Evaluation of Critical and Essential Concrete Highway Bridges in a Moderate 
Seismic Hazard 
 
by 
 
Jordan L. Panzer 
 
 
 
 
A thesis submitted to the Graduate Faculty of 
Auburn University 
in partial fulfillment of the 
requirements for the Degree of 
Master of Civil Engineering 
 
Auburn, Alabama 
May 3, 2014 
 
 
 
Copyright 2014 by Jordan L. Panzer 
 
Approved by 
 
Justin D. Marshall, Ph.D., P.E., Chair, Assistant Professor of Civil Engineering 
J. Brian Anderson, Ph.D., P.E., Associate Professor of Civil Engineering 
Robert W. Barnes, Ph.D., P.E., Associate Professor of Civil Engineering 
ii 
 
Abstract 
 
 
 The task of this thesis is to determine if the response of critical and essential 
bridges designed using the AASHTO LRFD Bridge Design Specifications would meet 
performance requirements when subjected to the moderate seismic hazard present in the 
state of Alabama.  The design procedures for these concrete superstructure bridges 
involve a displacement-based design found in the AASHTO Guide Specifications for 
LRFD Seismic Bridge Design.  Displacement-based design procedures in bridges can 
provide an economical and efficient design.  However, a caveat to using the 
displacement-based design procedures of the Guide Specifications relates to its 
inapplicability for critical or essential bridges.   This thesis will provide an analysis of 
designed bridge behavior in the state, and determine if this behavior is acceptable for 
bridges classified as critical or essential.  Recommendations and conclusions relating to 
design will also be provided for bridges subject to Alabama?s moderate hazard. 
  
iii 
 
Acknowledgements 
 
 
 I would like to thank the Alabama Department of Transportation for providing 
valuable bridge plans and designs for use in this project, as well as connection details that 
aided in the complicated modeling procedure undertaken in this project.  This project 
would have been all the more difficult without these valuable documents. 
 I would like to thank those involved in the maintenance and updating of resources 
like PEER and the USGS, resources that provide valuable records of ground motions and 
hazard maps that help to further our knowledge of seismic behavior and phenomena. 
 I would like to thank Dr. Justin Marshall for his leadership and knowledge 
throughout this project, without his guidance this undertaking would not have been 
possible.  He has helped many students learn to become better engineers and I am no 
exception. 
 I would like to thank the civil engineering faculty of Auburn University for 
always being available to answer questions especially Dr. Anderson, who provided 
valuable insight on all manner of geotechnical phenomena pertaining to this project. 
 I would like to thank all of my friends and peers that I have come to know these 
past six years, their support and guidance has helped me grow as a person, co-worker, 
and engineer.  Without them I do not know if I could have accomplished any of this. 
iv 
 
 Lastly I would like to thank my family, who has always supported me in life and 
school.  They motivate me to work hard and are a constant source of guidance.  I am 
grateful for everything they have given me. 
  
v 
 
Table of Contents 
 
 
Abstract ............................................................................................................................................ ii 
Acknowledgements ......................................................................................................................... iii 
List of Tables ................................................................................................................................ viii 
List of Figures ................................................................................................................................. xi 
Chapter 1: Introduction .................................................................................................................... 1 
1.1 Problem Statement ........................................................................................................... 1 
1.2 Problem Overview ........................................................................................................... 2 
Chapter 2: Literature Review ........................................................................................................... 4 
2.1 Bridges in Seismic Events ............................................................................................... 4 
2.1.1 Function of Bridges .................................................................................................. 4 
2.1.2 Bridge Components ................................................................................................. 4 
2.1.3 Seismic Events ......................................................................................................... 5 
2.1.4 Seismic Behavior of Bridges.................................................................................... 7 
2.1.5 Seismic Damage to Bridges ..................................................................................... 9 
2.2 Modeling and Analysis Method ..................................................................................... 16 
2.3 Ground Motion Selection ............................................................................................... 17 
2.3.1 Selection and Scaling ............................................................................................. 18 
2.4 Modeling Considerations ............................................................................................... 18 
2.5 Summary ........................................................................................................................ 18 
Chapter 3: Analysis Procedure ....................................................................................................... 20 
3.1 Introduction .................................................................................................................... 20 
3.2 Selection of Bridges ....................................................................................................... 20 
3.2.1 Seismic Hazard ...................................................................................................... 21 
3.2.2 Soil/Geologic Conditions ....................................................................................... 21 
3.2.3 Bridge Geometry .................................................................................................... 22 
3.2.4 Final Selection ....................................................................................................... 23 
3.3 Selection of Ground Motions ......................................................................................... 26 
vi 
 
3.3.1 Probable Seismic Hazard Analysis (PSHA) .......................................................... 26 
3.3.2 Uniform Hazard Spectrum (UHS) ......................................................................... 30 
3.3.3 Ground Motion ....................................................................................................... 32 
3.4 Selection of Modeling Elements .................................................................................... 37 
3.4.1 Superstructure Modeling ........................................................................................ 38 
3.4.2 Substructure Model ................................................................................................ 39 
3.4.3 Connection Modeling ............................................................................................. 44 
3.5 Analysis Method ............................................................................................................ 51 
3.5.1 Pre-Analysis Procedure .......................................................................................... 52 
3.5.2 Post-Analysis Procedure ........................................................................................ 52 
3.6 Overview of Bridge Models ........................................................................................... 52 
3.6.1 Little Bear Creek Bridge ........................................................................................ 53 
3.6.2 Oseligee Creek Bridge ........................................................................................... 56 
3.6.3 Norfolk Southern Railroad Bridge ......................................................................... 58 
3.6.4 Scarham Creek ....................................................................................................... 60 
3.6.5 Bent Creek Road .................................................................................................... 62 
3.7 Determination of Capacities .......................................................................................... 64 
Chapter 4: Analysis Results and Discussion .................................................................................. 69 
4.1 Introduction .................................................................................................................... 69 
4.1.1 Description of Results Presented ........................................................................... 69 
4.1.2 Limits of Bridge Behaviors .................................................................................... 71 
4.2 Bent Creek Road Bridge ................................................................................................ 72 
4.2.1 Transverse Motion ................................................................................................. 72 
4.2.2 Longitudinal Motion .............................................................................................. 76 
4.3 Scarham Creek Bridge ................................................................................................... 79 
4.3.1 Transverse Motion ................................................................................................. 79 
4.3.2 Longitudinal Motion .............................................................................................. 82 
4.4 Norfolk Southern Railroad Bridge ................................................................................. 84 
4.4.1 Transverse Motion ................................................................................................. 85 
4.4.2 Longitudinal Motion .............................................................................................. 88 
4.5 Oseligee Creek Bridge ................................................................................................... 91 
4.5.1 Transverse Motion ................................................................................................. 91 
4.5.2 Longitudinal Motion .............................................................................................. 94 
vii 
 
4.6 Little Bear Creek Bridge ................................................................................................ 96 
4.6.1 Transverse Motion ................................................................................................. 97 
4.6.2 Longitudinal Motion ............................................................................................ 101 
4.7 Summary of Results ..................................................................................................... 103 
Chapter 5: Summary, Conclusions and Recommendations ......................................................... 107 
5.1 Summary ...................................................................................................................... 107 
5.2 Conclusions .................................................................................................................. 108 
5.3 Recommendations ........................................................................................................ 109 
References .................................................................................................................................... 110 
APPENDIX A: GROUND MOTIONS ....................................................................................... 112 
viii 
 
List of Tables 
Table 3.1: Bridge Locations .............................................................................................. 23 
Table 3.2: Hazard Map Data ............................................................................................. 31 
Table 3.3: Ground Motion Data ........................................................................................ 34 
Table 3.4: Ground Motion Scaling ................................................................................... 37 
Table 3.5: Foundation Capacities ..................................................................................... 65 
Table 3.6: Calculated Bolt Strengths ................................................................................ 66 
Table 3.7: Bent Displacement Limits ............................................................................... 67 
Table 4.1: Results Overview for Lower Limit Friction Bent Creek Transverse Model ... 74 
Table 4.2: Results Overview for Upper Limit Friction Bent Creek Transverse Model ... 74 
Table 4.3: MCE-Level Results Overview for Lower Limit Friction Bent Creek 
Transverse Model.................................................................................................. 75 
Table 4.4: MCE-Level Results Overview for Upper Limit Friction Bent Creek Transverse 
Model .................................................................................................................... 76 
Table 4.5: Results Overview for Lower Limit Friction Bent Creek Longitudinal Model 77 
Table 4.6: Results Overview for Upper Limit Friction Bent Creek Longitudinal Model 77 
Table 4.7: MCE-Level Results Overview for Lower Limit Friction Bent Creek 
Longitudinal Model .............................................................................................. 78 
Table 4.8: MCE-Level Results Overview for Upper Limit Friction Bent Creek 
Longitudinal Model .............................................................................................. 78 
Table 4.9: Results Overview for Lower Limit Friction Scarham Creek Transverse    
Model .................................................................................................................... 80 
Table 4.10: Results Overview for Upper Limit Friction Scarham Creek Transverse  
Model .................................................................................................................... 81 
Table 4.11: MCE-Level Results Overview for Lower Limit Friction Scarham Creek 
Transverse Model.................................................................................................. 81 
Table 4.12: MCE-Level Results Overview for Upper Limit Friction Scarham Creek 
Transverse Model.................................................................................................. 82 
Table 4.13: Results Overview for Lower Limit Friction Scarham Creek Longitudinal 
Model .................................................................................................................... 83 
Table 4.14: Results Overview for Upper Limit Friction Scarham Creek Longitudinal 
Model .................................................................................................................... 83 
Table 4.15: MCE-Level Results Overview for Lower Limit Friction Scarham Creek 
Longitudinal Model .............................................................................................. 84 
Table 4.16: MCE-Level Results Overview for Upper Limit Friction Scarham Creek 
Longitudinal Model .............................................................................................. 84 
Table 4.17: Results Overview for Lower Limit Friction Norfolk Southern Railroad 
Transverse Model.................................................................................................. 86 
Table 4.18: Results Overview for Upper Limit Friction Norfolk Southern Railroad 
Transverse Model.................................................................................................. 86 
ix 
 
Table 4.19: MCE-Level Results Overview for Lower Limit Friction Norfolk Southern 
Railroad Transverse Model ................................................................................... 87 
Table 4.20: MCE-Level Results Overview for Upper Limit Friction Norfolk Southern 
Railroad Transverse Model ................................................................................... 87 
Table 4.21: Results Overview for Lower Limit Friction Norfolk Southern Longitudinal 
Model .................................................................................................................... 89 
Table 4.22: Results Overview for Upper Limit Friction Norfolk Southern Longitudinal 
Model .................................................................................................................... 89 
Table 4.23: MCE-Level Results Overview for Lower Limit Friction Norfolk Southern 
Longitudinal Model .............................................................................................. 90 
Table 4.24: MCE-Level Results Overview for Upper Limit Friction Norfolk Southern 
Longitudinal Model .............................................................................................. 90 
Table 4.25: Results Overview for Lower Limit Friction Oseligee Creek Transverse  
Model .................................................................................................................... 92 
Table 4.26: Results Overview for Upper Limit Friction Oseligee Creek Transverse  
Model .................................................................................................................... 93 
Table 4.27: MCE-Level Results Overview for Lower Limit Friction Oseligee Creek 
Transverse Model.................................................................................................. 93 
Table 4.28: MCE-Level Results Overview for Upper Limit Friction Oseligee Creek 
Transverse Model.................................................................................................. 94 
Table 4.29: Results Overview for Lower Limit Friction Oseligee Creek Longitudinal 
Model .................................................................................................................... 95 
Table 4.30: Results Overview for Upper Limit Friction Oseligee Creek Longitudinal 
Model .................................................................................................................... 95 
Table 4.31: MCE-Level Results Overview for Lower Limit Friction Oseligee Creek 
Longitudinal Model .............................................................................................. 96 
Table 4.32: MCE-Level Results Overview for Upper Limit Friction Oseligee Creek 
Longitudinal Model .............................................................................................. 96 
Table 4.33: Results Overview for Lower Limit Friction Little Bear Creek Transverse 
Model .................................................................................................................... 98 
Table 4.34: Results Overview for Upper Limit Friction Little Bear Creek Transverse 
Model .................................................................................................................... 99 
Table 4.35: Results Overview for Upper Limit Friction Little Bear Creek Transverse 
Model with Old Bolts............................................................................................ 99 
Table 4.36: MCE-Level Results Overview for Lower Limit Friction Little Bear Creek 
Transverse Model................................................................................................ 100 
Table 4.37: MCE-Level Results Overview for Upper Limit Friction Little Bear Creek 
Transverse Model................................................................................................ 100 
Table 4.38: MCE-Level Results Overview for Upper Limit Friction Little Bear Creek 
Transverse Model with Old Bolts ....................................................................... 100 
Table 4.39: Results Overview for Lower Limit Friction Little Bear Creek Longitudinal 
Model .................................................................................................................. 102 
Table 4.40: Results Overview for Upper Limit Friction Little Bear Creek Longitudinal 
Model .................................................................................................................. 102 
Table 4.41: MCE-Level Results Overview for Lower Limit Friction Little Bear Creek 
Longitudinal Model ............................................................................................ 103 
x 
 
Table 4.42: MCE-Level Results Overview for Upper Limit Friction Little Bear Creek 
Longitudinal Model ............................................................................................ 103 
Table 4.43: Summary of Critical Design-Level Analysis Behaviors ............................. 105 
Table 4.44: Summary of Critical MCE-Level Analysis Behaviors ................................ 106 
 
xi 
 
List of Figures 
 
Figure 2.1: Overview of damage to south abutment of Tubul bridge  .............................. 10 
Figure 2.2: Ground settlement and lateral movement of eastern approach to Llacolen 
bridge .................................................................................................................... 10 
Figure 2.3: Column settlement under approach to Juan Pablo II bridge  ......................... 11 
Figure 2.4: Differential settlement underneath first span over water at northern end of 
Juan Pablo II bridge  ............................................................................................. 11 
Figure 2.5: Lateral spreading and ground failure at old Ramadillas bridge  .................... 12 
Figure 2.6: The unseated slab of Baihwa Bridge  ............................................................. 13 
Figure 2.7: Unseating of simply supported span in eastern approach to Llacolen Bridge 13 
Figure 2.8: Unseated spans of the Tubul Bridge  .............................................................. 14 
Figure 2.9: Moorehouse Avenue Overbridge pier flexural buckling failure  ................... 16 
Figure 2.10: Shear failure in column at north end of Juan Pablo II Bridge  ..................... 16 
Figure 3.11: Bridge Locations .......................................................................................... 24 
Figure 3.2: Deaggregation Plot for Muscle Shoals, AL.................................................... 28 
Figure 3.3: Deaggregation Plot for Bridgeport, AL .......................................................... 29 
Figure 3.4: Targeted Design Spectrum ............................................................................. 31 
Figure 3.5: Targeted MCE Spectrum ................................................................................ 32 
Figure 3.6: Design-Level Response Spectra ..................................................................... 35 
Figure 3.7: MCE-Level Response Spectra........................................................................ 36 
Figure 3.8: Idealized Caltrans Hinge ................................................................................ 41 
Figure 3.9: Fiber Analysis Backbone Curve ..................................................................... 43 
Figure 3.10: Example Bearing Pad Configuration  ........................................................... 45 
Figure 3.11: Example Bearing Pad Force-Displacement Relationship ............................ 46 
Figure 3.12: Clip Angle Connection Detail ...................................................................... 48 
Figure 3.13: Little Bear Creek Bridge Model ................................................................... 55 
Figure 3.14: Oseligee Creek Bridge Model ...................................................................... 57 
Figure 3.15: Norfolk Creek Bridge Model ....................................................................... 59 
Figure 3.16: Scarham Creek Bridge Model ...................................................................... 61 
Figure 3.17: Bent Creek Bridge Model............................................................................. 63 
1 
 
Chapter 1: Introduction 
 
1.1 Problem Statement 
 
The evaluation and determination of geologic seismic behavior and hazard is a 
constantly updating process.  New research and updated geologic analysis technologies 
results in greater understanding of seismic hazards, resulting in updated codes and 
specifications.  Recent research has resulted in an increased understanding of the 
seismicity of the Southeast United States.  This region has been known to have a 
relatively low seismicity, resulting in its classification as a moderate seismic hazard zone.  
Alabama is one of the states in this region that has had its seismic hazard modified 
recently.  These modifications make their way into the seismic design specifications in 
the form of increased seismic loads and increased seismic design categories. 
The Alabama Department of Transportation (ALDOT) is currently using the 17th 
edition of the American Association of State Highway and Transportation (AASHTO) 
Standard Specifications (Standard Specifications for Highway Bridges, 2002) for the 
design of precast prestressed concrete bridges for the state.  Originally these 
specifications were designed using allowable stress design (ASD) theory and have been 
updated to incorporate principles of Load and Resistance Factor Design (LRFD).  These 
specifications, however, have not been updated since 2002.  As a result of this lack of 
updating, ALDOT has been forced to adopt a new bridge design specification.  The new 
design specifications selected were the AASHTO LRFD Design Specifications (LRFD 
2 
 
Bridge Design Specifications, 2009), these specifications are based on LRFD principles 
and regularly updated with the help of new research.  Major changes in the recent 
specifications revolve around the area of seismic design, prompting an update to 
ALDOT?s seismic design criteria. 
 
1.2 Problem Overview 
 
In an effort to update seismic bridge design procedures, the Alabama Department 
of Transportation (ALDOT) has investigated the use of the AASHTO Guide 
Specifications for LRFD Seismic Bridge Design (AASHTO, 2011).  This document uses 
displacement-based criteria in the seismic design of highway bridges.  The document was 
created for use in any region of the United States, and has been tailored to account for the 
most seismically active zones.  A downside to using this document is its exclusion of 
criteria for performance of ?critical? and ?essential? bridges.  These bridge classifications 
are determined based on the bridge?s role as a transportation lifeline after a major 
destructive event.   
ALDOT requested that a study and analysis be performed on various highway 
bridges designed according to the guide specifications.  These analyses assess the 
behavior and performance of each of these bridges during seismic events that are scaled 
to the hazard present in Alabama.  Results of these analyses provide a recommendation as 
to whether or not the Guide Specifications can be used as a design tool for critical and 
essential bridges in the state of Alabama.  The following document provides an overview 
of the work done leading up to the bridge analysis, as well as the analysis itself and the 
conclusions and recommendations obtained from the results of the analysis.  The next 
3 
 
chapter of this thesis provides a background for recent seismic bridge damage found 
around the world, as well as the bridge behaviors seen during seismic events.  This 
chapter also provides information regarding seismic hazards in Alabama.  The third 
chapter of this thesis provides an overview of the methodology and procedure used in the 
analyses performed for the presented research.  Modeling decisions and hazard 
evaluation/selection are also presented in this chapter.  The fourth chapter of this thesis 
presents the results taken from the analyses performed, as well as indicated some of the 
key pieces of data that relate to bridge performance during seismic events.  The final 
chapter of this thesis provides conclusions and recommendations obtained from analysis 
of the observed results of this research.  
4 
 
Chapter 2: Literature Review 
 
2.1 Bridges in Seismic Events 
 
2.1.1 Function of Bridges 
 
Bridges serve a variety of important purposes in the realm of transportation 
infrastructure.  Bridges provide a transportation connection between two points that could 
not be connected using standard roadways.  They are planned and engineered far in 
advance of construction.  They are often the most expensive sections of transportation 
systems per mile, and are likely to control the capacity of the transportation system 
(Barker & Puckett, 2007).  In order for a transportation system to function, all of its 
collective parts must function, thus if a bridge fails, the entire system fails. 
Highway systems provide transportation for all ground vehicles and are used by 
both civilian and military ground vehicles.  They provide transportation for goods, 
services, people and emergency services.  In the event of the failure of these systems, the 
flow of these aforementioned items would be drastically reduced, if not halted 
completely.  This fact illustrates the importance of preserving the functionality of these 
systems and their respective components. 
 
2.1.2 Bridge Components 
 
Typical components that are present in most bridges include girders, deck, 
abutments, foundations and bents/piers.  The focus of this paper is on structural concrete 
5 
 
highway bridges, which consist of all the aforementioned components.  Starting from the 
ground up, these bridges consist of a foundation system, connecting the soil beneath the 
bridge with the substructure of the bridge.  This substructure consists of abutments, 
elements that connect roadway to the bridge, as well as vertical columns and the beams, 
called bent caps, which span horizontally between and connect their tops.  These 
?Column-Bent Cap? frames are called ?Bents? and serve to support the superstructure of 
the bridge.  Standard Highway Bridge superstructures consist of reinforced concrete 
decking that rests on girders.  These superstructures are broken into spans, sections of 
superstructure that are supported on either end by substructure. 
In the case of the bridges in this research, a few more details can be included in 
the description of bridge components.  The prestressed concrete girders are simply 
supported at each end and rest on elastomeric bearing pads.  These bearing pads rest on 
the surface of abutments and bent caps.  These pads allow for movement of the girders 
relative to the substructure.  Girders are also given lateral stability and stiffness with the 
addition of cast in place reinforced concrete web walls.  Both the bent caps and the 
columns that support them consist of cast in place reinforced concrete.  These are the 
main components that are focused on in this paper. 
2.1.3 Seismic Events 
 
Among the many events that threaten to damage and possibly cause the failure of 
bridge systems, earthquakes and other seismic related phenomena have the capacity to do 
significant damage in many areas throughout the United States and around the world.  
Chen & Duan define earthquakes in Bridge Engineering Seismic Design as   
6 
 
?naturally occurring broad-banded vibratory ground motions that 
are due to a number of causes, including tectonic ground motions, 
volcanism, landslides, rockbursts, and man-made explosions, the most 
important of which are caused by the fracture and sliding of rock along 
tectonic faults within the Earth?s crust.?   
    
Ground motion is a primary result of earthquakes, but it is not the only impact 
worth mentioning.  Tsunamis, like the one that occurred off the coast of Japan as a result 
of the 2011 T?hoku Earthquake, can cause damage on par or greater than the resulting 
ground motions of earthquakes.  Fissures, like those caused in Taiwan during the Chi-Chi 
event (September 21st, 1999), open in the soil and rock near the site of an earthquake as a 
result of displacement that occurs along faults, deep underground.  These fissures can 
cause massive displacement differentials in any structure that rests above them, often 
resulting in large scale damage.  Liquefaction is also an important event that can result in 
certain soil types during an earthquake.  ?Liquefaction has been widely used to describe a 
range of phenomena in which the strength and stiffness of a soil deposit are reduced due 
to the generation of pore water pressure.  It occurs most commonly in loose, saturated 
sands? (Chen & Duan, 2000).  These occurrences are difficult to account for because they 
are hard to accurately predict in both frequency and scale. 
Earthquake ground motions also trigger other destructive events.  Landslides and 
lateral spreading can cause damage to structures and roadways, and are produced by 
ground motions in certain soil conditions.  Man-made disasters can also result from 
damage caused by ground motions, an example being the radiation leak of the Fukashima 
7 
 
nuclear reactor that occurred as a result of the 2011 T?hoku Earthquake.  Fires are also a 
common result of widespread damage in populated areas affected by large ground 
motions.  Structural collapse can also occur in severe ground motions, resulting in 
massive economic costs as well as the possibility of human casualties.   
Earthquakes present a large economic and safety hazard for many urban areas 
around the world.  Regions affected by severe seismic events usually require emergency 
aid as well as viable ways of transporting civilians away from the areas that have become 
unsafe or unfit for habitation.  The importance of having functioning transportation 
systems is highlighted in situations like this.  Bridges and roadways are not immune to 
damage however, and steps must be taken to ensure their survival and function after 
seismic events. 
2.1.4 Seismic Behavior of Bridges  
 
Bridges react to ground motions in much the same way as buildings and other 
structures.  Ground motions transfer forces into the foundation of the bridge due to the 
bridges natural inertial resistance to movement.  These foundation forces transfer through 
the substructure of the bridge and eventually reach the connection between super and 
substructure.  The forces are then transferred through the interface of the super and 
substructure.  The result of these forces is deformation of the effected components of the 
bridge.  This movement can be defined by a combination of displacements, velocities and 
accelerations.   
In the case of simple highway girder bridges, some generalized behavior can be 
described.  Due to the super-to-substructure connection consisting of friction bearing 
pads, it can be assumed that the superstructure slides with respect to the substructure.  
8 
 
The overall system behaves like two separate systems interfacing with one another at the 
bearing pad locations.  The superstructure behaves like a stiff beam (simply supported), 
while the substructure behaves like a rigid frame consisting of columns and beams.  The 
largest proportion of mass is located in the superstructure, giving it more inertial response 
to ground motion accelerations.   
Bridges, like all structures, have certain characteristics that can be used to predict 
seismic behavior.  Mass distribution gives an indication as to how the structure will 
respond to forces in the form of motion.  Stiffness helps to predict the distribution of 
forces throughout the structure.  A fundamental structural period, a parameter that 
dictates dynamic behavior, can be established using a relation of both of these 
parameters.  The fundamental structural period represents a time interval of structure 
oscillation.  If a loading function was applied to a structure whose period is similar to that 
of the fundamental structural period, the structure experiences a buildup of energy, often 
with little way of dissipating that energy.  These loading cases can continue until either 
the stiffness or the mass of the system changes, or the properties of the loading conditions 
change.  Cases like these can cause certain ground motions to become more damaging to 
certain bridges.  Many highway bridges have fundamental periods within the range where 
earthquakes motions contain significant energy.  This explains the existence of dissimilar 
bridges having very different responses to the same ground motion and it also highlights 
the importance of bridges having the ability to dissipate energy during a ground motion 
through the use of targeted damage zones.  
 
9 
 
2.1.5 Seismic Damage to Bridges 
 
Seismic damage in bridges can be divided into two classes, primary and 
secondary damage.  Primary damage is described as bridge damage caused directly by 
earthquake ground shaking or deformation.  Secondary damage is described as bridge 
damage that was caused by other failures or damage that resulted from ground shaking or 
deformation (Chen & Duan, 2000).  An example of primary damage would be a span 
becoming unseated as a result of ground motions, and the resulting girder damage caused 
by the spans impact with the ground would be considered secondary damage.   
This section has emphasis placed on three critical examples of primary damage 
that are typically associated with earthquake bridges failures: 1) Foundation Failure, 2) 
Span Unseating and 3) Column Failure.  Each of these failures have been documented in 
earthquakes around the world, and are a primary concern in bridge design, especially in 
regions with moderate to high seismic hazards. 
Foundation failure is a broad title meant to encompass any failure of soils or 
foundation elements.  According to Chen and Duan, reports of foundation failures during 
earthquakes are not very common, except for cases involving liquefaction.  They do 
speculate that underground foundation failures may remain undetected, keeping a 
significant number of failures unreported.  Despite being rarely reported, foundation 
failures remain a critical failure mode of bridges, whose importance should not be 
overlooked.  ?Foundation damage associated with liquefaction-induced lateral spreading 
has probably been the single greatest cause of extreme distress and collapse of bridges? 
(Chen & Duan, 2000).  Examples of foundation failure and lateral spreading, resulting 
from 2010 Maule Earthquake can be seen in the following figures presented by Yen et al. 
10 
 
(2011).  Figure 2.1 displays a bridge abutment that has displaced enough to have 
separated from one of its support piles.  Figure 2.2 contains a photograph taken of the 
settlement of a bridge foundation; this bridge, the eastern approach to Llacolen bridge, is 
discussed later in this section due to a span unseating.  Examples of foundation settlement 
that did not result in bridge collapse can be found in Figure 2.3 and Figure 2.4.  An 
example of lateral spreading can be seen in Figure 2.5, which can cause additional 
foundation loading pressures. 
 
 
Figure 2.1: Overview of damage to south abutment of Tubul bridge  (Yen, et al. 2011) 
 
 
Figure 2.2: Ground settlement and lateral movement of eastern approach to Llacolen bridge 
 (Yen, et al. 2011) 
 
11 
 
 
Figure 2.3: Column settlement under approach to Juan Pablo II bridge (Yen, et al. 2011) 
 
 
Figure 2.4: Differential settlement underneath first span over water at northern end of Juan Pablo II 
bridge (Yen, et al. 2011) 
 
12 
 
 
Figure 2.5: Lateral spreading and ground failure at old Ramadillas bridge (Yen, et al. 2011) 
 
 
Unseating of spans is a commonly observed phenomenon in bridges that fail 
during earthquakes.  This behavior is especially prevalent in simply supported bridges 
with shorter spans and seat widths.  Unseating refers to the girders of a span becoming 
displaced to a large enough degree resulting in a complete loss of vertical support 
provided by the bent cap that they are resting on.  This situation results in a massive loss 
in vertical strength and overall stability for the span, and can cause other elements of a 
bridge to become damaged or fail.  If enough of the girders experience unseating, a total 
collapse of the span may occur. 
Earthquakes that occurred in California during the 80s and 90s highlighted this 
downside of simply supported, simple span bridges and resulted in design changes away 
from these types of bridges.  These types of failures have also been observed around the 
world in recent years.  The Baihwa Bridge, a reinforced concrete bridge with a 500 meter 
13 
 
long viaduct and a height of 30 meters was constructed in 2004 in China, experienced a 
span unseating and collapse during the 2008 Wenchuan Earthquake.  This unseated slab 
can be seen in Figure 2.6.  Reconnaissance indicated that a lack of sufficient support 
length along with insufficient restrainers was observed for this bridge, as well as other 
damaged bridges in the region (Lin, et al. 2008).  An eastern approach span of the 
Llacolen Bridge and multiple spans of the Tubul Bridge in Concepci?n, Chile 
experienced span unseating during the 2010 Maule Earthquake as seen in Figure 2.7 and 
2.8 (Yen, et al. 2011). 
 
Figure 2.6: The unseated slab of Baihwa Bridge (Lin, et al. 2008) 
 
 
 
Figure 2.7: Unseating of simply supported span in eastern approach to Llacolen Bridge 
(Yen, et al. 2011) 
 
14 
 
 
Figure 2.8: Unseated spans of the Tubul Bridge (Yen, et al. 2011) 
 
Modern seismic design practices in bridge engineering seek to allow displacement 
and inelasticity to absorb and dissipate energy from ground motions.  Columns are 
designed to contain areas of high ductility that will form plastic hinges during ground 
motions.  These plastic hinge zones rely on strong connecting elements (e.g., bent caps, 
foundations) to allow yielding to the hinge zone without failure of other elements.  If this 
behavior does not occur properly a non-ductile column or connection failure could result. 
Column failure is the last damage case discussed in this Chapter.  Failure of a load 
carrying member, such as a column, can result in severe damage to total failure of a 
bridge. Column failure is often the primary cause of bridge collapse (Chen & Duan, 
2000).  In regions of high seismicity it is important to design and detail reinforced 
concrete columns with ductility and inelastic deformation in mind.  Most damage to 
columns can be attributed to inadequate detailing, resulting in limited deformation 
capacity and energy dissipation.  Detailing inadequacies can produce many different 
failures within a column, including a combination of flexural, shear, splice or anchorage 
failures (Chen & Duan, 2000).  Proper detailing is meant to provide both flexural 
strength, as well as shear strength to a column.  A flexural failure of a column occurs 
when the column rotational capacity is exceeded, either by strong ground motions or 
15 
 
excessive deflections of the bridge.  Flexural failures can be identified by broken or 
elongated rebar seen at hinge zones, as well as rebar completely uncoupled from any 
concrete.  An example of a flexural column failure can be seen in Figure 2.7 as a result of 
the 2011 Christchurch Earthquake, this particular column first experienced shear failure 
that weakened it, resulting in a flexural failure due to rebar buckling (Wotherspoon, et al. 
2011).  A shear failure occurs in a column when the shear resistance of the column is 
exceeded; a difficult failure to design for given that the shear strength of a column can be 
diminished during a bridges response to a ground motion.  Another factor that makes 
designing for shear failure difficult is the fact that the shear forces resisted by a column 
relate directly to the stiffness of the bridge, thus a stronger column will be subjected to a 
larger shear force.  Shear failures are identified by steep diagonal cracking throughout the 
center or ?core? of a reinforced concrete column.  These cracks extend throughout the 
entire diameter of the core, and in some cases completely dilate the core into discrete 
blocks of concrete (Chen & Duan, 2000).  An example of shear failure can be seen in 
Figure 2.8, in the case of a column supporting the Juan Pablo II Bridge in Concepci?n, 
Chile. 
16 
 
 
Figure 2.9: Moorehouse Avenue Overbridge pier flexural buckling failure (Wotherspoon, et al. 2011) 
 
 
Figure 2.10: Shear failure in column at north end of Juan Pablo II Bridge (Yen, et al. 2011) 
 
2.2 Modeling and Analysis Method 
 
The focus of this project involved analyzing the behavior of specific bridges 
modeled using computer software.  The software chosen was CSIBridge 15, a structural 
design and analysis program with the capability of nonlinear dynamic response history 
analysis (Computer and Structures Inc., 2012).  Given the need for an accurate model of 
17 
 
bridge behavior during ground motions, as well as accurate analysis of nonlinear 
elements, the Time-History Analysis method was selected for use in this project.  This 
method requires accurate bridge information (Structural and Geotechnical) from design 
documents, as well as an accurate suite of scaled ground motions meant to simulate 
earthquake events.     
 
2.3 Ground Motion Selection 
 
Time history analyses of nonlinear structures require an input of acceleration data 
in order to function.  Many factors are involved in the selection of acceleration data, chief 
among them being probable hazard associated with the location of the structure being 
analyzed.  Section 5.4.4 of AASHTO?s guide specifications states the requirements for a 
Nonlinear Time History Method Analysis (NTHM).  AASHTO requires at least three 
ground motions be used for NTHM, with less stringent performance requirements being 
placed on analysis involving seven or more (AASHTO, 2011).   For this particular project 
a range of seven to ten ground motions were selected for two different hazard levels, a 
design-level hazard, meant to represent the hazard of a 1,000 year event (ten selected 
ground motions), and a maximum credible earthquake (MCE) hazard, meant to represent 
the seismic hazard of a 2,500 year event (seven selected ground motions).  The 
probabilistic exceedance of these levels are 7% in 75 years and 2% in 50 years, 
respectively.  The selection process is similar to the process used by Rodriguez-Marek 
(Rodriguez-Marek, 2007). 
 
18 
 
2.3.1 Selection and Scaling  
 
Ground motions have a variety of unique characteristics that need to be accounted 
for when selected for a Time History Analysis.  Each ground motion needs to have 
similar parameters to that of the chosen hazard.  Parameters like Peak Ground 
Acceleration (PGA), distance from the fault and site class are all used in selection of 
ground motions.    Accepted ground motions are then scaled according to response 
factors discussed in the next chapter. 
2.4 Modeling Considerations 
 
Accuracy of bridge models is an important aspect of displacement-based and 
performance-based design.  Mass and stiffness make up the majority of calculation 
parameters used during nonlinear time history analysis, and taking account of these 
parameters and how they behave is paramount.  Changes in stiffness are the primary 
nonlinear behaviors that are accounted for in bridge models for these types of analysis.  
Without these behaviors, accurate results cannot be recorded. 
 
2.5 Summary 
 
This chapter reviewes the functions of bridges, and the role they play during a 
seismic event.  Seismic bridge behavior and damage is been highlighted in this chapter in 
an effort to focus attention on key design concepts, as well as provide an overview on the 
complexities of seismic bridge analysis.  The practices and guidelines set forth by 
AASHTO and other design aids take into account many different scenarios and factors in 
an attempt to ensure both survival and function of bridges under seismic loads.  The 
19 
 
methodology and philosophies outlined in this chapter continue to be used throughout the 
remainder of this project and thesis. 
20 
 
Chapter 3: Analysis Procedure 
3.1 Introduction 
 
The first part of this thesis is the overview of information that has been selected 
for use in the project detailed within.  The factors that led to the selection of ground 
motions, bridges, model connection designs and material behavior are all discussed, as 
well as simplifications that were chosen to expedite the analysis process.  Information 
was selected using research and procedures from projects performed in the past, as well 
as engineering judgment. 
3.2 Selection of Bridges 
 
All bridges analyzed in this project were selected from a list of existing highway 
bridges in the state of Alabama.  The seismic design details used in the analysis have 
been updated to the most recent edition of AASHTO?s Guide Specifications in order to 
analyze designs according to current code, as well as anticipate future bridge design in 
the state.  The bridge selection process revolved around three major factors: seismic 
hazard (based on location), soil condition and bridge geometry.  Seven highway bridges 
were considered for this analysis, but only five were chosen based on the aforementioned 
factors.  
 
 
 
21 
 
3.2.1 Seismic Hazard 
 
The geography of Alabama is not constant throughout, although some 
generalizations can be made for the purpose of this project.  The southern portion of the 
state tends to rest on deep layers of sandy soils.  These regions also tend to have little 
exposure to seismic activity.  The hazard originates in the Northwest region of Alabama 
(New Madrid Fault Zone), or in the mountainous region at the border of Alabama, 
Georgia and Tennessee (East Tennessee Seismic Zone).  Bridges in South Alabama tend 
to face a greater wind hazard than a seismic hazard.   This information, in combination 
with soil conditions, led to the elimination of most of the southern bridge designs.  
Bridges that were selected for use in this project tended to be located in the northern 
section of the state where the foundation conditions were similar to what would be 
expected in the areas of highest seismicity.  Selected bridges outside of this region still 
displayed characteristics consistent with bridges in the region. 
 
3.2.2 Soil/Geologic Conditions 
 
Geotechnical data became a defining factor in the selection of applicable bridges.  
Soil classifications at site locations can play a major role in overall earthquake hazard and 
structural performance.  Soil information was included along with the bridge plans 
provided by ALDOT for this project.  This soil information was used in the bridge 
selection process because it not only correlated to the bridge?s location in the state, but 
also correlated to the foundation types used in the bridges.   
Geological properties helped create expected soil conditions throughout the state 
in regions without specific data.  Basic generalizations were determined for different 
22 
 
regions of the state, and helped classify bridges, as well as estimate which bridges could 
be designed in specific regions.  The more mountainous regions of Alabama tended to 
have a soil layer between ten to thirty feet deep, resting on a hard rock layer.  
Foundations for these particular bridges consisted of drilled shafts that founded the bridge 
to the rock layer.  Other locations in northern Alabama consisted of deeper soil layers, 
but still utilized similar foundation systems.  Regions in central Alabama with similar soil 
conditions included bridges with pile group foundations, these bridges were also 
considered to be constructible in areas of higher seismicity, and thus were not dismissed 
in the selection process.  The bridges selected were bridges that could be constructed in 
higher hazard regions of Alabama, which also contained foundations that are suitable for 
soil found in these regions. 
 
3.2.3 Bridge Geometry 
 
A bridge?s geometry contributes significantly to the seismic behavior.  A bridge 
with tall columns and long spans will have a lower stiffness compared to a bridge with 
shorter spans and shorter columns.  In addition to component lengths, component 
dimensions are also a factor in bridge response.  For the purpose of this project it was 
important to collect a range of bridge geometry in order to realistically account for a 
spectrum of possible bridge designs.  Bridge configurations that included steel girders or 
concrete girders with changing thicknesses were not considered for this project.  Bridges 
that did not include bearing pad supports were also excluded from this project due to 
significant analysis effort being placed on bearing pad behavior and the girder-to-cap 
connection.   
23 
 
3.2.4 Final Selection 
 
A total of five bridge designs were selected for use in this project, all meeting the 
conditions required.  A list of these bridges and their locations can be seen in Table 3.1. 
A map of the bridge locations can be seen in Figure 3.1.  The Little Bear Creek Bridge 
was selected due to its location in the North West corner of the state, as well as its usage 
of two different girder sizes and bearing pad sizes along its three spans.  These 
discontinuities provided unique dynamic behavior, like a tendency for a structure to 
deviate from its first mode shape.  The soil conditions at this particular site consisted of a 
relatively shallow soil layer, as to be expected in this region of the state.  
  
Table 3.1: Bridge Locations 
Number Bridge Location 
1 Little Bear Creek Russellville 
2 Scarham Creek Albertville 
3 Norfolk Southern RR Gadsden 
4 Oseligee Creek Lanett 
5 Bent Creek Road Auburn 
 
24 
 
 
Figure 3.11: Bridge Locations 
 
The Scarham Creek Bridge was the second bridge selected for this project.  The 
Scarham Bridge contained the tallest and largest bent columns, making it one of the more 
flexible bridges.  Concrete struts were used in the bent design in an effort to reduce the 
unbraced length of the columns.  These struts presented a unique opportunity for dynamic 
element observation, since they appeared to have been overdesigned for the purpose of 
stability and safety.  Overdesigns like these can actually become harmful to earthquake 
25 
 
response if members are not detailed accordingly, so this condition made this particular 
bridge an important one to analyze.  The site conditions remain typical for the region, and 
the bridge is supported by drilled shaft foundations. 
The third bridge that was selected was the Norfolk Southern Railroad Bridge, a 
section of I-59 that spans above railroad tracks.  The geometry of the Norfolk Bridge 
represents a typical highway bridge.  It consists of two similar spans supported by a 
typical three column bridge bent.  The soil conditions present at this location resulted in 
the bridge being supported on a driven pile group footing foundations.  This was the first 
bridge selected with these foundations, as well as being the first bridge selected that uses 
rectangular columns instead of circular columns. 
The first bridge selected outside of the northern area of Alabama was the Oseligee 
Creek Bridge.  This bridge consists of three evenly spaced 80 ft. spans that are supported 
by two column bents.  The foundation of the bridge consists of driven pile foundations 
that extend all the way to bedrock.  Soil at this location was determined to have poor 
strength, and showed susceptibility to scour.  The geometry of this bridge was an 
important factor in the selection of this bridge, but its soil conditions may cause it to 
behave differently than expected.   
The last and largest bridge selected for these analyses was the Bent Creek Bridge, 
located in Lee County.  This bridge contains two bridge spans supported by fifteen 
girders each and a single five column bridge bent.  The mass and stiffness associated with 
this bridge made it an attractive option for analysis within this project.  The bridge itself 
is not located in a region of seismicity, but it was deemed plausible that this design could 
be reused in a more seismic region of Alabama. Some of the drawbacks regarding a 
26 
 
bridge this size involve modeling concerns.  This is the second bridge to be supported on 
a pile footing foundations. 
 
3.3 Selection of Ground Motions 
 
After the selection of bridges is performed, a determination of seismic hazard and 
the ground motions that represent them must be performed.  A total of ten scaled ground 
motions (GMs) were selected for the worst design-level hazard present in the state of 
Alabama.  The procedure used to select these ground motions involved the following 
tasks: 
1. Perform a probabilistic seismic hazard analysis (PSHA) for each hazard 
location.  
2. Collect GMs that contain values and properties similar to those highlighted by 
the PSHA. 
3. Obtain a design spectra for the locations from the PSHA based on a Uniform 
Hazard Spectrum (UHS) with a probability of exceedance of 7% in 75 years.  
4. Select 7-12 GMs with a spectral shape that best match the UHS.  
5. Apply scaling to the ground motions from (3) to fit the target spectra obtained 
from (2).  
3.3.1 Probable Seismic Hazard Analysis (PSHA) 
 
The PSH analyses for this project were performed using the tools provided by the 
United States Geological Survey (USGS) (https://geohazards.usgs.gov/deaggint/2008/).  
Details regarding the specifics of the procedures and calculations used by these tools can 
be found at the USGS website (USGS 2013).  A PSHA requires some information to run, 
27 
 
including site location, shear wave velocity of the top 30 meters of a soil (Vs30), and a 
seismic event return period for the hazard.  A Vs30 of 760 m/s, corresponding to a site 
class of B, was selected for Alabama, and a return period of 1000 years was selected, 
being input as a 5% exceedance probability in the next 50 years (similar return period 
correlating to the 7% exceedance in the next 75 years as specified in AASHTO).   The 
results of a PSHA analysis can be seen in the form of a deaggregation plot which allows 
the user to see the contributions from faults and other sources projected onto a 3?
dimensional plot.  The horizontal axis of the plot represents event magnitude; the vertical 
axis of the plot represents distance between event epicenter and location.  The height of 
each bar represents the portion of contribution an event has towards the hazard of a 
location.    Figure 3.2 displays the deaggregation plot for Muscle Shoals, Alabama. 
From a PSHA performed for four different locations in Alabama, it was 
determined that the two locations that had the highest seismic hazards were the northeast 
and northwest corners of the state, with a probabilistic peak ground acceleration (PGA) of 
0.09095g and 0.07557g, respectively.  It can be observed in Figure 3.3 that Bridgeport 
experiences a bimodal distribution meaning that the hazard is defined primarily by a 
magnitude 5-5.5 earthquake at a distance of 50 km (30 miles) or a magnitude 7.5-8 at a 
distance of 350 km (210 miles).  Using this information, a bin of about 40 earthquakes 
was selected using the Pacific Earthquake Engineering Research (PEER) Center 
database, a collection of ground motions recorded around the world.  These ground 
motions were then sorted by various parameters such as distance from fault, event 
magnitude, ground conditions, response spectrum behavior and PGA. 
 
28 
 
 
Figure 3.2: Deaggregation Plot for Muscle Shoals, AL 
 
29 
 
 
 
Figure 3.3: Deaggregation Plot for Bridgeport, AL
 
30 
 
3.3.2 Uniform Hazard Spectrum (UHS) 
 
A UHS was calculated for the two most critical locations in Alabama in the form 
of a Design Response Spectrum (DRS).  These DRSs were calculated using maps and 
equations present in Section 3.4 of AASHTO?s Guide Specifications.  The two locations 
selected for the creation of a UHS were Bridgeport, Alabama and Muscle Shoals, 
Alabama.  Both locations were assumed to have soil corresponding to a site class C.    
Table 3.2 provides the site data obtained from AASHTO?s ground motion maps.  Natural 
bridge periods were assumed to be in the range of zero to three seconds, and the UHS 
was plotted accordingly.  The target spectra for the bridges are displayed in Figure 3.4 
and Figure 3.5 which shows that the hazard spectra for the two sites and the resulting 
average response from the ground motions selected in the next section of this chapter.  It 
can be observed that the DRS for each site remain similar to the other, despite the 
geographic distance between the two sites.  The DRS for Muscle Shoals shows a greater 
hazard for bridges of higher periods, while the Bridgeport site shows slightly larger 
hazard for stiffer bridges.  This behavior can be explained by the expected event distance 
between event epicenter and site location, with close proximity events typically 
producing motions with higher frequencies.  These higher frequencies motions tend to 
produce higher responses in stiff structures; however these high frequency motions 
dissipate energy as they travel.  Longer frequency motions tend to travel longer distances 
and produce higher responses in flexible structures.  Due to the different hazards present 
for the state, both DRS?s were used in the ground motion scaling process. 
 
 
 
31 
 
Table 3.2: Hazard Map Data 
Hazard 
Location 
Values from map (g) Site coefficients Final Acceleration Values 
PGA Ss S1 FPGA Fa Fv As SDS SD1 
Bridgeport, AL 0.111 0.221 0.068 1.2 1.2 1.7 0.1332 0.2652 0.1156 
Florence, AL 0.089 0.213 0.078 1.2 1.2 1.7 0.1068 0.2556 0.1326 
 
 
Figure 3.4: Targeted Design Spectrum 
 
0 0.5 1 1.5 2 2.5 3
0
0.2
0.4
D e s i g n  S pe c t r u m  B r idg e po r t
S c a l e d A V G G r ou n d M ot i o n
D e s i g n  S pe c t r u m  M u s c l e  S h oa l s
T a rgete d  De s i gn R es po ns e  S p e c trum
N a t u r a l  p e r i od ( s )
A
c
c
e
l
e
r
a
t
i
on (
g)
 
32 
 
 
 
Figure 3.5: Targeted MCE Spectrum 
 
3.3.3  Ground Motion 
 
Initially a set of 40 ground motions were selected for possible use in this project 
and needed to be reduced to a set size of 7-12.  The initial constraints that led to the 
selection of these initial 40 included the following: 
1. A PGA between 0.007 and 0.020g 
0 0.5 1 1.5 2 2.5 3
0
0.2
0.4
0.6
M C E  S p e c t r u m  B r i d g e por t
S c a l e d A V G G r ou n d M ot i o n
M C E  S p e c t r u m  M u s c l e  S h oa l s
T a rgete d  MCE Res p ons e Sp ectrum
N a t u r a l  p e r i od ( s )
A
c
c
e
l
e
r
a
t
i
on (
g)
 
33 
 
2. A site Vs30 that did not correlate to a site class of A, E or F. 
3. An event magnitude between 5.5 and 7.5 
4.  Exclusion of any subduction zone events 
After an initial 40 were selected, properties of the GMs were analyzed in order to 
create a diverse suite of motions.  The first parameter used to eliminate motions was 
simply to limit the number of motions used from each event.  Oversaturation of a ground 
motion data set with many GMs taken from a single event can bias expected behaviors, as 
well as harm performance criteria of certain bridges that may have a unique response to 
the over selected event.  It was decided to limit the ground motions a single event could 
contribute to two.  This condition provided a decent reduction in the set of ground 
motions. 
The next parameter included in GM elimination was the analysis of response 
spectrum produced by each GM.  The behavior of each motion?s RS was examined, 
especially between a natural period of 0.5 and 3 seconds.  Values for these criteria were 
not specifically defined, and ground motions were eliminated based on engineering 
judgment.  The main factor involved in this selection process was the compilation of a 
diverse response spectrum.  Similar responses resulted in ground motions being 
eliminated, as well as responses deemed too weak or two severe to be feasibly 
incorporated into this analysis.  Any GM that needed to be scaled by a factor less than 0.5 
or greater than 2 was also eliminated.  Table 3.3 displays some of the properties 
associated with the ten GMs selected for this analysis.  The RS of each selected GM can 
be seen in Figure 3.6 and Figure 3.7, as well as their average and the DRSs for the two 
hazard sites in Alabama.  Ground motions that represent the hazard present in Alabama 
 
34 
 
can be difficult to acquire, due to a scarcity of recorded ground motions in geological 
conditions found in the Eastern United States.   
Table 3.3: Ground Motion Data 
NGA 
Event #  Event                Year   Station                 
 
Magnitude    Mechanism        
 Rupture 
distance 
(km)   Vs30(m/s)  
68 
 San Fernando 
1 1971 
 LA - Hollywood 
Stor FF  6.61  Reverse          22.8 317 
70 
 San Fernando 
2 1971  Lake Hughes #1          6.61  Reverse          27.4 425 
186  Imperial Valley  1979 
 Niland Fire 
Station     6.53  Strike-Slip      36.9 208 
333  Coalinga-01         1983 
 Parkfield - 
Cholame 8W  6.36  Reverse          51.8 257 
512 
 N. Palm 
Springs     1986 
 Anza - Tule 
Canyon      6.06 
 Reverse-
Oblique  52.1 685 
832  Landers             1992  Amboy                   7.28  Strike-Slip      69.2 271 
1741 Little Skull Mnt.           1992 
Station #2 NTS 
Control Point     5.65  Normal      24.7 660 
1147 
 Kocaeli- 
Turkey 1 1999  Ambarli                 7.51  Strike-Slip      69.6 175 
1165 
 Kocaeli- 
Turkey 2 1999  Izmit                   7.51  Strike-Slip      7.2 811 
1107  Kobe 1999 Kakogawa              6.9  Strike-Slip      22.5 312 
 
 
35 
 
 
Figure 3.6: Design-Level Response Spectra 
0 1 2
0
0.2
0.4
0.6
0.8
S a n  F e r n a n do- L A
S a n  F e r n a n do- L a ke  H u g h e s
I m p e r i a l  V a l l e y - 06
C O A L I N G A - 0 1
N  P a l m  S pr i n g s
L a n de r s - A m bo y
L i t t l e S ku l l M n t - r pr t
K o c a e l i , T u r ke y - A m ba r l i
K o c a e l i - r pr t
D e s i g n  S pe c t r u m  B r idg e po r t
S c a l e d G M  A V R G
K o be - r p r t
D e s i g n  S pe c t r u m  M u s c l e  S h oa l s
De s i gn S c al e d  R e s p ons e Sp ec t ra
N a t u r a l  P e r i od  ( s )
A
c
c
e
l
e
r
a
t
i
on (
g)
 
36 
 
 
Figure 3.7: MCE-Level Response Spectra 
 
An additional two ground motions were also initially included in the selected 
GMs, but have subsequently been eliminated.  These ground motions were synthetic 
ground motions created as a prediction of future ground motions associated with a large 
scale New Madrid fault event.  These GMs used soil conditions and behavior associated 
0 1 2
0
0.2
0.4
0.6
0.8
S a n  F e r n a n do L A
S a n  F e r n a n do- L a ke  H u g h e s
I m p e r i a l  V a l l e y - 06
L a n de r s - A m bo y
K o c a e l i , T u r ke y - A m ba r l i
K o c a e l i - r pr t
M C E  S p e c t r u m  B r i d g e por t
S c a l e d A V G G r ou n d M ot i o n
K o be - r p r t
M C E  S p e c t r u m  M u s c l e  S h oa l s
MCE Sc al ed R esp o nse S pec tru m
T  ( s )
S
A
 (
g)
 
37 
 
with Memphis, and the GMs indicated behavior associated with large columns of sand, a 
geotechnical feature that is not associated with the regions being analyzed in this project.  
A list of the final ground motions selected as well as their associated scaling factors can 
be seen in Table 3.4. 
 
Table 3.4: Ground Motion Scaling 
Ground Motion Design-Level Scale Factor MCE-Level Scale Factor 
 San Fernando 1 1.10 1.20 
 San Fernando 2 1.10 1.45 
 Imperial Valley  1.59 2.20 
 Coalinga-01         1.70 - 
 N. Palm Springs     1.95 - 
 Landers             1.40 1.95 
Little Skull Mnt.           1.50 - 
 Kocaeli- Turkey 1 0.65 1.10 
 Kocaeli- Turkey 2 0.95 1.50 
 Kobe 0.62 0.95 
   
3.4 Selection of Modeling Elements 
 
Computer models of existing structures, especially ones used for complicated 
analysis, require a balance of accuracy in geometry, behavior, and functionality.  A 
model that includes every detail of a structure may be considered accurate, but not 
practical for the purpose of analysis.  A structure simplified so much as to neglect 
important aspects of behavior may also be deemed impractical for analysis.  It is with 
these ideas in mind that the models for this project were designed.   
 
 
 
38 
 
3.4.1 Superstructure Modeling 
 
The bridge models were initially constructed using information taken from plans 
provided by ALDOT.  This information helps to build the initial geometry of the entire 
bridge using the Bridge Designer toolbar found in CSI Bridge.  This process includes the 
specifications of bridge spans, bents, abutments, deck geometry, girder geometry, bent 
cap dimensions, bent column dimensions and span support conditions.  Each parameter 
was imported into the bridge builder as the provided designs had indicated.  Once this is 
done a modeling option is presented that will determine the complexity of the model?s 
superstructure.  The first options present are a three dimensional meshed superstructure 
composed of smaller elements, representing each component of the deck, web walls and 
girders.  A second option is that of a spine model, a model that calculates the stiffness of 
the overall superstructure, and simplifies it into a single beam element.  The spine model 
also neglects inelastic behaviors that may be associated with large or deep elements.  
Typically elements of simply supported spans behave elastically in dynamic motions, 
elements like bridge decks and girders.  Attempts at using the complex mesh model 
resulted in slow analysis with an eventual failure.  This led to acceptance of the 
simplified spine model.  The overall performance of the spine model may not contain the 
accuracy present in the meshed three-dimensional model, but the meshed model also 
accounts for and records information that would not be used at the conclusion of the 
analysis.  Superstructure behavior was also not a cause for concern in this analysis 
compared to the behavior of the substructure and connections.  Thus it was acceptable to 
use a spine approximation for modeling of the bridges in this project. 
 
39 
 
Bridges that are designed as simply supported typically contain expansion joints 
at support locations.  These joints consist of a small gap between spans that allows for the 
expansion and contractions due to thermal and long term displacements as well as 
rotations at the end of the spans.  For modeling considerations these elements needed to 
have their behavior represented due to their importance in longitudinal motions.  This 
was accomplished by placing gap elements between span elements.  These gap elements 
are specified as linear elements that only activate after they have compressed a certain 
distance.  Once the springs are activated they function like a compression spring with a 
stiffness of 5000 kips/in.  This stiffness is meant to transfer a large force for a small 
amount of compressive displacement over the entire span element.  A similar procedure 
was also performed for the expansion joint located at the abutments. 
 
3.4.2 Substructure Model 
 
A primary focus of this project is the observation and analysis of substructure 
behavior under the effects of a design-level GMs.  With this in mind special precautions 
were taken to ensure accurate modeling and behavioral considerations pertaining to the 
substructure of each bridge, beginning with the foundation.  The analytical models 
incorporate geotechnical properties of each bridge in the form of multi-linear 
foundations.  These foundation springs have been calculated using a static pushover 
analysis detailed in work done by Kane (2013).  Stiffness associated with each of the six 
degrees of freedom is modeled at each foundation location.  These multi-linear springs do 
not contain a specific failure limit, due to the behavior of the spring elements in 
CSIBridge.  The foundation displacements must be noted for final analysis in order to 
 
40 
 
determine whether the displacements and rotations exceeded those generated in the soil-
structure interaction model.   
Columns represent the next aspect of the bridge models that needed to be 
carefully modeled.  Column behavior in bridges is highly dependent on accurate detailing 
of the column and its connections to other elements.  For this analysis it was assumed that 
detailing in the columns is consistent with that specified by AASHTO.  Accurate 
detailing within columns results in ductile plastic hinging at points determined by 
AASHTO.  These hinges and their behavior are accounted for in the bridge models using 
the ?define hinge function?.  Hinges are placed within the columns at select points and 
with defined lengths.  Hinges are specified to account for multiple directions of behavior, 
as well as different controlling behavior limits.  Shear hinges behave in the directions of 
shear and act as a sudden, non-ductile failure.  Shear hinges were considered later on in 
this project, but for a different element.  The behavior needed for plastic hinges in 
columns results from a ductile interacting flexural-axial hinge.  This hinge type accounts 
for rotational and axial behavior within the element.  This hinge type was selected for the 
purpose of this analysis; however this particular hinge type is subdivided further by 
which directional movement it should include as well as which method is used for 
behavior determination.  A biaxial direction was selected due to its more accurate 
application in a model that is subjected to dynamic GMs in two different global 
directions.  The selection of behavioral analysis method for the column is a more 
complex issue that required supplementary study and testing. 
The first option that was applicable for hinge analysis was the use of a Caltrans 
specified hinge.  This particular hinge type uses a simplification to determine various 
 
41 
 
capacities of a plastic hinge in an effort to alleviate design computations as well as 
provide a conservative estimation of hinge behavior.  The backbone curve for a Caltrans 
Hinge can be seen in Figure 3.8 and can be observed to display a linear elastic behavior at 
low rotations, followed by a sequence of inelastic behaviors meant to represent yielding 
and eventual failure of a hinge.  It is recommended that significant hysteresis should be 
avoided when using Caltrans Hinges in dynamic analysis (Computers & Structures inc. 
2012).  An analysis was performed on a simple reinforced concrete moment frame using 
Caltrans Hinges.  The frame was loaded with a lateral force and results were measured to 
determine the applicability of the Caltrans Hinge compared to the second hinge analysis 
option, the Fiber Hinge. 
 
Figure 3.8: Idealized Caltrans Hinge 
 
A Fiber Hinge analysis is a type of finite element analysis used in the 
determination of hinge behaviors and capacities.  It operates under the theory that a real 
structural element can be modeled as a collection of smaller individual elements, and that 
0 
200 
400 
600 
800 
1000 
1200 
1400 
1600 
0 0.0005 0.001 0.0015 
M
om
ent
 (K
ip-
ft)
 
Plastic Rotation (Rad) 
Origin 
Yielding 
Ultimate Capacity 
Residual Strength 
Total Failure 
 
42 
 
the smaller the individual elements get, the more accurate the behavior of the collection 
of elements becomes.  The fiber hinge model is the more complex option among the two 
hinge analyses.  Sections of the cross section of a member are divided into sections of 
concrete and steel.  Each section is analyzed based on the material properties and location 
of each fiber in order to determine cross-section behavior and capacities at given loading 
conditions.  Figure 3.9 displays a backbone curve resulting from a fiber hinge analysis.  
?The fiber-hinge model is more accurate in that the nonlinear material relationship of 
each fiber automatically accounts for interaction, changes in moment-rotation curve, and 
plastic axial strain. A trade-off is that fiber application is more computationally 
intensive? (Computers & Structures inc., 2012). 
  An analysis was performed on a simple reinforced concrete moment frame using 
plastic hinges.  The frame was loaded with a lateral force and results were compared to 
that of the Caltrans Hinge.  The resulting capacities were similar, indicating that both 
hinges estimated similar strengths; however, the displacements indicated that the fiber 
hinge was a much more rotationally flexible hinge.  It was determined that both hinge 
types were viable and later testing would show that fiber hinges resulted in more 
favorable numerical analysis behavior. 
 
43 
 
 
Figure 3.9: Fiber Analysis Backbone Curve 
 
In order to use a plastic hinge in modeling, characteristics of the hinging element 
need to be specified in detail.  A typical frame element in CSI Bridge and SAP 2000 can 
be specified with just cross-section geometry, concrete strength, reinforcement layout and 
steel properties.  An accurate hinge will require a more in-depth property specification.  
The Section Designer tool was used to create columns in the modeling for this project.  
Column designs included specification of longitudinal and shear reinforcement as well as 
spacing and clear cover distances.  Concrete was subdivided into confined and 
unconfined sections to be analyzed using the Mander-Confined and Mander-Unconfined 
concrete properties.  An axial load was also specified on the columns in the creation of a 
backbone curve for each section.  These properties help to establish nonlinear behavior 
where nonlinear behavior is expected.   
0 
10000 
20000 
30000 
40000 
50000 
60000 
70000 
0 0.0005 0.001 0.0015 0.002 0.0025 0.003 0.0035 0.004 
M
om
ent
 (K
ip-
in)
 
Curvature (?) 
 
44 
 
Bent caps and abutments, the last remaining substructure elements, were modeled 
as elastic frame elements.  They are assumed to have adequate detailing and design 
strength to avoid any nonlinear hinging that may cause instabilities in the bridge 
structure. The size of these elements depth and stiffness should allow them to resist 
moments transferred from the bent columns, allowing the use of this assumption.  
Abutments were also assumed to be supported with fixed foundation connections for the 
purpose of this project.  This property was determined when the geotechnical behavior of 
these elements was not considered a major concern for this project. 
3.4.3 Connection Modeling 
 
In addition to substructure behavior, connection behavior between the super- and 
substructure was also a primary focus in this analysis.  Special attention was paid to the 
behavior occurring in this region, and the impact each of these behaviors contributed to 
that of the overall bridge.  As stated in Section 2.1.2, the bridges in this project all contain 
a girder-to-elastomeric bearing pad connection at the end of each span.  This connection 
detail can be seen in Figure 3.10.  Each of these bearing pads is made from layers of 
elastomeric material interspersed with thin steel plates (steel shims).  These shims act to 
reduce bulging of the elastomeric material when subjected to vertical loads by limiting 
the thickness of each individual layer of elastomeric material.  It is with this in mind that 
all bearing pads are initially selected, meaning that the overall thickness of a single 
bearing pad is determined from the amount of vertical load that it is designed to resist.  
Additional design is performed in regards to rotational capacity of the girders along with 
bearing pad dimensions allowing for enough longitudinal girder expansion/contraction.  
 
 
45 
 
 
Figure 3.10: Example Bearing Pad Configuration (Alabama DOT, 2012) 
 
Elastomeric bearing pad thickness results in more flexible behavior during lateral 
shearing, regardless of the number of steel shims.  An elastomeric bridge bearing?s 
deformation in a lateral direction can be determined according to Equation 3.1.  The 
dimensions of each bearing pad are extrapolated from bridge design drawings, and a 
Shear Modulus (G) of 135 psi was selected from values found in a Caltrans design memo 
(Caltrans, 1994).  The deflection equation will eventually change at larger forces, but can 
be used for the purpose of modeling because the shear experienced in these bearing pads 
 
46 
 
are limited to a ceiling value determined by eventual slipping between the bearing pad 
and the girder.  A force-displacement relationship for a bearing in shear can be seen in 
Figure 3.11, with a plateau being reached at the point of slipping.   
                                                                                           Equation 3.1 
 
 
Figure 3.11: Example Bearing Pad Force-Displacement Relationship 
 
Research performed on a connection detail in use by the Illinois DOT, similar the 
one used by ALDOT, observed a friction coefficient between 0.2 and 0.5 in its bearing 
pad-steel plate friction connection; however, cyclical loading resulted in reduced values 
due to degradation of the rigidness of the bearing pad surface (Steelman, et al. 2012). 
Dynamic friction coefficients of 0.2 and 0.4 were selected for a lower and upper bound in 
this analysis. The process of implementation can be found later in this chapter.  All of 
these factors contributed to the creating of force-displacement plots for each bearing bad 
in use for this project, and these plots were implemented into the model using multi-
-20 
-16 
-12 
-8 
-4 
0 
4 
8 
12 
16 
20 
-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 
Fo
rce 
 (k
ips
) 
Displacement (in.) 
Slipping  
Elastic Behavior 
 
47 
 
elastic plastic kinetic link elements.  The ?kinetic? distinction of these link elements 
refers to the behavior these links exhibit during unloading.  A plastic link element 
recognizes unloading of a link and follows a different stiffness than the one used in the 
loading process.  The manner of which an unloading stiffness is determined varies 
depending on the type of plastic link, with ?kinetic? most closely describing the behavior 
of a friction connection.   
Each bridge model is meant to be loaded in a longitudinal direction (direction of 
travel) and a transverse direction (perpendicular to direction of travel).  The connection 
that resists motion in the longitudinal direction consists solely of a bearing pad, but there 
is an additional connection detail in the lateral direction that is also considered in the 
model.    Figure 3.12 displays a view of ALDOT?s typical connection, consisting of both 
a bearing pad and a clip angle with anchor bolts meant to limit movement in the lateral 
direction.  This clip angle system consists of steel clip angles fastened to the girder via 
small threaded inserts and fastened to either abutments or bent caps via anchor bolts.  For 
the purpose of this analysis it is assumed that the small threaded inserts that transfer 
longitudinal or tensile forces from girder to clip angle are not sufficiently embedded 
within the girder to provide any real resistance.  This configuration results in a single-
level longitudinal connection system; however, the anchor bolt provides the clip angles 
with sufficient stiffness to resist transverse forces when the movement is towards the 
angle.  The overall failure of this clip angle system is determined to occur when the 
anchor bolts have experienced nominal shear strength capacity.  This limit state was 
deemed conservative, and also selected above a failure state of the clip angles themselves 
due to the girders eventual collision with the anchor bolt in the event of a clip angle 
 
48 
 
failure.  Equation 3.2 was used to determine the shear capacity of an anchor bolt, 
assuming adequate embedment was provided for the bolt.   
 
Figure 3.12: Clip Angle Connection Detail 
 
                                                                                             Equation 3.2 
 
Modeling the anchor bolt?s contribution to the total bridge model connection 
proved to be a difficult challenge.  Several options were attempted that appeared accurate 
in both behavior and geometry.  The first option was the creation a frame element that 
acted as a fuse element, failing in shear.  The frame element would act in parallel with the 
bearing pad link, and contain a shear hinge.  This initial option was modified to 
incorporate two frame elements acting in parallel with one another, and connected to the 
bearing pad link via gap links.  These gap links allow only one frame element to be 
 
49 
 
loaded in shear for each direction of transverse movement.  Theoretically this system 
would mimic the real life behavior of the bridge during transverse movement in an 
accurate way.  The implementation of this modeling option proved to behave differently 
than anticipated.  Shear hinging resulted in the element being split into two different 
elements causing instabilities in the analysis, as well as the frame elements collecting 
additional forces and moments despite releases being implemented into the system as a 
measure of preventing that behavior from occurring. 
The second option implemented for modeling the transverse restrainer connection 
was the use of multi-linear elastic links in place of frame elements.  These elements 
require force displacement data in order to implement.  Initial shear stiffness equations of 
anchor bolts resulted in an incredibly stiff system.  Implementing the calculated shear 
stiffness values into elastic links created a system with large stiffness changes, resulting 
in numerical instability within models.  These stiffness equations were then modified to 
account for displacements within the clip angle configuration, resulting in the selection of 
an overall displacement of 0.55 inches between loading and failure of all clip angle 
systems.  This displacement was also meant to create a constant between all bridge 
models.  The resulting model functioned, but still resulted in some numerical instability 
during some of the ground motions.  It was also observed that this model did not 
accurately replicate shear failure behavior in the model, and a new link element was 
selected, the multi-linear plastic link element.  This new link element has the capability to 
apply failure and strength degradation that is expected with the behavior of this 
connection; as well as continue to function after loss of strength, instead of causing a 
numerical instability.  A ?kinematic? plastic hinge was initially selected, but its behavior 
 
50 
 
did not match up with the expected behavior of the system.  A ?Pivot? plastic link model 
was selected instead, based on its behavior matching that of the expected behavior of the 
system.  The Pivot link requires additional inputs in order to shape the unloading and 
reloading properties of the link; all Pivot factors for this link were set to zero in order to 
achieve desired behavior.  The implementation of this link resulted in reduced numerical 
failures in many of the bridge analyses, and was selected as the best option for modeling 
the clip-angle connection.   
The implementation of connection elements into the bridge models required an 
additional step.  A simplification was made to combine connection elements together to 
form a single connection element in place of multiple connection models.  This 
simplification was made in order to reduce overall complexity of the models.  It was 
assumed that the bridge deck remains nearly rigid in its plane during ground motions, 
allowing the use of this simplification.  Bridges that contain bents with more than two 
columns use an altered version of this simplification, multiple simplified connections at 
each column location. 
The final step in the modeling process for each bridge is the creation of behavioral 
limits.  The focus of the analysis in this project is centered on bridge behavior, but certain 
information cannot be accurately known.  This fact resulted in the creation of two 
different versions of each bridge model, a lower limit model and an upper limit model.  
The principal difference between each of these models is the friction coefficient between 
bearing pad and girder surface.  The lower bound model assumes slipping to occur at a 
friction coefficient of 0.2.  The upper bound model assumed a coefficient of 0.4.  The 
actual behavior of the bridge should fall somewhere between these two models.  Another 
 
51 
 
important range of uncertainty occurs in the prediction of system damping present within 
each bridge model.  A method for determining and specifying a damping factor was 
established and implemented based on Rayleigh damping.  The theory behind this 
damping method can be found in (Chopra 2007).  A model analysis was performed for 
each bridge model, with certain elements altered to remain stiff or flexible, and natural 
periods were recorded at the first and third mode shape for movement in lateral and 
longitudinal directions.  These natural periods were correlated to a damping factor of 2% 
and entered into a CSI table that calculated Raleigh damping factors.  These factors are 
applied to the model during the analysis procedure.   
3.5 Analysis Method 
 
The Time-History Analysis method was selected for use in this project.  This 
method of analysis functions by running a model, composed of stiffness and mass 
properties, through a series of subsequent analyses.  These analyses involve the 
application of a single acceleration value to the model and then calculating the models 
response to that acceleration using the equations of motion.  The resulting model is then 
subjected to a subsequent acceleration that represents the next acceleration value of a 
given time history.  This process is repeated until the model has experienced the entire 
acceleration time history.  A Hilber-Hughes-Taylor Time Integration Method was used in 
order to apply each ground acceleration history to the models. Time history analysis 
results are recorded as a response history. 
 
 
 
52 
 
3.5.1 Pre-Analysis Procedure 
 
Before the Nonlinear Time History Analysis was preformed, certain steps were 
taken to ensure accurate results.  The bridge models were preloaded with self weigh 
loading in order to account for p-delta effects.  A nonlinear Time History Analysis is able 
to incorporate p-delta effects into its analysis; thus this loading step results in the 
consideration of p-delta behaviors.  An additional 20 seconds of zero acceleration was 
added to the end of each GM used in analysis, in order to observe the free vibration of 
each bridge after an event.  This allowed for the observation and recording of residual 
response in each bridge.  With these steps preformed analysis was then able to be 
performed. 
3.5.2 Post-Analysis Procedure 
 
After analysis was complete, steps needed to be taken to ensure completed results.  
All analyses were examined to have run through the entirety of their associated ground 
motions.  If the examined results did not result in a complete analysis, their associated 
model was examined for adequate modal behavior.  If this check did not indicate a reason 
for failure, the nonlinear analysis parameters were altered and the analysis was performed 
again.  This process was iterated until a completed response was generated or consistent 
incomplete analyses were observed.  A final alteration of the time integration method was 
considered and applied for a few cases that did not result in complete analysis. 
3.6 Overview of Bridge Models 
 
The following section will include the details used to create each bridge model, 
and an overview of the modifications and features of each bridge.  Simplifications will be 
 
53 
 
explained and justified.  Expected behavior is also included in the description of each 
bridge as well as the bridge?s natural periods taken from modal analysis.   
3.6.1 Little Bear Creek Bridge 
 
This bridge carries the two lanes of State Road 24 over Little Bear Creek in 
Franklin County.  It is a three span bridge with spans of unequal lengths. The outer span 
lengths are 85 feet and the interior span is 130 feet. The outer spans support the 7 inch 
concrete deck with six Type III Girders and the interior span supports the deck with six 
BT-72 Girders.  Each of the six Type III girders rests upon 20.5? x 9? bearing pads. 
These bearing pads were calculated to slip after an average (0.3 friction coefficient) 
displacement of 0.567? correlating to a shear of 18.15 kips per bearing pad.  The clip-
angle/anchor bolt configuration for the transverse connection contained anchor bolts with 
a diameter of 1.25?.  These bolts were modeled with a shear capacity of 45 kips.  Each of 
the six BT-72 girders rests upon 24.5? x 9? bearing pads. These bearing pads were 
calculated to slip after an average (0.3 friction coefficient) displacement of 1.185? 
correlating to a shear of 33 kips per bearing pad.  The clip-angle/anchor bolt 
configuration for the transverse connection contained anchor bolts with a diameter of 
1.5?.  These bolts were modeled with a shear capacity of 61 kips.  A simplification was 
also modeled for this bridge using a redesigned clip-angle/anchor bolt configuration for 
the transverse connection at each span which contained anchor bolts with a diameter of 
1.75?.  These bolts were modeled with a shear capacity of 78 kips. The two bridge piers 
are 40? x 5? x 7? and supported by two circular columns 4.5 feet in diameter with 3 inches 
of concrete cover. The columns are reinforced longitudinally with 24 #11 bars and 
transversely with #5 hoops uniformly spaced at 10 inches from the bottom of the hinge 
 
54 
 
zone to the top of the foundation, and spaced at 6 inches inside the hinge zone. The 
average clear height of Bent 2 is 12.06 feet and 16.88 feet for Bent 3. All columns are 
supported on drilled shafts 5 feet in diameter. The concrete cover of the drilled shafts is 6 
inches but the longitudinal reinforcement in the drilled shaft still aligns with the 
longitudinal reinforcement of the column.   
This bridge model contains a single girder connection to represent six girder 
connections.   The selection of a single connection representing six total connections 
results in a miscalculation of the gravity moments imparted into the bent columns in a 
conservative fashion.  A non-rectangular member was selected to model the bent caps, 
and a gap link was established to model possible impact stiffness between the deeper 
girder and this abutment.  Plastic hinges were modeled at the top and bottom column 
locations.  The CSiBridge model for this bridge can be seen in Figure 3.13. 
 
55 
 
 
Figure 3.13: Little Bear Creek Bridge Model 
 
The overall model of this bridge was found to be moderately stiff in the context of 
the other bridges modeled due to shorter columns and stiff foundations.  Differences in 
mass between spans could result in secondary mode shapes dictating dynamic behavior in 
the transverse direction.  Modal analysis of the model revealed a first mode natural period 
of 0.470 seconds in the transverse direction and 0.818 seconds in the longitudinal 
direction.  The third mode natural periods of the fixed version of this bridge resulted in a 
natural period of 0.140 seconds in the transverse direction and 0.164 seconds in the 
longitudinal direction.  The ?fixed version? bridge model is a version of the bridge model 
in which all bearing pad connections and all expansion joints are fixed, and these 
connections do not allow movement along any degree of freedom.  This ultra stiff bridge 
model attempts to maximize the possible stiffness of the bridge?s geometry, resulting in 
the lowest possible natural period obtainable from the bridge.  The third mode natural 
 
56 
 
period of the ?fixed version? of the bridge represents the lowest acceptable natural period 
the bridge could obtain, and this number is used as a stiffness boundary for the 
calculation of damping.   Foundation behavior for this structure could result in hinging 
forming in the foundations, but that cannot be predicted in this particular model which 
specifies hinges in the column only.   
3.6.2 Oseligee Creek Bridge 
 
The Oseligee Creek Bridge is a two lane bridge that carries County Road 1289 
over Oseligee Creek in Chambers County. It is a three span bridge with equal span 
lengths of 80 feet. The 7 inch concrete deck is supported by four Type III girders.  Each 
of the four girders rests upon 20.5? x 9? bearing pads. These bearing pads were 
calculated to slip after an average (0.3 friction coefficient) displacement of 0.66? 
correlating to a shear of 21 kips per bearing pad.  The clip-angle/anchor bolt 
configuration for the transverse connection contained anchor bolts with a diameter of 
1.75?.  These bolts were modeled with a shear capacity of 60.9 kips.   The two bridge 
piers are 30? x 4? x 5? and supported by two circular columns 3.5 feet in diameter with 3 
inches of concrete cover. The columns are reinforced longitudinally with 12 #11 bars and 
transversely with #4 hoops uniformly spaced at 12 inches from the bottom of the hinge 
zone to the rock line, with the hinge zone being reinforced with #4 ties at 6 inch spacing. 
The average clear height of Bent 2 is 17.93 feet and 25.83 feet for Bent 3.  All columns 
are supported on drilled shafts 3.5 feet in diameter with concrete cover of 3 inches.  
This bridge model contains a single connection to represent four girder 
connections.   The selection of a single connection representing four total connections 
results in miscalculated gravity moments imparted into the bent columns in a 
 
57 
 
conservative fashion.  Plastic hinges were modeled at the top and bottom column 
locations.  The CSIBridge model for this bridge can be seen in Figure 3.14. 
 
Figure 3.14: Oseligee Creek Bridge Model 
 
The overall model of this bridge was found to be moderately flexible in the 
context of the other bridges modeled due to a 100% scour condition that was selected by 
geotechnical analysis of the bridge.  Initial modal analysis of this bridge shows stiff 
bridge behavior, however bridge geometry of the model results in more flexible behavior 
after hinge formation.  This behavior combined with second order effects could result in 
column buckling, however a bridge with these conditions helps to diversify the overall 
bridge suit.   Modal analysis of the model revealed a first mode natural period of 0.689 
seconds in the transverse direction and 1.066 seconds in the longitudinal direction.  The 
third mode natural periods of the fixed version of this bridge resulted in a natural period 
of 0.154 seconds in the transverse direction and 0.262 seconds in the longitudinal 
 
58 
 
direction.  The combination of tall, flexible substructure and short span lengths could 
result in dissimilar behaviors in super and substructure, resulting in possible higher forces 
at the connection location.  Foundation behavior for this could result in hinging forming 
in the foundations, but that cannot be predicted in this particular model which specifies 
hinges in the column only.  Large foundation rotations could result due to second order 
effects as well as large moments occurring in the substructure.   
3.6.3 Norfolk Southern Railroad Bridge 
 
The bridge over Norfolk Southern Railroad is the southbound I-59 bridge in 
Etowah County, a two-lane bridge that crosses over a Norfolk Southern railroad line and 
a state highway.  It is a two-span bridge with unequal span lengths of 125 feet and 140 
feet.  Nine modified BT-54 girders support a 6 inch concrete deck that is 46.75 feet wide.  
Each of the nine girders rests upon 24.5? x 9? bearing pads.  These bearing pads were 
calculated to slip after an average (0.3 friction coefficient) displacement of 1.132? 
correlating to a shear of 31.5 kips per bearing pad.  The clip-angle/anchor bolt 
configuration for the transverse connection contained anchor bolts with a diameter of 
1.375?.  These bolts were modeled with a shear capacity of 48 kips.  The only bridge pier 
is 53? x 4.5? x 4? and supported by three square columns 3.5 feet in width.  The columns 
are reinforced longitudinally with twelve #11 bars and transversely with #4 ties uniformly 
spaced at 9 inches from the bottom hinge zone to the top hinge zone, with a spacing of 4 
inches inside each hinge zone.  The average clear height of the columns is 25.25 feet.  
The bridge is supported on driven piles. The pile cap is 8.5? x 8? x 4.5? and each pile cap 
is supported by 7 HP 12x53 steel piles.    
 
59 
 
This bridge model contains three girder connections to represent nine girder 
connections.   The selection of multiple connections representing nine total connections 
results in the miscalculation of gravity moments imparted into the bent columns in a 
conservative fashion.  Plastic hinges were modeled at the top and bottom column 
locations.  The CSiBridge model for this bridge can be seen in Figure 3.15. 
 
Figure 3.15: Norfolk Creek Bridge Model 
 
The overall model of this bridge was found to be initially stiff due to the large 
column diameters and the strut members; however bridge geometry of the model results 
in more flexible behavior after hinge formation.  Modal analysis of the model revealed a 
first mode natural period of 0.415 seconds in the transverse direction and 0.705 seconds 
in the longitudinal direction.  The third mode natural periods of the fixed version of this 
bridge resulted in a natural period of 0.133 seconds in the transverse direction and 0.086 
 
60 
 
seconds in the longitudinal direction.  The combination of short, stiff substructure and 
large span lengths could result in dissimilar behaviors in super and substructure, resulting 
in possible higher forces at the connection location.  Foundation behavior for this bridge 
is difficult to predict, but a triple column bent should limit foundations rotations in the 
lateral direction.   
3.6.4 Scarham Creek 
 
The Scarham Creek Bridge is two lanes and carries State Route 75 over Scarham 
Creek in Marshall County. It is a four span bridge with equal span lengths of 130 feet. The 
7 inch concrete deck is supported by six BT-72 girders.  Each of the six girders rests upon 
24.5? x 9? bearing pads. These bearing pads were calculated to slip after an average (0.3 
friction coefficient) displacement of 1.336? correlating to a shear of 37.5 kips per bearing 
pad.  The clip-angle/anchor bolt configuration for the transverse connection contained 
anchor bolts with a diameter of 2.5?.  These bolts were modeled with a shear capacity of 
87 kips.  The bridge pier at bents 2 and 4 are 40? x 5.5? x 7.5? and the pier at bent 3 is 40? 
x 6.5? x 7.5.? Bents 2 and 4 are supported by two circular columns 5 feet in diameter with 
3 inches of concrete cover. Bent 3 is supported by two circular columns 6 feet in diameter 
with 3 inches of concrete cover.  All columns are reinforced transversely with #6 hoops at 
a spacing of 6 inches at the top and bottom hinge zones, and spaced at 12 inches at all 
other regions.  All columns are supported on drilled shafts, which are six inches larger in 
diameter than the columns. It is assumed that the plastic hinge will form at this transition, 
so the clear height of the columns is measured from the bottom of the bent cap to the 
transition between the column and drilled shaft. The average height of columns is 34.02 
feet at Bent 2, 59.17 feet at Bent 3, and 32.16 feet at Bent 4.  Because of the height of the 
 
61 
 
columns, struts are provided at approximately mid-height of the columns and span the 
full length between columns with a thickness of 3.5 feet.  The strut at bents 2 and 4 are 6 
feet deep and 10 feet deep at bent 3.   
This bridge model contained a single connection to represent six girder 
connections. The selection of a single connection representing six total connections 
results in the miscalculation of gravity moments imparted into the bent columns in a 
conservative fashion.  Plastic hinges were modeled at the column locations above and 
below strut locations due to the large depth and stiffness of the struts.  Behaviors at these 
locations are of particular concern during ground motions.  The CSiBridge model for this 
bridge can be seen in Figure 3.16. 
 
Figure 3.16: Scarham Creek Bridge Model 
 
The overall model of this bridge was found to be initially stiff due to the large 
column diameters and the strut members; however bridge geometry of the model results 
in more flexible behavior after hinge formation.  Modal analysis of the model revealed a 
first mode natural period of 0.760 seconds in the transverse direction and 1.080 seconds 
in the longitudinal direction.  The third mode natural periods of the fixed version of this 
 
62 
 
bridge resulted in a natural period of 0.300 seconds in the transverse direction and 0.340 
seconds in the longitudinal direction.  The combination of tall substructure and large span 
lengths could result in similar behaviors in super and substructure, resulting in possible 
lower forces at the connection location.  Foundation behavior for this bridge is difficult to 
predict, given the unpredictability of this bridges behavior as it enters the inelastic range.   
3.6.5 Bent Creek Road 
 
The Bent Creek Road Bridge in Lee County is a five-lane bridge that crosses over 
Interstate 85 with two spans of 135 feet. Each span is comprised of fifteen modified BT-
54 girders spaced approximately 5.33 feet apart that support a 6 inch concrete deck that is 
80.75 feet wide. The only bridge pier is 79? x 4? x 4.5? and supported by five square 
columns 3.5 feet in width. The columns are reinforced longitudinally with 12 #11 bars 
and transversely with #4 ties uniformly spaced at 9 inches from the bottom of the bent to 
the top of the pile cap foundation, except for plastic hinge zone region which are 
reinforced with #4 ties at a spacing of 4 inches. The average clear height of the columns 
is 20.1 feet. The bridge is supported on driven piles. The pile cap is 8.5? x 8? x 4.5? and 
each pile cap is supported by 9 HP 12x52 steel piles. Each of the fifteen girders rests 
upon 24.5? x 9? bearing pads.  These bearing pads were calculated to slip after an 
average (0.3 friction coefficient) displacement of 1.1855? correlating to a shear of 33 kips 
per bearing pad.  The clip-angle/anchor bolt configuration for the transverse connection 
contained anchor bolts with a diameter of 1.75?.  These bolts were modeled with a shear 
capacity of 61 kips. 
This bridge model contained five modeled connections to represent fifteen girder 
connections.  This simplification was chosen due to the bridge bent containing five 
 
63 
 
columns, instead of a configuration of two columns found in other bridges.  A selection 
of a single connection representing fifteen total connections would result in dramatically 
miscalculating the gravity loads and moments imparted into the bent columns.  The 
selection of five connections resulted in a more complex model then the other bridges, 
and it was predicted that analysis problems could arise.  The CSIBridge model for this 
bridge can be seen in Figure 3.17. 
 
Figure 3.17: Bent Creek Bridge Model 
 
The overall model of this bridge was found to be rather stiff, despite its large 
mass.  Modal analysis of the model revealed a first mode natural period of 0.717 seconds 
in the transverse direction and 0.933 seconds in the longitudinal direction.  The third 
mode natural periods of the fixed version of this bridge resulted in a natural period of 
0.108 seconds in the transverse direction and 0.133 seconds in the longitudinal direction.  
 
64 
 
The combination of stiff substructure and large superstructure mass could result in large 
forces at the connection location.  Foundation behavior for this bridge may also be 
inconsistent to the other bridges given a pile cap foundation group located under such a 
large and stout bridge bent. 
3.7 Determination of Capacities 
 
Overall dynamic bridge performance is measured by displacements and internal 
member forces as they relate to specified capacities.  Some of these capacities are 
specified in design documents and codes, and others vary from project to project.  This 
section will provide details regarding capacities selected for use in this project. 
Foundation information provided for this project was analyzed using the 
geotechnical analysis software FB-Multipier (BSI, 2013).  This information included a 
displacement limit on all foundation degrees of freedom.  These limits could not be 
implemented in the modeling of the bridges, but they are used in the bridge evaluation 
procedure after analysis.  The horizontal displacement and rotation capacities for each 
foundation are displayed in Table 3.5.  Any recorded foundation rotation or displacement 
exceeding these limits will be classified as a foundation failure. 
 
 
 
 
 
 
 
65 
 
Table 3.5: Foundation Capacities 
 
Foundation Displacement and Rotation Limits 
 
Transverse Direction Longitudinal Direction 
 
Displacement 
(in.) 
Rotation 
(rad.) 
Displacement 
(in.) 
Rotation 
(rad.) 
Little Bear Creek 0.54 0.005 0.52 0.0054 
Norfolk RR 2 0.022 3 0.0072 
Scarham (Bent 1 & 3) 0.58 0.002 0.55 0.002 
Scarham (Bent 2) 0.71 0.003 0.7 0.003 
Oseligee 0.22 0.0032 0.23 0.0033 
Bent Creek 1.4 0.035 1.4 0.036 
 
Bearing pad capacity is determined in terms of displacement.  Although there 
exists a limit of shear force that each bearing pad is capable of resisting, this limit does 
not intrinsically represent a failure criterion of the bearing pads, it simply represents the 
force in which slipping occurs within the system.  The actual capacity that is of concern 
in this element is the displacement that occurs between the girder and the bearing pad.  It 
was determined that the limit for differential movement between these two elements be 
set to the dimension of the bearing pad in that direction.  Any deflection beyond this limit 
would result in unseating of the girder. 
Bolt strength capacity was conservatively estimated using Equation 3.2.  Due to 
analysis concerns the limit specified for a bolt failure was selected as 95% of the 
calculated value.  This 95% attempts to account for any values not recorded by the 
analysis software that may trigger failure of the bolt link, and is not meant to account for 
any loss of strength or safety factor correlated to a failure mode.  Failure of the clip 
angle-anchor rod system does not result in overall structural failure, but its performance 
was monitored due to its importance in connection behavior.  Table 3.6 contains all 
calculated bolt strengths for the bridge designs selected. 
 
66 
 
Table 3.6: Calculated Bolt Strengths 
  Bolt Shear Capacities 
  Individual Bolts (kips) 
Little Bear Creek 61 
Norfolk RR 48 
Scarham 61 
Oseligee 78 
Bent Creek 87 
 
A commonly recorded criterion of seismic behavior in structures and structural 
elements is ductility.  Ductility in structures can refer to a variety of terms; however, in 
this analysis, it refers to the degree in which a structure can deform relative to its initial 
yield.  The ductility of bridge bents was selected as an indicator of overall bridge 
performance.  In order to determine the overall ductility of a bridge bent, a state of initial 
yielding needed to be determined within each bridge bent.  A backbone curve was created 
for each fiber hinge used in the bent columns, and a point of nominal strength was 
isolated.  The rotation that correlated to this point was recorded and used in Equation 3.3 
to determine the angle of rotation of the plastic hinges.  This calculated angle of rotation 
was applied at midpoint of each hinge location in each bridge bent, and the overall 
displacement differential between the center of the bent cap and the foundation was 
calculated.  This displacement was used as a measure of a ductility of 1.  A maximum 
ductility capacity was also determined for each bridge based on pushover analysis 
performed in a previous task of this project (Law, 2013).  The initial ductility 
displacements and the maximum ductility capacities can be seen in Table 3.7.  AASHTO 
specifies a range of acceptable ductility demands for bents in section 4.9, with a ductility 
limit of six or less for bents consisting of multiple columns.  The lesser of AASHTO and 
 
67 
 
Law?s ductility limit was selected as a maximum capacity for ductility in the bridge 
bents.    
                                                                      Equation 3.3 
Table 3.7: Bent Displacement Limits 
  Bent Displacements 
  Transverse Direction Longitudinal Direction 
  Ductility of 1 (in.) Pushover Limit (in.) Ductility Limit Ductility of 1 (in.) Pushover Limit (in.) Ductility Limit 
Little Bear Creek (Bent 1) 0.56 2.70 4.82 0.56 1.60 2.85 
Little Bear Creek (Bent 2) 0.91 13.5 6.00 0.91 2.52 2.77 
Norfolk RR 1.71 4.00 2.32 1.71 6.63 3.89 
Scarham (Bent 1 & 3) 1.00 9.80 6.00 2.05 2.20 1.08 
Scarham (Bent 2) 3.30 25.6 6.00 5.59 3.60 0.64 
Oseligee 0.41 6.75 6.00 0.41 6.30 6.00 
Bent Creek 1.79 2.45 1.37 1.79 4.57 2.56 
 
Lastly a serviceability limit was established for the residual displacements of the 
models.  Determining the state of functionality of a bridge after being subjected to a 
ground motion is a key factor in performance classification.  Critical and essential bridges 
should be open to emergency vehicles and be open for security and defense purposes 
immediately after the design earthquake, with critical bridges also being open to all 
traffic (AASHTO 2009).  The value for residual span displacements that was selected as 
an upper limit was one inch in either direction.  This value was determined from 
engineering judgment.  A residual displacement exceeding one inch could indicate 
substructure damage or a localized failure at a connection zone.   
This concludes the third Chapter of this thesis.  This chapter explained the process 
of bridge and ground motion selection for use in the Time History analysis that was 
performed in this research.  Ground motion scaling was described and the final scaling 
 
68 
 
values selected were presented.  An overview of the detailing of each bridge was given, 
along with information used in the modeling process in the form of strength and 
deflection capacities.  Each of these pieces of information were then used to create and 
analyze bridge models, and the results of these analyses can be found in the next chapter. 
 
 
69 
 
Chapter 4: Analysis Results and Discussion 
 
4.1 Introduction 
 
This section will present the results from the time history analyses performed on 
each bridge model.  Some sections will compare the results taken from specific bridges, 
while other sections will simply provide an overview of observed behaviors and trends.  
Unique behavior or abnormalities will also be examined.  Explanations for observed 
phenomena will also be provided when necessary.  This section will provide insight into 
the bridges analyzed as well as the ground motions used in the analysis. 
4.1.1 Description of Results Presented 
 
Each of the five bridges was analyzed in four ways.  Each bridge was modeled as 
an upper limit bearing pad configuration (coefficient of friction of 0.4), and a lower limit 
bearing pad configuration (coefficient of friction of 0.2).  Each bridge model incorporated 
soil interaction springs that provide foundation flexibility.  Both of these bridge models 
were then analyzed using ten design-level ground motions and seven MCE-level ground 
motions, applied in both the longitudinal and transverse direction.  The results of each 
analysis have been recorded and graphed using MathCAD (PTC, 2007).  Maximum span 
and bent displacements are recorded for each bridge analysis, as well as maximum 
foundation displacements and rotations.  Bolt shears and bearing displacements are also 
recorded at each connection location, with the largest value being used to calculate 
demand capacity ratios.  Each response history analysis performed is categorized based 
 
70 
 
on the results recorded.  Analyses that did not result in completion were rerun using 
altered analysis constraints in an effort to eliminate numerical instabilities that occur 
during analysis.  After numerous iterations of attempted analysis, motions that did not 
result in completion are recorded in their state of farthest completion.  The analysis is 
classified as one of the following: 1) Complete: indicating that the entire analysis was 
completed, 2) Numerical: indicating that no elements appeared to have failed but the 
analysis could not be completed due to numerical convergence issues and 3) Failure: 
indicating a numerical failure that either results from a structural failure within the model, 
or a set of recorded results that indicated possible future structural failure of the model 
during continued analysis.  Data taken from analyses deemed ?Numerical? is excluded 
from statistical analysis of each bridge due to the unknown behaviors of the model during 
analysis.  Data taken from analyses deemed ?Failure? is included in statistical analysis of 
each bridge given predictable behaviors or data related to failures that have already 
occurred.  All major structural component failures are recorded and highlighted for each 
analysis.  Foundation information is displayed as a demand/capacity value in reference to 
displacements and rotations.  The demand/capacity value is taken from each analysis as 
the maximum demand/capacity value out of each of the bents for the entirety of each 
ground motion.  The value associated with ?Bearing Capacity? represents a 
demand/capacity value that contains the highest bearing pad lateral or longitudinal 
displacement recorded from the displacements of every bearing pad in the bridge during 
the entirety of the ground motion.  ?Bent Ductility? is a value of the largest differential 
displacement between bent displacement and its respective foundation displacement.  The 
value representing ?Bent Ductility? is the quotient of the aforementioned displacement 
 
71 
 
divided by the displacement determined as the bent displacement relating to nominal 
moment capacity occurring at a hinge in that bent.  ?Bolt Status? indicates the highest 
shear demand present in all anchor bolts within the model compared to the shear capacity 
of the anchor bolts in the model.  All values indicated in Bold font represent values that 
exceed their acceptable limits. 
4.1.2 Limits of Bridge Behaviors 
 
Each bridge is analyzed for the purpose of determining behavior acceptability.  
This acceptable bridge behavior is based on many different criteria stemming from the 
limits specified in Section 3.6.  The average response of each model is used to determine 
that models performance.  In the event that a model contains numerous numerical 
failures, a prediction of overall behavior is be made, but recommendations for that bridge 
are not be made unless further completed analysis is performed. 
Each bridge result is checked for failures or exceedance of capacity limits.  A 
classification of each of these results is made in the event of a capacity limit being 
exceeded.  The following behaviors are classified as resulting in structural failure: 1) 
excessive span or bent deflections (values are dependent on bridges), 2) bearing pad 
capacities exceeding a value of ?1?, 3) Bent ductility exceeding the limits specified in 
Section 3.6.  A residual displacement of span or bents that exceeds 1 inch is classified as 
a failure on the grounds of a serviceability limit being exceeded.  Bolt failure is also 
addressed, but is not considered a structural failure due to the anchor bolts only providing 
a non-essential barrier to transverse displacement.  Bridges that contain multiple 
instances of structural failures are deemed to have inadequate behavior.  Bridges that 
contain only one or two structural failure analysis cases are classified as having 
 
72 
 
acceptable behavior.  This is deemed allowable due to overall analysis results being 
evaluated as an average of bridge response, and a small amount of failure cases are 
allowable in a large number of analyses.   
 
4.2  Bent Creek Road Bridge 
 
Results show that the average behavior of the Bent Creek Road Bridge was 
favorable for most aspects of the bridge components.  No design-level (1,000 year 
hazard) or MCE-level (2,500 year hazard) events resulted in specific structural failures; 
however, one event may lead to possible structural failures.  Foundation failures occurred 
in six of the design-level longitudinal events; however, the occurrence of these failures is 
not consistent with other structural behavior during the events.  Substructure behavior 
appears to be adequate at column and bent locations with the exception of foundation 
displacement capacities in the longitudinal direction of the bridge.  Bolt failure was the 
most frequent behavioral occurrence in the connection during transverse motion.  An 
additional observation regarding the analysis of this bridge is the large number of 
numerical failures present in the data.  This result is thought to have occurred due to the 
complexity of the bridge model.  Judgments needed to be made regarding the inclusion or 
exclusion of these results.  A more in-depth analysis of each set of analysis results is 
provided in the next sections.  
4.2.1 Transverse Motion 
 
The transverse analysis for this bridge resulted in stout, stiff structural behavior.  
See Table 4.1 and Table 4.2 for an overview of the design-level transverse results.  Bolt 
failure and low bent displacements indicate a stiff substructure able to transfer large loads 
 
73 
 
to the connection.  Both frictions models responded similarly to the design-level 
transverse ground motions, showing low foundation rotations and low residual 
displacements.  Overall design-level bent ductility did not exceed a value of 1, likely due 
to the stiff bridge bent.  Thirteen of the twenty design-level ground motions ran 
successfully with this bridge model, with a larger number of incomplete analyses 
occurring in the upper limit friction model.  Six of the seven design-level ground motions 
that did not run successfully were classified as numerical failures.  No data examined 
from these six ground motions indicated a structural failure before the analysis ceased.  
The one design-level analysis failure that was classified as such displayed bolt failure 
prior to the onset of larger ground motions during the Coalinga record.  It was estimated 
that larger shaking, given the existing bolt failure, may result in larger displacements of 
deck sections.  These displacements could result in failure of the structure, or at least a 
suspension of serviceability.  Another numerical failure (Imperial Valley) displayed 
similar bolt status to the Coalinga event, but had already withstood the majority of its 
associated ground motion, and was not classified as a ?failure? analysis. Successful 
design-level analysis results and the single ?failure? event were used to classify the 
design-level behavior of this model. 
 
 
 
 
 
 
 
74 
 
   
 
Table 4.1: Results Overview for Lower Limit Friction Bent Creek Transverse Model 
 
 
 
Table 4.2: Results Overview for Upper Limit Friction Bent Creek Transverse Model 
 
 
A total of fourteen MCE-level events were applied to this bridge model.  See 
Table 4.3 and Table 4.4 for an overview of the MCE-level transverse results.  Overall 
bent ductility did not exceed a value of 1.2, likely due to the stiff bridge bent.  Ten of the 
 
75 
 
Fourteen MCE-level ground motions ran successfully with this bridge model, with an 
equal number of incomplete analyses occurring in both friction models.  Three of the four 
ground motions that did not run successfully were classified as numerical failures.  No 
data examined from these three ground motions indicated a significant structural failure 
before the analysis ceased.  The one analysis failure that was classified as such displayed 
bolt failure prior to the onset of larger ground motions during the first Kocaeli record.  It 
was estimated that larger shaking, given the existing bolt failure, may result in larger 
displacements of deck sections.  These displacements could result in failure of the 
structure, or at least a suspension of serviceability.  Successful analysis results and the 
single ?failure? event design-level ground motions were used to classify the MCE-level 
behavior of this model.  All successful analyses resulted in acceptable bridge behavior in 
the transverse direction of loading, however the complexity of the model makes it 
difficult to analyze and thus state any solid conclusions regarding overall bridge 
behavior.  A solution to this would be to continue simplifying the model by reducing the 
number of modeled connections. 
Table 4.3: MCE-Level Results Overview for Lower Limit Friction Bent Creek Transverse Model 
 
 
 
 
76 
 
Table 4.4: MCE-Level Results Overview for Upper Limit Friction Bent Creek Transverse Model 
 
4.2.2 Longitudinal Motion 
 
Analysis of the Bent Creek Bridge under longitudinal ground motions provided 
mixed behavioral results.  Table 4.5 and Table 4.6 provide an overview of Bent Creek?s 
design-level longitudinal analysis.  Both friction models exhibited numerical failures 
(nine in total) as well as conflicting completed results.  Six of the eleven completed 
design-level analyses exhibited large foundation displacements.  This foundation 
behavior does not seem to be supported by forces within the recorded model data 
however, and its existence may be the result of errors propagated within the analysis.  
The remaining successful design-level analyses indicate very small displacements; 
however, the overall behavior of the structure may not be adequately described with the 
results obtained. 
 
 
 
 
 
 
77 
 
Table 4.5: Results Overview for Lower Limit Friction Bent Creek Longitudinal Model 
 
Table 4.6: Results Overview for Upper Limit Friction Bent Creek Longitudinal Model 
 
A total of fourteen MCE-level events were applied to this bridge model.  See 
Table 4.7 and Table 4.8 for an overview of the MCE-level longitudinal results.  Bent 
ductility was inconsistent throughout the analysis, with large and small values being 
recorded at the MCE-level.  Eight of the fourteen MCE-level ground motions ran 
successfully with this bridge model, with an equal number of incomplete analyses 
occurring in both friction models.  All of the ground motions that did not run successfully 
 
78 
 
were classified as numerical failures.  No data examined from these ground motions 
indicated a large structural failure before the analysis ceased.  One case of residual 
deformation exceedance was recorded, and seven cases of foundation displacement 
exceedance were recorded.  These large foundation displacements were recorded as a 
result of unexpected and sudden foundation movement.  These displacements could result 
in failure of the structure, or at least a suspension of serviceability.   
 
Table 4.7: MCE-Level Results Overview for Lower Limit Friction Bent Creek Longitudinal Model 
 
 
Table 4.8: MCE-Level Results Overview for Upper Limit Friction Bent Creek Longitudinal Model 
 
 
 
79 
 
4.3 Scarham Creek Bridge 
 
Results show that the average behavior of the Scarham Creek Bridge was 
favorable for design-level events.  This behavior does not carry over to the MCE-level 
behaviors.  Design-level events resulted in behaviors typical with flexible bridges, 
including larger span displacements and bent deflections.  Foundation performance for 
this bridge was mixed, and dependent on event direction.  All design-level analyses 
performed for this bridge succeeded in completing except for the second San Fernando 
ground motion applied to the upper limit friction model in the longitudinal direction.  
Substructure behavior appears to be adequate at column, strut and bent cap locations with 
the exception of foundation displacement capacities in the longitudinal direction of the 
bridge.  Connection behavior also appears to be adequate from the results of each design-
level analysis.  Bearing pad slippage was minimal in both directions of loading, and did 
not approach levels indicative of unseating.  A more in-depth analysis of each bridge 
model?s results is proved in the next sections.  
4.3.1 Transverse Motion 
 
The transverse analysis for this bridge resulted in flexible structural behavior.  See 
Table 4.9 and Table 4.10 for an overview of the design-level transverse results.  High 
bent displacements are indicative of a flexible substructure able to accommodate large 
deflections.  This behavior is in keeping with tall column bridge bents, like the ones 
present in the Scarham Bridge.  Excessive deflections were a concern for the transverse 
motion of this bridge, in both span and bent locations.  A hallmark of long period 
structures is large deflections, especially during certain ground motions.  Results from   
both friction models indicated design-level deflections that were within acceptable limits 
 
80 
 
for this bridge, however the magnitude of these deflections may not be acceptable for 
other bridges in this project.  All transverse design-level analyses of the Scarham Creek 
Bridge ran successfully.  One very important observation of the recorded data from both 
friction models of this bridge is the fact that every single data point is the same for the 
transverse direction motions.  The only difference between both bridge models is the 
change in bearing pad friction factor.  Both models behave in exactly the same manner 
during events that do not allow friction slips to occur.  This information points to a 
variety of possible conclusions.  The first conclusion assumes a very flexible bearing pad 
in use at each connection, one that is able to accommodate all connection motion without 
reaching a slipping point.  A second conclusion could be made indicating that the 
stiffness of both the substructure and superstructure were so similar that the resulting 
differential movement between the two at the connection zones was minimal and did not 
cause slipping.  This second conclusion is not supported by shear bolt behavior however; 
which did carry a significant load at the connection location.  Results seem to favor the 
first conclusion, but the lack of slipping is still noted.   
 
Table 4.9: Results Overview for Lower Limit Friction Scarham Creek Transverse Model 
 
 
81 
 
Table 4.10: Results Overview for Upper Limit Friction Scarham Creek Transverse Model 
 
 
A total of fourteen MCE-level events were applied to this bridge model.  All 
transverse MCE-level analyses of the Scarham Creek Bridge ran successfully.  Only one 
case of limit exceedance was reported and the results can be found in Table 7.11 and 
Table 7.12.  Results between the two friction models are the same with the exception of 
the Landers MCE event.  This event in the upper friction limit model results in a high 
foundation rotation case similar to the cases recorded for longitudinal motion.  All 
residual displacements are below the 1? limit, indicating acceptable serviceability of the 
structure post event.  
Table 4.11: MCE-Level Results Overview for Lower Limit Friction Scarham Creek Transverse 
Model 
 
 
82 
 
Table 4.12: MCE-Level Results Overview for Upper Limit Friction Scarham Creek Transverse 
Model 
 
4.3.2 Longitudinal Motion 
 
The longitudinal analysis for this bridge resulted in fairly consistent design-level 
results.  See Table 4.13 and Table 4.14 for an overview of the design-level transverse 
results.  Maximum span displacements of about 2.5 inches occurred for most ground 
motions in each bridge model, and bent displacements hovered around 2 inches.  These 
results are to be expected in bridges with such tall columns.  Bent ductility appears to 
remain below the established maximum for the second and fourth bent (1.076 inches), 
with only one design-level ground motion, (San Fernando 2) causing a ductility in excess 
of this limit.  Foundation capacity for this bridge exceeds the displacement limit set by 
geotechnical data.  This foundation displacement is present in all design-level events, and 
is thus a conclusive failure in the longitudinal direction of this bridge.  These 
deformations at foundation-level are a result of shear forces carried by the substructure, 
and not a product of excessive moments or structural failures.   
 
 
 
 
 
 
 
83 
 
Table 4.13: Results Overview for Lower Limit Friction Scarham Creek Longitudinal Model 
  
 
Table 4.14: Results Overview for Upper Limit Friction Scarham Creek Longitudinal Model 
 
 
A total of fourteen MCE-level events were applied to this bridge model.  See 
Table 4.15 and Table 4.16 for an overview of the MCE-level longitudinal results.  Only 
six of the fourteen analyses were completed with all incomplete analyses showing signs 
of possible structural failure.  It is difficult to classify the MCE-level of behavior for this 
bridge as acceptable given the large amount of unknowns found in the results.  
Conservative estimates would suggest that this bridge?s behavior is unsuitable for an 
MCE-level event, however the classification of ?Failure? in many of these results stems 
 
84 
 
from the conservative judgment, rather than an observed failure.   Additional study of 
taller bent bridges at the MCE-level may be required in order to gather more information 
on cases similar to this analysis case. 
Table 4.15: MCE-Level Results Overview for Lower Limit Friction Scarham Creek Longitudinal 
Model 
  
 
Table 4.16: MCE-Level Results Overview for Upper Limit Friction Scarham Creek Longitudinal 
Model 
 
 
4.4 Norfolk Southern Railroad Bridge 
 
Results show that the average behavior of the Norfolk Southern Bridge was 
favorable for design and MCE-level events.  Design-level events resulted in behaviors 
typical with stiff elastic bridges, including smaller span displacements and bent 
deflections.  Foundation performance for this bridge was acceptable in both directions of 
 
85 
 
motion.  All design-level analysis performed for this bridge succeeded in completing in 
the lateral direction.  Three design-level analyses performed in the longitudinal direction 
failed to result in completion.  Substructure behavior appears to be adequate at column 
and bent cap locations.  Connection behavior also appears to be adequate from the results 
of each analysis with anchor bolts failing, but no cases of unseating.  Bearing pad 
slippage was minimal in both directions of loading, and did not approach levels indicative 
of unseating.  A more in-depth analysis of each bridge model?s results is proved in the 
next sections.  
4.4.1 Transverse Motion 
 
The transverse analysis for this bridge resulted in stiff structural behavior.  See 
Table 4.17 and Table 4.18 for an overview of the design-level transverse results.  Lower 
bent displacements are indicative of a stiff substructure.  This behavior is in keeping with 
triple column bridge bents, like the ones present in the Norfolk Southern Railroad Bridge.    
Design-level results from both friction models indicated residual deflections that were 
within acceptable limits for this bridge.  All transverse design-level analyses of the 
Norfolk Southern Railroad Bridge ran successfully.  Many anchor bolt failures can be 
observed in the transverse analyses.  These failures are a result of high levels of forces 
being transferred between the super and substructure of the bridge.  This is more common 
in bridges with stiffer substructures, since the superstructure of each bridge is simply 
supported with long span lengths and generally behaves in a more flexible manner.  
Overall these results are in keeping with the predicted behaviors of this bridge. 
 
 
 
86 
 
Table 4.17: Results Overview for Lower Limit Friction Norfolk Southern Railroad Transverse Model 
 
 
Table 4.18: Results Overview for Upper Limit Friction Norfolk Southern Railroad Transverse Model 
 
 
A total of fourteen MCE-level events were analyzed with this bridge model.  See 
Table 4.19 and Table 4.20 for an overview of the MCE-level transverse results.  A 
majority of these MCE results indicated lower bent displacements and span 
displacements with a single exception.  This behavior is in keeping with triple column 
bridge bents, like the ones present in the Norfolk Southern Railroad Bridge.  The Landers 
event for the lower friction limit model resulted in behavior that is indicative of span 
 
87 
 
unseating.  A high bearing pad displacement was recorded as well as a high residual 
displacement.  High residual displacement was also recorded in the upper bound friction 
model during the same ground motion.  All transverse design-level analyses of the 
Norfolk Southern Railroad Bridge ran successfully.  Many anchor bolt failures were 
observed in the MCE-level transverse analyses.  This is more common in bridges with 
stiffer substructures, since the superstructure of each bridge is simply supported with long 
span lengths and generally behaves in a more flexible manner.  Overall these results are 
in keeping with the predicted behaviors of this bridge, and the average behavior of these 
models indicates acceptable behavior at the MCE-level. 
 
Table 4.19: MCE-Level Results Overview for Lower Limit Friction Norfolk Southern Railroad 
Transverse Model 
 
Table 4.20: MCE-Level Results Overview for Upper Limit Friction Norfolk Southern Railroad 
Transverse Model 
 
 
88 
 
4.4.2 Longitudinal Motion 
 
The longitudinal analysis for this bridge resulted in fairly consistent results.  See 
Table 4.21 and Table 4.22 for an overview of the design-level longitudinal results. 
Maximum span displacements of about 1 inch occurred for most design-level ground 
motions in each bridge model, and bent displacements hovered around 1.1 inches.  These 
results are to be expected in bridges with such stiff substructure.  Bent ductility appears to 
remain well below the established maximum (3.89).  Foundation capacity for this bridge 
does not exceed the displacement limit set by geotechnical data.  Recorded column shears 
for this model appear to be incorrectly recorded due to very small recorded values when 
compared to every other analysis performed in the chapter.  Three of the design-level 
analyses performed on the models resulted in analysis failure, with the Kocaeli 1 event 
resulting in possible structural failure.  The results from this event only last for 22.52 
seconds of the applied ground motion, but they already exhibit reasonable bridge 
response before the largest accelerations are even applied.  A judgment was made to 
classify these results as a possible structural failure.  Despite this, overall behavior of this 
bridge appears to be sufficient in the design-level longitudinal direction. 
 
 
 
 
 
 
 
 
 
 
89 
 
Table 4.21: Results Overview for Lower Limit Friction Norfolk Southern Longitudinal Model 
  
 
Table 4.22: Results Overview for Upper Limit Friction Norfolk Southern Longitudinal Model 
 
 
 
A total of fourteen MCE-level events were applied to this bridge model.  See 
Table 4.23 and Table 4.24 for an overview of the MCE-level longitudinal results.  Ten of 
the fourteen analyses ran completely.  Three of the four that did not result in completion 
were determined to cause possible failure in the form of eventual span unseating.  The 
displacements of the bearing pads in these analyses were relatively high at the time of 
analysis failure.  This combined with the fact that these analysis ceased before sections of 
 
90 
 
the largest ground motion accelerations were applied to the models lead to the prediction 
of possible unseating.  This was determined due to high recorded bearing displacements 
during sections of the ground motion that proceeded larger ground motion accelerations 
that were never applied in the analysis.   The overall behavior of this bridge appears to be 
acceptable for critical and essential bridges, and the problem of unseating should become 
reduced due to the adoption of increased seat widths for future designs.  
   
Table 4.23: MCE-Level Results Overview for Lower Limit Friction Norfolk Southern Longitudinal 
Model 
 
 
 
Table 4.24: MCE-Level Results Overview for Upper Limit Friction Norfolk Southern Longitudinal 
Model 
 
 
 
 
 
 
91 
 
4.5 Oseligee Creek Bridge 
 
Results show that the average behavior of the Oseligee Creek Bridge was mixed 
for design and MCE-level events.  Design-level events resulted in behaviors typical with 
flexible bridges, including larger span displacements and bent deflections.  These large 
span deflections and bent deflections also lead to some cases of collapse of the bridge 
model.  Foundation performance for this bridge was also mixed, and varied from event to 
event.  All design and MCE-level analysis performed for this bridge succeeded in 
completing except for one transverse analysis at each level.  Substructure behavior 
appears to be inadequate at column and foundation locations due to the large scour 
situation recommended by geotechnical information. Connection behavior appears to be 
adequate from the results of each analysis, with design-level unseating only occurring in 
cases of bridge collapse.  A more in-depth analysis of each bridge model?s results is 
proved in the next sections.  
4.5.1 Transverse Motion 
 
The transverse analysis for this bridge resulted in flexible structural behavior.  See 
Table 4.25 and Table 4.26 for an overview of the design-level transverse results.  High 
bent displacements are indicative of a flexible substructure able to accommodate large 
deflections.  This behavior is in keeping with tall column bridge bents, like the ones 
present in the 100% scour condition of the Oseligee Creek Bridge.  Excessive deflections 
were a concern for the transverse motion of this bridge, in both span and bent locations.  
These deflections led to the collapse of the lower limit friction bridge model during the 
design-level Landers event.  A hallmark of long period structures is large deflections, 
especially during certain ground motions.  Results from both friction models indicated a 
 
92 
 
majority of design-level deflections that were within acceptable limits for this bridge.  
The ductility for both models exceeds the limit stated in Table 3.7 of Section 3.6 during 
the second San Fernando event, however the limit established for ductility of this bridge 
did not incorporate a 100% scour case.  This 100% scour case could alter the limits 
established by the static pushover analysis. All transverse design-level analyses of the 
Oseligee Creek Bridge ran successfully with the exception of the Landers ground motion 
applied to the upper limit friction model.  This design-level event was assumed to cause a 
collapse failure due to the results observed from the previous model, as well as the results 
recorded before the analysis ceased.  Excessive foundation rotations were also observed 
in 14 of the 20 design-level transverse analyses, with a few other recorded rotations 
approaching the rotational limits.  Ground motion analyses that exhibit these large 
foundation rotations also tended to exhibit bolt failures at the connection locations.  
 
Table 4.25: Results Overview for Lower Limit Friction Oseligee Creek Transverse Model 
 
 
 
 
 
93 
 
Table 4.26: Results Overview for Upper Limit Friction Oseligee Creek Transverse Model 
 
A total of fourteen MCE-level events were analyzed with this bridge model.  See 
Table 4.27 and Table 4.28 for an overview of the MCE-level transverse results.  High 
bent ductilities can be observed in these results, as well as high foundation rotations.    
Four of the fourteen analyses resulted in residual displacements in excess of the 1? limit.  
All of these undesired behaviors result in questions with this bridge?s adequacy but it is 
important to consider the 100% scour condition present in this bridge.  A reduction in this 
scour amount would result in smaller deflections and a more rigid bridge.   
 
Table 4.27: MCE-Level Results Overview for Lower Limit Friction Oseligee Creek Transverse 
Model 
 
 
 
94 
 
 
Table 4.28: MCE-Level Results Overview for Upper Limit Friction Oseligee Creek Transverse 
Model 
 
4.5.2 Longitudinal Motion 
 
The longitudinal analysis for this bridge resulted in fairly consistent results.  See 
Table 4.29 and Table 4.30 for an overview of the design-level transverse results. 
Maximum span displacements of about 2 inches occurred for most ground motions in 
each bridge model, and bent displacements hovered around 1 inch.  These results are to 
be expected in bridges with such tall columns.  Bent ductility remains below the 
established maximum (6), with larger ductility occurring in the upper limit friction 
model.   Foundation capacity for some design-level analyses of this bridge exceeds the 
rotational limit set by geotechnical data.  This foundation rotation is present in three 
design-level events, two of which also cause collapse of the bridge.  These collapse 
events also show span unseating and high residual displacements, both behaviors support 
the conclusion of bridge collapse.   
 
 
 
 
95 
 
 
Table 4.29: Results Overview for Lower Limit Friction Oseligee Creek Longitudinal Model 
 
 
 
 
Table 4.30: Results Overview for Upper Limit Friction Oseligee Creek Longitudinal Model 
 
 
 
A total of fourteen MCE-level events were applied to this bridge model.  See 
Table 4.31 and Table 4.32 for an overview of the MCE-level longitudinal results.  All 
MCE-level analyses of this bridge were completed, and most resulted in acceptable 
bridge behavior.  Four cases of residual displacements exceeding the limit of 1? were 
recorded, including three cases of bridge collapse.  An additional case of span unseating 
 
96 
 
was also recorded.  The recorded foundation rotations of the upper limit friction model 
were also very high in some cases. 
 
Table 4.31: MCE-Level Results Overview for Lower Limit Friction Oseligee Creek Longitudinal 
Model 
 
 
 
Table 4.32: MCE-Level Results Overview for Upper Limit Friction Oseligee Creek Longitudinal 
Model 
 
 
 
4.6 Little Bear Creek Bridge 
 
Results show that the average behavior of the Little Bear Creek Bridge was 
favorable for design and MCE-level events.  Super and substructure behaviors of this 
bridge were found to have acceptable behavior.  Foundation performance for this bridge 
was mixed, and dependent on event direction.  Most design-level analysis performed for 
 
97 
 
this bridge succeeded in completing except for four ground motions applied in the 
transverse direction.  Substructure behavior appears to be adequate at column, strut and 
bent cap locations with the exception of foundation displacement capacities in the 
transverse direction of the bridge.  Connection behavior also appears to be adequate from 
the results of each analysis with one design-level case of unseating being recorded.  This 
bridge had three separate models analyzed in the transverse direction, two of which 
contain a newly designed anchor bolt configuration, while the third contains a previously 
designed anchor bolt configuration.  A more in-depth analysis of each bridge model?s 
results is proved in the next sections.  
4.6.1 Transverse Motion 
 
The transverse analysis for this bridge resulted in stiff structural behavior.  See 
Table 4.33, Table 4.34 and table 4.35 for an overview of the design-level transverse 
results.  Each of the three connection details used in this model behaved in an acceptable 
manner during all design-level ground motions.  No bolt failure was recorded, and the old 
bolt configuration consists of smaller anchor bolts than the new configuration.  This old 
configuration results in a connection that is less stiff, and some behavioral effects of this 
stiffness change can be observed.  The model containing the old configuration 
experienced higher bent displacements, higher span displacement and higher bolt 
demands then that of the newer bolt configuration.  Higher residual displacements were 
also observed in the older bolt configuration.  These results make sense given that 
structures with less strength and/or stiffness have a tendency for higher displacements.  
Results from all transverse models indicated design-level deflections that were within 
acceptable limits for this bridge.  Small residual displacements occurred in each ground 
 
98 
 
motion as a result of the survival of the anchor bolts as well as the lower levels of 
ductility that occurred during each ground motion.  Transverse design-level analyses of 
the Little Bear Creek Bridge ran successfully with the exception of the Landers event.  
There was also an analysis failure in the upper limit friction model containing the new 
bolt configuration.  This failure occurred during the first San Fernando event.  Large 
foundation deformations were recorded in the transverse direction for every design event 
except the Little Skull Mountain event.  This Little Skull Mountain event has generally 
resulted in lower bridge response throughout this project.  The exact cause of these large 
foundation displacements can be assumed to have been caused by weaker foundation 
capacity and a bridge with a relatively large mass.  The inertia of the bridge?s 
components may have caused large forces to develop at the foundation level during 
motion. 
Table 4.33: Results Overview for Lower Limit Friction Little Bear Creek Transverse Model 
 
 
 
 
 
 
99 
 
Table 4.34: Results Overview for Upper Limit Friction Little Bear Creek Transverse Model 
 
Table 4.35: Results Overview for Upper Limit Friction Little Bear Creek Transverse Model with Old 
Bolts 
 
 
A total of twenty-one MCE-level events were analyzed with this bridge model.  
See Table 4.36, Table 4.37 and table 4.38 for an overview of the MCE-level transverse 
results.  Each of the three connection details used in this model behaved in an acceptable 
manner during most MCE-level events.  Two cases of collapse and span unseating were 
recorded for MCE-level events; however these results are taken from analyses that did 
not result in completion.  A total of seven events did not result in a complete analysis, 
 
100 
 
with three resulting in data that suggests structural failure.  Foundation displacement 
capacity is exceeded in all transverse MCE events, just like the design-level results. 
 
Table 4.36: MCE-Level Results Overview for Lower Limit Friction Little Bear Creek Transverse 
Model 
 
Table 4.37: MCE-Level Results Overview for Upper Limit Friction Little Bear Creek Transverse 
Model 
 
Table 4.38: MCE-Level Results Overview for Upper Limit Friction Little Bear Creek Transverse 
Model with Old Bolts 
 
 
101 
 
4.6.2 Longitudinal Motion 
 
The longitudinal analysis for this bridge resulted in fairly consistent results.  See 
Table 4.39 and Table 4.40 for an overview of the design-level transverse results. 
Maximum span displacements of about 2.5 inches occurred for most ground motions in 
each bridge model, and bent displacements hovered around 1.5 inches.  These results are 
larger than expected, and may point to possible amplified bridge response.  Bent ductility 
appears to remain below the established maximum for the second bent (2.77), with only 
one ground motion, (Landers) causing a ductility in excess of this limit.  This excess in 
the upper limit friction model?s ductility appears to have been caused due to a collapse of 
a bridge bent.  Results show that this design-level event also contains large foundation 
displacements and a large case of span end displacements, resulting in unseating.  Despite 
the results of the aforementioned event, all other events appear to behave in an acceptable 
manner.  Foundation behavior in the longitudinal direction appears to suggest low shears 
and moments within bent columns.  Bearing pad displacements also indicate constant 
movement and slipping, given maximum demand capacity ratios of about 0.4 for each 
model.  Lastly the residual displacements in this direction are below the acceptable limit 
in all but two cases, the case of collapse and the Coalinga analysis of the lower limit 
bearing pad model. 
 
 
 
 
 
 
102 
 
Table 4.39: Results Overview for Lower Limit Friction Little Bear Creek Longitudinal Model 
  
 
Table 4.40: Results Overview for Upper Limit Friction Little Bear Creek Longitudinal Model 
  
 
A total of fourteen MCE-level events were applied to this bridge model.  See 
Table 4.41 and Table 4.42 for an overview of the MCE-level longitudinal results.  All of 
the MCE-level analyses were completed, resulting in acceptable bridge behavior.  A 
single case of residual displacement exceedance was observed in the lower bound friction 
model during the Landers event. 
 
103 
 
 
Table 4.41: MCE-Level Results Overview for Lower Limit Friction Little Bear Creek Longitudinal 
Model 
  
 
Table 4.42: MCE-Level Results Overview for Upper Limit Friction Little Bear Creek Longitudinal 
Model 
 
 
4.7 Summary of Results 
 
Overall each bridge in these analysis contained some flaw or behavior that will 
need to be improved upon, but the majority of analyses provided results deemed 
acceptable.  A summary of critical events taken from all design-level analyses can be 
found in Table 4.43.  A summary of critical events taken from all MCE-level analyses 
can be found in Table 4.44.  The Bent Creek Road Bridge consistently produced partially 
complete analyses, but the few that did result in completion suggested adequate 
performance.  The Scarham Creek Bridge demonstrated expected behaviors in regards to 
 
104 
 
overall design-level performance, utilizing large displacements as a method of surviving 
the earthquake.  The Scarham model did however present large foundation 
displacements.  The MCE-level behavior for the Scarham Bridge proved to be less than 
acceptable.  Incomplete analyses were present in most MCE-level longitudinal events 
along with high displacements.  These results indicate the need for additional design 
considerations of tall substructure long span bridges like the Scarham Creek Bridge, 
design considerations regarding P-delta and stability concerns. The Norfolk Southern 
Railroad Bridge behaved as a stiff bridge would, results of these analyses include lower 
displacements and anchor bolt connection failure.  Despite this, bridge behavior still 
remained within the range of acceptable behavior at the design level and MCE-level.  The 
Oseligee Creek Bridge presented the challenge of a 100% scour condition during an 
earthquake event.  Total bent collapse occurred in both directions of motion, at both 
design and MCE-level events.  Special design actions may need to be implemented in 
order to prevent occurrences like this.  The Little Bear Creek Bridge performed 
acceptably under most ground motions.  This bridge was also used to compare old bolt 
configurations with code specified configurations.  The new bolt configurations caused a 
slight increase in overall bridge performance in that it reduced overall deflections of the 
structure.  This reduction of overall deflection does not necessarily result in better seismic 
bridge behavior for all bridges, and greater connection stiffness can actually lead to 
damage in other components of the bridge.  The main reason for selection of the new 
connection anchor bolt is the redundancy that it provides for the connection, as well as an 
increase in the connection shear capacity.  An isolated occurrence of bent collapse was 
 
105 
 
observed in the design-level analysis of the Little Bear Creek Bridge. Conclusions and 
design recommendations will be provided in the next chapter of this thesis. 
Table 4.43: Summary of Critical Design-Level Analysis Behaviors 
Bridge Case Number of Analyses Critical Analysis Behaviors 
Bent Creek Bridge 40 16 cases of unfinished analysis 6 cases of foundation displacement exceedance 
Scarham Creek Bridge 40 Foundation Displacement exceedance in all longitudinal events 
Norfolk Southern Bridge 40 3 cases of unfinished analysis 1 case of residual displacement exceedance 
Oseligee Creek Bridge 40 
17 cases of Foundation Rotation exceedance 
3 cases of residual displacement exceedance 
3 cases of bent collapse 
Little Bear Creek 50 
4 cases of unfinished analysis 
25 cases of foundation displacement exceedance 
1 case of bent collapse 
2 cases of residual displacement exceedance 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
106 
 
 
Table 4.44: Summary of Critical MCE-Level Analysis Behaviors 
Bridge Case Number of Analyses Critical Analysis Behaviors 
Bent Creek Bridge 28 
10 cases of unfinished analysis 
1 case of residual displacement exceedance 
7 cases of foundation displacement exceedance 
Scarham Creek Bridge 28 
Foundation Displacement exceedance in all 
longitudinal events 
8 cases of unfinished analysis 
1 case of residual displacement exceedance 
Norfolk Southern Bridge 28 
4 cases of unfinished analysis 
1 case of bent collapse 
1 case of span unseating 
2 cases of residual displacement exceedance 
Oseligee Creek Bridge 28 
17 cases of Foundation Rotation exceedance 
7 cases of residual displacement exceedance 
4 cases of bent collapse 
Little Bear Creek 35 
7 cases of unfinished analysis 
21 cases of foundation displacement exceedance 
2 cases of bent collapse 
1 case of residual displacement exceedance 
 
 
107 
 
Chapter 5: Summary, Conclusions and Recommendations 
 
5.1 Summary 
 
The Alabama Department of Transportation (ALDOT) is in the process of 
updating their seismic design procedure by adopting AASHTO?s LRFD Bridge Design 
Specifications and the AASHTO?s LRFD Seismic Bridge Design Guide Specification.  
The use of the Guide Specifications is limited to bridges classified as ?non-critical? and 
?non-essential.?  Analysis was requested by ALDOT regarding the applicability of the 
Guide Specifications design on current bridges in the state, and the behavior of these 
redesigned bridges when subjected to the seismic hazards present in the state.  If this 
analysis was able to conclude that the behavior of these redesigned bridges was adequate 
for critical and essential bridge performance, then it could be concluded that the Guide 
Specifications could be adopted for use in the design of critical and essential bridges in 
the state of Alabama due to the moderate hazard. 
In this thesis, analyses were performed on redesigned bridge models in an attempt 
to determine these bridges behaviors when subjected to seismic hazards found in the 
state.  An initial literature review was performed in order to determine the function and 
behavior of concrete bridges during seismic events, as well as determine the most critical 
failure modes and damage associated with seismic loading of bridges.  Three main 
behaviors were selected as the primary cause of bridge failure during seismic events: 
foundation failure, unseating of spans, and column failure.  A suite of five bridges was 
 
108 
 
selected from a list of bridges designed and constructed in Alabama.  These bridges were 
redesigned according to the Guide Specifications for use in this analysis.  Next an 
analysis of Alabama?s seismic hazard was performed, and ground motions were selected 
and scaled to represent this hazard.  Bridges were then modeled in CSIBridge and 
analyzed using a nonlinear time history analysis.  The results of these analysis concluded 
that the average behavior for most of the bridges is acceptable for use in critical and 
essential bridges.  Design recommendations for each of these bridges are provided in the 
next section in order to improve overall seismic performance. 
5.2 Conclusions 
 
Based on the analysis performed in this project, the following conclusions 
regarding the performance and implementation of the Guide Specifications design on 
Alabama bridges were reached: 
 Ground Motions that represent the hazard for northern Alabama are difficult to 
acquire, and scaling of worldwide ground motions is required in order to simulate 
this hazard. 
 The two largest seismic hazards locations in Alabama contain unique hazards that 
each control different natural period bridges.  One hazard location controls longer 
period bridges (Northwest Alabama), and the other controls shorter period bridges 
(Northeast Alabama). 
 Current connection details at end span locations of highway bridges exhibit 
adequate strength and ductility for design-level seismic events.  Failures of these 
connections were only observed in cases of excessive bent and span 
displacements, indicating structural collapse of bridge bents.  
 
109 
 
 Highway bridges designed with stiffer substructures using the Guide 
Specifications behave in a manner that is considered safe and result in usability 
after a seismic event. 
 Bridges resting on weaker or scour-able soils will require additional analysis in 
order to determine lateral foundation and soil capacities, as well as capacities of 
substructure displacement.  Increase in foundation and column sizes may be 
required if soil cannot provide adequate lateral resistance. 
 Bridges with tall substructures will require additional p-delta analysis for larger 
displacements.   
 Critical and Essential bridges can be designed using the AASHTO Guide 
Specifications in the state of Alabama, with minor additional analysis. 
5.3 Recommendations 
 
The following are a few additional tasks that have been recommended for 
completion after the culmination of the work performed in this thesis: 
 The comparison of results obtained from the models described in this thesis with 
the results obtained from the geotechnical foundation models analyzed and 
discussed by Kane (2013). 
 Performance of a dynamic lateral loading test on a real bridge bent designed to 
conform to the Guide Specifications.  This test would help evaluate inelastic 
behavior of bridge bents. 
 
 
 
 
110 
 
References 
 
 
AASHTO. (2002). Standard Specifications for Highway Bridges (17th ed.). Washington 
DC: American Association of State Highway and Transportation Officials. 
AASHTO. (2007). AASHTO Ground Motion Parameters Calculator Version 2.1. 
Washington DC. 
AASHTO. (2009). LRFD Bridge Design Specifications (4th ed.). Washington DC: 
American Association of State Highway and Transportation Officials. 
AASHTO. (2011). Guide Specifications for LRFD Seismic Bridge Design (2nd ed.). 
Washington DC: American Association of State Highway and Transportation 
Officials. 
AISC. (2005). Steel Construction Manual (13th ed.). United States: American Institute of 
Steel Construction. 
Alabama DOT. (2012, November 30). ALDOT Standard Details. Alabama Department of 
Transportation. 
ATC/MCEER Joint Venture. (2003). Recommended LRFD Guidelines for the Seismic 
Design of Highway Briges. Federal Highway Administration. 
Barker, R. M., & Puckett, J. A. (2007). Design of Highway Bridges An LRFD Approach. 
Hoboken, NJ: John Wiley & Sons. 
BSI. (2013). FB-MultiPier v4.18. Gainesville, Florida: BSI. 
Caltrans. (1994, June). Bridge Bearings. Memo to designers 7-1, pp. 1-12. 
Caltrans. (2011, October). Bridge Design Specifciations: Chapter 1. Retrieved February 
18, 2013, from California Department of Transportation: 
http://www.dot.ca.gov/hq/esc/techpubs/manual/bridgemanuals/bridge-design-
practice/pdf/bdp_1.pdf 
Chen, W.-F., & Duan, L. (2000). Bridge Engineering Seismic Design. Boca Raton: CRC 
Press. 
Chopra, A. K. (2007). Dynamics of Structures. Upper Saddle River, NJ: Pearson Prentice 
Hall. 
Computer and Structures Inc. (2012). CSI Bridge 15. Berkely, CA: Computer and 
Structures, Inc. 
Computers & Structures inc. (2012, August 02). Caltrans vs. fiber hinge. Berkley, 
California, United States of America. 
Corporation, P. T. (2007). Mathcad 14. Needham, MA: Parametric Technology 
Corporation. 
Coulston, P. J., & Marshall, J. D. (2011). Influence of the LRFD Bridge Design 
Specifications on Seismic Design in Alabama. Auburn University, Auburn, AL. 
EERI. (2001, November). Preliminary Observations on the Southern Peru Earthquake of 
June 23, 2001. Learning from Earthquakes, pp. 3-11. 
 
111 
 
Kane, K. M. (2013). Response of Deep Foundations to Seismic Loads in Alabama. 
Auburn University, Auburn. 
Law, J. D. (2013). Update of Bridge Design Standards in Alabama for AASHTO LRFD 
Seismic Design Requirements. Auburn, AL: Auburn University. 
Lin, C.-C. J., Hung, H.-H., Liu, K.-Y., & Chai, J.-F. (2008). RECONNAISSANCE 
REPORT OF 0512 CHINA WENCHUAN. The 14th World Conference on 
Earthquake Engineering (pp. 1-9). Bejing: NCREE. 
PEER. (2010). PEER Ground Motion Database. Retrieved September 20, 2013, from 
PEER Ground Motion Database: 
http://peer.berkeley.edu/peer_ground_motion_database/spectras/21713/unscaled_
searches/new 
PTC. (2007). Mathcad. Needham, Massachusetts, USA. 
Rodriguez-Marek, A. (2007). DYNAMIC RESPONSE OF BRIDGES TO NEAR-FAULT, 
FORWARD DIRECTIVITY GROUND MOTIONS. Pullman: Washington State 
Transportation Center. 
Steelman, J. S., Fahnestock, L. A., Filipov, E. T., LaFave, J. M., Hajjar, J. F., & Foutch, 
D. A. (2012, April 24). Shear and Friction Response of Non-Seismic Laminated 
Elastomeric Bridge Bearings Subject to Seismic Demands. Journal of Bridge 
Engineering, pp. 1-15. 
StructurePoint. (2012). spColumn. StructurePoint, LLC. 
Tobias, D. H., Anderson, R. E., Hodel, C. E., Kramer, W. M., Wahab, R. M., & Chaput, 
R. J. (2008, August). Overview of Earthquake Resisting System Design and 
Retrofit Strategy for Bridges in Illinois. Practice Periodical on Structural Design 
and Construction, pp. 148-152. 
USGS. (2013, May 17). 2008 Interactive Deaggregations. Retrieved September 17, 2013, 
from USGS: https://geohazards.usgs.gov/deaggint/2008/ 
Wight, J. K., & MacGregor, J. G. (2009). Reinforced Concrete: Mechanics and Design. 
Upper Saddle River, NJ: Pearson Prentice Hall. 
Wotherspoon, L., Bradshaw, A., Green, R., Wood, C., Palermo, A., Cubrinovski, M., & 
Bradley, B. (2011). Performance of bridges during the 2010 Darfield and 2011 
Christchurch earthquakes. Seismological Research Letters Volume 82, 961-962. 
Yen, W.-H. P., Chen, G., Buckle, i., Allen, T., Alzamora, D., Ger, J., & Arias, J. G. 
(2011). Post-Earthquake Reconnaissance Report on Transportation 
Infrastructure:. McLean, VA : U.S. Department of Transportation. 
 
112 
 
APPENDIX A: GROUND MOTIONS 
 
 
 
 
 
 
  
 
0 3.5 7 10.5 14 17.5 21 24.5 28
0.3
0.15
0
0.15
0.3
S A N F E R NA ND O  1 ( NG A  EV ENT # 68)
T i m e  ( s )
A
c
c
e
l
e
r
a
t
i
on (
g)
0 3.75 7.5 11.2 5 15 18.7 5 22.5 26.2 5 30
0.2
0.1
0
0.1
0.2
S A N F E R NA ND O  2 ( NG A  EV ENT # 70)
T i m e  ( s )
A
c
c
e
l
e
r
a
t
i
on (
g)
 
113 
 
 
 
  
 
 
 
 
 
 
 
 
 
 
 
0 5 10 15 20 25 30 35 40
0.1
0.03 8
0.02 5
0.08 8
0.15
IM P ERIAL VAL LEY  (NGA EV ENT # 18 6)
T i m e  ( s )
A
c
c
e
l
e
r
a
t
i
on (
g)
0 4 8 12 16 20 24 28 32
0.1
0.05
0
0.05
0.1
C OALI NG A  (NGA E VEN T  # 3 33)
T i m e  ( s )
A
c
c
e
l
e
r
a
t
i
on (
g)
 
114 
 
 
 
 
 
 
  
 
 
 
 
 
 
 
 
 
0 2. 55 5. 1 7. 65 10 .2 12 .7 5 15 .3 17 .8 5 20 .4
0. 15
0. 08 8
0. 02 5
0. 03 7
0. 1
NOR T H  P A LM S P RING S  (N G A E VEN T  #512)
T i m e  ( s )
A
c
c
e
l
e
r
a
t
i
on
 (
g)
0 6. 25 12 .5 18 .7 5 25 31 .2 5 37 .5 43 .7 5 50
0. 2
0. 12 5
0. 05
0. 02 5
0. 1
LA NDE R S  (NG A EV ENT # 83 2)
T i m e  ( s )
A
c
c
e
l
e
r
a
t
i
on
 (
g)
 
115 
 
 
 
  
 
 
 
 
 
 
 
 
 
 
 
 
0 14 .0 19 28 .0 38 42 .0 56 56 .0 75 70 .0 94 84 .1 13 98 .1 31 11 2 .1 5
0. 11 8
0. 06 6
0. 01 5
0. 03 7
0. 08 8
LITT LE  S K ULL  MOU NTAI N  (NG A EV EN T  # 1 741)
T i m e  ( s )
A
c
c
e
l
e
r
a
t
i
on (
g)
0 18.8 03 37.6 06 56.4 09 75.2 13 94.0 16 112 .819 131 .622 150 .425
0.24 87
0.13 24
0.01 61
0.10 02
0.21 64
KOC AE LI-TUR K EY  1 (N G A E VEN T  #1147 )
T i m e  ( s )
A
c
c
e
l
e
r
a
t
i
on (
g)
 
116 
 
 
 
 
 
 
 
 
0 3.5 7 10.5 14 17.5 21 24.5 28
0.2
0.07 5
0.05
0.17 5
0.3
KOC AE LI-TUR K EY 2 (NGA E VEN T  # 1 165)
T i m e  ( s )
A
c
c
e
l
e
r
a
t
i
on (
g)
0 5. 12 5 10 .2 5 15 .3 75 20 .5 25 .6 25 30 .7 5 35 .8 75 41
0. 2
0. 07 5
0. 05
0. 17 5
0. 3
KOBE (NG A  EV ENT # 11 07)
T i m e  ( s )
A
c
c
e
l
e
r
a
t
i
on (
g)