beams
Figure 3.49. Distribution of bond and flexural failures for Strand D (0.5 in.).
A and B. Both of these strands can be considered “high bond- Figure 3.51 shows the distribution of bond and flexural
ing,” since the NASP Bond Test value was so high. Strand B failures for 0.6 in. strand, Strand A6, with respect to concrete
was cast in the 4 ksi rectangular beams and I-shaped beams, strength and embedment lengths. As in Figure 3.50, the pro-
and Strand A was used in the higher strength rectangular posed design equation is shown in Figure 3.51 along with the
beams. The chart shows that the high-bonding strand was de- beam test results. There are no bond failures occurring in the
veloped in all concrete strengths, even in embedment lengths region where embedment length exceeds the calculated
as short as 46 in. The proposed design equation is shown on development length using the proposed equation. The tests
the chart along with the beam test results. support the proposed equation for development length.
16000
14000
12000
10000
8000
6000
Proposed Design Equation
4000
2000
0
0 20 40 60 80 100
Flexural Failures = R-beams Flexural Failures = I-beams

Proposed Equation Curve Shear Failures = I-beams
Figure 3.50. Distribution of bond and flexural failures for Strands A/B (0.5 in.).

75
16000
14000
12000 Proposed Design
Equation
10000
f'c (psi)
8000
6000
4000
2000
0
0 20 40 60 80 100 120
Bond Failures = R-beams Flexural Failures = R-beams

Flexural Failures = I-beams Proposed Equation Curve
Figure 3.51. Distribution of bond and flexural failures for

Strand A6 (0.6 in.).
3.6.4 NASP Value and Bond Performance worst performer of the four strands in both single strand
and double strand beams, with bond failures at the AASHTO
Along with the recommendation for the development development length of 73 in.
length design expression, it is important to recommend a Strand FF from Russell and Brown’s research (2004) is the
minimum value from the NASP Bond Test. First of all, how- same strand labeled Strand D in the NCHRP research. As seen
ever, it was important to establish a correlation between the in Tables 3.31 and 3.32, Russell and Brown reported a NASP
NASP pull-out test values and the bond performance of the Bond Test value of 7,300 lb for Strand FF. This compares with
same strands in transfer and in development length tests. a NASP Bond Test value of 6,890 lb in the NCHRP testing.
Russell and Brown (2004) measured transfer lengths and per- Strand FF demonstrated the ability to develop adequate ten-
formed flexural tests on rectangular-shaped beams. Table 3.31 sion in an embedment length of 73 in. in the rectangular
and Table 3.32 summarize the test results and the failure beams. However, if one looks at the results of the I-shaped
modes obtained from flexural tests performed by Russell and beams in Table 3.27, one can see that Strand D or Strand FF
Brown (2004).The NASP pull-out test values are also given. was unable to develop adequate strand tension at the devel-
Strand II had the lowest NASP Bond Test value, only 4,140 opment length of 73 in.
lb. Strand II is the same strand as that labeled Strand E in the
NCHRP research. One can see also that Strand II was the
Table 3.32. Failure mode on beams made with two
strands (Russell and Brown 2004).
Table 3.31. Failure modes on single-strand beams
(Russell and Brown 2004).
fc Average NASP Embedment Length
Beam No.
56 Days Pull-Out Value (in)
fc Average NASP Embedment Length (psi) (lb)
Beam No.
56 Days Pull-Out Value (in) 58 73
(psi) (lb) II21 6,290 4,140 B F
58 73
II22 6,280 4,140 B B
II11 6,290 4,140 B F
FF21 6,260 7,300 F F
II12 6,280 4,140 B B
FF22 6,070 7,300 F F
FF11 6,260 7,300 V F
HH21 6,330 10,700 F F
FF12 6,070 7,300 B F
HH22 6,300 10,700 F F
HH11 6,330 10,700 F F
AA21 6,220 14,950 F F
HH12 6,300 10,700 B F
AA22 6,160 14,950 F F
AA11 6,220 14,950 F F
AA12 6,160 14,950 F F F = Flexural Failure
F = Flexural Failure V = Shear Failure
V = Shear Failure
B = Bond Failure B = Bond Failure

76
Also, in Russell and Brown’s research (2004), NASP Round Table 3.33. Bond failures at 58 in. and 73 in.
IV testing, Strand HH demonstrated the ability to develop ad- for all 0.5 in. strands—I-shaped beams and
equate strand tension at the development length of 73 in. The rectangular beams.
NASP Bond Test value was 10,700 lb. One bond failure oc-
curred at an embedment length of 58 in. This occurred in a Number of Bond Failures
Strand NASP Value
single strand beam. The results from NASP Round IV testing (lb)
Name 58-in 73-in
reported by Russell and Brown (2004) indicate that the bond Embedment Embedment
performance of Strand HH was adequate. Length Length
Figure 3.52 shows the distribution of bond and flexural II 4,140 4 2
failures for Strand HH (0.5 in.) with respect to the concrete D 6,590 3 2*
strength and the provided embedment lengths. There are no
FF 7,300 1 0
bond failures occurring in the region where provided em-
bedment length exceeds the calculated development length HH 10,700 1 0
using the proposed equation. The tests support the proposed AA 14,950 0 0
equation for development length and also indicate that bond
B 20,210 0 0
performance of strand HH was adequate.
No bond failure was recorded on the beams with Strand A 20,950 0 0
AA. Comparing the NASP values of these strands, the follow- * Embedment lengths were 72 in. instead of 73 in.
ing observation can be made: as the NASP value increases,
chances of bond failure at provided embedment length de-
crease. In other words, Strand II had the lowest NASP value The number of bond failures is lower for strands with
and the highest number of bond failures, Strand FF and higher NASP pull-out values. Strand HH, with NASP pull-
Strand HH had NASP values lying between those of Strand II out value of 10,700 lb, lies at a critical position (boldfaced in
and Strand AA, and bond failures were noted on fewer occa- Table 3.33): strands with NASP pull-out values lower than
sions for Stand FF and Strand HH than for Strand II. Strand Strand HH’s pull-out value sustained bond failures, but no
AA had the highest NASP value and no bond failures, sug- strands with NASP pull-out values higher than Strand HH’s
gesting that it was capable of developing enough anchorage pull-out value suffered bond failure. Embedment lengths of
to achieve flexural failures. A higher NASP value seems to in- 58 in. and 73 in. correspond to 80 percent and 100 percent,
dicate better bonding qualities for the strand. respectively, of the code provision for development length.
Table 3.33 presents the number of failures obtained for all Strand HH suffered a bond failure at an embedment length
types of strands (0.5 in.) including NASP Round III Strands. of 58 in., but none at 73 in. These data show that a NASP
In Table 3.33, strands are arranged in the order of increasing pull-out value of 10,700 lb is adequate to develop enough an-
NASP pull-out values. The number of bond failures obtained chorage for achieving flexural failures at the code-specified
at 58-in. and 73-in. embedment lengths is shown. development length.
16000
14000
Proposed Design
12000 Equation
10000
8000
6000
4000
Flexural Failures - Round III Single Strand Beams
2000 Bond Failures - Round III Single Strand Beams
Flexural Failures - Round III Double Strand Beams
0
0 20 40 60 80 100
Figure 3.52. Distribution of bond and flexural failures for Strand

HH (Russell and Brown 2004).

77
3.7 The Effect of Concrete Strength • Current AASHTO code provisions may overestimate the
on Bond Performance— required development length of prestressing strands in
Summary, Conclusions, higher strength concretes.
and Recommendations • I-shaped beams were more susceptible to bond failures
than rectangular beams because of the higher incidence of
The research program involved development length tests web shear cracks developing in I-shaped beams.
on two types of beam specimens. Four types of strands were
employed to cast 43 rectangular-shaped beams and 8 I-shaped Finally, on the basis of the study findings, the following
beams. Both 0.5 in. and 0.6 in. diameter strands were included recommendations are made:
in the testing program. The beam specimens had concrete re-
lease strengths varying between 4 ksi and 10 ksi for both types • AASHTO code equations for transfer length should in-
of beams. Transfer lengths were measured on all beam speci- clude a parameter reflecting the reduced transfer length
mens using the strand end slip of the strands with the help of with increasing concrete release strength. The recom-
clamps attached to the strands. Transfer lengths were also mended equation for transfer length, lt (in.), is
measured using the concrete surface strain measurements.
Fifty flexural tests were carried out on the rectangular beams, ⎡ 120db ⎤
lt = ⎢ ⎥ ≤ 40db (3.9)
and 14 flexural tests were carried out on the I-shaped beams. ⎣ fci′ ⎦
Values of load, deflection, and strand end slip were recorded
electronically and manually along with photographic records where
of failure stages and crack patterns. I-shaped beam specimen f c′i = release concrete strength in ksi, and
concrete surface strains were measured at 36 in. from the end db = diameter of prestressing strands in inches.
of the beam and vertically at the center of the web. • AASHTO code equations for development length should
Prestressing strand anchorage requirements were assessed include a parameter reflecting the reduced transfer length
using the data collected from the development length tests. with increasing concrete release strength. Further, the flex-
Results from the development length tests were compared ural bond length is reduced by higher strength concrete as
with the NASP pull-out values of corresponding strands. well. The recommended equation for development length
Based on the failure modes during the development length is the following:
tests, the effect of concrete strength on bond performance was ⎡ 120 225 ⎤
analyzed. The current AASHTO code requirements for de- ld = ⎢ + ⎥ db ≥ 100db (3.10)
⎣ f ci′ f c′ ⎦
velopment length of prestressing strands were assessed for
their effectiveness in predicting accurate anchorage require- where
ments. The conclusions from this research are the following: ld = development length (in.),
f c′i = release concrete strength in ksi,
• Development length tests can be used to assess the bond f c′ = design concrete strength in ksi, and
performance of prestressing strands. db = diameter of prestressing strands in inches.
• The ability of a prestressing strand to bond with concrete is • A relatively large database has been collected during the
affected by concrete strength. Increasing concrete strength course of this research project. The data include crack pat-
improves the bond-ability of a given prestressing strand. terns, crack spacing, and surface strain measurements on
• The development length requirement for a particular I-shaped beams. A more detailed analysis should be made
strand is reduced if cast in higher strength concrete. using the information embedded in the summary reports
• The NASP Bond Test provides a good indicator of strand for a better understanding of the failure mechanisms. It is
bond performance in a pretensioned concrete beam. recommended that the Standardized Test for Strand Bond
• The required development length shows a clear relation- be adopted into the AASHTO LRFD Bridge Design Specifica-
ship with the NASP Bond Test values of the prestressing tions. The Standard Test for Strand Bond, formerly known
strand. Higher NASP Bond Test values result in shorter de- as the NASP Bond Test, requires an average pull-out value
velopment lengths. of 10,500 lb with no single test out of a sample of six tests
• Rectangular beams with all types of strands were able to falling below 9,000 lb. These values are established from the
achieve flexural failures at embedment lengths less than or review of the data obtained from the testing reported herein.
equal to the AASHTO-specified development length. Supporting data is found in the NASP Round III test report,
• With increased concrete strength, it is possible to achieve which is incorporated into this report via discussion in pre-
flexural failures at an embedment length less than the vious sections. The Standard Test Method for the Bond of
AASHTO-specified value. Prestressing Strands is recommended to ensure adequate

78
anchorage at embedment lengths equal to or higher than top and side clear cover to the bar being developed or spliced.
AASHTO code development length provision for normal- Details of typical specimens are shown in Table 3.34 and Fig-
strength concretes. ure 3.53. In Specimens I-4, I-5, and I-6, transverse reinforce-
• The effect of admixtures on the transfer and development ment was used in the splice region to confine the concrete as
length tests should be studied, with more development shown in Figure 3.54(b). The splice length shown in Column
length tests carried out while changing the proportions of 7 of Table 3.34 was selected to provide a direct link with com-
different admixtures in the concrete. panion test specimens containing epoxy-coated bars, which
were reported on in a separate paper and other tests in the lit-
erature. The same splice lengths were used for the specimens
3.8 Experimental Program—Mild
with transverse reinforcement so that the confining effect of
Steel Anchorage of Uncoated
this reinforcement could be evaluated. Rearranging Equation
Bars in Tension
12-1 of the 318 Code (ACI 2005) with appropriate modifica-
An extensive literature review of test data was conducted, tion factors and with a splice class factor of 1.0, it was possi-
and the results were reported in Chapter 2. The findings of ble to estimate a design stress and force in the bars for various
the literature review indicated the need to supplement the anchorage conditions, as shown in Equation 3.11. To deter-
data with six additional tests of top cast uncoated bar splices mine the calculated stress, fy (specified yield strength of rein-
in order to extend the use of the AASHTO LRFD Bridge forcing bars [psi]) is replaced with fs and ld is replaced by the
Design Specifications for development and splice length of splice length provided, 16, 24, and 36 in. for specimens with
uncoated bars to higher strength concretes. The variables #6 (#19M) bars, specimens with #8 bars, and specimens with
considered were bar size (#6, #8, and #11) and amount of #11 bars, respectively. Note that all the specimens had more
transverse reinforcement over the splice length. All six speci- than 12 in. of concrete cast below the splice. As shown in
mens tested had clear concrete cover of db. Table 3.34, all the bars in the specimens with transverse rein-
forcement had a calculated stress over the design stress of 60
ksi. These values are shown in Column 8 of Table 3.34 next to
3.8.1 Specimen Design
the yield design value.
Six beam splice specimens were tested. The specimen
⎡ ⎛ c + K tr ⎞ ⎤
dimensions and variables are shown in Table 3.34. The test 40 fc′ ⎜ b
variables were bar size and the presence of transverse rein- ld ⎢ ⎝ db ⎟⎠ ⎥
fs = ⎢ ⎥ (3.11)
forcements in the splice region in higher strength concretes. db ⎢ 3(1.3)ψ e ψ s ⎥
The cover value given in Column 3 of Table 3.34 is for both ⎢⎣ ⎥⎦
Table 3.34. Specimen dimensions and variables.
(1) (2) (3) (4) (5) (6) (7) (8) (9)

Specimen Bar Cover Beam Effective Number Splice 318-05 Test Date
Size (in.) Size (B x Depth of Spliced Length Cal. Compressive
H) (in.) (in.) Bars (in.) Stress Strength
(ksi) ( f c , ksi)
I-1 #6 0.75 9 x 18 16.88 3 16 52.10 16.2
I-2 #8 1.00 12 x 18 16.50 3 24 44.68 14.6
I-3 #11 1.50 18 x 18 15.75 3 36 49.89 16.2
I-4* #6 0.75 9 x 18 16.88 3 16 60 (66.91) 15.1
I-5* #8 1.00 12 x 18 16.50 3 24 60 (64.54) 14.6
I-6* #11 1.50 18 x 18 15.75 3 36 60 (63.55) 15.1
1 in. = 25.4 mm; 1 ksi = 6.89 MPa.

B = specimen width. H= specimen height.
(* shows specimens with transverse reinforcement in the splice region)

79
Splice Region: only I-4, I-5, and I-6 have stirrups in the splice region
18”
(12”, 9”)
Ls = 36” (24”, 16”)
54” 48” 54”
#11 Bars: #[email protected]” #4@8” #[email protected]”

#8 Bars: #3+#4@8” #4@8” #3+#4@8”
#6 Bars: #3@8” #3@8” #3@8”
Note. Ls = length of splice.
Figure 3.53. Specimen details (1 in. = 25.4 mm).
The factor representing the contribution of confining re- ters in both I-5 and I-6. Figure 3.54 shows the specimen rein-
inforcement across potential splitting planes is Ktr. The vari- forcing cages.
able cb represents the spacing or cover dimension, calculated
using either the distance from the center of the bar (or wire) 3.8.2 Test Set-Up
to the nearest concrete surface or one-half the distance of the
The beam splice setup used in this investigation is shown
center-to-center spacing of the bars being developed. ψe is a
in Figure 3.55. In all specimens, the distance between the
coating factor of 1.5 for cases with cover less than 3db, or clear
loading points and the support was 48 in. The constant mo-
spacing less than 6db, and 1.2 for all other cases. The param-
ment region was also 48 in. Splices were located within the
eter ψs is a reinforcement size factor: 0.8 for #6 bars and
constant moment region. To investigate the characteristics of
smaller and 1.0 for all other cases.
spliced beams, the applied loads, the resulting deflections at
The specimens were checked and reinforced in the over-
each beam end and midspan, and strains developed in longi-
hang region to prevent premature shear failures outside of the
tudinal bars and stirrups were monitored using load cells, lin-
test region. To prevent shear failure, a stress of 1.25 times the
ear variable differential transducers (LVDTs) anchored to a
yield strength of the bar was assumed in the overhang for pur-
reference frame, and electrical resistance strain gages attached
poses of estimating the required shear reinforcement to resist
to the bars, as shown in Figure 3.55 (b) and (c).
the maximum shear associated with the moment capacity of
the section at the support. The shear reinforcement in the
3.8.3 Materials
overhang region consisted of #3 @8 in., #3 + #4 @8 in., and
#4 @4.5 in., in Specimens I-1 and I-4, I-2 and I-5, and I-3 and Concrete and reinforcing steel were the materials used.
I-6, respectively. The shear reinforcement in the splice region Table 3.35 shows a typical concrete mix for the specimens.
consisted of #3 @8 in. on centers in I-4 and #4 @8 in. on cen- This mix was designed for a compressive strength of at least
(a) Specimen I-1 (b) Specimens I-4 & I-6
Figure 3.54. Specimen fabrication.

80
(a) Loading & Supporting (b) Measuring by LVDT
G2=15” G1=39” (Plan view for gage location)
Ls = 36”
6” 48” 48” 48” 6”

(Side view for Supporting)
Specimen Bar Size Ls (in.) G1 (in.) G2 (in.)

I-1, I-4 #6 (#19M) 16 19 13
I-2, I-5 #8 (#25M) 24 27 9
I-3, I-6 #11 (#35M) 36 39 15
(c) Location of gages and support (Specimen I-3)
Note: G1 = Gage 1. G2 = Gage 2.
Figure 3.55. Test setup (1 in. 25.4 mm).
15 ksi. The water to cement ratio was 0.20. The average mod-
ulus of rupture was 834 psi at 28 days. Typical maximum
compressive stress versus age data are shown in Figure 3.56.
Table 3.35. Typical concrete mix ratio The concrete strength continued to increase after 28 days and
(per 1 cubic yard). achieved a strength of 17 ksi at 56 days. The specimens began
to be tested after they reached a 15-ksi uniaxial compressive
Contents 15-ksi Mix strength.
Cement (lb) 900 The reinforcing bars were ASTM A615 Grade 60 steel and
Silica fume (lb) 200 had a yield strength based on tests of samples of the reinforc-
Water (lb) 220 ing bars of 78.3 ksi, 70.3 ksi, and 66 ksi for the #6, #8, and #11
Coarse aggregate (lb) 1800 bars, respectively. Stress versus strain curves for #6, #8 and
(1/2” crushed limestone)
#11 bars are shown in Figure 3.57.
Fine aggregate (lb) 1000
High-range water reducer (oz) 520 3.8.4 Cracking and Failure Mode
Normal-range water reducer (oz) 38
In nearly all tests, the cracking sequence was similar. First,
1 lb = 0.454 kg; 1 oz = 28.35 gr; 1 yd3 = 0.765 m3; 1 ksi = 6.89 MPa a flexural crack appeared in the constant moment region.

81
20
18
Compressive Strength (ksi)

16
14
12
10
8
6
4
2
0
0 20 40 60 80 100
Age (days)
Figure 3.56. Concrete stress versus age relationship

(1 ksi 6.89 MPa).
With the increase of beam end loads, a shear crack appeared 3.8.5 Beam End Displacement
in the overhang region and was arrested by the presence of the
shear reinforcement. Near the peak load, horizontal cracks The applied load versus deflection at the tip of the overhang
appeared along the longitudinal bars within the splice region. response for Specimens I-1 to I-6 is shown in Figure 3.59.
Finally, the deformations pushed the concrete away from the Load represents the average of the two values from the actua-
bar by wedge action. Failure crack patterns of all the speci- tors. Deflections were calculated by averaging displacements
mens are shown in Figure 3.58. All the specimens failed in at both ends of the beam. The test results are summarized in
splitting mode following yielding of the spliced bars in the Table 3.36. In the specimens without transverse reinforcement
constant moment region. in the splice region (Specimens I-1, I-2, and I-3), the end
#6 Black Bars
100
90
80
70
Stress (ksi)
60
50
40
30
20
10
0
0 2000 4000 6000 8000
Strain (x10^-6)
#8 Black Bars #11 Black Bars

100 100
90 90
80 80
70 70
Stress (ksi)
Stress (ksi)
60 60
50 50
40 40
30 30
20 20
10 10
0 0
0 2000 4000 6000 8000 0 2000 4000 6000 8000
Strain (x10^-6) Strain (x10^-6)
Figure 3.57. Tensile stress versus strain relationship (1 ksi 6.89 MPa).

82
(a) Specimen I-1 (b) Specimen I-2
(c) Specimen I-3 (d) Specimen I-4
(e) Specimen I-5 (f) Specimen I-6
Figure 3.58. Failure crack patterns for all the specimens for the #6, #8, and #11 bars.

83
I-1 I-2
30 50
40
Load (kips)
Load (kips)
20
30
20
10
10
0 0
0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00
Displ. (in.) Displ. (in.)
I-3 I-4
100 30
90
80
70 20
Load (kips)
Load (kips)
60
50
40
10
30
20
10
0 0
0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00
(c) Specimen I-3 (d) Specimen I-4
I-5 I-6
70 100
90
60
80
50 70
Load (kips)
Load (kips)
40 60
50
30 40
20 30
20
10
10
0 0
0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00
(e) Specimen I-5 (f) Specimen I-6
Figure 3.59. End load—end displacement curves for Specimens I-1 through I-6 (1 in. 25.4 mm; 1 kip 4.448 kN).
displacements at the peak load were 0.5 to 0.7 in. In the spec- constant moment region of Specimen I-6 are shown. Yield
imens with transverse reinforcement over the splice region strain in the longitudinal reinforcement was first recorded at
(Specimens I-4, I-5, and I-6), the end displacements at the around one-third of the peak load. Table 3.37 shows the
peak load were 0.8, 1.6 and 0.8 in., respectively. measured maximum strains on all of the specimens. All the
gages on the longitudinal reinforcement showed strains in ex-
cess of the bar yield strain before reaching peak load. In the
3.8.6 Bar Strains
gages placed on the stirrups in the constant moment region,
In Figure 3.60, the typical end concentrated load versus the measured maximum strain was around half of the bar
measured longitudinal bar and transverse bar strains in the yield strain in Specimens I-4 and I-5 and almost equal to the

84
Table 3.36. Summary of test results for uncoated bar specimens.
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)
Spec. Max. Displ. 318-05 318-05* AASHTO Test (7)/(4) (7)/(5) (7)/(6)
Load at Cal. Cal. Cal. Stress Max.
(kips) Peak Stress Stress (ksi) Stress
(in) (ksi) (ksi) (ksi)
I-1 28.2 0.506 52.10 41.68 38.10 78.55 1.51 1.88 2.06
I-2 39.6 0.429 44.68 44.68 42.86 70.93 1.59 1.59 1.65
I-3 88.6 0.654 49.89 49.89 45.59 67.65 1.36 1.36 1.48
I-4** 29.5 0.805 60 53.54 38.10 81.24 1.21 1.52 2.13

(66.91)
I-5** 59.4 1.572 60 60 42.86 91.88 1.42 1.42 2.14

(64.54) (64.54)
I-6** 96.4 0.800 60 60 45.59 71.94 1.13 1.13 1.58

(63.55) (63.55)
1 in. = 25.4 mm; 1 kip = 4.448 kN; 1 ksi = 6.89 MPa

* Shows stress calculated by removing bar size factor
** Shows specimens with transverse reinforcement in the splice region
bar yield strain in Specimen I-6. The use of stirrups in the Jirsa, and Breen study—together with contributions on bond
splice region of Specimens I-4, I-5, and I-6 resulted in an in- of reinforcement from ACI Committee 318 and ACI Com-
crease in the displacement capacity when compared with mittee 408 that were meant to simplify the provisions for
companion specimens I-1, I-2, and I-3, respectively. calculating development length of straight bars in tension—
led to Equation 12-1 in the 318 Code (ACI 2005), which is
3.8.7 U.S. Design Specifications Equation 3.12 herein:
3.8.7.1 318 Code (ACI 2005) ⎡ ⎤

⎢ 3 f y ψt ψe ψ sλ ⎥
Orangun, Jirsa, and Breen (1977) evaluated the results of a ld = ⎢ ⎥ db (3.12)
⎢ 40 fc′ ⎛ cb + K tr ⎞ ⎥
large number of bond and splice tests. The evaluation high-
⎢⎣ ⎜⎝ ⎟
db ⎠ ⎥⎦
lighted the importance of parameters such as bar diameter,
stress in the bar to be developed ( fc′ ), cover or bar spacing, In Equation 3.12, fy is the specified yield strength of rein-
and the amount of transverse reinforcement. The Orangun, forcing bars (psi), ψt is the reinforcement location factor of
I-6 (longitudinal reinforcement gage) I-6 (transverse reinforcement gage)

100 100
90 90
80 80
70 70
Load (kips)
Load (kips)
60 60
50 50
40 40
30 30
20 20
10 10
0 0
0 1000 2000 3000 4000 5000 6000 7000 0 1000 2000 3000 4000 5000 6000 7000
Strain (µ) Strain (µ)
(a) Longitudinal gage (b) Transverse gage
Figure 3.60. Beam end load versus measured strain relationship in Specimen I-6 (1 kip 4.448 kN).

85
Table 3.37. Measured maximum strains () in specimens I-1

through I-6.
Gage Location I-1 I-2 I-3 I-4 I-5 I-6
Longitudinal Bar 3,405 2,370 3,100 10,300 10,750 5,960
Transverse N/A N/A N/A 1100 875 1,910

Reinforcement
1.3 to reflect the adverse effects on top casting position on the 1963. A brief description of the background of the 1963 ACI
bond strength of the reinforcement. The parameter ψe is a specifications is provided below.
coating factor of 1.5 for cases with cover less than 3db, or clear The 1963 edition of the 318 Code provisions for bond and
spacing less than 6db, and 1.2 for all other cases. These factors anchorage for ultimate strength design were stated on the
are consistent with a ratio of bond strength of coated bars to basis of the ultimate flexural bond stress at the sections of
bond strength of uncoated bars observed in the literature of interest (ACI 1963), μu
1/1.5 = 0.67 and 1/1.2 = 0.82. However, the product of ψt and
Vu
ψe need not be taken greater than 1.7. The parameter ψs is a μu = (3.13)
reinforcement size factor: 0.8 for #6 bars and smaller and 1.0 ϕ ∑ o jd
for all other cases. The factor reflecting the lower tensile Critical sections were stated to occur at the face of support,
strength of lightweight concrete is λ. Bar diameter is db. The at each point of inflection, and at each point where tension
factor representing the contribution of confining reinforce- bars were terminated within a span. Vu was the factored shear
ment across potential splitting planes is Ktr. The variable cb at the section, ∑o, which represented the sum of bar perime-
represents the spacing or cover dimension, calculated using ter(s) at the same section, and jd was the flexural lever arm.
either the distance from the center of the bar (or wire) to the To prevent bond failure or splitting, the calculated tension
nearest concrete surface or one-half the distance of the cen- or compression force in any bar at any section had to be de-
ter-to-center spacing of the bars being developed. The ratio veloped on each side of that section by proper embedment
of (cb+ Ktr)/db should not be taken greater than 2.5. length or end anchorage, or, for tension only, by hooks.
However, the development length, ld, so calculated, cannot Anchorage, or development bond stress (μu), was to be de-
be less than 12 in. In addition, when calculating anchorage termined as the bar force, computed from M (moment at the
length requirements for tension lap splices, these should be as section due to factored loads) /ϕ, divided by the product of
required for a Class A or B splice, but not less than 12 in., ∑o times the embedment length. The two values so calcu-
where lated—ultimate flexural bond stress and anchorage bond
Class A splice..................1.0 ld stress—were not to exceed the limits given below, except that
Class B splice..................1.3 ld flexural bond stress did not have to be considered in com-
pression or in those cases of tension where anchorage bond
It must be noted that this factor is associated with the was less than 0.8 of the permissible stress given below. For
potential mode of failure when multiple bars are spliced at tension, there were two equations given for each of the two
the same location and does not speak to the actual strength of types of steel included: ASTM A 305 and ASTM A 408. For
the spliced bar. instance, for ASTM A 408, the permissible values were the
following:
3.8.7.2 2004 AASHTO Specifications (Section 5.11:
Development and Splices of Reinforcement) • Top bars (more than 12 in. of concrete below the bar)—
4.2 fc′ ;
The bond provisions for mild reinforcement in the
AASHTO LRFD Bridge Design Specifications mirrored the 318 • Bars other than top bars—6 f c′ ; and
Code provisions first introduced in the 1963 edition of the • For all deformed bars in compression—13 f c′ or 800 psi.
ACI Standard (ACI 1963). At the end of the last decade, the
ACI 318 provisions for development and splices of reinforce- In 1971, there was a complete revamping of the bond spec-
ment were extensively modified; however, the AASHTO ifications in ACI’s 318 Code. In the new format, a basic devel-
provisions for development and splices of reinforcement opment length, ldb, was determined and then modified by ap-
continued to mirror the ACI provisions first introduced in propriate factors to obtain the required anchorage length, ld.

86
ld = ldb * f1 * f 2 .... (3.14) which was revised on the basis of a review of available test re-
sults on large bars. The revised version for #18 bars was the
The development length concept replaced the dual system following:
contained in the 1963 ACI Code. It was no longer necessary
ldb = 0.125 * f y / fc′ (3.16)
to use the flexural bond concept, which placed an emphasis
on the computation of nominal peak bond stresses. The av- with fy and f c′ in psi. If put in ksi units,
erage bond resistance over the full development length of the
ldb = 3.95 * f y / fc′ (3.17)
bar is more meaningful in part because of the highly empiri-
cal nature of the design provisions and because bond tests in- This is an increase of 12 percent over the values given by
volve averaging of bond resistance. The current minimum the current AASHTO LRFD Bridge Design Specifications for
development length for bars in tension and in compression is the same size bars. Another important change introduced in
based on the attainable average bond stress over this length. the 1989 ACI Code was the limitation that fc′ cannot be
The various ld lengths in the 1971 ACI Code were based di- taken greater than 100 psi. This limitation meant that devel-
rectly on the 1963 ACI Code permissible bond stresses. opment lengths would no longer decrease with concrete
Slightly modified versions of the 1971 provisions in ACI’s 318 strengths greater than 10,000 psi. It was noted that research
Code (due to the fact that fy and f c′ are stated in terms of ksi) on development of bars in high-strength concretes was
are the current provisions for these design situations in the not sufficient to substantiate a reduction beyond the limit
AASHTO LRFD Bridge Design Specifications. imposed.
The basic tension development length, ldb (in.), for #11 bar While these provisions were based on extensive research
and smaller bars shall be taken as Equation 3.15: and professional judgment, many found them overly com-
plex in application. In 1999, Committee 318 of the ACI re-
ldb = 1.25 Abfy/ fc′ but not less than . . . 0.4 dbfy
examined these procedures with the goal of formulating a
For #14 bars: ldb = 2.7 fy/ fc′ more user-friendly format while maintaining general agree-
(3.15)
For #18 bars: ldb = 3.5 fy/ fc′ ment with the research results and professional judgment
that produced the changed provisions. The revision was
and for deformed wire: ldb = 0.95 dbfy/ fc′
based on the same general equation for development length
In Equation 3.15, Ab is the area of bar or wire (in.2), fy is the that served as the basis for the 1989 provisions. This equation
specified yield strength of reinforcing bars (ksi), f c′ is the spec- was Equation 12-1 in the 2005 version of the 318 Code (ACI
ified compressive strength at 28 days unless another age is 2005) and Equation 3.11 in this report.
specified (ksi), and db is the diameter of bar or wire (in.). In 1977, provisions for tension lap splices of deformed bars
The tension development length, ld, shall not be less than and deformed wire encouraged the location of splices away
the product of the basic tension development length, ldb, and from regions of high tensile stresses to locations where the
modification factor specified in Article 5.11.2.1.2 (for epoxy- area of steel provided at the splice location is at least twice that
coated bars with cover less than 3db or with clear spacing be- required by analysis. A lap splice of any portion of the total
tween bars less than 6db . . . 1.5, For epoxy-coated bars not area of steel in regions where (As provided/As required) was
covered above . . . 1.2 ). The tension development length shall less than 2.0 had to be at least 1.3 times the development
not be less than 12.0 in., except for lap splices specified in length of the individual bar in tension (Class B splice) in
Article 5.11.5.3.1 (Class A splice . . . 1.0 ld, Class B splice . . . length. If more than one-half of the reinforcement was
1.3 ld, Class C splice . . . 1.7 ld). spliced in such regions, lap splices had to be at least 1.7 times
In the 1989 ACI Code, major changes were made in the the development length of the individual bar (Class C splice)
procedures for calculating development lengths for deformed in length. Class A splices where the length of bar was equal to
bars and deformed wire in tension. This represented a major the development length of the individual bar were only per-
departure in approach between the ACI Code and the current mitted in regions where (As provided/As required) was less
AASHTO LRFD Bridge Design Specifications. These changes than 2.0 and no more than 25 percent of the total area was
resulted in an increase in the development lengths for closely spliced within one lap length. These same provisions are in
spaced bars and bars with small covers. The basic develop- the current AASHTO LRFD Bridge Design Specifications.
ment length was modified to reflect the influence of cover, When the changes in development in tension that eliminated
spacing, transverse reinforcement, casting position, type of many concerns regarding tension splice due to closely spaced
aggregate, and epoxy coating. The basic development lengths bars were introduced in the 1989 version of the 318 Code
remained essentially the same as in the 1971 edition of the (ACI 1989), Class C splices were eliminated.
ACI Code and the current AASHTO LRFD Bridge Design In summary, there are a few major differences between the
Specifications with the exception of the equation for #18 bars, ACI Code and the AASHTO LRFD Bridge Design Specifica-

87
tions with respect to development and splice length of tension specimens with bottom bars (478 specimens) was 0.51 to
reinforcement: 3.02, and some specimens had a ratio of less than 1, which
means the test bond strength was lower than the strength cal-
• The AASHTO LRFD Bridge Design Specifications don’t culated using the 318 Code (ACI 2005) without bar size
have bar size factor for smaller bars. factor. The bond strength of specimens with top bars (111
• The AASHTO LRFD Bridge Design Specifications don’t specimens) was 1.04 to 3.27. In tests for this study, the ratio
consider the role of confining reinforcement over the splice of test result to calculated result was 1.13 to 1.88. The design
region; however, in the ACI Code, the Ktr factor represents equation without the bar size factor conservatively estimated
the contribution of confining reinforcement across poten- bar stress for the specimens with top bars. However, it over-
tial splitting planes in the case of closely spaced bars with estimated the bar stress in many specimens with bottom bars,
small covers. especially for specimens with concrete compressive strength
• The AASHTO LRFD Bridge Design Specifications still con- higher than 10 ksi. These are tests with values greater than 100
tain Class C splices. The second and third differences are, psi along the horizontal axis. Figure 3.61(d) shows the com-
of course, related. This parameter is especially important parison of test maximum stress to calculated stress in the bar
because bars are being developed in higher strength using Equation 3.15 (AASHTO LRFD Bridge Design Specifi-
concretes. cations) on uncoated bars for the specimens reported by ACI
Committee 408 (2003). The bond efficiency (the ratio of test
stress to calculated stress using AASHTO LRFD Bridge Design
3.8.8 Bond Strength Comparisons
Specifications) of specimens with bottom bars (478 speci-
Table 3.36 shows the comparison of calculated stress in the mens) was 0.50 to 2.63, and 85 specimens had less than 1,
bar using Equations 3.12 and 3.15 and test results. In the spec- which means the test bond strength was lower than the
imens with transverse reinforcement (Specimens I-4 through strength calculated by Equation 3.15.
I-6), the 318 Code (ACI 2005) calculated stress was higher ACI Committee 408 (ACI 408R-03) proposed a new design
than the calculated stress in specimens without transverse re- equation for the bond and development of straight reinforc-
inforcement (Specimens I-1 to I-3). Also, the use of transverse ing bars in tension based on research by Zuo and Darwin
reinforcement over the splice region increased deflection at (2000). Figure 3.61(e) shows the comparison of stress in the
failure. The ratio of test maximum stress to ACI-calculated bar calculated using Equation 3.18 and the previous test
stress in the bar ranged from 1.13 to 1.59. The ratio of test results reported by ACI Committee 408 (2003). The result
maximum stress to AASHTO-calculated stress in the bar shows that the ratio of bond efficiency of specimens with bot-
ranged from 1.48 to 2.14. It should be noted that the second tom bars was 0.79 to 2.26, and only 12 specimens showed a
part of Equation 3.15 controlled the basic development length ratio of less than 1.
in the entire specimen, and the calculated flexural capacity was
⎡ ⎛ fy ⎞ ⎤
greater than the moment at failure. The failure moment ⎢ ⎜⎝ f ′1/4 − 2200ω ⎟⎠ αβλ ⎥
ranged between 60 and 98 percent of the flexural capacity. ld = ⎢ c
⎥ db (3.18)
Column 5 in Table 3.36 shows the calculated stress with the ⎢ 70 ⎛ cω + K tr ⎞ ⎥
⎢⎣ ⎜⎝ ⎟ ⎥⎦
bar size factor removed. Even though the test results of db ⎠
NCHRP Project 12-60 do not result in ratios of test maxi- In Equation 3.18, α is a factor reflecting the lower tensile
mum stress to calculated stress less than 1.0, on the basis of strength of lightweight concrete, β is 1.2 for all epoxy-coated
the analysis of the entire database, it is proposed that the 0.8 bars, λ is a factor reflecting the lower tensile strength of light-
bar size factor not be used for smaller bars. In Figure 3.61(a), weight concrete, and c, ω, and Ktr are defined as follows:
the comparison of test maximum stress to the stress calcu-
c = cmin + 0.5db (3.19)
lated using 318 Code (ACI 2005) for uncoated bottom bars
(reported by ACI Committee 408 [2003] and discussed in where
Chapter 2) is shown. It can be seen that many of the speci- c = spacing or cover dimension
mens had ratios less than 1. = cmin + db/2;
Figure 3.61(b) and (c) show the comparison of test maxi- cmin = minimum concrete cover or one-half of the clear
mum stress to calculated stress in the bar using Equation 3.11 spacing between bars, whichever is smaller,
without bar size factor for test results on uncoated bars = minimum (cb, cs);
reported by ACI Committee 408 (2003). In these figures, the cb = bottom concrete cover for reinforcing bar being de-
specimens are divided by casting position. The bond veloped or spliced;
efficiency (the ratio of test maximum stress to stress calcu- cs = minimum [cso , csi + 0.25 in.];
lated using 318 Code [ACI 2005] without bar size factor) of cso = side concrete cover for reinforcing bar;

88
2.5
Bond Efficiency (Test/ACI-05)

2
1.5
0.5
0
0 20 40 60 80 100 120 140
Concrete Strength (√fc', psi)
(a) Specimens with Bottom Bars (ACI-05)
3.5
3.5
3
3
Bond Efficiency (Test/ACI-05*)
Bond Efficiency (Test/ACI-05*)
2.5 2.5
2 2
1.5 1.5
1 1
0.5 0.5
0 0
0 20 40 60 80 100 120 140 0 20 40 60 80 100 120 140
Concrete Strength (√fc', psi) Concrete Strength (√fc', psi)
(b) Specimens with Top Bars (ACI-05*) (c) Specimens with Bottom Bars (ACI-05*)
2.5
3
2.5 2
Bond Efficiency (Test/ACI-408)
Bond Efficiency (Test/AASHTO)
2
1.5
1.5
1
0.5
0.5
0 0
0 20 40 60 80 100 120 140 0 20 40 60 80 100 120 140
Concrete Strength (√fc', psi) Concrete Strength (√fc', psi)
(d) Specimens with Bottom Bars (AASHTO) (e) Specimens with Bottom Bars (ACI-408)
(*shows stress calculated by removing bar size factor, 1 psi = 6.89 kPa)
Figure 3.61. Comparison of bond efficiency with concrete strength.

89
csi = one-half of the bar clear spacing; and imum displacement capacity of the beam end at failure for
db = diameter of bar. higher strength concretes.
c max
ω = 0.1 + 0.9 ≤ 1.25 (3.20)
c min 3.9 Anchorage of Epoxy-Coated
Bars in Tension
where cmax = maximum (cb, cs)
The object of this phase of NCHRP Project 12-60 was to
Ktr = (0.52trtdAtr/sn)f c′1/2 (3.21)
evaluate the bond strength of epoxy-coated bar lap splices in
where concrete with strengths up to 15 ksi. An extensive literature
tr = 9.6 Rr + 0.28 ″ 1.72; review of test data was supplemented with 12 additional tests
Rr = relative rib area of the reinforcement; of top cast epoxy-coated bar splices. The variables considered
td = 0.78db + 0.22; in the experimental program included bar size (#6 and #11),
Atr = area of each stirrup or tie crossing the potential plane concrete strength (12 to 17 ksi), and the amount of transverse
of splitting adjacent to the reinforcement being de- reinforcement over the splice length.
veloped, spliced, or anchored;
n = number of bars being developed or spliced; and
3.9.1 Literature Review
s = spacing of transverse reinforcement.
Epoxy-coated bars have been used as an economical
method of protection against deterioration of reinforced con-
3.8.9 Summary and Conclusions
crete structures associated with corrosion of steel reinforce-
On the basis of the analysis of results from the tests of six ment. Treece and Jirsa (1989) tested 21 beams in 9 series. The
beam specimens with lap-spliced uncoated bars embedded in variables were bar size (#6 and #11), concrete strength (4, 8,
higher strength concretes conducted as part of NCHRP and 12 ksi), casting position, and coating thickness (5 and 12
Project 12-60 and the evaluation of an extensive database of mils). The splice lengths were selected so that the bars would
test results compiled by ACI Committee 408, the following fail in bond before reaching yield, and no transverse rein-
conclusions can be drawn: forcement was provided in the splice region. Test results
showed that epoxy-coated bars with an average coating thick-
• The ratios of test maximum stress on the top spliced bars to ness above 5 mils developed 67 percent of the bond strength
the stress calculated from the design equation in the 318 of black bars.
Code (ACI 2005) ranged from 1.13 to 1.59. A similar ratio DeVries, Moehle, and Hester (1991) reported the test re-
of test maximum stress to stress calculated from the sults of 36 beams. The variables were casting position, bar size
AASHTO specifications ranged from 1.48 to 2.14. Thus, the (#6 and #9), and the presence of an antibleeding agent in the
procedure in the 318 Code (ACI 2005) and the AASHTO concrete. The range of concrete strengths was 8 to 15 ksi. Test
specifications for top bar uncoated splice and development results indicated that the ratio of bond strength of epoxy-
length in tension can be extended to normal-weight con- coated bars to black bars was 0.84. Based on the test results,
crete with uniaxial cylinder strength up to 16 ksi. De Vries and Moehle indicated that the effects of casting po-
• The design equation in the 318 Code (ACI 2005) and the sition and epoxy coating were not cumulative and that the
design equation in the AASHTO LRFD Bridge Design Spec- modification for top cast epoxy-coated bars relative to bot-
ifications, Equations 3.11 and 3.15, respectively, overesti- tom cast epoxy-coated bars was not needed. Also, the results
mated the bar stress in several of the bottom cast specimens showed that the presence of an antibleeding agent in the con-
in the ACI 408 Committee Database, especially for speci- crete did not significantly alter the bond stress of the splice for
mens with concrete compressive strength higher than 10 either top cast or bottom cast bars.
ksi. However, the calculated result proposed by ACI 408 Choi et al. (1991) reported on the tests of 15 beams. The
Committee, Equation 3.18, resulted in fewer cases where variables were bar size (#5, #6, #8, and #11), average coating
the ratio of test to calculated stress was less than 1.0. It also thickness (3 to 17 mils), and deformation patterns (three pat-
resulted in more conservative estimates of the bond terns designated S, C, and N). The concrete strength was
strength defined by the stress of spliced bars embedded in around 6 ksi. Test results indicated that the ratio of the bond
higher strength concrete beams. strength of epoxy-coated bar splices to that of black bar
• Based on the maximum bar stress and beam end displace- splices varied from 0.71 to 0.94 with an average value of 0.82.
ment at peak load in all the specimens, the use of stirrups They reported that all splice specimens exhibited extensive
in the amount of Ktr from 0.37 to 0.67 in the splice region longitudinal and transverse cracking in the region of the
resulted in increases in both maximum bar stress and max- splices at failure. The salient conclusion was that differences

90
in coating thickness have little effect on the amount of the strengths vary from 4 to 10 ksi. The relationship between the
bond strength reduction for #6 bars and larger with coating bond efficiency (the ratio of test stress to stress calculated using
thicknesses between 5 and 12 mils. 318 Code [ACI 2005]) of the spliced bars and the square root of
Hamad and Jirsa (1993) reported on an experimental concrete compressive strength in this literature review is shown
study in which 12 beams were tested. The main variables were in Figure 3.62. Note that the upper limit on the fc′ of 100 psi
bar size, bar spacing, and the amount of transverse reinforce- was removed in this calculation. Generally, the calculated stress
ment in the splice region. The concrete strength was around was conservative in the range of higher strength concretes.
4 ksi. Failure of all beams was governed by splitting of the
concrete cover in the splice region. Test results indicated that
3.9.2 U.S. Design Specifications
the presence of transverse reinforcement in the splice region
increased the deformation capacity of the beams and im- 3.9.2.1 318 Code (ACI 2005)
proved anchorage strength of epoxy-coated bar splices rela-
318 Equation 12-1 (ACI 2005) for estimating tension splice
tive to black bar splices more than 10 percent.
and development length requirements, Equation 3.12 in this
Cleary and Ramirez (1993) reported on an experimental
report, contains several factors. One of these is ψt, the tradi-
study in which 23 beam splice tests were subjected to repeated
tional reinforcement location factor of 1.3 to reflect the
loadings and then tested to failure to compare the service and
adverse effects of the top reinforcement casting position.
ultimate load behavior of beams with coated and uncoated
Parameter ψe is the specific coating factor to deal with epoxy-
reinforcement. The range of concrete strengths was 4 to 7 ksi.
coated bars. It is 1.5 with cover less than 3db or clear spacing
They reported that the differences in crack widths, deflec-
less than 6db, and it is 1.2 for all other cases. These factors are
tions, and reinforcement stresses in beams with coated and
consistent with the ratio of bond strength of coated bars to
uncoated reinforcement were reduced with repeated loading.
bond strength of uncoated bars reported in the literature of
The ratio of the average bond stress at failure for a beam
1/1.5 = 0.67 and 1/1.2 = 0.82. However, the product of ψt and
containing epoxy-coated bars to its companion specimen
ψe need not be taken greater than 1.7. All other factors are the
containing uncoated reinforcement ranged from 0.82 to 0.96,
same as for uncoated bars. In addition, as for uncoated bars,
with an average of 0.88.
when calculating anchorage length requirements for tension
Hester et al. (1993) tested 65 beam and slab splice specimens
lap splices, these should be as required for Class A or B splice
containing #6 and #8 bars. The average coating thickness
but not less than 12 in., where
ranged from 6 to 11 mils, and concrete strength ranged from 5
to 6.5 ksi. The Hester et al. study concluded that transverse re- Class A splice..................1.0 ld
inforcement improved the strength of splices containing both Class B splice..................1.3 ld
coated and uncoated bars, and the percentage increase in
strength was approximately the same for both coated and un-
3.9.2.2 2004 AASHTO Specifications (Section 5.11
coated bars with an equal amount of transverse reinforcement.
Development and Splices of Reinforcement)
A maximum development length modification factor of 1.35
was proposed for design with epoxy-coated reinforcement. In 1989, on the basis of several test programs that showed
Grundhoffer et al. (1998) reported on a series of 94 in- that the bond strength of epoxy-coated bars is reduced
verted half-beam specimens. The variables were bar size (#6,
Choi et al. (1991) Hamad and Jirsa (1993) Treece and Jirsa (1989)
#8, and #11), bar surface (epoxy and uncoated), concrete
DeVries, Moehle, and Hester (1991) Grundhoffer et al. (1998)
strength (6, 10, 12, and 14 ksi), and the addition of micro-
silica to concrete. A comprehensive review of the effect of 3.0
Bond Efficiency (Test / ACI-05)
epoxy-coating on bond strength was conducted using the 2.5

results of this study and 151 test results from seven other 2.0
research studies. They concluded that ACI’s 1989 318 Code
1.5
was more conservative than the 1995 318 Code for all the test
results based on the comparison between experimental re- 1.0
sults and the values of design bond strength calculated using 0.5
ACI’s 1989 and 1995 318 Code equations. 0.0
The review of past work shows that only two specimens of 0 50 100 150
Treece and Jirsa (1989), eight specimens of DeVries, Moehle, Concrete Strength (√f'c)
and Hester (1991) and some specimens of two groups out Figure 3.62. Bond efficiency of the spliced
of eight groups in Grundhoffer et al. (1998) used concrete bars with concrete strength relationship
strengths greater than 10 ksi. Other researchers’ concrete (1 psi-6.89 kPa).

91
because coating prevents adhesion between the bar and the In summary, although the factors are the same in both the
concrete, two factors—1.5 and 1.2 (function of the amount ACI Code and AASHTO LRFD Bridge Design Specifications
of concrete cover or bar spacing)—were introduced in the with respect to development and splice length of tension of
318 Code provisions for development length of bars in ten- epoxy-coated reinforcement, the same differences observed
sion. No factors were stated for similar bars in compression in the case of uncoated bars for the calculation of tension
or epoxy-coated bars terminated by means of standard hooks development length remain.
anchored to resist tension. Similar factors are currently em-
ployed in the AASHTO LRFD Bridge Design Specifications. 3.9.3 Experimental Program
The rest of the approach is the same as for uncoated bars. No
3.9.3.1 Test Specimens
factors were stated for similar bars in compression or epoxy-
coated bars terminated by means of standard hooks anchored The experimental program covers the testing of 12 beam
to resist tension. splice specimens reinforced with epoxy-coated bars. The spec-
The tension development length, ld , shall not be less than imen dimensions and variables are shown in Table 3.38. The
the product of the basic tension development length, ldb (see test variables are bar size, concrete cover, concrete strength,
Equation 3.15), and the modification factor specified in Arti- and transverse reinforcements in the splice region in higher
cle 5.11.2.1.2 (1.5 for epoxy-coated bars with cover less than strength concretes. The cover value given in Column 3 is both
3db or with clear spacing between bars less than 6d and 1.2 for top and side clear cover to the bar being developed or spliced.
epoxy-coated bars not covered above). The tension develop- Details of typical specimen are shown in Figure 3.63. In Spec-
ment length shall not be less than 12.0 in., except for lap imens II-15 through II-18, transverse reinforcement was used
splices specified in Article 5.11.5.3.1 (Class A splice . . . 1.0 ld, in the splice region to confine the concrete as shown in Figure
Class B splice . . . 1.3 ld, Class C splice . . . 1.7 ld). When the 3.64(b). In the splice region, the transverse reinforcement
changes that eliminated many concerns regarding develop- consisted of #3 @8 in. for Specimens II-15 and II-17 and #4 @8
ment length of tension lap splices due to closely spaced bars in. for Specimens II-16 and II-18, respectively.
were introduced in the 1989 version of the 318 Code, Class C The splice length shown in Column 7 of Table 3.38 was
splices were eliminated. selected to provide a direct link with previous tests in order to
Table 3.38. Specimen dimensions and variables.
(1) (2) (3) (4) (5) (6) (7) (8) (9)

Specimen Bar Cover Beam Effective Number Splice 318-05 Compressive
Size (in.) Size Depth of Length Cal. Strength
(B x H) (in.) Spliced (in.) Stress ( f c , ksi)
(in.) Bars (ksi)
II-7 #6 0.75 9 x 18 16.875 3 16 34.82 12.4
II-8 #11 1.50 18 x 18 15.750 3 36 33.34 12.3
II-9 #6 2.25 18 x 18 15.375 3 16 66.67 13.6
II-10 #11 4.50 24 x 18 12.750 2 36 63.83 13.6
II-11 #6 0.75 9 x 18 16.875 3 16 40.78 16.8
II-12 #11 1.50 18 x 18 15.750 3 36 39.05 16.8
II-13 #6 2.25 18 x 18 15.375 3 16 73.50 16.6
II-14 #11 4.50 24 x 18 12.750 2 36 70.38 16.6
II-15* #6 0.75 9 x 18 16.875 3 16 54.51 17.2
II-16* #11 1.50 18 x 18 15.750 3 36 51.76 17.2
II-17* #6 2.25 18 x 18 15.375 3 16 72.93 16.4
II-18* #11 4.50 24 x 18 12.750 2 36 69.83 16.4

1 in. = 25.4 mm; 1 ksi = 6.89 MPa (* denotes specimens with transverse reinforcement in the splice
region). B = specimen width. H = specimen height.

92
Specimens with transverse reinforcement in the splice region:

#4@8” with #11 Bars (#3@8” with #6 Bars)
18” (9”)
Ls = 36” (16” )
54” 48” 54”

Specimens with #11 Bars:
#[email protected]” Splice Region #[email protected]”
Specimens with #6 Bars: #3@8”
Figure 3.63. Typical beam-splice specimen reinforced with

epoxy-coated bars (1 in. 25.4 mm).
extend the specifications to higher strength concretes for for purposes of estimating the required shear reinforcement
epoxy-coated bars and to permit a more straightforward to resist the maximum shear associated with reaching the mo-
cover effect evaluation among specimens. The splice lengths ment capacity of the section at the support. In the overhang
have been selected to get a yielding stress in the basic speci- region, the spacing of shear reinforcement was #3 @8 in. on
mens with 3db concrete cover (II-9 and II-10) as shown in centers and #4 @4.5 in. on centers for specimens with #6 bars
Column 8 of Table 3.38. Using Equation 3.11 with appropri- and specimens with #11 bars, respectively. Figure 3.64 depicts
ate modification factors, including the epoxy-coated bar the construction of the specimens.
factor, and with a splice class factor of 1.0, it was possible to
calculate stress and force in the bar for various anchorage
3.9.3.2 Test Setup and Loading Protocol
conditions. To determine the calculated stress, fs, ld is replaced
by the splice length provided, 16 and 36 in. Note that all the The test setup is shown in Figure 3.65(a). In all speci-
specimens were cast with more than 12 in. below the splice. mens, the distance between the loading points and the sup-
As shown in Table 3.38, all the bars in specimens with 3db port was 48 in., and the distance between supports was also
concrete cover had a calculated stress greater than 60 ksi. 48 in. To investigate the characteristics of spliced beams,
The specimens were reinforced in the overhang region to the applied loads, resulting deflections at each beam end
prevent premature shear failures outside of the test region. and midspan, and strains developed in longitudinal bars
For safety against shear failure, a stress of 1.25 times the yield and stirrups were monitored using load cells, LVDTs
strength of the longitudinal bar was assumed in the overhang attached to an external reference frame, and electrical
(a) Specimens II-7 & II-8 (b) Specimens II-17 & II-18
Figure 3.64. Construction of beam-splice specimen with epoxy-coated bars.

93
(a) Loading & Supporting (b) Measuring by LVDT
#6 Bar Specimens (Ls = 16”, G1=19”, G2=13”)

G2=15” G1=39” (Plan for gage location and bars)
Ls = 36”
6” 48” 48” 48” 6”

(Elevation for Supporting)
(c) Location of gages and support (Specimen II-8)
Figure 3.65. Test setup for beam-splice specimens reinforced with epoxy-coated bars (1 in. 25.4 mm).
resistance strain gages affixed to the bars as shown in Fig- uniaxial compressive stress by age are shown in Figure 3.66(b).
ure 3.65 (b) and (c). As shown, the strength of Mix II continued to increase after 28
days and achieved a strength of 17 ksi at 56 days.
ASTM A615 Grade 60 reinforcing bars were used for both
3.9.3.3 Materials
longitudinal and transverse reinforcement. The yield
Table 3.39 shows the design concrete mixes. The water-to- strength, calculated by a 0.2-percent offset from tensile tests
cement ratio was 0.32 for the 10-ksi Mix I and 0.20 for the of samples of the reinforcing bars, was 70.3 ksi and 74 ksi for
14-ksi Mix II. A sample of the uniaxial stress versus strain re- the #6 and #11 bars, respectively. The average thickness of
lationship for the concrete is shown in Figure 3.66(a). The epoxy coating was 12.5 mils and 11.5 mils for the #6 and #11
average modulus of rupture was 566 psi and 834 psi at 28 days bars, respectively. The relative rib area was 0.091 and 0.135
for Mix I and Mix II, respectively. Also, typical data for for the #6 and #11 bars, respectively. The measured tensile
Table 3.39. Concrete mix (per cubic yard).
Contents Mix I: 10 ksi Mix II: 14 ksi

Cement (lb) 780 900
Silica fume (lb) 50 200
Water (lb) 265 220
1,600 1,800
Coarse aggregate (lb)
(3/8” pea gravel) (1/2” crushed limestone)
Fine aggregate (lb) 1,240 1,000
High-range water reducer (oz) 190 520
Normal-range water reducer (oz) 35 38
3 3
1 lb = 0.454 kg; 1 oz = 28.35 gr; 1 yd = 0.765 m ; 1 ksi = 6.89 MPa

94
Concrete Stress vs. Strain Relationship

20
Compressive Strength, ksi

15
10
5
10 ksi Mix
14 ksi Mix
0
0.0000 0.0005 0.0010 0.0015 0.0020
Strain, in/in
(a) Concrete Stress vs. Strain Relationship
20
18
16
14
12
10
8
6
4 10 ksi
2 14 ksi
0
0 20 40 60 80 100
Age (days)
(b) Concrete Strength vs. Age Relationship
100 100
90 90
80 80
70 70
Stress (ksi)
Stress (ksi)
60 60
50 50
40 40
30 30
20 20
10 10
0 0
0 2000 4000 6000 8000 10000 0 2000 4000 6000 8000 10000
Strain (x10^-6) Strain (x10^-6)
(c) # 6 (#19M) Bars (d) #11 (#35M) Bars
Figure 3.66. Material properties for beam-splice specimens reinforced with

epoxy-coated bars (1 in. 25.4 mm; 1 ksi 6.89 MPa).
stress versus strain curves for #6 and #11 bars are shown in With the increase of beam end loads, a shear crack appeared
Figure 3.66(c) and (d). in the overhang region. Near the peak load, splitting hori-
zontal cracks appeared along the longitudinal bars in the
3.9.4 Experiment Results splice region. Finally, the deformations pushed the concrete
away from the bar by wedge action. Typical failure crack pat-
3.9.4.1 Cracking Pattern and Mode of Failure
terns are shown in Figure 3.67. All the specimens failed in
In nearly all tests, the cracking sequence was similar. First, splitting mode after yielding of the spliced bars in the con-
a flexural crack appeared in the constant moment region. stant moment region.

95
(a) Specimen II-7 (b) Specimen II-8
(c) Specimen II-9 (d) Specimen II-10
(e) Specimen II-17 (f) Specimen II-18
Figure 3.67. Typical failure crack pattern for beam-splice specimens reinforced with epoxy-coated bars.

96
3.9.4.2 Load versus End Displacement 3.9.4.3 Summary of Test Results

Characteristics
The test results are summarized in Table 3.40 and findings
The applied load versus deflection at the tip of the over- from these results are presented on the basis of three main
hang response for Specimens II-7 to II-10 and II-17 and II-18 parameters.
is shown in Figure 3.68. Load was calculated by averaging the
two values from the actuators, and deflection was obtained Concrete Cover (db). Comparison of Specimens II-7 and
averaging displacements at both ends of the beam. II-9 (#6 bars) and comparison of Specimens II-11 and II-13
II-7 II-8
30 80
70
25
60
Load (kips)
Load (kips)
20 50
15 40
30
10
20
5 10
0 0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Displ. (in) Displ. (in)
(a) Specimen II-7 (b) Specimen II-8
II-9 II-10
35 80
70
30
60
25
Load (kips)
Load (kips)
50
20
40
15
30
10 20
5 10
0 0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
(c) Specimen II-9 (d) Specimen II-10
II-17 II-18
35 80
30 70
25 60
Load (kips)
Load (kips)
20 50
40
15
30
10
20
5
10
0 0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0
(e) Specimen II-17 (f) Specimen II-18
Figure 3.68. Applied load versus deflection at the tip of the overhang response (Specimens II-7
through II-10 and II-17 through II-18).

97
Table 3.40. Summary of test results.
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)
Speci- Max. Displ. 318-05 318-05* AASHTO Test (7)/(4) (7)/(5) (7)/(6)
men Load at Peak Cal. Cal. Cal. Stress Max.
(kips) Load Stress Stress (ksi) Stress
(in) (ksi) (ksi) (ksi)
II-7(db) 20.7 0.311 34.82 27.86 31.37 63.81 1.83 2.29 2.03
II-8(db) 61.5 0.423 33.34 33.34 37.55 65.50 1.96 1.96 1.74
II-9 29.0 0.687 66.67 48.94 31.75 78.39 1.18 1.60 2.47
(3db)
II-10 49.4 0.701 63.83 58.57 37.99 66.19 1.04 1.13 1.74
(3db)
II-11 21.0 0.315 40.78 32.63 31.37 65.50 1.61 2.01 2.09
(db)
II-12 64.5 0.395 39.05 39.05 37.55 65.00 1.66 1.66 1.73
(db)
II-13 32.1 1.161 73.50 53.96 31.75 83.45 1.14 1.55 2.63
(3db)
II-14 52.9 0.793 70.38 64.58 37.99 69.31 0.98 1.07 1.82
(3db)
II-15** 28.8 0.602 54.51 43.60 31.37 65.34 1.20 1.50 2.08
(db)
II-16** 92.0 0.662 51.76 51.76 37.55 65.96 1.27 1.27 1.76
(db)
II-17** 32.4 1.185 72.93 53.54 31.75 84.80 1.16 1.58 2.67
(3db)
II-18** 67.4 1.924 69.83 64.08 37.99 86.41 1.24 1.35 2.27
(3db)
1 in. = 25.4 mm; 1 kip = 4.448 kN; 1 ksi = 6.89 MPa
* shows stress calculated by removing bar size factor and using one epoxy-coated bar factor of 1.5.
** shows specimens with transverse reinforcement in the splice region.
(#6 bars) show that increasing the concrete cover increased maximum stress or deflection at failure. In Specimens II-10
both maximum stress and deflection at failure. This result can and II-14 with larger cover (3db), increasing the concrete
also be seen in comparison of Specimens II-8 and II-10 (#11 strength resulted in increases in both maximum stress and
bars) and comparison of Specimens II-12 and II-14 (#11 deflection at failure.
bars). However, the increase in maximum stress for #11 bar
specimens was less than the increase in maximum stress for Effect of Minimum Amount of Transverse Reinforce-
#6 bar specimens. ment in Higher Strength Concretes. A comparison of
Specimens II-11 and II-15 (#6 bars) shows that the use of
Effect of Concrete Strength. For Specimens II-7 and II-11 transverse reinforcement over the splice region did not re-
(#6 bars) with small cover (equal to db), increasing the con- sult in an increase in the maximum stress but more than
crete strength led to an increase in maximum stress, but did doubled the deflection at failure when the small cover (db)
not significantly increase the maximum deflection at failure. was used. When the large cover (3db) was used, it resulted
When larger cover (3db) was used, increasing the concrete in increases to both maximum stress and deflection at fail-
strength increased both maximum stress and deflection at ure, as can be seen by comparing Specimens II-13 and II-17.
failure as can be seen by comparing Specimens II-9 and II-13. Comparison of Specimens II-12 and II-16 (#11 bars) and
For Specimens II-8 and II-12 (#11 bars) with small cover comparison of Specimens II-14 and II-18 (#11 bars) show
(db), an increase in concrete strength did not increase the that the use of transverse reinforcement over the splice

98
region resulted in increases to both maximum stress and mens II-17 and II-18 (with larger covers and transverse rein-
deflection at failure. forcement) to the ratios for Specimens II-13 and II-14 (with
larger covers but no transverse reinforcement) is due to the
3.9.5 Comparison of Calculated Stress requirement that the ratio of (cb+ Ktr)/db in Equation 3.11
and Test Results should not be taken greater than 2.5.
The comparison of stress in the bar calculated using Equa-

tions 3.11 and 3.15 with appropriate modification factors and 3.9.6 Design Recommendation
test maximum stress is also shown in Table 3.40. The ratios of When the spliced bar stress was calculated using 318 Code
test maximum stress to ACI-calculated stress in the bar (ACI 2005), without a limitation on the square root of the
ranged from 0.98 to 1.96, as shown in Column 8. The ratios compressive concrete strength, only Specimen II-14 with 3db
of test maximum stress to AASHTO-calculated stress in the concrete cover had a ratio of test maximum stress to calculated
bar ranged from 1.73 to 2.67, as given in Column 10. It should stress of less than 1. However, the average ratio of test maxi-
be noted that the second part of Equation 3.15 controlled the mum stress to 318 Code (ACI 2005) calculated stress for the
basic development length. The findings from these compar- specimens with 3db cover was less than average of the same ratio
isons are discussed on the basis of three main parameters for the specimens with db cover. Therefore, it is possible to con-
studied: effect of concrete cover, effect of concrete strength, clude that the current cover contribution may be overestimated
and effect of minimum amount of transverse reinforcement in the case of higher strength concrete specimens with large cov-
in higher strength concretes. ers. Using only one coating factor of 1.5 may be the simplest way
to handle the possible overestimation of cover contribution.
Effect of Concrete Cover. In the specimens without trans-
Regarding bar size factor, no stress ratios less than 1 were
verse reinforcement, the ratio of test maximum stress to
obtained when the calculated stress included the 0.8 bar size fac-
ACI-calculated stress (see Column 8 in Table 3.40) in the
tor within the range of specimens covered in this study. The
specimens with 3db concrete cover was 0.98 to 1.18. The ratio
three specimens shown in Figure 3.62 with a ratio of test maxi-
of test maximum stress to ACI-calculated stress in the speci-
mum stress to calculated stress (defined in Figure 3.62 as “bond
mens with db concrete cover was 1.61 to 1.96, much higher
efficiency”) of less than 1 were specimens reinforced with #11
than in the specimens with 3db concrete cover. This tendency
(#35M) bars. However, following the position of ACI Commit-
was consistent regardless of other parameters, such as con-
tee 408, the authors of this report also suggest not using the 0.8
crete strength and bar size.
bar size factor. Column 5 of Table 3.40 shows the calculated
Effect of Concrete Strength. In higher strength concrete stress without the bar size modification factor. Figure 3.69
specimens without transverse reinforcement, the average shows a comparison of bond efficiency (defined as the ratio of
ratio of test maximum stress to ACI-calculated stress for the
specimens with db and 3db concrete cover was near 1.64 and Purdue No Stirrups With Stirrups Purdue(S)
1.06, respectively (see Column 8 in Table 3.40). These obser-
vations point to the possibility that the current cover contri- 3.5
bution in the code may be overestimated in the case of higher
3.0
strength concrete specimens for larger covers.
Bond Efficiency (Test / ACI-05*)
2.5
Effect of Minimum Amount of Transverse Reinforce-
ment in Higher Strength Concretes. The ratios of test max- 2.0
imum stress to ACI-calculated stress for Specimens II-11 and
II-14 are 1.61 and 0.98, respectively; the ratios of test maxi- 1.5
mum stress to ACI-calculated stress for Specimens II-15 and
II-18 are 1.20 and 1.24, respectively. These data show that for 1.0
specimens with transverse reinforcement over the splice re-

0.5
gion the difference between ratios of test maximum stress to
ACI-calculated stress was smaller than the difference between 0.0
ratios for specimens without transverse reinforcement (see 600 2080 40 100 120 140
Column 8, Table 3.40). The lower ratios in Specimens II-15 Concrete Strength (√f'c)
*Stress calculated by removing bar size factor and the epoxy-coated bar factor of 1.5
and II-16, with small covers, came from higher calculated
stress, considering the contribution factor of confining rein- Figure 3.69. Comparison of bond efficiency with
forcement. However, the similarity of the ratios for Speci- concrete strength.

99
test maximum stress to calculated stress) using 318 Code (ACI • The ratios of measured maximum stress on the spliced bars
2005) without a limit on the square root of the concrete com- to stress calculated using 318 Code (ACI 2005) ranged
pressive strength, without a bar size factor, and with a single from 0.98 to 1.96. Ratios calculated using 318 Code (ACI
epoxy-coated bar factor of 1.5 with concrete strength (defined 2005) with these two proposed modifications—no bar size
as the square root of the concrete compressive strength). In factor and a single epoxy-coated bar factor of 1.5—ranged
Figure 3.69, the specimens of Hamad, Jirsa, and D’Abreu de from 1.07 to 2.29. Thus, the procedure in Chapter 12 of the
Paulo (1993), Treece and Jirsa (1989), Choi et al. (1991), De- 318 Code (ACI 2005) for splice and development length of
Vries, Moehle, and Hester (1991), and the tests on epoxy-coated epoxy-coated bars in tension can be extended up to 17 ksi
bar splice specimens carried out under NCHRP Project 12-60 with the modifications suggested in this section.
(Purdue [S]) were separated into two groups: test results from • The ratios of measured maximum stress on the spliced bars
specimens with stirrups and test results from specimens with- to the stress calculated using the AASHTO specification
out stirrups over the splice region. As can be seen from Figure ranged from 1.73 to 2.67.
3.69, the bond efficiency values ranged from 0.86 to 3.11 for the • The use of transverse reinforcement over the splice region
specimens without stirrups and from 1.07 to 2.20 for the spec- resulted in increases both in the test maximum stress and
imens with stirrups. The ranges for each of the studies, except deflection at failure.
for NCHRP Project 12-60, are listed in Table 3.41. The use of • The current contribution of the cover in the 318 Code
transverse reinforcement over the splice region increased the (ACI 2005) can be overestimated in the case of higher
ACI-calculated stress, causing a decrease in the ratio of test max- strength concrete specimens with large covers.
imum stress to ACI-calculated stress, and this tendency was
consistent with the tendency of the tests conducted under
NCHRP Project 12-60. It is interesting to note as well that for 3.10 Anchorage of Bars Terminated
the studies in the literature, the range of stress ratios in the spec- with Standard Hooks
imens with epoxy-coated bars and companion specimens with in Tension
uncoated bars was similar, as shown in Table 3.41. Thus, on the
This section deals with the tensile strength of black and
basis of the maximum concrete compressive strength included
epoxy-coated reinforcing bars terminated in 90-deg hooks
in the experimental evaluation and the evaluation of the data in
with and without transverse reinforcement under monotonic
the literature, the procedure in Chapter 12 of the 318 Code (ACI
loading in normal-weight concrete with uniaxial compressive
2005) for splice and development length of epoxy-coated bars
strength up to 16 ksi. As part of this examination, in addition
in tension could potentially be extended up to 17 ksi without a
to 43 previous tests, the test results of 21 beam-column joint
limit of 100 psi to the square root of the concrete compressive
type specimens are reported. Variables in the tests conducted
strength and with these two modifications—removal of the bar
under NCHRP Project 12-60 included bar size (#6 and #11),
size factor and use of a single epoxy-coated bar factor of 1.5.
concrete strength (10, 14, and 16 ksi), and amount of trans-
verse reinforcement in the anchorage region. Codes and spec-
ifications have limits to their applicability to higher strength
From the test results of 12 beam splice specimens rein- concretes (ACI 2005, AASHTO 2004). These limits are justi-
forced with epoxy-coated bars, the following conclusions can fied on the basis of the empirical nature of code and specifi-
be drawn: cation requirements. The requirements for bars in tension
Table 3.41. Comparison of test results to calculated results.
Research Study Without Stirrups With Stirrups

Hamad and Jirsa (1993) (Black) 1.37 to 3.11 1.14 to 1.83
Hamad and Jirsa (1993) (Epoxy) 1.31 to 2.73 1.20 to 1.77
Treece and Jirsa(1989) (Black) 1.07 to 1.82 –
Treece and Jirsa (1989) (Epoxy) 0.86 to 1.71 –
Choi et al. (1991) (Black) 1.05 to 1.61 –
Choi et al. (1991) (Epoxy) 1.19 to 2.01 –
DeVries, Moehle, and Hester (1991) – 1.07 to 2.21
(Black)
DeVries, Moehle, and Hester (1991) – 1.22 to 2.20
(Epoxy)
–no data available.

100
anchored by means of a standard hook are an example of such other cases, these two factors are taken equal to 1.0. Other pa-
specifications. rameters are db, which is the bar diameter of the hooked bar;
In 1975, Marques and Jirsa reported a series of tests to de- f c′, which is the concrete compressive strength in psi; and the
termine capacities of uncoated hooked bars. Twenty-two square root of the concrete compressive strength, which shall
specimens simulating exterior beam-column joints were not exceed 100 psi as per Section 12.1.2 of the 318 Code (ACI
tested to evaluate the capacity of uncoated anchorage beam 2005). The modification factors of Section 12.5.3 of the 318
reinforcements subjected to varying degrees of confinement Code (ACI 2005) are all less than 1.0 and thus reduce the cal-
at the joint. The types of confinement included vertical col- culated length on the basis of cover, presence of ties where the
umn reinforcement, lateral reinforcement through the joint, first tie encloses the bent portion of the hook within 2db of
side concrete cover, and column axial load. To simulate beam the outside of the bend, and where anchorage or development
moment acting on the column, tension was applied to an- for specified minimum yield strength, fy, is not specifically
chored bars and a reaction assembly transferred compression required. These modification factors are the following.
load to the specimen. Failure in most tests was sudden and
resulted in the entire side cover of the column spalling away
For #11 bar and smaller hooks with side
to the level of the hooked anchorage. The maximum concrete
cover (normal to the plane of the hook)
compressive strength in these tests, which served as the basis
not less than 2.5 in. and for 90-deg hooks
for the current anchorage requirements, was 5.1 ksi.
with cover on the bar extension beyond
Anchorage of epoxy-coated hooked bars was evaluated by
the hooks that are not less than 2 in.: 0.7
Hamad, Jirsa, and D’Abreu de Paulo (1993) in a series of tests.
Twenty-four hooked-bar specimens simulating exterior For 90-deg hooks of #11 and smaller bars
beam-column joints were tested. It was reported that #11 that are enclosed within ties or stirrups
hooked bars (coated or uncoated) were consistently less stiff perpendicular to the bar being developed,
than #7 hooked bars. Epoxy-coated hooked bars consistently spaced not greater than 3db along ldh; or
developed lower anchorage capacities and load-slip stiffness enclosed within ties or stirrups parallel to
than companion uncoated hooked bars. The companion the bar being developed, spaced not
hooked-bar specimens that had ties in the beam-column greater than 3db along the length of the
joint region improved both the anchorage capacity and load- tail extension of the hook plus bend: 0.8
slip behavior of both coated and uncoated bars. For 180-deg hooks of #11 and smaller bars
To date, there has been little work on the anchorage that are enclosed within ties or stirrups
performance of hooked bars, black and epoxy-coated, in perpendicular to the bar being developed,
high-strength concrete. In the 2005 ACI Building Code, the spaced not greater than 3db along ldh: 0.8
equation for the basic development length (lhb) of a hooked
bar is limited to concrete strength of 10 ksi. Therefore, fur- Where anchorage or development for fy is
ther investigation on anchorage strength of hooked bars in not specifically required, reinforcement in (As required /
high-strength concrete is needed. excess of that required by analysis: As provided)
3.10.1 U.S. Design Specifications The factor As required/As provided, also referred to as the
factor for excess reinforcement, applies only where anchor-
3.10.1.1 318 Code (ACI 2005) age for full fy is not specifically required because the area of
Development length for deformed bars in tension termi- steel required to resist the factored flexural moment at the
nating in a standard hook, ldh, is determined using Section section, As required, is less than the area of steel provided, As
12.5.2 and applicable modification factors of 12.5.3, as shown provided, at the same section.
in Equation 3.22. However, ldh shall not be less than the larger
of 8db and 6 in. as indicated in Section 12.5.1 of the 318 Code 3.10.1.2 2004 AASHTO Specifications (Section
(ACI 2005). 5.11.2.4 Standard Hooks in Tension)
⎛ 0.02 ψ e λf y ⎞ The 1995 318 Code provisions for anchorage of bars termi-
ldh = ⎜ db (3.22)
⎝ fc′⎟⎠ nated in a standard hook in tension are the current procedure
in the AASHTO LRFD Bridge Design Specifications (ACI 1995).
In Equation 3.22, ψe is the coating factor, taken as 1.2 for The 1983 provisions for development of standard hooks in
epoxy-coated reinforcement; λ is the factor reflecting the tension in the 318 Code were a major departure from the 1977
lower tensile strength of lightweight concrete, which is 1.3. In 318 Code in that they uncoupled hooked bar anchorages from

101
straight bar development provisions and measured the

Reinforcement has a yield strength
hooked bar embedment length from the critical section to the
exceeding 60 ksi: fy/60
outside end or edge of the hook. The development length of
the hooked bar is represented by the product of a basic devel- Side cover for #11 bar and smaller,
opment length and appropriate modification factors. In the normal to the plane of the hook, is not
1995 edition of the 318 Code, a factor of 1.2 was introduced in less than 2.5 in., and cover on bar
the calculation of development lengths of epoxy-coated bars extension beyond 90-deg hooks is not
terminated in a standard hook (ACI 1995). less than 2 in.: 0.7
The development length, ldh (in.), for deformed bars in ten-
Hooks for #11 bar and smaller that are
sion terminating in a standard hook specified in Article
enclosed vertically within ties or stirrup
5.10.2.1 shall not be less than the following:
ties spaced along the full development
length, ldh, at a spacing not exceeding 3db: 0.8
• The product of the basic development length and the
applicable modification factor or factors, as specified in Anchorage or development of full
Article 5.11.2.4.2; yield strength is not required, or
• 8.0 bar diameters; or reinforcement is provided in excess (As required /
• 6.0 in. of that required by analysis: As provided)
Lightweight concrete is used: 1.3
Basic development length, lhb, for a hooked-bar with yield
strength, fy, not exceeding 60.0 ksi shall be taken as: Epoxy-coated reinforcement is used: 1.2
lhb = 38.0db (3.23)

f c′ 3.10.2 Experimental Program
where 3.10.2.1 Test Specimens
db = diameter of the hooked bar (in.) and
The experimental program reported in this research con-
f c′ = specified compressive strength of concrete at 28 days,
sisted of the monotonic loading in tension only (see Figures
unless another age is specified (ksi).
3.70 and 3.71) of 20 specimens with two bars terminated in
Below, cases in which basic hook development length, lhb, 90-deg standard hooks (see Figure 3.72).
should be multiplied by a factor are given, as well as the ap- Key test parameters are given in Table 3.42. The test
plicable factor. specimens were cast using normal-weight concrete (see
Concrete Column
(15”x15”) Strong Column
(W14x99)
Loading Plate
Load Cell 17.5”
Hydraulic Ram 15”

Stiffener
Bar Lock
Anchorage Plate Compression

Plate 17.5”
Strong Girder 11”

(2-MC10x33.6)
Figure 3.70. Test setup for beam-column-type specimens

(1 in. 25.4 mm).

102
Column Section (9”x15”) Column Section (15”x15”)

Ldh=6.5” Ldh=12.5”
3” 3”
3” 3”
6.5” 9” 6.5” 9”
3” 3”
3” 3”
9” 15”
Anchorage Plate Anchorage Plate
Figure 3.71. Specimen details for anchorage tests of bars terminated with
standard hooks (1 in. 25.4 mm).
Table 3.43). Specimens I-1 to I-6 contained black hooked epoxy-coated bars. In NCHRP Project 12-60, a similar test
bars. Specimens II-7 to II-12 had epoxy-coated hooked setup was used in the evaluation of these provisions in
bars. In Specimens III-13 to III-20, transverse reinforce- higher strength concretes.
ment was provided in the joint area to confine the concrete
along the anchorage length of the hooked bars. The speci-
3.10.2.2 Test Setup and Procedure
mens were provided with an anchorage length, ldh, as per
318 Code (ACI 2005) (see Table 3.42). The test setup used in this investigation is shown in Figures
Test specimens contained two #6 hooked bars or two #11 3.70 and 3.73. A force couple consisting of a tensile force in
hooked bars. The concrete column size of the specimens the test bars (applied by two center-hole hydraulic rams) and
with the #6 hooked bars was 9 by 15 in. The column cross a compressive force concentrated at a distance of 15 in. below
section of the specimens reinforced with the #11 bars (such the centerline of the bars was applied. The compression force
as I-2) was 15 by 15 in. The width of the column was kept at the face of test specimen was applied through two plates
the same in all specimens, but the depth was changed to ac- (3 in. and 3/4 in. thick) attached to the reaction column sim-
commodate the different development lengths. In both ulating a 6 in. deep compression zone of the assumed beam.
types of specimens, concrete cover was 2.5 in. Each concrete The reaction column consisted of a W14x99 column
column was reinforced with five or seven #8 main vertical welded to a base plate 1 in. thick and bolted to the strong
bars and 4 stirrups spaced at 6 in.—two at the top and two girder on the floor. Pull-out load was applied in 3.5-kip in-
at the bottom of the column as shown in Figure 3.72. The crements to the #6 bar specimens and in 10-kip increments
318 Code (ACI 2005) anchorage requirements for uncoated to the #11 bar specimens until failure occurred. Two strain
bars anchored by a combination of standard hook and gages were affixed to each bar, and the slip of the anchored
straight embedment length were based on the test results of reinforcing bar relative to the concrete surface was measured
Marques and Jirsa (1975). These provisions were later using LVDTs.
extended by Hamad, Jirsa, and D’Abreu de Paulo (1993) to
3.10.2.3 Materials
The two concrete mixes ordered from a concrete ready-
mix company were proportioned to yield a concrete com-
pressive strength of 10 ksi (Mix I) and at least 14 ksi (Mix II).
Table 3.43 shows a typical concrete mix. The water-to-
cement ratio was 0.32 for Mix I and 0.20 for Mix II. A stress
versus age relationship is shown in Figure 3.74. The modulus
of rupture was 566 psi and 834 psi at 28 days for Mix I and
Mix II, respectively.
Each size of reinforcing bar was from the same heat of steel,
and all bars had the same deformation pattern. The relative
rib area of #6 and #11 bars was 0.091 and 0.135, respectively.
Grade 60 steel was used for both black and epoxy-coated bars.
The yield strength obtained from tensile tests was 81.9 ksi and
63.1 ksi for the #6 and #11 black bars, respectively. For the
Figure 3.72. Detail of Specimen I-2. epoxy-coated bars, the yield strength calculated by 0.2-

103
Table 3.42. Description of key test parameters.
Name Bar Concrete Bar Type ldh Concrete Stirrup

Size Strength (psi) (in.) Cover (in.) Spacing
I-1 #6 8,905 Black 6.5 2.5 None
I-2 #11 8,905 Black 12.5 2.5 None
I-2’ #11 9,535 Black 15.5 2.5 None
I-3 #6 12,455 Black 6.5 2.5 None
I-4 #11 12,455 Black 12.5 2.5 None
I-5 #6 12,845 Black 6.5 2.5 None
I-6 #11 12,845 Black 12.5 2.5 None
II-7 #6 9,535 Epoxy-coated 6.5 2.5 None
III-13 #6 13,980 Black 6.5 db 3db
III-14 #11 13,980 Black 12.5 db 3db
III-15 #6 16,350 Black 6.5 db 3db
III-16 #11 16,500 Black 12.5 db 3db
III-17 #6 13,670 Epoxy-coated 6.5 db 3db
1 in. = 25.4 mm; 1 psi = 6.89 kPa
percent offset from tensile tests was 72.5 ksi and 74.7 ksi for in Figure 3.76. Load was measured using a load cell attached
the #6 and #11 bars, respectively. The average coating thick- to each bar terminated with a 90- deg standard hook.
ness measured with a dry film thickness gage for all epoxy- In almost all the specimens, the gages placed on the bar at
coated bars was around 12 mils. a distance of 2 in. from the column surface showed yielding
before reaching the maximum pull-out load, with less than
3.10.3 Experimental Results 0.05-in. slip between hooked bar and concrete surface on the
loaded side. As can be seen from Figure 3.75(a) and (b), spec-
3.10.3.1 Load versus Slip Behavior and Cracking
imens without shear reinforcement in the test region, II-9 and
Pattern
II-10, had a significant decrease in load at a 0.2-in. relative slip
Pull-out load versus slip responses for Specimens II-9, II-10, between hooked bar and concrete surface. However, in the
III-17, and III-18 are shown in Figure 3.75. The pull-out load specimens with shear reinforcement in the test region, III-17
versus slip responses for Specimen III-19 and III-20 are given and III-18 (see Figure 3.75[c] and [d]), the load decrease (20
Table 3.43. Typical concrete mix ratio (per 1 cubic yard).
Contents 10-ksi Mix 14-ksi Mix

Cement (lb) 780 900
Silica fume (lb) 50 200
Water (lb) 265 220
Coarse aggregate (lb) 1,600 1,800
(3/8” pea gravel) (1/2” crushed limestone)
Fine aggregate (lb) 1,240 1,000
High-range water reducer (oz) 190 520
Normal-range water reducer (oz) 35 38
3 3
1 lb = 0.454 kg. 1 oz = 28.35 gr. 1 yd = 0.765 m . 1 ksi = 6.89 MPa.

104
the behavior for the #6 bar specimens, III-17 and III-19,

recorded in Figure 3.75(c) and Figure 3.76(a), respectively, it
can be observed that the increase in concrete compressive
strength from about 13.5 ksi to 16.5 ksi resulted in an increase
in pull-out strength. The same increase in concrete strength in
the case of the specimens anchoring #11 bars, III-18 and III-20,
also resulted in an increase in pull-out strength (see Figure
3.75[d] and Figure 3.76[b]).The same type of finding was ob-
served for the specimens anchoring uncoated bars.
In almost all of the tests, the cracking sequence was simi-
lar. The first flexural (horizontal) crack occurred on the back
face of the specimens at a load of 20 kips for the #6 bar spec-
imens and 60 kips for the #11 bar specimens. The crack ap-
peared near the tail end of the hook. After the initial flexural
crack, a shear crack appeared on the side of the specimen as
Figure 3.73. Beam-column-type specimen test setup shown in Figure 3.77. With the increase in the pull-out load,
and instrumentation. the gage near the hook showed signs of yielding. At 90 per-
cent of the peak load, the vertical cracks appeared along the
column main bar. Finally, the concrete block near the hooked
percent of the peak load) was not as severe as the load
bar pushed out in Type I and II specimens that had no stir-
decrease observed in the specimens without shear reinforce-
rups in the joint (see Figure 3.78). In the Type III specimens
ment (almost 50 percent of the peak load). In Specimens
containing stirrups over the anchorage length, with the fail-
III-17 and III-18, at 0.2-in. slip, the sustained anchorage force
ure of the concrete near the hook, some of the side concrete
was more than 80 percent of the maximum pull-out force. It
cover spalled off (see Figure 3.79). Within the range of these
can be concluded that the #6 epoxy-coated bar specimen with
tests, there was no significant difference on the pull-out char-
shear reinforcement in the test region and smaller cover was
acteristics of the hooked bars with different concrete com-
able to reach a higher peak load than its companion specimen
pressive strength up to 16 ksi. It must be noted that with large
without shear reinforcement but with a larger cover (see Fig-
slips and with the tendency of the bar to straighten under ten-
ure 3.75[a] and [c]). This was not the case for specimens with
sion, the tail end of the hook tended to kick out, thus splitting
#11 epoxy-coated bars anchored by standard hooks (Figure
the concrete behind the hook. However, these cracks were
3.75[b] and [d]). However, the specimens with shear rein-
very small, implying that a cover of 2.5 in. over the tail end of
forcement (Figure 3.75[c] and [d]) were able to sustain
the hook used was sufficient for design purposes in the range
almost 80 percent of the peak load at a slip of 0.2 in. regard-
of concrete strengths considered in this study.
less of the bar size.
The load versus slip behavior of Specimens III-19 and III-20
are shown in Figure 3.76(a) and (b), respectively. Comparing 3.10.3.2 Maximum Pull-Out Stress and
Failure Mode
20
Table 3.44 shows a comparison of maximum pull-out
18
stress and the calculated stress on the basis of Equation 3.22.

16
This equation gives the straight embedment length calculated
14
in accordance with the 318 Code (ACI 2005) and measured
12
from the critical section to the outside portion of the hook. In
10
this equation, fy is the yield strength of hooked bar, ψe is the
8
6
coating factor (epoxy-coated reinforcement = 1.2, uncoated
4 10 ksi reinforcement = 1.0), λ is the lightweight aggregate concrete
2 factor (for lightweight concrete = 1.3, for normal concrete
14 ksi
0 =1.0), and f c′ represents the concrete compressive strength.
0 20 40 60 80 100 Substituting fs in place of fy, stress in the bar for a given
Age (days) anchorage length, and solving for s with a given design an-
Figure 3.74. Concrete compressive strength chorage length, as in Equation 3.24, it is possible to obtain the
versus age relationship in standard hook tests calculated stress shown in Table 3.44. The maximum stress
(1 ksi 6.89 MPa). corresponding to the peak pull-out load is obtained by

105
Load- Slip of II-9 (Right) Load - Slip of II-10 (Right)

30000 100000
25000 80000
Load (lb)
20000
Load (lb)
60000
15000
40000
10000
5000 20000
0 0
0 0.1 0.2 0.3 0.4 0.5 0 0.1 0.2 0.3 0.4 0.5
Slip (in.) Slip (in.)
(a) II-9 (#6, fc' = 13.5 ksi, cover = 2.5”) (b) II-10 (#11, fc' = 13.5 ksi, cover = 2.5”)
Load- Slip of III-17 (Right) Load - Slip of III-18 (Right)

40000 80000
35000 70000
30000 60000
Load (lb)
50000
Load (lb)
25000
20000 40000
30000
15000
20000
10000
10000
5000
0
0 0 0.1 0.2 0.3 0.4 0.5
0 0.1 0.2 0.3 0.4 0.5
Slip (in.)
Slip (in.)
(d) III-18 (#11, fc' = 13.5 ksi, cover = db)
(c) III-17 (#6, fc' = 13.5 ksi, cover = db)
Figure 3.75. Effect of transverse reinforcement in the anchorage region on the pull-out load versus
slip response (1 in. 25.4 mm; 1 kip 4.448 kN).
averaging the values from the two load cells attached to each 1.23). However, most of the specimens anchoring # 11 bars at
hooked bar divided by the area of the bar. failure reached a stress less than or equal to the calculated
stress (ratio of test result to calculated result of 0.83 to 1.02).
( )
⎡ ldh
⎢
f s = ⎣ db
* fc′ ⎤⎥
⎦ (3.24) In the case of specimens anchoring epoxy-coated bars (Series
II), the tendency was the same as in Series I specimens rein-
0.02ψ e
forced with black bars. In Series III, the specimens anchoring
Specimens with #6 bars, except Specimens II-9 and III-15, #11 bars reached failure stress levels less than or equal to
at failure reached a stress equal to or greater than the calcu- the calculated stress values, yielding a ratio of test to calcu-
lated stress (ratio of test result to calculated value of 0.99 to lated stress ranging from 0.83 to 0.98 while the specimens
Load- Slip of III-19 (Left) Load - Slip of III-20 (Right)

45000 140000
40000 120000
35000
100000
30000
Load (lb)
Load (lb)
25000 80000
20000 60000
15000
40000
10000
5000 20000
0 0
0 0.1 0.2 0.3 0.4 0.5 0 0.1 0.2 0.3 0.4 0.5
Slip (in.) Slip (in.)
(a) III-19 (#6, fc' = 16.5 ksi, cover = db) (b) III-20 (#11, fc' = 16.5 ksi, cover = db)
Figure 3.76. Effect of high-strength concrete in the anchorage region on the pull-out load
versus slip response (1 in. 25.4 mm; 1 kip 4.448 kN).

106
(c) Specimen III-13 (d) Specimen III-14
(e) Specimen III-19 (f) Specimen III-20
Figure 3.77. Crack patterns.

107
Figure 3.78. Concrete block push off in specimens Figure 3.79. Failure region in the case of Specimen
without stirrups in the anchored hooked bar III-13 with stirrups.
specimens.
anchoring #6 bars reached failure stresses greater than the cal- the case of #6 bars, and 15 percent more force in the case of
culated values, i.e., ratios ranging from 1.01 to 1.14, except #11 bars. Each pair of specimens had the same dimensions
Specimen III-15, which failed at a lower level. A similar com- and material properties, but had different details, such as hav-
parison was conducted in terms of force developed in the bar. ing ties, having no ties, or having different concrete cover.
Comparison of the results of Specimens II-9 and II-10 with Taking into account these similarities and differences in the
Specimens III-17 and III-18 indicated that the use of ties over specimens, it can be concluded that the confinement (with
the joint region developed 56 percent more force in the bar in ties) of the anchorage region produced stronger bond char-
Table 3.44. Comparison of maximum pull-out bar stress compared

with calculated stress using the 318 Code (ACI 2005) method with
a modification factor of 0.7 (ksi).
Specimen db (in.) f c (psi) ldh (in.) Calculated Test Ratio (T/C) Force (kips)
I-1-9 0.75 8905 6.5 58.4 68.2 1.17 30.0
I-2-9 1.41 8905 12.5 59.8 56.4 0.94 88.0
I-2'-10 1.41 9535 15.5 76.7 67.3 0.88 105.0
I-3-12 0.75 12455 6.5 69.1 68.2 0.99 30.0
I-4-12 1.41 12455 12.5 70.7 63.5 0.90 99.1
I-5-13 0.75 12845 6.5 70.2 69.3 0.99 30.5
I-6-13 1.41 12845 12.5 71.8 73.1 1.02 114.0
II-7-10 0.75 9535 9.5 73.6 90.9 1.23 40.0
II-8-10 1.41 9535 15.5 63.9 56.4 0.88 88.0
II-9-14 0.75 13670 6.5 60.3 56.1 0.93 24.7
II-10-14 1.41 13670 12.5 61.7 53.5 0.87 83.5
II-11-15 0.75 14800 6.5 62.8 64.8 1.03 28.5
II-12-15 1.41 14800 12.5 64.2 54.5 0.85 85.0
III-13-14 0.75 13980 8.3 92.9 93.8 1.01 41.3
III-14-14 1.41 13980 13.5 80.9 67.3 0.83 105.0
III-15-16 0.75 16350 8.3 100.5 87.5 0.87 38.5
III-16-16 1.41 16500 13.5 87.8 76.9 0.88 120.0
III-17-14 0.75 13670 8.3 76.6 87.5 1.14 38.5
III-18-14 1.41 13670 13.5 66.6 61.5 0.92 95.9
III-19-16 0.75 16350 8.3 83.7 89.8 1.07 39.5
III-20-16 1.41 16500 13.5 73.2 71.8 0.98 112.0
1 in. = 25.4 mm; 1 kip = 4.448 kN; 1 psi = 6.89 kPa; 1 KSI = 6.89 MPa

108
120 Project 12-60 test results follow in general the same trend as
those of previous researchers. Thus, it is plausible to propose
to extend the current ACI procedure for hooked bars up to 16
Maximum Experimental Stress (ksi)
100
ksi without a limit on the fc term.
80 In Figure 3.81, the ratio of test to calculated stress for
hooked bars is shown versus the concrete compressive
strength of the specimen. It can be seen that the ratio de-
60
creases as the concrete compressive strength is increased in
both black and epoxy-coated bars terminated with a standard
40
hook and subjected to direct tension. To increase the values
of the ratio of test to calculated stress in specimens with
20 Marques & Jirsa (1975) higher concrete strengths, it is proposed that a 0.8 modifica-
Kim
Hamad, Jirsa & D’Abreu de Paulo (1993)
tion factor be used instead of the current factor of 0.7 [for
0 hooks with side cover not less than 2-1/2 in. and for 90-deg
0 20 40 60 80 100 120 hooks with cover on bar extension beyond hook not less than
Calculated Stress (ksi)
2 in. in ACI Code 12.5.3(a) in concrete strengths above 10
Figure 3.80. Maximum experimental ksi]. The calculated results using the proposed modification
stress versus 318 Code calculated stress factor and current factor are shown in Figure 3.82.
(1 ksi 6.89 MPa).

acteristics in hooked bars than no ties with a 2.5 in. cover. Based on the review of over 40 specimens in the literature
This tendency is observed with both the epoxy-coated bars and the results from 21 tests of hooked bar anchorages in
and black bars. beam-column specimens with normal-weight concrete
strengths up to 16 ksi, the following conclusions can be drawn:
3.10.3.3 Comparison with Other Tests
• The approach in the 318 Code (ACI 2005) provision for
and Recommendations
anchorage of bars terminated in standard hooks in tension,
The comparison of the stress calculated using Equation black and epoxy-coated, can be extended to concrete com-
3.24 for NCHRP Project 12-60 tests (Kim) and test results pressive strengths up to 15 ksi. However, a minimum
reported by Hamad, Jirsa, and D’Abreu de Paulo (1993) and transverse reinforcement (3db spacing) should be provided
Marques and Jirsa (1975) are plotted against the experimen- in higher strength concretes to improve the bond charac-
tal values in Figure 3.80. The comparison shows that NCHRP teristics of both epoxy-coated and black #11 hooked bars.
1.6
1.6
1.4
1.4
1.2
1.2
Test / Calculated
Test / Calculated
1.0
1.0
0.8
0.8
0.6
Marques & Jirsa (1975) 0.6
0.4 0.4
Kim
0.2 0.2
Hamad, Jirsa & D’Abreu de Paulo (1993)
0.0 0.0
0 5000 10000 15000 20000 0 5000 10000 15000 20000
Concrete Compressive Strength (psi) Concrete Compressive Strength (psi)
(a) Black Bar Specimens (b) Epoxy-Coated Bar Specimens
Figure 3.81. Test to calculated stress ratio versus concrete compressive strength (1 psi 6.89 KPa).

109
120 • The epoxy-coated hooked bars developed lower anchorage

capacities than uncoated hooked bars. In the #11 hooked
Maximum Test Stress (ksi) 100 bar specimens, the ratios of measured stress to calculated
stress were 0.85 to 0.88.
80 • Transverse reinforcement in the anchorage length of a
bar terminated with a standard hook improves the max-
60 imum pull-out strength and load versus slip behavior.
In the #11 epoxy-coated hooked bar specimens, the
40 ratios of measured stress to calculated stress increased up
to 0.98.
20 Cal.-0.7 • While the minimum concrete cover of 2.5 in. at the end of
Cal.-0.8 the hook appeared to be adequate to prevent kicking out of
0 the tail end of the hooked bar, it is proposed that a modi-
0 20 40 60 80 100 120
fication factor of 0.8 be used instead of 0.7. The use of a 0.8
Calculated Stress (ksi)
modification factor eliminated almost the entire test to cal-
Figure 3.82. Maximum test stress versus culated stress ratios less than 1.0. This value of minimum
calculated stress using factors of 0.7 and concrete cover can be reduced to db if transverse reinforce-
0.8 for Marques and Jirsa (1975), Hamad ment is used in the anchorage length of a bar terminated
et al. (1993), and those tested in NCHRP with a standard hook.
Project 12-60 (1 ksi 6.89 MPa).

110
CHAPTER 4
Design Recommendations
4.1 Introduction • Article 5.11.4.3 Partially Debonded Strands

• Article 5.11.5 Splices of Bar Reinforcement
Article 5.4.2.1 of the AASHTO LRFD Bridge Design Specifi- • Article 5.11.5.3.1 Lap Splices in Tension
cations (2004) limits the applicability of the specifications for
concrete compressive strengths of 10 ksi or less unless physi-
cal tests are made to establish the relationships between 4.2 Design Recommendations
concrete strength and other properties. A comprehensive
For prestressing strands, there are five essential recom-
article-by-article review of Section 5 of the AASHTO LRFD
mendations stemming from the research:
Bridge Design Specifications (2004), pertaining to transfer and
development of prestressing strand and splice length and
1. Adoption of the Standard Test Method for the Bond of
anchorage of free ends by means of standard hooks for mild
Prestressing Strands. Heretofore, this test has been known
reinforcement was performed under NCHRP Project 12-60 to
as the NASP Bond Test.
identify all provisions that would have to be revised directly or
2. Adoption of a transfer length expression incorporating a
indirectly to extend their use to high-strength, normal-weight
factor to account for improved bond with increasing con-
concrete up to 15 ksi. These articles are the following:
crete strength. The recommended expression reflects the
• Article 5.4.4 Prestressing Steel decrease in transfer lengths as concrete release strengths
• Article 5.5.4.2. Resistance Factors increase. For release strengths of 4 ksi, the recommended
• Article 5.11 Development and Splices of Reinforcement expression would provide a transfer length of 60 strand di-
• Article 5.11.2 Development of Reinforcement ameters, which is the same value found in prior editions
• Article 5.11.2.1 Deformed Bars and Deformed Wire of the AASHTO LRFD Bridge Design Specifications. For
• Article 5.11.2.1.1 Tension Development Length release strengths of 6 ksi, the design transfer length would
• Article 5.11.2.1.2 Modification Factors that Increase ld be approximately 50 db. Transfer lengths would be limited
• Article 5.11.2.1.3 Modification Factors which Decrease ld to a minimum of 40 strand diameters.
• Article 5.11.2.2 Deformed Bars in Compression 3. Adoption of a development length expression that incor-
• Article 5.11.2.2.1 Compressive Development Length porates factors to account for improved strand bond as
• Article 5.11.2.2.2 Modification Factors concrete strengths increase. The recommended code
• Article 5.11.2.3 Bundled Bars expression is founded on the same principles as prior edi-
• Article 5.11.2.4 Standard Hooks in Tension tions of the AASHTO LRFD Bridge Design Specifications,
• Article 5.11.2.4.1 Basic Hook Development Length i.e., the development length is the sum of a transfer length
• Article 5.11.2.4.2 Modification Factors expression plus a flexural bond expression. At “normal”
• Article 5.11.2.5 Welded Wire Fabric concrete strengths, the development length expression
• Article 5.11.2.5.1 Deformed Wire Fabric requires 60 strand diameters for transfer length and ap-
• Article 5.11.2.5.2 Plain Wire Fabric proximately 90 strand diameters for the flexural bond
• Article 5.11.2.6 Shear Reinforcement length, for a total of 150 strand diameters. For a concrete
• Article 5.11.4 Development of Prestressing Strand release strength of 6 ksi and design concrete strength of
• Article 5.11.4.1 General 10 ksi, the development length expression provides a
• Article 5.11.4.2 Bonded Strand development length of 120 strand diameters. For higher

111
concrete strengths, the development length is limited to a Prestressing Strands. The reproducibility of the test is summa-
minimum of 100 strand diameters. rized in Figure 3.7 where the Standard Test Method for the
4. Adoption of a modified bilinear build-up of strand stress Bond of Prestressing Strands was performed on identical strand
consistent with the recommended transfer and develop- samples at Purdue University and at OSU. Altogether, the data
ment length expressions. illustrated in Figure 3.7 come from tests performed on five dif-
5. Adoption of additional restrictions regarding the use of ferent 0.5-in.-diameter strand samples and two different 0.6-
debonded, or shielded, strands. in.-diameter strand samples. Figure 3.7 shows that there is a
high degree of statistical correlation between the results from
The following recommendations, discussed in detail in the two sites. The coefficient of determination, R2, is 0.92.
Section 4.3.3, stemming from the work conducted under Perhaps even more importantly, Figure 3.7 helps to show that
NCHRP Project 12-60 address the anchorage of Grade 60 the test results fall very near to the “perfect” line, where nearly
mild steel in tension: identical results are obtained at the two sites independently. The
tests at Purdue University were completely blind, as strand des-
1. Development length of black and epoxy-coated reinforc- ignations for the strands tested at Purdue University did not
ing bars anchored by means of straight embedment length match the strand designations on the same samples at OSU.
and splices. The protocols for the NASP Bond Testing used in NCHRP
2. Anchorage of black and epoxy-coated bars terminated Project 12-60 were based on the NASP Bond Test from May
with standard hook. 2004 and are found in Appendix I. However, some refine-
ments were made during the early part of the NCHRP testing
4.3 Details of the Design at OSU, and the round-robin testing between Purdue Univer-
Recommendations sity and OSU more closely matched the protocols that were
further refined during the NCHRP testing program. These
4.3.1 Prestressing Strand—Adoption protocols are now titled, “Standard Test Method for the Bond
of the Standard Test Method for Bond of Prestressing Strands,” and are recommended as the basis for
of Prestressing Strands the material requirements to ensure “bond-ability” of pre-
Table 5.4.4.1-1 reiterates the requirements for mechanical stressing strands with concrete. The final recommended stan-
properties for the prestressing steels that are found in the two dard is included in Appendix H. Some of the modifications
AASHTO material specifications. Mechanical properties in- included in the April 2006 edition of the Standard Test
clude the grade or type, the size, the tensile strength (ksi) and Method for the Bond of Prestressing Strands include provi-
the yield strength (ksi). Research findings from NCHRP Proj- sions for minimum acceptance and frequency of testing.
ect 12-60 support the conclusion that the bond of prestressing Therefore, it is recommended that the Standard Test
strand should be recognized as a material property of the Method for the Bond of Prestressing Strands be adopted and
strand and included in Section 5.4.4, “Prestressing Steel,” of the made part of the LRFD specifications. Notably, the North
AASHTO LRFD Bridge Design Specifications. The research American Strand Producers Committee of the American
described in Chapter 3 of this report provides supporting Wire Producers Association has formally, by unanimous
evidence that the Standard Test Method for the Bond of Pre- vote, adopted the test procedure as their standard for bond.
stressing Strands, found in Appendix H, should be adopted The recommendation from this report is that the Standard
into the LRFD specifications for the purpose of qualifying Test Method for the Bond of Prestressing Strands be incor-
strand for use in prestressed concrete structures. (Please note porated into the AASHTO LRFD Bridge Design Specifications
that the Standard Test Method for the Bond of Prestressing in Section 5.4.4. The LRFD specifications text should state
Strands is also known as the Standard Test for Strand Bond.) that the material supplier is required to provide certification
There are two issues that require resolution in order to adopt that the bonding ability of the prestressing strand is accept-
the Standard Test Method for the Bond of Prestressing Strands. able for use in pretensioned and prestressed concrete appli-
First, the repeatability of the test procedure must be clearly cations and that the Standard Test Method for the Bond of
shown. Second, minimum threshold values for bond perfor- Prestressing Strands “shall be permitted” to provide accept-
mance need to be established. Both items are addressed below. ability of the prestressing strand product.
4.3.1.1 Standard Test Method for the Bond 4.3.1.2 Minimum Acceptance Value for
of Prestressing Strands Strand Bond
Section 3.2 in this report addresses the repeatability and Minimum acceptance values for strand bond are incorpo-
reproducibility of the Standard Test Method for the Bond of rated into the Standard Test Method for the Bond of

112
Prestressing Strands in its current form. The minimum values Strand D at embedment lengths of either 72 in. or 88 in. Bond
were determined by analysis of data measured from both failures occurred when web shear cracking in I-shaped beams
transfer length and development length beams. Development propagated through the transfer zones of Strand D. These
length beams were tested with three different 0.5-in. strands results from the tests on I-shaped beams indicate that bond
for their ability to develop the tension force necessary to sup- performance of Strand D is inadequate.
port flexural failures in the prescribed development length. Comparison of data in the two tables also shows that
Strand A had a NASP Bond Test value of 20,950 lb. Strand B Strands A and B had superior bond performance when com-
had a NASP Bond Test value of 20,210 lb. The bond values of pared with Strand D. The data in the tables show that the cor-
Strands A and B are contrasted with Strand D, which had a rect threshold value for the NASP Bond Test lies somewhere
NASP Bond Test value of 6,890 lb. From the testing that was between the NASP Bond Test value for Strand D and the
done and discussed in Chapter 3 of this report, it can be de- value for Strands A and B.
termined that beams made with Strands A and B were able to The data from NASP Round III research, also discussed in
develop their nominal flexural capacity in embedment lengths Chapter 3, provide additional data points for strands with
much shorter than the current requirements for strand devel- varying bonding properties (Russell and Brown 2004). In
opment length. Conversely, tests on beams made with Strand these tests, four different 0.5-in.-diameter strands were used
D indicate that the bond-ability of this strand is marginal for in a testing program that used the NASP Bond Test (August
the rectangular beams and insufficient for the I-shaped beams 2001) to assess the bonding properties of the strands. The
with respect to the strand’s ability to satisfy the current and strands were also cast into beams where transfer lengths and
recommended design provisions for development length. development lengths were measured. Tables 3.31 through
NCHRP Project 12-60 testing indicates that the minimum 3.33 summarize the results from the Round III testing. Strand
threshold number resides somewhere between 6,890 lb and FF from the NASP Round III report is the same as Strand D
20,210 lb, and that the strand with a bond value of 6,890 lb was in NCHRP Project 12-60. The data from NASP Round III
acceptable in some of the beams but not in others. testing match the data from NCHRP Project 12-60, where
Table 3.28 reports the results of development length tests beams made with Strand D failed in flexure at 73 in. of em-
on rectangular beams made with Strand D. From the table, bedment but sometimes failed in bond at 58 in. The data
the following can be seen: from NASP Round III also include testing performed on
Strand HH, which had a NASP Bond Test value of 10,700 lb.
1. For concrete strengths with release strengths as low as Beams made with Strand HH failed in flexure at both 73 in.
4 ksi, Strand D was developed in the AASHTO-prescribed and 58 in., except for one bond failure at an embedment
development length of 73 in.; length of 58 in. In this bond failure, the beam achieved nearly
2. For the same beams, bond failures regularly occurred at 90 percent of its fully developed nominal flexural capacity,
58 in.; and exhibited significant ductility, and achieved a flexural
3. Bond performance as measured by strand development strength in excess of that calculated for the beam considering
improved dramatically with increases in concrete the bilinear stress curves now found in the AASHTO LRFD
strength. Bridge Design Specifications and the 318 Code. The perfor-
mance of beams made with Strand FF (NCHRP Project 12-60
Similarly Table 3.29 reports the results from development Strand D) and Strand HH provides important data points
length tests on rectangular beams made with Strands A and in determining the minimum acceptance values for the
B. From the table, the following can be seen: Standard Test Method for the Bond of Prestressing Strands.
1. Flexural failures were reported on all concrete strengths Therefore, from these test results that incorporate the find-
and at all development lengths, and ings of the NASP Round III tests with those of NCHRP Proj-
2. Strand A was able to develop adequate tensile strength in ect 12-60 testing, it is recommended that the minimum
only 46 in. of embedment for concrete strengths as low as threshold shall be 10,500 lb for 0.5-in. strands. In other words,
6,180 lb at release and 8,500 lb at design. the performance of Strand HH would be minimally acceptable,
and the value of the NASP Bond Test of 10,700 lb can be
When considering whether the performance of Strand D is rounded to 10,500 lb for simplification. Further, the data sup-
adequate for development length, it is important to also con- port the overall conclusion that the development length per-
sider the development length test results from the I-shaped formance improves for strand with improving NASP Bond Test
beams. The results from the I-shaped beams are reported in values and therefore support the recommendation that the
Table 3.26. In this table, Beams ID-6-5-1-N, ID-6-5-1-S, and Standard Test Method for the Bond of Prestressing Strands be
ID-10-5-1-S all failed in bond. All three of these tests had adopted into the AASHTO LRFD Bridge Design Specifications.

113
Please note the protocols described in the Standard Test found in Figure 3.9 corresponds with a NASP Bond Test
Method for the Bond of Prestressing Strands require that a value (in concrete) of about 6,600 lb. Strand D, tested in
single bond test consist of six individual tests and that the re- mortar using the Standard Test Method for the Bond of
ported “Bond Test Value” is the average value from those six Prestressing Strands, had an NASP value of 6,890 lb. In con-
tests. Therefore, the 10,500-lb recommended minimum crete with a strength of 10 ksi, Strand D had a NASP Bond
threshold is an average value in conformance with the proto- Test value of nearly 12,000 lb. The data from the specimens
cols found in the Standard Test Method for the Bond of in concrete support the conclusion that increasing concrete
Prestressing Strands. In addition to the requirement for a strength improves the bond between strand and concrete.
minimum average Bond Test Value, the test includes a re- Figures 3.12 and 3.13 illustrate the results from all of the
quirement for the minimum value for the lowest value of the NASP Bond Tests in concrete versus concrete strength on four
set of six individual tests. This second criteria is established to different strand samples. One of the remarkable features of
avoid excessive variations in strand bond quality within the this dataset is that it includes data from both 0.5-in. and
same sample of strand. The requirement for the minimum 0.6-in. strands. There are more than 20 data points repre-
single test value is 9,000 lb for 0.5-in. strands. This second sented in the figures, all cast within a wide sample of concrete
requirement effectively limits the standard deviation, or the strengths. The unifying factor is that the Standard Test
statistical variance, for strand produced that may have mod- Method for the Bond of Prestressing Strands (in concrete) was
erate bonding properties. performed on each of the samples, and each data point repre-
In the NCHRP Project 12-60 research, transfer and devel- sents the average from at least six individual tests. Further-
opment length testing was also performed on 0.6-in. strands. more, the coefficient of determination, R2, was a remarkable
The NASP Bond Test was also conducted on 0.6-in.-diameter 0.80 for this dataset, which includes two different strand sizes,
strands. The results indicated that the NASP Bond Test or the four different strand samples, and a variety of concrete mix-
Standard Test Method for the Bond of Prestressing Strands tures. The figures show clearly how the bond of concrete and
was suitable in predicting the bond behavior of 0.6-in strands. sand is directly improved by increases in concrete strength.
Therefore, the NASP Bond Test is recommended for use for This result is important in determining recommendations for
0.6-in. strands. For 0.6-in. strands, the minimum threshold transfer length and development length equations.
values are an average value of 12,600 lb and a single test min- Figures 3.12 and 3.13 show two very important things: (1)
imum of 10,800 lb. that the Standard Test Method for the Bond of Prestressing
Strands is very robust and (2) that it can be modified with
concrete to assess bond performance of strand of all sizes and
4.3.2 Transfer and Development Length
in various concrete mixtures. As illustrated in Figure 3.12, the
Expressions for Prestressing Strand
power regression between bond strength and concrete
The current ACI and AASHTO design equations for pre- strength suggests the relationship:
tensioned transfer length and development length do not in-
( NASPconcrete )
clude concrete strength as a parameter in the design equation. = 0.49139 f ci′0.51702 (4.1)
However, test results obtained during NCHRP Project 12-60 NASP
strongly suggest that the anchorage ability of the strands is The results from these tests further attest to the suitability
improved as concrete strength increases. The results from of the Standard Test Method for the Bond of Prestressing
both transfer and development length testing support the Strands as a vehicle to measure the bond between prestress-
conclusion that concrete strength is an important factor. ing strand and concrete. Further, these results support the ar-
These results are supported independently by the results from gument that the variation in transfer and development
NASP Bond Tests in varying concrete strengths. lengths should vary with the square root of concrete strength.
The Standard Test Method for the Bond of Prestressing The results also support the recommendation for the adop-
Strands was used to assess the impact of varying concrete tion of the Standard Test Method for the Bond of Prestress-
strengths on the bond between strand and concrete. In these ing Strands by demonstrating that the Standard Test Method
tests, the NASP Bond Test was modified simply by casting the for the Bond of Prestressing Strands is useful for various sizes
strand in concrete (with varying concrete strengths) instead of strands and can predict differences in bond based on vari-
of the sand-cement mortar required in the Standard Test ations in concrete strength.
Method for the Bond of Prestressing Strands protocols. Fig- Testing showed that it is possible to describe the relation-
ure 3.9 illustrates the data collected from the NASP Bond Test ship between concrete strength and bond strength using the
performed on Strand D, in concrete, and the effects of vary- correlations found directly from the NASP Bond Tests in
ing concrete strengths. The figure shows that at a concrete concrete. In those tests, it was found that bond strength var-
strength of about 4.5 ksi, the prediction curve for Strand D ied nearly in proportion with the square root of concrete

114
strength. It can be observed from the data shown in Figures to include the effects of concrete strength in the design ex-
3.12 and 3.13 that the bond strength of concrete is almost di- pression for development length.
rectly proportional to the square root of concrete strength. Current expressions for development length (both ACI and
The linear regression shown in Figure 3.13 gives the follow- AASHTO) were developed from the addition of the transfer
ing relationship: length and a “flexural bond length.” This approach is supported
by past and current research. Therefore, it is recommended to
( NASPconcrete )
= 0.51 f ci′ (4.2) continue with the approach of splitting the development length
NASP into a transfer length component and a flexural bond length
Testing also demonstrated that both transfer length mea- component. Both components, however, are affected by the
surements and development length requirements were short- concrete strength. Following is an outline of the rationale for
ened as concrete strength increased. Figures 3.24 through developing the recommendation for development length:
3.32 show that for all sizes of strand and for all the varying
qualities of strand bond, the transfer lengths shorten as con- 1. The current AASHTO development length equation can
crete release strengths increase. Therefore, the recommended be used to adequately predict required development
code equation includes the square root of concrete strength lengths for “normal strength concrete” with release
at release in order to be consistent with testing results that strengths in the range of 4 ksi to 6 ksi, provided that the
show shorter transfer lengths with higher concrete strengths. strand itself is qualified by the Standard Test Method for
The transfer length recommendation provides a transfer the Bond of Prestressing Strands. The results presented in
length equal to the current design expression of 60 strand di- Chapter 3 of this report demonstrate flexural failures at
ameters when the concrete release strength is 4 ksi. Using the embedment lengths of 73 in. for all 0.5-in. strands tested
recommended expression, the transfer length shortens as in this research program. The embedment length of 73 in.
the concrete release strength increases. Additionally, the ex- corresponds to 100 percent of the current code provision
pression provides a minimum limit of 40 strand diameters. for development length for these specimens. The results
Thus, the recommended expression for transfer length ex- included tests on beams made with concrete strength at
pression is the following: release of approximately 4 ksi and approximately 6 ksi
at the time of the beam test.
120 2. The data demonstrate that development length require-
lt = db ≥ 40db (4.3)
fci′ ments are shortened as concrete strength increases.
3. The required development length calculated from the cur-
where
rent code provisions is approximately 150 db, although
lt = transfer length (in.),
under the 2004, 3rd edition of the AASHTO LRFD Bridge
f c′i = release concrete strength (ksi), and;
Design Specifications, some variations will exist due to
db= diameter or prestressing strand (in.).
variations in strand stressing, beam geometry, and subse-
If the concrete release strength is 4 ksi, then the equation quent variations in computed prestress losses. Only a few
results in a transfer length of 60 db. As release strength in- specimens contained prestressed strands at the tops of the
creases, the transfer length decreases. The limit of 40db cross sections. For these, beam tests were conducted for
ensures that all designs consider transfer lengths of some rea- development length. Therefore, no comment can be made
sonable value. The recommendation effectively limits an regarding the “top bar effect” for prestressing strands.
additional decrease in transfer length from concrete release Discussions regarding the κ factor are not included, al-
strengths greater than 9 ksi. This minimum limit on transfer though the testing demonstrates that the κ factor should be
length is consistent with the testing conducted as part of this discarded for all sizes of strand. Instead all strands can be
research in which the highest release strength achieved on qualified for bond through the Standard Test Method for
rectangular beams was 9.7 ksi. the Bond of Prestressing Strands.
The current ACI and AASHTO design equations for de- 4. Using a transfer length of approximately 60 db, and a de-
velopment length do not include concrete strength as a velopment length of approximately 150 db, the flexural
parameter. As with the results on transfer length, results from bond length must be approximately 90 db.
development length testing strongly suggest that the bond-
ability of the strands is improved as concrete strength Thus, the development length expression can then be
increases. The experimental results clearly demonstrate that written as
the required development length is shortened as concrete
225db
strength increases; higher concrete strength results in shorter ld = lt + (4.4)
development length requirements. Therefore, it is important fc′

115
where In contrast to the results from Strand D, Figure 3.50 plots

ld = development length, (in.), results from development length tests on Strands A/B versus
lt = transfer length, (in.), the proposed development length equation. The chart clearly
db = nominal diameter of the prestressing strand, (in.), and shows that all of the beams tested resulted in flexural failures
f c′ = design concrete strength, (ksi). except for one single shear failure that occurred in an
I-shaped beam. Thus, Strands A/B, high performers with an
If a concrete design strength, f c′, of 6 ksi is seen as a reason-
NASP Bond Test value in excess of 20,000 lb, can develop
able approximation for “normal concrete strength,” then a
adequate tension force even in relatively short distances.
coefficient can be computed to correspond with a flexural
In Figure 3.52, the results obtained during NASP Round III
bond length of 90 strand diameters. That coefficient is 225
with Strand HH are shown. Strand HH had a NASP Bond
(225/ 6 ≈ 90).
Test value of 10,700 lb. The chart shows that in the beams
A minimum value is also recommended for the develop-
with two strands, flexural failures occurred at 73 in. (the
ment length expression. The recommended expression for
AASHTO development length expression) and 58 in. The
development length, therefore, is based on a limiting concrete
chart indicates that one bond failure occurred at an embed-
strength of approximately 14 ksi, which is slightly less than
ment length of 58 in. on a single strand beam. An embedment
the maximum concrete strength attained in beams tested in
length of 58 in. corresponds to about 80 percent of the
the research program (14.9 ksi). Thus, the recommended
AASHTO requirement for development length. So in fact, a
development length equation is as follows:
bond failure is the expected result. These beam test results
⎡ 120 225 ⎤ from Round III provide important support to the recom-
ld = ⎢ + ⎥ db ≥ 100db (4.5)
mendation that 10,500 lb should be the minimum average
⎣ f ci′ f c′ ⎦
bond value for acceptance of prestressing strand. Clearly,
The proposed development length equation is plotted Strand HH performed adequately when the ACI and
against the development length test results in Figures 3.49 to AASHTO development lengths are provided.
3.51. The result of each beam test, whether flexural failure or Finally, shown in Figure 3.51 are the results of development
bond failure is plotted on a chart showing concrete strength length tests on the 0.6-in. strand, Strand A6. Strand A6 demon-
versus embedment length. The curves representing the rec- strated good bond performance with a NASP Bond Test value
ommended design equations for development length are also of 18,290 lb. That NASP Bond Test value is comparable with
shown on each of the charts. Note that the development the recommended minimum average value for 0.6-in. strands
length expression is now dependent on the concrete strength. of 12,600 lb. Figure 3.51 indicates that rectangular beams made
For the purpose of computing the values within the equation with Strand A6 experienced bond failures at an embedment
and charting the results, the release strength is taken as 66.7 length of 58 in., for concrete strengths of about 8 ksi and lower.
percent of the design strength. Note that in the case of 0.6-in. strand these bond failures oc-
Figure 3.49 shows the results of development length tests curred when the embedment length provided was only 66 per-
on Strand D. Strand D demonstrated below average bond cent of the development length required. At an embedment
performance with a relatively low NASP Bond Test value of length of 73 in. or so, Strand A6 was able to achieve adequate
6,890 lb. Strand D also had measurably longer transfer tension for all concrete strengths tested.
lengths than Strands A/B. Figure 3.49 indicates that beams In summary, the results on beam tests clearly demonstrate
made with Strand D experienced bond failures in rectan- support for the proposed development length expression for
gular beams with embedment lengths of 58 in. for concrete 0.5-in. and 0.6-in. strand. The inclusion of concrete strength
strengths of about 8 ksi and lower. Note that in Figure 3.49, is an important parameter that should be included in the
the flexural failures are represented by open symbols (tri- design expressions for transfer and development lengths.
angles for R-beams and diamonds for I-Beams) and the
bond failures are represented by solid symbols. Figure 3.49
4.3.2.1 Additional Requirements for the Use
clearly shows that as concrete strength increased, Strand D
of Debonded Strands
was able to move from bond failures to flexural failures a
58-in. embedment length. Furthermore, Figure 3.49 indi- Past research includes some behavioral models that more ac-
cates that I-beams made with Strand D failed in bond at curately describe the behavior of beams containing debonded
embedment lengths where the proposed design equation strands. Russell and Burns (1994) and Russell et al. (1994) de-
would predict adequate development. Therefore, Figure scribed the behavior of beams containing debonded strands.
3.49 shows that Strand D, with an NASP Bond Test value of Their research indicates that some additional provisions limit-
only 6,890 lb, does not provide adequate bond-ability with ing the overall length of the debonded strands should be incor-
concrete. porated into the LRFD Specifications. Early testing performed

116
by Rabbat et al. (1979) and Kaar and Magura (1965) also sup- provisions and the AASHTO LRFD Bridge Design Specifications
port a behavioral approach toward limitations on debonded (2004). In the 1995 edition of the 318 Code, the procedures for
strands. The experimental programs were reviewed, and rec- calculating development lengths for deformed bars and de-
ommendations are made to amend the current code provisions. formed wire in tension were extensively modified (ACI 1995).
The experimental program did not incorporate testing on The changes resulted in an increase in the development lengths
beams containing debonded strands. Recommended changes for closely spaced bars and bars with small covers. The basic
are based on experimental work already performed and identi- development length was modified to reflect the influence of
fied within the existing body of knowledge. cover, spacing, transverse reinforcement, casting position, type
The current AASHTO LRFD Bridge Design Specifications of aggregate, and epoxy coating. The basic development
provide limitations on debonding strands. There are limita- lengths remained essentially the same as they were in the 1971
tions on the total percentage of strands that can be debonded edition of the ACI Code and the AASHTO LRFD Bridge Design
and the total number of debonded strands per row, as well as Specifications, with a few exceptions. One exception is the equa-
limitations on which strands within the cross section can be tion for the basic development length in tension for No. 18
debonded. In addition to these existing limitations, the bars. This equation was revised on the basis of a review of avail-
authors recommend three more: able test results on large bars. This change resulted in an
increase of 12 percent over the values given by the current
1. Where pretensioned beams are not simply supported, AASHTO LRFD Bridge Design Specifications for bars of the
debonding shall not be permitted except where it can be same size. Also, the top bar factor, which is 1.3 in the 318 Code,
shown that cracking will not occur through the regions is 1.4 in the LRFD specifications. Another important change to
where debonding is placed nor through the transfer zones the 318 Code, introduced in 1989, was to limit the fc′ value
of debonded strands. to a maximum of 100 psi, regardless of the compressive design
2. Debonding shall be limited in length from the end of a strength of the concrete (ACI 1989). This limitation meant that
member to a distance equal to 0.15 times the span mea- development lengths would no longer decrease with concrete
sured from center of bearing to center of bearing. strengths greater than 10 ksi. It was noted that research on de-
3. Debonding shall be limited in number and in length to velopment of bars in high-strength concretes was not sufficient
sections where debonding is required to meet the require- to substantiate a reduction beyond the limit imposed. This is
ments of Article 5.9.4.1. At sections where debonding is also the reason given for the limitation imposed in Section
not required to meet the requirements of Article 5.9.4.1, 5.4.2.1 of the current LRFD specifications.
debonding shall not be permitted. In 1977, the 318 Code provisions for tension lap splices of
deformed bars and deformed wire encouraged the location
of splices away from regions of high tensile stresses and to
4.3.3 Mild Reinforcement Development
places where the area of steel provided at the splice location
and Splice Lengths and Anchorage
is at least twice that required by analysis. A lap splice of any
with Standard Hooks in Tension
portion of the total area of steel in regions where (As pro-
In Section 12.1.2 of the 318 Code (ACI 2005) an upper vided/As required) is less than 2.0 had to be at least 1.3 times
limit of 100 psi on the fc′ term in the anchorage and devel- the development length of the individual bar in tension
opment length provisions is imposed; in Section 5.4.2.1 of the (Class B splice) in length. If more than one-half of the rein-
Interim 2008 AASHTO LRFD Bridge Design Specifications, it forcement was to be spliced in such regions, lap splices had
is stated that design concrete strengths above 10 ksi shall be to be at least 1.7 times the development length of the indi-
used only when allowed by specific articles or when physical vidual bar (Class C splice) in length. Class A splices in which
tests are made to establish the relationships between concrete the length of bar is equal to the development length of the
strength and other properties. The experimental program on individual bar were only permitted in regions where (As
mild reinforcement described in NCHRP Project 12-60 was provided/As required) is less than 2.0, and no more than
designed to determine, in conjunction with the data already 25 percent of the total area is spliced within one lap length.
available in the literature, whether these limitations can be re- These same provisions are currently in the AASHTO LRFD
moved for concrete compressive strengths up to 15 ksi. Bridge Design Specifications. The Class C splice has been re-
The 1971 provisions in the 318 Code (with slight modifica- moved from the 318 Code. It must be noted that the splice
tions because fy and f c′ are stated in terms of ksi and the intro- factor is associated with the potential mode of failure when
duction of epoxy-coated bar factors) are the current provisions multiple bars are spliced at the same location and does not
for development and splice length of mild reinforcing bars in relate to the actual strength of the spliced bar.
tension in the AASHTO LRFD Bridge Design Specifications. The work by Treece and Jirsa (1989) is the basis of the de-
Thus, there are differences between the current 318 Code velopment length modification factors for epoxy-coated bars

117
in the 318 Code and the AASHTO LRFD Bridge Design Speci- lightweight concrete, db is the bar diameter, Ktr is 40 Atr/sn and
fications. In both, development length is multiplied by a factor represents the contribution of confining reinforcement
of 1.5 for the epoxy-coated bars with a cover of less than 3db across potential splitting planes, where Atr is the area of trans-
or clear spacing between the bars that is less than 6db. Devel- verse reinforcement within the spacing s that crosses the
opment length is multiplied by 1.2 for other cases. In either potential splitting plane, s is the spacing of stirrups and n is
case, the product of the top-bar factor and epoxy-coating fac- the number of bars being spliced or developed along the
tor should not exceed 1.7. The 1.2 factor was selected based on plane of splitting; cb is the spacing or cover dimension using
the work of Johnston and Zia (1982). DeVries, Moehle, and the smaller of either the distance from the center of the bar or
Hester (1991), Hadje-Ghaffari, Darwin, and McCabe (1991), wire to the nearest concrete surface or one-half the center-to-
and Hadje-Ghaffari et al. (1994) found the current maximum center spacing of the bars being developed. The ratio of (cb+
of 1.7 for the product of the top-bar factor and epoxy-coating Ktr)/db is not to be taken greater than 2.5. However, ld shall not
factor to be too conservative and recommended a value of 1.5. be less than 12 in. In addition, when calculating anchorage
Cleary and Ramirez (1993), based on experimental observa- length requirements for tension lap splices, these should be as
tions on slab-type specimens, noted that since the experimen- required for Class A or B splices but not less than 12 in., where
tal data on splitting type failures included only up to maxi- a Class A splice is 1.0 ld and a Class B splice is 1.3ld.
mum cover to bar diameter ratio of 2.67 due to the increase of Development length for deformed bars in tension terminat-
rib-bearing forces with epoxy-coated reinforcement, the limit ing in a standard hook (as per Section 7.1 of the 318 Code), ldh,
of 3 used in the 318 Code for transition between splitting and is determined using the equation below, found also in Section
pull-out failures should be examined experimentally. 12.5.2 of the 318 Code and applicable modification factors of
The 1983 provisions in the 318 Code for development of Section 12.5.3, but ldh shall not be less than the smaller 8db and
bars in tension terminated with standard hooks were a major 6 in. as indicated in Section 12.5.1 of the 318 Code (ACI 2005).
departure from the 1977 version of the 318 Code as hooked- ⎛ 0.02 ψ e λf y ⎞
ldh = ⎜ db (4.7)
fc′⎟⎠
bar anchorage provisions were uncoupled from provisions
⎝
for straight-bar development. In the 1983 version of the 318
Code, the hooked-bar embedment length was measured from ψe is a coating factor taken as 1.2 for epoxy-coated reinforce-
the critical section to the outside end or edge of the hook. The ment; λ is a factor reflecting the lower tensile strength of
development length of the hooked bar was calculated as the lightweight concrete taken as 0.75 (for other cases, these two
product of a basic development length and appropriate mod- factors are taken equal to 1.0). Other parameters include the
ification factors. In the 1995 edition of the 318 Code, a factor bar diameter of the hooked bar, db, concrete compressive
of 1.2 was introduced in the hooked anchorage requirements strength, f c′, in psi (the square root of the concrete compres-
(ACI 1995). The requirements in the 3rd edition of the sive strength shall not exceed 100 psi as per Section 12.1.2 of
AASHTO LRFD Bridge Design Specifications (2004) for an- the 318 Code). The modification factors of Section 12.5.3 are
chorage of hooked bars in tension are the same as those in the all less than 1.0 and thus reduce the calculated length on the
2005 318 Code (2005). basis of cover, presence of ties where the first tie encloses the
Equation 12-1 of the 318 Code (ACI 2005), used for calcu- bent portion of the hook within 2db of the outside of the bend,
lating tension splice and the development length require- and when anchorage or development for specified minimum
ment, is the following: yield strength, fy, is not specifically required:
⎡ ⎤
⎢ 3 f y ψt ψe ψ sλ ⎥ 1. For #11 bar and smaller hooks with side cover (normal to
ld = ⎢ ⎥ db (4.6) the plane of the hook) not less than 2.5 in. and for 90-deg
⎢ 40 fc′ ⎛ cb + K tr ⎞ ⎥
⎢⎣ ⎜⎝ ⎟ hook with cover on bar extension beyond hook not less
db ⎠ ⎥⎦
than 2 in, the factor is 0.7;
In Equation 12-1, ψt is a reinforcement location factor of 2. For 90-deg hooks of #11 and smaller bars that are either en-
1.3 to reflect the adverse effects of the top reinforcement cast- closed within ties or stirrups perpendicular to the bar being
ing position; ψe is a coating factor—1.5 with cover less than developed, spaced not greater than 3db along ldh or enclosed
3db or clear spacing less than 6db and 1.2 for all other cases. within ties or stirrups parallel to the bar being developed,
The product of ψt and ψe need not be taken greater than 1.7. spaced not greater than 3db along the length of the tail
The parameter ψs is a reinforcement size factor—0.8 for No. extension of the hook plus bend, the factor is 0.8;
6 bars and smaller and 1.0 for all other cases (the square root 3. For 180-deg hooks of #11 and smaller bars that are en-
of the concrete compressive strength shall not exceed 100 psi closed within ties or stirrups perpendicular to the bar being
as per Section 12.1.2). Other parameters are defined as developed, spaced not greater than 3db along ldh, the factor
follows: λ is a factor reflecting the lower tensile strength of is 0.8; and

118
4. Where anchorage or development for fy is not specifically Beam splice specimens with bottom cast bars were not eval-
required, reinforcement in excess of that required by uated in this study. The ACI Committee 408 (2003) indicated
analysis, the factor is (As required/As provided). that the current approach in the 318 Code overestimated the
bar force at failure in many specimens with bottom bars that
are available in the ACI Committee 408 database, especially
4.3.3.1 Development and Splice Length
for specimens with concrete compressive strengths greater
of Uncoated and Coated Bars
than 10 ksi. The ACI Committee 408 proposed a modified ex-
Article 5.11 of the 3rd edition of the AASHTO LRFD Bridge pression for development and splice length in addition to the
Design Specifications (2004), “Development and Splices of removal of the bar size factor to address this issue. In the eval-
Reinforcement,” contained provisions for development uation of test data conducted under NCHRP Project 12-60,
length of reinforcement that were essentially the same as the researchers found that the use of a bottom cast modifica-
those included in editions of the 318 Code up to the 1989 tion factor of 1.2 for uncoated bars anchored in concrete with
edition. The provisions of the 318 Code were extensively compressive strengths greater than 10 ksi appeared to address
modified for the 1995 edition with a view to formulating a the safety concerns raised by ACI Committee 408. This factor
more “user friendly” format while maintaining the same gen- would not be needed for bottom cast epoxy-coated bars (be-
eral agreement with professional judgment and research re- cause of the single modification factor of 1.5) or for uncoated
sults. Tests on splices of uncoated bars (Azizinamini et al. top bars. This approach could be used as an alternative to the
1993, 1999a) have indicated that in the case of high-strength approach suggested by ACI Committee 408. The researchers
concrete some minimum amount of transverse reinforce- note that additional testing of bottom cast uncoated splices is
ment is needed to ensure adequate ductility out of the splice justified with higher strength concretes.
at failure. Based on these tests, a proposed modification
(Azizinamini et al. 1999b) to the 1999 318 Code calls for the
4.3.3.2 Anchorage in Tension of Uncoated
determination of a basic, straight development length for bars
and Coated Mild Reinforcement Using
in tension, without including the presence of transverse rein-
Standard Hooks
forcement, together with a minimum area of transverse steel
in the form of stirrups, Asp, crossing potential splitting planes. Article 5.11.2.4 of the AASHTO LRFD Bridge Design Specifi-
In these studies, over 70 specimens were tested with concrete cations was verified for high-strength concrete in the proposed
compressive strengths ranging between 5 ksi and 16 ksi. The experimental Work Plan for NCHRP Project 12-60 with the
experimental work conducted in NCHRP Project 12-60 exception of the lightweight aggregate factor. Based on the analy-
aimed to fill the gaps in the existing data to extend the appli- sis of tests conducted during NCHRP Project 12-60 (21 full-scale
cability of the LRFD provisions for development and splice tests of hooked-bar anchorages) and of tests of additional spec-
length of uncoated and epoxy-coated bars (ASTM A 775) imens in the literature, it is possible to support the extension of
in tension to normal-weight concrete with compressive the approach in the 318 Code (ACI 2005) provision for anchor-
strengths up to 15 ksi. age of bars terminated with standard hooks, black and epoxy-
Based on the observations from tests conducted during coated, to normal-weight concrete with concrete compressive
NCHRP Project 12-60 on 18 top-cast beam splice specimens strengths of up to 15 ksi, with the following modifications:
and the examination of an extensive database of previous
tests compiled by ACI Committee 408 (presented in previous • A minimum amount of transverse reinforcement (at least
chapters of this report), it is proposed to extend the AASHTO #3 U bars at 3db spacing) should be provided in the anchor-
LRFD Bridge Design Specifications to concrete strengths up to age length to improve the bond strength of both uncoated
15 ksi using the approach in the 318 Code (ACI 2005), with and epoxy-coated No. 11 and larger bars terminated in a
the following exceptions: standard hook.
• A modification factor of 0.8 instead of the current factor of
• Remove the bar size factor for No. 6 and smaller bars; thus, 0.7 for No. 11 and smaller hooks with side cover (normal
ψs = 1.0 in all cases. to the plane of the hook) not less than 2.5 in. and for 90-
• Use a single factor for epoxy-coated bars of 1.5 regardless degree hooks with cover on bar extensions beyond the
of the cover-to-bar diameter ratio. hook of not less than 2 in.

119
References
AASHTO (1998). AASHTO LRFD Bridge Design Specifications, 2nd ed., Azizinamini, A., Pavel, R., Hatfield, E., and Gosh, S.K. (1999a). “Behav-
including 1999, 2000, and 2001 interim revisions. AASHTO, ior of Spliced Reinforcing Bars Embedded in High-Strength
Washington, DC. Concrete.” Structural Journal of the American Concrete Institute,
AASHTO (2004). AASHTO LRFD Bridge Design Specifications, 3rd ed., Vol. 96, No. 5, September-October, 826–835.
with 2005 and 2006 interim revisions. AASHTO, Washington, DC. Azizinamini, A., Stark, M., Roller, J.J., and Gosh, S.K. (1993). “Bond
AASHTO (2007). AASHTO LRFD Bridge Design Specifications, 4th ed., Performance of Reinforcing Bars Embedded in High-Strength
Customary U.S. Units. AASHTO, Washington, DC. Concrete.” Structural Journal of the American Concrete Institute,
Abrishami, H. H. and Mitchell, D. (1993). “Bond Characteristics of Vol. 90, No. 5, September-October, 554–561.
Pretensioned Strand.” ACI Materials Journal, Vol. 90, No. 3, May- Barnes, R.W. and Burns, N.H. (2000). “Anchorage Behavior of 15.2
June, 228–235. mm (0.6 in) Prestressing Strand in High Strength Concrete.”
ACI (1963). Building Code Requirements for Reinforced Concrete Presentation at PCI/FHWA/FIB International Symposium on
(ACI 318-63). American Concrete Institute, Redford Station, High Performance Concrete, September 25–27, Orlando, FL,
Detroit, MI. 484–493.
ACI (1989). Building Code Requirements for Reinforced Concrete (ACI Buckner, C.D. (1994). An Analysis of Transfer and Development
318-89) and Commentary (ACI 318R-89). American Concrete Lengths for Pretensioned Concrete Structures, FHWA-RD-94-049.
Institute, Farmington Hills, MI. FHWA, U.S. DOT, Washington DC.
ACI (1995). Building Code Requirements for Structural Concrete (ACI Choi, O.C., Hadje-Ghaffari, H., Darwin, D., and McCabe S.L. (1990).
318-95) and Commentary (ACI 318R-95). American Concrete “Bond of Epoxy-Coated Reinforcement to Concrete: Bar Parame-
Institute, Farmington Hills, MI. ters.” SM Report No. 25, University of Kansas Center for Research,
ACI (1999). Building Code Requirements for Structural Concrete (ACI Lawrence, KS.
318-99) and Commentary (ACI 318R-99). American Concrete Choi, O.C., Hadje-Ghaffari, H., Darwin, D., and McCabe, S.L. (1991).
Institute, Farmington Hills, MI. “Bond of Epoxy-Coated Reinforcement: Bar Parameters.” ACI
ACI (2002). Building Code Requirements for Structural Concrete (ACI Materials Journal, Vol. 88, No. 2, 207–217.
318-02) and Commentary (318R-02). American Concrete Institute, Clark, A.P. (1946). “Comparative Bond Efficiency of Deformed Con-
Farmington Hills, MI. crete Reinforcing Bars.” ACI Journal Proceedings, Vol. 43, No. 4,
ACI (2005), ACI 318-05: Building Code Requirements for Structural December, 381–400.
Concrete and Commentary, American Concrete Institute, Farming- Cleary, D.B. and Ramirez, J.A. (1993). “Epoxy-Coated Reinforcement
ton Hills, MI. Under Repeated Loading.” ACI Structural Journal, Vol. 90, No.4,
ACI Committee 408 (1964). “A Guide for the Determination of Bond July-August, 451–458.
Strength in Beam Specimens.” Journal of the American Concrete Cooke, D.E., Shing, P.B., Frangopol, D.M. (1998), “Colorado Study on
Institute, Vol. 61, No. 2, February, 129–135. Transfer and Development of Prestressing Strand in High Perfor-
ACI Committee 408 (1966). “Bond Stress—The State of the Art.” ACI mance Concrete Box Girders,” University of Colorado Research
Journal Proceedings, Vol. 63, No. 11, November, 1161–1188. Series No. CU/SR-97/4, April, 168 pp.
ACI Committee 408 (1970). “Opportunities in Bond Research.” ACI Cousins, T.E., Johnston, D.W., and Zia, P. (1990). “Transfer and Devel-
Journal Proceedings, Vol. 67, No. 11, November, 857–867. opment Length of Epoxy-Coated and Uncoated Prestressing
ACI Committee 408 (2003). “Bond and Development of Straight Rein- Strands.” PCI Journal, Vol. 35, No. 3, July-August, 92–106.
forcing Bars in Tension (408R-03). Technical Documents, Ameri- Cousins, T.E., Stallings, J.M, and Simmons, M.B (1993). “Effect of
can Concrete Institute, Farmington Hills, MI. Strand Spacing on Development of Prestressing Strands,” Research
Azizinamini, A., Darwin, D., Eligehausen, R., Pavel, R., and Gosh, S.K. Report, Alaska Department of Transportation and Public Facilities,
(1999b). “Proposed Modifications to ACI 318-95 Tension Devel- August 1993.
opment and Lap Splice for High-Strength Concrete.” Structural Darwin, D. and Graham, E.K. (1993a). “Effect of Deformation Height
Journal of the American Concrete Institute, Vol. 96, No. 6, November- and Spacing on Bond Strength of Reinforcing Bars.” SL Report
December, 922–926. 93-1, University of Kansas Center for Research, Lawrence, KS.

120
Darwin, D. and Graham, E. K. (1993b). “Effect of Deformation Height Hester, C. J., Salamizavaregh, S., Darwin, D., and McCabe, S.L. (1993).
and Spacing on Bond Strength of Reinforcing Bars.” ACI Structural “Bond of Epoxy-Coated Reinforcement: Splices.” ACI Structural
Journal, November-December, Vol. 90, No. 6, 646–657. Journal, Vol. 90, No. 1, 89–102.
Darwin, D., Tholen, M.L., Idun, E.K., and Zuo, J. (1996). “Splice Hyett, A.J., Dube, S., and Bawden, W.F. (1994). “Laboratory Bond
Strength of High Relative Rib Area Reinforcing Bars.” ACI Struc- Strength Testing of 0.6” 7-Wire Strand from 7 Different Manufac-
tural Journal, Vol. 93, No. 1, January-February, 95–107. turers,” Final Report. Department of Mining Engineering, Queen’s
Deatherage, J.H., Burdette, E.G., and Chew, C.K. (1994). “Development University, Kingston, Ontario, November.
Length and Lateral Spacing Requirements of Prestressing Strand Jacob J. (1998). “Improved Shear Reinforcement Details for the End
for Prestressed Concrete Bridge Girders.” PCI Journal, Vol. 39, Regions of Prestressed Concrete Bridge Girders,” Masters Thesis.
No. 1, January-February, 70–83. University of Oklahoma, Norman, OK.
DeVries, R.A., Moehle, J.P., and Hester, W., 1991. “Lap Splice of Plain Jirsa, J.O. and Breen, J.E. (1981). Influence of Casting Position and Shear
and Epoxy-Coated Reinforcements: An Experimental Study Con- on Development and Splice Length—Design Recommendation,
sidering Concrete Strength, Casting Position, and Anti-Bleeding Research Report No. 242-3F. Center for Transportation Research,
Additives,” UCB/SEMM-91/02 Structural Engineering Mechanics The University of Texas at Austin, Austin, TX.
and Materials, University of California, Berkeley, CA, January. Johnston, D.W. and Zia, P. (1982). Bond Characteristics of Epoxy-Coated
Ferguson, P.M. and Thompson, J.N. (1965). “Development Length of Reinforcing Bars, FHWA-NC-82-002. FHWA, U.S. DOT, Washing-
High Strength Reinforcing Bars.” ACI Journal Proceedings Vol. 62, ton, DC.
No. 1, January, 71–94. Kaar, P.H., LaFraugh, R.W., and Mass, M.A. (1963). “Influence of
Furlong, R.W., Fenves, G.L., and Kasl, E.P. (1991). “Welded Structural Concrete Strength on Strand Transfer Length.” PCI Journal, Vol. 8,
Wire Reinforcement for Columns.” ACI Structural Journal, Vol. 88, No. 5, October, 47–67.
No. 5, September-October, 585–591. Kaar, P.H. and Magura, D.D. (1965). “Effect of Strand Blanketing on
Griezic, A., Cook, W.D., and Mitchell, D. (1994). “Tests to Determine Performance of Pretensioned Girders.” Portland Cement Association
Performance of Deformed Welded-Wire Fabric Stirrups.” ACI Research and Development Laboratory Bulletin, December, 24–35.
Structural Journal, Vol. 91, No. 2, March-April, 211–220. Kahn, L.F., Dill, J.C., and Reutlinger, C.G. (2002). “Transfer and Develop-
Gross, S.P. and Burns, N.H. (1995). Research Report 580-2: Transfer and ment Length of 15-mm Strand in High Performance Concrete Gird-
Development Length of 15.2 mm (0.6 in) Diameter Prestressing ers.” Journal of Structural Engineering, Vol.128, No.7, July, 913–921.
Strand in High Performance Concrete—Results of the Hoblitzell- Kaufman, M.K. and Ramirez, J.A. (1988). “Reevaluation of the Ultimate
Buckner Beam Tests. The University of Texas at Austin, Austin, TX. Shear Behavior of High Strength Concrete Prestressed I-Beams.”
Grundhoffer, T., Menis, P.A., French, C.W., and Leon, R. (1998). ACI Structural Journal, Vol. 85, No. 3, May-June, 295–303.
“Bond of Epoxy-Coated Reinforcement in Normal and High- Leonhardt, F. and Teichen, K. (1972). “Druck-Stosse von Be-
Strength Concrete.” ACI Special Publication 180-2, 261–298. wehrungsstaben.” Deutscher Ausschuss fur Stahlbeton Bulletin No.
Guimaraes, G.N., Kreger, M.E., and Jirsa, J.O. (1992). “Evaluation of 222, W. Ernst and Sohn, Berlin, FRG, 1–53.
Joint-Shear Provisions for Interior Beam-Column-Slab Connec- Logan, D.R. (1997). “Acceptance Criteria for the Bond Quality of
tions Using High Strength Materials.” ACI Structural Journal, Vol. Strand for Pretensioned Prestressed Concrete Applications.” PCI
89, No. 1, January-February, 89–98. Journal, Vol.42, No.2, March-April, 52–90.
Hadje-Ghaffari, H., Darwin, D., and McCabe, S.L. (1991). “Effects of Maeda, M., Otani, S., and Aoyama, H. (1991). “Bond Splitting Strength
Epoxy-Coating on the Bond of Reinforcing Steel to Concrete.” SM in Reinforced Concrete Members.” Transactions of the Japan
Report No. 28, University of Kansas Center for Research, Inc., Concrete Institute, Vol. 13, 581–588.
Lawrence, KA. Mansur, M.A., Lee, C.K., and Lee, S.L. (1987). “Deformed Wire Fabric
Hadje-Ghaffari, H., Choi, O.C., Darwin, D., and McCabe, S.L. (1994). as Shear Reinforcement in Concrete Beams.” ACI Structural Jour-
“Bond of Epoxy-Coated Reinforcement to Concrete: Cover, Cast- nal, Vol. 84, No. 5, September-October, 392–399.
ing Position, Slump, and Consolidation.” ACI Structural Journal, Marques, J.L. and Jirsa, J.O. (1975). “A Study of Hooked Bar Anchor-
Vol. 90, No. 1, January-February, 59–68. ages in Beam-Column Joints.” Journal of the American Concrete
Hamad, B.S., and Itani, M.S. (1998). “Bond Strength of Reinforcement Institute, Vol. 72, No. 5, May, 198–209.
in High-Performance Concrete: The Role of Silica Fume, Casing Mathey, R.G. and Clifton, J.R. (1976). “Bond of Coated Reinforcing
Position, and Superplasticizer Dosage.” ACI Materials Journal, Vol. Bars in Concrete.” ASCE Journal of the Structural Division, Vol. 102,
95, No. 5, September-October, 499–511. No. ST1, 215–229.
Hamad, B.S., Jirsa, J.O., and D’Abreu de Paulo, N. (1993). “Anchorage Meyer, K.F. (2002). “Transfer and Development Length of 0.6-inch
Strength of Epoxy-Coated Hooked Bars.” ACI Structural Journal, Diameter Strand in High Strength Light Weight Concrete.” Ph.D.
Vol. 90, No. 2, March-April, 210–217. dissertation, Georgia Institute of Technology, Georgia.
Hamad, B.S., and Jirsa, J.O. (1993). “Strength of Epoxy-Coated Rein- Mitchell, D., Cook, W.D., Khan, A.A., and Tham, T. (1993). “Influence
forcing Bar Splices Confined with Transverse Reinforcement.” ACI of High Strength Concrete on Transfer and Development Length of
Structural Journal, Vol. 90, No.1, 77–88. Pretensioning Strand.” PCI Journal, Vol. 38, No. 3, May-June, 52–66.
Hanson, N.W., and Kaar, P.H. (1959). “Flexural Bond Tests of Preten- Nilson, A.H. (1997). Design of Concrete Structures, 12th ed., McGraw-
sioned Prestressed Beams.” ACI Journal, Vol.55, No.7, January, Hill, New York, NY.
783–802. Orangun, C.O.; Jirsa, J.O.; and Breen, J.E. (1977). “A Reevaluation of
Hawkins, N.M. and Kuchma, D.A. (2002). “Application of the LRFD Test Data on Development Length and Splices,” ACI Journal
Bridge Design Specifications to High-Strength Structural Concrete: Proceedings, Vol. 74, No. 3, 114–122.
Shear Provisions, Phase 1,” NCHRP Project 12-56 Interim Report, Ozyildirim, C. and Gomez, J.P. (1999). High-Performance Concrete in a
Transportation Research Board of the National Academies, Wash- Bridge in Richlands, Virginia. Virginia Transportation Research
ington, DC, February. Council, Charlottesville, VA.

121
Pfister, J.P. and Mattock, A.H. (1963). “High Strength Bars as Concrete Russell, B.W., and Paulsgrove, G.A. (1999a). “NASP Strand Bond
Reinforcement, Part 5. Lapped Splices in Concentrically Loaded Testing Round One Pull-Out Tests and Friction Bond Tests of
Columns.” Journal of the PCA Research and Development Laborato- Untensioned Strand,” Final Report 99-03. The University of Okla-
ries, Vol. 5, No. 2, May, 27–40. homa, Fears Structural Engineering Laboratory, Norman, OK.
Pincheira, J. A., Rizkalla, S.H., Attiogbe, E.K. (1989). “Performance of Russell, B.W., and Paulsgrove, G.A. (1999b). “Assessing Repeatability
Welded Wire Fabric as Shear Reinforcement under Cyclic Load- and Reproducibility of the Moustafa Test, the PTI Bond Test and
ing.” ACI Structural Journal, Vol. 86, No. 6, November-December, the NASP Bond Test,” Final Report 99-04. The University of Okla-
728–735. homa, Fears Structural Engineering Laboratory, Norman, OK.
Rabbat, B.G., Kaar, P.H., Russell, H.G., and Bruce, R.N. (1979). “Fa- Russell, H.G., Miller, R.A., Ozyildirim, H.C., and Tadros, M. K. (2006).
tigue Tests of Pretensioned Girders with Blanketed and Draped Compilation and Evaluation of Results from High-Performance Con-
Strands.” PCI Journal, Vol. 24, No. 4, July-August, 88–114. crete Bridge Projects, Volume I: Final Report, FHWA-HRT-05-056.
Roberts-Wollmann, C., Cousins, T., Gomez, J., and Ozyildirim, C. FHWA, U.S. DOT.
(2000). “Use of High Performance Concrete in a Bridge Structure Sakurada, T., Morohashi, N., and Tanaka, R. (1993). “Effect of Trans-
in Virginia.” Presented at the PCI/FHWA/FIB International Sym- verse Reinforcement on Bond Splitting Strength of Lap Splices.”
posium on High Performance Concrete, September 25–27, Or- Transactions of the Japan Concrete Institute, Vol. 15, 573–580.
lando, FL. Shahawy, M.A., Issa, M., and Batchelor, B. (1992). “Strand Transfer
Rose, D.R. and Russell, B.W. (1997). “Investigation of Standardized Lengths in Full Scale AASHTO Prestressed Concrete Girders.” PCI
Tests to Measure the Bond Performance of Prestressing Strand.” Journal, Vol. 37, No. 3, May-June, 84–96.
PCI Journal, Vol. 42, No. 4, July-August, 56–80. Shing, P.B., Frangopol, D.M., McMullen, M.L., Hutter, W., Cooke,
Russell B.W. (1992). “Design Guidelines for Transfer, Development, D.E., and Leonard, M.A. (2000). “Strand Development and Trans-
and Debonding of Large Diameter Seven Wire Strands in Preten- fer Length Tests on High Performance Concrete Box Girders.” PCI
sioned Concrete,” Ph.D. dissertation, University of Texas at Austin, Journal, Vol. 45, No. 5, September-October, 96–109.
Austin, TX. Tabatabai, H. and Dickson, T.J. (1993). “The History of the Preten-
Russell, B.W. (1994). “Impact of High Strength Concrete on the Design sioned Strand Development Length Equation.” PCI Journal,
and Construction of Pretensioned Girder Bridges.” PCI Journal, Vol.38, No. 6, November-December, 64–75.
Vol. 39, No. 4, 1994, 76–89. Tessema Eden. (2006). “The Effect of High Strength Concrete on Bond-
Russell, B.W., and Brown, M.D. (2004). “Evaluation of Test Methods in ability of Prestressing Strands.” Masters Thesis, Oklahoma State
Assessing Bond Quality of Prestressing Strands,” Final Report, University, Stillwater, OK.
NASP Round III Strand Bond Testing. Oklahoma State University, Treece, R.A. and Jirsa, J.O. (1989). “Bond Strength of Epoxy-Coated
Stillwater, OK, August. Reinforcing Bars.” ACI Materials Journal, Vol. 86, No. 2, March-
Russell, B.W., and Burns, N.H. (1993). Design Guidelines for Transfer, April, 167–174.
Development and Debonding of Large Diameter Seven Wire Strands Xuan, X., Rizkalla, S., Maruyama, K. (1988). “Effectiveness of Welded
in Pretensioned Concrete Girders. Report No. 1210-5F. Center for Wire Fabric as Shear Reinforcement in Pretensioned Prestressed
Transportation Research, The University of Texas at Austin. Concrete T-Beams.” ACI Structural Journal, Vol. 85, No. 4, July-
Russell, B.W., and Burns, N.H. (1994). “Fatigue Tests on Prestressed August, 429–436.
Concrete Beams Made with Debonded Strands.” PCI Journal, Vol. Zhongguo, (J.) M., Tadros, M.K., Baishya, M. (2000). “Shear Behavior
39, No. 6, 99–106. of Pretensioned High-Strength Concrete Bridge I-Girders.” ACI
Russell, B.W., and Burns, N.H. (1996). “Measured Transfer Lengths of Structural Journal, Vol. 97, No. 1, January-February, 185–192.
0.5 in. and 0.6 in. Strands in Pretensioned Concrete.” PCI Journal, Zia, P. and Moustafa, T. (1977). “Development Length of Prestressing
Vol.41, No.5, September-October, 44–63. Strands.” PCI Journal, Vol. 22, No. 5, 54–65.
Russell, B.W., and Burns, N.H. (1997). “Measurement of Transfer Zuo, J. and Darwin, D. (1998). “Bond Strength of High Relative Rib
Lengths in Pretensioned Concrete Elements.” Journal of Structural Area Reinforcing Bars.” SM Report, No. 46, University of Kansas
Engineering, Vol.123, No.5, May, 541–549. Center for Research, Lawrence, KA.
Russell, B.W., Burns, N.H., ZumBrunnen, L.G. (1994). “Predicting the Zuo, J. and Darwin, D. (2000). “Splice Strength of Conventional and
Bond Behavior of Prestressed Concrete Beams Containing High Relative Rib Area Bars in Normal and High Strength Con-
Debonded Strands.” PCI Journal, Vol. 39, No. 5, 60–77. crete.” ACI Structural Journal, Vol. 97, No.4, July-Aug., 630–641.

122
APPENDICES
Appendix A of the contractor’s final report for NCHRP • Appendix C: Rectangular Beam Summaries—Strand D
Project 17-25 contained the figures and tables for the report, • Appendix D: Rectangular Beam Summaries—Strands A&B
and Appendix B contained the report reference list. In this • Appendix E: Rectangular Beam Summaries—Strand A
publication, these materials have been incorporated into the (0.6 in.)
text of the report. • Appendix F: I-Beam Summaries—0.5-in. Strand
Appendices C through I of the contractor’s final report are • Appendix G: I-Beam Summaries—0.6-in. Strand
available on the TRB website at http://trb.org/news/blurb_ • Appendix H: AASHTO Mxxx—Standard Test Method for
detail.asp?id=9210. Titles of Appendices C through I are the the Bond of Prestressing Strands
following: • Appendix I: NASP Test Protocols

Abbreviations and acronyms used without definitions in TRB publications:

AAAE American Association of Airport Executives
AASHO American Association of State Highway Officials
AASHTO American Association of State Highway and Transportation Officials
ACI–NA Airports Council International–North America
ACRP Airport Cooperative Research Program
ADA Americans with Disabilities Act
APTA American Public Transportation Association
ASCE American Society of Civil Engineers
ASME American Society of Mechanical Engineers
ASTM American Society for Testing and Materials
ATA Air Transport Association
ATA American Trucking Associations
CTAA Community Transportation Association of America
CTBSSP Commercial Truck and Bus Safety Synthesis Program
DHS Department of Homeland Security
DOE Department of Energy
EPA Environmental Protection Agency
FAA Federal Aviation Administration
FHWA Federal Highway Administration
FMCSA Federal Motor Carrier Safety Administration
FRA Federal Railroad Administration
FTA Federal Transit Administration
IEEE Institute of Electrical and Electronics Engineers
ISTEA Intermodal Surface Transportation Efficiency Act of 1991
ITE Institute of Transportation Engineers
NASA National Aeronautics and Space Administration
NASAO National Association of State Aviation Officials
NCFRP National Cooperative Freight Research Program
NCHRP National Cooperative Highway Research Program
NHTSA National Highway Traffic Safety Administration
NTSB National Transportation Safety Board
SAE Society of Automotive Engineers
SAFETEA-LU Safe, Accountable, Flexible, Efficient Transportation Equity Act:
A Legacy for Users (2005)
TCRP Transit Cooperative Research Program
TEA-21 Transportation Equity Act for the 21st Century (1998)
TRB Transportation Research Board
TSA Transportation Security Administration
U.S.DOT United States Department of Transportation

Transfer, Development, and Splice Length For Strand/Reinforcement in High-Strength Concrete (2008)

Uploaded by

Transfer, Development, and Splice Length For Strand/Reinforcement in High-Strength Concrete (2008)

Uploaded by

This PDF is available at http://nap.nationalacademies.

Transfer, Development, and Splice Length for

BUY THIS BOOK

FIND RELATED TITLES SUGGESTED CITATION

NATIONAL COOPERATIVE HIGHWAY RESEARCH PROGRAM

NCHRP REPORT 603

TRANSPORTATION RESEARCH BOARD

Copyright National Academy of Sciences. All rights reserved.

NATIONAL COOPERATIVE HIGHWAY NCHRP REPORT 603

Published reports of the

NATIONAL COOPERATIVE HIGHWAY RESEARCH PROGRAM

Transportation Research Board

and can be ordered through the Internet at:

Copyright National Academy of Sciences. All rights reserved.

Copyright National Academy of Sciences. All rights reserved.

COOPERATIVE RESEARCH PROGRAMS

CRP STAFF FOR NCHRP REPORT 603

NCHRP PROJECT 12-60 PANEL

Copyright National Academy of Sciences. All rights reserved.

This report documents research performed to develop recommended revisions to the

Copyright National Academy of Sciences. All rights reserved.

Copyright National Academy of Sciences. All rights reserved.

Transfer, Development, and Splice Length

Transfer Length and Development Length for Strand

Copyright National Academy of Sciences. All rights reserved.

Development Length and Splice Length

1. Development length of black and epoxy-coated reinforcing bars anchored by means of

Copyright National Academy of Sciences. All rights reserved.

Copyright National Academy of Sciences. All rights reserved.

Introduction and Research Approach

Copyright National Academy of Sciences. All rights reserved.

Copyright National Academy of Sciences. All rights reserved.

Copyright National Academy of Sciences. All rights reserved.

Copyright National Academy of Sciences. All rights reserved.

Copyright National Academy of Sciences. All rights reserved.

Figure 2.1. Comparison of Moustafa pull-out values from

Figure 2.2. Comparison of Moustafa pull-out values from

Copyright National Academy of Sciences. All rights reserved.

OU SERIES TWO DATA (kips)

Figure 2.3. Comparison of NASP Bond Test results at OU in

Figure 2.4. Comparison of NASP Bond Test results at two

Copyright National Academy of Sciences. All rights reserved.

Copyright National Academy of Sciences. All rights reserved.

Copyright National Academy of Sciences. All rights reserved.

1 MPa0.5 = 0.381 ksi0.5; 1 MPa-0.5 = 2.63 ksi0.5

Transfer Length (db)

Figure 2.5. Data comparing transfer lengths to concrete

1 MPa0.5 = 0.381 ksi0.5; 1 MPa-0.5 = 2.63 ksi-0.5

Figure 2.6. Data highlighting differences among strand

Copyright National Academy of Sciences. All rights reserved.

Copyright National Academy of Sciences. All rights reserved.

Bottom Cast Uncoated

Copyright National Academy of Sciences. All rights reserved.

Top Cast Uncoated

Epoxy-Coated Bottom Bars

Copyright National Academy of Sciences. All rights reserved.

Epoxy-Coated Top Cast

Copyright National Academy of Sciences. All rights reserved.

terminated with a standard hook are similar to those in the

2.3 Identification of Issues C1 C2

2.3.1 Member Properties

Copyright National Academy of Sciences. All rights reserved.

Copyright National Academy of Sciences. All rights reserved.

Copyright National Academy of Sciences. All rights reserved.

Copyright National Academy of Sciences. All rights reserved.

Copyright National Academy of Sciences. All rights reserved.

Figure 2.14. Testing methods to evaluate bond strength.

Copyright National Academy of Sciences. All rights reserved.

large differences in code-required anchorage lengths for

Copyright National Academy of Sciences. All rights reserved.

Copyright National Academy of Sciences. All rights reserved.

Experimental Program and Results

Copyright National Academy of Sciences. All rights reserved.

Copyright National Academy of Sciences. All rights reserved.

Table 3.1. Results of NASP Bond Tests at OSU.

Copyright National Academy of Sciences. All rights reserved.