Professional Documents
Culture Documents
The National Academies Press: Proposed Modification To AASHTO Cross-Frame Analysis and Design (2021)
The National Academies Press: Proposed Modification To AASHTO Cross-Frame Analysis and Design (2021)
DETAILS
CONTRIBUTORS
GET THIS BOOK Matthew Reichenbach, Joshua White, Sunghyun Park, Esteban Zecchin, Matthew
Moore, Yangqing Liu, Chen Liang, Bal zs K vesdi, Todd Helwig, Michael
Engelhardt, Robert Connor, Michael Grubb, Ferguson Structural Engineering
FIND RELATED TITLES Laboratory, The University of Texas at Austin; National Cooperative Highway
Research Program; Transportation Research Board; National Academies of
SUGGESTED CITATION
Sciences, Engineering, and Medicine
Visit the National Academies Press at NAP.edu and login or register to get:
Distribution, posting, or copying of this PDF is strictly prohibited without written permission of the National Academies Press.
(Request Permission) Unless otherwise indicated, all materials in this PDF are copyrighted by the National Academy of Sciences.
N AT I O N A L C O O P E R AT I V E H I G H W AY R E S E A R C H P R O G R A M
Matthew Reichenbach
Joshua White
Sunghyun Park
Esteban Zecchin
Matthew Moore
Yangqing Liu
Chen Liang
Balázs Kövesdi
Todd Helwig
Michael Engelhardt
Robert Connor
Michael Grubb
Ferguson Structural Engineering Laboratory
The University of Texas at Austin
Austin, Texas
Subscriber Categories
Bridges and Other Structures
Research sponsored by the American Association of State Highway and Transportation Officials
in cooperation with the Federal Highway Administration
2021
The National Academy of Sciences was established in 1863 by an Act of Congress, signed by President Lincoln, as a private, non-
governmental institution to advise the nation on issues related to science and technology. Members are elected by their peers for
outstanding contributions to research. Dr. Marcia McNutt is president.
The National Academy of Engineering was established in 1964 under the charter of the National Academy of Sciences to bring the
practices of engineering to advising the nation. Members are elected by their peers for extraordinary contributions to engineering.
Dr. John L. Anderson is president.
The National Academy of Medicine (formerly the Institute of Medicine) was established in 1970 under the charter of the National
Academy of Sciences to advise the nation on medical and health issues. Members are elected by their peers for distinguished contributions
to medicine and health. Dr. Victor J. Dzau is president.
The three Academies work together as the National Academies of Sciences, Engineering, and Medicine to provide independent,
objective analysis and advice to the nation and conduct other activities to solve complex problems and inform public policy decisions.
The National Academies also encourage education and research, recognize outstanding contributions to knowledge, and increase
public understanding in matters of science, engineering, and medicine.
Learn more about the National Academies of Sciences, Engineering, and Medicine at www.nationalacademies.org.
The Transportation Research Board is one of seven major programs of the National Academies of Sciences, Engineering, and Medicine.
The mission of the Transportation Research Board is to provide leadership in transportation improvements and innovation through
trusted, timely, impartial, and evidence-based information exchange, research, and advice regarding all modes of transportation. The
Board’s varied activities annually engage about 8,000 engineers, scientists, and other transportation researchers and practitioners from
the public and private sectors and academia, all of whom contribute their expertise in the public interest. The program is supported by
state transportation departments, federal agencies including the component administrations of the U.S. Department of Transportation,
and other organizations and individuals interested in the development of transportation.
AUTHOR ACKNOWLEDGMENTS
The research reported herein was performed under NCHRP Project 12-113 by the Department of
Civil, Architectural, and Environmental Engineering at the University of Texas at Austin (UT Austin).
UT Austin was the contractor for this study.
The Texas Department of Transportation (TxDOT) provided substantial assistance to the research
team throughout the field experimental program. TxDOT organized lane closures and provided dump
trucks for controlled live load tests as well as general traffic control to assist in the instrumentation of
the bridges. The research team specifically acknowledges Dr. Ken Lin, Lynn Champagne, Chris Baker,
David Jeffreys, Adam Galland, Melody Galland, Maria Aponte, Ray Castillo, Jaime Castaneda, and
Larry Whittington for their contributions to the field studies. Significant assistance was also provided
by Jamie Farris, Dennis Johnson, and Walter Fisher in identifying bridges for possible instrumentation.
Texas Advanced Computing Center (TACC) provided computing resources to conduct the extensive
parametric studies associated with this research. The research team specifically acknowledges Bob Garza
and Dr. Ritu Arora for their assistance throughout the process.
The research team thanks Dr. Tom Murphy, Dr. John Kulicki, and Dr. Andrzej Nowak for informa-
tion provided from the SHRP 2 R19B study that was invaluable for understanding weigh-in-motion data
gathered and related reliability studies.
FOREWORD
By Waseem Dekelbab
Staff Officer
Transportation Research Board
NCHRP Research Report 962 presents proposed AASHTO bridge design specifications
for the analysis and design of cross-frames in straight and curved steel I-girder bridges.
The proposed specifications are based on comprehensive analytical and testing programs
for investigating the effects of support skew on the cross-frame behavior, fatigue design
forces, the strength and stiffness requirements for stability bracing, and the influence of
cross-frame member end connection upon the cross-frame stiffness. This report will be of
immediate interest to bridge engineers.
Developments in bridge design and analysis in recent years have created the need for
improvements to cross-frame analysis and design for steel girder bridges. In the past, the
configuration of cross-frame systems was generally based upon standard designs in which
member sizes and layouts were dependent upon geometry and minimum member cross-
section requirements. The opportunities for improvements to cross-frame analysis and
design cover a variety of topics including: (1) improved definition of fatigue loading for
cross-frames in curved and/or severely skewed steel girder bridges, analyzed using refined
analysis methods; (2) implementation of stability bracing strength and stiffness require-
ments in the context of AASHTO Load and Resistance Factor Design (LRFD) bridge
design; and (3) additional guidance for adjustment of the effective stiffness of cross-frame
members in refined analysis models to reflect the influence of end connections on cross-
frame member stiffness. Addressing these topics could result in a dramatic improvement in
reliability and economy of cross-frames for steel I-girder bridges.
Research was performed under NCHRP Project 12-113, “Proposed Modification to
AASHTO Cross-Frame Analysis and Design” by the University of Texas at Austin and
included: (1) controlled live load tests and measurement of in-service stress cycle counts
for three steel I-girder bridges and (2) an extensive analytical parametric study. The results
of the live load tests and analytical study were used to develop the proposed modifica-
tion to the AASHTO specifications.
Several deliverables, provided as appendices, are not included in the report itself but
are available on the TRB website at trb.org by searching for NCHRP Research Report 962.
The appendices are as follows:
• Appendix B: Cross-Frame Design Example (Straight Bridge)
• Appendix C: Cross-Frame Design Example (Curved Bridge)
• Appendix D: Phase I Summary
• Appendix E: Phase II Summary
• Appendix F: Phase III Summary
CONTENTS
1 Summary
4 Chapter 1 Background
4 1.1 Problem Statement and Current Knowledge
5 1.1.1 Best Practices for Cross-Frame Layout and Detailing
6 1.1.2 Cross-Frame Fatigue Behavior
7 1.1.3 Cross-Frame Analysis
7 1.1.4 Cross-Frame Stability Bracing Requirements
8 1.2 Research Objectives
10 1.3 Scope of Study
10 1.4 Organization of Report
143 References
A-1 Appendix A Proposed Modifications to AASHTO LRFD
Note: Photographs, figures, and tables in this report may have been converted from color to grayscale for printing.
The electronic version of the report (posted on the web at www.trb.org) retains the color versions.
SUMMARY
Proposed Modification
to AASHTO Cross-Frame
Analysis and Design
Cross-frames are important structural components that serve many functions throughout
the service life of steel I-girder bridges. They primarily act as stability braces to enhance
the lateral-torsional buckling resistance of the bridge girders during erection and deck
construction. Among other functions, they also distribute live loads to girders across
the bridge width in the final composite condition. Under repetitive load cycles caused
by heavy truck passages, cross-frames and their connections can be susceptible to load-
induced fatigue cracking if not properly designed. Despite recent changes to design speci-
fications that signal cross-frames as potential fatigue-sensitive details given substantial
live load force effects, little load-induced fatigue cracking has been observed in existing
steel I-girder bridges across the country.
Cross-frames have historically been detailed and fabricated based on general rules-of-
thumb and experience. In recent years, however, developments in bridge design specifica-
tions have necessitated updates to cross-frame design and analysis practices. Cross-frames
are now designed and detailed based on rational analysis for all stages of construction and
service life, which has further emphasized the importance of accurate and reliable analysis
techniques and design criteria.
Although considerable research over the past several decades has improved cross-
frame design and analysis, bridge designers have generally lacked adequate guidance
on the appropriate (i) fatigue loading criteria, (ii) analysis procedures, and (iii) stability
bracing requirements for cross-frame systems. With that in mind, NCHRP Project 12-113,
“Proposed Modification to AASHTO Cross-Frame Analysis and Design,” was under-
taken to address many of these knowledge gaps in an attempt to improve the reliability
and economy of cross-frames in steel I-girder bridges. The fundamental objectives of
NCHRP Project 12-113 were to produce quantitatively based methodologies and design
guidelines for the following items:
a. Improved definition of fatigue loading for cross-frames in skewed steel I-girder bridges
(straight or curved) analyzed using refined analysis methods;
b. Investigation of the influence of girder spacing, cross-frame stiffness and spacing
(including staggered layouts), and deck thickness on the force effects in cross-frame
systems;
c. Additional guidance on how to evaluate the influence of end connections on cross-frame
member stiffness in refined analysis models;
d. Evaluation of commercial software programs and their ability to accurately predict
cross-frame forces for various bridge geometries and cross-frame configurations; and
e. Evaluation of stability bracing strength and stiffness requirements for implementation in
the AASHTO Load and Resistance Factor Design (LRFD) Bridge Design Specifications.
1
The objectives were systematically addressed through a series of field experiments, para-
metric finite element analyses, and weigh-in-motion studies (considering high-resolution
data from 16 sites across the country). This report summarizes the results of the research
effort undertaken as part of this project.
An important step in the research was to first validate the computational models
utilizing field data from representative bridges. As such, three steel I-girder bridges of
varying geometry (i.e., a straight bridge with normal supports, a straight bridge with
skewed supports, and a horizontally curved bridge with radial supports) were instru-
mented and experimentally tested in two different ways. First, a controlled live load test
was conducted. Dump trucks of known axle configuration and weight were statically
positioned at different locations on the bridge deck, and the stress response in critical
cross-frame members and girder flanges was subsequently measured. Second, stress cycle
counts induced in the same critical cross-frame members from truck traffic were moni-
tored for a one-month span. The data collected from the controlled live load test were
primarily used to validate a finite element modeling approach. The in-service stress range
data served as a physical benchmark for computational analysis and weigh-in-motion
studies conducted in later phases of the project.
Upon validating a finite element approach, a series of extensive analytical parametric
studies were conducted to expand the breadth and depth of knowledge gained from the
field studies. In general, the load-induced fatigue behavior and stability bracing charac-
teristics of conventional X- and K-type cross-frames were examined for a variety of bridge
geometries commonly found in the United States. These analyses were performed with
different levels of computational refinement, ranging from sophisticated, three-dimensional
(3D) approaches to simplified, two-dimensional (2D) approaches. Based on the results
of the experimental and analytical studies, there were several key findings related to the
five major objectives identified above, as follows:
• The current fatigue loading model implemented in AASHTO LRFD, which was calibrated
based on girder response, does not accurately represent the cross-frame fatigue response
to the U.S. truck spectrum. The current design loads and factors generally overestimate
the force effects in critical cross-frame members when compared to measured field
data or weigh-in-motion records, largely due to the sensitivity of cross-frames to load
position on the bridge deck.
• Load-induced cross-frame forces are highly variable and depend on a number of
parameters. However, the results of this study indicate that significant live load-induced
cross-frame forces are generally correlated with large support skews and tight horizontal
curves. Implementing discontinuous, staggered cross-frame layouts represents a practical
means to mitigate excessive load-induced forces for bridges with large support skews.
In contrast, cross-frame fatigue force effects in straight bridges with little to no support
skews are generally small.
• The load-induced deformational response of cross-frames in composite systems (appli-
cable to in-service bridges) is much different than the deformational response in non-
composite systems (applicable to bridges during construction). Additionally, traditional
3D modeling approaches of cross-frames commonly used by commercial software
programs (i.e., cross-frames treated as pin-ended planar trusses) can misrepresent the
true stiffness of a cross-frame panel and its connections. As such, in-service stiffness
modification factors are proposed to improve analysis procedures for cross-frames.
• Two-dimensional models, which represent cross-frame panels as equivalent beam
elements, are prone to predict erroneous top strut forces and substantial errors in
diagonal member forces when evaluating the composite condition. This is largely
Summary 3
attributed to the assumptions inherent with the postprocessing procedures (i.e., con-
verting equivalent beam forces into cross-frame member forces). Although these dis
crepancies were observed for bridges of all geometries, 2D analysis methods tend to
perform worse for heavily skewed and/or curved bridge systems.
• Previously established bracing stiffness and strength requirements published in the
American Institute of Steel Construction (AISC) Specifications were reviewed and
modified for implementation into the AASHTO LRFD Specifications. Additionally,
stability-related force effects in cross-frames can be combined with other construction-
related load cases via linear superposition. Design examples were developed to demonstrate
these procedures.
Based on the major conclusions of NCHRP Project 12-113, changes to the specification
and commentary language in select AASHTO LRFD Articles are proposed. Design examples
were also developed to support and demonstrate the major findings of this research.
Ballot items were subsequently prepared for consideration by the AASHTO Highway
Subcommittee on Bridges and Structures. In general, the guidance developed from this
study is expected to result in improved reliability and economy of cross-frame design in
steel I-girder bridges.
CHAPTER 1
Background
Background 5
As structural analysis tools and bridge design practices advanced in the following years,
updates to the cross-frame analysis and design procedures became increasingly necessary.
This process was initiated in the 1st Edition of the American Association of State Highway and
Transportation Officials LRFD Bridge Design Specifications (1994). This document, along
with its subsequent updates, is referred to as AASHTO LRFD hereafter. In the 1994 AASHTO
LRFD, the 25-foot spacing limit was removed in favor of a rational design approach. Thus,
rather than using a prescriptive or standard design, the need for cross-frames was to be evaluated
for all stages of construction and in the final condition of the bridge.
In the past two to three decades since the rational design approach was implemented,
considerable research has been conducted to improve the body of knowledge related to cross-
frame behavior in steel bridge systems. Many of these past research findings are documented
in the current 9th Edition of the AASHTO LRFD Specifications (2020). Note that any specific
reference to AASHTO LRFD articles in this report implies the 9th Edition specifications,
unless noted otherwise. Despite those advancements, several gaps in knowledge have been
identified, particularly in the areas of fatigue loading and analysis of cross-frames. Before
identifying the major objectives of NCHRP Project 12-113 in Section 1.2, these past research
advancements and the corresponding updates to AASHTO LRFD are first reviewed.
connection effects. While these concepts show promise, neither has been widely adopted in
the steel bridge design and construction industry.
Background 7
the current fatigue loading model (i.e., AASHTO LRFD Articles 3.4.1, 3.6.1.4, and 6.6.1.2.1) for
cross-frame analysis and design, as further outlined in Section 1.2.
the Marcy Pedestrian Bridge. Although the Marcy Pedestrian Bridge was a steel tub girder, the
LTB failure is similar to the system buckling mode for which many narrow I-girder systems
are susceptible. AASHTO Article 6.10.3.4.2 summarizes the major findings of these studies
and provides design guidance to estimate the resistance of this system-level failure mode.
Substantial work has also been performed in the area of lean-on bracing (Helwig and Wang
2003, Romage 2008). Lean-on bracing concepts, which have been utilized in building frame
design for decades, were adapted for steel straight I-girder bridge applications. These braces
include only a top and bottom strut and “lean on” adjacent cross-frames to stabilize the
longitudinal girders. Several newly constructed straight skewed I-girder bridges in Texas
have successfully implemented lean-on braces, which subsequently alleviate load-induced force
effects in cross-frames as well as improve fabrication economy by eliminating several full
cross-frame panels.
Despite these recent advancements related to the topic of steel bridge stability, stability
bracing requirements for cross-frames are absent from the AASHTO LRFD Specifications as
of this writing. In contrast, the American Institute of Steel Construction (AISC) Specification
commonly used for building design was the first major specification in the world to include
guidance on stability bracing design. In the current AISC Specification (2016), the stability
bracing provisions are located in Appendix 6, and Article 6.3.2a provides specific strength
and stiffness requirements for torsional point braces such as cross-frames. These require-
ments are largely based on the work conducted by Yura (2001), Helwig and Yura (2015), and
Prado and White (2015). Note that the fundamental work that led to the stability bracing
requirements in the AISC Specification was actually developed from studies on bridge girders
(Yura et al. 1992).
Considering that many of these stability bracing design provisions are applicable for both
building and bridge applications, another major objective of NCHRP Project 12-113 was
to assess the stability bracing guidance in the AISC Specification for implementation into
AASHTO LRFD.
Background 9
• [Stability bracing – Article N/A] Can the AISC design guidelines for stability bracing be
incorporated into AASHTO LRFD? Are special requirements needed for negative moment
regions of continuous systems?
• [Stability bracing – Article N/A] How are these stability bracing requirements combined with
other load conditions such as wind?
Background 11
Chapters 2 and 3 of the main report are not intended to be comprehensive in nature. Rather
than overwhelm the reader with the full set of experimental and analytical data, sample results
are selectively presented to provide the proper context to the major conclusions summarized in
Chapter 4. For a more detailed overview of the experimental and analytical studies and results,
several appendices have been prepared.
Based on the key findings summarized in Chapter 4, Appendix A includes the draft speci-
fication revisions prepared for consideration by the AASHTO T-14 Technical Committee for
Structural Steel Design and the AASHTO Committee on Bridges and Structures. Appendices B
and C present cross-frame design examples for a straight bridge with normal supports and a
horizontally curved bridge with radial supports, respectively. These examples articulate how
the various load cases and conditions independently investigated in Chapters 2 and 3 can be
combined for the various stages of construction and service. Rather than develop representative
examples from scratch, the research team elected to adopt and modify previously published
design examples that are commonly referenced by design engineers. As such, the design
examples generally follow Steel Bridge Design Handbook (SBDH) Examples 2A (Barth 2015)
and 3 (Rivera and Chavel 2015).
Appendices D through F cover much of the same material as in the main body of the report.
However, these appendices provide a more thorough and exhaustive overview of the research
methodology, results, and critical findings. Appendix D specifically addresses Phase I of NCHRP
Project 12-113, whereas Appendices E and F address Phases II and III, respectively. The various
phases of the projects are detailed in Chapter 2.
CHAPTER 2
Research Approach
To address the major objectives of NCHRP Project 12-113, four distinct phases were system-
atically carried out by the research team. The four phases, as identified in the project RFP, are as
follows: (I) Planning, (II) Part 1 of the Analytical Program and Field Experiment, (III) Part 2 of
the Analytical Program and Proposed AASHTO Specifications, and (IV) Final Products. Each
successive phase built on the findings and lessons learned from the previous phase. This is
demonstrated schematically with a workflow diagram in Figure 2-1. As noted in Section 1.4,
comprehensive summaries of Phases I, II, and III are provided in the respective Appendices D, E,
and F. The information provided in this chapter is intended only to provide a general overview
of the research approach and methodologies.
In Phase I, an extensive literature review and industry survey was conducted, and preparations
were made for the Phase II field experiments. Phase I consisted of six tasks, which are briefly
summarized below:
• Task 1: Conduct a literature review, document current design practices, and examine how
design software incorporates these design practices.
• Task 2: Synthesize the results of the literature review to identify the knowledge gaps related
to the research objectives.
• Task 3: Propose an analytical program to be executed in two parts. Part 1 (executed in
Phase II) includes preliminary analysis to aid in the development of the experimental
program, and Part 2 (executed in Phase III) includes comprehensive parametric studies to
achieve the research objectives outlined in Chapter 1.
• Task 4: Propose field experiments (executed in Phase II) that include one straight bridge,
one skewed bridge, and one horizontally curved bridge. Measurements include (i) cross-
frame member fatigue force ranges under controlled application of live load and (ii) in-service
effective and maximum stress ranges for the same cross-frame members.
• Task 5: Identify existing articles of AASHTO LRFD that require modification.
• Task 6: Prepare Interim Report No. 1.
Note that several of these tasks were previously summarized in Chapter 1, including the
review of pertinent research and AASHTO LRFD articles. Section 2.1 briefly details the industry
survey performed by the research team that facilitated the planning of the experimental and
analytical studies and also aided in the development of the research scope. Appendix D also
outlines this material in a more thorough manner. The experimental and analytical programs
are covered under the Phase II and III tasks discussed in the following paragraphs.
In Phase II, field experimental studies and model validation were performed, which allowed
the analytical studies in Phase III to commence. In total, Phase II consisted of four tasks, which
are briefly summarized below:
12
• Task 7: Execute Part 1 of analytical program related to the specific bridges evaluated in the
field experiments.
• Task 8: Execute the field experiments, as introduced in Task 4.
• Task 9: Validate the analytical modeling approach based on the results of the field experiments.
• Task 10: Prepare Interim Report No. 2.
The methodology used to accomplish Tasks 7 through 9 is briefly outlined in Section 2.2.
Appendix E provides a more detailed overview of these Phase II tasks.
Phase III expanded on the work completed in Phase II. In order to properly evaluate the
behavior and design criteria of cross-frames in common highway I-girder bridge systems,
a much larger representative sample than the three unique bridge geometries and traffic condi-
tions from the field studies was necessary. As such, the validated modeling approach established
in Phase II was used to conduct extensive analytical studies that ultimately improve the breadth
and depth of knowledge pertaining to cross-frame behavior. Phase III included three major
tasks, as summarized with the following:
• Task 11: Execute Part 2 of the analytical program. Perform parametric studies related to live
load-induced fatigue performance and stability bracing characteristics of cross-frames.
• Task 12: Based on the analytical and experimental investigations, develop specification and
commentary language for proposed changes to AASHTO LRFD.
• Task 13: Prepare Interim Report No. 3.
Sections 2.3 through 2.6 describe the methodology used to carry out Task 11 and the analy
tical studies. The project culminated in Phase IV, for which the proposed specification and
commentary language was finalized for potential balloting by AASHTO. These proposed modi-
fications are summarized in Chapter 4 and provided in full in Appendix A.
As outlined in Section 1.2, there were five major objectives of the NCHRP Project 12-113
investigation. Task 11 was structured in such a way to address these objectives through a series
(a) X-Type (b) X-Type; No Top Strut (c) X-Type; No Struts (d) K-Type
Figure 2-2. Various cross-frame configurations used across the United States.
• Cross-frame connection details vary substantially from state to state. Figure 2-2 graphically
summarizes the range of cross-frame configurations used in the United States.
• In almost all instances where gusset plates are used to attach cross-frame members to
the connection plates, the gusset plates are bolted to the connection plates and welded to the
member. The lone exception to this is the Texas DOT (TxDOT) cross-frame details, where
cross-frames often utilize erection bolts and the final connections are welded prior to deck
construction.
• In instances where gusset plates are not used to connect members to the connection plates,
the diagonals are connected in one of two ways. They are either bolted directly to the connec-
tion plate or they are fillet welded to the top and bottom chords.
As indicated above, many DOTs specify bolted connections between the cross-frame member
and the gusset plate with TxDOT being a key exception. In these instances, slip-critical (SC)
connections are commonly used. Given that the present study is generally focused on the
cross-frame response within the elastic range (i.e., the fatigue limit state), it is assumed that
SC bolted connections do not slip at these service-level loads, such that the response of a cross-
frame is relatively unaffected by the connection method. Therefore, the studies presented herein
do not differentiate between bolted and welded connections, as cross-frame behavior is more
influenced by its member size and overall geometry.
As noted in Section 1.3, the industry survey allowed the research team to develop a manageable
scope for the analytical studies. Based on these results, only single-angle sections in traditional
X-type and K-type cross-frames were assessed [i.e., cross-frames (a) and (d) in Figure 2-2].
It should also be noted that cross-frames without top and/or bottom struts can potentially lead
to stability issues during construction. Although zero-force members in many stability applica-
tions (i.e., construction condition), struts are important since they prevent buckling mode in
which the girders rotate out-of-plane in opposite directions. Thus, it is always recommended to
include top and bottom struts for both X- and K-type frames, particularly in horizontally curved
bridges and/or bridges with significant skews.
supports, a straight bridge with skewed supports, and a horizontally curved bridge. The research
team assessed the pros and cons for a variety of existing bridge structures throughout Texas
but ultimately elected to study three in the greater Houston area.
This section provides a brief overview of the field experimental program and the general
procedures used to validate the FEA modeling approach. Section 2.2.1 succinctly describes the
key geometric parameters and traffic conditions of each instrumented bridge. Section 2.2.2
outlines the two experimental tests performed with additional guidance on how the data was
processed. Note that the commentary provided herein is not intended to cover every aspect of
the field tests. For a more detailed synopsis, refer to Appendix E.
on numerous factors including field access, traffic control considerations, the capacity of the
data acquisition system (DAQ), and preliminary analysis. The preliminary analysis, which was
conducted prior to the field studies, identified the cross-frames likely to experience the largest
stress ranges, as well as the critical truck positions to evaluate in the controlled live load test. The
results of these preliminary studies are further detailed in Appendix E.
A wireless monitoring system was designed to collect continuous strain time-histories, as well
as simultaneously calculate stress cycle counts in accordance with commonly used rainflow-
counting algorithms (Downing and Socie 1982). For each instrumented cross-frame member,
four quarter-bridge strain gages were installed at various locations along a cross-section. By
measuring the strain at various points on a cross-section, the axial component of stress could
be isolated from the biaxial bending effects in the member via the linear regression method
developed by Helwig and Fan (2000). This procedure was validated by laboratory studies prior
to implementation in the field.
Given that the Category E′ designation established for these welded connection details was
explicitly based on axial stresses (i.e., P/A), reliably measuring axial stress was critical. It
is important to note, however, that significant bending stresses were measured as part of this
study to verify the impact of eccentric end connections. Two strain gages were also installed at
each edge of the girder bottom flange, such that longitudinal bending stress and warping stress
components could be independently evaluated.
Upon completing the strain gage installation, two different experimental tests were performed
for each bridge. The controlled live load test consisted of four dump trucks loaded with sand
(of known axle configuration and weight) positioned at various longitudinal and transverse
points on the concrete deck. The in-service monitoring consisted of recording live load stress
range cycles in the instrumented elements over a one-month period. The subsections herein
briefly address each test separately.
experiences. Note that the one-month monitoring period is sufficiently long to capture stabi-
lized data as well as hourly, daily, and weekly trends that may deviate from normal or average
conditions (Connor and Fisher 2006; Fasl 2013).
In addition, three supplementary metrics derived from the histogram plots were used by the
research team to compare data between the various instrumented components of the different
bridges. These three metrics are effective stress range (Sre), maximum stress range (Srm), and
index stress range (Sri).
All three, which can be derived from simple postprocessing of the measured data, are
effective tools for evaluating the fatigue performance of structural components and details.
Effective and maximum stress ranges are defined in the current 9th Edition AASHTO LRFD
Specifications (2020). Index stress range was a metric developed by Fasl (2013) but is not
currently adopted by standard codes. These metrics are further explained in the context of
measured data in Section 3.1 and Appendix E.
load position and Fatigue I and II load factors [which have been calibrated for longitudinal
girders only through the research efforts documented in SHRP 2 R19B (Modjeski and Masters
2015)] can also be evaluated with respect to cross-frame systems.
Following this introduction, the modeling approach adopted in the Fatigue Loading Study is
outlined (Section 2.3.1). The procedures used to evaluate the bridges for current fatigue design
loading are presented (Section 2.3.2), as well as the procedures related to the WIM study (Sec-
tion 2.3.3). Major results and outcomes related to the studies are summarized in Section 3.3.
shells but instead use line elements. While the present study utilized line elements with R-factors,
additional studies were also carried out making use of shell-element models. A discussion of
these modeling techniques is provided in the following paragraphs.
In terms of the more rigorous approach, each component of the cross-frame panel is
explicitly modeled with a shell-element: connection plates, gusset plates, cross-frame members
(typically single angles), and fill plates commonly found at the intersection of diagonals in
X-frames. The eccentric load path is also explicitly considered. These models are simply
referred to as shell-element models herein. Although generally more accurate, this refined
technique (especially with a fine mesh) rapidly increases the computational demands and
run time.
The second alternative, which is more common in design practice and even in many research
studies, simplifies the cross-frames as pin-ended truss elements. Therefore, the eccentric load
path and the stiffness contributions of connection and gusset plates are neglected. The actual
stiffness of the cross-frame system, though, can be approximated with a simple modifica-
tion factor (R-factor) assigned to the cross-sectional area of the truss element or the elastic
modulus of the material. These types of models are referred to as truss-element models in
this section.
Given the scope of the study, the shell-element modeling approach was deemed too
computationally intensive. Thus, the research team elected to perform the Fatigue Loading
Study with truss-element models to balance computational speed and modeling accuracy. To
better understand the behavior, preliminary studies were performed using the general purpose
FEA program, Abaqus, to determine which variation of truss-element models produces the
most accurate results relative to the validated, shell-element models. Based on these preliminary
studies, the following observations were made, which influenced the approach adopted for the
Fatigue Loading Study (Figure 2-3):
• Explicitly modeling the offset dimension between cross-frame working line and the girder
flanges better represents the stiffness of the panel and the inclination angle of the diagonal
members (as opposed to connecting the cross-frame member into a shared node at the
web-flange juncture);
• Without the connection plates, cross-frame forces are greatly affected by web distortion caused
by the concentrated forces acting on an unstiffened, flexible web (note that distortion effects
would not be impactful if cross-frames were connected directly into the girder flange); and
• Terminating the cross-frame truss member at the edge of or in the middle of the connection
plate affected the force distribution in the cross-frame members.
Note that the vertical offset dimension and the presence of connection plates are further
discussed in Sections 2.4 and 2.5, which focus heavily on the approaches traditionally adopted
by 3D commercial design software programs. Prior to conducting the parametric study, the
accuracy of the assumed truss-element modeling approach and its simplifications outlined
above was validated with the field-measured data. This procedure is detailed in Appendix F.
Lastly, it was determined that a first-order, elastic analysis was suitable for the fatigue load-
ing conditions that are representative of service-level load magnitudes. This approach is also
consistent with common design practice and AASHTO LRFD recommendations. To verify this
assumption, supplementary FEA studies that considered second-order effects and initial
imperfections in cross-frame members were performed, and their influence on the cross-frame
force demands was found to be negligible.
was filtered to generally provide only the cases that are likely most critical. Rules were estab-
lished to eliminate permutations deemed unnecessary or less important than others. Many
of these rules were based on current AASHTO provisions related to geometric limits. For
example, AASHTO Article 2.5.2.6.3 recommendations were implemented to establish reason-
able bounds on the span-to-depth-ratios considered in the study. To further clarify Table 2-2,
the following bulleted items provide key discussion points. The following list represents an
abbreviated version; refer to Appendix F for a more thorough overview and a comprehensive
list of every bridge parameter studied.
• The differentiation between independent variables and constants is based on prioritization
of the parameters, the information gained from the validation studies, data collected from
the industry survey, and common bridge design practice. For example, the presence of a
concrete barrier was deemed a pertinent variable based on the validation studies in Phase II,
as is documented in Section 3.1.1. However, for purposes of the Phase III parametric study,
the research team elected to simply maintain constant dimensions and parameters for barriers
rather than evaluate variable dimensions.
• In the same vein, overhang dimensions were taken as a uniform 3 feet (measured from
the centerline of the fascia girder to the outer edge of the deck) to simplify the calculations.
For bridges with a wider girder spacing, overhang lengths are commonly greater than 3 feet.
As is discussed later in the report, fatigue forces in cross-frames are often maximized by
truck passages where a wheel line rides along the overhang portion of the deck. It was found,
however, that force effects in a cross-frame member due to an “overhang” live load are gener-
ally linearly dependent on the outer wheel distance to the centerline of the fascia girders.
• Staggered cross-frame layouts were considered in the study, but skewed cross-frames were
not. For skewed bridges, cross-frames were either: (i) contiguous lines normal to the longitu-
dinal girders or (ii) discontinuous lines that run parallel to the support skew angle but placed
at a line normal to the longitudinal girders. Because both support skew angles considered in
the study (30 and 60 degrees) exceed 20 degrees, skewed cross-frames parallel to the support
skew were not considered in accordance with AASHTO LRFD Article 6.7.4.2.
• In general, the research team ensured AASHTO Article 6.7.4.2 was satisfied for the layout of
cross-frames. In particular, the cross-frame depth was selected to exceed 75% of the girder
depth to preclude any significant web-distortional effects. Additionally, the cross-frame
layout recommendations near obtuse and acute corners of skewed bridges given in AASHTO
Article C6.7.4.2 were followed.
• A constant stiffness modification factor (R-factor) is assumed for all cross-frame systems,
regardless of the assumed geometry and connection details. Because the focus of the Fatigue
Loading Study is not on stiffness modification factors, the research team elected to normalize
the effects of eccentric load paths for all cross-frame systems in the 4,104-model parametric
study. That way, any potential variability associated with the R-factors is eliminated, and
the relative impact is then consistent across all bridge models. A uniform factor of 0.60
was selected, as it was shown to be a reasonable representation of the “average” value in
preliminary R-factor studies.
• The decision to use R = 0.6 was made early in Phase III, several months before the final results
were obtained from the R-Factor Study (i.e., proposed R = 0.75). Clearly, consistently
larger cross-frame force effects would have been determined had the parametric study been
conducted with the larger R-factor. As noted in Section 3.3, an increase in R-factor from 0.60
to 0.75 would only increase those predicted force effects about 5% to 7%. As such, it was
justified that this deviation did not warrant a repeat of the parametric study since no major
conclusions outlined herein would be affected.
• The girder sections of each bridge were evaluated with respect to proportion limits (AASHTO
Article 6.10.2) and strength limit state resistance requirements (AASHTO Article 6.10.6).
In other words, it was ensured that the bridges studied represent realistic structures that have
been properly designed and detailed.
• Aside from the parameters listed in the tables, the research team also conducted spot checks
to examine the influence of several parameters including different flange dimensions and
the removal of bridge barriers.
Based on the findings of this preliminary study, the research team preselected cross-frames
to evaluate prior to running the full parametric study. Given the unique geometries and
cross-frame layouts, the selected cross-frames are not necessarily in the exact same location
for each bridge. Rather, general locations/regions that are consistent across all 4,104 models
were selected. These four general locations/regions, which apply equally to both straight and
horizontally curved bridges, included:
• Edge cross-frame bay closest to the point of maximum positive dead load moment (i.e.,
midspan for single-span bridges and approximately 0.35L measured from the end support in
end spans of continuous units);
• Interior cross-frame bay closest to the point of maximum positive dead load moment;
• Interior cross-frame bay nearest to the end support (skewed or non-skewed); and
• Interior cross-frame bay nearest to the intermediate support (skewed or non-skewed).
For these regions, every cross-frame member in the panel was examined (i.e., top strut,
bottom strut, and diagonals). These cross-frame panels were selected given the propensity
for more significant differential girder displacements in the surrounding regions. For a more
detailed evaluation of these assumptions, refer to Appendix F.
To illustrate these general locations, a representative sketch is provided in Figure 2-4. The
figure presents the critical cross-frame panels evaluated for representative single-span, two-
span continuous, and three-span continuous straight bridges. Note that these figures are simply
schematic and not drawn to scale. In addition, only intermediate cross-frames are illustrated;
cross-frames at end and intermediate supports are hidden for clarity, as these are not the focus
of the study.
tasks: (i) develop the 3D FEA models for all geometries, (ii) run the simulated influence-
surface analysis, (iii) output the axial-force and displacement response of the preselected
cross-frame members of interest, and (iv) create summarized output files of all pertinent data
(i.e., influence-surface results for every critical cross-frame member). These extensive calcula-
tions were performed using the Texas Advanced Computing Center (TACC) supercomputers.
Once the summarized data files were obtained, further postprocessing was performed in an
external Excel program, which is documented in the next subsection.
375
350
350
325
325
300
300
275
275
250
250
225
125
125
100
100
75
75
50
50
25
25
-25 0
0
-13
13
-13
13
-0.2 (C)
0.2 (T)
-0.2 (C)
In total, the research team produced thousands of influence-surface plots similar to those in
Figure 2-5. Influence-surface plots are useful visual tools and provide important insights on the
load-induced response of cross-frames, particularly in the case of this study where evaluating
over 65,000 plots in an efficient and meaningful way is virtually impossible given the number
of parameters investigated and the variability in the response. While the contour graphs show
the response for a 1-kip force applied anywhere on the deck, the actual force generated in
the cross-frame due to a truck on the bridge is the sum of the wheel loads multiplied by the
corresponding influence-surface values at the wheel locations. In order to assess the correla-
tion between truck placement, fatigue stresses, and the various bridge geometries, additional
processing of the data was necessary. Thus, Sections 2.3.2 and 2.3.3 introduce realistic loading
conditions for these 4,104 bridge models.
More specifically, Section 2.3.2, utilizes the over 65,000 influence-surface plots to assess
current AASHTO fatigue loading criteria. Section 2.3.3 expands on that analysis and inves-
tigates the effects of measured WIM traffic streams.
420
11.8 kips
0
370
-5
-10
-30 20 70 120 170 220 270 320 370 420
320
Longitudinal Position of Truck (ft)
10
270
Longitudinal Position (ft)
Axial Force (kips)
5
3.7 kips
220
0
170
-10
-30 20 70 120 170 220 270 320 370 420
Longitudinal Position of Truck (ft)
120
10
Axial Force (kips)
70
5
9.4 kips
0 20
-5
-10
-30
magnitudes are smaller, but the response is more complex. Whereas the other two truck passages
resulted in one primary force cycle, the centerline passage results in additional secondary
cycles of lesser magnitude. The primary cycle (3.7 kips) and the two secondary cycles (0.5 kips
and 0.2 kips) are illustrated in the figure.
Given that the permanent stress states in the cross-frames of these representative bridges are
unknown (e.g., locked-in fit-up forces, dead loads, residual stresses), only live load force/stress
cycles entirely in tension or subject to reversal, regardless of how small the tension component
is, were considered. It is recognized that, in actuality, this assumption would only be valid when
the Fatigue I factored tensile stress component exceeds the compressive stress due to unfactored
permanent loads or the permanent loads are tensile in nature. To simplify the analysis, the
tensile force component was always considered to be large enough to propagate a crack.
Thus, truck passages that induce a purely compressive force/stress cycle with no tensile force
component are disregarded from the evaluation, as they are not a fatigue-sensitive loading
condition. Also note that, since the primary focus of this study is related to the fatigue limit
state, permanent (dead) loads and locked-in stresses are not explicitly addressed herein,
except as described above.
The computationally intensive procedure introduced in Figure 2-6 was performed for all
influence-surface plots obtained from the 4,104 bridge models analyzed in Abaqus, which
resulted in over 3 million influence-line plots similar to the three examples presented in the
figure. A script was developed to automate these moving-load simulations. The script was
designed to perform rainflow counting on the various influence-line plots (i.e., axial-force
time-histories), and the following output was recorded for each cross-frame member evaluated:
(i) the transverse position of the AASHTO fatigue truck that maximizes the force range, (ii) the
magnitude of the primary force cycle caused by the critical lane passage, and (iii) the number
and magnitude of any secondary cycles caused by that same critical lane passage.
The maximum force range for a given cross-frame member represents the unfactored design
force for which the engineer evaluates the Fatigue I or II limit state. Note that many of the results
presented in Chapter 3 are based on the unfactored design force ranges obtained from this
specified procedure.
The unfactored fatigue force ranges were then factored in accordance with current 9th Edition
AASHTO LRFD load factors (as well as the cross-frame-specific load factors proposed and
outlined in Section 3.2) and converted into an axial stress with consideration of shear lag effects
[i.e., the U factor specified in AASHTO Table 6.6.1.2.3-1 (Condition 7.2)]. Dynamic load
allowance (1.15 for fatigue limit state) was also subsequently applied to produce factored axial-
stress ranges for fatigue design and evaluation.
The resistance side of the AASHTO fatigue design criteria (i.e., load-induced fatigue) was then
introduced into the computational studies. Based on the guidance provided in Article 6.6.1.2.3,
it was assumed that the critical Category E′ welded cross-frame details were all governed by the
finite-life calculations and Fatigue II limit state, which is a reasonable assumption for Category E′
details and practical traffic conditions. Given that finite fatigue life is inherently a function of
both stress magnitude and frequency of load occurrence, the anticipated number of stress cycle
counts over the service life of these “fictional” bridges must be considered in some capacity.
The 4,104 bridges evaluated in this study, however, are not representative of any particular
location or traffic conditions. As such, the research team elected to investigate different repre-
sentative traffic conditions to bound the problem. In other words, these “fictional” bridges were
effectively constructed along different highway corridors in Texas, both in rural areas with
low traffic volumes and congested urban areas with high traffic volumes. Rather than utilize
measured WIM data (Section 2.3.3) at this stage, the research team utilized ADT maps readily
available on the TxDOT website (2020), as well as simplifying assumptions recommended
by AASHTO Article C3.6.1.4.2. A state highway system in rural Llano, Texas, and a heavily
trafficked corridor in Houston, Texas, were selected as the extreme conditions, but the general
process would be identical for other highway systems in other states.
Although not outlined in the report herein, AASHTO LRFD guidance was followed to obtain
an estimate of realistic single lane, average daily truck traffic (ADTTSL) values for the purposes
of computing the Fatigue II resistance stress ranges for the two extreme traffic conditions
considered. A detailed overview of these calculations is provided in Appendix F.
As noted previously in this section, the design stress ranges and resistances (as computed
by the procedures outlined above) are used throughout Chapter 3 to establish generalized
observations about load-induced cross-frame behavior in composite systems. More specifically,
these analytical data are processed and investigated in several different ways, as listed below.
Note that many of these concepts were previously introduced in Section 1.2 as the major
questions to be answered in NCHRP Project 12-113.
• Maximum stress ranges are compared for the various bridge geometries, such that the
impact of each parameter can be evaluated with respect to cross-frame behavior.
• Maximum stress ranges are compared for the various cross-frame members, such that
generalized observations about which members typically govern fatigue design can be made.
• The governing lane positions corresponding to those maximum stress ranges are compiled
for each representative bridge, such that the AASHTO fatigue loading model can be assessed in
terms of truck positioning.
• Factored stress ranges (based on current AASHTO LRFD fatigue loading criteria) are compared
to factored resistances, such that the overall efficiency of cross-frame design can be assessed.
• Factored stress ranges, determined by analytical methods, are compared to effective and
maximum stress ranges obtained from WIM records and the field experiments, such that the
accuracy and appropriateness of the current Fatigue I and II load criteria for application to
cross-frames can be assessed.
Table 2-3. SPS sites from which WIM data was obtained by FHWA.
vehicles. The records used for the fatigue study contain approximately 11 million truck measure-
ments after all appropriate filters are applied.
Figure 2-7 shows the cumulative distribution functions (CDFs) of GVWs captured by the
WIM sensors for all SPS sites. The GVWs are shown on a normal probability plot, in which
the horizontal axis is GVW (in kips), and the vertical axis is the standard normal variable
(i.e., axis values are “Z-values” indicating the number of standard deviations the GVW value
is from the mean of the distribution). A normal probability plot can be used to determine how
well the data represents a normal distribution; nonlinear data sets indicate departures from
normality. As clearly shown, the GVW populations are not normally distributed. This likely
is the result of natural groupings of different vehicle types and payload. The mean GVWs for
the 11 million truck records range from 40 to 62 kips, with a maximum GVW of 220 kips. As
documented in Appendix F, the shape of the CDFs appears to be generally consistent with
the CDFs of WIM data used in the SHRP 2 R19B project (Modjeski and Masters 2015), which
were recreated by the research team with the same filters described above. Thus, the 2014
data obtained from FHWA that were used on this project were deemed acceptable given their
good agreement with the SHRP 2 R19B data.
2
Standard Normal Variable
-1
-2
-3
-4
-5
0 50 100 150 200 250
GVW (kips)
Figure 2-7. CDF of GVWs from FHWA 2014 data (excluding light vehicles).
The research team created scripts to perform the following basic load configuration routines:
• Load Configuration 1 – Single Traffic Stream: This routine steps a stream of user-defined WIM
traffic on a bridge deck along an influence surface at 1-foot longitudinal increments in any
defined transverse position (also 1-foot increments). The script uses a cluster analysis to
include the effects of groups of vehicles in the same traffic stream, provided that any wheel
of the following truck is on the bridge during the time window in which the leading vehicle
is still on the bridge. Time windows are calculated based on the respective vehicle speeds.
• Load Configuration 2 – Two Traffic Streams: Similar to Load Configuration 1, a WIM traffic
stream is stepped along a bridge deck in a defined transverse position. The script uses a cluster
analysis to include the effects of groups of vehicles in any adjacent, user-defined transverse
position, provided any tire on the following truck is on the bridge during the time window
in which the leading vehicle is still on the bridge. Time windows are calculated based on the
respective vehicle speeds.
• Load Configuration 3 – Realistic Meandering Traffic Stream: This routine systemically steps
a stream of WIM traffic along a defined influence surface at 1-foot longitudinal increments
in any defined 12-foot-wide lane position. The routine randomly selects a transverse position
(within the 12-foot-wide lane) for each vehicle record based on an assumed distribution,
such that effects of lane meandering are considered due to cross-frames being highly sensitive
to transverse position. This script uses a cluster analysis to include the effects of groups of
vehicles, provided any of a following vehicle tires are on the bridge during the time window
that a leading vehicle is still on the bridge. Time windows are calculated based on the respec-
tive vehicle speeds.
Recalling that the influence-surface output from Abaqus provides results that are relatively
sparsely gridded (Section 2.3.1.4), the scripts use bi-linear interpolation methods to re-mesh
the influence surfaces to a 1-foot by 1-foot grid. This procedure is similar to that described
in Section 2.3.1 for the single AASHTO fatigue truck.
The output of each load configuration routine above includes the following:
• A sample load event history (see Figure 2-8);
• The total number of stress cycles (using rainflow-counting techniques discussed in Appendix F);
• The average number of cycles per passage;
• The maximum stress and stress cycle recorded;
• The equivalent stress range using the Palmgren-Miner damage accumulation model (i.e.,
stress range corresponding to the Fatigue II limit state); and
• The lowest stress range of the top 0.01% of all stress ranges (i.e., the 99.99th percentile criteria
corresponding to the Fatigue I limit state).
Past research has indicated that eliminating smaller stress cycles of a variable-amplitude
loading source has negligible effect on the damage accumulated in a fatigue detail. Connor
and Fisher (2006) showed that stress cycle magnitudes less than 25% of the constant ampli-
tude fatigue limit (CAFL) generally have little impact on the long-term fatigue performance,
2.0
Cross-Frame Axial Stress (ksi)
1.8
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0
2000 2200 2400 2600 2800 3000 3200
Load Step (1-ft Increment)
Figure 2-8. Sample load event history showing three vehicles
back to back.
which is outlined in Appendix E. To study the effects of this lower stress range truncation
on cross-frames, the research team compiled the results of each loading iteration (as listed
above) for two different conditions: (i) including the effects of stress magnitudes less than
25% of the CAFL and (ii) filtering out those effects. For the purposes of this study, a detail
Category E′ is assumed based on the results of McDonald and Frank (2009) and AASHTO
LRFD Table 6.6.1.2.3-1 (Condition 7.2). The corresponding CAFL value for this detail is 2.6 ksi;
thus, all stress cycles less than 0.65 ksi are truncated when the filter is applied. For the remainder
of this report, the data discussed refers to the unfiltered data (i.e., no stress ranges are truncated).
A more detailed discussion of the effects of stress truncation is provided in Appendix F.
The routines discussed in the previous section systematically and sequentially apply a defined
traffic stream in a specific load configuration over the influence surface corresponding to one
cross-frame member in one of the 20 representative bridges. Considering the computational
time for each iteration (depending on the total number of vehicle entries and bridge length) the
number of potential iterations becomes unmanageable. For this reason, the studies presented
in this report were obtained by positioning the WIM traffic stream in a realistic drive lane,
rather than a worst-case drive lane.
The worst lane position is often when the truck is positioned at the edge of the deck (e.g., the
right tire of a truck is positioned along the deck edge where a traffic barrier would be placed).
As a worst-case lane position would lead to unnecessary conservatism, the research team opted
for defining a realistic lane position based on the actual bridge width and recommendations
from AASHTO Article 3.6.1.1.1. As illustrated in Figure 2-9, this realistic truck position does
not produce the largest tensile stress range for the selected cross-frame; however, this position
would be consistent with realistic loading and lane striping on the bridge (and thus consistent
with the intent of the fatigue limit state).
With that in mind, the research team deemed Loading Configuration 3 introduced above
to be more pertinent with respect to fatigue design. As such, the focus of the study presented
herein is on realistic traffic streams as opposed to considering all possible lane positions (i.e.,
overhang loads).
For the actual location of each 6-foot vehicle track width (i.e., transverse distance between
left and right wheel lines) within the realistic lane, the research team assumed a distribution for
which the vehicle is located in the center of the lane 55% of the time; 30% of the time, the vehicle
is located plus or minus 1 foot of the lane centerline; 10% of the time, the vehicle is located plus
or minus 2 feet of the lane centerline; and 5% of the time, the vehicle is riding along one
of the lane edges. This is illustrated in Figure 2-10 for a sample 30-foot wide bridge, although
the procedure is similar for different bridge widths and lane configurations. The intent of this
assumed distribution, although not based on past precedents or measurements, is to consider
the inherent variability in driving ability and consistency.
2.3.3.3 WIM Drive Lane Truck Traffic Positioned in Realistic Drive Lanes
The traffic streams discussed herein exclude vehicles with GVW less than 20 kips and utilized
a drive lane defined by AASHTO Article 3.6.1.1.1, along with an assumed vehicle position
distribution discussed in Section 2.3.3.2.
For Fatigue I, the maximum stress ranges produced by the WIM traffic streams were normal-
ized to the maximum stress range produced by applying the unfactored AASHTO design load
(i.e., HS-20) to the same bridge in all possible transverse positions within the clear distance of
the barriers. An impact factor was not included since the WIM stations attempt to correct
for dynamic effects and relate measured drive-by weights to static weights. Note that it may
be prudent to include a portion of the typical 0.15 impact factor (AASHTO Article 3.6.2.1)
in subsequent studies, since the effects of bridge dynamics are not accounted for explicitly in
this FEA.
The governing Fatigue I maximum stress ranges calculated for each individual WIM site
and bridge combination were determined based on a 99.99th percentile criteria. This stress
range corresponds to the lowest magnitude of the top 0.01% of all stress ranges recorded (for
the governing cross-frame for each bridge given the defined lane position). In other words,
one critical cross-frame produced WIM stress ranges that exceeded the remainder of the cross-
frames in the bridge and thus served as the governing case.
Similarly, the governing Fatigue II stress ranges calculated for each WIM site and bridge
combination were determined based on the maximum effective stress range. Recall that
the Fatigue II limit state is intended to represent the average effect of the traffic spectrum.
These effects are generally characterized by the effective stress range (or equivalent stress range)
of the variable-amplitude response. Appendix F discusses how the effective stress range is
obtained from a variable-amplitude spectrum via rainflow-counting techniques and the use of
Palmgren-Miner’s rule. In order to compare the WIM force effects to the unfactored AASHTO
design truck force effects, it is more appropriate to compare the accumulated fatigue damage,
rather than just the effective stress ranges. The accumulated fatigue damage metric inherently
considers both the variable stress range magnitudes and the number of cycles. Thus, the total
damage accumulated by the various WIM streams on the critical, governing cross-frame
members is compared directly to the damage caused by the AASHTO fatigue truck.
Passing Truck
Passing Truck
indicate the front axles of the passing truck are “ahead of ” the rear axles of the drive-lane truck.
This provides a smooth, continuous function of clear distances, with increasing clear distances
(positive or negative) indicating a larger separation between vehicles.
Note that a clear distance of zero corresponds to a staggered configuration, which was
deemed critical to cross-frame force effects in the 7th Edition AASHTO LRFD Specifications.
This provision was removed in the 2016 Interims to the 7th Edition Specifications, citing too
infrequent occurrences. This study herein evaluates the frequency of occurrence in the context
of cross-frame response. Results of the multiple presence study are subsequently summarized
in Section 3.2.3.
Differential Vertical
Equal Rotation Differential Rotation
Displacement
θy θx2
θx
θx θx1
θy
In-
Plane
Out-of-
Plane
Figure 2-13 divides the deformations into in-plane and out-of-plane deformations. The
in-plane rotations depicted in Figure 2-13 are attributed to the in-plane rigidity typically provided
by the connection and gusset plates. In order to accommodate the girder displacements and
compatibility of the system, the cross-frame members have to flex given the rotational restraint
provided by the end connections. Given that the in-plane rotational stiffness of the connection
plate and welds is generally large (i.e., full-depth web stiffener welded along three sides),
this behavior is likely observed in all practical steel I-girder systems in service. This behavior,
however, is typically ignored in the truss-element modeling approach.
Out-of-plane rotations, on the other hand, are attributed to the eccentric nature of the connec-
tions. As internal force is transmitted from the girder webs into the cross-frame members, the
load must pass several “eccentric jumps.” These “jumps” of potentially several inches in magnitude
can have a considerable impact on the stiffness response of cross-frame members and systems.
The stiffness modification factors (R-factors), detailed in the next subsection, were primarily
developed to account for this modeling oversight in truss-element models.
for the softening effect of eccentric end connections (i.e., the R-factor is often assigned to
reduce the cross-sectional area of the truss element or the elastic modulus of the material).
As such, this aspect is a well-researched behavior.
• The truss-element modeling approach neglects the in-plane bending effects caused by the connec-
tion rigidity. Unlike the out-of-plane effects, the stiffness modification factors formulated
by Battistini et al. (2016) did not explicitly address these effects. Instead, in-plane rotational
behavior is implicitly considered in those previous studies.
• The truss-element modeling approach neglects the axial stiffness of the connection and gusset
plates in favor of the cross-frame member axial rigidity. In other words, truss-element models
replace the stiffer connection and gusset plates with an additional length of an often, more
flexible cross-frame member. Again, this behavior was not explicitly considered in the
development of the R-factors but was implicitly considered in the corresponding laboratory
and analytical studies.
To expand on the previous work, two generalized solutions were explored in NCHRP
Project 12-113. First, stiffness modification factors similar to those derived by Battistini et al.
(2016) were analytically developed in the context of composite bridge systems. To accomplish
this, a series of 3D analyses in Abaqus and approximate hand calculations were conducted at
three different model-scale conditions: (i) member-level, (ii) panel-level, and (iii) system-level
studies. The second generalized solution is a proposed alternative to the truss-element modeling
approach that makes use of eccentrically loaded beam elements for the cross-frame members.
These procedures are briefly outlined herein.
Pin-Ended Truss-Element
Shell-Element Eccentric-Beam
with R-Factor
y–
Aa
RAa
from more refined to relatively simple. Each variation (i.e., the refined, simplified, and angle-
only models) is presented schematically in Figure 2-15. This figure schematically depicts
these different modeling methods for a sample cross-frame panel. Note that the “eccentric
jumps” only apply to the out-of-plane direction, as the behavior is assumed concentric in
the in-plane axis.
To illustrate some of the assumptions used in the development of these models, an abbre-
viated set of notes is provided below. This discussion specifically highlights key aspects of the
“refined” model, as defined by Figure 2-15, but are also directly applicable to the “simplified”
and “angle-only” approaches, where appropriate. For a more detailed overview, as well as
sample calculations demonstrating the proposed methodologies, refer to Appendix F.
• In the “refined” model, the cross-frame member, gusset plate, connection plate, and
individual overlapped sections are treated independently when developing the eccentric-
beam model.
96″
Connection plate
Overlap 1
Gusset plate
Overlap 2
L4x4x3/8
60″ Girder web shell (typ.)
Refined Model
Connection plate
L4x4x3/8
8″
Simplified Model
9/16″ PL.
4″
1″ L4x4x3/8
L4x4x3/8
16¼″ ½″ PL.
Pinned or fixed
L4x4x3/8
Angle-Only Model
7¼″
Note: Out-of-plane view shown; all
“offset” links are rigid elements
• The out-of-plane “jumps” in the neutral axis, or the eccentricities, are represented by rigid
offsets, which imply that the welded or bolted connections between the plates and angles
perfectly constrain these elements together (i.e., no slip is assumed in the bolted connections).
• The lengths of the rigid elements are based on the distances between the neutral axes of the
connected components.
• The length of each individual beam component is assigned assuming there are no shear lag
effects. In other words, the length of the angle member starts precisely at the termination of
the gusset plate and ends at the gusset plate edge on the opposite side of the angle.
• Section properties (i.e., A, out-of-plane I, in-plane I, and length L) of the connection and
gusset plates are based on a Whitmore approach. Section properties of the equivalent
overlapped portions are taken from an assumed composite section. A detailed calculation
demonstrating this process is provided in Appendix F.
• Unlike the pin-ended truss models, rotations are not released at both ends of the eccentric-
beam model. The distortional stiffness of the girder web is assumed substantial enough to
warrant in-plane and out-of-plane rotational restraint of the cross-frame system.
• In X-type cross-frames, the diagonal members are connected on opposite faces of the cor
responding gusset plates to avoid interference at the crossover point. These different
eccentric effects were considered accordingly.
• Although not shown here, a similar procedure was also used to develop equivalent beam
properties for K-type cross-frames.
This proposed 3D modeling technique was evaluated for a variety of bridge geometries,
as outlined in the next subsection.
bridge structures with various gusset plate thicknesses. More specifically, the sensitivity of
the load-induced cross-frame response due to the assigned R-factor was examined.
To make those assessments, a variety of analytical studies were conducted as part of the
system-level study. A parametric study (consisting of only straight and normal bridges) was
performed to emphasize the influence of connection and gusset plate dimensions on this
behavior. Additionally, a second study was conducted to further investigate the effects of
support skew and horizontal curvature on the R-factor and eccentric-beam approaches.
For each individual study, the research team developed and compared multiple versions of
the same 3D bridge model. Each iteration investigated a different approach to modeling the
cross-frame elements, while the remainder of the 3D model, as outlined in Section 2.3.1, remained
unchanged. In general, the rigorous shell-element model served as the benchmark model to
which the rest of the iterations were compared. Truss-element models were also developed where
the assigned R-factor was varied. Lastly, the proposed eccentric-beam modeling approach was
also implemented for the various bridge models. Thus, for each individual study, the following
model iterations were considered:
• Shell-element model (representing the most accurate solution),
• Truss-element model with different stiffness modification factors {R = 0.5, 0.6, 0.7, 0.8,
0.9, 1.0}, and
• Eccentric-beam model.
For the shell-element and eccentric-beam modeling approaches, different values of the gusset
plate thickness were also considered as a variable. Note that the gusset plate thickness and the
corresponding effects on the load eccentricity were explicitly considered in the shell-element
model. In the eccentric-beam model, the eccentric offset dimension, as illustrated in Figure 2-15,
and the equivalent beam properties were adjusted accordingly. Gusset plates were not explicitly
represented in the truss-element models.
Ultimately, the effectiveness of the simplified 3D model (i.e., the truss-element model with
an assigned R-factor or the eccentric-beam model) is evaluated as a ratio between the predicted
axial force in select cross-frame members from the simplified models to the predicted force in
the same members from the shell-element model. This is described algebraically as Fsimplified/Fshell.
Rather than compare the stiffness of the models indirectly, it is more appropriate to evaluate the
ability of the simplified model to accurately obtain design forces, which is of most importance
to designers.
A ratio of unity represents perfect agreement between the shell-element model and the
simplified models. This implies that the approximated stiffness is identical to the “true”
stiffness of the panel. Values below unity indicate that either: (i) the assigned R-factor is
likely too low for the truss-element model or (ii) the equivalent section properties assigned for
the eccentric-beam model are too low (i.e., the cross-frame attracts less force). In other words,
the simplified models underpredict the cross-frame force when compared to the more accurate
shell-element model (i.e., unconservative estimates), and an increase in the modification factor
or equivalent beam properties is needed. The opposite is true for force ratios above unity.
In all cases, the applied load on these composite bridge systems represented fatigue design
load conditions. Thus, the cross-frame force effects compared between the various modeling
approaches were based on the same corresponding load cases (i.e., same truck configuration,
weight, and lane position). For more information on the parameters considered and the perti-
nent loading conditions, refer to Appendix F. The results of this computational study are sub
sequently summarized in Section 3.3.
from comparing specific commercial design software packages and instead highlighted the
modeling approach, which is of more general value to designers and software developers.
In total, three distinct types of analyses were performed in Software A as part of this study,
including 3D models (as a direct comparison with the 3D truss-element models produced in
Abaqus and documented in Section 2.4), 2D plate and eccentric-beam (PEB) models, and 2D
grid/grillage models. The following subsections outline key attributes related to these 2D and 3D
modeling procedures. Section 2.5.1 briefly introduces the major assumptions and features related
to PEB and grillage models, as well as explains the subtle differences between the 3D models
developed in Software A and Abaqus (Section 2.4). Section 2.5.2 then expounds on the equiva-
lent beam approach for modeling cross-frames that is inherent with any 2D analysis. Finally, Sec-
tion 2.5.3 offers techniques commonly used to improve the 2D equivalent beam approach, and
Section 2.5.4 outlines the parametric studies used to evaluate these analytical techniques. Note
that the general approach for these simplified methods is largely based on the guidelines docu-
mented in AASHTO G13.1 Guidelines for Steel Bridge Analysis (2019) and White et al. (2012).
2.5.1.1 3D Models
Three-dimensional modeling of bridge systems in any commercial design software (e.g.,
Software A) can vary from package-to-package and engineer-to-engineer. The accuracy of the
model as it pertains to cross-frame forces can be sensitive to the assumptions made by the engi-
neer. With this in mind, the intent was to provide a modeling approach that is representative
of most 3D software packages and assumptions likely utilized by bridge engineers. However, it
is important to note that variability in the results should be anticipated for a different software
package or a different set of assumptions.
The 3D models in Software A were developed similarly to the 3D Abaqus model (Section 2.3.1)
with a few notable exceptions. First, girder flanges were modeled as beam elements rather than
shells. Second, rather than framing cross-frame members into the stiffened girder web (i.e., how
cross-frames are fabricated and erected in practice), cross-frames were framed into the shared
node at the girder web-to-flange juncture to maintain consistency with most 3D bridge-related
software programs. Upon developing the model, the built-in influence-surface feature of Soft-
ware A was utilized to move a 1-kip load along the length and width of the bridge deck, similar
to the procedures outlined previously.
the beam element representing the girder between two adjacent cross-frame intersections
on that beam. This is particularly important for horizontally curved bridge systems, which
generally model the girders as a series of chorded, straight-line segments (White et al. 2012).
The influence-surface loads were applied directly to the deck shell elements, in the same manner
as the 3D model.
Analysis Method
Element
Control 3D 2D PEB 2D Grillage
Concrete deck Shells Shells Shells --a
Girders Shells Shells/beams Beam elementsb Beam elementsb
Cross-frames Truss elements Truss elements Equivalent beamsc Equivalent beamsc
Notes:
a
Grillage models do not explicitly consider the concrete deck.
b
In the 2D PEB model, beam elements represent the steel section alone; in the 2D grillage model, beam elements
represent the effective composite section.
c
There are several methods by which the equivalent beam section properties are computed, as discussed in
Section 2.5.2.
and converted back into an idealized truss system for obtaining internal member forces. With
respect to this transformation, there are two major questions that often arise when imple-
menting these procedures into 2D models. The first is related to the section properties of the
equivalent beam elements. Several different methods are commonly used to compute the
equivalent moment of inertia, torsional constant, and shear area as outlined in Section 2.5.3.
The second question is related to the postprocessing procedures once the shear forces and
end moments in the equivalent cross-frame beams are obtained from the 2D models. Typically,
these end moments and shears are applied as external loads on an idealized truss model, from
which internal axial forces in struts and diagonals are computed. The end moments are often
resolved as a force couple between the top and bottom nodes of the truss, resulting in equal-
and-opposite forces in the top and bottom struts. For X-type cross-frames, the end shears are
equally distributed between top and bottom nodes resulting in equal-and-opposite diagonal
member forces. In many cases, these behaviors are seldom observed. For K-type cross-frames,
the vertical shear component is resisted entirely by the single diagonal member framing into a
given end, such that no distribution assumption is required.
Aside from the general procedures, this equivalent beam approach is also thought to produce
cross-frame force results with varying levels of accuracy, particularly for heavily skewed and/or
curved bridges. With the background information outlined, the next subsection offers and
explores specific improvements on these simplified 2D methodologies.
force predictions. For completeness, the models and corresponding results presented later in
this report include equivalent torsional properties.
3. Consider the transverse stiffness of the concrete deck [Analysis]. Given that 2D PEB models
represent the deck as a shell element, this technique is only applicable to grillage models. In
this modification, the transverse stiffness of the deck can be simulated using (i) equivalent
section properties or (ii) notional beams.
For the first approach, the section properties of the equivalent cross-frame beam can be
modified by simply summing the contributions from both the cross-frame (as previously
determined by the Timoshenko beam approach) and the deck, whose effective width is taken
as half the distance to the nearest cross-frame on each side. Thus, the additional moment
of inertia, shear area, and torsional constant contributed by the applicable concrete deck
strip (as modified by the appropriate modular ratio) is added to the equivalent cross-frame
beam properties. In addition to the effective section property approach, consideration of
the transverse deck stiffness can also be implemented with separate, notional transverse
beam elements in the grillage model per AASHTO Article C4.6.3.3.4.
4. Reconsider the distribution of shear forces and moments when postprocessing results [Post
processing]. As introduced previously, the end moments obtained from 2D analyses are often
resolved as equal-and-opposite force couples to the top and bottom nodes of the truss, and
end shears are assumed equally distributed to top and bottom nodes for X-type cross-frames.
These assumptions, although simple to implement, can produce significant errors for com-
posite bridge systems as is shown by the results herein. Three-dimensional FEA results and
measured data have shown that top and bottom struts generally do not have equal-and-
opposite forces; similarly, diagonal members generally do not have equal-and-opposite forces
in X-frames. As such, a unique postprocessing tool is proposed for both 2D grillage and
2D PEB models, as illustrated in Figure 2-17.
The grillage improvement technique (left side of Figure 2-17.) is related to the equivalent
section properties approach outlined in item #3 above (i.e., considering the contributions
of the transverse deck stiffness). By effectively increasing the stiffness of the equivalent beam
to account for the contributions of the concrete deck, considerations for the deck must also be
made when processing 2D analysis results. Otherwise, the force demands on the cross-frame
elements will be artificially amplified due to the increase in assigned stiffness. In contrast,
the methodology below is not applicable to the notional beam approach introduced in
item #3. For that case, the end moments and shears obtained for the equivalent cross-frame
beam can be resolved directly to the cross-frame truss model given that the contributions
of the concrete deck are explicitly and independently considered in analysis.
As such, rather than applying the end moments and shear forces on the cross-frame
panel alone, the applied loads are resolved on a pseudo-composite system including the
cross-frame and deck. The end moment is resolved between the centroid of the deck and
the centroid of the bottom strut, which results in a moment arm of H (or the distance
M1 / H
M1 / H FBS M1 / hb FBS
between the respective centroids). This modification is intended to remedy the problem of
having equal-and-opposite strut forces in composite systems. It should be noted that in
noncomposite systems, equal-and-opposite force distribution in the struts is anticipated.
Thus, it is apparent from this simple discussion that the behavior of cross-frames in non
composite and composite systems is different and should be treated as such in the cor
responding analysis. This discussion is advanced further in subsequent sections.
Additionally, in this proposed technique for 2D grillage models, the shear force acting
on the equivalent beam must be distributed to the concrete deck and cross-frame panel.
The research team utilized a rational approach in which the shear force is distributed
based on the relative stiffness of the deck and cross-frame to determine Vdeck (sheer force
component resisted by the concrete deck) and VCF (shear force component resisted by the
cross-frame), although other methodologies are possible. These expressions can be found in
Appendix F. The distribution of VCF to the top and bottom nodes of the cross-frame truss is
discussed below.
The PEB postprocessing technique (right side of Figure 2-17) differs from the grillage
technique in that the internal moments and shears obtained from the analysis (via equiva-
lent beams) only consider the contributions from the cross-frame panel. Recall that the
stiffness of the concrete deck is explicitly considered by shell elements in PEB models.
Thus, in the process of converting end moments and shears into cross-frame force effects,
the added discussion about the moment arm H above is not relevant.
With that in mind, the remaining discussion relates to how the VCF acting on the cross-
frame panel is distributed to the top and bottom nodes. This significantly impacts the
assumed force effects in the cross-frame members, particularly for the diagonals since the
force effects in these members are directly related to the vertical component of the nodal
forces. The research team explored several solutions to this problem and compared the
behavior of noncomposite and composite systems. Ultimately, the equal distribution “50-50”
assumption (i.e., 0.5VCF to the top node and 0.5VCF to the bottom node) is compared to
a “100-0” assumption (i.e., conservatively and independently evaluate the cases in which
100% of VCF is resisted by the top node and 100% of VCF is resisted by the bottom node).
This is demonstrated schematically in Figure 2-17, where the percentage of shear force
distribution is represented by the variable n.
Table 2-5 summarizes the analysis-related improvement techniques outlined above and the
relative impact on deck, girder, and cross-frame elements. Blank table entries indicate that no
modifications are made to that particular element.
2.4 LM r2
b T,req = 2.1
fnEI yeff Cb2
where
bT,req = required system torsional brace stiffness,
L = span length,
Mr = maximum factored moment within the critical unbraced segment,
n = number of intermediate braces within the span,
E = the modulus of elasticity of steel (29,000 ksi),
Iyeff = effective moment of inertia,
f = resistance factor, taken as 0.75, and
Cb = moment gradient factor assuming the beam buckles between the brace points.
These variables, in the context of single- and reverse-curvature conditions, are discussed
in Section 3.5.
Although this “twice the ideal stiffness” assumption works well for columns, studies on
beam torsional bracing have shown a larger value may be warranted. In response, the research
team conducted a parametric study that investigated the effects of girder cross-section, loading
conditions, intermediate bracing schemes, girder spacing, and number of girders on the
required stiffness of a torsional beam brace (i.e., a cross-frame). A brief overview of the research
methodology used in this study is provided in Section 2.6.3.
While a reasonable view of stiffness requirements might focus solely on the stiffness of the
cross-frame, in reality, the total torsional brace stiffness is a function of several components.
The total torsional stiffness of a brace is a combination of three main components including:
(i) brace stiffness, b b, (ii) cross-sectional distortion stiffness, bsec, and (iii) in-plane girder stiff-
ness of the beams, bg. The AISC bracing provisions do not include the expression for in-plane
girder stiffness since that component is not a major factor in most building applications.
However, for bridge applications, bg can significantly impact the behavior, particularly for
relatively narrow girder systems (Yura et al. 2008; Han and Helwig 2016). The individual
stiffness components tend to follow the expression for springs in series, as demonstrated
mathematically with the following expression:
1 1 1 1
= + + 2.2
b T b b b sec b g
A review of Eq. 2.2 demonstrates that the total stiffness of the system, b T, is always less than
the smallest of the three individual terms on the right of the equation. For example, a stiff cross-
frame with flexible connection details will severely limit the stiffness response of the system;
or a flexible in-plane girder stiffness, which effectively equates to a narrow superstructure, will
diminish the adequacy of a cross-frame. The previous research related to these topics is sum-
marized by Helwig and Yura (2015). While Eq. 2.2 provides the total stiffness of the system, in
design, the actual brace stiffness must exceed the required stiffness, as determined by Eq. 2.1
(i.e., b T ≥ b T,req).
In addition to stiffness requirements, stability bracing must also satisfy strength require-
ments. In general terms, the required brace strength is a function of the required stiffness and
the assumed initial imperfection. A distinct change in bracing strength requirements, however,
occurred between the 14th Edition AISC Specifications (2010) and the current 15th Edition
(2016). Applying some simplifications which are documented in the commentary, the original
2010 version of the torsional brace strength requirement (based primarily on elastic buckling
behavior) was as follows:
0.024 M r L
M br = b Tq o = 2.3
nCb Lb
where
Mbr = required strength of a torsional brace,
b T = required system torsional brace stiffness (based on twice the ideal stiffness),
qo = initial imperfection in terms of a twist angle, and
Lb = unbraced length of the critical segment.
The remainder of the variables were previously introduced in Eq. 2.1. Note that the initial
imperfection is based on the critical imperfection shape, which for torsional bracing consists
of a lateral sweep of the compression flange equal to Lb/500 (sweep tolerance) while the tension
flange remains straight (Wang and Helwig 2005). This produces an initial twist equal to Lb/500ho,
where Lb and ho are the respective unbraced length and distance between flange centroids.
In contrast, the latest AISC Specification (2016) introduced a change in the torsional brace
moment equation based upon a study conducted by Prado and White (2015). The researchers
carried out a detailed investigation on the stability bracing requirements with an emphasis on
inelastic buckling of relatively short unbraced lengths. This research prompted the revision in
the torsional brace moment equation, which is also given as Eq. A-6-9 in the current AISC
Specifications:
M br = 0.02 M r 2.4
This expression is simply a function of the internal factored moment in the critical unbraced
segment. Although the simplicity of Eq. 2.4 is attractive, the applicability of the expression for
general design situations is questionable when compared to the longstanding strength equations
predicted by Eq. 2.3. To determine which strength design expression is more appropriate for
implementation into AASHTO LRFD, an additional parametric study was conducted in tandem
to NCHRP Project 12-113 (Liu and Helwig 2020).
As noted above, LTB in bridge applications is most critical during girder erection and deck
construction. Although the bending moments during construction are smaller in magnitude
than the live load moments in the completed structure, the noncomposite girders are most
susceptible to instability at this stage. During construction, the steel section alone generally
supports the entire load, and all permanent bracing may not be installed. Additionally,
stay-in-place forms, commonly used to support wet concrete during deck construction, have
connections to the girder that potentially can introduce significant flexibility and therefore are
not considered bracing elements in bridge applications (Egilmez, Helwig and Herman 2016).
While bracing demands are often considered most critical in positive moment regions, both
the positive and negative moment regions in continuous girder systems need to be considered.
In the finished structure, the composite deck provides significant continuous lateral and
torsional restraint to the top flange. For simple spans, for which only the top flange is in
compression, girders are not prone to LTB. In continuous spans, the stability behavior of
continuous girders in the negative moment region (i.e., the bottom flange is in compression) is
often questioned by designers. As a result, cross-frames are often provided in negative moment
regions to control LTB, and the substantial restraint provided by the deck to the bottom flange
is conservatively neglected in design. There are, however, a number of beneficial restraints
that can be considered around interior supports of the finished structure.
As discussed by Yura (2001), the composite deck not only continuously braces the top
flange but also provides additional bracing benefits to the bottom compression flange in the
negative moment regions, assuming web distortion is prevented. Additionally, the girder
bearings themselves provide lateral restraint and some torsional restraint to the bottom
compression flange in these regions, which further mitigates an LTB problem. Therefore, the
stability bracing requirements for cross-frames in negative moment regions of composite
systems are not critical for design.
To demonstrate the beneficial effects of continuous top flange restraint, Figure 2-18
schematically depicts the results of an eigenvalue FEA buckling analysis on a bridge girder
with stiffened webs subjected to reverse-curvature bending, which is representative of many
continuous systems near the interior supports. In all three cases presented, twist of the cross-
section is prevented at the girder ends. However, the level of top flange restraint is modified in
each successive case.
From the buckled shapes, it is evident that cases (i) and (ii), which are representative of the
steel section alone and a noncomposite girder condition (i.e., a concrete deck without shear
studs), respectively, tend to buckle in a traditional LTB (or similar) mode. In these cases, the
unrestrained compression flange(s) displaces laterally in the out-of-plane direction along the
full unbraced length. In case (iii), the buckled shape is indicative of a distortional mode in the
web, which is not sensitive to the unbraced length but is sensitive to the web slenderness and
transverse stiffener details. The buckling capacity of these cases generally increases from left to
right, as presented in Figure 2-18. More specifically, the bucking capacity associated with case
(iii) is typically substantially larger provided that transverse stiffeners are included and are
properly detailed. As such, girder segments in the negative moment regions are often controlled
by yielding or the distortional buckling mode, for which bracing demands in the cross-frames
are much less significant than the construction condition. This phenomenon is supported with
additional analyses outlined in Section 2.6.1.
The critical stage, therefore, occurs during erection and deck construction with the non-
composite steel girder system supporting the entire construction load. With that in mind, the
studies outlined herein are focused primarily on noncomposite systems under various moment
gradients including single- and reverse-curvature bending, as these conditions represent the
most critical in terms of bracing requirements.
Before outlining the stability bracing studies, a cursory overview of a computational study
investigating buckling in composite conditions is provided in Section 2.6.1. An overview of the
parametric studies conducted to address these bracing strength and stiffness requirements are
then provided in the following subsections. Section 2.6.2 outlines the bracing strength study,
and Section 2.6.3 outlines the bracing stiffness study.
flange restraint, which is representative of the steel section alone during deck construction,
(ii) continuous lateral restraint to the top flange, which is representative of noncomposite
finished bridges not utilizing shear connectors, and (iii) continuous lateral and torsional restraint
to the top flange, which is representative of a composite bridge in its finished state. Case
(iii) represents an upper-bound condition in terms of restraint fixity, as the composite deck
does not supply a completely rigid torsional restraint to the top flange.
For the various parameters and moment gradients, the critical buckling moment was
evaluated and compared as a function of the top flange restraint provided. The critical eigen
vector either corresponded to an LTB mode or a web-distortional buckling mode, similar to
the illustrations provided in Figure 2-18. The results of this study are subsequently presented in
Section 3.5.1.
Flanges:
8"×0.5" 12"×0.75" 16"×1"
0.75"
48"
supported at the ends. For the cases involving reverse curvature, the girders were still simply
supported, but girder continuity was simulated with applied end moments (positive and
negative). Twist was also restrained at the girder ends, but the sections were free to warp.
For the single-curvature bending cases, it was observed that the critical segment for buckling
that consistently resulted in the largest brace forces was near midspan, where positive moment
and out-of-plane girder displacements were maximized. As such, a critical asymmetric imper-
fection consistent with Prado and White (2015) (i.e., the critical compression flange is dis-
placed laterally in accordance with the discussion in the preceding section) was assumed in these
critical areas.
For reverse-curvature bending, it is not always as clear which segment is critical for buckling.
In continuous girder systems, the negative moment regions generally have the largest moment
magnitudes. However, these regions are aided by steeper moment gradients (i.e., larger Cb factors)
and restraint provided by the nearby girder supports. In contrast, positive moment regions have
smaller moment magnitudes but typically have Cb values close to unity, especially as additional
intermediate braces are included. For example, in the interior span of a three-span continuous
unit, the maximum negative moment is approximately double the maximum positive moment.
However, the moment gradient factor in the negative moment region is often greater than 2.0
(depending on the unbraced length) based upon published values or AISC expressions, compared
to a Cb factor near unity for the positive moment region.
Thus, it was observed that the critical brace (i.e., maximum bracing forces) depended on
the location of the critical imperfection. For instance, the midspan brace forces in the positive
moment region were maximized when the imperfection was assumed along the compression top
flange at that same location. In contrast, the brace forces at the first intermediate brace line in the
negative moment region were maximized when the imperfection was assumed along the bottom
(compression) flange at that location. Note that the imperfection in the negative moment case
was not implemented directly at the support condition, where negative moment is actually
largest, due to the presence of bearings that benefits the torsional bracing requirements.
Permutations of different girder cross-sections, bracing configurations, and loading conditions
were systematically analyzed by a series of independent eigenvalue and incremental analyses.
Eigenvalue analyses were initially performed to obtain the ideal stiffness of the cross-frame
required to buckle the girders between the brace points. The large-displacement incremental
analysis, performed on an imperfect system and cross-frames with twice the predetermined
ideal stiffness, was subsequently performed to obtain the relationship between internal girder
moments and critical bracing moments. An additional discussion on modeling convergence
and inferring cross-frame brace forces from the results is provided in Appendix F for reference.
The results of this stability bracing strength study are subsequently presented in Section 3.5.2.
CHAPTER 3
As noted in Chapters 1 and 2, the five major objectives of NCHRP Project 12-113 are
addressed with both experimental and analytical studies. The experimental portions included
the instrumentation and monitoring of three bridges, while the analytical studies included four
independent studies (Fatigue Loading Study, R-Factor Study, Commercial Design Software
Study, and Stability Study). While Chapter 2 outlined the general means and methods for each
study, this chapter summarizes the key findings and results of those studies.
Similar to the previous chapters, the data presented herein is not intended to provide a
comprehensive overview of the experimental and analytical results. Rather, sample results are
provided that highlight the key observations to contextualize the conclusions and proposed
modifications to AASHTO LRFD presented in Chapter 4. For a more detailed overview of the
results, refer to Appendices E and F.
This chapter is divided into five major sections similar to the organization of Chapter 2.
Section 3.1 outlines the major results from the field experimental program. The controlled live
load test and the model validation studies are summarized, followed by the in-service stress
data that supplements the computational results. Section 3.2 presents the key results from the
Fatigue Loading Study with a particular emphasis on AASHTO fatigue loading criteria and
WIM records. Section 3.3 summarizes the investigation of the eccentric connections for cross-
frames and the proposed R-factors. More specifically, the reliability of the stiffness modification
approach for 3D modeling of cross-frames is explored, as well as a proposed eccentric-beam
approach. Section 3.4 reviews the data related to the Commercial Design Software Study, where
the limitations of various 2D modeling methods are examined. Lastly, Section 3.5 provides
a summary of the major findings related to the Stability Study. Note that the results of the
industry survey were previously discussed in Section 2.1, so no additional commentary is pro-
vided in this chapter.
60
Given the distinct differences in the two experimental tests, this section of the report is divided
into two major subsections. Section 3.1.1 highlights key findings from the controlled live load
tests, and Section 3.1.2 summarizes the measured rainflow-counting data considering one month
of measured traffic data on the instrumented components of the bridges.
Brg.
CF
2.40
1.60 TS2
D2-2 D2-1
0.80 BS2
TS2
0.00
-0.80
2.40
1.60
Axial Stress (ksi)
D2-2
BS2
0.80
TS2
0.00
D2-1
-0.80
2.40
1.60
0.80
D2-1
BS2
0.00
TS2
D2-2
-0.80
0 50 100 150 200
Distance from Start of Bridge (ft)
Cross-frame gages
Bottom flange gage
TS2
D4-2 D4-1 D3-2 D3-1 D2-2 D2-1 D1-1
D1-2
BS3 BS2
the figures presented in this section, both the static and moving-load cases, are based on instru-
mented cross-frames in Bridge 1 unless noted otherwise. The full set of field data related to all
three bridges can be found in Appendix E.
Three distinct loading scenarios are presented sequentially in the figure. In the first load
case (top plot), a three-axle dump truck (with an approximate gross vehicle weight of 50 kips)
slowly traversed the full length of the three-span continuous unit along the inside edge of the
left bridge barrier. Similar dump trucks subsequently traversed the full length of the bridge at
different transverse lane positions. For each scenario, the axial-stress time-history in the cross-
frame members of interest was recorded by the DAQ system.
The time-history responses were then converted into influence-line plots by aligning the
time component of the measured data with the longitudinal position of the truck relative to the
start of the span. This is demonstrated in the horizontal axis of Figure 3-1. Additional bench-
mark distances, such as cross-frame (CF) line 4 where the instrumented members are located,
and bridge supports (abbreviated as “Brg.” in the figures for “bearings”) are included for
reference. Note that only the instrumented 194-foot end span of the three-span continuous unit
is plotted along the horizontal x-axis. The influence of applied load on the adjacent two spans
with respect to cross-frame force effects was negligible.
It is also important to note that the data obtained from the strain gages illustrated in this fig-
ure simply provide the change in strain/stress during the applied loading. The data does not
indicate the state of stress prior to the gage being installed. As such, the measured response
does not provide information on the permanent state of stress due to dead loads or residual
stresses.
In review of the sample influence-line results provided in Figure 3-1, there are several key
observations:
• Load-induced cross-frame response is highly sensitive to the transverse position of the
truck. For instance, the cross-frame diagonal designated as “D2-2” experiences a tensile
stress cycle when the truck traverses along the left-hand side of the panel but experiences a
compressive stress cycle when the truck transverses along the right-hand side.
• The influence of longitudinal load position is localized, as demonstrated by the fact that the
measured axial stresses are nearly zero when the truck is positioned beyond 50 feet from
the panel of interest, regardless of the lane position.
• The various cross-frame members in the instrumented panel also have highly varied
responses to a given load condition. In general, top strut (TS) force effects are negligible,
which is attributed to the composite nature of the superstructure. Diagonal and bottom strut
(BS) forces are generally more substantial, but the stress magnitude and sign are dependent
on truck position and the corresponding load-induced deformation pattern of the cross-
frame panel.
• Although not explicitly shown in the figure, similar behavior was observed from the results
for the straight bridge with skewed supports (Bridge 2) and the horizontally curved bridge
with radial supports (Bridge 3). The overall load path and load-induced response of the cross-
frames in those more complex framing systems, however, produces more interesting results.
The effects of skewed supports and horizontal curvature are examined computationally in
Section 3.2.
In terms of the static load cases performed during the field experimental program, four
different dump trucks were incrementally positioned on the bridge deck one at a time. Thus,
for each individual load case, the result was a stepped time-history response of increasing
stress. An example of this is presented in Figure 3-3, where measured bottom flange stresses
in Girder 1 (measured at cross-frame line 4 in accordance with Figure 3-2) are graphed as a
function of time for a specific static load case. There are six distinct steps that are apparent
in the time-history plot, as follows:
A. No trucks positioned on bridge; no live load-induced stress.
B. Truck 1 (lead truck) positioned on the bridge; live load-induced stress increases.
C. Truck 2 positioned on the bridge behind Truck 1; live load-induced stress increases.
D. Truck 3 positioned on the bridge behind Trucks 1 and 2; live load-induced stress increases.
E. Truck 4 positioned on the bridge behind Trucks 1, 2, and 3; live load-induced stress increases.
F. All four trucks are removed from bridge simultaneously; live load-induced stress returns
to zero.
Due to the relatively static nature of this test, the stress response for each instrumented
element essentially follows a step function. Small spikes were periodically recorded because the
7.0
6.0 E
D
Bottom Flange Stress (ksi)
5.0
4.0 C
3.0
2.0 B
1.0
A F
0.0
-1.0
0 200 400 600 800 1000 1200 1400
Time (sec)
load was not purely static. For example, a small spike was recorded at the beginning of plateau B
in Figure 3-3. This spike occurs as the truck passes over the maximum point on the influence
line for the girder bottom flange, which happens to not coincide with the final static position of
Truck 1 in this sample load case. Had the final position of Truck 1 coincided with the maximum
point on the influence-line plot, the spike would not have been measured. There are also a few
spikes as the trucks are moved off the bridge, which can be explained the same way as the spike in
plateau B. Given that these load cases were performed at very low speeds, the research team does
not believe these spikes are related to a dynamic impact effect of the truck entering the bridge.
The magnitudes of the plateaus are the major focus of the model validation process. These
values indicate the stress imposed on the instrumented elements under static loading condi-
tions. In order to cancel out potential effects of electromechanical noise, an average stress
value, which served as the metric by which the FEA models were validated, was obtained from
each of these load-step plateaus.
An example of static loading results is presented in Figure 3-4. The measured axial stresses
in the instrumented cross-frame members along line 4 are graphically depicted for each
intermediate stage of a load case. The positioning of the trucks relative to the cross-frame line
of interest is depicted in Figure 3-2 for reference. From these results, there are a few notable
observations with regard to the cross-frame response:
• Similar to Figure 3-3, the single instrumented top strut member developed very little axial
force, even for the final stage of the load case that represented over 200 kips of GVW.
0.09
1 Truck
0.47 0.43
0.43
2 Trucks
1.42 1.30
0.59
3 Trucks
2.01 1.80
0.65
4 Trucks
2.24 1.99
• Interestingly, all but two instrumented cross-frame members experienced a net tensile stress
under the applied load. This behavior seems to indicate that the simplified postprocessing
methods commonly utilized in 2D analysis programs as outlined in Section 2.5, do not accu-
rately represent realistic conditions. In other words, the diagonal members and strut members
do not have equal-and-opposite force effects under a given applied load.
Cross-frame stresses (e.g., Figure 3-4), girder flange stresses, and girder deflections were
compiled from every static load test performed at all three instrumented bridges. This measured
data served as the metric by which the FEA modeling approach utilized throughout the Phase III
computational studies was validated.
Before the experimental program was executed, preliminary 3D FEA models were developed
based largely on the properties and dimensions specified in the design plans as well as con-
servative design code approaches. In general, these models consistently overpredicted girder
stresses, deflections, and cross-frame forces. Using the measured data as a benchmark, the
research team was able to adjust several of the key modeling assumptions. Ultimately, it was
shown that three modeling parameters, when adjusted, had the most impact in achieving
good agreement with the measured results. These three parameters, which are briefly intro-
duced herein, included boundary conditions, contribution of concrete barriers, and the elastic
modulus of the concrete deck. For a more detailed review of these modeling parameters, refer
to Appendix E.
In short, the boundary conditions were modified from pure pinned and roller conditions
to a more accurate linear spring representation of the bridge bearings. The stiffness of the
elastomeric bridge bearings was estimated based on AASHTO LRFD Equation 14.6.3.1-2
and the dimensions and parameters specified on the design plans. The stiffness of the bridge
bearings was most influential for instrumented Bridge 3 (the horizontally curved system), which
is intuitive given the torsional response of the horizontally curved superstructure to vertically
applied loads.
Conventional analysis and design practices generally neglect the contributions of a bridge
barrier, continuous or discontinuous. The preliminary analyses, much like the conventional
approach, also neglected the barriers in the 3D model. After several model iterations, it was
evident that including the discontinuous bridge barriers markedly improved the cross-frame
force predictions in the FEA models, particularly for Bridge 3, which is a narrow system. By
providing another load path for truck loads, barriers reduce the force demands in the cross-
frame elements. It is acknowledged, however, that including the barriers in analysis models
increases the computational efforts by both designers and programmers. Thus, the conventional
approach of not modeling barriers typically results in a conservative estimate of cross-frame
force effects.
Lastly, the assumed modulus of elasticity of the concrete deck and rails was also shown
to significantly affect girder and cross-frame response in 3D FEA models. In the preliminary
analyses, the research team based the assumed elastic modulus on the minimum concrete
strength specified by the design plans and the traditional American Concrete Institute (ACI)
equation (i.e., fc′ where f ′c is the concrete strength in psi). It was recognized, however, that
design codes inherently provide lower-bound estimates. As such, the research team explored
different publications and experimental data to justify a higher elastic modulus for concrete
(i.e., a value more representative of mean conditions as opposed to lower-bound conditions)
(Tadros et al. 2003), which in turn improved comparisons between the analytical solutions
produced by the models and the experimentally measured data. In design practice, making
these assumptions would be challenging given the uncertainty in the material properties. Thus,
similar to the bridge barrier discussion above, the conventional approach of using code-based
material properties typically results in a conservative estimate of cross-frame force demands.
It is also important to note that in general terms, a user can manipulate a model in many
ways to achieve the target solution; however, those changes may not be a good representation
of the actual structural system. The team was interested in achieving good agreement between
measurements and FEA predictions but not at the expense of using unreasonable assumptions.
Based on a literature review and examining many different model configurations, the research
team was able to select a consistent set of parameters and assumptions that not only improved
the accuracy of the results compared to the measured data, but also made sense based on reason-
able engineering assumptions.
Based on the commentary above, the modeling approach was fine-tuned to improve the
agreement between the measured and finite element results. To demonstrate these improve-
ments for instrumented Bridge 1, Table 3-1, Table 3-2, and Table 3-3 compare the preliminary
and validated analytical results with the measured data for critical girder deflections, girder
stress, and cross-frame force effects, respectively. Only the load case illustrated in Figure 3-2
is presented in these tables; however, the results are representative of all load cases performed
at Bridge 1. For reference, the percent error (relative to the measured data) associated with the
model is also presented.
From these tables, it is apparent that increasing the stiffness of the concrete deck and
including a discontinuous concrete rail stiffened the bridge overall and consistently improved
the analytical results. Critical girder deflections, which were once uniformly overestimated,
improved to errors typically within 10%. Critical girder stresses were also consistently
improved to within 10% to 20% of the measured data. Given the complexity of the bridge
model and the potential uncertainty in the field measurements, these validated discrepancies
are deemed acceptable.
In general, Table 3-3 also demonstrates the improved accuracy of the model with respect
to cross-frame forces. It is important to note that the error associated with cross-frame forces
is noticeably higher than what is observed for girder stresses or deflections. Even for the most
sophisticated full-shell, 3D FEA model and a relatively simple bridge geometry, the errors
associated with the critical cross-frame forces still ranged from 0 to 60%. Larger discrepancies
were tabulated, but those correspond to less critical cross-frame members.
Load paths and flexural behavior of girders are generally more straightforward than that
of cross-frames. Furthermore, it should be noted that many of the stress magnitudes were
relatively small such that any slight variations can produce very large percent differences
between measured and predicted stresses. In general, these tables highlight the difficulty with
trying to improve all of the measurements (i.e., girder deflections, girder stresses, and cross-frame
stresses). The system is highly indeterminant when the various components are considered—
multiple girders, many cross-frames, and variations in the concrete deck thickness along the
length and width. Given the complexity of the bridge model and the potential uncertainty in
the field measurements, these validated discrepancies were deemed acceptable.
As noted above, the data presented in Table 3-1 through Table 3-3 represent just one static
load considered at Bridge 1. Similar results were synthesized for every live load case at every
instrumented bridge, which is summarized in Appendix E. By achieving good agreement
between the measured and finite element results in this validation study, it ensured that the
parametric studies executed in Phase III produced reliable results that were consistent with real
load-induced behavior of cross-frames.
The primary goal for obtaining these data was to establish effective and maximum stress
range metrics by which the computational studies and current AASHTO fatigue criteria could
be assessed. Effective stress ranges are related to finite-life behavior (Fatigue II limit state), and
maximum stress ranges are related to infinite-life behavior (Fatigue I). As such, this section
presents representative response spectra data and summarizes the measured effective and
maximum stress range values obtained from the field studies. These metrics are subsequently
used for comparison in later sections of the report.
Measured response spectra are best illustrated as histogram plots, for which the number
of stress cycle counts, determined by rainflow-counting algorithms, are compiled and sorted
into different stress magnitude bins, Sr,j (i.e., jth stress range bin). The effective stress range then
mathematically represents the response of the bridge component to the entire truck population
by equating the fatigue damage caused by the variable-amplitude response spectrum to a
constant amplitude spectrum of equal cycle count. The maximum stress range represents the
upper tail of the spectra. A more detailed discussion on the calculation of effective and maxi-
mum stress range of the spectrum, as well as the truncation process and bin size parameters,
is provided in Appendix E.
A sample histogram illustrating the variable-amplitude stress range spectrum for a cross-
frame member (D1-2) in instrumented Bridge 1 is presented in Figure 3-5 (left graph). The right
graph provides the mathematical representation of the effective stress. That is, 2,683 cycles at
0.91 ksi produces fatigue damage equivalent to the spectra of stress magnitudes and cycle counts
shown on the left graph. The maximum stress range is also approximately 2.7 ksi based on the
upper tail of the spectrum on the left. For clarity, the vertical axis is presented using a log scale.
Aside from the calculation of effective and maximum stress range metrics, there are several
additional observations from Figure 3-5:
• This instrumented cross-frame member in Bridge 1 experienced load-induced stress magni-
tudes ranging from 0 to 2.7 ksi over the entire monitoring period. Note that the stress cycles
below 0.65 ksi were truncated based on the discussion provided in Appendix E, which is
consistent with the method outlined in Connor and Fisher (2006).
• The vast majority of the stress cycles are well below the CAFL for a Category E′ detail (i.e.,
2.6 ksi). In fact, 90% of the recorded stress cycles corresponded to magnitude less than 1.1 ksi.
100,000
Stress Range Spectrum, Sr, j
Total Cycles = 2,683
10,000
1,000
Number of Cycles
100 100
10
0
0 1 2 3 0 1 2 3
Stress Range, Sr, j (ksi) Stress Range, Sr, j (ksi)
Similar histograms were produced for girder flanges as well. Figure 3-6 presents a side-
by-side comparison of the stress range spectra measured for a cross-frame member (D1-2 in
Figure 3-2) and girder flange (Girder 2 in Figure 3-2) for instrumented Bridge 1. Both instru-
mented elements were subjected to the same load spectrum (i.e., same truck population and
lane positions); however, the response of each is significantly different. Not only are the stress
range magnitudes higher for the girder flanges, but the number of cycles above the correspond-
ing truncation stress was significantly more. Given that the cross-frames were designed for a
much more stringent fatigue detail (i.e., lower CAFL value), it is intuitive that the stress magni-
tudes in girder flanges exceed those in cross-frames. The variation in stress cycle counts, on
the other hand, can be attributed to the sensitivity of cross-frames to transverse load position.
Depending on the precise transverse position of a passing truck, the stress range response in
a particular cross-frame element could be significant or negligible (i.e., below the established
truncation stress or even hidden by the electromagnetic noise of the strain gage). Bending
stresses in girders, however, are less influenced by lane position, as evidenced by the sample
data in Figure 3-6.
Given that (i) the current AASHTO LRFD fatigue load factors were only calibrated for
girder response to truck traffic and (ii) transverse load distribution effects were not explicitly
considered in that study, this measured data gives an initial indication that load cases specific
to cross-frame fatigue are perhaps warranted. This behavior is explored in greater depth
computationally in Section 3-3.
As noted above, the research team produced and processed a number of histogram plots
similar to those presented in this report. In order to evaluate the measured in-service data with
AASHTO fatigue design criteria in subsequent sections, only the effective and maximum stress
ranges are examined. As such, Table 3-4 summarizes the range of effective and maximum stress
ranges computed for the instrumented cross-frame members in Bridges 1, 2, and 3. In other
words, these stress range metrics were computed for every instrumented cross-frame member,
and the range presented in the table represents the maximum and minimum values of the data
set. A similar table with respect to girder flange data is provided in Appendix E for reference.
From Table 3-4, it is obvious that critical cross-frames in Bridge 2 (skewed) consistently
experienced higher effective stress ranges in comparison to critical cross-frames in the other
100,000 100,000
Cross-Frame Girder Flange
10,000 10,000
1,000 1,000
Number of Cycles
Number of Cycles
100 100
10 10
1 1
0 0
0 1 2 3 4 5 0 1 2 3 4 5
Stress Range, Sr, j (ksi)
1 {0.70–0.98} {1.35–3.61}
2 {0.78–1.19} {0.99–3.77}
3 {0.77–0.95} {0.99–2.84}
two bridges. These results imply that the cross-frames in this skewed system, especially
with a contiguous layout, generally experience slightly higher stress ranges than the normal
and horizontally curved systems. Although not explicitly addressed in the table, it was also
observed in several instances that cross-frame stress ranges were dependent on its location
relative to the right lane (i.e., the typical drive lane for heavy truck traffic). The discussion
related to load-position effects and bridge geometry is expanded in Section 3.2 with respect to
the finite element studies.
fatigue load factors for cross-frame design. Section 3.2.3 reviews the WIM data and explores
the frequency at which a “double truck” case (i.e., a loading condition that maximizes the force
reversal in a cross-frame member) occurs. Section 3.2.4 provides a summary of the key findings.
Figure 3-7. Unfactored design force ranges compiled for every cross-frame
member evaluated.
the compiled design force ranges for all top strut members evaluated near the maximum
positive dead load moment region (edge bay) of straight and normal bridges. In total,
312 data points (corresponding to the unfactored design force ranges for 312 unique, straight
and normal bridges) are represented.
• For each box-and-whiskers plot, six key statistical parameters are displayed: (i) the minimum
value in the data set, (ii) the first quartile or the 25th percentile, (iii) the second quartile or
the median value, (iv) the third quartile or the 75th percentile, (v) the maximum value, and
(vi) the mean value. The maximum and minimum values are represented by the whiskers, the
quartile values are represented by the bounds of the box, and the “X” represents the mean.
A graphical representation of this procedure is presented in Appendix F for reference.
• The critical lane passage that corresponds to the design force ranges presented are different
for each point in the data set. In this figure, only the magnitudes of the design force ranges are
of interest. The critical lane positions are covered in a subsequent section.
• A reference sketch is also provided in the figure to clarify the relative position of the cross-
frame considered; the results for the color-coded cross-frame locations are plotted. Note
that the exact location of the cross-frame, the number of girders, the number of spans, the
support skew angle, and the radius of curvature in the sketch are for visual reference only;
the bridge geometries of the various data points may differ from the representative sketch.
By presenting the box-and-whiskers side-by-side, the cross-frame members that generally
experience the largest force ranges due to the passage of the AASHTO fatigue truck in its critical
lane can be evaluated. By examining the results independently for bridge types, the impacts of
support skewness and bridge curvature can also be assessed.
From Figure 3-7, the following observations can be made about governing cross-frame
members:
• There is substantial variability observed in all the box-and-whiskers plots presented. For
example, the design forces for bottom struts near end supports of horizontally curved and
skewed bridges (curved, skewed) range from a minimum of about 0.5 kips to a maximum of
nearly 18 kips. This scatter is attributed to the large number of bridge parameters inherently
considered in the results.
• Top strut members, as anticipated, generally have the lowest force demands observed since
the concrete slab tends to restrain out-of-plane deformations at the top of the girder.
• The force ranges observed in skewed and/or horizontally curved bridges are generally greater
than equivalent straight and normal bridges, particularly in bottom strut and diagonal
members near skewed supports. For example, the mean force range for bottom struts at end
supports (BS; ES) increases from 3 kips for straight and normal bridges to over 4 kips for
straight and skewed bridges.
• This same relationship is not as impactful for cross-frames not in the vicinity of supports
(i.e., near the maximum positive dead load moment region), as cross-frames in these regions
are less impacted by the effects of support skew.
• In general, bottom strut members near M-I regions (interior bay near maximum positive
dead load moment region) have the largest design force ranges for non-skewed (straight or
horizontally curved) bridges. The maximum, 75th percentile, and mean values are all largest
for this condition.
• For skewed bridges, the bottom strut members near end supports tend to have the largest
force demands, but there is still significant scatter.
From all the data obtained from this study, it is apparent that bottom strut members tend to
be the most critical in terms of fatigue force demands. In past studies that focused on stability
bracing applications of noncomposite girders, it has been shown that the top and bottom struts
generally behave as zero-force members with both diagonal members equally effective (i.e.,
adjacent girders rotating equally). However, for cross-frames in composite systems and sub-
jected to live loads, bottom struts are often heavily engaged. In many cases, the cross-frames
behave similarly to floorbeams in a stringer-floorbeam system to distribute loads from girder-
to-girder, where adjacent girders rotate differentially in this case. The bottom strut is analo-
gous to the bottom tension flange of a composite floorbeam, the concrete deck is the top
compression flange, and the top strut is in close proximity to the pseudo-neutral axis.
At skewed supports, the global, torsional response of the superstructure tends to engage the
nearby bottom struts. Contiguous lines of cross-frame panels act like a stiff, closed section that
resists the torsional moments on the bridge cross-section.
Even with these discernible trends, there is still significant variability in cross-frame response.
The governing cross-frame member is still a function of many bridge parameters, albeit support
skew and curvature are the most important. Given a random bridge geometry, the critical
cross-frame panel could likely be identified within reasonable limits before any analysis is
performed. Still, it is recommended that the critical cross-frame is not “missed” by taking a
shortcut. Rather, conducting an influence-surface analysis and performing a comprehensive
design of all intermediate cross-frames in the bridge ensures a fully vetted design.
Many commercial design software packages have the built-in functionality to analyze and
design all cross-frames for design loads. If the program does not automatically address this,
manually developing a spreadsheet to evaluate each cross-frame due to AASHTO fatigue loading
criteria is possible.
Reference Sketch:
ES M-I
IS Key:
Non-Overhang Load
M-E Overhang Load
1.0
312 Models 1,440 Models 624 Models 1,728 Models
0.9
0.8
Frequency of Occurrence
M-I
0.7
ES
0.6
M-I
M-E
0.5
IS
0.4
ES
M-E
0.3
M-I
M-E
0.2
ES
M-E
IS
M-I
0.1
IS
ES
IS
0.0
Straight, Normal Straight, Skewed Curved, Normal Curved, Skewed
Bridge Type
Figure 3-8. Bar graph presenting the critical lane position associated
with the governing cross-frame member.
To clarify the intent of the figure, the straight, normal data set is considered as an example.
For all 312 straight and normal bridges evaluated, the governing cross-frame was located in an
interior bay near the maximum positive dead load moment region on 204 occasions (65%). Of
those 204 occasions, the critical cross-frame member was always governed by a non-overhang
load (i.e., 100% non-overhang and 0% overhang). This is reflected in the “M-I” bar in the
straight, normal portion of Figure 3-8.
A more interesting example is the intermediate support “IS” case of the straight and skewed
data set. In total, 1,440 straight and skewed bridges were analyzed in the Fatigue Loading
Study. Of those 1,440 bridges, 565 were governed by cross-frame members near the intermediate
support (39%). Of those 565 cases, 514 were governed by overhang truck passages and only
51 by non-overhang loads.
From Figure 3-8, the following observations can be made about bridge type and critical
lane passages:
• Non-overhang loads, although not overly descriptive, generally correspond to “localized”
load effects. In other words, the force effects in the critical cross-frame were maximized by
a truck passing just to the left or right of the panel, similar to the trends observed in the
influence-surface plots in Section 2.3.
• For non-skewed bridges, the critical cross-frame members are almost always governed by
non-overhang loads. In fact, the forces in critical cross-frame members in straight, normal
bridges were maximized 100% of the time by non-overhang loads. The exception to this rule
is M-I cross-frames in horizontally curved bridges (interior bay near the maximum positive
dead load moment region), which were often governed by truck passages along the outer
radius of the curve. Thus, it can be concluded that cross-frame fatigue design in bridges
without support skews is largely governed by truck passages in close transverse proximity to
the cross-frame panel.
• For skewed bridges, the overhang loads represent a large percentage of the critical truck
passages, particularly for “ES” and “IS” cross-frames. The torsional response of the super-
structure and the corresponding girder rotations often heavily engage the bottoms struts
and diagonals in these areas. As such, loads that maximize the induced torque on the bridge
cross-section are often critical. For straight bridges, the critical loads correspond to those
applied along either deck edge where the moment arm about the bridge centroid is largest.
For horizontally curved bridges, those loads correspond to either the inner or outer radius
about the straight-line chord of the curved segment.
It is evident that force demands in cross-frames of skewed bridges, especially those near the
end or intermediate supports, are sensitive to overhang loads. Because these cases are rather
infrequent, basing a fatigue limit state design on an overhang truck passage could potentially
result in overly conservative design loads. With that in mind, overhang and non-overhang truck
passages are evaluated in the context of measured WIM data in Section 3.2.2, where the appro-
priate lane position for AASHTO implementation is examined in greater depth.
D2
8
750-ft Radius Radius of Force Range
1500-ft Radius Curvature (ft) (kips)
6
750 7.00
Axial Force (kips)
-2
0 20 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320 340 360
Longitudinal Position of Truck (ft)
straight bridge, the resulting force range from the passing truck increases 10% when a 1,500-ft
radius is introduced and an additional 9% when the radius is halved to 750 feet. The general
response of the cross-frame is the same, except for the magnitudes. Thus, based on this limited
data set, it could be stated that bridge curvature is positively correlated to cross-frame force
demands (i.e., bridge radius is negatively correlated).
Figure 3-9 represents only a select number of unique conditions. The response could also
have been presented for different lane positions, different cross-frame members, and differ-
ent bridges altogether. Ultimately, a primary interest of most designers is knowledge of how a
bridge parameter impacts the overall fatigue design of cross-frames, not the localized effects on
non-governing members that are deemed less important. Therefore, the results presented here-
after are focused on the governing cross-frame members and lane positions (i.e., the unfactored
design forces in accordance with current AASHTO fatigue loading criteria).
To expand on the sample results presented in Figure 3-9, critical cross-frame force effects
and lane passages were evaluated for every bridge in the analytical testing matrix. That is to say,
box-and-whiskers plots that represent the spectrum of design forces were generated and sub
sequently categorized by bridge parameter (e.g., all cross-frame force effects in straight bridges,
bridges with a 1,500-foot radius of curvature, and bridges with a 750-foot radius of curvature
were evaluated and compared).
Figure 3-10 presents the summarized results from this exercise highlighting several key
parameters, including skew index (defined below), horizontal curvature, cross-frame layout,
number of girders, and concrete deck thickness. These parameters were selected, as they signi
ficantly influence the load-induced force effects in the critical cross-frames. In these figures,
box-and-whiskers plots and response spectra are represented as a series of three distinct line
graphs. In the figure, the solid black line represents the mean of each data set, and the dashed
lines represent the 95th and 5th percentile values. The respective percentile lines give an
indication where 90% of the data set is populated. Despite only presenting five parameters
in Figure 3-10, every independent variable that was previously identified in Table 2-2 was
evaluated similarly. For additional information on the development of these plots and the
results for all bridge parameters (including spot checks of bridge models without barriers),
refer to Appendix F.
Before introducing the results, it is prudent to define skew index, as well as connectivity
index. Skew (Is) and connectivity (Ic) indexes are respective measures of bridge skewness and
curvature, as defined in the AASHTO G13.1 Guidelines for Steel Bridge Analysis (2019). These
indexes are used to categorize bridge geometries for purposes of recommended analysis practices.
Bridges with larger skew and connectivity indexes require more advanced analysis procedures,
either improved 2D or 3D techniques. Although not presented in Figure 3-10, the connectivity
index is utilized in subsequent sections of the report. The skew and connectivity indexes are
defined by the following expressions:
wg tanq
Is = 3.1
Ls
15,000
Ic = 3.2
R (ncf + 1)m
where
Is, Ic = skew index and connectivity index, respectively,
wg = width of bridge, measured between the centerlines of the fascia girders,
18 18
Skew Index Curvature
15 15
12 12
9 9
6 6
3 3
0 0
0 0.35 0.7 0 5 10 15
Skew Index Curvature (10–4/ft)
12 12
9 9
6 6
3 3
0 0
3 5 7
Cross-Frame Layout No. of Girders
18 Deck Thickness
15
12
0
8 10
Deck Thickness (in)
q = largest skew angle on bridge relative to the axis normal to the longitudinal girders,
Ls = span length at the bridge centerline,
R = minimum radius of the horizontal curvature,
ncf = number of intermediate cross-frames in the span, and
m = constant, taken as 1 for simple-span bridges and 2 for continuous-span bridges.
As the equations show, the skew index increases for shorter, wider spans with larger skew
angles. Similarly, the connectivity index increases for tighter curves (i.e., smaller radius of
curvature) and fewer cross-frames connecting the girders together.
From Figure 3-10, the following observations can be made about cross-frame force demands
and the select bridge parameters outlined above:
• Skew index: It is evident from the results that skew index substantially impacts the response of
cross-frames. A +0.52 correlation coefficient was obtained, indicating a strong, positive
relationship between skew index and cross-frame force effects. In simpler terms, a +0.1 increase
in skew index generally resulted in a 25% increase in the critical cross-frame force demand.
In general, these observations are consistent with the field experimental data outlined in
Section 3.1.
• Curvature: A positive correlation is observed between bridge curvature and cross-frame force
effects (i.e., correlation coefficient of +0.32). On average, an increase in curvature of 10-4/ft
results in an increase of 3% for the governing cross-frame force effect. Thus, the trend observed
in Figure 3-9 for a small sample set is consistent with the full data set.
• Cross-frame layout: Staggering the cross-frame layout in skewed bridges significantly reduces
the governing design force in cross-frames (35% on average). This trend is attributed to the
stiffness reduction associated with a discontinuous line of cross-frames. These results are
consistent with the guidance provided in AASHTO LRFD Article C6.7.4.2.
• Number of girders: A positive correlation coefficient of +0.60 was observed between number
of girders (i.e., bridge width) and cross-frame force demands. A more variable response
was also observed as the number of girders increased, as evidenced by the distance between
the 95th and 5th percentile lines. On average, increasing the overall width of the bridge
results in a 30% increase in the governing force effect. These trends are attributed to the
overhang truck loads outlined in Section 3.2.1.2. Eccentric loading and torque on a straight
or curved system are amplified for wider bridges, which in turn amplifies the governing
cross-frame response.
• Deck thickness: Deck thickness and cross-frame force effects are negatively correlated
(–0.22 correlation coefficient). On average, a 1-inch increase in deck thickness results in a
14% decrease in force demands. A reduction in deck thickness often indicates that the
cross-frames will attract a higher proportion of the load distribution, which increases member
force effects. Additionally, a thinner deck corresponds to a reduction in overall superstructure
stiffness, which subsequently increases girder displacements and cross-frame deformations.
In general terms, it is evident that skewed and/or curved composite bridge systems typically
result in larger load-induced force demands in their cross-frames. Although this behavior
has been widely recognized by designers throughout the last few decades, the results above
validate these assertions quantitatively. To mitigate these force effects and potential load-
induced fatigue cracking problems in new designs, designers can utilize these trends to avoid
an iterative “chase-your-tail” design solution (i.e., increase the size of the cross-frame member to
accommodate the design forces, re-analyze, and redesign for larger forces in the second pass, etc.).
For instance, using a discontinuous, staggered cross-frame layout is a practical and economical
solution for skewed systems. Increasing the deck thickness or decreasing girder spacing can also
lessen load-induced cross-frame forces in composite bridges, but adjusting those parameters
has significant impacts on the rest of the design.
200
Bridge 1 Key:
Bridge 3 Fatigue II “Analytical” Design Stress
160
Number of Occurrence
Bridge 2 Bridge 2 Measured Effective Stress
120 Bridge 1
Bridge 3
80
40
0
0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0
Governing Fatigue II Design Stress, γΔf (ksi)
analytical and experimental results of instrumented Bridges 1, 2, and 3, a few key observations
are established.
First, the governing design stress ranges for Bridges 1 and 2 are close to the mean response
(2.35 ksi), and the stress range for Bridge 3 is in the lower tail of the spectrum. Second and
more importantly, it is evident that the analytical results generally exceed the measured results
by a considerable margin. This is a preliminary indication that the AASHTO fatigue model for
the Fatigue II limit state may be conservative when compared to real loading conditions expe-
rienced by the instrumented bridges in the context of cross-frame behavior. This assertion is
verified for a wider range of bridges herein; however, note that the physical evidence from
measured data, rather than just relying entirely on analytical data, strengthens this observation.
Figure 3-12 then illustrates the governing demand-to-capacity ratio (i.e., DFn/γDf) with
respect to the Fatigue II limit state. The factored stress ranges from Figure 3-11 are compared
to the factored resistances computed based on the procedure outlined in Section 2.3.2. Demand-
to-capacity (D/C) ratios below unity indicate designs in conformance with AASHTO LRFD,
0.1
Llano, TX, Traffic
0.08 Mean = 0.54
% Above 1.0 = 7.4%
0.06
0.04
Probability of Occurrence
0.02
0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
0.1
Houston, TX, Traffic
0.08 Mean = 1.40
% Above 1.0 = 65%
0.06
0.04
0.02
0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
Demand-to-Capacity Ratio
whereas ratios exceeding unity represent designs in violation of AASHTO LRFD. As outlined
previously, it is assumed that the design of all intermediate cross-frame members is governed
by the critical, maximum case. Thus, any bridge exceeding a D/C of 1.0 means that one
(i.e., the critical cross-frame member) or more members are inadequate in terms of Fatigue II
design criteria.
Figure 3-12 presents D/C histograms for the two extreme traffic conditions: rural Ranch
Road 152 in Llano, Texas, and urban I-10 in Houston, Texas. As outlined in Section 2.3.2, in these
examples, the Llano and Houston bridges are just hypothetical locations of a rural bridge with
relatively light traffic versus a busy urban environment with relatively high traffic. Based
upon TxDOT data (2020), the single-lane ADTTSL for the Llano location can be estimated as
182 compared to 3,000 for the Houston location. For each individual histogram, the factored
demands are identical; however, the factored capacities differ based on the large disparities
in projected truck traffic volumes.
From Figure 3-12, it is apparent that the projected traffic conditions have a significant
impact on the results. For the low-volume conditions, the results indicate that the majority
of the 4,104 bridges are conservative in terms of cross-frame fatigue design. In fact, the average
D/C ratio was 0.54, which implies substantial reserve capacity for the majority of bridges.
Only 7.4% of the models exceeded a D/C ratio of unity.
In contrast, the results indicate a more severe trend for the high-volume conditions under
the Fatigue II limit state (i.e., the 75-year ADTTSL does not exceed 8,485 trucks per day for a
Category E′ detail per AASHTO LRFD Table 6.6.1.2.3-2). On average, the D/C ratios exceeded
1.0 (mean of 1.40), and 65% of the bridges resulted in a design in violation of AASHTO
LRFD. If these bridges were to be implemented for a real construction project, modifications
to the cross-frame properties (e.g., an increase in cross-sectional area) or layout (e.g., use a dis-
continuous cross-frame layout in heavily skewed bridges) would likely be required to bring the
design in conformance with AASHTO LRFD. To expand further on Figure 3-12, Figure 3-13
highlights the differences between straight and curved bridges, as well as bridges with normal
and skewed supports. The full histogram related to the Houston, Texas, traffic in Figure 3-12 is
broken down into four bridge types: straight bridge with normal supports, straight bridge with
skewed supports, curved bridge with normal supports, and curved bridge with skewed supports.
Each subset histogram (shown in black) is overlaid on the full histogram (shown in light blue)
to demonstrate how bridge curvature and support skewness affect the D/C ratios.
The average D/C ratio increases when skewed supports and/or horizontal curvature are
introduced. For instance, the mean D/C for straight, normal bridges is 0.88; that value increases
to 1.23 for straight, skewed bridges and 1.15 for curved, normal bridges. A similar trend is
observed for the percent unconservative metric.
In broader terms, the results from Figure 3-12 and Figure 3-13 suggest the following:
• For roadways with low truck traffic, load-induced fatigue does not appear to be a major
concern for the 4,104-model bridge sample (which represents the most common bridge
conditions in the United States) according to AASHTO 9th Edition criteria.
• For bridges serving heavy truck traffic, current AASHTO criteria indicate a potential load-
induced fatigue problem.
Development of appropriate design criteria for cross-frames, like other structural elements,
requires achieving acceptable structural safety but must also consider economy. As learned from
the industry survey discussed in Section 2.1 and Appendix D, actual load-induced fatigue prob-
lems in cross-frame members have seldom been documented by bridge owners.
Thus, experience has shown that cross-frames are largely satisfying structural safety
requirements in terms of load-induced fatigue. A question of interest then is whether further
0.04
Straight, Normal Straight, Skewed
Mean = 0.88 Mean = 1.23
0.03 % Above 1.0 = 35% % Above 1.0 = 54%
0.02
0.01
Probability of Occurrence
0
0.04
Curved, Normal Curved, Skewed
Mean = 1.15 Mean = 1.73
0.03 % Above 1.0 = 57% % Above 1.0 = 83%
0.02
0.01
0
0 1 2 3 4 5 0 1 2 3 4 5
Demand-to-Capacity Ratio
economies are possible in cross-frame design, while still providing adequate structural safety.
The absence of observed load-induced fatigue failures does not necessarily imply the current
AASHTO LRFD design criteria are overly conservative and wasteful from a cost perspective.
Nonetheless, with a goal of developing improved design criteria for cross-frames, the question
of whether current AASHTO LRFD design criteria for load-induced fatigue of cross-frames
are too conservative is important to examine. The analysis presented earlier in this section
provides at least some indication that AASHTO LRFD may be too conservative and merits
closer scrutiny.
When considering whether AASHTO LRFD may be too conservative for load-induced
fatigue design of cross-frames, the following three factors may be potentially considered:
1. The AASHTO fatigue resistance model is potentially too conservative (DFn is too low). The
resistance model primarily consists of detail Category E′ and its associated constants
(Table 6.6.1.2.3-1).
2. 3D FEA models perhaps consistently produce overly conservative force predictions. Among
many modeling assumptions, the primary focus is the stiffness modification factors
(R-factors) for cross-frames in 3D models (Article C4.6.3.3.4).
3. The AASHTO fatigue loading model is potentially too conservative (γDf is too high). The load-
ing model comprehensively consists of the following: fatigue truck (Article 3.6.1.4.1), truck
positioning (Article 3.6.1.4.3a; Article C6.6.1.2.1), dynamic load allowance (Article 3.6.2),
load factors (Article 3.4.1), and shear lag factor [Table 6.6.1.2.3-1 (Condition 7.2)].
Note that, by simply listing the above factors, it is not advocated that changes to the design
specifications are necessary related to all these factors. The list is intended to be comprehensive
in nature and identify all possible sources. For instance, this research does not suggest that
welded cross-frame connections be reclassified as a higher fatigue category.
As noted previously, fatigue resistance is beyond the scope of the project and is not discussed
further. Instead, the research team narrowed its efforts to examining the modeling approach
(stiffness modification and simplified analysis techniques) and the fatigue loading model herein.
Section 3.3 studies the applicability of the current R-factor approach (0.65AE per AASHTO
Article C4.6.3.3.4) for composite systems. Section 3.4 explores the limitations of simplified 2D
modeling approaches. Sections 3.2.2 through 3.2.3 examine the loading model (primarily the
fatigue truck, truck positioning, and load factors) with respect to high-resolution WIM data.
1.0
0.9
0.8
Cumulative Design Function
0.7
0.6 0.99995
0.5 0.99990
0.4
0.99985
0.3
0.2 0.99980
0.1
0
0 1 2 3 4 5 6 7
Stress Range (ksi)
distribution of all stress ranges produced by the 18 WIM site records). Load bias refers to the
ratio between the actual load effects and the design load effects (i.e., the WIM load effects
divided by the load effects caused by the fatigue design truck).
As documented in the SHRP 2 R19B report (Modjeski and Masters 2015), the value
associated with the “mean plus 1.5 standard deviations” was taken as both the bias of the data
(SHRP 2 R19B states this was a conservative measure, since it was unknown how accurately the
WIM records used in the project reflected truck traffic across the nation) as well as an appro
priate load factor. This resulted in a proposed load factor of 2.0 for Fatigue I; however, the load
factor adopted in the 8th Edition of AASHTO LRFD was 1.75. Based on discussions with indi-
viduals knowledgeable on the selection of this load factor, it appears that the value of 1.75 was
ultimately selected by using the arithmetic mean of the normalized stress ranges produced
by the WIM site records (i.e., the bias and load factor are both taken to be equal to the mean
bias value). This was rationalized by acknowledging perceived conservatism inherent in the
resistance data for Fatigue I and the fact that the WIM data reviewed was deemed to be
sufficiently abundant to represent national traffic loads.
Figure 3-15 summarizes the ratios of the 99.99th percentile WIM cross-frame stress ranges
for each of the 20 bridges divided by the unfactored fatigue truck stress range for each bridge.
Each bin value in this figure represents the average of all Fatigue I stress ranges for all WIM sites
applied to the identified bridge (as explained in the preceding paragraph). The solid line in this
figure represents the arithmetic mean of all Fatigue I stress ranges for all WIM sites applied to
the identified bridge. The dashed line in this figure represents the mean value of all “mean plus
1.5 standard deviations” (i.e., the mean plus 1.5 standard deviation for all values represented by
the individual bars).
Consistent with the approach that led to the 8th Edition AASHTO LRFD updates to the
Fatigue I load factor, the arithmetic mean value of all Fatigue I stress ranges for all WIM sites
would correspond to an appropriate Fatigue I load factor for cross-frames. In other words,
the stress ranges produced by the unfactored fatigue truck must be amplified by this factor
to produce an equivalent 99.99th percentile stress range that represents the 99.99th percentile
load effects caused by the WIM traffic streams. Without consideration of material fatigue resis-
tances (and therefore the reliability index), the mean values associated with the solid lines in
Figure 3-15 imply an appropriate load factor for the Fatigue I limit state is 1.01. The coefficient
1.75 Key:
Mean
1.50 Mean + 1.5 Standard Deviations
Maximum Normalized SR99.99%
1.25
(WIM/DesignTruck)
1.00
0.75
0.50
0.25
of variation associated with this value is 0.23. This load factor is less than the current Fatigue I
load factor of 1.75, which indicates a potential source of conservatism in the design load criteria
for cross-frame fatigue.
Recall from Section 2.3.3.1 that the GVWs for the WIM records are not normal; rather, many
WIM sites appear to have multi-modal distributions that likely correspond to natural groupings
of different vehicle types and payload. Appendix F discusses examining the distribution on a
normal probability plot in order to assess the normality of the data and applying a best fit line
to the CDF to estimate the mean and standard deviations. This procedure is performed for all
distribution summaries to compare directly to the means and standard deviations computed
using standard arithmetic formulas. Table 3-5 compares the standard deviations and means
obtained by both methods.
0.8 Key:
Mean
0.7
0.6
(WIM/DesignTruck)
0.5
0.4
0.3
0.2
0.1
factor for Fatigue II. This suggests that the 0.8 load factor may be conservative for characterizing
the response of cross-frames to the U.S. truck spectra.
Similar to the comparison made in Section 3.2.2.1, the fatigue damage ratios for all WIM records
and all bridges were plotted on a normal probability plot to assess the normality of the damage
ratios (further discussed in Appendix F). Table 3-6 compares the standard deviations and
means obtained by standard arithmetic formulas and the normal probability plot approach.
Coefficient of
Limit State Bias for Load Data Load Factor
Variation of Load
Fatigue I 0.26 1.01 1.01
Fatigue II 0.21 0.40 0.53
In order to obtain reliability indices close to the assumed target reliability of unity, adjust-
ments are required to either the load or resistance factors. As an alternative to altering the
resistance factors [or the associated changes in constant amplitude fatigue thresholds, (F)TH,
for Fatigue I, or detail constants, A, for Fatigue II] it is also possible to adjust the load factor
for fatigue such that the minimum target reliability of unity is achieved for each category.
Additionally, as an alternative to introducing two new load factors for the fatigue limit state
for cross-frames, it is possible to apply a single adjustment factor to the existing load factors,
similar to the procedure used for fatigue design of orthotropic decks established in AASHTO
LRFD (2020) Article 3.4.4.
Using an adjustment factor of 0.65 applied to the Fatigue I load factor of 1.75 (i.e., a resul-
tant load factor for cross-frames of 1.14 for Fatigue I) and the Fatigue II load factor of 0.8 (i.e.,
a resultant load factor for cross-frames of 0.52), the resulting reliability indices are calculated
via Monte Carlo simulation and shown in Table 3-9. With this adjustment, each detail category
satisfies the minimum assumed target reliability of unity.
Table 3-8. Reliability of cross-frames for Fatigue I and II limit states using
the load factors introduced in Sections 3.2.2.1 and 3.2.2.2.
Reliability Index
Detail Category
Fatigue I Fatigue II
A 1.01 1.77
B 0.88 1.74
B′ 0.98 2.47
C 0.93 1.82
C′ 0.93 1.82
D 1.27 2.89
E 0.60 1.95
E′ 1.41 2.31
Table 3-9. Reliability of cross-frames for Fatigue I and II limit states using
adjustment factor of 0.65 to existing load factors.
Reliability Index
Detail Category
Fatigue I Fatigue II
A 1.32 1.74
B 1.22 1.70
B′ 1.49 2.39
C 1.29 1.79
C′ 1.30 1.77
D 1.82 2.81
E 1.08 1.87
E′ 1.77 2.25
0.08
Straight, Normal Slightly Curved
IS = 0 IS = 0
IC = 0 IC ≤ 1
0.06
Mean = 0.55 Mean = 0.62
% Above 1.0 = 5% % Above 1.0 = 9%
0.04
0.02
0.08
Heavily Curved Slightly Skewed
IS = 0 IS ≤ 0.3
Probability of Occurrence
IC > 1 IC = 0
0.06
Mean = 0.78 Mean = 0.66
% Above 1.0 = 19% % Above 1.0 = 16%
0.04
0.02
0.08
Heavily Skewed Skewed & Curved
IS > 0.3 IS > 0.1
IC = 0 IC > 0.5
0.06
Mean = 1.02 Mean = 1.19
% Above 1.0 = 44% % Above 1.0 = 56%
0.04
0.02
0
0 1 2 3 4 5 0 1 2 3 4 5
Demand-to-Capacity Ratio
First, the Fatigue II load factors inherently built into the force demands (i.e., the numerator
of the D/C ratio) are reduced from 0.80 to 0.52 based on the findings outlined above. Note that
the resistance model (i.e., the denominator in the D/C ratio) remains unchanged. The second
major exception is in how skewed and curved bridges are organized and presented. Rather
than simply grouping the bridges based on the presence of support skew and/or horizontal
curvature, the indexes developed in NCHRP Research Report 725 (White et al. 2012) and
previously defined in Eqs. 3.1 and 3.2 are utilized.
Thus, each bridge in the 4,104-model data set is grouped based on the skew and connectivity
index bounds established in NCHRP Research Report 725. For instance, heavily curved bridges
are differentiated from moderately curved bridges by Ic > 1 and Ic ≤ 1, respectively. By orga-
nizing the figure in this manner, the criticality of a particular bridge type in terms of load-
induced fatigue forces in cross-frames can be clearly delineated. It is also important to note
that Figure 3-17 only investigates fatigue performance in the 4,104 bridges with respect to the
high-volume truck traffic in Houston, Texas. The results for low-volume Llano traffic would
be much less severe.
With consideration of the proposed cross-frame-specific load factors and I-10 Houston
traffic, the following observations can be made from Figure 3-17:
• In straight bridges with normal supports, the average D/C ratio was 0.55, where only 5% of
the 312 qualifying bridges exceeded unity. This indicates that load-induced fatigue would not
govern the design of cross-frames in these bridge types even for heavy truck traffic volumes.
This implies that conducting a 2D or 3D refined analysis to obtain live load force effects
in cross-frames is likely not warranted. This observation is consistent with the current
AASHTO LRFD design approach.
• For slightly curved bridges (Ic ≤ 1) the mean D/C ratio (0.62) increased slightly compared
to straight bridges, and only 9% of the qualifying data points exceeded unity. This indicates
that design of cross-frames in moderately curved bridges with normal supports is also not likely
to be governed by load-induced fatigue. In contrast, heavily curved bridges (Ic > 1) showed
an increase in mean D/C ratio to 0.78 and an increase in percent exceeding unity to 19%.
• It is evident that support skew affects cross-frame force demands more significantly than
horizontal curvature. Slightly skewed bridges (Is ≤ 0.3) and heavily skewed bridges (Is > 0.3)
produced mean D/C ratios of 0.66 and 1.02, respectively. Additional statistical analysis
showed that obtaining fatigue loads and designing for load-induced fatigue should be
considered for cross-frames in bridges with a skew index exceeding 0.15.
• For bridges with both support skew and horizontal curvature (i.e., Ic > 0.5 and Is > 0.1), the
mean D/C reported was 1.19, which indicates a significant load-induced fatigue issue even
with the reduced load factors. Consequently, these bridges would likely need to be redesigned
or the cross-frames reconfigured to mitigate the fatigue force demands. A refined 2D and
3D analysis would be necessary to accurately obtain those design forces.
400
100%
350
Frequency of Occurrence
300 80%
Key:
250 Frequency
Cumulative % 60%
200
150 40%
100
20%
50
0 0%
-
-
-
-
-
-
-
-
-
Clear Distance Between Trucks (ft)
-
As discussed in Section 2.3.3.5, a cluster analysis was performed on the multi-lane WIM data
to consider if a bridge may be loaded with other truck traffic during a primary drive lane load
event (i.e., passage of one or more axles of a vehicle). Typical results of the adjacent lane
scenarios are summarized in Figure 3-18 through Figure 3-20, which provide a sample histo
gram illustrating various aspects of multiple presence studied. For various load-position
parameters (e.g., clear distance between drive lane and passing truck), the number of occur-
rences for each multi-lane WIM site is compiled and plotted. Specifically, Figure 3-18 illustrates
the clear distances between passing lane trucks and drive lane trucks, Figure 3-19 illustrates
the spectrum of GVW for passing lane trucks, and Figure 3-20 illustrates the ratio of passing
lane truck GVW to drive lane truck GVW. Appendix F provides additional histograms from
these studies. Note that the occurrences on the primary vertical axis for these histograms are
the number of total occurrences for the 12-month data set. This value should be taken into
context with the total annual truck traffic (ATT) for the specific site. Table 3-10 summarizes
100%
2000
Frequency of Occurrence
Key: 80%
1500 Frequency
Cumulative % 60%
1000
40%
500
20%
0 0%
1600 120%
Indiana US 31 SB
1400
100%
1200
Frequency of Occurrence
Key: 80%
1000
Frequency
Cumulative % 60%
800
600
40%
400
20%
200
0 0%
how often (for the year of data considered) any truck was within a certain window of the
drive lane truck, relative to the total volume of traffic in the drive lane for each specific site.
Based on these results, it is evident that a passing truck in close proximity to a drive lane
truck is a rare occurrence. The frequency of occurrence is less than the assumptions used in the
original calibration studies (Nowak 1999). The largest frequency of occurrence is demonstrated
by the Tennessee IH-40 westbound (WB) data, where approximately 30% of traffic is accom-
panied by another vehicle located with a headway distance of less than 1,000 feet, and 1.8% of
traffic is accompanied by another vehicle located with a headway distance less than 20 feet. For
the most critical passing lane position for cross-frames (one directly behind another or zero
clear distance), the largest frequency of occurrence is demonstrated to be 0.03%. Recalling that
cross-frame forces generally reduce rapidly as a truck moves away from the cross-frame in the
Table 3-10. Summary of multiple presence statistics for adjacent lane loaded.
Frequency of Occurrence within Clear Distance Window, Relative to Drive Lane Annual
Truck Traffic (ATT) a
WIM Site
+/- 1,000 ft +/- 280 ft +/- 50 ft +/- 20 ft +/- 0 ft
Indiana
4.0% 2.1% 0.5% 0.3% 0.01%
US-31 NB
Indiana
4.2% 2.2% 0.5% 0.3% 0.00%
US-31 SB
Tennessee
16.0% 7.4% 1.7% 1.0% 0.02%
IH-40 EB
Tennessee
29.9% 13.2% 3.0% 1.8% 0.03%
IH-40 WB
Virginia
3.9% 2.0% 0.5% 0.3% 0.01%
US-29
a
Reference Figure 2-11 for illustration of positive and negative clear distances.
longitudinal direction, it is apparent that, for cross-frames, the largest frequency of occurrence
for passing vehicles occurring simultaneously is appreciably low (i.e., larger headway distances
generally do not result in superimposed cross-frame forces). Another observation is that the
distribution of the passing lane truck GVW appears to be bi-modal for several sites, with
the heavier truck mode being equal to or heavier than the drive-lane truck GVW. Sensitivity
studies were conducted by White (2020) that explicitly accounted for the cross-frame force
effects caused by passing lane vehicles, and the effects on the accumulated fatigue damage were
found to be negligible.
The WIM study performed in this research confirmed that the dual truck event initially
considered in the 7th Edition AASHTO LRFD is a rare occurrence. As such, the current load
criteria (i.e., a single design truck positioned in all longitudinal and transverse positions) is more
appropriate. This multiple presence study indicates that even when considered, the effects of
passing lane traffic on both the Fatigue I and Fatigue II limit state parameters are negligible.
A comprehensive multiple lane analysis using all available two-lane traffic records demonstrates
that the probability of a single truck record (regardless of GVW) being located in a critical
position for magnifying cross-frame force effects is exceptionally low—less than 0.02% for the
most critical location (i.e., the steering axle of one truck is located just behind the rear axles
of another truck in an adjacent lane) and less than 3% when the truck is located within 20 to
50 feet from the rear axle of another truck in an adjacent lane.
variable to pinpoint which cross-frames are critical without some refined, influence-surface
analysis (i.e., to examine all possible truck positions and cross-frame force effects).
• The correlation between specific variables and cross-frame fatigue design forces was
explored. In general, increased skew and tighter bridge curvature tends to increase cross-
frame force demands in composite systems. These general rules-of-thumb can potentially be
utilized by engineers to mitigate cross-frame forces in the design phase of a project. In other
words, the engineer can potentially avoid an iterative “chase-your-tail” design solution by
adjusting other key parameters. For instance, using discontinuous, staggered cross-frame
layout is a practical and economical solution, whereas increasing the deck thickness or
decreasing girder spacing has significant impacts on the rest of the design.
• From the fatigue truck study, it was concluded that additional economies in the AASHTO
fatigue criteria are possible for cross-frame design. This is largely attributed to the sensi
tivity of cross-frame response to transverse load position. Current design criteria require
all possible lane positions be considered in accordance with AASHTO Article 3.6.1.4.3a,
regardless of design lanes or actual lane striping. A more realistic scenario is to consider only
drive lanes (i.e., lanes within the limits of the fascia girders, for which most truck traffic
traverses). Alternatively, reduction factors specific to cross-frames could be applied to
the existing load factors. Based on feedback received from the review panel, cross-frame-
specific load factors outlined in the subsequent bulleted item were explored as the preferred
approach.
• To maintain the current guidance in terms of truck position (i.e., the AASHTO fatigue
truck is located in every transverse position on the bridge deck between the vehicle barriers),
cross-frame-specific fatigue load factors were investigated to eliminate the source of conser-
vatism in the current load factors. The modified load factors developed in this study are
based on a comprehensive WIM study utilizing calibrated bridge models representative of
a variety of straight bridges with normal supports, straight bridges with skewed supports,
and horizontally curved bridges. Assuming the WIM records represent typical truck weights
throughout the country, the improved load factors reflect more realistic force effects experi-
enced by cross-frames in these types of bridges. Using published statistical data for resistance,
the reliability of using these improved load factors was investigated. The Monte Carlo
simulation using the available resistance data and cross-frame statistics summarized in this
report result in reliability indices as shown in Table 3-9 when using an adjusted load factor
of 0.65 applied to both of the current Fatigue I and II limit state load factors (i.e., a respec-
tive Fatigue I and II load factor for cross-frames of 1.14 and 0.52). Based on this analysis,
the resulting reliability indices for cross-frames using these adjusted load factors meets or
exceeds the assumed target reliability of unity (Modjeski and Masters 2015).
• For the analytical and experimental analyses conducted in this study, the number of design
stress cycles per truck passage (n in AASHTO LRFD Table 6.6.1.2.5-2) was taken as 1.0.
The proposed load factors above were subsequently calibrated using this n = 1.0 assumption.
• In implementing the reduced cross-frame-specific load factors proposed, it was observed
that load-induced fatigue problems are likely only an issue in moderately to heavily skewed
and/or curved bridges (i.e., Ic > 1, Is > 0.15, or Ic > 0.5 & Is > 0.1) in areas of significant truck
traffic volume. For bridges that do not fall into one of these categories, a refined 2D or 3D
analysis is not warranted. These limits are similar to those established in NCHRP Research
Report 725 (White et al. 2012), which focused primarily on analysis of noncomposite systems,
except that the skew index limit proposed in that study is lowered herein from 0.3 to 0.15.
A discussion on the appropriate analysis method is outlined in the subsequent sections.
These findings, which are further summarized in Chapter 4, served as the basis for many of
the proposed modifications to AASHTO LRFD. Refer to Appendix A for the proposed language
and commentary.
These questions are systematically addressed throughout this section of the report. Section 3.3.1
summarizes key findings from a panel-level study, which analytically examined the response of
noncomposite cross-frame panels to various load-induced deformation patterns anticipated in
in-service bridges. Section 3.3.2 presents the results of a 3D FEA parametric study that inves-
tigated the response of cross-frames in composite bridge systems. The accuracy of the stiffness
modification factors and the eccentric-beam approach are assessed in terms of predicted cross-
frame force effects for a variety of bridge geometries. Lastly, Section 3.3.3 summarizes the key
findings from the study.
1
0.9
0.8
0.7
Stiffness Modification Factor, R
0.6
0.5
0.4
Equal Rotation Differential Vertical Displacement
0.3
1
0.9
0.8
0.7
0.6
0.5
0.4
Differential Rotation – Same Side Differential Rotation – Opposite Side
0.3
1/4 3/8 1/2 5/8 1/4 3/8 1/2 5/8
Angle Thickness (in)
From Figure 3-21, the following observations can be made with respect to the panel-level
study:
• In general, the parameters of the single-angle section (thickness, leg width) and the aspect
ratio of the panel (girder spacing, panel height) affect the computed modification factor, R.
The R-factor equation presented in Battistini et al. (2016) shows that increasing the S/hb ratio
(girder spacing-to-panel height), –y (distance between the face of the connected leg and the
neutral axis of the angle section), and t (angle leg thickness) results in a decreased R-factor for
the equal rotation condition. Similar trends for composite systems were generally observed
from the results of the studies described in the next section.
• Significant scatter in the results is observed for the differential-rotation case depending on
which cross-frame diagonal is engaged. The same side connections generally result in a more
flexible response due to the additive effect of the end connection eccentricities, whereas the
opposite side connections are much stiffer.
• Even for parameters dictated by designers (e.g., angle thickness, angle leg width, girder
spacing, panel height, gusset thickness), there can be substantial variability in the R-factor
response. Additional variability is also introduced by parameters not controlled by designers
(e.g., deformation pattern and the force distribution due to truck position).
Additional spot checks were performed to investigate the effects of different connection
plate dimensions (i.e., slenderness ratios) and K-type cross-frames. In brief, it was evident
that the connection plate slenderness had minimal effect on the computed R-factor for all
loading conditions evaluated. In terms of the K-type cross-frames, the “same and opposite
side” effects were not observed given that diagonals in K-frames typically connect to the
same face of the corresponding gussets. As such, it can be concluded that K-frames are not as
sensitive to the deformation pattern given the symmetry of the end connections. For a more
thorough review of these spot check analyses, refer to Appendix F.
Given the variability and uncertainty in the response due to different geometries and load-
ing conditions, developing a closed-form, general-use solution similar to that suggested in
Battistini et al. (2016) would be challenging and impractical for design. From a simplicity stand-
point, assigning a single modification factor for each cross-frame member, similar to the current
approach in AASHTO LRFD, is desirable. This is particularly true since the orientation of the
cross-frame connections (i.e., increasing or decreasing eccentricities) may not be known during
the design process. With that in mind, the system-level studies in the next subsection evaluate
the most appropriate modification factor to assign for the majority of bridge superstructures in
service. An independent parametric study that quantifies the impact of an assumed R-factor on
the cross-frame force prediction in composite bridge systems is outlined.
Key:
3
1 2
1.6
1/2-inch Gusset
1.4
Diagonal 1 (R = 0.63)
1.2
1
Fsimplified / Fshell (Critical Loading)
1
Bottom Strut (R = 0.95)
0.8
Diagonal 1 (R = 0.70)
0.6
0.4
0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 1
Assigned R-Factor (Truss Model)
but neglected in the figures for clarity. For more information regarding the pertinent parameters
of the superstructure, refer to Appendix F.
The following observations can be made from Figure 3-22:
• The cross-frame force effect is not 1:1 dependent on the stiffness modification due to the
high degree of indeterminacy in the bridge system. In other words, a 10% reduction in the
R-factor does not result in a 10% reduction in the estimated cross-frame force.
• Rather, a 10% reduction in the R-factor generally results in approximately a 5% reduction
on average in the estimated cross-frame force. That relationship typically holds true for
modification factors above 0.6. Below R = 0.6, the rate at which the Fsimplified /Fshell approaches
zero increases. Ultimately, a modification factor of zero results in Fsimplified /Fshell = 0, despite
not being shown in the figure. This assumed 2:1 relationship (x% reduction in R ≈ x/2%
reduction in force) has design implications. For example, if an acceptable level of error in the
predicted cross-frame force is established as 10%, the R-factor likely needs to be within 20%
of the exact value (recall that it is conservative from a fatigue perspective to assign higher
R-factors).
• Even for cross-frame members of similar geometries and eccentricities, the “exact” R-factor
for these critical loading conditions can vary. For example, the exact R-factor for the bottom
strut is 0.73 compared to 0.63 for the top strut in the same cross-frame panel. This variability
highlights the impact of load position and deformation patterns on the assumed stiffness.
• The gusset plate thickness also impacts the response of the cross-frame in the shell-element
model. As the gusset plate thickness increases, the stiffness of the cross-frame tends to also
increase. As noted earlier, although the eccentricity is larger with a thicker gusset plate, the
bending stiffness of the plate increases. The reason for this is that the eccentricity is a linear
function while the bending stiffness changes as a cubic function. As such, the exact R-factor
for the bridge with 1-inch thick gusset plates tends to exceed the exact R-factor for the
bridge with 1⁄2-inch thick gussets. For example, the exact R-factor for the diagonal member
evaluated was 0.63 for 1⁄2-inch gussets and 0.70 for 1-inch gussets.
• In general, the sample results presented in Figure 3-22 are representative of the full data set
in terms of the observed variability and trends.
Figure 3-22 represents the results of just one bridge in the sample set. Rather than evaluate
the effectiveness of the R-factor approach for every cross-frame member and every transverse
lane passage, it is more important to just focus on the governing cross-frame member in every
bridge (i.e., the cross-frame with the maximum force range). In this manner, more generalized
observations about the stiffness modification factors can be made. Note that a similar approach
was taken when compiling the data obtained from the Fatigue Loading Study.
Thus, for each iteration of a given bridge model, Fsimplified /Fshell ratios were computed and
compiled in the form of box-and-whiskers plots. The results of the eccentric-beam approach
(“refined model” only; refer to Figure 2-15) were also included to examine its potential benefits.
Note that, while the results herein focus on the refined eccentric-beam model only, an examina-
tion of the simplified versions is presented later in this section.
For Figure 3-23, the box-and-whiskers components are organized by the assigned R-factor
in the truss-element model, as well as the eccentric-beam model (which is independent of an
assigned R-factor). Therefore, each box-and-whiskers represents a data set of Fsimplified /Fshell ratios
corresponding to the truss-element models with R = {0.5, 0.6, 0.7, 0.8, 0.9, and 1.0} and the
eccentric-beam model. By presenting the box-and-whiskers of the results for various R-factors
side-by-side, the research team can evaluate which fixed factor generally produces the most
accurate representations of the “true” cross-frames stiffness (i.e., the shell-element model).
Similarly, the variability of the R-factor approach can be compared to the eccentric-beam
approach, which more accurately represents the stiffness of the cross-frame and its connections.
Statistical parameters such as the mean, minimum, and maximum values of the data set are
also provided. For the R = 0.5 example, the respective mean and maximum Fsimplified /Fshell ratios
were 0.82 and 0.98, which indicates that assigned R = 0.5 always underpredicted the design
force when compared to the benchmark shell-element model.
1.4
1/2-inch Gusset
Fsimplified / Fshell (Governing Member)
1.3
1.2
1.1
1.0
0.9
25th Percentile
0.8
0.7
0.6
0.50 0.55 0.60 0.65 0.70 0.75 0.80 0.85 0.90 0.95 1.0 Ecc.
Beam
Assigned R-Factor (Truss Model)
The figure only presents the results related to the bridges analyzed with 1⁄2-inch-thick gusset
plates. However, bridges with 1-inch-thick gusset plates were evaluated, and the results are
subsequently provided in Appendix F. In addition to the box-and-whiskers, a line graph
representing the 25th percentile of each data set is provided, whose meaning is outlined below.
Similar to the sample bridge in Figure 3-22, the following observations can be made about
the stiffness modification (R) factor approach with respect to the full data set presented in
Figure 3-23:
• As the assigned R-factor increases, the design force in the truss-element model increases,
thereby increasing the Fsimplified /Fshell ratio. In other words, the R-factor and Fsimplified /Fshell ratio
are positively correlated.
• Despite the positive correlation, there is still significant variability in the observed response.
The accuracy of the truss-element model in terms of predicted cross-frame forces is not
only a function of the assigned modification factor but also the bridge geometry and loading
conditions. For example, the respective maximum and minimum Fsimplified /Fshell ratios for the
R = 0.5 case are 0.98 and 0.73.
• Although not explicitly shown in this report, the Fsimplified /Fshell ratio decreases as the gusset plate
thickness increases. This is attributed to the fact that an increased gusset thickness results in a
stiffer cross-frame, which then attracts more force to the shell-element model.
From the observations above, it is apparent that developing an expression that precisely
predicts the exact R-factor for any general bridge model (i.e., the R-factor for which Fsimplified /Fshell
always achieves unity) is not practical or feasible. The exact R-factor is a function of many
parameters, including bridge geometry, loading conditions, cross-frame details (e.g., gusset
plate thickness), and member type (e.g., bottom strut versus diagonal). Thus, it is more appro-
priate to assign a uniform R-factor to all cross-frame members that statistically represents the
majority of bridge conditions, similar to the R = 0.65 approach currently adopted in AASHTO
LRFD that was developed for the construction condition of the bridge.
Based on these results, there are three potential approaches for handling the variable response
of the R-factor approach in terms of AASHTO LRFD. These are listed below from most
sophisticated to most simplistic as well as potential pros and cons to each approach:
• Option 1: Develop a member-specific and bridge-specific expression that provides a precise
R-factor. Given the variability in the response and the uncertain nature of live loads on bridge
structures, this approach is neither feasible nor practical for implementation in AASHTO.
The expression would inevitably be extremely complex and would still likely produce
scattered results.
• Option 2: Develop an R-factor that is a direct function of the gusset plate thickness. Rather
than develop an expression that explicitly considers all pertinent bridge and cross-frame
parameters (e.g., girder spacing, angle thickness), this approach simply focuses on gusset
plate thickness, which has been shown to substantially impact the exact R-factor. This
procedure could be easily implemented as an equation or a table. However, given that the
gusset plate thickness may be variable during the design process, this approach potentially
lends itself to an iterative analysis procedure. Even with this extra level of refinement, the
results could still be variable.
• Option 3: Develop a single R-factor that generally represents the behavior of all cross-frame
members in a conservative manner (similar to the current 0.65 approach in the 9th Edition AASHTO
LRFD that was developed for the construction condition). As demonstrated with all previous
results, pinpointing an exact R-factor is very difficult given its inherent dependence on the
loading conditions and the associated deformation patterns. Regardless of how the cross-
frames are modeled and R-factors are assigned, some error in cross-frame force predictions
is to be expected when compared to the shell-element modeling approach. With that said,
assigning a single R-factor for all cross-frame members, regardless of connection details,
bridge geometry, and loading conditions, is a viable option.
After weighing the options, it was elected to recommend the use of a single R-factor
(Option 3) similar to the current approach taken in AASHTO Article C4.6.3.3.4. With that in
mind, it is also not appropriate to ensure every bridge model results in conservative estimates
of cross-frame fatigue forces (Fsimplified /Fshell ≥ 1). Instead, the research team investigated different
statistical parameters but ultimately elected to base the proposed modification factor on the
25th percentile value of the data sets. In other words, the intersection of the 25th percentile line
and Fsimplified /Fshell = 1.0 represents the assigned R-factor that provides a conservative estimate
of cross-frame design forces for 75% of the bridges. For the 25% of the bridges that result in
unconservative estimates, it has been shown that the truss-element model is still within 10%
of the rigorous shell-element model, which was deemed reasonable by the research team.
Using that metric, the most appropriate R-factor to assign for the bridges with 1⁄2-inch gussets
is 0.75. As such, in addition to the stiffness modification factor of 0.65 that was derived for a
noncomposite system during construction, a second factor of 0.75 is proposed for structural
analysis of the composite condition.
Table 3-11 tabulates the results outlined above, as well as additional data corresponding to
different gusset plate thickness values and statistical parameters. For instance, if the in-service
R-factor is based on the 50th percentile (i.e., mean) of the data set with 1⁄2-inch-thick gusset
plate, then R = 0.66 is more appropriate, which is approximately the same as the established
modification factor in AASHTO LRFD (R = 0.65).
Although the R-factor approach provides a method of correcting truss-element models for
the reduction in stiffness due to eccentric connections, some designers or software producers
may opt to make use of a model that directly considers the connection eccentricity. In terms of
the proposed eccentric-beam results in Figure 3-23, it was observed that the refined eccentric-
beam models predict the cross-frame fatigue design forces with reasonable levels of accuracy.
For the 1⁄2-inch gusset case, the Fsimplified /Fshell ratio is 0.96 on average with a maximum of 1.03 and
a minimum of 0.86. In other words, the eccentric-beam produces results within 4% of the
shell-element model on average.
As outlined above, the results presented in Figure 3-23 are based on the “refined” eccentric-
beam model only. To illustrate the impacts of the more simplified eccentric-beam models,
Figure 3-24 presents sample influence-line results from the same two-span, continuous skewed
bridge from Figure 3-22. In this figure, two different responses are examined: (i) the axial force
in the bottom strut member near the skewed intermediate support when the AASHTO fatigue
truck traverses the bridge along the barrier and (ii) the axial force in the bottom strut member
(interior bay) near the maximum positive dead load moment region when the AASHTO fatigue
truck traverses the bridge along its centerline.
Appropriate R-factor
Gusset thickness [in]
Mean 25th Percentile
BS
BS Case 1
Case 2
-3
-6
Axial Force (kips)
Case 1
-9
0 210 420
6 CF Model FRa (kips) % Error
R = 1.0
Shell-Elements (Control) 4.8 --
Truss-Elements; R={0.5, 1.0} {3.9, 5.6} {-21%, +14%}
Ecc. Beam; Refined 4.6 -6%
3 Ecc. Beam; Simplified 4.1 -16%
Ecc. Beam; Angle (Fixed Ends) 4.8 -2%
R = 0.5 Ecc. Beam; Angle (Pinned Ends) 3.2 -35%
For each cross-frame response, several different results are presented, including the shell-
element model (which serves as the control), the truss-element model range (i.e., R = 0.5 and
R = 1.0), the refined eccentric-beam model, the simplified eccentric-beam model, and the
truss-only eccentric-beam models (both fixed-end and pinned-end). The various responses are
superimposed on the same graph. Additionally, the total force ranges and the percent error
relative to the control model are tabulated for reference. Note the scale along the vertical axis
is different for the top and bottom plots in Figure 3-24. Refer to Appendix F for a detailed
description of each eccentric-beam modeling approach.
The following observations can be made from the sample data set presented in Figure 3-24:
• Similar to the results presented above, the truss-element model approach has significant
variability depending on which R-factor is assigned. Note that a modification factor of
approximately 0.75 would produce the most accurate results in both cases shown in
Figure 3-24, which is consistent with the conclusions of the R-Factor Study presented in
the previous section.
• In both loading scenarios, the refined eccentric-beam model produces relatively accurate
results when compared to the shell-element model (6% conservative). This is a similar
observation to the results of the parametric study presented above. For instance, these two
data points are close to the mean of the eccentric-beam data set (box-and-whiskers) shown
in Figure 3-23 (i.e., a Fsimplified /Fshell ratio of 0.94 compared to 0.96 in Figure 3-23).
• The simplified eccentric-beam model, which neglects the individual eccentric offsets caused
by the connection and gusset plates, has more variability but generally produces reasonably
accurate and conservative results. This approach is much simpler to employ in practice and
requires much less information about the cross-frame connection details.
• The angle-only models are largely dependent on the assumed end restraints. When the beam
ends are fixed rotationally to the girder web, the stiffness of the cross-frame panel tends
to be overestimated, which results in slightly overconservative force effects (but still less
conservative than the truss-element model assuming R = 1.0). Despite the eccentric offset
being explicitly modeled, the out-of-plane flexural rigidity is greatly overestimated at the end
connections by using the angle section properties for the full length of the modeled element.
• When the beam ends are pinned (to account for the rotational flexibility of the connection
plates), the overall cross-frame stiffness tends to be vastly underestimated (even more under-
estimated than the truss-element model with R = 0.5). Thus, the cross-frame force effects are
significantly unconservative with respect to the shell-element model. It can be observed from
this sample data that an eccentric-beam approach, for which rotation is released at its ends,
is not an appropriate 3D modeling technique for cross-frames.
• In review of the angle-only models, the “real” rotational restraint at the ends of the beam
elements is somewhere between the idealized fixed and pinned scenario. Although not
studied here, a rotational spring with finite stiffness could be explored as an alternative as well.
The most significant observation from the sample results in Figure 3-24 is that accurately
representing the out-of-plane flexibility associated with the connection and gusset plates is
vital for eccentric-beam models. The angle-only models generally do a poor job of representing
that end flexibility (assuming idealized fixed- or pinned-end conditions), whereas the simplified
and refined models showed much more promise.
In general, the eccentric-beam approaches eliminate several sources of uncertainty associated
with the R-factor approach (e.g., eccentric distances, relative stiffness of gusset and connection
plates, effect of load position on the deformed shape). However, modeling a cross-frame based
on the refined methodology requires knowledge of the gusset and connection plate thicknesses
prior to designing those elements. This potentially lends itself to an iterative process where the
engineer starts with an assumed geometry (likely based on standard DOT details) and updates
the model based on project-specific design decisions. This iterative process, though, affects
the shell-element and refined eccentric-beam approaches equally. In contrast, the simplified
methodology, although less precise and accurate, reasonably approximates the flexibility of the
connection plate without having full detailed dimensions of the cross-frame connections.
Based on these perceived advantages, this modeling technique has potential to serve as
an alternative for bridge designers and commercial design software packages. As a reference,
the current commentary language in AASHTO LRFD (C4.6.3.3.4) that discusses the use of
the R-factor is provided herein:
“In addition, the axial rigidity of single-angle members and flange-connected tee-section cross-frame
members is reduced due to end connection eccentricities (Wang et al., 2012). In lieu of a more accurate
analysis, (AE)eq of equal leg single angles, unequal leg single angles connected to the long leg, and flange-
connected tee-section members may be taken as 0.65AE.”
As underlined in the provision above, the eccentric-beam approaches (namely the refined
and simplified approaches) qualify as a “more accurate analysis” but do not require bridge
designers to assemble labor-intensive, 3D FEA models with cross-frames modeled as shell
elements. The overall conclusions with regards to this proposed analysis procedure are provided
in Section 3.3.3 below.
By comparing the performance of 3D models, 2D PEB models, and 2D grillage models, these
questions are systematically addressed throughout this section of the report. Similar to the
Fatigue Loading Study, an abundance of cross-frame data was obtained from the Commercial
Design Software Study. To simplify the results and discussion for the reader, a similar outline
with respect to data presentation is followed for the discussion of all models. Representative
data are first shown in Sections 3.4.1 and 3.4.2. These sections address different aspects of the
results. Section 3.4.1 presents sample influence-line results to introduce the general limita-
tions of 2D analysis methods, and Section 3.4.2 investigates the influence of composite action
on the accuracy of these 2D models. Section 3.4.3 compiles and summarizes the results for the
full data set. Section 3.4.4 proposes an alternative postprocessing procedure for 2D PEB analyses,
and Section 3.4.5 summarizes the major outcomes.
1 2
8 Key:
Case 1
[Approach; Equiv. Beam]
6 A Control (Abaqus)
B 3-D
4
A/B D-2 C-1 2-D PEB; Shear
2 C-3 D-3 C-2 2-D PEB;
D-1
Flexural
C-2
0 C-3 2-D PEB;
C-1 Timoshenko
-2
D-1 2-D Grid;
2 0 A/B
240
Flexural
D-3
0 D-2 2-D Grid;
Axial Force (kips)
C-1 Timoshenko
D-1
-2 C-2 C-3 D-3 2-D Grid;
D-2 Composite
-4 Note: Framing plan
sketch above not to scale
-6
Case 2
-8
4 0 240
Case 3
2 A/B
D-1/D-2/D-3
0
C-2/C-3
-2 C-1
-4
-6
0 240
Longitudinal Position of Truck (ft)
to capture the torsional response of the bridge and the complex load paths through the structure
in each case. These geometric effects are addressed in subsequent sections.
The axial-force response of the three select cross-frame locations is shown: (i) a bottom strut
near midspan referred to as Case 1, (ii) a top strut from the same cross-frame panel referred
to as Case 2, and (iii) a diagonal near the support referred to as Case 3. This is demonstrated
schematically in the framing plan sketch in Figure 3-25.
The AASHTO fatigue truck was moved longitudinally along the centerline of the bridge as is
depicted. Figure 3-25 plots the axial-force response of the respective cross-frame members as
a function of the truck position along the bridge length, measured from the front axle to the
start of the bridge. These graphs are similar to the influence-line plots presented in Figure 3-24.
This process was repeated for several variations of the same bridge model including the 3D
validated approach in Abaqus (control) and the following methods in the commercial design
software program Software A (recall that Software A has the ability to create either a 3D model
or a variety of 2D models including PEB and grillage models):
• 3D model,
• 2D PEB model with equivalent cross-frame beams based on the shear-analogy approach,
• 2D PEB model with equivalent cross-frame beams based on the flexural-analogy approach,
• Improved 2D PEB model with equivalent cross-frame beams based on the Timoshenko
approach,
• 2D grid/grillage model with equivalent cross-frame beams based on the flexural-analogy
approach,
• 2D grid/grillage model with equivalent cross-frame beams based on the Timoshenko
approach, and
• Improved 2D grid/grillage model with equivalent cross-frame beams considering the
transverse stiffness of the concrete deck and the modified postprocessing procedure per
Figure 2-17 (abbreviated as 2D Grid Composite in the Figure 3-25).
The specifics of these 2D modeling approaches (e.g., equivalent cross-frame beam analogies)
were outlined in Section 2.5. Note that, for the PEB and grillage models, equal distribution
of cross-frame shear was assumed between top and bottom nodes when postprocessing the
respective analysis results. The implications of using a different distribution assumption are
addressed in subsequent sections. For more information regarding the overall geometry and
cross-frame layout of this sample bridge as well as additional influence-line results, refer to
Appendix F.
From Figure 3-25, the following observations can be made when comparing the results of a
specific cross-frame location across each bridge model:
Case 1 (bottom strut near the maximum positive dead load moment region):
• The 3D commercial design software model produces excellent results when compared to the
validated 3D FEA model in Abaqus (i.e., estimated force range within a few percent). Given
that the commercial design software modeling approach is nearly identical to that of the
control model, this trend was anticipated. This also indicates that the connection point of
the cross-frame members (i.e., into a shared node on the web-flange junction versus at an
offset distance along the web) is less impactful for 3D model results.
• The 2D PEB results were not significantly influenced by the equivalent beam method employed
for these composite conditions. In general, these models produced accurate estimates of
bottom strut forces when compared to the refined 3D models.
• The 2D grid models generally did a poor job at predicting the cross-frame forces, except when
the Timoshenko beam approach was implemented.
Case 2 (top strut near the maximum positive dead load moment region):
• Similar to Case 1, the 3D commercial design software model generally produces excellent
results when compared to the validated 3D FEA model in Abaqus.
• For the 2D analyses that do not consider the contributions of the deck when postprocessing
the analysis results, the top strut results are equal in magnitude and opposite in sign to those
results of Case 1. This is a function of the simplified postprocessing used when converting
the equivalent beam back into a truss. As illustrated in this example, this simplification
generally results in erroneous results for top strut forces.
• This error is corrected in the 2D grillage model using cross-frame beams with equivalent
section properties to account for the contributions of the deck. Albeit the correct sign was
obtained, the magnitudes are still generally overestimated.
Case 3 (diagonal near the intermediate/end support):
• In general, the observations for Case 3 are similar to those for Case 1 except that the 2D
approaches consistently underpredicted the cross-frame force effects.
In reviewing data beyond these sample results (specifically for bridges with X-type cross-
frames), it was evident that the predicted response of bottom strut member forces in 2D models
was much more accurate than the predicted response of diagonal or top strut members. This
behavior was apparent in both simple and complex bridge geometries. In other words, the
variability observed in the response was not only a function of the bridge geometry (i.e., straight
versus curved, normal supports versus skewed supports) but also the cross-frame member type
(i.e., bottom strut versus diagonals).
The variability associated with bridge geometry is attributed more to analysis error (i.e., how
the analysis determines end shears and moments on the equivalent cross-frame beam), whereas
variability associated with member type is attributed more to postprocessing error (i.e., how
those end shears and moments are converted into cross-frame member forces). With that in
mind, Section 3.4.2 examines this behavior in the context of noncomposite and composite sys-
tems to determine the root cause of these discrepancies.
100 k
2.6 k
4.2 k 4.2 k
8.4 k
120” 120”
2D PEB (Noncomposite System) 3D (Noncomposite System)
21.1 k
14.7 k 14.7 k
20.8 k
• 2D PEB model of a composite system utilizing the Timoshenko beam approach for equiva-
lent cross-frame beam properties and the equivalent torsion constant (Jeq) for girder beam
properties (top left quadrant),
• 2D PEB model of a noncomposite system utilizing the Timoshenko beam approach for
equivalent cross-frame beam properties and the equivalent torsion constant (Jeq) for girder
beam properties (bottom left quadrant).
Given that the preliminary results presented in Figure 3-25 demonstrated the benefits of
PEB models over grillage models, the results in this section are focused on just PEB methods.
For the 3D analyses (right side of figure), the axial-force output from the cross-frame members
is presented, followed by the nodal forces required to equilibrate the panel. For the 2D PEB
analyses (left side of figure), the shear and moments produced from the equivalent beam
analysis are presented, followed by the cross-frame force effects computed based on the common
postprocessing assumptions previously outlined.
For each figure, there are three metrics by which the results can be compared, as listed below:
1. The total shear force acting on the cross-frame panel relative to the shear force in the concrete
deck: This value provides an indication of how accurate the equivalent beam properties
in the 2D model are. For instance, if the total shear force determined from the 3D model
exceeds that of the 2D PEB model, then the stiffness of the equivalent cross-frame is likely
underestimated. The opposite would be true if the total shear force of the 2D PEB model
exceeds the 3D model.
2. The distribution of that total shear force to the top and bottom nodes: Girder displacements
in noncomposite systems, particularly rotations, are different than girder displacements in
composite systems. To demonstrate this behavior in simple terms, an idealized deformation
pattern is provided in Figure 3-27. Figure 3-27 schematically depicts a common displaced
shape of an X-type cross-frame panel under applied vertical loads; a noncomposite system
is compared to a composite system. Note that the displacements are greatly exaggerated and
that rigid-body motion has been removed for clarity.
In comparing the noncomposite and composite systems, it is evident that noncomposite
girders generally rotate out of plane about their centroid under applied vertical loads.
Consequently, the top and bottom struts tend to deform equal magnitudes but in oppo-
site directions, which results in equal-and-opposite axial-force effects in those members.
In contrast, composite girders tend to rotate out-of-plane about an axis higher on the
cross-section due to composite action with the concrete deck. As a result, the deformation
demand on the top strut is small, as it is located near the point of girder rotation. In relative
terms, the bottom strut deforms significantly more, which is illustrated in Figure 3-27 with
the original position of the system lightly shadowing the deformed position. Under this
idealized displaced shape, the top strut sees little to no force effects whereas the bottom
strut sees substantial axial tension. To maintain compatibility with the girder displacements,
both diagonal members in the X-frame must also deform under axial tension.
Thus, when this idealized differential-rotation deformation pattern is paired with differen-
tial vertical displacement (which generates substantial diagonal forces approximately equal-
and-opposite in magnitude), the net effect is unbalanced diagonal forces. Cases in which the
diagonals are both in tension or both in compression are, therefore, not uncommon. This
generally explains why the equal-and-opposite force distribution in X-type cross-frames is
observed in noncomposite systems but not in composite systems where realistic deformation
patterns are highly complex. For K-type cross-frames, this discussion is less influential as
equal-and-opposite force distribution in diagonal members is anticipated given the layout
of the members.
Assuming equal shear force distribution in the postprocessing phase of 2D PEB models
generally leads to poor results for cross-frame diagonals. With that in mind, evaluating the
“true” shear force distribution from the 3D models provides an indication of how significant
the effects illustrated in Figure 3-27 are.
3. The relative difference between top strut forces and bottom strut forces: Similar to the discussion
in Item #2 above, it is anticipated that bottom strut force effects will greatly exceed top strut
force effects in composite systems. By evaluating the relative difference between these two
magnitudes, the effects of composite action can be quantified. The sign of the top strut force
can also provide an indication as to where the point of girder rotation exists along the depth
of the cross-section (e.g., if the top and bottom struts are both in tension, then the point of
rotation is likely above the top strut).
In the context of the three comparative metrics established above, the following observations
can be made from Figure 3-26:
• Total shear force: For the noncomposite system, the total cross-frame shear force estimated
from the 2D PEB model is nearly identical to that of the 3D model (i.e., 14.8 kips versus
14.6 kips). This indicates that the Timoshenko beam analogy accurately represented the
effective cross-frame stiffness by considering both shear and flexural deformations. For
reference, the shear-analogy approach produced a total cross-frame shear of 10.6 kips for this
same example, despite not being shown in the figure. This indicates that the shear-analogy
underestimates the cross-frame stiffness. For the composite system, the total cross-frame
shear force is underestimated in the 2D PEB models (e.g., 4.2 kips versus 5.7 kips). This
relationship is explored further in this section.
• Shear force distribution: For the noncomposite system, the 3D model demonstrates that
shear force is equally distributed to the top and bottom struts (subsequently resulting in
equal-and-opposite diagonal force effects). As a result, the “50-50” distribution assumption
commonly utilized in 2D postprocessing is valid for noncomposite conditions. Thus, the
2D models accurately predict the force effects in the diagonal members for the specified
loading condition.
For the composite system, the 3D analysis model demonstrates that load-induced forces
in diagonal members are not equal and opposite. Figure 3-26 shows 2.3 kips of compression
in one diagonal and 10.4 kips of tension in the other, resulting in an “80–20” distribution of
the total shear force. Thus, when comparing cross-frame forces results, it is evident that the
2D PEB postprocessing procedure produces inaccurate results.
• Moment force couple distribution: For the noncomposite system, the 3D analysis results
indicate that the load-induced forces in the top and bottom struts are nearly equal and
opposite, as expected (i.e., 21.1 kips of compression in the top strut and 20.8 kips of tension
in the bottom strut). Therefore, the 2D PEB model accurately predicts top and bottom strut
forces when compared to the 3D counterpart model.
For the composite system, this equal-and-opposite relationship does not hold true. The
3D model indicates that the top strut force effects share the same sign as the bottom strut
force effects, but the magnitude is significantly less. Based on the postprocessing assumption,
the 2D PEB model produces equal-and-opposite forces in the top and bottom struts.
Thus, there is substantial error associated with the force prediction in the top strut member
in the composite system for the 2D PEB model. However, the error in the force prediction
in the bottom strut member is minimal (i.e., within 5% conservative in both cases). Despite
the inconsistencies and oversimplifications in the 2D analysis and postprocessing, this same
general trend related to bottom strut members is typically observed assuming the equivalent
beam properties are accurate (i.e., the Timoshenko beam approach is utilized). The bottom
strut results in Figure 3-25 are additional examples of this trend.
Given that 2D PEB models generally produce reasonably accurate estimates for bottom
strut force effects, especially for simple bridge geometries, it is important to note when these
limitations and shortcomings become critical. If a bottom strut cross-frame member maximizes
load-induced force effects and governs fatigue design, then this simplified analysis and post-
processing procedure likely produces reasonable estimates of the governing design forces.
However, if diagonal members (or much less likely, top struts) govern the fatigue design,
then 2D PEB approaches may grossly underestimate the governing force effects. To expand
on the findings from Figure 3-26, which focuses on a simple straight bridge with normal
supports, the research team repeated these same studies to examine more complex bridge
geometries and loading scenarios.
Two additional bridges were examined similar to the sample bridge presented above: a straight
bridge with skewed supports (X-type cross-frames) and a curved bridge with normal supports
(X-type cross-frames). Additional studies were performed on different bridge geometries and
cross-frame configurations (i.e., K-type versus X-type); those results are presented in Appendix F.
For the straight and skewed bridge, an interior-bay cross-frame panel near the intermediate
skewed support is examined. For the curved bridge, an interior-bay cross-frame panel near
the maximum positive dead load moment region is examined.
The results outlined above investigated only two different extreme conditions: noncomposite
and full in-service composite (i.e., concrete modulus, Ec, taken as 5,000 ksi). Rather than
evaluate just the two extremes, the research team expanded the study by investigating different
variations of Ec to better quantify how composite action affects cross-frame behavior. The
concrete modulus, Ec, was varied as opposed to the deck thickness as to maintain a consistent
centroid location for the concrete deck and to simplify discussions. The Software A analysis
procedures previously described in this subsection are otherwise unchanged. As such, a total
of 272 unique models were developed as part of this study, including different bridges, various
3D and 2D modeling approaches, and multiple iterations of Ec. Note that the Ec = 0 ksi case
represents noncomposite conditions. Instead of completely removing the deck from the model,
the deck was simulated with virtually no stiffness to expedite the repetitive calculations.
The response of the specified cross-frame members in each bridge was evaluated for two
different static loading cases (100 kips in magnitude) that are arbitrarily identified as Load
Case 1 and 2. For reference, the location of load application relative to the framing plan is
sketched in Figure 3-28 and Figure 3-29. The results presented herein follow the same orga-
nization as the three comparative metrics outlined above: total cross-frame shear, shear force
distribution, and moment force couple distribution.
Figure 3-28 and Figure 3-29 show how the total shear force developed in the cross-frame
panel of interest varies as a function of the concrete deck modulus (or the degree of composite
action, in more general terms). Figure 3-28 focuses on the straight bridge with skewed supports
and X-type cross-frames, and Figure 3-29 focuses on the curved bridge. In these figures, only
the results of 2D PEB models associated with the flexural-analogy and the Timoshenko beam
approach are presented, given the inherent inaccuracies with the shear-analogy approach
outlined previously.
Figure 3-30 demonstrates how the total shear force is distributed to the top and bottom
nodes per the 3D analysis model. Since the 2D models assume a “50-50” distribution, results
related to 2D models are excluded as they do not add value to the graph. In Figure 3-31, the ratio
between the top strut force and the bottom strut force is plotted as a function of the assumed
concrete deck modulus. The results of the two sample bridges are compiled and provided. Note
that these results correspond to the 3D models only, as the 2D models assume an equal-and-
opposite force couple in the postprocessing phase. Thus, the 2D analysis results do not add
significant value to the graph.
Case 1 (100 k)
5
Case 1 Case 2
Cross-Frame Shear, VCF (kips)
The following additional notes provide important insight to the development and construc-
tion of Figure 3-28 through Figure 3-31:
• A cross-frame diagram is provided in each plot to illustrate the nodal force or internal axial
force examined in the graph. For instance, Figure 3-28 and Figure 3-29 present the definition
of the total cross-frame shear force, VCF.
• As stated above, a modulus value of 0 ksi represents a noncomposite system, whereas values
between 3,000 and 5,000 ksi are representative of most in-service, composite bridges in the
United States. The horizontal axis is presented on a log scale to clarify the noticeable change
in behavior between 100 ksi and 1,000 ksi.
VCF VCF
30
Case 1 Case 2
Cross-Frame Shear, VCF (kips)
25
20
15
1.0
0.5
0.0
-0.5
-1.0
0 1 10 100 1000 10000 0 1 10 100 1000 10000
Concrete Modulus (ksi; log scale)
• In Figure 3-30, the shear force acting on the top node (Vtop) and the bottom node (Vbot) are
compared independently with the total cross-frame shear (VCF). Thus, the sum of all Vtop /VCF
and Vbot /VCF ratios equals one. Note that Vtop and Vbot are arbitrarily taken relative to the
left side of the cross-frame panel as shown in the sketch. Had the shear forces been taken
from the opposite side, the values of Vtop and Vbot would be flipped. The general behavior,
however, would remain the same given that the shear force is constant across the width of
the cross-frame panel.
• In Figure 3-31, the ratio of the top strut force (PTS) and the bottom strut force (PBS) is
compared to a similar ratio determined by elastic beam theory. In other words, the concrete
deck, top strut, and bottom strut are treated as three force couples that develop the moment
strength of a pseudo-composite transverse beam, similar to a reinforced concrete beam.
The elastic neutral axis of this pseudo-composite beam is derived after transforming the
PTS
Key: Elastic Beam Theory
Load Case 1 Results
Load Case 2 Results PBS
1.0
Straight, Skewed
Normal Curved, Normal
Cross-Frame Force Couple
0.5
Distribution, PTS / PBS
0.0
-0.5
-1.0
-1.5
0 1 10 100 1000 10000 0 1 10 100 1000 10000
Concrete Modulus (ksi; log scale)
effective concrete area with the corresponding modular ratio. Assuming an elastic stress
distribution, the PTS /PBS ratio is related to beam curvature and is simplified as yTS /yBS, where
yTS and yBS are the distances measured from the centroid of the respective strut to the
computed neutral axis. Therefore, by comparing the analysis results to this idealized beam
theory approach, the impact that composite action has on the distribution of bending
moments through the deck and cross-frame panel can be evaluated.
The following observations can be made with regard to the total shear force estimated by
the various analysis methods (Figure 3-28 and Figure 3-29):
• As the concrete modulus increases beyond zero (noncomposite), it is intuitive that the
cross-frame forces generally decrease. Not only does the relative stiffness of the concrete
deck increase and thus attracts more force in the highly indeterminate system, but the girder
displacements diminish as the global stiffness increases. However, as the concrete modulus
approaches values exceeding 3,000 ksi or more, the relative change is less significant.
• In general, the Timoshenko beam approach is accurate for noncomposite systems when
compared to the 3D analysis, but the variability increases slightly for composite systems.
Still, it is consistently the most accurate 2D equivalent beam method in terms of cross-frame
force predictions, as is presented later in this subsection.
• The observed behavior for Load Case 2 in Figure 3-28 is less predictable. Note that the vertical
axis scale is much smaller than in Figure 3-29 due to the nature of the loading. In Figure 3-29,
the applied load is in close proximity to the cross-frame panel of interest, such that the shear
demands are more significant. In Figure 3-28, though, the applied loads induce a torsional
response to the superstructure, such that the net shear on the cross-frame of interest is
generally small. Consequently, the discrepancy in the total cross-frame shear is less influential
in terms of predicted cross-frame force effects, as is presented later.
The following observations can be made regarding the shear force distribution determined
by 3D analyses (Figure 3-30):
• For noncomposite systems (i.e., Ec = 0 ksi), especially for the non-skewed bridge, the shear
force is equally distributed top and bottom (same result as Figure 3-26 before). The distribu-
tion for the straight bridge with skewed supports is slightly less balanced given the nature of
the cross-frame deformation pattern.
• As the concrete deck becomes stiffer and composite action becomes more substantial, it is
evident that the shear force distribution gradually diverges from the “50-50” assumption. For
instance, the top shear reaction in the straight, skewed bridge for Load Case 2 is approximately
1.6 VCF, whereas the bottom shear reaction is larger than 0.6 VCF in the opposite sign. In this
particular instance, the observed behavior indicates that both diagonals were in significant
compression.
The following observations can be made regarding the force couple distribution determined
by 3D analyses (Figure 3-31):
• For the noncomposite systems, the PTS /PBS ratio consistently converges to -1, which represents
an equal-and-opposite force distribution in the struts. As the deck stiffens and composite
action become more significant, the ratio tends to a small positive value. This indicates that
the top strut shares the same sign as the bottom strut, but the magnitude is substantially
smaller (i.e., a clear indication of composite behavior). In fact, the transition between a nega-
tive ratio and a positive ratio as Ec increases represents the condition in which the center of
out-of-plane girder rotation moves above the top strut. At that point, the top and bottom
struts would be both compressive or both tensile forces.
• In general, the analysis results follow the same trends as the “elastic beam theory” bench-
mark solution, but not exactly. Thus, the flexural action assumed in the development of this
idealized curve is apparent in the realistic data, but the deviation from that curve is due to the
fact that the “true” deformation patterns are not identical to the idealized one.
With general observations established about the validity of common analysis and postpro-
cessing procedures, the next step is to compare the predicted force effects between 3D and 2D
models. Utilizing those common assumptions (i.e., equivalent beam moment resolved as a force
couple and equivalent beam shear equally distributed to top and bottom nodes of X-frames),
2D PEB results are compared to the 3D solutions. The results of the two sample bridges are
compiled and presented in Figure 3-32. In total, four different graphs are presented. The two
columns represent the two distinct load cases examined for each bridge. The two rows represent
the different bridges evaluated as part of this study.
The horizontal axis represents the elastic modulus of the concrete deck and is presented on
log scale. The vertical axis represents the relative error of the 2D PEB model relative to the
corresponding 3D model. Note that only the results using the Timoshenko beam analogy are
shown. Relative error values of zero indicate perfect agreement between the 3D model and
the simplified 2D model. Relative error values exceeding 1.0 indicate the 2D models over
estimate the force effects in excess of 100%; the opposite is true for relative errors below –1.0.
For clarity, data points beyond -1.0 and 1.0 are eliminated from view.
The relative errors associated with the diagonal members and the bottom strut member
are also graphed separately. The lines are color-coordinated based on the cross-frame sketch
provided in the corners of each graph. The results for top struts are excluded given that they
are generally lightly loaded members.
The following observations can be made regarding the relative error associated with the
simplified 2D modeling procedures (Figure 3-32):
• For noncomposite bridges, relative error approaches zero for all member types and all bridge
types. Although not explicitly shown in the report, this is especially true for straight and
Case 1 Case 2
1.0 Key:
0.5
Straight, Skewed
Relative Error (Compared to 3D Model)
0.0
Key:
-0.5
-1.0
1.0 0.1 1 10 100 1000 0.1 1 100 1000 10000
0.5
Curved, Normal
0.0
Key: Key:
-0.5
-1.0
0 1 10 100 1000 10000 0 1 10 100 1000 10000
normal bridges. This is largely attributed to two ideas: (i) the Timoshenko beam approach
accurately represents the cross-frame stiffness in the analysis, and (ii) the common post
processing procedures are valid for noncomposite systems. The error increases slightly with
more complex bridge systems (i.e., skewed or curved bridges) but is still relatively accurate.
For example, the relative error associated with the “red” X-frame diagonal in the skewed
bridge for Ec = 0 ksi is approximately 20% conservative under Load Case 1. Under Load
Case 2, the relative error is not shown largely due to the fact the absolute magnitude of the
force is small. In this case, any slight discrepancies typically result in large error values.
• For the composite systems evaluated (i.e., Ec exceeding 3,000 ksi), the error associated with
bottom strut members is typically small (within 5%–10%). Despite its limitations and
oversimplifications, 2D methods produce accurate force estimates for bottom strut members.
This is consistent with the results presented previously in this section.
• For diagonal members in X-type cross-frames, the errors generally become excessive as Ec
increases. This trend was also observed for straight bridges with normal supports. One diagonal
tends to a large positive value, and the other tends to a large negative value. For example, the
relative error associated with the diagonal members of the curved bridge under Load Case 2
approach +100% and -100% error, respectively, for Ec = 5,000 ksi. This is attributed mainly
to the equal shear force distribution assumed with the postprocessing methods.
• Although not explicitly shown in the report, the errors for diagonals in K-type cross-frames
are much more manageable than those with X-type cross-frames. For these cross-frame types,
recall that the distribution of shear force distribution is not a meaningful discussion.
• Although not explicitly shown, relative error magnitudes associated with 2D PEB models
using the flexural-analogy approach for equivalent beams are more substantial than those
shown in Figure 3-32 for the Timoshenko beam approach. Thus, it can be concluded that the
Timoshenko beam approach is more appropriate for noncomposite and composite systems.
From Figure 3-32, it is apparent that 2D analysis methods produce poor cross-frame force
estimates in two areas: (i) in top strut members and (ii) in X-frame diagonal members regardless
of bridge type. Erroneous, overly conservative results in top struts are not generally impactful
in terms of cross-frame fatigue design. These members typically do not govern fatigue design
such that those results, accurate or not, can be disregarded. Erroneous diagonal member forces,
on the other hand, can have significant impact on the cross-frame fatigue design of a bridge.
As such, it is prudent to reevaluate how designers typically handle the postprocessing of 2D
analysis models to ensure more reliable results.
The major challenge in developing a simple postprocessing modification is the degree in
which the shear force distribution varies. In noncomposite systems, diagonal forces typically
are nearly equal-and-opposite. In the composite systems evaluated as part of Figure 3-30, shear
force distribution was shown to vary from nearly equal to 150% VCF at one node and –50% at
the other. The distribution is a function of bridge type, loading, and subsequently the induced
deformation pattern, which is difficult to quantify without developing a 3D model. An alter-
native method is explored in Section 3.4.4.
member under a specific transverse lane passage maximized force effects, whereas the cor
responding 2D model indicated that the force in the same member was maximized by a
different lane passage. To maintain consistency, the governing lane position was established
based on the control model (i.e., the 3D validated model in Abaqus); the results of the simplified
analysis models were then obtained from those established parameters.
Figure 3-33 presents this data in the form of a box-and-whiskers graph, similar to those pre-
viously shown in the report. Each box-and-whiskers plot represents a set of unfactored design
force effects obtained from the various analysis models. Along the vertical y-axis, the governing
force effects derived for each iteration of the 20 representative bridges are compared to the
control model. The percent error with respect to the results of the control model is reported.
Zero percent error indicates perfect agreement between the simplified commercial design software
models and the control model. Positive error indicates that the simplified analysis conservatively
overestimates the predicted force effects, and negative error indicates the opposite.
In terms of accuracy, it is desirable that the simplified methods provide solutions within about
10% of the validated solution with marginal variance in the box-and-whisker. In terms of a
conservative design approach, it is preferred to have positive error over negative error.
In Figure 3-33, the results corresponding to bottom strut members are differentiated from
diagonal members based on the discussion in Section 3.4.2. In both plots, the results are catego
rized by the modeling technique along the horizontal x-axis (3D; 2D grillage with flexural-
analogy equivalent beams; and 2D PEB with shear-analogy, flexural-analogy, and Timoshenko
equivalent beams).
From Figure 3-33, the following generalized observations can be made when comparing the
analysis error reported across the full data set:
• The 3D commercial design software analysis generally provides cross-frame results (struts
and diagonals) that agree well with the control model. On average, the predicted design force
range was 2% conservative. The observed scatter in the results was also small.
• For the 2D analysis methods, significant variability is observed. This behavior is attributed to:
(i) skewed and/or curved geometry and (ii) the postprocessing errors inherent with the 2D
analysis methods outlined in Section 3.4.2.
20%
0%
–20%
–40%
–60%
–80%
–100%
• The 2D PEB models (shear-analogy) consistently underpredict the governing design force
range (-80% error on average). The 2D PEB models (flexural-analogy) improve upon the
shear-analogy approach but still underpredict the design force ranges.
• The 2D PEB models (Timoshenko beams) provide the most accurate solutions with respect
to 2D models but still result in significant errors. These models are generally conservative
(between 0% and 20% conservative) for bottom struts but are unconservative for diagonals
(between 40% and 60%). This is attributed to the equal shear assumption used in the post-
processing procedure.
• 2D grillage models using the flexural-analogy for the development of equivalent cross-
frame beams show considerable scatter in the predicted cross-frame response. In some cases,
the model predicted design forces above 100% conservative and below 60% unconservative.
The upper tail of the bottom strut box-and-whiskers results is beyond the scale of the vertical
axis, given the large discrepancies observed.
Figure 3-34 expands on Figure 3-33 by breaking up the bottom strut results based on bridge
geometry. For the 2D PEB results, only models utilizing the Timoshenko beam approach are
included for clarity. Note that the skew index and connectivity index established in NCHRP
Research Report 725 (White et al. 2012) and defined in Eqs. 3.1 and 3.2 are used to differentiate
straight and curved bridges with or without support skews.
To explore the effects of bridge geometry, Figure 3-34 compares the results in terms of skew
and connectivity indices. Recall that only bottom strut results are presented in this figure for
clarity. The following generalized observations can be made from that figure with an emphasis
on PEB methods:
• 3D models generally perform well for all cases. Thus, 3D models offer consistently reliable
results in terms of cross-frame forces primarily because the superstructure depth is accurately
represented.
• 2D PEB models are generally conservative in terms of their bottom strut force estimates.
However, scatter in the results is more evident as the bridges become more skewed
and/or more curved. For instance, the percent errors range from –20% unconservative to
20% conservative for straight and normal bridges, whereas the percent error ranges from
–60% conservative to 80% conservative for curved and skewed bridges.
• Although not presented, the results for diagonal members are more variable than what is
shown in Figure 3-34 for bottom struts.
100%
80%
60%
40%
Percent Error
20%
0% 2D
Grid
-20% 3D 2D 3D 3D
-40% PEB 2D
2D 3D
PEB
-60% 2D PEB 2D 2D
2D
-80% Grid Grid PEB
Grid
-100%
Straight & Normal Curved & Normal Straight & Skewed Curved & Skewed
IS ≤ 0.1 IS = 0 IS > 0.1 IS > 0.1
IC ≤ 0.5 IC > 0.5 IC = 0 IC > 0.5
It is evident from the results that 3D models offer increased repeatability and reliability
in terms of cross-frame force effects. However, the biggest setback to simplified 2D analyses
(particularly PEB models) is related to diagonal members and the postprocessing assumptions.
This statement is independent of bridge type, although analysis error becomes more apparent
for complex systems. Consequently, Section 3.4.4 explores a simplified method to enhance
the postprocessing procedures for 2D analysis methods.
100 k
Analysis Output:
756 k-in 258 k-in
4.2 k 4.2 k
2.1 k 2.1 k
60”
2.1 k 2.1 k
8.5 k 2.6 k
12.6 k 4.3 k
120”
4.2 k
4.2 k
Combining Results:
8.5 k
8.5 k
develop than the 2D counterparts, these models offer solutions with improved accuracy,
reliability, and repeatability. The traditional approach of connecting cross-frame members
into a shared node along the web-to-flange juncture is also acceptable.
• 2D PEB models generally produce more accurate results than 2D grillage models because the
concrete deck is explicitly considered. Although there are methods to consider the effective
transverse stiffness of the deck, it is more accurate to represent the deck as a thin shell element
that is rigidly connected to the girder to simulate composite action.
• Common postprocessing practices for 2D methods tend to produce erroneous results for
top strut members and inaccurate results for diagonal members in X-frames, regardless of
bridge type. A simple, conservative approach has been presented, but alternative methods
can and should be explored by designers or programmers.
• If a designer elects to use a 2D modeling approach for these complex cases, it is recommended
to incorporate simple modifications to the model (e.g., utilizing the Timoshenko approach
for equivalent cross-frame beams and considering the transverse stiffness of the concrete for
grillage-type models) and postprocessing (e.g., conservatively applying the full cross-frame
shear force in the design of X-frame diagonals and considering the concrete deck in post-
processing of grillage models similar to Figure 2-17). Still, even after implementing these
improvements to 2D analyses, substantial error is likely in cross-frame force predictions,
especially in bridges with significant horizontal curvature and/or support skews.
• Based on the conclusions above, it is recommended that any bridge satisfying the geometric
limits established in Section 3.2.4 (i.e., bridge geometries for which live load force effects in
cross-frames should be considered) should be evaluated using 3D modeling techniques. In
general terms, this implies that heavily skewed and/or curved systems are best suited to be
analyzed with 3D models.
These findings, which are further summarized in Chapter 4, served as the basis for many of
the proposed modifications to AASHTO LRFD. Refer to Appendix A for the proposed language
and commentary.
section, several sample results are presented that demonstrate the benefits of a composite con-
crete deck on the buckling capacity of the steel girder section, even in the region around interior
supports when shear studs may not be provided and the moments cause compression in the
bottom flange.
Figure 3-36 presents sample eigenvalue buckling results of a girder with an unbraced-length-
to-depth ratio (Lb/d) of 5 and a well-stiffened web with various top flange restraints and moment
gradients. For the three different top flange restraints considered (i.e., no additional restraint,
continuous lateral restraint, and continuous lateral and torsional restraint), the critical buckling
moment, Mcr, was obtained from the FEA analysis and is plotted. To demonstrate the magni-
tude of these buckling loads compared to a limit state governed by cross-sectional yielding,
Mcr is normalized by the moment at first yield of the section, My, where the yield strength
is taken as 50 ksi. Along the horizontal axis, the results are organized by the corresponding
straight-line moment gradient, which is represented as the ratio between applied end moments,
MR/ML (applied moment at the right end of the unbraced beam segment/applied moment
at the left end). For example, MR/ML = 1.0 represents a uniform-moment case in which the
bottom flange is subjected to uniform compressive stresses. Sketches are provided above the
figure to graphically illustrate these moment gradients.
From Figure 3-36, the following observations can be made with regards to the effects of top
flange restraint on girder buckling behavior:
• In general, the critical buckling moment increases as MR/ML decreases due to the benefits of
moment gradient along an unbraced segment in terms of buckling capacity. The intent of
the moment gradient factor, Cb, adopted by many design specifications is to account for
these effects in design.
• In comparing the case with no additional top flange restraint and the case with continuous
lateral restraint, it is evident that adding continuous lateral support increases the buckling
capacity, particularly for cases with reverse-curvature bending (i.e., MR/ML = -1). These results
are consistent with the modified Cb factor expressions in the AISC Commentary Section F1
(2016) that account for continuous lateral restraint.
• When adding continuous lateral and torsional restraint, the critical buckling moments
consistently exceeded the yield moment of the section. For these cases, the buckled shapes
of the beams were consistent with the web distortion mode presented in Figure 2-18. Even
for the uniform-moment condition in which the restrained top flange is purely and uniformly
ML MR ML ML MR
Normalized Buckling Moment, Mcr / My
3.0
Key: Top Flange Restraint
2.5
Continuous Lateral &
2.0 Torsional
1.5
Continuous Lateral
1.0
0.5
None
0.0
–1 0 1
Applied End Moment Ratio, MR/ML
in tension, Mcr exceeded My by approximately 50%. Thus, these unbraced segments would
yield prior to reaching an instability, which in turn mitigates the bracing demands on the
cross-frames in these regions.
To expand on Figure 3-36, Figure 3-37 evaluates two different conditions independently:
an unbraced-length-to-depth ratio of 10 and an unstiffened web. The intent of this figure is
to examine the effects of increased unbraced length and web slenderness on the LTB and web
distortion buckling modes introduced schematically in Figure 2-18. Recall that the cases for
which no top flange restraints are provided (blue line in Figure 3-37) and continuous lateral
restraints are provided (red line) generally correspond to more conventional lateral-torsional
buckled shapes. In contrast, the case simulating composite systems (i.e., continuous lateral
and torsional restraint to the top flange) produced buckled shapes more consistent with web
distortion modes.
From Figure 3-37, the following observations can be made with regards to the effects of
unbraced length and transverse stiffeners on LTB and web distortion buckling:
• As demonstrated by the left figure, increasing the unbraced length impacts the LTB modes
(i.e., girders with no top flange restraint and girders with continuous lateral top flange
restraint) much more significantly than the web distortion modes (i.e., girders with con-
tinuous lateral and torsional top flange restraint). For instance, a 70% reduction in buckling
capacity was consistently observed when the Lb /d ratio was doubled. These results verify the
findings in Helwig and Yura (2015) that state that web distortion buckling is not as sensitive
to the unbraced length.
• As demonstrated by the right figure, eliminating the web stiffeners drastically reduces the
buckling capacity of the web distortion buckling modes. A nearly 70% decrease in Mcr on
average was observed when compared to the finite element solutions for the girders with
well-stiffened webs. For the LTB-related cases, this trend is less pronounced, particularly
for the uniform-moment case where no shear force is present. For the cases with moment
gradient, however, a slight reduction in Mcr is observed due to the effects of web distortion
on the LTB response, which were intrinsically neglected in the development of the tradi-
tional LTB solutions developed by Timoshenko and Gere (1961). Note that the solution
derived by Timoshenko assumes a perfectly straight web (i.e., absent of any distortion) in
its buckled shape. Studies have shown that the interaction between web distortion and LTB
is most pronounced with smaller unbraced lengths, which are often controlled by yielding
over buckling.
-20%
-40%
Continuous Lateral
-60%
None
-80%
-1 0 1 -1 0 1
Applied End Moment Ratio, MR/ML
• Despite the nearly 70% reduction demonstrated in the right figure for the continuous lateral
and torsional restraint condition, it is important to note that these results are presented
in relative terms. In absolute terms, a 70% reduction in M cr for this restraint condition
still produces buckling capacities that exceed the lateral-only or no top flange restraint
conditions.
With these factors in mind, it is evident that instability related to conventional LTB behavior
is generally not critical once the composite deck has cured, and it is able to provide continuous
lateral and torsional restraint to the girder top flange. In these instances, the critical unbraced
segment in negative flexure is generally controlled by yielding of the cross-section, for which
stability bracing demands in the cross-frames are greatly diminished.
While including minimal shear studs in the negative moment region improves the twist
restraint of the girders, the deck still provides considerable restraint even without shear con-
nectors. The restraint comes in the form of “tipping restraint,” in which twist of the girder
shifts the contact point between the deck and the girder out to the edge of the flange. Tipping
restraint is very significant in most practical problems such that conventional LTB in the
composite girder is not generally a critical mode in the finished bridge. As such, the subsequent
sections focus primarily on bracing requirements for the noncomposite system, where stability
is most critical.
In general, moment demands in the critical bracing elements were monitored as a function
of increasing girder moments. That is to say, the out-of-plane girder deformations and cor
responding brace moments resulting from second-order effects on the initial imperfection were
recorded during these incremental analyses. Thus, the figures presented in this section graph
the brace moment, Mb, as a function of the maximum girder moment, M.
As an example, Figure 3-38 illustrates this relationship for two different uniformly distributed
loading cases. Uniform-moment conditions were also evaluated as part of this study. However,
it was observed that uniformly distributed loads that introduce shear on the cross-section
produced larger moment demands on the critical braces. As such, uniform-moment results
are included in Appendix F for reference but are not included in the report.
The graphs on the left represent top flange loading, whereas the graphs on the right represent
mid-height loading. The figure also normalizes girder moments M to Mcr, which represents the
theoretical Timoshenko and Gere (1961) solution for critical elastic buckling moment between
brace points. Additionally, the brace moments, Mb, are normalized to the girder moments, M,
such that a direct comparison to Eq. 2.3 and Eq. 2.4 can be made. In other words, the cor-
responding Mb /M ratio as the applied moment, M, approaches Mcr (i.e., M/Mcr → 1.0) is the
primary focus of the FEA results.
1.0
0.8
0.6
0.0
0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.00 0.02 0.04 0.06 0.08 0.10 0.12
1.2
n = 3, top flange loading n = 3, mid-height loading
1.0
0.8
M / Mcr
0.6
0.0
0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.00 0.01 0.02 0.03 0.04 0.05 0.06
0.8
0.6
0.0
0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.00 0.01 0.02 0.03 0.04 0.05 0.06
Mb / M
Figure 3-38 also represents the results related to different girder cross-sections and bracing
schemes. Girder Cross-sections 1, 2, and 3 are evaluated independently as denoted on the line
graphs. In terms of bracing, note that only an odd number of intermediate braces (i.e., 1, 3,
and 5 intermediate braces) between the end supports is presented, as these cases align a brace
point at the location of maximum girder moment (i.e., at midspan). Cases with an even number
of braces were also analyzed, but those conditions resulted in less severe brace forces.
As noted in Section 2.6.2, it was observed that elastic materials (which were controlled by
stability limit states and not yielding) always resulted in larger buckling-related deformations
and thus larger brace moments compared to brace moments that occurred at the limit state of
yielding in the girder. As such, Figure 3-38 and subsequent figures presented in this section are
based on elastic material properties.
From Figure 3-38, the following observations can be made with regards to brace strength
requirements and uniformly distributed loading:
• For these single-curvature bending cases, the brace moments ranged from 3% to 10% of
the girder moment depending on the position of the load on the cross-section (load-height
effects) and the specific cross-section. As such, the current 15th Edition AISC (2016) brace
strength equation (Eq. 2.4) that recommends 2% of the design moment is unconservative
(i.e., 3% versus 2% is a 100% error).
• Top flange loading is generally more critical than mid-height loading, and girders with fewer
intermediate braces result in larger normalized brace moments. In other words, the relative
magnitude of the brace forces at the location of the maximum girder moment decreases with
an increased number of intermediate braces.
• For most analyses, slender sections (i.e., smaller bf /d ratios) resulted in larger brace forces;
however, in some cases the stockiest section (Cross-section 3) resulted in the largest brace
force as a percentage of the applied moment. The much wider flange for Cross-section 3 leads
to a larger critical brace moment compared to Cross-sections 1 and 2, such that the applied
moment M is relatively large.
• Some of the analyses did not reach convergence beyond 90%–98% of Mcr due to excessive
girder deflections. However, the original study focused on the current AISC stiffness recom-
mendation. As noted later, the recommended stiffness for the AASHTO LRFD provisions is
higher, which will result in better control (and analysis convergence) of brace forces.
As Figure 3-38 presented sample results from the parametric study, Table 3-12 summarizes
the torsional bracing strength requirements for the entire data set. As noted above, different
cross-sections, numbers of intermediate braces, and loading conditions were evaluated. The pre-
vious figure demonstrated the curves tend to flatten as M/Mcr approaches 1.0, and convergence
was not reached above 90%–98% of Mcr in some instances. In Table 3-12, the brace moments
are presented for either M/Mcr = 1 or the largest load that convergence was achieved, which was
typically capped at M/Mcr = 0.9.
Although some of the brace moments tabulated are approximately 2% at M/Mcr = 0.9 (for
analyses that did not converge), it should be emphasized that brace moments tend to increase
dramatically as the applied moment increases from M/Mcr = 0.9 to M/Mcr = 1. The brace moment
has been shown to grow between 50% and 100% beyond an applied moment of 90% of Mcr.
As such, these cases, which may not appear critical, can be rather significant.
Another point to note is that the results documented in this section were from a com-
panion study to NCHRP Project 12-113 that is documented in Liu and Helwig (2020). A later
companion study that is discussed in the next subsection modified the stiffness requirements.
Loading n 1 3 5
Conditions bf / d 1/6 1/4 1/3 1/6 1/4 1/3 1/6 1/4 1/3
Uniform Mb / M 6.9% 3.9% 3.7% 2.7% 2.2% 2.2% 2.0% 2.0% 3.0%
Moment M / Mcr 100% 100% 100% 90% 90% 90% 90% 90% 90%
Top Flange Mb / M 10.4% 7.3% 7.7% 2.5% 2.0% 2.0% 4.0% 3.4% 2.0%
Loading M / Mcr 100% 100% 90% 90% 90% 90% 100% 100% 90%
Mid-Height Mb / M 7.3% 4.0% 3.3% 2.5% 2.0% 2.0% 4.0% 3.4% 2.0%
Loading M / Mcr 100% 100% 100% 90% 90% 90% 100% 100% 90%
The larger stiffness requirements produce much better agreement of the strength behavior
with regards to Eq. 2.3.
In general, the brace moments reported are consistently much larger than the 0.02M
requirement given by the current AISC expression (Eq. 2.4). In many cases, the FEA results
were more than 100% larger. Therefore, it is recommended that designers use the brace strength
equation from the 14th Edition Specification with the modifications outlined in Section 3.5.3.
Even with the potential implementation of Eq. 2.3 in AASHTO LRFD, it is evident from
Table 3-12 that the required brace moments can still exceed the brace moment given by Eq. 2.3
(i.e., a minimum brace moment of 2.6% of Mcr). However, there are mitigating factors that
lessen the demands reported in the table. As previously noted, stability brace moments are
extremely sensitive to the shape and distribution of the initial imperfection in the girders.
In most situations, though, the actual imperfection will not match the critical shape considered
in this study, resulting in smaller brace forces. Also, in many situations, the brace sizes often
utilized in design will exceed the stiffness required by Eq. 2.1, which will typically result in
smaller stability-induced forces. Lastly, a companion study related to bracing stiffness require-
ments that is documented in Section 3.5.3 proposes an increase of the stiffness requirements
such that better agreement is achieved with Eq. 2.3. With that in mind, Eq. 2.3, as modified in
the next subsection, is deemed suitable for implementation into AASHTO LRFD.
Given that Table 3-12 focuses entirely on single-curvature bending cases, it is prudent to
verify these conclusions for reverse-curvature bending. To demonstrate bracing demands in
beams subjected to significant moment gradient, sample results are presented in Figure 3-39.
1.00
0.80
Key: Key:
0.60
M M
0.40
0.20
Imperfect TF Imperfect TF
0.00
M / Mcr
1.20
1.00
0.80
Key: Key:
0.60
0.40
0.20
Imperfect TF Imperfect BF
0.00
0.00 0.01 0.02 0.03 0.04 0.05 0.00 0.01 0.02 0.03 0.04 0.05
Mb / M
Beams with Cross-section 1 properties were subjected to the reverse-curvature moment gradient
outlined in Section 2.6. Five intermediate braces were incorporated to maximize the negative
moment magnitude at the first intermediate cross-frame line. For the sake of comparison,
Figure 3-39 presents the results for four different loading and critical imperfection scenarios:
• Uniform moment with the critical imperfection along the top flange at midspan;
• Single-curvature, uniformly distributed mid-height loading with the critical imperfection
along the top flange at midspan;
• Reverse-curvature, uniformly distributed mid-height loading with the critical imperfection
along the top flange at midspan; and
• Reverse-curvature, uniformly distributed mid-height loading with the critical imperfection
along the bottom flange at the first intermediate cross-frame line.
Note that in the third and fourth scenarios listed, the loading conditions were identical.
Only the location of the critical imperfection was varied to study its impact on the results.
In Figure 3-38, only the midspan cross-frame results were provided, as bracing demands
were maximized at this location for single-curvature bending (due to girder moment magni-
tudes and girder displacements). For reverse-curvature bending, however, that behavior is
less clear. As such, Figure 3-39 presents the bracing force demands on three different cross-
frames in the five-brace scheme. It should be noted that the results are not symmetric (e.g.,
cross-frame lines 1 and 5 produce different results) due to the asymmetric nature of the
initial imperfection. Thus, only the results related to the more critical half of the beam are
presented.
Figure 3-39 is otherwise constructed similarly to Figure 3-38 with one key exception. Along
the horizontal and vertical axes, the maximum applied moment, M, is taken as the maximum
negative moment at the support for consistency. Mcr is then based on the end segment, which
was shown to buckle before the segments in the positive moment region by eigenvalue analysis.
This is clearly denoted at the top of the figure. Sketches are also provided for each loading and
imperfection condition, for reference.
The following observations can be made from the results presented in Figure 3-39:
• In the uniform-moment case, it is evident that the midspan brace experiences the largest
bracing demands. This validates the assumption used throughout the development of the
single-curvature cases (e.g., Figure 3-38). Additionally, the midspan brace moment approaches
2% of the maximum applied moment at M/Mcr = 0.9 (i.e., analysis did not converge at 100%
of Mcr). This result is consistent with Table 3-12.
• In the single-curvature uniformly distributed load case, it is also evident that the midspan
brace experiences the largest bracing demands. Note that the critical values are also consistent
with what is reported in Table 3-12.
• For the reverse-curvature bending cases, the critical cross-frame coincides with the location
of the critical imperfection, as expected. In the bottom left plot (imperfection along the top
flange at midspan), the bracing moment approaches 1% of the maximum applied moment
(taken as the negative moment value at the support) as Mcr is reached. The other two braces
that are presented pick up nearly no bracing forces, even at the elevated load magnitudes.
In the bottom right plot (imperfection along the bottom flange at the first brace line), the
critical bracing moment approaches 1.5% of the maximum applied moment. These cases
still produce less force demands than the comparable single-curvature conditions. Note that
the beams at the supports in these analyses were assumed to have zero initial twist. This is
consistent with actual practice where the most precise control on geometry will generally be
achieved at the support region.
• For these spot check conditions, it was generally observed that bracing demands are more
critical for single-curvature bending conditions with less moment gradient (i.e., a smaller
Cb factor) than reverse-curvature conditions with significant moment gradient. This is
reflected in Eq. 2.3, in which the Cb factor of the unbraced segment of interest is in the
denominator.
Based on these results, it can be concluded that Eq. 2.3 is valid and generally conservative
for cases with reverse-curvature bending in noncomposite systems. That is to say, the bracing
strength requirements are not necessarily a function of positive or negative moment. For cases
with variable unbraced segment lengths and moment gradient factors, the bracing moment is to
be based on the brace point and unbraced segment that maximizes the Mr/CbLb component of
the equation regardless of which flange is in compression. In doing so, this ensures the critical
condition is covered. All cross-frame braces in the span would be subsequently designed for
this worst-case condition. These three variables are segment-specific, whereas the other vari-
ables and constants are consistent for the given span (e.g., the span length L is taken as the
full span length under consideration). One note to the above discussion is the often overly
conservative expression for Cb that is currently in AASHTO LRFD as of this writing. There is
significant work currently underway related to the moment gradient factor in the AASHTO
Specification that is likely to result in improved accuracy over the current Cb equation given
in the specification.
Note that Lb in Eq. 2.3 need not be taken as less than the maximum unbraced length permitted
for the girder segment based upon the required flexural strength, Mr. This criterion accounts for
scenarios in which a closer spacing of the braces is provided compared to the spacing required
to develop the required strength. In essence, this provision lessens the demands on cross-frames
for girders that would not buckle under the specified cross-frame layout (i.e., the girder would
partially or fully yield prior to buckling). In these instances, out-of-plane deformations would be
limited such that cross-frame bracing demands are relatively small. Hence, using the increased
Lb term in the denominator reflects that behavior.
As outlined in Section 3.5.1, bracing demands in the composite condition were deemed
less critical than in the noncomposite condition, given the lateral and torsional restraint pro-
vided by the concrete deck as well as the bridge bearings in the negative moment region. Thus,
the procedure introduced above is only necessary for the construction condition, where gird-
ers are far more susceptible to LTB. Note that, although cross-frame bracing requirements are
not critical for the composite condition, web stiffeners must be adequately proportioned to
prevent web distortion buckling effects in the negative moment regions per Yura (2001).
1.2
1.0
0.8
M / Mcr
0.6
βb = β2i
0.4 βb = 2.5 βi
βb = 3 βi
0.2 βb = 4 βi
0.0
1.0 1.5 2.0 2.5 3.0
θ / θo
For clarity, these results are not included herein. However, the following items provide a
cursory overview of the major observations:
• The relative girder twist induced at critical buckling loads decreases as the brace stiffness
increases. In all cases, assigning stiffer cross-frame elements mitigates girder deformations.
• In general, adding intermediate braces (regardless of the stiffness) decreases the relative
cross-section twists.
• Increasing the girder spacing, particularly in twin-girder systems, not only improves the
system buckling capacity but also decreases the stiffness demands for the cross-frame braces.
Similarly, systems with increased girder redundancy generally resulted in an increase in the
in-plane stiffness of the bridge system, which effectively led to smaller girder deformations
at critical buckling loads.
• Top flange loading and the mid-height loading resulted in similar deformation behavior.
The effect of the transverse loading locations on the stiffness requirement for torsional
bracing was marginal.
• In general, assigning three times the ideal stiffness tends to result in girder cross-sectional
twists closer to two times qo than assigning twice the ideal stiffness.
Table 3-13 summarizes the torsional bracing stiffness requirements based on the results of
this parametric study. The final girder twist (as a ratio of the initial imperfection) is tabulated
for various cross-frame stiffness values, loading conditions, and girder cross-sections. Note
that these results correspond to twin girders only, because this system has been shown to
be most critical. For each q/qo ratio reported, the corresponding M/Mcr value is also shown. In
most cases, the applied moment reached the critical buckling moment (i.e., M/Mcr = 100%).
However, there were a few instances in which the critical buckling moment was not achieved;
in those cases, the girder twist at the maximum girder moment recorded is presented instead.
Notes:
a
Given the uniformly distributed loading (wL2/8 type loading), the moment gradient factor, Cb , is a function of the
bracing scheme.
In general, the table shows that providing twice the ideal stiffness (2b i) typically results in
load-induced girder twists exceeding the assumed twist in Eq. 2.3 (i.e., q/qo > 2). This implies
that the current assumption built into AISC Specifications (i.e., Eqs. 2.1 and 2.3) is invalid.
However, the cross-sectional twist can be reduced to magnitudes more consistent with the
assumed values by requiring braces to possess three times the ideal stiffness (3b i). Therefore,
for implementation into AASHTO LRFD, it is advised that Eq. 2.1 be modified to inherently
consider three times the ideal stiffness as a means to limit girder deformations. This is presented
mathematically by the following expression:
3.6 LM r2
b T,req = 3.3
f nEIyeff Cb2
A cross-frame must then be designed and detailed such that its total stiffness, as computed
by Eq. 2.2, exceeds the required stiffness from Eq. 3.3. Note that the constant in the equation
has increased 50% (i.e., from 2.4 to 3.6) to account for these recommended changes. This
modification not only impacts the brace stiffness requirements, but also the brace strength
requirements. Recall from Section 2.6 that the brace strength equations are effectively the
required brace stiffness multiplied by the assumed initial imperfection. With that in mind, it is
also recommended to modify Eq. 2.3 to account for the 3b i assumption outlined above. This
is reflected with the revised expression below:
0.036 M r L
M br = b Tq o = 3.4
nCb Lb
To determine the axial-force demands in the cross-frame members, the bracing moment
computed by the expression above can be resolved as a force couple to the top and bottom
nodes. This procedure is demonstrated in Appendices B and C. It is important to note that, as
an alternative, designers are also permitted to perform a large-displacement analysis to estimate
stability bracing force demands, provided that initial imperfections are considered (Helwig and
Yura 2015). For curved systems though, the effect of initial imperfections has been shown to
be less impactful given the curved geometry of the girders.
construction. Given that cross-frames are generally not required as stability braces in the
finished bridge, it is only necessary to evaluate stability-related force effects with other con-
struction force effects, including dead loads (e.g., wet concrete, formwork, fit-up forces),
construction loads (e.g., overhang construction, screed), and wind loading acting on the
bridge fascia. Cross-frame force effects related to the finished structure, such as live loads
and wearing surface dead loads, are then to be considered independently in their own set of
load combinations as currently established in AASHTO LRFD.
Because the stability-related force effects are critical during construction, at which point the
bridge presumably exhibits fully elastic behavior, the principle of superposition is applicable for
these specific load combinations. That is, the bracing strength forces computed from Eq. 3.4 or
a large-displacement analysis can be directly added to the force effects from dead, construction,
and wind loads derived from an elastic analysis.
Before introducing each load case, it is important to note that construction sequencing must
be considered when applying principles of superposition. Depending on the responsible party
at the specific construction condition, either the engineer-of-record (EOR) or the erector or
contractor’s engineer may need to evaluate stability bracing requirements at different stages of
construction. The behavior is often dependent on the state of cross-frame installation and/or
external bracing systems (i.e., guy cables, shoring, or holding cranes) at the specific erection
or construction stage. Therefore, the additive nature of these force effects is contingent on
consistent and concurrent bridge conditions. For example, stability bracing force effects related
to an intermediate phase of girder erection should not be combined with dead load force effects
related to wet concrete, as these two activities occur at different stages of construction. It is also
imperative that the EOR or the erector or contractor’s engineer maintain the proper sign when
summing the force effects of each load case (i.e., tension and compression).
With that in mind, dead loads consider both (i) internal force effects due to gravity loads and
(ii) “fit-up” forces due to externally applied loads by the erector to assemble the structural steel
during erection. The fit-up forces are largely dependent on which fit condition is selected by the
engineer and constructed by the steel erector (i.e., no-load fit, steel dead load fit, or total dead
load fit). Dead load force effects in cross-frames due to gravity loads on noncomposite systems
are commonly estimated using 2D or 3D analysis models. Many commercial design software
programs have also been designed and programmed to perform staged construction analysis
to handle the various states of stress throughout the construction process. This topic is covered
extensively in White et al. (2012). Guidance for estimating fit-up forces (when necessary),
on the other hand, is provided in AASHTO LRFD Article 6.7.2, which is largely based on the
work of White et al. (2015).
Construction-related force effects in cross-frames, such as those induced by overhang deck
construction, are often estimated with hand solutions as documented in Appendices B and C.
Lastly, wind loads on a steel superstructure during construction can be estimated with refined
analyses or the simplified procedures provided in AASHTO LRFD Article 4.6.2.7.
With the individual load cases outlined, the final step is related to load factors. Because
stability bracing force effects are a function of the factored girder moment (i.e., Mr in Eq. 3.4),
no additional load factors are needed when combining these effects with the other construction-
related force effects. If large-displacement analysis is used to estimate the bracing force demands,
then the designer must give special consideration to how load factors are applied in analysis and
design checks. The remainder of the load cases are subsequently factored based on the guidance
and tables provided in AASHTO LRFD Articles 3.4.1 and 3.4.2.
This entire process, as summarized above, is illustrated through two design examples in
Appendices B and C.
CHAPTER 4
Up until the 1990s, specifications included a spacing limit that effectively standardized the
design of cross-frames. AASHTO LRFD Specifications have since eliminated a spacing limit
for straight bridges and instead require the design to be based on a rational analysis. Although
many bridge owners still utilize typical details for cross-frames, software tools have made it
possible for designers to develop improved predictions of cross-frame behavior and design
forces. However, to adequately make use of these software tools, there has become an increas-
ingly important need to modernize design guidance for cross-frames.
Although considerable research over the past several decades has improved cross-frame
design in the areas of stability bracing, load-induced fatigue resistance, erection fit-up, and
analysis for the noncomposite construction condition, the design industry has generally
lacked quantitatively based guidance on several other topics related to fatigue loading criteria,
analysis techniques, and stability bracing. In response, NCHRP Project 12-113 identified
several gaps in knowledge, including those generalized concepts as follows: (i) fatigue loading,
(ii) analysis, and (iii) stability bracing requirements of cross-frame systems in steel I-girder
bridges. These gaps in knowledge served as the primary objectives of this project, previously
identified as Objectives (a) through (e) in Chapter 1. To further clarify the intent of the
research, those five objectives were subsequently posed as a series of questions that have been
systematically addressed throughout the report.
This chapter summarizes the major findings from the experimental and analytical studies
performed in relation to those questions. Section 4.1 outlines the key conclusions of NCHRP
Project 12-113. Section 4.2 serves as the link between the major findings of the studies and the
proposed specification and commentary language provided in Appendix A. In other words,
the conclusions outlined in Section 4.1 are synthesized in the context of AASHTO LRFD, and
Section 4.2 explains the justification and rationale for the proposed language presented in the
appendix. Lastly, although this project significantly advances cross-frame analysis and design
practices, there are a number of related studies that merit additional investigation. Section
4.3 identifies research topics beyond the scope of NCHRP Project 12-113 that could further
improve the understanding of cross-frame behavior in steel I-girder systems.
134
synthesized from the “Major Outcomes” sections in Chapter 3, are then provided to address
each question individually. Proposed changes to the corresponding AASHTO LRFD Articles are
subsequently presented in Section 4.2 based on the commentary herein.
[Fatigue loading – Articles 3.6.1.4, 6.6.1.2.1] Is the current fatigue load model in terms of truck
position (i.e., single design truck passages positioned in various transverse lane positions) appro-
priate for cross-frame analysis and design? Or do multiple presence effects need to be considered?
• The WIM study (Section 3.2) confirmed that the “dual truck” event, which maximizes force
reversal in critical cross-frames and was initially considered in the 7th Edition AASHTO
Specifications, is a rare occurrence. As such, the current 9th Edition load criteria (i.e., a single
design truck positioned in all longitudinal and transverse positions with the truck confined
to one critical transverse position per each longitudinal position) is more appropriate.
• In evaluating the AASHTO fatigue criteria, it was concluded that additional economies are
possible for cross-frame design. This is largely attributed to (i) the sensitivity of cross-frame
response to transverse load position and (ii) the current fatigue load factors. The fatigue load
factors introduced in the 8th Edition AASHTO LRFD, which were calibrated based on girder
response, do not accurately represent the response of cross-frames. As such, one or both of
these items above could be addressed to maximize the efficiency of the design criteria.
• In terms of item (i) above, current design criteria require all possible lane positions be
considered, regardless of design lanes or actual lane striping. A more realistic scenario would
be to apply fatigue loading only in anticipated drive lanes. Item (ii) and cross-frame-specific
load factors are addressed with the next question.
[Fatigue loading – Article 3.4.1] Are the current AASHTO LRFD Fatigue I and II load factors,
which were calibrated for girder force effects and recent WIM data, appropriate for cross-frame
analysis and design?
• The WIM study summarized in this report indicates that the existing load factors of 1.75 for
Fatigue I and 0.8 for Fatigue II are conservative for the fatigue limit state design of cross-
frames by at least 35%, specifically when the AASHTO fatigue design truck is placed in every
possible transverse position during analysis and design. To improve economy in the AASHTO
LRFD guidance, this study demonstrated that either the fatigue design truck could be posi-
tioned in striped lanes (rather than permitting design trucks in unrealistic lane positions
such as adjacent to the bridge barriers), or cross-frames could be designed to adjusted load
factors. Based on feedback from the NCHRP Project 12-113 panel, it was deemed more
desirable to propose adjusted cross-frame-specific load factors.
• As an alternative to introducing two new load factors for the Fatigue I and II limit states for
cross-frames, it is possible to apply a single adjustment factor to the existing load factors,
similar to the current use of an adjustment factor in the fatigue design of orthotropic decks in
AASHTO LRFD Article 3.4.4.
• Using a single adjustment factor of 0.65 applied to the Fatigue I load factor of 1.75 (i.e.,
a resultant load factor for cross-frames of 1.14) and the Fatigue II load factor of 0.8 (i.e.,
a resultant load factor of 0.52), this study demonstrated that the resulting reliability indices
calculated via Monte Carlo simulation satisfy a minimum assumed target reliability of unity,
as recommended in SHRP 2 R19B (Modjeski and Masters 2015).
[Fatigue loading – Table 6.6.1.2.5-2] Is the “n” value (i.e., number of cycles per truck passage)
currently assumed for the generic “transverse member” designation applicable for cross-frames?
• The reduced fatigue load factors specific to cross-frames proposed in the preceding question
are calibrated assuming the number of cycles per design truck passage, n, is equal to unity,
regardless of cross-frame spacing. This is consistent with the analytical and experimental
analyses conducted in this study.
[Fatigue loading – Article 6.7.4.2] What is the influence of composite bridge geometry (e.g., support
skew, horizontal curvature) and cross-frame layout (e.g., cross-frame spacing, staggered layout)
on the load-induced fatigue behavior of cross-frame systems (including governing force effects,
critical lane loading, and critical members under AASHTO fatigue loading criteria)?
• In general, cross-frame response is sensitive to transverse truck placement. The sensitivity
of cross-frame response to longitudinal truck placement depends heavily on support skew
and horizontal curvature. Similarly, critical lane position also depends on a variety of
parameters. For instance, truck passages closer to the outer edges of the deck tend to
maximize cross-frame forces in skewed and curved bridges, but not in straight bridges with
normal supports.
• In terms of implementation into AASHTO LRFD, the results of this study indicate that the
current specification language in Article 3.6.1.4.3a, “a single design truck shall be positioned
transversely and longitudinally to maximize stress range at the detail under consideration,”
is the best way to ensure the critical load position is considered by the designer. This require-
ment is best executed with an influence-surface analysis. Thus, performing an influence-
surface analysis for the entire bridge deck, regardless of design or striped lanes, in conjunction
with the proposed cross-frame-specific fatigue load factors more accurately represents the
response of cross-frames to live load traffic.
• The governing cross-frame member in a given bridge depends on many variables, including
girder spacing, support skew, and striped lane positions. Bottom struts tend to see the
most substantial load-induced force effects in composite systems, especially those near
the supports of heavily skewed bridges. However, to ensure the critical force effects are
considered in design, an influence-surface analysis is necessary (i.e., to examine all possible
truck positions and cross-frame force effects).
• In general, heavily skewed and/or heavily curved bridges, particularly those with contiguous
lines of braces, generally produce the most significant force effects in cross-frame members.
To avoid an iterative “chase-your-tail” design solution, whereby the designer uses increasingly
large trial sizes for cross-frame members which in turn attract larger cross-frame forces, other
key bridge parameters can be adjusted to mitigate force effects. For instance, using a dis
continuous, staggered cross-frame layout is a practical and economical solution for skewed
systems. Increasing the deck thickness or decreasing girder spacing can also reduce cross-
frame forces. However, these measures have significant impacts on the rest of the design and
may therefore be less economical.
[Fatigue loading – Articles 6.7.4.1, 6.7.4.2] Is it necessary to perform a refined analysis, either
simplified 2D or 3D methods, for straight and non-skewed bridges? In other words, are the mini-
mum design requirements outlined in Article 6.7.4.1 appropriate?
• When comparing the response of cross-frames to measured WIM records, it was observed
that load-induced fatigue forces are most significant in skewed and/or curved bridges, for
which Ic > 1, Is > 0.15, or Ic > 0.5 and Is > 0.1. For bridges that do not fall into one of these
categories, a refined 2D or 3D analysis is likely not warranted. In these instances, other
design considerations such as stability bracing during construction often govern the design.
Therefore, the minimum design requirements with respect to straight bridges with normal
supports in Article 6.7.4.1 (in addition to stability bracing requirements) are appropriate.
[Analysis – Article 4.6.3.3.4] Is the current established R-factor (0.65AE), which was based on
analytical and experimental studies of a noncomposite system, appropriate for cross-frames in the
composite condition? Are there alternative 3D modeling approaches for cross-frames?
• 3D commercial design software programs commonly represent cross-frames as pin-ended
truss elements, which inherently misrepresent the in-plane and out-of-plane stiffness of
the system. The stiffness modification approach accounts for these limitations by adjusting
the axial stiffness of the cross-frame members in the analysis model. The proposed eccentric-
beam approach considers flexural deformations by explicitly representing the eccentric load
path through the cross-frame connections.
• Cross-frame deformational response in composite systems (i.e., in-service) is different than
the response in noncomposite systems (i.e., during construction). As such, the established
R-factor in AASHTO LRFD that was developed specifically for the noncomposite condition
should be reconsidered for composite conditions and live load force effects.
• Independent stiffness modification factors are proposed for the construction stages (Rcon = 0.65)
and for in-service conditions (Rser = 0.75). Note that the appropriate R-factor to be assigned
for the in-service condition is largely a function of bridge geometry, cross-frame details,
and uncertain loading conditions. Although scatter was observed in the results, the proposed
R-factor is a simple solution to a complex problem that produces reasonably accurate
approximations of the true cross-frame stiffness. Considering that many designers often
prefer simple alternatives over sophisticated refined analyses, it serves an important role
in AASHTO LRFD guidance moving forward.
• Analytical results demonstrated that the proposed eccentric-beam analysis method improves
the repeatability and reliability of 3D models by eliminating several sources of uncertainty
associated with the R-factor approach. Although the R-factor approach serves a vital purpose
in practice due to its ease of use, the proposed eccentric-beam method offers an alternative
to engineers seeking a more refined solution.
[Analysis – Article 4.6.3.3.2] What are the limitations of simplified 2D analysis techniques
commonly used by popular commercial bridge design software programs in terms of predicting
cross-frame force effects in composite systems? Are there methods available to improve these
simplified techniques?
• In general terms, designers must be cognizant of the potential trade-offs between sophis-
ticated and simplified analyses. A central theme explored throughout this report is related
to the balance between increased computational complexity and improved reliability versus
simplified modeling and improved ease of use. The results presented in this report clearly
highlight this trade-off. The conclusions below are intended to provide practicing engineers
the technical and quantitative background to help them make informed decisions, while not
necessarily advocating one method or the other for a specific project need.
• Although 3D models are more time-consuming to develop than 2D counterparts, these
models offer solutions with improved accuracy, reliability, and repeatability. 2D analysis
methods have traditionally been shown to produce reasonable approximations of girder
force effects, but this study demonstrated the significant limitations of 2D analyses asso
ciated with cross-frame force effects due to truck loads.
• The traditional approach of connecting cross-frame members into a shared node along
the web-to-flange juncture in 3D models is also acceptable, as long as the true stiffness is
approximated with R-factors or the eccentric-beam approach.
• 2D PEB models generally produce more accurate cross-frame results than 2D grillage
models because the concrete deck is explicitly considered. Although there are methods to
consider the effective transverse stiffness of the deck, it is more accurate to represent the
deck as a thin shell element that is rigidly connected to the girder to simulate composite
action.
• Common postprocessing practices for 2D methods tend to produce erroneous results for
top strut members and potentially inaccurate results for diagonal members in X-frames,
regardless of bridge type. If a designer elects to use a 2D modeling approach for these
complex cases, it is recommended to incorporate simple modifications to the model
(e.g., utilizing the Timoshenko approach for equivalent cross-frame beams) and to the
postprocessing (e.g., considering various shear force distributions in the design of X-frame
diagonals or considering the contributions of the concrete deck when evaluating grillage
model results similar to Figure 2-17). Still, even after implementing these improvements to
2D analyses, substantial error is likely in cross-frame force predictions, especially in bridges
with significant horizontal curvature and/or support skews.
• Based on the conclusions above, it is recommended that cross-frame force effects—
particularly those related to live loads—for any bridge satisfying the skew and connectivity
index limits established above should be evaluated using 3D modeling techniques. In general
terms, this implies that obtaining live load cross-frame forces for heavily skewed and/or
curved systems is best suited for 3D analysis. Although the focus of this study is related to the
fatigue limit state, this statement is also directly applicable to the strength limit state design
(and associated analysis) of cross-frames.
[Stability bracing – Article N/A] Can the AISC design guidelines for stability bracing be incorporated
into AASHTO LRFD? Are special requirements needed for negative moment regions of continuous
systems?
• Given that a composite deck provides continuous restraint to the top flange and substantial
restraint to the bottom flange, cross-frame bracing requirements are only necessary to be
evaluated for construction conditions.
• Through FEA parametric studies, it was concluded that providing three times the ideal
stiffness for cross-frames better limits girder deformations and cross-frame forces at critical
buckling loads. As such, a proposed modification to the current AISC expression was
provided in Eq. 3.3 for implementation into AASHTO LRFD.
• The general form of the torsional bracing strength equations from the 14th Edition of AISC
(2010) is more appropriate than the current 15th Edition (2016) for steel bridge applications.
Based on the revisions to the torsional bracing stiffness requirements above, Eq. 3.4 was
proposed for implementation into AASHTO LRFD.
• For cases with reverse-curvature bending, it is recommended to conservatively base the
required bracing moment on the unbraced segment that maximizes Mr /CbLb in Eq. 3.4. This
ensures that, for each span (with a corresponding length, L, and number of intermediate
braces, n), the critical bracing moment is considered.
[Stability bracing – Article N/A] How are these stability bracing requirements combined with other
load conditions such as wind?
• Bracing strength demands, as computed by Eq. 3.4 or from a large-displacement analysis,
can be combined with other construction-related force effects via linear superposition. These
bracing forces must be evaluated at all stages of steel erection and deck construction. It
is also important that the designer evaluates these load cases under the same concurrent
construction conditions. For example, stability bracing force effects related to an intermediate
phase of girder erection should not be combined with dead load force effects related to wet
concrete, as these two activities occur at different stages of construction. Because stability
bracing force effects are a function of the factored girder moment (i.e., Mr in Eq. 3.4), no
additional load factors are needed when combining these effects with the other construction-
related force effects.
In terms of organization, each specific AASHTO LRFD Article is listed in italicized font herein
followed by bulleted items that expound on the proposed modifications.
Articles 3.4.1 and 3.4.5 (proposed)
• As noted in Section 4.1, cross-frame-specific load factors that more accurately represent the
live load response of cross-frames to the U.S. truck spectra have been proposed. Rather
than introduce two new Fatigue I and II load factors, an additional adjustment factor (0.65)
to be applied to the existing load factors is proposed, similar to the approach taken for the
fatigue design of orthotropic decks in AASHTO LRFD Section 3.4.4. As such, a new article is
proposed (Article 3.4.5) to introduce this additional adjustment factor.
Article 3.6.1.4
• As noted in Section 4.1, conservatism in the current AASHTO fatigue loading model has been
eliminated by proposing reduced cross-frame-specific fatigue load factors as opposed to
limiting design trucks to a specified lane. As such, the findings of this study concur with and
validate the current commentary language in this article, which addresses the uncertainty
of future lane striping changes and deck widening. Thus, no additional specification or
commentary language is recommended.
Article 4.6.1.2.4
• Article 4.6.1.2.4 currently specifies that cross-frames and diaphragms in straight and slightly
curved I-girder bridges shall be designed, at a minimum, for wind loads and slenderness
requirements. Based on the findings of the Stability Study, it is recommended to include
stability bracing as an additional minimum design requirement for these I-girder bridges,
especially during deck construction. As such, a reference to proposed Article 6.7.4.2.2 is
added to the specification language in this article.
Article 4.6.3.3.2
• Based on the findings of this study, the commentary language in Article C4.6.3.3.2 with regards
to equivalent torsional stiffness to account for warping should be maintained. It was observed
that neglecting warping-torsional rigidity in composite bridges is much less impactful than
in noncomposite bridges as documented by White et al. (2012). However, many designers
utilize commercial design software packages that perform staged construction. Thus, it is
beneficial to incorporate warping rigidity through all stages of construction and service, even
if it does not significantly influence the cross-frame response to live loads.
• As noted in Section 4.1, 2D PEB and grillage model approaches may lead to more significant
errors in the prediction of cross-frame forces regardless of bridge geometry. This was largely
attributed to the postprocessing procedures commonly implemented that convert equiva-
lent beam shear and moments into cross-frame member forces. Thus, commentary language
has been added to help designers make informed decisions regarding analysis for their
project needs.
Article 4.6.3.3.4
• Article 4.6.3.3.4 has been divided into three separate sub-articles to clarify the differences
between 2D and 3D modeling techniques of cross-frames. Proposed Article 4.6.3.3.4a
focuses on current 9th Edition specification and commentary language as well as proposed
language with respect to 2D analysis. Proposed Article 4.6.3.3.4b outlines 3D analysis
methods of cross-frames, including the traditional truss-element modeling approach and the
proposed eccentric-beam approach in the commentary. Lastly, proposed Article 4.6.3.3.4c
discusses the use of stiffness modification factors to be considered in both 2D and 3D models
of cross-frames.
connectivity index limits established previously (Ic > 1, Is > 0.15, or Ic > 0.5 and Is > 0.1)
are introduced in the commentary language to expound on the minimum design consid-
erations in this article. These limits also complement the general approach and language
in Article 4.6.3.3.2.
• In general, this proposed language helps clarify the design considerations and what consti-
tutes a “rational analysis.”
Article 6.7.4.2
• Article 6.7.4.2 is divided into two independent subheadings based on the addition of
stability bracing requirements for cross-frames. The specification and commentary language
in the original Article 6.7.4.2 are now contained in proposed Article 6.7.4.2.1. Stability bracing
requirements are provided in proposed Article 6.7.4.2.2.
Article 6.7.4.2.1 (proposed)
• Similar to the discussion for Article 4.6.1.2.4, the minimum design requirements for cross-
frames in I-girder bridges should include the applicable stability bracing checks at all critical
stages of construction. Specification language is added accordingly.
• This article currently highlights good design and detailing practices of cross-frames, particu-
larly in skewed and/or curved systems. Given that contiguous versus discontinuous cross-
frame lines are already discussed in the commentary with respect to noncomposite systems,
the major findings related to the Fatigue Loading Study are best suited here.
• As documented in Section 4.1, cross-frame force effects are highly variable and depend on
a number of parameters. Ultimately, the two most important aspects of the findings are
that: (i) bottom struts and diagonal members near supports tend to govern fatigue design in
skewed and/or curved bridges and (ii) there are a number of ways that designers can mitigate
design forces in cross-frame elements other than increasing the cross-sectional area. The
most practical and economical way to mitigate load-induced cross-frame forces in skewed
systems is to arrange the braces in a discontinuous, staggered layout. With that in mind,
proposed commentary language is provided to emphasize these findings.
Article 6.7.4.2.2 (proposed)
• On the specification side, Eqs. 3.3 and 3.4 are included, which present the required stiff-
ness and strength of cross-frames or diaphragms serving as torsional braces, respectively.
Additionally, it is important to include guidance on how the total stiffness of a cross-frame
system is calculated (i.e., Eq. 2.2). As such, guidance from Yura (2001) and specifically the
equations to estimate the stiffness of the torsional brace considering connection flexibility,
in-plane girder effects, and cross-sectional distortion effects (also documented in the AISC
Specifications) are provided for reference. The geometric limits, at which in-plane girder
stiffness and cross-section distortion effects become negligible, are also established.
• On the commentary side, much of the information included in the AISC Specifications is
adapted given that it directly applies to steel bridge applications. For instance, the basis of the
original strength and stiffness requirements (Taylor and Ojalvo 1966) is outlined, as is the
critical shape imperfection (Wang and Helwig 2005). Furthermore, expressions that consider
the effect of skewed cross-frames relative to a line normal to the longitudinal axis of the bridge
girders is provided based on work conducted by Wang and Helwig (2008).
• Given the bracing strength requirements specified in proposed Article 6.7.4.2.2 herein, it is
important to provide designers with guidance on how to utilize these forces in the context
of strength limit state design for all stages of construction. As noted in Section 4.1, stability
bracing force effects are to be combined with other construction-related load cases via
principles of superposition. Additional specification and commentary language have been
proposed to explain this procedure as well as the pertinent load factors. Design examples in
Appendices B and C demonstrate these calculations.
References
143
Helwig, T., K. Frank, and J. Yura. 1997. “Lateral-Torsional Buckling of Singly Symmetric I-Beams.” Journal of
Structural Engineering 123 (9): 1172–1179.
Keating, P., D. Mertz, J. Kulicki, and C. Hess. 1990. Economical and Fatigue Resistant Steel Bridge Details.
FHWA-H1-90-043, Washington, D.C.: Federal Highway Administration.
Kulicki, J., Z. Prucz, C. Clancy, D. Mertz, and A. Nowak. 2007. Final Report, NCHRP Project 20-7/186,
“Updating the Calibration Report for AASHTO LRFD Code.” Washington, D.C.: Transportation Research
Board.
Liu, Y., and T. Helwig. 2020. “Torsional Brace Strength Requirements for Steel I-Girder Systems.” Journal of Bridge
Engineering (American Society of Civil Engineers) 146 (1): 1–11.
McDonald, G., and K. Frank. 2009. The Fatigue Performance of Angle Cross-Frame Members in Bridges. MS Thesis,
The University of Texas at Austin.
Modjeski and Masters. 2015. SHRP 2 Report S2-R19B-RW-1: Bridges for Service Life Beyond 100 Years: Service
Limit State Design. Washington, D.C.: Transportation Research Board of the National Academies.
Nowak, A. 1999. NCHRP Research Report 368: Calibration of LRFD Bridge Design Code. Washington, D.C.:
Transportation Research Board.
Prado, E., and D. White. 2015. “Assessment of Basic Steel I-Section Beam Bracing Requirements by Test
Simulation.” Report to the Metal Building Manufacturers Association.
Quadrato, C., A. Battistini, T. Helwig, M. Engelhardt, and K. Frank. 2014. “Increasing Girder Elastic Buckling
Strength using Split Pipe Bearing Stiffeners.” Journal of Bridge Engineering (American Society of Civil
Engineers) 19 (4).
Reichenbach, M., Y. Liu, T. Helwig, and M. Engelhardt. 2020. “Lateral-Torsional Buckling of Singly Symmetric
I-Girders with Stepped Flanges.” Journal of Structural Engineering (American Society of Civil Engineers).
Rivera, J., and B. Chavel. 2015. Steel Bridge Design Handbook Design Example 3: Three-Span Continuous Hori-
zontally Curved Composite Steel I-Girder Bridge. Washington, D.C.: Federal Highway Administration.
Romage, M. 2008. Field Measurements on Lean-On Bracing System for Steel I-Girder Bridges with Skewed Supports.
Thesis, The University of Texas at Austin.
Sivakumar, B., M. Ghosn, and F. Moses. 2011. NCHRP Research Report 683: Protocols for Collecting and Using
Traffic Data in Bridge Design. Washington, D.C.: Transportation Research Board.
Tadros, M., N. Al-Omaishi, S. Seguirant, and J. Gallt. 2003. NCHRP Research Report 496: Prestress Losses in
Pretensioned High-Strength Concrete Bridge Girders. Washington, D.C.: Transportation Research Board.
Taylor, A., and M. Ojalvo. 1966. “Torsional Restraint of Lateral Buckling.” Journal of the Structural Division
(American Society of Civil Engineers) 92 (ST2): 115–129.
Texas Department of Transportation. 2020. Transportation Planning Maps. https://www.txdot.gov/inside-txdot/
division/transportation-planning/maps.html.
Timoshenko, S., and Gere, J. 1961. Theory of Elastic Stability. New York: McGraw-Hill.
Wang, L., and T. Helwig. 2005. “Critical Imperfections for Beam Bracing Systems.” Journal of Structural Engi-
neering (American Society of Civil Engineers) 131 (6): 933–940.
Wang, L., and T. Helwig. 2008. “Stability Bracing Requirements for Steel Bridge Girders with Skewed Supports.”
Journal of Bridge Engineering (American Society of Civil Engineers) 13 (2): 149–157.
Wang, W. 2013. A Study of Stiffness of Steel Bridge Cross Frames. PhD diss., The University of Texas at Austin.
White, D., D. Coletti, B. Chavel, A. Sanchez, C. Ozgur, J. Jimenez Chong, R. Leon, et al. 2012. NCHRP Research
Report 725: Guidelines for Analysis Methods and Construction Engineering of Curved and Skewed Steel Girder
Bridges. Washington, D.C.: Transportation Research Board.
White, D., T. Nguyen, D. Coletti, B. Chavel, M. Grubb, and C. Boring. 2015. Final Report, NCHRP 20-07/
Task 355, “Guidelines for Reliable Fit-Up of Steel I-Girder Bridges.” Washington, D.C.: Transportation
Research Board.
White, J. 2020. Evaluation of Fatigue Design Load Models for Cross-Frames in Steel I-Girder Bridges. PhD diss.,
The University of Texas at Austin.
Winter, G. 1960. “Lateral Bracing of Columns and Beams.” Journal of Structural Engineering (American Society
of Civil Engineers) 125: 809–825.
Yura, J. 2001. “Fundamentals of Beam Bracing.” Engineering Journal (American Institute of Steel Construction)
11–26.
Yura, J., B. Philips, S. Raju, and S. Webb. 1992. Bracing of Steel Beams in Bridges. Research Report 1239-4F, Austin,
TX: Center for Transportation Research, University of Texas at Austin.
Yura, J., T. Helwig, R. Herman, and C. Zhou. 2008. “Global Lateral Buckling of I-Shaped Girder Systems.” Journal
of Structural Engineering (American Society of Civil Engineers) 134 (9): 1487–1494.
APPENDIX A
Proposed Modifications
to AASHTO LRFD
A-1
Washington, DC 20001
500 Fifth Street, NW
Transportation Research Board
ISBN 978-0-309-67373-0
90000
9 780309 673730