STRUCTURAL SYSTEMS

RESEARCH PROJECT

Report No. EXPERIMENTAL AND ANALYTICAL

SSRP–07/12
INVESTIGATION ON STIFFNESS AND
FINAL
ULTIMATE CAPACITY OF BRIDGE
ABUTMENTS

by

AZADEH BOZORGZADEH

SCOTT A. ASHFORD

JOSÉ I. RESTREPO

NONTAPAT NIMITYONGSKUL

Final Report Submitted to the California Department

of Transportation Under Contract No. 59A0337.

Department of Structural Engineering

June 2008 University of California, San Diego

La Jolla, California 92093-0085
 University of California, San Diego
Department of Structural Engineering
Structural Systems Research Project

Report No. SSRP–07/12

FINAL

Experimental and Analytical Investigation on

Stiffness and Ultimate Capacity of Bridge Abutments

by

Azadeh Bozorgzadeh
Graduate Student Researcher

Scott A. Ashford
Adjunct Professor of Geotechnical Engineering, and
Professor and School Head, Civil and Construction Engineering, Oregon State University

José I. Restrepo
Professor of Structural Engineering

Nontapat Nimityongskul
Graduate Student Researcher, Oregon State University

Final Report Submitted to the California Department of Transportation

Under Contract No. 59A0337.

Department of Structural Engineering

University of California, San Diego
La Jolla, California 92093-0085

June 2008
 Technical Report Documentation Page
1. Report No. 2. Government Accession No. 3. Recipient’s Catalog No.

4. Title and Subtitle 5. Report Date

Experimental and Analytical Investigation on Stiffness and Ultimate June 2008
Capacity of Bridge Abutments
6. Performing Organization Code

7. Author(s) 8. Performing Organization Report No.

Azadeh Bozorgzadeh, Scott A. Ashford, José I. Restrepo, Nontapat UCSD / SSRP-07/12
Nimityongskul
9. Performing Organization Name and Address 10. Work Unit No. (TRAIS)

Department of Structural Engineering

School of Engineering
11. Contract or Grant No.
University of California, San Diego
La Jolla, California 92093-0085
59A0337
12. Sponsoring Agency Name and Address 13. Type of Report and Period Covered
Final Report – July 2002 / June 2006
California Department of Transportation
14. Sponsoring Agency Code
Division of Engineering Services
1801 30th St., MS #9-2/5I
Sacramento, California 95816
15. Supplementary Notes
Prepared in cooperation with the State of California Department of Transportation.

16. Abstract
Abutments are earth-retaining structures which provide resistance to deformation and earthquake
induced inertial forces from the bridge superstructure. In order to limit the inertial forces transmitted into the
abutment walls and piles, the abutment walls are designed to be sheared off in major seismic events.
Therefore, the force-resistance mechanism of bridge abutments in the longitudinal direction is mainly provided
by backwall-soil interaction, and the passive earth pressure of the structure backfill of the abutments.
Current design practice in California makes use of the bi-linear load-deformation curve and does not
account for the structure backfill properties. This experimental and analytical research program was conducted
to investigate the effect of the structure backfill properties, area of structure backfill, backfill height, and vertical
wall movement on the abutment capacity and stiffness. The experimental program included five large-scale
tests.
The study shows that soil properties, abutment geometry, and the area of structure backfill play
important roles in ultimate capacity and stiffness of bridge abutments, and an evaluation of the current
Caltrans procedure is provided in this context. In addition, forcases where post-peak softening behavior is
deemed to be important in system modeling efforts, an improved soil spring model for predicting the stiffness
and capacity of the bridge abutments in longitudinal direction is proposed.

17. Key Words 18. Distribution Statement

abutments, stiffness, ultimate capacity, experimental test, bridges No restrictions

19. Security Classification (of this report) 20. Security Classification (of this page) 21. No. of Pages 22. Price

Unclassified Unclassified 211

Form DOT F 1700.7 (8-72) Reproduction of completed page authorized

-i-
 DISCLAIMER

The contents of this report reflect the views of the authors who are responsible for the
facts and accuracy of the data presented herein. The contents do not necessarily reflect
the official views or policies of the California Department of Transportation, the Federal
Highway Administration, and National Science Foundation. This report does not
constitute a standard, specification or regulation.

- ii ­
 ACKNOWLEDGEMENTS

This study was made possible by funding from the California Department of
Transportation under contract No. 59A0337 and the Pacific Earthquake Engineering
Research Center, under NSF contract No. 9701568.

- iii ­
 TABLE OF CONTENTS

DISCLAIMER .................................................................................................................... ii
ACKNOWLEDGEMENTS............................................................................................... iii
TABLE OF CONTENTS................................................................................................... iv
LIST OF FIGURES .......................................................................................................... vii
LIST OF TABLES............................................................................................................ xii
ABSTRACT..................................................................................................................... xiii
1 INTRODUCTION .......................................................................................................... 1
1.1 BACKGROUND ................................................................................................ 1
1.2 SCOPE OF RESEARCH .................................................................................... 2
2 LONGITUDINAL STIFFNESS AND CAPACITY OF BRIDGE ABUTMENTS ....... 6
2.1 LITERATURE REVIEW ON SOIL-STRUCTURE INTERATION ................. 6
2.2 PASSIVE EARTH PRESSURE THEORIES................................................... 10
2.2.1 Rankine Theory........................................................................................... 10
2.2.2 Coulomb Theory ......................................................................................... 11
2.2.3 Log Spiral Theory ....................................................................................... 12
2.3 PASSIVE RESISTANCE FORCE-DISPLACEMENT BEHAVIOR .............. 14
2.4 FACTORS THAT CONTROL THE PASSIVE RESISTANCE...................... 15
3 FIELD AND EXPERIMENTAL PROGRAM............................................................. 24
3.1 REVIEW OF EXISTING ABUTMENTS ........................................................ 25
3.2 OVERALL TEST SETUP AND DESIGN OF TEST UNITS ......................... 27
3.2.1 SOIL PROPERTIES ................................................................................... 28
3.2.2 TEST SETUP-PHASE I ............................................................................. 29
3.2.2.1 Test Setup.............................................................................................. 30
3.2.2.2 Material Testing .................................................................................... 33
3.2.2.3 Instrumentation ..................................................................................... 33
3.2.2.4 Loading Protocol................................................................................... 34

- iv -
 3.2.3 TEST SETUP-PHASE II ............................................................................ 35
3.2.3.1 Test Setup.............................................................................................. 36
3.2.3.2 Material Testing .................................................................................... 37
3.2.3.3 Instrumentation ..................................................................................... 37
3.2.3.4 Loading Protocol................................................................................... 38
4 EXPERIMENTAL RESULTS ..................................................................................... 77
4.1 OBSERVATIONS DURING THE TESTS ...................................................... 77
4.2 FORCE-DISPLACEMENT CURVE ............................................................... 80
4.3 DISCUSSION OF RESULTS........................................................................... 83
4.3.1 EFFECT OF VERTICAL MOVEMENT OF THE WALL ........................ 83
4.3.2 EFFECT OF EXCAVATED AREA FOR STRUCTURE BACKFILL ..... 84
4.3.3 EFFECT OF BACKFILL HEIGHT............................................................ 85
5 ANALYSIS OF LONGITUDINALLY LOADED BRIDGE ABUTMENTS ........... 114
5.1 ABUTMENT WALL MODEL....................................................................... 114
5.1.1 PASSIVE EARTH PRESSURE RESISTANCE ...................................... 114
5.1.2 ABUTMENT WALL FORCE-DISPLACEMENT RELATIONSHIP .... 118
5.2 COMPARISON OF MEASURED AND CALCULATED FORCE-
DISPLACEMENT RESULTS ............................................................................... 121
5.3 COMPARISON OF MEASURED AND CALCULATED PASSIVE FAILURE
WEDGE ................................................................................................................. 126
5.4 COMPARISON OF MEASURED AND CALCULATED FORCE-
DISPLACEMENT FOR MARONEY (1995) TEST ............................................. 127
5.5 COMPARISON OF MEASURED AND CALCULATED FORCE-
DISPLACEMENT ON ROLLINS AND COLE’S (2006) TESTS ........................ 128
5.6 SUMMARY OF DESIGN METHOD ............................................................ 128
5.7 ASSESSMENT OF CALTRANS METHOD................................................. 129
5.8 PERFORMANCE-BASED LIMIT STATES ................................................. 132
6 SUMMARY AND CONCLUSIONS......................................................................... 160
6.1 EXPERIMENTAL INVESTIGATION .......................................................... 160

-v-
 6.2 ANALYTICAL STUDY ................................................................................ 163
6.3 SIMPLIFIED CALTRANS BI-LINEAR MODEL ........................................ 164
6.4 RECOMMENDATIONS FOR FUTURE RESEARCH ................................. 165
REFERENCES ............................................................................................................... 167
APPENDIX A - CU TRIAXIAL TEST RESULTS ....................................................... 174
APPENDIX B - REINFORCING DETAILS OF THE TEST BASE............................. 191
APPENDIX C - MAXIMUM DENSITY AND OPTIMUM MOISTURE .................... 194
APPENDIX D - EQUATIONS FOR DEFINING THE CENTER OF LOG SPIRAL
PASSIVE WEDGE..................................................................................................... 197
APPENDIX E - COMPUTING THE ABUTMENT WALL INITIAL ELASTIC
STIFFNESS ................................................................................................................ 202
APPENDIX F - DISTRIBUTION OF SOIL PRESSURE AT WALL SURFACE VS.
DEPTH OF THE WALL............................................................................................ 206

- vi -
 LIST OF FIGURES
Figure 1-1. Bridge seat-type abutment, side view .............................................................. 5
Figure 1-2. Bridge diaphragm abutment, side view ............................................................ 5
Figure 2-1. Elevation and plan views of the large-scale diaphragm abutment tests (after
Maroney 1995).......................................................................................................... 18
Figure 2-2. Soil wedge, Rankine theory ........................................................................... 19
Figure 2-3. Soil wedge, Coulomb earth pressure theory .................................................. 19
Figure 2-4. Log Spiral failure mechanism ........................................................................ 20
Figure 2-5. Force-displacement behavior of bridge abutment (SDC 2006) ..................... 21
Figure 2-6. Hyperbolic passive force versus displacement (adapted from Duncan et al.
2001) ......................................................................................................................... 21
Figure 2-7. Structure and soil movements in passive pressure (after Duncan et al. 2001)
................................................................................................................................... 22
Figure 2-8. Compaction Curve for clayey soils of Yolo Loam tests (after Romstad et al.
1995) ......................................................................................................................... 23
Figure 3-1. Backfill area placed behind diaphragm and seat-type abutment, side view .. 45
Figure 3-2. Map of the site locations for field investigation on abutment structure backfill
material (after EMI report 2005)............................................................................... 46
Figure 3-3. Soil-structure-interaction test facilities, Englekirk Structural Engineering
Center........................................................................................................................ 47
Figure 3-4. Overall view of abutment wall during the construction ................................. 47
Figure 3-5. Map of site location of material imported from Gilman Dr., La Jolla (Photo
from Google Earth) ................................................................................................... 48
Figure 3-6. Map of site location of material imported from Naval Training Center, San
Diego (Photo from Google Earth)............................................................................. 49
Figure 3-7. Excavated area before starting the construction of the abutment wall........... 50
Figure 3-8. Test base during the construction................................................................... 50
Figure 3-9. Overall view of the abutment test unit, Phase I ............................................. 51

- vii ­
Figure 3-10. Reinforcing details of the abutment test unit, Phase I .................................. 52
Figure 3-11. Overall view of installing the formwork of the abutment wall, Phase I ...... 53
Figure 3-12. Overall view of the abutment wall during the construction, Phase I ........... 54
Figure 3-13. Overall view of the abutment wall, ready for concrete pour, Phase I .......... 54
Figure 3-14. Overall view of the abutment after construction, Phase I ............................ 55
Figure 3-15. Excavated area before backfilling, abutment Phase I-Test 1 ....................... 55
Figure 3-16. Overall test setup for abutment Phase I-Test 1 ............................................ 56
Figure 3-17. Top view of the abutment during the backfilling, Phase I-Test 1 ................ 56
Figure 3-18. Overall view of the test unit after backfilling, Phase I-Test 1 ..................... 57
Figure 3-19. Excavated area before backfilling Phase I-Test 2 ........................................ 57
Figure 3-20. Overall test setup before backfilling, abutment Phase I-Test 2 ................... 58
Figure 3-21. Overall test setup for abutment Phase I-Test 2 ............................................ 58
Figure 3-22. Overall view of the test specimen after backfilling, Phase I-Test 2............. 59
Figure 3-23. Excavated area prior of backfilling Phase I-Test 3 ...................................... 59
Figure 3-24. Excavated area in abutment Phase I-Test 3.................................................. 60
Figure 3-25. Overall test setup of abutment Phase I-Test 3 .............................................. 60
Figure 3-26. Overall view of the test specimen after backfilling, Phase I-Test 3............. 61
Figure 3-27. Excavated area prior of backfilling Phase I-Test 4 ...................................... 61
Figure 3-28. Overall view of the excavated area in abutment Phase I-Test 4 .................. 62
Figure 3-29. Overall test setup for abutment Phase I-Test 4 ............................................ 62
Figure 3-30. Overall view of the test specimen after backfilling, Phase I-Test 4............. 63
Figure 3-31. Layout of the instrumentation in the abutment specimen, Phase I .............. 64
Figure 3-32. Overall view of the instrumentation on the test unit, Phase I ...................... 64
Figure 3-33. Locations of the strain gauges of the backwall and wingwalls, Phase I ...... 65
Figure 3-34. Loading protocol of Phase I, Test 1 ............................................................. 65
Figure 3-35. Plastic layers arrangement at wingwalls interface in Phase II, System Test 66
Figure 3-36. Excavated area for building the System Test (seat-type abutment) ............. 66
Figure 3-37. Plan view of the seat-type abutment test unit, Phase II................................ 67
Figure 3-38. Elevation view of the seat-type abutment test unit, Phase II ....................... 68

- viii -
Figure 3-39. Details of reinforcing steel for wingwalls, plan view, Phase II ................... 68
Figure 3-40. Details of reinforcing steel for stemwall and shear keys, Phase II .............. 69
Figure 3-41. Details of reinforcing steel for wingwalls, Phase II ..................................... 70
Figure 3-42. Details of reinforcing steel for backwall, Phase II ....................................... 70
Figure 3-43. Steel fiber used in concrete mix of stemwall and wingwalls, Phase II ........ 71
Figure 3-44. Overall view of the System Test during the construction, Phase II ............. 71
Figure 3-45. One inch thick Styrofoam placed at the interface of backwall-wingwalls,
Phase II...................................................................................................................... 72
Figure 3-46. Overall view of the System Test after construction, Phase II ...................... 72
Figure 3-47. Overall view of the excavated area in Phase II, System Test ...................... 73
Figure 3-48. Overall test setup of the Phase II, System Test ............................................ 73
Figure 3-49. Backfilling of the abutment Phase II, System Test ..................................... 74
Figure 3-50. Overall view of the test unit after backfilling, Phase II ............................... 74
Figure 3-51. Layout of the instrumentation in the abutment specimen, Phase II ............. 75
Figure 3-52. Locations of the strain gauges of the backwall, Phase II ............................. 76
Figure 4-1. Test observation at different displacement level in Phase I-Test 1 ................ 87
Figure 4-2. Test observation at different displacement level in Phase I-Test 2 ................ 88
Figure 4-3. Test observation at different displacement level in Phase I-Test 3 ................ 89
Figure 4-4. Test observation at different displacement level in Phase I-Test 4 ................ 90
Figure 4-5. Test observation at different displacement level in Phase II-System Test..... 91
Figure 4-6. Observed failure plane, Phase I-Test 1 .......................................................... 92
Figure 4-7. Observed failure plane, Phase I-Test 2 .......................................................... 92
Figure 4-8. Observed failure plane, Phase I-Test 3 .......................................................... 93
Figure 4-9. Observed failure plane, Phase I-Test 4 .......................................................... 93
Figure 4-10. Observed failure plane, Phase II-System Test ............................................. 93
Figure 4-11. Horizontal force-displacement, Abutment wall-Test 1 ................................ 94
Figure 4-12. Horizontal stiffness-displacement, Abutment wall-Test 1........................... 95
Figure 4-13. Horizontal force-displacement, Abutment wall-Test 2 ................................ 96
Figure 4-14. Horizontal force-vertical displacement, Abutment wall-Test 2 ................... 97

- ix -
Figure 4-15. Horizontal stiffness-displacement, Abutment wall-Test 2........................... 98
Figure 4-16. Horizontal force-displacement, Abutment wall-Test 3 ................................ 99
Figure 4-17. Horizontal force-vertical displacement, Abutment wall-Test 3 ................. 100
Figure 4-18. Horizontal stiffness-displacement, Abutment wall-Test 3......................... 101
Figure 4-19. Horizontal force-displacement, Abutment wall-Test 4 .............................. 102
Figure 4-20. Horizontal force-vertical displacement, Abutment wall-Test 4 ................. 103
Figure 4-21. Horizontal stiffness-displacement, Abutment wall-Test 4......................... 104
Figure 4-22. Horizontal force-displacement, Abutment wall-System Test .................... 105
Figure 4-23. Horizontal force-vertical displacement, Abutment wall-System Test ....... 106
Figure 4-24. Horizontal stiffness-displacement, Abutment wall-System Test ............... 107
Figure 4-25. Horizontal force-displacement response, Phase I and II ............................ 108
Figure 4-26. Passive failure mechanisms of a wall......................................................... 109
Figure 4-27. Horizontal stiffness-displacement response of Test 2 and 4...................... 110
Figure 4-28. Horizontal force-displacement response of Test 2 and 4 ........................... 111
Figure 4-29. Horizontal stiffness-displacement response of Test 3 and 4...................... 112
Figure 4-30. Horizontal force-displacement response of Test 3 and 4 ........................... 113
Figure 5-1. Shape of failure surface-composed of a Log Spiral and a straight line ....... 141
Figure 5-2. Forces on slice element (after, Shields et al. 1973) ..................................... 141
Figure 5-3. Force-displacement relationship- sheared abutment backwall .................... 142
Figure 5-4. Computed and measured force-displacement curves for the abutment wall in
Phase I-Test 1.......................................................................................................... 143
Figure 5-5. Computed and measured force-displacement curves for the abutment wall in
Phase I-Test 2.......................................................................................................... 144
Figure 5-6. Computed and measured force-displacement curves for the abutment wall in
Phase I-Test 3.......................................................................................................... 145
Figure 5-7. Computed and measured force-displacement curves for the abutment wall in
Phase I-Test 4.......................................................................................................... 146
Figure 5-8. Computed and measured force-displacement curves for the abutment wall in
Phase II-System Test............................................................................................... 147

-x-
Figure 5-9. Computed and measured stiffness-displacement curves for the abutment wall
in Phase I-Test 1...................................................................................................... 148
Figure 5-10. Computed and measured stiffness-displacement curves for the abutment
wall in Phase I-Test 2.............................................................................................. 149
Figure 5-11. Computed and measured stiffness-displacement curves for the abutment
wall in Phase I-Test 3.............................................................................................. 150
Figure 5-12. Computed and measured stiffness-displacement curves for the abutment
wall in Phase I-Test 4.............................................................................................. 151
Figure 5-13. Computed and measured stiffness-displacement curves for the abutment
wall in Phase II-System Test................................................................................... 152
Figure 5-14. Computed and observed passive failure plane ........................................... 153
Figure 5-15. Computed and measured force-displacement curves for the west abutment
wall in Maroney’s Test ........................................................................................... 154
Figure 5-16. Computed and measured force-displacement curves of clean sand, Cole and
Rollin’s Tests .......................................................................................................... 155
Figure 5-17. Computed and measured force-displacement curves of fine gravel, Cole and
Rollin’s Tests .......................................................................................................... 156
Figure 5-18. Computed and measured force-displacement curves of coarse gravel, Cole
and Rollin’s Tests ................................................................................................... 157
Figure 5-19. Computed and measured force-displacement curves of silty sand, Cole and
Rollin’s Tests .......................................................................................................... 158
Figure 5-20. Performance-based limit-states of abutment backwall .............................. 159

- xi -
 LIST OF TABLES

Table 2-2. Values of Δmax /h for different backfill soil types (after Cole et al. 2006) ...... 17
Table 3-1. Gradation requirements for structure backfill ................................................. 39
Table 3-2. Representative soil types for abutment structure backfill from field
investigation (after EMI report 2005) ....................................................................... 40
Table 3-4. Summary of index test results on sample of structure backfill materials........ 41
Table 3-5. Summary of CU test results on samples of abutment structure backfills........ 42
Table 3-6. Structure backfill strength parameters based on CU triaxial tests................... 42
Table 3-7. Bridge abutment research program test matrix, Phase I .................................. 43
Table 3-8. Concrete compressive strengths in Phase I ..................................................... 43
Table 3-9. Bridge abutment research program test matrix, Phase II ................................ 44
Table 3-10. Concrete compressive strengths in Phase II .................................................. 44
Table 5-1. Typical range of soil adhesion factor, β (after Mokwa 1999) ....................... 134
Table 5-3. Input parameters for calculating the force-displacement curves of bridge
abutments ................................................................................................................ 135
Table 5-4. Summary of computed and measured passive resistance .............................. 136
Table 5-5. Computed the initial stiffness of the abutment wall by Caltrans method...... 136
Table 5-6. Summary of Yolo Loam soil properties used for computing passive resistance
of Maroney’s Test. .................................................................................................. 137
Table 5-7. Summary of soil properties used for computing passive resistance of Rollin
and Cole’s Tests...................................................................................................... 138
Table 5-8. Interface friction angles, δ, between different backfill materials and concrete
(after U.S. Navy 1986)............................................................................................ 139
Table 5-9. Computed force and displacement at Limit-State I ....................................... 140
Table 5-10. Computed force and displacement at Limit-State II.................................... 140
Table 5-11 Computed force and displacement at Limit-State III ................................... 140

- xii ­
 ABSTRACT

Abutments are earth-retaining structures which provide resistance to deformation

and earthquake induced inertial forces from the bridge superstructure. In order to limit the
inertial forces transmitted into the abutment walls and piles, the abutment walls are
designed to be sheared off in major seismic events. Therefore, the force-resistance
mechanism of bridge abutments in the longitudinal direction is mainly provided by
backwall-soil interaction, and the passive earth pressure of the structure backfill of the
abutments.
Current design practice in California makes use of the bi-linear load-deformation
curve and does not account for the structure backfill properties. This experimental and
analytical research program was conducted to investigate the effect of the structure
backfill properties, area of structure backfill, backfill height, and vertical wall movement
on the abutment capacity and stiffness. The experimental program included five large­
scale tests.
The study shows that soil properties, abutment geometry, and the area of structure
backfill play important roles in ultimate capacity and stiffness of bridge abutments, and
an evaluation of the current Caltrans procedure is provided in this context. In addition, for
cases where post-peak softening behavior is deemed to be important in system modeling
efforts, an improved soil spring model for predicting the stiffness and capacity of the
bridge abutments in longitudinal direction is proposed.

- xiii -
1 INTRODUCTION

There are two types of bridge abutments generally used in state of California.
Seat-type abutments support the bridge superstructure on a stemwall or “seat”. Typically,
short seat abutments include a backwall which retains the structure backfill material
above the seat. Figure 1-1 shows the schematic of a seat-type bridge abutment. The
backwall is considered to be sacrificial during major seismic event. The backwall is
intended to break off and mobilize the longitudinal resistance of the approach fill (SDC
2006) while protecting the foundation from inelastic response. Diaphragm abutments
consist of an end diaphragm cast integrally with the superstructure and the abutment
stemwall (see Figure 1-2). The diaphragm supports the abutment approach fill under
service conditions, and mobilizes the passive pressure longitudinally during seismic
response. During an earthquake, the bridge superstructure moves longitudinally. Once the
gap is closed, the superstructure collides with the backwall which induces deformation
and inertial force to bridge abutments.

Bridge abutments play an important role in the magnitude and distribution of

earthquake-induced forces transmitted into the bridge structure. Factors such as nonlinear
soil behavior, soil properties, abutment dimensions, and soil-structure-interaction must be
considered for a realistic characterization of abutment capacity and/or stiffness. However,
for simplicity, in many existing models for abutment capacity and stiffness, the effects of
soil nonlinearity and soil properties are not considered.

1.1 BACKGROUND

For earthquake design of bridges, most specifications (e.g. ATC 1996 and
AASHTO 2004) require that the abutment–soil system be included in the analytical
model as a discrete equivalent linear spring. Current design practice in California (SDC
2006) makes use of the bi-linear load-deformation curve established from the large scale
abutment testing conducted by Maroney (1995) and does not account for the structure

1
backfill properties. A review of typical bridge plans showed that a wide range of soil
types are actually used for abutment structure backfills in practice and, therefore different
soil strength and stiffness should be expected in field conditions. It should be noted that
the current bridge design procedure (ATC 1996, SDC 2006) considers the backwall in
seat-type abutments as a sacrificial element in order to protect abutment walls and piles
from damage by limiting the inertial forces that can be transmitted into the abutment.
Therefore, the force-resistance system of the bridge abutments in longitudinal direction
during major seismic events is mainly provided by backwall-soil interaction, and the
passive earth resistance behind the bridge abutments. Neglecting the structure backfill
properties in calculating the abutment stiffness and capacity may result in an unrealistic
prediction of abutment behavior. Based on performance-based seismic design, it is
necessary to have an accurate characterization and prediction of the bridge abutment at
the system and component level. It is also important to consider the interaction of the soil
with structure in evaluating the performance of the structures which is strongly
influenced by the behavior of the soils that support them. Therefore soil-structure­
interaction must be considered for a realistic characterization of abutment capacity and/or
stiffness.

Several researchers have studied the passive earth resistance on pile caps, and
retaining/abutment walls both experimentally and analytically. The results of the
literature review showed that only limited research has been done in the area of abutment
capacity, and there is much uncertainty regarding appropriate modeling of bridge
abutments.

1.2 SCOPE OF RESEARCH

Proper analytical idealizations of boundary conditions in any structural analysis

are essential to obtain accurate results. Therefore, it is important to use proper soil springs
to model abutment-structure backfill resistance in analytical modeling of bridges. In order
to provide an appropriate abutment-soil spring used in bridge analytical model, an

2
experimental research program was conducted at the University of California, San Diego,
funded by the California Department of Transportation (Caltrans). The main goal of this
research program was to investigate the effect of structure backfill properties, the area of
structure backfill, the backfill height, and the vertical wall movement on abutment
capacity and stiffness from field tests.

The research effort is aimed at developing a simplified soil dependent model to

estimate the abutment capacity and stiffness accurately. In order to meet the objectives of
the research project, both field and experimental programs were carried out. The
objective of the field program was to develop a proper characterization of the soil types
used as abutment structure backfill, and then to design the experiments to allow
development of soil springs appropriate for abutment modeling as a function of soil type.

The original objective of the research project was aimed at investigating the
response of diaphragm-type abutments, similar to Maroney (1995), using both
experimental and analytical modeling. At the request of Caltrans (and after the first
experiment was completed), the objective of the research was modified to also investigate
the response of seat-type abutments. Therefore, the experimental part of this research
program was conducted in two phases. In the first phase of the experiment, a diaphragm
abutment (without a foundation) was tested at 50% scale of a prototype abutment (full­
scale height, 50% width), to study the longitudinal stiffness and capacity. In this
experiment, the desired failure mode was geotechnical, not structural. Therefore, the
abutment wall was designed and built to remain in the elastic range during the tests. Four
tests were performed in Phase I with the key variables of soil type, area of the structure
backfill, backfill height, and direction of the wall movement. The test setup was designed
to model the longitudinal behavior of bridge abutments and restrained from translational
and rotational movements.

The second phase of this research program investigated the stiffness and strength
of a seat-type abutment in longitudinal direction. In this test, it was assumed that the

3
backwall was already sheared off from the seat (stemwall) and wingwalls and the only
force-resistance system of the abutment was provided by backwall-soil interaction.
Similar to Phase I, the backwall was designed and built to remain elastic to develop
geotechnical failure during the test.

This report presents recommendations for design and construction details of

bridge abutments based on the experimental results. The results of the experimental
research program and development of the simple analytical model for capacity evaluation
of bridge abutments in longitudinal direction are included in this report.

4
 Superstructure Backwall Wingwall

Structure
Backfill Backfill
Material

Seat Embankment Material

Pile cap
Ground level
Piles

Figure 1-1. Bridge seat-type abutment, side view

Backwall

Superstructure
Structure
Backfill Backfill
Material

Wingwall

Embankment Material
Pile cap

Ground level
Piles

Figure 1-2. Bridge diaphragm abutment, side view

5
2 LONGITUDINAL STIFFNESS AND CAPACITY OF
BRIDGE ABUTMENTS

This chapter addresses four topics. The first is a review on published experimental
and analytical research related to the capacity of bridge abutments, the passive earth
resistance on pile caps, retaining/abutment walls, and soil-structure interaction. The
second topic discusses about the passive earth pressure theories, addressing the
advantages and disadvantages of each theory. As part of studying the behavior of
longitudinally loaded bridge abutments, it is essential to review the knowledge regarding
force-displacement behavior of the soil under passive pressure. Therefore, third topic is
addressing the most recognized analytical models to predict the passive force­
displacement curves of soils. Finally the forth section discusses about the factors that
control the magnitudes of passive earth pressure.

2.1 LITERATURE REVIEW ON SOIL-STRUCTURE INTERATION

In the new design procedure, the abutment backwall is designed to shear off (seat-
type abutment) or respond in a flexible manner during strong earthquakes in order to
protect abutment piles from severe damage under earthquake induced forces (SDC 2006).
Therefore, the capacity of the bridge abutment in a seismic event is developed mainly
from the mobilized passive pressure behind the abutment wall.

A comprehensive literature review was performed to establish the state of

knowledge regarding the bridge abutment capacity in longitudinal direction. The
literature search was focused in the direction towards experimental and analytical studies
related to the capacity of bridge abutments; the passive earth resistance on pile caps, and
retaining/abutment walls; soil-structure interaction.

Various analytical models have been proposed by several researchers to predict

the behavior of bridge abutments in longitudinal directions. The first approach is to use
linear soil springs to model bridge abutments. Many advanced approaches, developed

6
subsequently, consider soil-structure interaction, and effect of nonlinear soil behavior.
Wilson (1988) proposed a theoretical model for determining abutment stiffness based on
the abutment dimensions and soil properties. The model allows an evaluation of discrete
elastic spring stiffness and does not include the effect of nonlinear soil behavior.

Maroney (1995) tested two large scale diaphragm bridge abutments to failure at
University of California, Davis. Both abutments were tested under cyclic loading, the
shorter (west) one tested to failure. The taller (east) abutment was tested in longitudinal
and transverse directions while the west abutment was tested only in the longitudinal
direction. Both abutments had 9-inch cast-in-drilled hole (CIDH) piles and wingwalls as
shown in Figure 2-1.

The soil used for the west embankment was compacted clay referred to as Yolo
Loam, having a plastic limit (PL) of 23.5% and a plasticity index (PI) of 10.5. The shear
strength of compacted clayey soils is very sensitive to water content. Thus, the
unconfined compression tests were done on samples varying water contents. The shear
strength was obtained by dividing the unconfined compression strength by 2. Table 2-1
shows the results of the unconfined compression tests. The compaction curve of the
clayey soils for the Yolo Loam tests is shown in Figure 2-8. Unwashed well graded silty
sand was used for the structure backfill of the west abutment. The embankment and
structure backfill consisted of poorly graded, and coarse clean sand compacted at 90%
relative compaction, and its friction angle was estimated to be 40º. It was recognized that
these materials were a reasonable representation of what is used in California to construct
embankments.

The test setup was designed to apply the load at mid-height on the abutments in
longitudinal direction; where as in the transverse test the load was applied near the top. In
order to provide the secant stiffness of the abutment wall, the lateral resistance
contribution from the piles were subtracted from the recorded results. Maroney (1995)
provided two load responses in longitudinal direction against two different embankment

7
materials (Yolo loam, clean sand) and one in transverse direction against a clean sand
embankment. The details on the test setup, instrumentation, and results are provided by
Romstad et al. (1995) and Maroney (1995). Important results from the longitudinal west
abutment test are listed below:

1. Based on the ultimate static force developed in west abutment test, the maximum
habut
passive resistance is equal to 5.5 ksf. The height proportionality factor, ,
5.5 ft
should be considered in design of abutment with different height.

2. Stiffness of 200 kip/in/ft was identified to be significantly high for typical

deformation. The measured stiffness was approximately 300 kip/in at one half
inch superstructure displacement where, as-practiced predicted stiffness was 1375
kip/in.

The current Caltrans method to calculate the stiffness and capacity of bridge
abutments is based on the results of Maroney’s tests. This method captures the basic
abutment backfill load-deformation behavior reasonably well for clayey soil with a wall
height of 5.5 feet.

A simplified model to evaluate nonlinear longitudinal bridge abutment spring

stiffness has been developed by Siddharthan et al. (1995). Martin et al. (1997) conducted
advanced theoretical studies using a two-dimensional explicit finite difference computer
program (FLAC) to characterize the load-deformation behavior of bridge abutments
under cyclic loading.

Extensive experimental work associated with passive earth pressure has been
conducted by several researchers such as Rowe and Peaker (1965); Narain et al. (1969);
James and Bransby (1970); Fang et al. (1994 and 1997). There have been several full­
scale tests conducted with interest in passive resistance recently (Romstad et al. 1995;
Gadre 1997; Rollins and Sparks 2002; Duncan and Mokwa (2001)). Duncan and Mokwa

8
(2001) conducted several full-scale tests on pile caps and found that the passive
resistance of the backfill soil contributes up to 50% of the total pile cap resistance.
Rollins and Cole (2006) performed a series of static cyclic lateral load tests on a full­
scale on a pile cap. They proposed a new model to account for the passive p-y
relationship for cyclic loading condition where the stiffness of the backfill soil type was
degraded by repeated loading and a gap between the pile cap and backfill soil.

Shamsabadi et al. (2005) proposed a method to predict the mobilized force­

displacement-capacity for the seismic design of a bridge abutment-embankment system.
The method is based on a limit equilibrium logarithmic spiral, method of slices, coupled
with characterization of the stress-strain behavior of the soil.

Finite element is a numerical approach to model soil-structure interaction by

considering the soil as quasi-elastic continuum. Several researches used finite element
method to study soil-structure interaction problems. Clough and Duncan (1971)
developed procedure for representing the interface between a structure and adjacent soil
in finite element analysis of soil-structure interaction. Briaud et al. (1997) performed a
three-dimensional nonlinear finite element analysis to study the behavior of soil-nailed
walls constructed under a bridge abutment built on piles. The piles and nails were
modeled with beam elements, while the abutment and the pile cap were modeled with
solid 3D elastic elements. A modified hyperbolic model with unloading hysteresis was
used to simulate soil in the numerical analysis. Faraji et al. (2001) implemented a full 3D
finite element model of an integral abutment bridge system which was incorporated the
nonlinear soil response. Yang and Boris Jeremic (2003) conducted a numerical study on
the effective stiffness for pile groups. The OpenSees finite element framework was used
to simulated the response of 3x3 and 4x3 pile groups founded in loose and medium dense
sands. Shamsabadi et al. (2006) used finite element program, Plaxis to study the stiffness
degradation (nonlinear load-deformation curve) for the earth pressure behind bridge
abutments. The Hardening Soil Model which has an expendable yield surface was used to
simulate the nonlinear abutment backfill force-deformation relationship.

9
2.2 PASSIVE EARTH PRESSURE THEORIES

A wide variety of methods are available to determine the capacity provided by

passive pressure against the retaining structures. These methods include the classical
approaches such as Log Spiral (Terzaghi 1943, Terzaghi et al. 1996), Rankine, and
Coulomb, which are ultimate capacity predictors and do not attempt to capture stiffness
behavior. In Rankine’s and Coulomb’s theory, it is assumed that the failure surface in the
backfill is planar. The fundamental assumptions of Rankine and Coulomb theories are
that the retained soil is homogenous, isotropic, semi infinite (wall is very long and soil
goes back a long distance), and well drained to avoid formation of pore pressure.
However, Terzaghi (1943) and Terzaghi et al. (1996) showed that, due to the wall
interface friction the real failure surface consists of a curved lower part and a straight
upper part (see Figure 2-4). The soil within the area ECD is considered to be in the
passive Rankine state. The curved part of the surface was assumed to have a logarithmic
spiral shape. A summary of these three earth pressure theories are described in following
paragraphs. A good discussion and more details about these earth pressure theories are
offered by Bowles (1988).

2.2.1 Rankine Theory

The Rankine theory assumes that there is no wall friction (δ=0). For a level
backfill, the resultant force acts horizontally. For the more general case of a sloping
backfill, the resultant force is assumed to act parallel to the backfill slope (Das, 2005).
The Rankine coefficient of passive pressure is given by the following expression:

⎡ cos β +
K p = cos β ⎢
(cos 2
β − cos 2 φ ) ⎤
⎥ (2.1)
⎢⎣ cos β − (cos 2
β − cos 2 φ ) ⎥⎦

10
where β is the embankment slope angle and φ is the soil internal friction angle. Figure 2-2
shows the schematic of the Rankine’s passive soil wedge. The coefficient of passive earth
pressure is the term used to express the ratio between the effective lateral passive earth
pressure to the vertical earth pressure in the condition that the wall is gradually pushed
into the soil mass. For cohesionless soils, the resultant force per unit width, Fp, on a wall
is given by the equation:

F p = 0.5γh 2 K p (2.2)

And for cohesive soils:

F p = 0.5γh 2 K p + 2ch K p (2.3)

where γ=total unit weight of the backfill; h=wall height; c= cohesive strength of the soil;
and Kp= coefficient of passive earth pressure. The resultant passive earth resistance is
assumed to act at a distance of h/3 measured from bottom of the wall. It should be noted
that the passive pressure computed by Rankine method tends to be conservative
especially for the cases with large wall friction values.

2.2.2 Coulomb Theory

The coulomb theory provides a method that computes the resultant force on a
retaining system for any slope of wall, wall friction, and slope of backfill provided β <φ.
The Coulomb theory assumes that the soil shear resistance develops along the wall and
failure plane. The Coulomb coefficient of passive pressure is given by:
cos 2 φ
Kp = (2.3)
sin (φ + δ )sin (φ + β ) ⎤
2
⎡
cos δ ⎢1 − ⎥
⎣ cos δ cos β ⎦

11
where β is the angle of backfill slope, φ is the soil internal friction angle, and δ is the
angle of wall friction. Figure 2-3 shows the soil wedge computed by Coulomb earth
pressure theory. Coulomb’s method accounts for the wall friction; however it assumes a
planar failure surface. As a result, values of Kp computed using the Coulomb theory are
too high when the wall friction angle is greater than half of the backfill soil friction angle
(Duncan and Mokwa 2001).

For cohesionless soils, the resultant force per unit width, Fp, on a wall is given by
the equation:
F p = 0.5γh 2 K p (2.4)

And for cohesive soils:

F p = 0.5γh 2 K p + 2ch K p (2.5)

where γ=total unit weight of the backfill; h=wall height; δ=angle of wall friction; c=
cohesive strength of the soil; and Kp= coefficient of passive earth pressure. Based on Eq.
(2.4), the passive pressure distribution is linear and the resultant passive earth resistance
acts at a distance of h/3 measured from bottom of the wall. Fang et al. (1994) showed
that, for a wall under translational movement, the passive earth pressure is linear and in
good agreement with Terzaghi’s prediction based on the general wedge theory.

2.2.3 Log Spiral Theory

The Log Spiral method is generally considered to be the most theoretically correct
method for computing lateral earth pressure (Fang et al. 1994; Gadre 1997; Duncan and
Mokwa 2001; Rollins and Sparks 2002). The method accounts for the curved failure
surface and for friction between the backfill and the wall (Figure 2-4). Log Spiral earth
pressure forces can be computed by the developed charts and tables (available in many
foundation engineering text books and manuals) to estimated the earth pressure
coefficient, Kp. The alternative approach is to calculate passive forces by a trial and error

12
Log Spiral graphical process (Terzaghi 1943) based on the assumption that a surface
traction makes an angle of φ with the tangent to the spiral surface, and the lines of the
surface traction pass through the center of the spiral (see Figure 2-4). The advantage of
this approach is that it can also account for cohesion in c-φ soil. However, it requires
considerable time and effort. Numerical analysis is another way of applying the Log
Spiral theory in computing the passive pressure.

An EXCEL spreadsheet PYCAP was developed by Duncan and Mokwa (2001) to

calculate the passive pressure using the Log Spiral theory. The passive earth pressure
force per unit width of the wall, Ep, includes three components: 1) the component due to
soil weight and friction, Ppφ, 2) the component due soil cohesion, Ppc and 3) the
component due surcharge, Ppq. The passive earth pressure component Ppφ, Ppc, and Ppq can
be determined by using equilibrium. The individual components can be combined as
follows:

E p = Ppφ + Ppc + Ppq (2.6)

Each passive pressure component can be expressed in more traditional form by the
passive earth pressure coefficient. Combining these three components, the ultimate
passive force per unit width, Fp, on a wall as can be expressed as:

1 2
Fp = γh K pφ + 2chK pc + qhK pq (2.7)
2

where the Kpφ is the passive earth pressure coefficient based on soil friction and weight;
Kpc is the passive earth pressure coefficient based on soil cohesion; Kpq is the passive
earth pressure coefficient for surcharge; γ=total unit weight of the backfill; h=wall height;
δ=angle of wall friction; c= cohesive strength of the soil. A good discussion and
summary of this numerical analysis are described by Mokwa (1999).

13
2.3 PASSIVE RESISTANCE FORCE-DISPLACEMENT BEHAVIOR

The variation of passive resistance with displacement can be modeled by

analytical expressions. Several methods have been proposed to characterize the
development of passive pressure with displacement; including the Caltrans method, and
the hyperbolic model given by Duncan and Mokwa (2001).

The Caltrans method is based on the results from large scale abutments testing at
University of California Davis (Maroney 1995). Caltrans (SDC, 2006) suggests the initial
kip / in
longitudinal abutment stiffness to be equal to 20 .
ft

The initial stiffness must be adjusted proportional to the backwall height as:

kip/in ⎛h ⎞
k abut = k i (= 20.0 ) × w × ⎜ abut ⎟ (2.8)
ft ⎝ 5.5 ⎠

where, w is the width of the backwall. The ultimate capacity of the abutment is given by
Eq. (2.9). The maximum passive resistance of 5 ksf in Eq. (2.9) is based on the ultimate
static force developed in large scale abutment testing at University of California, Davis.
habut
The height proportionality factor, , is based on the abutment test specimen
5.5 ft
height (5.5 ft) used at UC Davis (Maroney 1995). In Eq. (2.9), Ae is the effective
abutment area. The passive pressure resisting the movement at the abutment increases
linearly with the displacement, as shown in Figure 2-5.

habut
Pabut = Ae × 5.0 ksf × (2.9)
5.5

14
 Figure 2-6 shows the hyperbolic representation of passive resistance-displacement
relationship developed by Duncan and Mokwa (2001).The hyperbolic p-y curve
developed by Duncan and Chang (1970) is expressed as:

y
P= (2.10)
1 y
+ Rf
k max Pult

where P is the load at any displacement y, Pult is the ultimate passive force (using Log
Spiral method); kmax is the initial stiffness which corresponds to the initial slope of the
load deflection curve. This value can be approximated using elasticity theory. The
approach used to compute the kmax is given in Appendix E. The failure ratio, Rf, is defined
as the ratio between the actual failure force and the hyperbolic ultimate force, which is an
asymptotic value that is approached as y approaches infinity. For soil stress-strain curves,
Rf is always smaller than one, and varies from 0.5 to 0.9 for most soils (Duncan et al.
2001). The value of Rf can be estimated by substituting Pult for P, and Δmax for y. Re­
arranging the term in Eq. (2.10) results in the following expression for Rf:

Pult
Rf = 1− (2.11)
Δ max k max

The required movement to mobilize the maximum passive resistance, Δmax, has been
investigated by several researchers experimentally and numerically. Movement necessary
to mobilize the maximum passive earth pressure suggested for different types of backfill
are given in Table 2-2.

2.4 FACTORS THAT CONTROL THE PASSIVE RESISTANCE

The magnitude of the passive earth pressure that resists the movement of the wall
is controlled by several factors (Duncan and Mokwa 2001): (1) the direction in which the
wall moves, (2) strength and stiffness of the soil, and (3) wall-soil interface friction and

15
adhesion. For a given amount of movement, a soil type with a greater stiffness and
strength develop more passive resistance. Cohesion and friction are the soil properties
that contribute in stiffness and strength of soils. Interface friction and adhesion create
additional shear stresses on the soil-structure interaction which should be considered in
determining the passive resistance.

If the wall moves horizontally (vertical restraining force in the wall is greater than
the vertical component of the passive pressure force), slip will occur on the interface
between the structure and soil, and the value of the interface friction, δ, will be controlled
by the properties of the soil-structure interface (Figure 2-7a). This would be the case of
diaphragm abutment that is restrained vertically by the pile foundation. If the vertical
restraining force in the wall is smaller than the vertical component of the passive pressure
force (the backwall is sheared off from the stemwall in seat-type abutment and there is no
vertical restraining force in the backwall), smaller relative displacement across the
interface will occur which result in only partial mobilization of the interface friction
(Figure 2-7b). Therefore, the value of δmob is controlled by the requirements to satisfy
vertical equilibrium. The value of δmob must satisfy the vertical equilibrium as following:

Wab = E p sin(δ mob ) (2-12)

where δmob is the mobilized friction angle, Wab is the weight of the structure, and Ep is the
developed passive pressure.

When the backfill soil is cohesive, the interface adhesion may cause additional
shear resistance on the soil-structure interface. The adhesion, αc, is expressed as a factor
of maximum adhesion which is equal to the cohesion of the soil, c. The conventional
value for the term, α, varies from 0.5 for stiff soils to about 0.9 for soft soils (Duncan and
Mokwa 2001).

16
 Table 2-1. Laboratory unconfined compressive strength tests of Yolo Loam to obtain
undrained shear strength (after Romstad et al. 1995)

γd Water Content cu=qu/2

(pcf) (%) (tsf)

104 16 0.4
104 16.5 1
106 15.7 1.9
109 11.8 4.6
115 12.6 3.55

Table 2-2. Values of Δmax /h for different backfill soil types (after Cole et al. 2006)

Backfill Soil Type Δmax/h*

Dense Sand 0.01
Medium-Dense Sand 0.02
Loose Sand 0.04
Compacted Silt 0.02
Compacted Lean Clay 0.05
Compacted Fat Clay 0.05
*
Δmax is usually expressed as a function of the height of the retained structure (h)

17
 CL CL
9” 9”
8’ 8’
Hydraulic Ram 1’6”

5’6” 6’ 9”
1’6”

Yolo Loam Sand Embankment

Embankment

CL 10”
9”

Yolo Loam Sand Embankment 15’ 6”

10’
Embankment

9”
10”

West Abutment East Abutment

Figure 2-1. Elevation and plan views of the large-scale diaphragm abutment tests (after
Maroney 1995)

18
 Figure 2-2. Soil wedge, Rankine theory

Figure 2-3. Soil wedge, Coulomb earth pressure theory

19
 o

A α = 45 − φ/2
C E
θ α
α
EP W
h δ hd

D Rankine zone
B
c
φ f = surface traction

Log Spiral
surface

Figure 2-4. Log Spiral failure mechanism

20
 Force

Pabutment

kabutment

Δeffective Displacement

Figure 2-5. Force-displacement behavior of bridge abutment (SDC 2006)

Force Kmax

1
Pult

Hyperbolic:
y
P=
⎡ 1 y ⎤
⎢ + Rf ⎥
⎣ k max Pult ⎦

Displacement

Figure 2-6. Hyperbolic passive force versus displacement (adapted from Duncan et al.
2001)

21
 Mech
Mechanism
anism 1 Mec
Mech
hanism 2

Soi
Soil Soil
mov
move ement mo
movveme
ment
nt
Stru
Structur
cture
e Struct
Structur
ure
e
mov
move ement
ment move
mov ement

Ep Ws Ep Ws
δmob δmob
mob
R mob R
φmob
mob
φmob
mob

T Ep
Wab W ab δmob
mob
Ep T

(a) (b)

Figure 2-7. Structure and soil movements in passive pressure (after Duncan et al. 2001)

22
 116

114
Dry unit weight, γd (pcf)

112

110

108

106

104

102
10 11 12 13 14 15 16 17

Moisture content, w (%)

Figure 2-8. Compaction Curve for clayey soils of Yolo Loam tests (after Romstad et al.
1995)

23
3 FIELD AND EXPERIMENTAL PROGRAM

In order to meet the objectives of the research project, both a field and
experimental program were carried out. The objective of the field program was to
develop a proper characterization of the soil types used for abutment structure backfills
and its potential variation in the field. The result of the field investigation helped
determine the type structure backfill soils to be used for the tests properly.
As mentioned earlier, the original objective of the research project was aimed at
investigating the response of diaphragm-type abutments, similar to Maroney (1995),
using both experimental and analytical modeling. At the request of Caltrans, the
objective of the research was modified to also investigate the response of seat-type
abutments. Therefore, the experimental part of this research program was conducted in
two phases. In the first phase, an abutment wall (without a foundation) was tested at full
height-scale and 50% width-scale of a prototype diaphragm abutment. In the second
phase, a large scale seat-type abutment was tested. The test units were modeled to reflect
typical diaphragm and seat-type bridge abutments commonly used in practice. The result
of field investigation, modeling, design, and test setup for the experimental testing of the
abutment test units are presented in this chapter. The specific aims of the experimental
program were to examine the effect of:

1) structure backfill soil type

2) backwall height
3) restraining the vertical movement of the wall
4) pre-existing weak planes (pre-existing cut slope) for typical diaphragm abutment
construction

on stiffness and capacity of the abutments in longitudinal direction. The bridge abutment
tests were conducted at the field test facility from early January 2006 through June 2006.

24
3.1 REVIEW OF EXISTING ABUTMENTS

The current design practice for ordinary bridges assumes that the abutment
backwall shears off in the early stage of a major seismic event and that the abutment wall
does not contribute to resist external loads. The only load-resisting source is the soil
behind the wall; therefore, it is important to understand what types of soil have been used
as structure backfill in bridge abutments. The goal of the field program was to collect
sufficient abutment backfill material properties to develop “typical” soil parameters used
statewide.

The structure backfill material should be placed in the area between the
wingwalls. The structure backfill area and embankment area for a diaphragm abutment
and a seat-type abutment are shown in Figures 3-1(a) and 3-1(b), respectively. The dash
line shows the border between the embankment material and structure backfill material.
The current Caltrans Standard Specifications (Caltrans, 2002) contains gradation
requirements (Table 3-1), and a 95% relative compaction requirement for the
construction of an abutment structure backfill. An additional requirement calls for the
structure backfill to be a sandy soil is the Sand Equivalent (SE), which must be greater
than twenty, determined by Caltrans Test Method 217 (CTM, 1999). However, Caltrans
Standard Specifications (2002) requires a 90% relative compaction requirement for the
construction of an embankment, and recommends using embankment material with SE
greater than ten. Based on specified requirements, it can be noticed that there is a wide
range of soil types can be used for structure backfills and especially for embankment
materials based on requirement presented in Caltrans Standard Specifications (2002).

A review of several hundred bridge plans showed there are many different soil
types have been used for structure backfills and embankments. Proper characterization of
the soil types used for abutment structure backfill and its potential variation in the field is
important to improve the current abutment design procedures. One-hundred bridges,
throughout the state, were selected to develop “typical” abutment structure backfill

25
characteristics. Fifteen bridge sites were selected for field investigation. Figure 3-2 shows
the map of site locations. The selection process considered several factors, including:
1) abutment type
2) geographic distribution
3) structure backfill soil type
4) accessibility and safety for soil boring
5) fill height that could be sampled and tested
6) ease of proving permits considering the time constraint

Site investigations were performed by Earth Mechanics Inc. at eleven bridge sites
with a total of twenty-six soil borings, twenty-two CPT’s and fourteen PMT’s. Caltrans
provided bulk samples from four additional sites which were tested in the laboratory as
well. Each site included one to three soil borings with soil sampling, laboratory testing,
CPT and pressuremeter tests. For each site, representative soil types were determined
from the soil samples in the upper fifteen feet of the boreholes. The results from the field
investigation showed a quite wide range of soil types have been used as structure backfill
materials in bridge abutments in state of California. The results from field and laboratory
data were then evaluated and soil types were sorted into four soil categories with
representative geotechnical properties. Table 3-2 defines these four categories (clean
sands, dirty sands containing fines, clayey sands, and cohesive soils). The field
investigation report by Earth Mechanics Inc. (EMI Report 2005) provided the details of
the geotechnical properties measured in the field and laboratory. After review the report,
the data was grouped based on grain size distribution, sand equivalent, and plasticity
index into three key soil categories: sands, silty/clayey sands, and lean clay. Table 3-3
defines these three categories. It can be noticed that a wide range of soil types have been
used for structure backfills.

26
3.2 OVERALL TEST SETUP AND DESIGN OF TEST UNITS

A soil-structure-interaction test facility has been constructed at the University of

California, San Diego Englekirk Structural Engineering Center. As shown in Figure 3-3,
the test facility consists of a refillable soil pit with dimensions of 70 ft by 70 ft and a
depth of 20 ft. A movable reaction wall was constructed next to the soil pit (Figure 3-4).
The movable reaction wall with a height of 16 feet consists of four 4-foot concrete
reaction blocks which are post-tensioned to a deep pile foundation. The reaction wall has
a moment capacity of 8000 kip-ft. The facility is capable of investigating the factors that
control the longitudinal resistance of bridge abutments.

Prior to construction, it was decided to fill the pit with the same material used to
construct the abutment embankment. Caltrans geotechnical engineers and field material
inspectors were consulted in the selection of the materials to use for the embankment and
structure backfill. It was decided to select the structure backfill materials which match
with the second category of proposed soil types in Table 3-3, given that more bridges
with silty/clayey sand backfill found in the field invetigation. It should be noted that
Maroney (1995) performed a large-scale abutment test using clayey silt material for
backfill which matches with the third category in Table 3-3. In order not to repeat the test
with same material for backfilling, the third soil category in Table 3-3 was not considered
in the selection of material for the structure backfill.

Two different type of soils was imported to be used as structure backfill in this
research project, clayey sand and silty sand. Besuase of geographic, time, and budget
limitation in selecting the structure backfill materials, it was decided to consider just
materials which provides the lower bound capacity and stiffness. However, it was found
that the clayey sand material which was high in shear strength and stiffness provided the
upper bound capacity. Thus, the silty sand material was used as structure backfill material
for fours tests, where the clayey sand was used as structure backfill material just in Test1,
Phase I. In addition, as can be seen in Table 3-2, sandy soil was found in just two bridges

27
structure backfill. It is believed that using sandy soil as structure backfill with SE of 40 or
more is very costy and hard to find. The two different types of soil imported for the
project were:
• Clayey sand imported from Gilman Dr., La Jolla close to UCSD campus (Figure
3-5). This soil was used for filling the soil pit, the abutment embankment in all the
tests, and the abutment structure backfill in Test 1, Phase I
• Silty sand imported from Naval Training Center in Rosecrans Street in San Diego
(see Figure 3-6). This soil was used for the abutment structure backfill in Test 2,
3, and 4 in Phase I and abutment System Test in Phase II.

The soil pit was filled with clayey sand which was used as the structure backfill
material in Test 1. This material was compacted to a relative compaction of 90% or more.
The embankment was constructed in accordance with 2002 Caltrans Standard
Specifications. The loose thickness of each layer of embankment material was 6 to 8
inches. The material was compacted to 95% relative compaction according to ASTM
D1557. After compacting each layer nuclear density gage tests were conducted to
monitor the density of the compacted layers. The embankment was constructed using
clayey sand for all of the tests, but the structure backfill material was changed through
Test 1 to Test 2. The embankment was constructed during the period of April 24-27,
2005.

3.2.1 SOIL PROPERTIES

The purpose of the laboratory testing program was to develop soil parameters that
will be useful in performing analysis of the large-scale abutment tests. The laboratory
tests included soil classification, unit weight, sand equivalent tests, strength tests (direct
shear test and CU triaxial test) and Atterberg tests. Laboratory tests were performed on
soil samples obtained from the site during the compaction and on samples taken from the
failure plane after the tests. Table 3-4 describes the Index properties and unit weight for
each soil type. The properties of clayey sand had a good agreement with the second
category of the proposed soil in Table 3-3. The silty sand material was non-plastic and

28
Plasticity Index was lower than the proposed one for the second group of soil in Table 3­
3.
A total of three triaxial tests were performed on each soil type on remolded
samples to 95% relative compaction. The CU specimens were saturated by applying back
pressure. Summaries of the CU test results are presented in Table 3-5. Values of φ' and c'
determined from these tests are shown in Table 3-6. The complete report of laboratory
tests is included in the Appendix A.

3.2.2 TEST SETUP-PHASE I

The first phase of the experimental program was conducted on the performance of
a component of a bridge abutment. An abutment wall (without a foundation) was built
approximately at full height-sclae and 50% width-scale of a prototype diaphragm
abutment. The abutment wall was used for all four tests in Phase I. In this experiment, the
desired failure mode was geotechnical, not structural. Therefore, the abutment wall was
designed and built to remain in the elastic range during these tests. The key variables in
Phase I were structure backfill type, backfill height, the area of structure backfill, and the
vertical wall movement as shown in Table 3-7. The wall was restrained from rotational
movement about three directions. Also, in Test 1, the wall was restrained vertically by
means of proper configuration of actuators.

Geometry and Reinforcement Details of the Abutment Wall Test Unit

The test setup was designed to model the longitudinal behavior of bridge
abutments, restrained from translational and rotational movements. The test base,
consisting of two concrete blocks, was post-tensioned to the reaction wall to prevent any
movement of the test base. In detailing the test base, minimum reinforcement quantities
were found to be sufficient. The abutment wall was placed on the test base, with no
structural connection between the abutment wall and test base. Appendix B contains
detailed information about the test base (concrete blocks).

29
 It was intended to follow the same construction phases of real bridge abutments in
this project. The purpose was to study the influence of construction phases on bridge
abutment behavior. Based on common practice, the structure backfill material is placed at
a very limited area restrained between the wingwalls. There are two key issues related to
have a small structure backfill area. First, it is important to ensure that the passive failure
wedge would occur within the structure backfill area and it is not extended to the
embankment material (it happened in west abutment in Maroney’s test 1995). Second, as
it was mentioned previously, the cut slope resulting from the excavation prior to
placement of the structure backfill may introduce a potential weak plane of failure to the
system which mobilizes less passive resistance in the system. Therefore, it was decided to
extend the excavation area of structure backfill in Test 3 and 4 to prevent any potential
weak plane of failure in the test setup.

Prior to the construction of the test base and test unit, the contractor was asked to
excavate the south part of the embankment in order to open sufficient space for building
the test unit (Figure 3-7). The test base reinforcement cages were built and completed by
casting using standard concrete mix with a target compressive strength of 4000 psi.
Figure 3-8 shows the test base during the construction. The abutment test unit steel
reinforcement was instrumented with strain gages. The geometry and general reinforcing
steel details of the abutment wall are shown in Figure 3-9 and Figure 3-10, respectively.
The test unit was built against the cut slope of the excavated area (Figures 3-11 to 3-14).
The construction of the abutment test unit was completed by August 2005. No load was
applied to the concrete structure until the 28 day concrete curing period was complete.

3.2.2.1 Test Setup

Phase I, Test 1

The lateral load was applied to the abutment wall by set of hydraulic actuators. In
Test 1, the setup of the actuators restrained any upward movement of the abutment wall
to simulate the diaphragm abutment wall with fixed connection to its foundation. The

30
actuators’ setup in the rest of the tests in Phase I, allowed the vertical movement of the
abutment wall to simulate the backwall sheared off from the stemwall in seat-type
abutments. Figure 3-15 shows the test setup before placing the structure backfill, and
Figure 3-16 shows the overall test setup after placing structure backfill in Test 1. The
dash line in dark brown color in Figure 3-16 shows the top of the existed cut slope before
placing structure backfill.

In order to resemble the interface condition of the abutment wall in the field, the
inside surface of the test unit backwall and wingwalls were covered with prefabricated
plastic sheet drainage system suggested by Caltrans. However, after the first test, it was
decided not to use these sheets for the rest of the tests since there was no drainage
problem in San Diego in the period of testing. Also it helped to eliminate the unknown
effect of the sheets on wall interface friction angle. The structure backfill material for
Test 1 was placed on December 15 and 16, 2005.

Prior to placement of structure backfill, the existing cut slope from the excavation
was roughened by the contractor using a backhoe. The loose thickness of each layer of
structure backfill material was 4 inches. The structure backfill material was compacted
using a front loader and a hand operated compactor (“Wacker”). A moisture-density
curve for the clayey sand, shown in Appendix C, indicated a maximum dry density of
126 pcf at water content of 10.5%. Nuclear density gauge tests were performed after
compacting each lift of backfill. Figure 3-17 shows the test unit during the backfilling.
During the backfilling the actuators were connected to the abutment wall to hold the wall
in place and prevent any movement of the wall. Figure 3-18 shows the overall view of the
test setup after placing the structure backfill. Actuators were covered by a tarp for
protection against moisture.

Phase I, Test 2

After completing Test 1, the structure backfill material was excavated as shown in
Figure 3-19 and replaced by silty sand. After consultation with the Caltrans abutment

31
committee, it was decided to focus further testing on seat-type abutments. It was found
that the built specimen tested in Test 1 was suitable for rest of the tests and the wingwalls
did not have any significant effect on test results. In order to simulate the condition in
which the backwall is sheared off from the stemwall in seat-type abutments, it was
necessary to design the setup to allow for vertical movement of the abutment wall. Figure
3-20 illustrates the test setup before placing structure backfill. The overall test setup for
abutment Test 2 is depicted in Figure 3-21. The lateral load was applied to the abutment
wall by two 220-kip actuators, and two 500-kip actuators. Backfilling procedure was
identical to that in Test 1. Appendix C provides a moisture-density curve for the silty
sand with a maximum dry density of 127 pcf, at water content of 8.6%. Nuclear density
gauge tests were performed after compacting each lift of backfill. During the backfilling
the actuators were holding the wall in place and preventing any movement of the wall.
Figure 3-22 shows the overall view of the test setup after placing structure backfill in the
field.

Phase I, Test 3

As mentioned previously, it was decided to extend the excavation area for

structure backfill in Test 3 and 4 to prevent any potential weak plane of failure in the test
setup. Figures 3-23 and 3-24 show the test setup after excavation. It is believed that
resting of wingwalls on the embankment slope (see Figure 3-23) did not affect the test
results since the passive earth resistance was provided by a very large homogenous area
of structure backfill material and the pre-existing embankment slope just beneath the
wingwalls could be neglected. The setup of actuators in Test 3 was similar to that in Test
2 which allowed the upward movement of the abutment wall. Figure 3-25 shows the
overall test setup of the Test 3.

The structure backfill material was compacted in accordance with Caltrans

Standard Specification (2002), using the similar procedure to Test 1 and 2. A moisture­
density curve was just the same as the one in Test 2, Test 3, and Test 4 since similar soil
was used for structure backfill in these tests. Nuclear density gauge tests were performed

32
after compacting each lift of backfill. During the backfilling, the wall was held in place
by connecting the actuators to the abutment wall and then locking the actuators. Figure 3­
26 shows the overall view of the test specimen after placing structure backfill material.
The height of the backfilling in Test 3 was 7.5 feet.

Phase I, Test 4

In Test 4, like Test 3, the excavated area prior to placement of structure backfill
was extended to the larger area. Figures 3-27 and 3-28 show the test setup after
excavation. The setup of actuators in Test 4 allowed the upward movement of the
abutment wall to simulate the backwall being sheared off from the stemwall in seat-type
abutments. Figure 3-29 shows the overall test setup of abutment Test 4. The construction
process was similar to that in Test 3. The only difference between Test 3 and 4 was the
height of the backfill, changing from 7.5 feet to 5.5 feet. Figure 3-30 shows the overall
view of the test specimen after backfilling.

3.2.2.2 Material Testing

The concrete properties used in the test unit-Phase I was established from testing
at UCSD’s Charles Lee Powell Laboratory. The compression strength of concrete was
measured at 7 days, 28 days, and on the Day Of the Testing (D.O.T). Results are
presented in Table 3-8. Each value in the table represents an average strength of three
unconfined concrete cylinders (6.0 inch x 12.0 inch cast during the concrete pour).

3.2.2.3 Instrumentation

External

External instrumentation consisting of soil pressure transducers, linear

potentiometers, and tiltometers were attached to the abutment wall. The lateral
displacement of the top and bottom of the abutment wall was recorded by set of four
linear potentiometers, also allowing determination of the rotation of the wall. Additional
linear potentiometers were placed in Test 2, Test 3, and Test 4 to record the vertical and

33
horizontal displacement of the wall. Tiltometers were also mounted to each wingwall to
record any rotation in the abutment wall during the tests. Twelve soil pressure transducers
were placed on the abutment backwall to monitor soil pressures developed on the wall.
These transducers had a sensitive face with a 7.87 inch (200 mm) diameter and were
glued to the backwall.

In order to capture the location of the passive failure plane in the backfill
developed during the tests, twelve to sixteen columns of Styrofoam (in two rows) with 3­
foot spacing were placed in the structure backfill area extended to embankment backfill
for each test prior to the testing. The external instrumentation is illustrated in Figures 3­
31 and 3-32.

Internal

The abutment wall was internally instrumented with electrical resistance strain
gauges. The strain gauges were mounted on the reinforcing steel of the abutment wall to
track strains in the backwall and wingwalls. The strain gauge with gauge length = 0.2
inch (5 mm), gauge resistance = 120 Ω, and gauge factor = 2.13 was used. The primary
locations of strain gauges are shown in Figure 3-26.

3.2.2.4 Loading Protocol

Figure 3-12 shows the test setup of the Test 1 with a total of six actuators. Four
220-kip actuators were inclined at ±20º with respect to horizontal. Two 500-kip actuators
were inclined by 9.0º respect to horizontal. The setup restricted effectively the vertical
movement of the wall. The loading protocol in Test 1 consisted of displacement­
controlled cycles where amplitude increased as the test progressed, as shown in Figure 3­
33. Three cycles were performed at each displacement level. In Test 2, 3, and 4 the load
was applied monotonically, using a displacement increment of 0.001 inch per second.

34
3.2.3 TEST SETUP-PHASE II

Although the setup for Tests 2, 3, and 4 in Phase I replicated the seat-type
abutment behavior, the backwall of the test specimen was built monolithically with its
wingwalls as a diaphragm abutment. Therefore, the second phase of this research
program was conducted on a seat-type abutment which had a backwall separated from the
seat and wingwalls. The abutment test unit consisting of the seat (stemwall), shear keys,
wingwalls, and backwall was built at a large-scale of a prototype abutment. The desired
failure mode was geotechnical not structural, like the Phase I testing, so the abutment
backwall was designed and built to remain in the elastic range during the test. The details
of this System Test are shown in Table 3-9.

The test setup was designed to study the behavior of seat-type abutments in
longitudinal direction, restrained from translational and rotational movements (Figure 3­
47). The supporting block A (Appendix B) from the previous tests was kept in place to
use as part of the test base in Phase II. The test base, consisting of two concrete blocks,
was post-tensioned to the reaction wall to prevent any movement of the test base,
wingwalls, and stemwall (seat). In detailing the supporting block B and the seat
(stemwall); instead of using minimum reinforcement, steel fiber (Figure 3-43) concrete
was used. In Phase II, the wingwalls were detached from the backwall and just resisting
the soil pressure caused by the backwall movement during the test. Therefore, minimum
amount of steel reinforcement was used in wingwalls. Supporting block B, seat
(stemwall), wingwalls and shear keys were cast in a single pour using steel fiber concrete
with a target compressive strength of 3000 psi at the age of 28 days. The backwall was
cast against the hardened and smooth concrete surface of the seat (stemwall). A bond
breaking film was applied to the backwall-stemwall interface. One inch thick Styrofoam
was placed at the interface of backwall-wingwalls to reduce the friction at the interface
(Figure 3-45). There was no structural connection between the abutment backwall and
seat (stemwall).

35
 To model the plane-strain condition in the abutment test, the frictional resistance
between the soil and wingwalls was minimized by using four layers of plastic sheets at
the interface, as suggested by Fang et al. (2004). In this experiment, two types of
polyethylene sheeting were used, “thick” and “thin”. The “thick” sheet was 0.006 inch
thick, and it is commonly used for construction (e.g. concrete curing). The “thin” sheet
was 0.001 inch thick, and it is widely used for protection during painting. The inner
surface of the wingwalls was covered with a thick plastic sheet (see Figure 3-35). No
slippage was allowed at this interface. Two layers of thin plastic sheeting were used as
middle layers. Thick sheeting was placed next to the soil particles, as an outer layer, in an
attempt to smooth out the rough surface that might result from particle penetration during
the backfilling.

Prior to the construction, the Phase I abutment test unit and one of the concrete
block was removed from the site. The contractor was asked to excavate the portion of the
embankment to open sufficient space for building the test unit (Figure 3-36). Prior to the
construction of the abutment backwall, the steel reinforcement was instrumented with
strain gages. The geometry and general reinforcing steel details of the test unit are shown
in Figures 3-37 to 3-42. The wingwalls were built against the cut slope of the excavated
area (Figure 3-44). The construction of the abutment test unit was completed by June
2006.

3.2.3.1 Test Setup

The test setup in the System Test allowed the upward movement of the abutment
backwall. Figure 3-47 shows the overall test setup for the System Test before placing the
structure backfill. The lateral load was applied to the abutment backwall by two 220-kip
actuators and two 500-kip actuators. Prior to placement of the structure backfill material,
the existing cut slope resulting from the excavation was roughened by the contractor
using a backhoe. The construction process was similar to the construction procedure
explained in Phase I. During the backfilling, the backwall was fixed in place by
connecting the actuators to the wall. Figure 3-48 shows the overall view of the test

36
specimen after placing the structure backfill. The height of the backfill in System Test
was 5 feet. Figures 3-49 and 3-50 show the abutment system Test during the backfilling
and after backfilling, respectively.

3.2.3.2 Material Testing

The concrete properties used in the test unit-Phase II was established from testing
at UCSD’s Charles Lee Powell Laboratory. The compression strength of concrete was
measured at 7 days, and on the Day Of the Testing (D.O.T). Three unconfined concrete
cylinders (6.0 inch x 12.0 inch cast during the concrete pour) were tested. The values in
Table 3-10 represent the average strength of three samples.

3.2.3.3 Instrumentation

External

External instrumentation consisting of soil pressure transducers, and linear

potentiometers were attached to the abutment backwall. The lateral displacement of the
top and bottom of the abutment backwall was recorded by set of four linear
potentiometers, also allowing determination of the rotation of the wall. Additional linear
potentiometers were placed to record the vertical and horizontal displacement of the
backwall. Eleven soil pressure transducers were glued into the abutment backwall to
monitor soil pressures developed on the wall. These transducers had a sensitive face with
a 7.87 inch (200 mm) diameter and a 290 psi capacity. The measurements from the
pressure transducers are reported in Appendix F. The external instrumentation is
illustrated in Figure 3-51. In addition, fourteen columns of Styrofoam were placed in the
backfill area to capture the passive failure wedge (Figure 3-50).

Internal

The abutment backwall was instrumented internally with electrical resistance

strain gauges. The strain gauges were mounted on the reinforcing steel of the abutment
backwall to track strains in the wall. The strain gauge with gauge length = 0.2 inch (5

37
mm), gauge resistance = 120 Ω, and gauge factor = 2.13 was used. The locations of strain
gauges are shown in Figure 3-52.

3.2.3.4 Loading Protocol

Figure 3-48 shows the test setup of the System Test where two 220-kip actuators and
two 500-kip actuators were connected to the backwall horizontally. The load test was
performed monotonically, using a displacement increment of 0.001 inch per second.

38
Table 3-1. Gradation requirements for structure backfill
Sieve Sizes Percentage Passing
75-mm 100
4.75-mm 35-100
600-μm 20-100

39
 Table 3-2. Representative soil types for abutment structure backfill from field investigation (after EMI report 2005)

Grain Size Sand Expansion Shear strength of compacted

Number of Plasticity Compaction
Number of Depth Sand grain Field dry unit Distribution Equivalent Index soil (95% R.C.)
Soil Type Samples Plasticity NSPT w (%)
Bridges (ft) size weight (pcf) γd,max wopt
Tested gravel sand fines PI SE EI pd (pcf) ϕ (deg) c (ksf)
(pcf) (%)
SP-Dense to very dense <5 37-43 125-130 35-43 0.13-0.3
2 2 2.5-7.5 coarse none-low 52-59 3-6 120 8-25 70-79 5-13 0-7 NM NM
sand with gravel (0) (40) (125) (38) (0)
SM-Medium dense silty 0-9.5 5-10 21-52 116-127 24-35 0.25-1.5
9 10 fine-coarse none-low 9-15 5-14 105-120 1-28 60-87 12-35 0-2 109-127 8-12
sand, some with gravel (0-5) (8) (34) (120) (33) (0.5)

40
SC-Medium dense clayey medium­ 6-14 116-125 22.5-24 1.3-6
4 4 5-10 low 20-44 7-15 113-114 0-8 50-60 35-50 11 0 NM NM
sand, some with gravel coarse (10) (120) (23) (2)
CL-Stiff-hard clay, some
medium­ 12-41 0-3 106-122 1-8
with fine to coarse-grained 5 10 5-12.5 fine-coarse 18-35 14-29 90-116 0-3 25-35 62-75 20-77 114-122 17-22 0-17 (6)
high (22) (0) (110) (3.5)
sand and/or silt

1. Values shown in () are typical values observed from the test results by EMI
2. NM = Not measured
 Table 3-3. Proposed soil type for abutment structure backfill
Grain Size Distribution
Soil Type (Percentage Passing, %) SE PI
75 mm 4.75 mm Fines
Sands 100 >75 5-12 40+ <5
Silty/Clayey Sands 100 >80 20-40 20-30 5-15
Lean Clay 100 100 60-80 <10 >15

Table 3-4. Summary of index test results on sample of structure backfill materials
Grain Size Distribution Dry Unit d
PIe
Soil Type USCSa (Percentage Passing, %) Weightc SE
(%)
75 mm 4.75 mm 74 μmb (pcf)
Clayey Sand SC 100 93-100 35-40 126.0 16-22 10-13

Silty Sand SM 100 82-85 25-30 127.0 20-22 <4

Notes
a
ASTM D2487
b
ASTM D1140
c
ASTM D1557
d
ASTM D2419
e
ASTM D4318

41
 Table 3-5. Summary of CU test results on samples of abutment structure backfills

Cell Strain
p' max q max
Soil type Sample B Pressure Rate
(psi) (psi)
(psi) (% / min)
SM-1 96 3.5 0.3 64.9 40.6
Silty Sand SM-2 96 6.9 0.3 73.8 43.1
SM-3 96 13.9 0.3 117.1 70
SC-1 96 3.5 0.3 39.0 22.0
Clayey Sand SC-2 96 6.9 0.3 57.1 32.4
SC-3 96 13.9 0.3 71.2 37.7

Table 3-6. Structure backfill strength parameters based on CU triaxial tests

Silty Sand (SM) Clayey Sand (SC)

φ' (deg) c' (psf) φ' (deg) c' (psf)
34.5 600 28 1000

42
 Table 3-7. Bridge abutment research program test matrix, Phase I

Phase I
Variables
Test 1 Test 2 Test 3 Test 4
Soil Type clayey sand silty sand silty sand silty sand
Structure Backfill Height 5.5 ft 5.5 ft 7.5 ft 5.5 ft
Structure Backfill Area small small large large
Vertical Movement of restrained allowed allowed allowed
Wall

Table 3-8. Concrete compressive strengths in Phase I

D.O.T* D.O.T D.O.T D.O.T

7 Days 28 Days
Test 1 Test 2 Test 3 Test 4
psi psi
psi psi psi psi
Abutment Wall 3,153 4,752 5.97 4,770 4,937 5,550

*
The concrete cylinders in Test 1 were from Truck1, and in Test 2, 3, and 4 were from Truck 2

43
 Table 3-9. Bridge abutment research program test matrix, Phase II
Phase II
Variables
System Test
Soil Type silty sand
Backfill Height 5.0 ft
Backfill Area large
Vertical Movement of the Wall allowed

Table 3-10. Concrete compressive strengths in Phase II

7days D.O.T* (17 days)

psi psi
Abutment wingwalls and seat
4807 5213
(steel fiber concrete)
Abutment backwall 3840 4510
*
D.O.T: Day of the Testing

44
 (a) Diaphragm abutment
Backwall

Superstructure
Structure
Backfill Material

Wingwall

Embankment Material
Pile cap

Piles
Ground level

(b) Seat-type abutment

Superstructure Backwall Wingwall

Structure
Backfill Material

Seat Embankment Material

Pile cap
Ground level
Piles

Figure 3-1. Backfill area placed behind diaphragm and seat-type abutment, side view

45
 HUMBOLDT

ALAMEDA
SAN MATEO
SANTA CLARA

FRESNO

KERN

SANTA SAN BERNARDINO

BARBARA
LOS
ANGELES
RIVERSIDE

ORANGE
IMPERIAL

SAN DIEGO

Figure 3-2. Map of the site locations for field investigation on abutment structure backfill
material (after EMI report 2005)

46
 Reaction Wall Control Room Shake Table
Facilities

70' 20'

70'

Soil Pit

Figure 3-3. Soil-structure-interaction test facilities, Englekirk Structural Engineering Center

Embankment

Abutment Wall

Reaction Wall

Figure 3-4. Overall view of abutment wall during the construction

47
Figure 3-5. Map of site location of material imported from Gilman Dr., La Jolla (Photo
from Google Earth)

48
Figure 3-6. Map of site location of material imported from Naval Training Center, San
Diego (Photo from Google Earth)

49
Figure 3-7. Excavated area before starting the construction of the abutment wall

Figure 3-8. Test base during the construction

50
Figure 3-9. Overall view of the abutment test unit, Phase I

51
Figure 3-10. Reinforcing details of the abutment test unit, Phase I

52
Figure 3-11. Overall view of installing the formwork of the abutment wall, Phase I

53
Figure 3-12. Overall view of the abutment wall during the construction, Phase I

Figure 3-13. Overall view of the abutment wall, ready for concrete pour, Phase I

54
Figure 3-14. Overall view of the abutment after construction, Phase I

15.5' 4'

11.5'
2'
7.5'

Figure 3-15. Excavated area before backfilling, abutment Phase I-Test 1

55
 3' 7'
18"

2' 15.5'

7.5'

Figure 3-16. Overall test setup for abutment Phase I-Test 1

Figure 3-17. Top view of the abutment during the backfilling, Phase I-Test 1

56
Figure 3-18. Overall view of the test unit after backfilling, Phase I-Test 1

Figure 3-19. Excavated area before backfilling Phase I-Test 2

57
 4'
14.5'
15.5'
2'
7.5'

Figure 3-20. Overall test setup before backfilling, abutment Phase I-Test 2

7' 18"

15.5'
6'

2'
7.5'

Figure 3-21. Overall test setup for abutment Phase I-Test 2

58
Figure 3-22. Overall view of the test specimen after backfilling, Phase I-Test 2

Figure 3-23. Excavated area prior of backfilling Phase I-Test 3

59
 Side and back slopes are 1:1

3'
15.5'

10'
3'
31'

Figure 3-24. Excavated area in abutment Phase I-Test 3

18"
15.5'
10"
8.5'
7.5'

22.5' 21.5'

Figure 3-25. Overall test setup of abutment Phase I-Test 3

60
Figure 3-26. Overall view of the test specimen after backfilling, Phase I-Test 3

Figure 3-27. Excavated area prior of backfilling Phase I-Test 4

61
 Side and back slopes are 1:1

15.5' 3'

29' 8'
3'

Figure 3-28. Overall view of the excavated area in abutment Phase I-Test 4

18"
15.5'
10"

2' 8.5' 7.5'

21.5' 20.5'

Figure 3-29. Overall test setup for abutment Phase I-Test 4

62
Figure 3-30. Overall view of the test specimen after backfilling, Phase I-Test 4

63
 Pressure transducers

Tiltometer
Linear Potentiometers

Figure 3-31. Layout of the instrumentation in the abutment specimen, Phase I

16 columns of
Styrofoam placed
in two rows

Figure 3-32. Overall view of the instrumentation on the test unit, Phase I

64
Figure 3-33. Locations of the strain gauges of the backwall and wingwalls, Phase I

6
5
Displacement (inch)

4
3
2
1
0
-1
-2
0 1000 2000 3000 4000 5000 6000 7000
Time (sec)

Figure 3-34. Loading protocol of Phase I, Test 1

65
 1" thick Wingwall
Thick plastic sheet
Styrofoam
Two layers of thin plastic sheet
Thick plastic sheet
Backwall
Backfill

Figure 3-35. Plastic layers arrangement at wingwalls interface in Phase II, System Test

Figure 3-36. Excavated area for building the System Test (seat-type abutment)

66
 1' 18' 1'

11.5'

Wing Wall Wing Wall

Back Wall
3'
1.5'

3.5' Concrete Reaction Slab

Figure 3-37. Plan view of the seat-type abutment test unit, Phase II

67
 3' 14' 3' Shear Key
2' A 2'

Back Wall
ELEVATION

9'

Stem Wall

1.5' Concrete Reaction Slab

Backwall
3.5' 1.5' 1.5' 11.5'

Shear Key
4'
5' Wing Wall
SECTION A-A

4'
Stem Wall

1.5' Concrete Reaction Slab

6.5'

Figure 3-38. Elevation view of the seat-type abutment test unit, Phase II

#4
Wing wall 18' Wing wall

3'
1.5'

3.5'

Reaction slab

Figure 3-39. Details of reinforcing steel for wingwalls, plan view, Phase II

68
 A

3' 14' 3'

2' 2'

Shear Key
#4

#4
5'

#5 @9"
4' Stem Wall

1.5' 1'

20'

A 7- Steel Pipe Sleeve (2.5" ID)@ 2' Concrete Reaction Slab

ELEVATION

3.5' 3'
1.5'

#4

#5 @9"

1'
Concrete Reaction Slab

6.5'

SECTION A-A

Figure 3-40. Details of reinforcing steel for stemwall and shear keys, Phase II

69
 WING WALL 3'
#4

2" clearance 9- #4 @18" (for each face) #5 @9"

3'
2'
1'

#5

14.5'

PLAN

9- #4 @18" (for each face) 2" clearance

2" clearance 1'
A
#4
#5 @9"

4'
#5 @9"
9'

#4 (for each face)

2- #4

12- #5 @9" (for each face)

2" clearance
A
ELEVATION SECTION A-A
2" clearance
#4
ELEVATION

Figure 3-41. Details of reinforcing steel for wingwalls, Phase II

BACK WALL A
2" clearance
18' 1.5'

#8 @ 6"

5' #4 stirrup @ 6"

#9

10- #8 @ 6" 36- #9 @ 6" 2" clearance 2" clearance

A
ELEVATION SECTION A-A

Figure 3-42. Details of reinforcing steel for backwall, Phase II

70
Figure 3-43. Steel fiber used in concrete mix of stemwall and wingwalls, Phase II

Figure 3-44. Overall view of the System Test during the construction, Phase II

71
Figure 3-45. One inch thick Styrofoam placed at the interface of backwall-wingwalls,
Phase II

Figure 3-46. Overall view of the System Test after construction, Phase II

72
 3'
4'
17.5'
24'
3'
9'

Figure 3-47. Overall view of the excavated area in Phase II, System Test

3' 2'
1'

24' 20'

14.5'
4.5' 5'

4'

Figure 3-48. Overall test setup of the Phase II, System Test

73
 Figure 3-49. Backfilling of the abutment Phase II, System Test

14 columns of
Styrofoam placed
in two rows

Figure 3-50. Overall view of the test unit after backfilling, Phase II

74
 Pressure transducers

Linear Potentiometer

Figure 3-51. Layout of the instrumentation in the abutment specimen, Phase II

75
Figure 3-52. Locations of the strain gauges of the backwall, Phase II

76
4 EXPERIMENTAL RESULTS

This chapter describes test observations recorded during the testing of the bridge
abutment test units and summarizes the experimental results. The test observations
include crack growth, maximum recorded capacity of the wall, the failure surface, and the
measured failure angle. The force-displacement response measured during the tests was
depicted in this chapter. The effect of key variables in this experimental program
including the backfill height, the vertical movement of the wall, and the area of structure
backfill were discussed.

4.1 OBSERVATIONS DURING THE TESTS

TEST 1, PHASE I:

The loading protocol in Test 1 consisted of displacement-controlled cycles where

amplitude increased as the test progressed (see Figure 3-34). During the test, the first
crack at the surface of the backfill occurred at the displacement of 1 inch, at
approximately 11.0 feet from the face of backwall (Figure 4-1(b)). A small bulge was
observed at some parts of the crack. Various ground cracks occurred around the end of
the wingwalls and also close to the back side of the backwall. As the test continued, the
first crack extended along the surface laterally, parallel to the backwall and became a
major crack (Figures 4-1(c) and (d)). The cracks at end of the wingwalls also grew and
joined to the major crack. The maximum horizontal load of 788 kips was reached at 3.6
inch displacement. Figure 4-1 shows the surface cracks at different displacement levels.
After the test, the backfill material was excavated to observe the failure plane and
measure its angle. Figure 4-6 shows the broken columns of foam along the failure plane
during the test. The failure plane was approximately linear and coincident with the
previously excavated slope during construction as expected. The measured failure angle
was 26.8º where the structure backfill slope had an angle of approximately 26º to 27º.

77
TEST 2, PHASE I:

The load test was performed monotonically, using a loading rate increment of
0.001 inch per second. During the test, the first crack at the surface of the backfill
occurred at the displacement of 0.4 inch, around the front corner of the wingwalls (see
Figure 4-2(b)). As the test continued, the crack extended along the surface laterally,
parallel to the backwall (Figures 4-2(c) and (d)). The observed 3D effect was very small
because the wingwalls confined both sides. The maximum horizontal load of 334.7 kips
was reached at 1.4 inch horizontal and 0.5 inch vertical displacement of the backwall.
Figure 4-2 shows the test at the different displacement levels. The wall and the soil
wedge moved upward together during the test. As shown in Figure 4-2(b), a distinct
bulge developed on the surface of the structure backfill, 13.5 ft from the face of the
backwall and parallel to the backwall. After the test, the failure wedge was excavated to
observe the failure plane. Figure 4-7 shows the broken columns of foam along the failure
plane after the test. Similar to Test 1, the failure plane was approximately linear, along
the structure backfill slope. The measured failure angle was 21º, the same as the angle of
structure backfill slope.

TEST 3, PHASE I:

The load test was performed monotonically, with the same rate of the loading as
Test 2. The first crack at the surface of the structure backfill occurred at a displacement
of 0.6 inch, around the front corner of the wingwalls (see Figure 4-3(b)). As the test
continued, the crack extended on the surface laterally. In Test 3, the surface crack was not
parallel to the backwall, with distance varied from one side of the backwall to the other
side making an oblique shape (Figure 4-3(d)). No significant 3D effect was observed
during the test because of wingwalls confinement. The maximum horizontal load of
576.5 kips was reached at 1.7 inch horizontal and 0.5 inch vertical displacement of the
backwall. Figure 4-3 shows the test at the different displacement levels. The wall moved
vertically with the soil wedge during the test. As shown in Figure 4-3(d), a distinct bulge
developed on the surface of the structure backfill, 10 ft to 15 ft from the face of the

78
backwall. After the test, the failure wedge was excavated to observe the failure plane.
Figure 4-8 shows the broken columns of foam along the failure plane. The failure plane
was approximately linear with angle of 21.8º. It should be noted that in Test 3 the area of
structure backfill was much larger than the structure backfill area in Test 1 and 2. The cut
slope prior to backfilling was 31 feet away from the face of backwall, far enough not to
affect the failure zone.

TEST 4, PHASE I:

Test 4 was similar to Test 2, except the excavated area for the structure backfill
was larger in Test 4. The load test with same rate as that in Test 2 was performed in Test
4. The first crack was observed at the displacement of 0.3 inch, at the front corner of the
wingwalls as shown in Figure 4-4(b). As the test continued, the crack extended and
became a major crack. In Test 4, the surface crack was parallel to the backwall as a
straight line (see Figure 4-4(d). The observed 3D effect was very small. The maximum
horizontal load of 381.8 kips was reached at 1.3 inch horizontal and 0.4 inch vertical
displacement of the backwall. Figure 4-4 shows the test at the different displacement
levels. Similar to Test 2 and 3, the wall moved vertically with the soil wedge during the
test. As shown in Figure 4-4(d), a distinct bulge developed on the surface of the structure
backfill, 9.0 ft from the face of the backwall and parallel to the backwall. After the test,
the backfill material was excavated to observe the failure plane and measure the angle of
failure plane. Figure 4-9 shows the broken columns of foam along the failure plane
during the test. The failure plane was approximately linear with an angle of 24º from
horizontal. The cut slope prior to backfilling was far enough (29 feet away from the face
of the backwall) not to introduce the weak plane of failure to the system.

SYSTEM TEST, PHASE II:

In Phase II, the backwall was separated from seat and wingwalls to simulate the
condition in which the backwall being sheared off. The test setup and load test in Phase II
was similar to those in Test 2, 3, and 4 in Phase I. During the test, the first crack was
observed at the displacement of 1.2 inch, 7 ft from the backwall, close to the wingwalls

79
(Figure 4-5(b)). As the test continued, the crack extended along the surface. In the
System Test, the surface crack was essentially parallel to the backwall. There was no 3D
effect since the cracks occurred in the area confined by wingwalls. The maximum
horizontal load of 415.7 kips was reached at 2.2 inch horizontal and almost zero (0.01
inch) vertical displacement of the backwall. It should be noted that in Phase II, the
vertical movement of the backwall at the maximum force capacity was very small, unlike
the observed vertical movement of the backwall in Test 2, 3, and 4 in Phase I. It is
believed that the existence of friction at the bottom and side of the backwall prevented
the vertical movement of the wall at the early stages of the test. Figure 4-5 shows the test
at the different displacement levels. The vertical movement of the wall along with the
passive soil wedge was observed after the passive pressure reached to its maximum. As
shown in Figure 4-5(d), a distinct bulge developed on the surface of the structure backfill,
9.0 ft from the face of the backwall and parallel to the front of the backwall. After the
test, the failure wedge was excavated to observe the failure plane. Figure 4-10 shows the
broken columns of foam along the failure plane during the test. The failure plane was
approximately linear with angle of 23.3º.

4.2 FORCE-DISPLACEMENT CURVE

TEST 1, PHASE I:

The measured horizontal force-displacement response of the abutment wall is

shown in Figure 4-11. The load cells mounted on the actuators measured the force during
the test. By knowing the orientation of each actuator, the horizontal force was computed.
The abutment wall in Test 1 was just allowed to move in horizontal direction, so the
developed failure mechanism was Mechanism 1 as described in Chapter 2 (see Figure 2­
7). The abutment force-displacement behavior was nonlinear up to the peak point and
after the peak; it became approximately a horizontal line. The test was stopped after four
inch displacement due to reaching the maximum capacity of two actuators. Figure 4-11
shows that the abutment in Test 1 was degrading with each cycle by comparing the loads

80
at cycles with equal displacement peak. The permanent displacements at the end of each
half loading cycle show the plastic behavior of the structure backfill soil. Therefore, the
behavior of the abutment wall is dependent upon the prior loading of the abutment system
and the stress-strain history of the structure backfill material. Figure 4-12 is a plot of the
abutment horizontal secant stiffness as a function of displacement. The secant horizontal
stiffness plotted in Figure 4-12 is defined as the peak horizontal load of each half cycle
divided by the corresponding displacement. Figure 4-12 demonstrates that the secant
stiffness is strongly dependent upon displacements and softened as the displacement was
increased.

TEST 2, PHASE I:

Figure 4-13 shows the measured horizontal force-displacement response of the

abutment wall Test 2. Recorded horizontal and vertical displacements of the wall during
the test were used to measure the orientation of the actuators at each displacement level.
The orientation angle of actuators was needed to measure the horizontal force. The load
test was under displacement control and performed monotonically. In Test 2, the
abutment wall was free to move in the vertical direction. The developed failure
mechanism in this test was failure Mechanism 2 (see Chapter 2, Figure 2-7). The
abutment force-displacement behavior was nonlinear during the test. After reaching the
peak point, the load started degrading. The inflection point occurred at approximately two
times of the displacement at the maximum capacity. Figure 4-14 shows the measured
horizontal force vs. vertical displacement. The vertical displacement was relatively small
and increased as the horizontal displacement was increased. The test was stopped at five
inch displacement after developing the distinct failure crack at the top surface of the
structure backfill. Figure 4-15 is a plot of the abutment horizontal secant stiffness as a
function of displacement. Figure 4-15 demonstrates that the secant stiffness is strongly
dependent upon displacements.

81
TEST 3, PHASE I:

The measured horizontal force-displacement response of the abutment wall Test 3

is shown in Figure 4-16. The loading was under displacement control and performed
monotonically. Similar to Test 2, the abutment wall was allowed to move in the vertical
direction. The developed failure mechanism in this test was also failure Mechanism 2 as
described in Chapter 2 (see Figure 2-7). The abutment force-displacement behavior was
nonlinear up to the peak point and after the peak, it started degrading. The inflection
point of the graph occurred at approximately two times of the displacement at the
maximum capacity. Figure 4-17 shows the measured horizontal force vs. vertical
displacement. The vertical displacement was relatively small and increased as the test
was progressed. The test was stopped at five inch displacement after the distinct failure
cracks observed at the top surface of the structure backfill. Figure 4-18 is a plot of the
abutment horizontal secant stiffness as a function of displacement. Figure 4-18
demonstrates that the secant stiffness is strongly dependent upon displacements and
softened nonlinearly as the displacement was increases.

TEST 4, PHASE I:

The measured horizontal force-displacement response of the abutment wall Test 4

is shown in Figure 4-19. The load test was similar to Test 2, and 3, and under
displacement control and performed monotonically. The abutment wall was allowed to
move in the vertical direction during the test. The developed failure mechanism in this
test was failure Mechanism 2 (see Figure 2-7). The abutment force-displacement
behavior was nonlinear up to the peak point. After the peak, the softening behavior was
observed and the inflection point occurred at approximately two times of the
displacement at the maximum capacity. Figure 4-20 shows the measured horizontal force
vs. vertical displacement. The vertical displacement was increased as the horizontal
displacement was increased. The test was stopped at seven inch displacement after the
distinct bulges were formed at the top surface of the structure backfill. Figure 4-21 is a
plot of the abutment horizontal secant stiffness as a function of displacement. Figure 4-21
demonstrates that the secant stiffness is strongly dependent upon displacements.

82
SYSTEM TEST, PHASE II:

The measured horizontal force-displacement response of the abutment backwall in

the System Test is shown in Figure 4-22. The load test was under displacement control
and performed monotonically. In System Test, the abutment wall was free to move in the
vertical direction. The developed failure mechanism in this test was also failure
Mechanism 2 (see Figure 2-7). The abutment force-displacement behavior was nonlinear
up to the peak point and after the peak, it started degrading. The inflection point occurred
at approximately two times of the displacement at the maximum capacity. Figure 4-23
shows the measured horizontal force vs. vertical displacement. The vertical displacement
was very small at early stage of the test and it started increasing after reaching the
maximum force capacity. The test was stopped at six inch displacement after developing
the distinct bulges at the top surface of the structure backfill. Figure 4-24 is a plot of the
abutment horizontal secant stiffness as a function of displacement. Figure 4-24
demonstrates that the secant stiffness is strongly dependent upon displacements and
degraded nonlinearly as the displacement is increased. For comparison purposes, the
force-displacement responses of all tests are shown in Figure 4-25. Except Test 1 which
the cyclic loading was applied, the rest of the tests were loaded monotonically.

4.3 DISCUSSION OF RESULTS

4.3.1 EFFECT OF VERTICAL MOVEMENT OF THE WALL

As discussed in Chapter 2, one of the factors that controls the magnitude of the
passive earth pressure is the direction of the wall movement. If the wall movement
restrained vertically (vertical restraining force in the wall is greater than the vertical
component of the passive pressure force), slip will occur on the interface between the
structure and soil, and the value of the interface friction, δ, will be controlled by the
properties of the soil-structure interface (Figure 4-26(a)).Test 1 was performed on the
abutment wall which was restrained to move in the vertical direction (diaphragm type
abutment). In this test, the failure Mechanism 1 was developed and the value of the soil­

83
wall interface friction angle, δ, was controlled by the properties of the interface (Figure 4­
26(a)). The rest of the tests in Phase I and II were performed on the abutment wall which
was free to move vertically (the backwall being sheared off from the stemwall in seat­
type abutment). Failure Mechanism 2 was developed in these tests where the smaller
relative displacement across the interface occurred resulting in only partial mobilization
of the interface friction (Figure 4-26(b)). In this condition, the value of δmob was
controlled by the requirements of vertical equilibrium. The calculated mobilized interface
friction angle, δmob, was small and less than 10º in Test 2, 3, 4, and the System Test. The
value of the δ has a considerable effect on the amount of maximum passive pressure
(Coulomb theory and Log Spiral theory). Therefore, in cases where the interface friction
is partially mobilized, the developed passive pressure is much less than the cases with
fully mobilized interface friction.

In cases where the failure Mechanism 2 was developed, the force-displacement

relationship showed significant degradation after reaching the peak force, where, in case
of having failure Mechanism 1 the force degradation was not substantial. It was found,
that when the wall is allowed to move just in horizontal direction, it engages more soil to
participate in passive resistance resulting larger passive pressure and less degradation in
capacity after reaching maximum passive force (see Figure 4-25).

4.3.2 EFFECT OF EXCAVATED AREA FOR STRUCTURE BACKFILL

Tests 2 and 4, Phase I, were performed with two different structure backfill areas
to evaluate the effect of structure backfill area on capacity of the abutment wall. All the
variables, except the structure backfill area were kept the same in Test 2 and 4. In Test 2,
excavated area was small, where in Test 4 the excavated zone was extended to a larger
area. Even though the contractor roughened the surface of the slope before backfilling, it
was believed the existing cut slope from excavation would affect the development of the
failure plane if the cut slope was within the failure zone. The stiffness of the abutment in
Test 2 at small displacements (smaller than 1.0 inch) was much lower than the stiffness in
Test 4 (see Figure 4-27). The maximum capacity of the abutment in Test 2 and 4 were not

84
the same and was higher in Test 4 as shown in Figure 4-28. Therefore, the cut slope from
the excavation in Test 2 introduced a weak plane of failure to the system and sliding
occurred at that slope. It should be noted that by coincidence the cut slope in Test 2 had
the same angle as the passive failure angle of the silty sand (which is a function of soil
properties). Therefore, the passive failure wedge in Test 4 was matched with the Test 2
failure wedge.

4.3.3 EFFECT OF BACKFILL HEIGHT

The effect of backfill height can be studied by comparing the results of Test 3 and
4. Tests 3 and 4 were performed with two different backfill heights to evaluate the effect
of backfill height on capacity of the abutment wall. All the other variables were same.
The secant stiffness of the abutment in Test 3 and 4 are shown in Figure 4-29. The
stiffness ratio is not constant and varying from 1.3 to 1.6 while the displacement is
changing from 0.5 inch to 3.0 inch. For instance, the ratio of the secant stiffness at
maximum capacity in Test 3 and Test 4 was calculated and equal to 1.5 while the backfill
height ratio is equal to 1.36.

The ratio of the maximum passive pressure in Test 3 to the maximum passive
pressure in Test 4 was calculated and compared with the ratio of the backfill height in
Test 3 and 4. Figure 4-30 shows the force-displacement response of Test 3 and 4. The
ratio of maximum passive pressure of corresponding tests is:

p max,Test 3 4.96
= = 1.11 4.1
p max,Test 4 4.48

85
and the ratio of the backfill height is equal to:

hTest 3 7.5
= = 1.36 4.2
hTest 4 5.5

The comparison between the Eq. (4.1) and Eq. (4.2) shows that the maximum
passive pressure ratio is not equal to backfill height ratio. This single comparison
suggests that using a linear proportionality constant similar to Eq. (2.9) to estimate the
passive pressure for various abutment heights may not be appropriate. This study
indicates that the response of bridge abutments in the longitudinal direction is non-linear
and a function of several factors. Therefore, further testing of seat-type abutments of
various dimensions is recommended to investigate the effect of backfill height and the
validity of this method.

86
 (a) Overall side view of the test setup

(b) Surface crack at 1.0 inch displacement (c) Surface crack at 2.0 inch displacement

(d) Observation at failure, side view (e) Observation at failure, plan view

Figure 4-1. Test observation at different displacement level in Phase I-Test 1

87
 (a) Overall side view of the test setup

(b) Surface crack at 1.0 inch displacement (c) Surface crack at 2.0 inch displacement

(d) Observation at failure, side view (e) Observation at failure, top view

Figure 4-2. Test observation at different displacement level in Phase I-Test 2

88
 (a) Overall side view of the test setup

(b) Surface crack at 1.2 inch displacement (c) Surface crack at 2.1 inch
displacement

(d) Observation at failure, top view (e) Observation at failure, side view

Figure 4-3. Test observation at different displacement level in Phase I-Test 3

89
 (a) Overall side view of the test setup

(b) Surface crack at 1.2 inch displacement (c) Surface crack at 2.1 inch
displacement

(d) Observation at failure, top view (e) Observation at failure, side view

Figure 4-4. Test observation at different displacement level in Phase I-Test 4

90
 (a) Overall side view of the test setup

(b) Surface crack at 1.2 inch displacement (c) Surface crack at 2.4 inch
displacement

(d) Observation at failure, top view (e) Observation at failure, top view

Figure 4-5. Test observation at different displacement level in Phase II-System Test

91
 Distance from backwall, inch
0 30 60 90 120 150
0
-10
-20

Depth, inch
-30
-40
26.8º -50
-60
-70

(a) Broken foam columns after (b) Profile of the broken foam columns
excavation

Figure 4-6. Observed failure plane, Phase I-Test 1

Distance from the backwall, inch

0 40 80 120 160
21º 0
-10
-20
Depth, inch

-30
-40
-50
-60
-70

(a) Broken foam columns after (b) Profile of the broken foam columns
excavation

Figure 4-7. Observed failure plane, Phase I-Test 2

92
 Distance from the backwall, inch
0 30 60 90 120 150
0

-20

Depth, inch
21 8º -40

-60
-80

-100

(a) Broken foam columns after (b) Profile of the broken foam columns
excavation

Figure 4-8. Observed failure plane, Phase I-Test 3

Distance from the backwall, inch
0 30 60 90 120 150
0
24º
-10
-20
Depth, inch

-30
-40
-50
-60
-70

(a) Broken foam columns after (b) Profile of the broken foam columns
excavation

Figure 4-9. Observed failure plane, Phase I-Test 4

Distance from the backwall, inch

0 30 60 90 120 150
0
-10
-20
Depth, inch

23.3º
-30
-40
-50
-60
-70
(a) Broken foam columns after (b) Profile of the broken foam columns
excavation
Figure 4-10. Observed failure plane, Phase II-System Test

93
Figure 4-11. Horizontal force-displacement, Abutment wall-Test 1

94
Figure 4-12. Horizontal stiffness-displacement, Abutment wall-Test 1

95
Figure 4-13. Horizontal force-displacement, Abutment wall-Test 2

96
Figure 4-14. Horizontal force-vertical displacement, Abutment wall-Test 2

97
Figure 4-15. Horizontal stiffness-displacement, Abutment wall-Test 2

98
Figure 4-16. Horizontal force-displacement, Abutment wall-Test 3

99
Figure 4-17. Horizontal force-vertical displacement, Abutment wall-Test 3

100
Figure 4-18. Horizontal stiffness-displacement, Abutment wall-Test 3

101
Figure 4-19. Horizontal force-displacement, Abutment wall-Test 4

102
Figure 4-20. Horizontal force-vertical displacement, Abutment wall-Test 4

103
Figure 4-21. Horizontal stiffness-displacement, Abutment wall-Test 4

104
Figure 4-22. Horizontal force-displacement, Abutment wall-System Test

105
Figure 4-23. Horizontal force-vertical displacement, Abutment wall-System Test

106
Figure 4-24. Horizontal stiffness-displacement, Abutment wall-System Test

107
Figure 4-25. Horizontal force-displacement response, Phase I and II

108
 Soi
Soil Soil
mov
move ement movement
Structure
Struct ure Structtur
Struc ure
e
mov
moveement
ment mov
moveement

Mechanism 1 Mechanism 2

(a) (b)

Figure 4-26. Passive failure mechanisms of a wall

109
Figure 4-27. Horizontal stiffness-displacement response of Test 2 and 4

110
Figure 4-28. Horizontal force-displacement response of Test 2 and 4

111
Figure 4-29. Horizontal stiffness-displacement response of Test 3 and 4

112
Figure 4-30. Horizontal force-displacement response of Test 3 and 4

113
5 ANALYSIS OF LONGITUDINALLY LOADED BRIDGE
ABUTMENTS

This chapter describes a procedure for analyzing the response of bridge abutments
to longitudinal loads. The procedure was validated by comparing the computed response
of the bridge abutments tested at UCSD Englekirk Structural Engineering Center to the
results of the tests discussed in Chapter 4. The maximum passive earth pressure was
calculated using the Method of Slices, which follows the Terzaghi’s suggestion and uses
a failure surface composed of a logarithmic spiral and a straight line as shown in Figure
5-1 (Shields et al. 1973).

5.1 ABUTMENT WALL MODEL

In order to model the behavior of a bridge abutment loaded longitudinally, it is

necessary to know the maximum soil resistance developed behind the wall. In addition, it
is required to have a proper model to predict the behavior of the bridge abutment during
the loading and after reaching the maximum capacity. This section describes the Method
of Slices in details, which was used to predict the maximum passive pressure. Also, it
explains the proposed model to predict the force-displacement relationship of
longitudinally loaded bridge abutments.

5.1.1 PASSIVE EARTH PRESSURE RESISTANCE

The Log Spiral theory was used to estimate the passive earth pressure developed
back of the abutment wall. Based on Terzaghi’s (1943) suggestion the failure surface
composed of a logarithmic spiral and a straight line. The shape of log spiral failure
surface is shown in Figure 5-1. In order to calculate the total passive earth pressure the
Method of Slices was used. The forces acting on a slice element are shown in Figure 5-2.
The equilibrium equations of a slice in horizontal and vertical direction are as follows:

114
 Summation of forces in the horizontal direction yields to:

E + dE − E − R sin (α + φ ) − S cos α = 0 5.1

By substituting S = cL in Eq. (5.1)

dE = R sin (α + φ ) + cL cos α 5.2

Summation of forces in the vertical direction yields to:

w + x + dx − x − R cos(α + φ ) + S sin α = 0 5.3

By substituting S = cL in Eq. (5.3) and solving for R:

w + dx + cLsin α
R= 5.4
cos(α + φ )

By substituting Eq. (5.4) into Eq. (5.2):

dE = (w + dx + cLsin α ) tan (α + φ ) + cL cos α 5.5

The total passive force created by ABCD in Figure 5-1 is:

∑ dE = ∑ [(w + dx + cL sin α ) tan(α + φ ) + cL cos α ]

BD
5.6

115
The direction of the failure surface at the wall is determined by the roughness angle, δ, of
the wall (Shields 1973). The angle between the horizontal and the failure surface at the
wall is αw. Shields et al. (1973) expressed the αw in terms of φ and δ as follows:

1⎡ ⎛ sin (φ − δ ) ⎞ ⎤
α w = ⎢arccos⎜⎜ cos(φ − δ ) − ⎟⎟ − φ − δ ⎥ 5.7
2⎣ ⎝ tan φ ⎠ ⎦

The logarithmic spiral must leave the wall at the angle of αw. In addition, the ray from the
center of the logarithmic spiral must intersect the failure surface at the corner of the wall
with an angle of 90-φ. Therefore, the direction of the ray OB is fixed and the center of
logarithmic spiral, O, can be defined precisely. Unlike other methods which search the
location of the failure surface by repeated trials, this method can establish the failure
surface by knowing the φ and δ. Appendix D describes the approach used to calculate the
center of the logarithmic spiral of the passive failure wedge.
The Rankine zone soil wedge, CDE, has a passive force of:

1 ⎛ φ⎞ φ
PR = γH R2 tan 2 ⎜ 45 D + ⎟ + 2cH R tan(45 D + ) 5.8
2 ⎝ 2⎠ 2

where γ is the unit weight of the soil. Therefore, the total passive force is:

PP = PR + ∑ dE 5.9

Substituting Eq. (5.6) in Eq. (5.9) gives:

PP = PR + ∑ [(w + dx + cL sin α ) tan (α + φ ) + cL cos α ] 5.10

BD

116
It is assumed that the shear forces across a slice can be considered to be zero ( dx → 0 for
slices other than the one closest to the wall) and maximum at the wall ( dx = PP tan δ
at α = α w ). For cohesive soil, the adhesion between the soil and wall should be added to
the shear force at the wall (Mokwa 1999) and becomes:

dx = PP tan δ + βcH 5.11

where β.c is the adhesion between the wall and soil. Typical values of β are shown in
Table 5-1. By substituting Eq. (5.11) into Eq. (5.10):

PP = PR + ∑ [(w + cL sin α ) tan (α + φ ) + cL cos α ] + PP tan δ tan (α w + φ ) + cβH tan (α w + φ )

BD

Therefore:

PP (1 − tan δ tan (α w + φ )) = PR + ∑ [(w + cL sin α ) tan (α + φ ) + cL cos α ] + cβH tan (α w + φ )

BD

The maximum passive force, Pp, can be calculated by:

PR + ∑ [(w + cL sin α ) tan(α + φ ) + cL cos α ] + cβ H tan (α w + φ )

PP = BD
5.12
1 − tan δ tan (α w + φ )

It should be noted that the failure zone was divided into vertical slices and the slices were
assumed to be trapezoid when calculating the area.

The passive pressure developed behind the abutment wall, corresponds to a long
wall moving against the soil (plane strain condition). For this case, there is no need to
consider 3D effects. In the experiments carried out as part of this research, however, the
length of the wall was not long enough to be considered as a 2D condition. However, the

117
confinement provided by the wingwalls and with the failure wedge within or close to the
confined zone made the three-dimensional effect very small. In the analysis of the test
results using the Ovesen-Brinch Hansen method (Duncan, 2001), an estimated 3-D shape
factor of 1.1 was used for Phase I. In this approach, the 3D effect was estimated for that
part of the failure wedge beyond the zone confined by the wingwalls. In Phase II, the
failure wedge was entirely within the confined area, and 3D effects did not need to be
considered.

5.1.2 ABUTMENT WALL FORCE-DISPLACEMENT RELATIONSHIP

There are several analytical methods for predicting the passive resistance­
displacement response in a longitudinally loaded wall such as p-y approach (see Chapter
2). There are a number of formulations available for developing p-y curves. These are
often empirically related to soil strength parameters measured in the laboratory or in the
field. Some of these methods use hyperbolic equations to represent the load-deflection
curves. Most of these methods consider asymptotic curves to model the relationship
between p and y after reaching the maximum passive pressure. After analyzing the
experimental curves (Section 4-2) and previously proposed relationships, it was found
that the force-displacement relationship of the abutment wall (free to move in vertical and
horizontal directions) can be approximated by rational function and exponential
functions. The force-displacement relationship of the abutment wall (restrained in vertical
movement) can be approximated as a rational function.

To model the part of the force-displacement curve, the following general

expression (Sargin et al. 1968) was used:

a + bx + cx 2
P= (5.13)
1 + dx + ex 2

118
 Δ
where a, b, c, d, and e are the parameters, x = is the non-dimensional displacement,
Δ max

and Δ max is the displacement at Pmax , the maximum passive force predicted by one of the
earth pressure theories. If the following boundary conditions:

dP
P = 0 at Δ = 0 and = k max at Δ = 0
dΔ

dP
P = Pmax at Δ = Δ max and =0 at Δ = Δ max
dΔ

are substituted, Eq. (5.13) becomes:

Ax + (D −1)x 2
P = Pmax (5.14)
1 + ( A − 2)x + Dx 2

k max
in which A = Δ max
Pmax

Pmax = the maximum passive force, calculated by one of previously mentioned

theories
k max = the initial elastic stiffness, see Appendix E
D = a parameter which affects the slope of the descending branch
By substituting Pult = RPmax in Eq. (5.14) and solving for D:

1
( A − RA + 2R ) −1 − R2
D= x x (5.15)
R −1

There is no conventional way to determine the residual shear strength of the soil,
but the Pres can be estimated based on experience. Based on the abutment test results and
published research in area of soil residual strength (Eid et al. 1999, Stark et al. 1994,
Fang et al. 2002, Salgado et al. 2000), the ratio of residual passive force to the maximum

119
passive force can be estimated ranging from 0.6 to 0.7 (R= 0.6 to 0.7). Based on field
observation, the ratio of the displacement at residual passive force to the displacement at
maximum passive force was varied from 2.5 to 3 ( x = 2.5 to 3).

In Test 2, 3, 4, and the System Test, the results showed that the inflection point
occurred at approximate displacement of 2Δmax (see Figures 4-13, 4-16, 4-19, and 4-22).
The softening behavior of the force-displacement relationship can be expressed using the
exponential softening rules in the plasticity model (Shing et al. 2001) as follows:

⎡ ⎛ ⎞⎤
( )
P = P2 y0 + RPmax − P2 y0 ⎢1 − exp⎜
−m
⎜ RPmax − P2 y
( x − xinf )⎟⎥
⎟⎥
(5.16)
⎢⎣ ⎝ 0 ⎠⎦
where P2yo is the passive force at 2Δmax and m is the slope at the inflection point:
′
dp ⎛ Ax + (D −1)x 2 ⎞
m= = ⎜ Pmax ⎟ (5.17)
dx ⎝⎜ 1 + ( A − 2)x + Dx 2 ⎟⎠

By substituting xinf=2 (inflection point at 2Δmax), Eq. (5.17) becomes as:

3A + 4D − 4
m(x = 2 ) = −Pmax (5.18)
(2A + 4D − 3) 2
Therefore, the rational function formulation in conjunction with the exponential
formulation is capable of predicting the nonlinear force-displacement relationship of the
abutment wall with the post-peak softening behavior. Note that the force-displacement
relationship of the abutment wall which is not expected to shear off during the earthquake
can be approximated using just the rational function formulation. In the following
section, the proposed force-displacement relationship of the abutment wall is calculated
and compared with test results.

120
5.2 COMPARISON OF MEASURED AND CALCULATED FORCE-
DISPLACEMENT RESULTS

Force-displacement curves for the abutment walls were estimated using the
proposed model described in previous section. Table 5-3 shows the input values of the
primary variables and dimensions used in the proposed model for each test.

To begin, it was needed to calculate the maximum passive force developed back
of the wall. In order to calculate the maximum passive force using Log Spiral theory, the
shear strength parameters of the soils were required to be determined. Triaxial tests were
performed on each soil type to measure the internal friction angle φ, and the cohesion, c.
Table 3-6 shows the shear strength soil parameters resulted from triaxial testing.

The maximum interface friction angle is commonly characterized in terms of the

ratio δmax/φ. Note that in Test 1, the maximum interface friction δmax, was assumed to be
2/3φ which is consistent with the commonly adopted for concrete-soil interface friction
angle (Potyondy 1964, U.S. Navy 1986). In Test 2, 3, 4, and the System Test, small
relative displacements across the interface resulted in partial mobilization of the interface
friction. Therefore, the values of δmob were controlled by vertical equilibrium in each test
(see in Figure 2-7). The values of δmob were determined by trial and error. A value of δmob
was assumed in the beginning of the process and then the corresponding value of passive
resistance force, P, was calculated. The vertical component of P was compared to the
weight of the structure (weight of the wall plus half of the weight of the actuators). The
process was repeated until the difference between the vertical component of P, and
weight of the structure was close to zero. Table 5-2 shows the computed δmob for each
test. The computed mobilized interface friction angle, δmob, was small and less than 10º in
Test 2, 3, 4, and the System Test, as expected. The Poisson’s ratio was estimated from
Eq. (E.1) in Appendix E and the initial tangent soil modulus, Ei, was obtained using the
stress-strain results from triaxial testing. The adhesion factor, β, was estimated using
Table 5-1. According to Table 5-1, the adhesion factor, β, was assumed to be equal to

121
0.85 for the clayey sand with cohesion of 1000 psf in Test 1. In Test 2, 3, 4, and System
Test, the interface adhesion was neglected due to the small relative displacements across
the interface.

In order to compare the absolute value of the passive earth force in Test 1 with
measured one, the sliding friction force at the interface of the abutment wall and the
concrete slab existing during testing, was back-calculated from the recorded data and
added to the computed passive force. The normal force at the interface of the abutment
wall and concrete slab (weight of the wall ± the vertical component of actuator force) was
back-calculated from the recorded data. It was assumed that the coefficient of friction at
the interface was approximately 0.4 (Bozorgzadeh et al. 2006) for a smooth concrete
interface. The back-calculated sliding friction force was equal to 86 kips in Test 1. The
wingwall-soil interface friction (side and bottom) was negligible and assumed to be equal
to zero in this test.

In Tests 2, 3, and 4, the abutment wall displacement was in horizontal and vertical
directions from the beginning of the test, so there was no sliding friction, except at the
bottom of wingwall-cut slope interface which was negligible. However, in the System
Test the recorded vertical displacement was close to zero until reaching the maximum
passive force, declaring that the sliding force was existed in the system at the maximum
passive force. When the wall was pulled back at the end of the test, the only resistance
force was the sliding friction force which was approximately 20 kips (see Figure 5-8). In
the System Test, since the wingwalls were fixed and not moving with the wedge of the
soil, side friction was introduced to the system. The sliding friction and the side friction
forces were added to the predicted capacity so that the predicted passive force could be
compared with that measured. As noted earlier, in order to reduce the amount of friction
between the wingwalls and soil in the System Test, four layers of plastic sheets were used
at the interface. Fang et al. (2004) conducted interface shear tests for measuring the
friction at the interface between soil and plastic sheets. The experimental results obtained
in the tests on different interface friction arrangements are shown in Table 5-4. Based on

122
Table 5-4, the interface friction angle of 12.0º was used for back-calculating the side
friction in the abutment System Test. It is believed that the passive pressure at the corner
of the wingwall-backwall is much higher due to the high confinement of the soil in that
region. However, in order to simplify the calculation, an evenly distributed passive
pressure along the wingwalls was assumed. An average side friction force may be
introduced by the following equation:
1
Fs = ( K p γh 2 l) tan δ s (5.19)
6
where Kp is the coefficient of passive earth pressure and is computed by dividing the
1
passive force by ( γh 2 b) which is equal to 12.9 in the System Test. Due to the triangular
2
shape of the soil wedge (see Appendix F), the location for equivalent reaction of the
pressure distribution was assumed at 2/3 of the soil wedge. The length of the soil wedge,
l, calculated from the Log Spiral theory and (tan δs) is the coefficient of friction. By
substituting the values in Eq. (5.19), Fs is equal to:
⎛1 ⎞
Fs = ⎜ *12.93*121* 5 2 *10.2 ⎟ tan12 = 28 kips
⎝6 ⎠
By adding the sliding friction of 20 kips, the total friction force in the system was 48 kips
and was extracted from recorded data points during the test.

It was found that values of ymax/H ranging from 2% to 4% of the wall height were
consistent with the commonly adopted displacement at the maximum mobilized passive
pressure (Clough and Duncan 1971, Mokwa and Duncan 2001, Cole 2006). For each test,
the maximum resistance, Pmax was calculated using the Method of Slices. The initial
backwall stiffness, kmax, was determined using the elasticity theory by Douglas and Davis
(1964), described in Appendix E. By knowing all variables, the rational function and
exponential formulations were used to develop the p-y curves for the backwall.

Table 5-4 shows comparisons of the measured maximum passive pressure of the
tests with calculated values using different earth pressure theories and methods discussed
in Chapter 2 (Rankine, Coulomb, Log Spiral, and Caltrans method). Since Rankine

123
theory does not take into account the interface friction and adhesion, it underestimated
the capacity in Test 1 significantly. In addition, this theory underestimated the abutment
capacity for rest of tests where there was very small interface friction and no interface
adhesion.

Estimating the capacity of bridge abutments using the Coulomb theory showed a
better agreement with test results than using the Rankine theory. It should be noted that in
computing passive pressure using Coulomb theory, the cohesion contribution along the
sliding surface and the mobilized interface friction angles between soil and the wall
(illustrated in Table 5-2) were considered. Coulombs’ method underestimated the passive
resistance in Test 1, 3, 4, and System Test and was 6-21% lower than measured values,
and overestimated by 5% in Test 2.

It can be seen that the Log Spiral theory has the most reasonable agreement with
the experimental results. The Log Spiral method estimated the Pmax within 7% for the
Test 1, 3, 4, and the System Test, but overestimated the passive resistance in Test 2 by
18%. As expected, the existence of the cut slope for placement of the wingwalls created a
weakened plane in Test 2 and resulted in overestimating the capacity by Log Spiral and
Coulomb Methods.

Table 5-4 shows that the Caltrans approach underestimated the abutment capacity
in Test 1 by 35% whereas overestimating in Test 2, 3, 4 and System Test by 10%-37%.
The underestimation in Test 1 is partially a result of restraining the vertical movement of
the abutment during the test, resulting in a high mobilized wall friction angle as
compared to the subsequent tests, were vertical movement of the wall resulted in low
mobilized wall friction.. Nevertheless, the Caltrans approach, which does not account for
wall friction nor soil type, tends to overestimate capacity more than any of the methods
that do account for wall friction, when the wall is allowed to move vertically relative to
the soil.

124
 Computed horizontal force-horizontal displacement plots are compared to
measured response curves. The hyperbolic curve given by Duncan and Mokwa (2001)
was constructed using Eq. (2.10). The parameters used in this approach were similar to
the proposed approach. However, the proposed methods provided the post-peak softening
behavior where the Duncan et al. (2001) model assumes a plateau behavior after the peak
force. The Caltrans force-displacement plot was constructed using Eqs. (2.8) and (2.9)
including the geometry of the wall in the calculation. The initial stiffness suggested by
this approach was adjusted proportional to the backwall height and illustrated in Table 5­
5.

Figures 5-4 through 5-8 compare the measured passive force-displacement curves
for each test with the predicted force-displacement behavior using the proposed method,
as well as Duncan et al. (2001) model, and Caltrans model. The post-peak behavior in
Test 2, 3, 4, and System Test showed substantial degradation in force capacity resulted
from the characteristics of the failure mechanism developed in these tests. It should be
noted that the post-peak softening behavior was observed slightly in Test 1 which was
consistent with the observed results from other large-scale tests with the same failure
mechanism (Maroney 1995, Rollins and Cole 2006). In the case of having the wall
restrained in vertical direction (Test 1), the force-displacement relationship should be
modeled just with the rational function. The more commonly used force-displacement
models such as Duncan et al. (2001) model do not take into account the descending
branch. However, the proposed model has been developed to account for the passive
force-displacement relationship with post-peak degradation behavior. The Caltrans
method significantly underestimated the capacity of the abutment wall in Test 1, but
overestimated in Test 2, 3, 4, and System Test. The initial stiffness of the wall computed
by the Caltrans approach (Figs 5-4 to 5-8) was significantly underestimated. The test
results noted that the Caltrans method did not provide a reasonable estimate for the
abutment capacity and stiffness.

125
 The predicted p-y curves using previously mentioned models overestimated the
maximum capacity in Test 2. However, there is a good agreement between the predicted
curves and measured one in Test 4. It should be mentioned that the only difference
between Test 2 and Test 4 was having the larger area of backfilling in Test 4. The
comparison between these two tests shows that it is important to have a large enough
structure backfill area to ensure the failure occurs within the structure backfill material
not at the embankment/structure backfill interface or within the embankment material.
The proposed model was the only model to provide a good prediction in post-peak
behavior for each test, especially for the tests (Test 2, 3, 4, and System Test) with
significant post-peak degradation in capacity. Overall, the proposed model provides the
most consistent estimate of force-displacement behavior of each abutment test.

Figs 5-9 to 5-13 compare the measured secant stiffness for each test with the
stiffness estimated using the proposed model and Duncan et al. (2001) model. The secant
stiffness was defined by the slope of the line from the origin to the current point of
interest on a force-displacement curve. Very good agreement was found between the
proposed model stiffness prediction and measured stiffness of the abutment wall tests. It
may be observed that the Caltrans approach underestimated the stiffness in each test
(shown in Table 5-5) at displacements of less than 2 inches significantly. Although the
predicted secant stiffness by Caltrans approach and Duncan et al. (2001) model shows a
good agreement with the measured secant stiffness at displacements of larger than 2
inches, but it should be noted these models ignore the post-peak softening behavior,
which was substantial in Test 2, 3, 4, and System Test.

5.3 COMPARISON OF MEASURED AND CALCULATED PASSIVE

FAILURE WEDGE

The failure plane using Log Spiral theory was predicted and compared with the
observed failure plane. In Test1, the Log Spiral theory predicted the logarithmic failure
surface extending 12.4 feet from the face of the backwall; however the observed failure
surface (from broken columns of foam) was linear along the pre-existing slope prior to

126
placement of the structure backfill which is indicated that instead of developing
logarithmic failure surface, sliding occurred at that slope.

In Test 2, the cut slope due to excavation prior to placement of the structure
backfill had almost the same angle as the predicted angle by Log Spiral theory.
Therefore, the predicted failure surface had a good agreement with the observed one. It
should be noted that when the interface friction angle is small (like when the wall is free
in vertical movement) the developed failure wedge doesn’t have a logarithmic surface
and it is almost linear. This is due to the positive value of αw when the interface friction
angle is small. In Test 2, 3, 4, and System Test, the failure plane was almost linear,
consistent with the predicted plane of failure using the Log Spiral theory.

5.4 COMPARISON OF MEASURED AND CALCULATED FORCE-

DISPLACEMENT FOR MARONEY (1995) TEST

Maroney (1995) tested two large scale diaphragm bridge abutments to failure at
University of California, Davis. Both abutments were tested under cyclic loading, the
shorter (west) one tested to failure. The west abutment was tested only in the longitudinal
direction. The schematic of the test units are shown in Figure 2-1. The soil used for the
west embankment was compacted clay of Yolo Loam. The soil properties (Romstad et al.
1995) are summarized in Table 5-6. The measured force-displacement curve of the west
abutment wall was compared with the proposed model, the hyperbolic model given by
Duncan and Mokwa (2001), and the Caltrans (2006) method (Figure 5-15). The current
Caltrans method is based on the results of this test. It can be observed that this method
captures the force-displacement behavior of Maroney’s test reasonably well. The
proposed model and hyperbolic model (using Log Spiral to estimate the passive
resistance) provides a reasonable prediction of the passive resistance (within 15%). The
proposed model is capable of capturing the shape of the passive resistance curve
appropriately.

127
5.5 COMPARISON OF MEASURED AND CALCULATED FORCE-
DISPLACEMENT ON ROLLINS AND COLE’S (2006) TESTS

A series of full-scale static load tests were performed on a concrete pile cap by
Rollins and Cole (2006). The pile cap was 3.67 ft high, 17 ft wide, and 10 ft deep. Four
different soils (sand, silty sand, fine gravel, and coarse gravel) were selected as backfill in
front of the pile cap. The backfill was compacted to 95% of the modified Proctor
maximum density. Backfill soil properties are summarized in Table 5-7. The measured
force-displacement curve for each backfill test was compared with the proposed model,
the hyperbolic model given by Duncan and Mokwa (2001), and the Caltrans (2006)
method (Figures 5-16, 5-17, 5-18, and 5-19). It should be noted that in calculating the
maximum passive resistance, the Method of Slices was not used since this method is
underestimating the maximum passive pressure for large value of fiction angles (Shields
et al. 1973). The maximum passive pressure was calculated from Log Spiral theory using
the moment equilibrium (Mokwa 1999). The Caltrans method significantly
underestimates the observed stiffness. The proposed model and hyperbolic model (using
Log Spiral to estimate the passive resistance) provides a reasonable estimate of the
passive resistance (within 15%), and the shape of the passive resistance curve.

5.6 SUMMARY OF DESIGN METHOD

This section summarizes the approach developed for analyzing the capacity and
stiffness of a bridge abutment wall forced in longitudinal direction. The soil parameters
required for the analysis are: φ, c, δ, ν, Ei, and γm.

Triaxial testing should be used to obtain φ and c. If the abutment wall is designed
to develop the failure Mechanism 1 (the backwall would not shear off from the seat
during the earthquake, Figure 2-7), the wall friction angle, δ, can be estimated based on
type of structure backfill soil and type of interface material using Table 5-8. The adhesion
factor, β, can be estimated using Table 5-1. Poisson’s ratio can be estimated from the

128
empirical equations (e.g. Appendix E, Eq. (E.1)). The initial tangent soil modulus, Ei, can
be obtained using the stress-strain results from triaxial testing, described in Appendix E.
Soil unit weight, γm, can be measured in the lab using the ASTM D1557, the modified
proctor compaction test. After determining the required soil parameters, the residual
strength of the structure backfill material is determined using the Log Spiral earth
pressure theory and Method of Slices. Section 5.1.1 describes the equations and
approaches used to calculate the maximum passive earth pressure force.

The force-displacement relationship of the abutment backwall was developed

using the rational function and exponential formulations. The required parameters for the
analysis are the initial stiffness of the backwall, kmax, the displacement at maximum
resistance, ymax, displacement at residual resistance, yresidual, and R, residual strength
factor. The initial stiffness of the backwall is approximated using the elasticity theory.
Soil parameters needed for kmax calculation are Ei and ν and the backwall dimensions.
Appendix E describes the details of the procedure to calculate the kmax. It was observed
from the test results that the ymax is ranging from 2% to 3% of the backwall height which
is consistent with the commonly adopted for the required displacement to fully mobilize
the passive earth pressure. It was also found that the residual strength of 60% to 70% of
the maximum strength reached at 3ymax displacement (yresidual=3ymax). The details of the
proposed force-displacement relationship are described in section 5.1.2. The force­
displacement relationship for the vertically restrained abutment backwall is developed
using only the rational function formulation since there was no considerable softening in
post-peak behavior of failure Mechanism 1 (Figure 2-7).

5.7 ASSESSMENT OF CALTRANS METHOD

This section assesses Caltrans approach for estimating the capacity and stiffness
of bridge abutments in the longitudinal direction. The details of each test setup are
presented earlier in Chapter 3. In summary, Phase I-Test 1 was built to simulate
diaphragm abutments similar to large scale abutment test conducted by Maroney (1995).

129
Phase I-Tests 2, 3 and 4, at the request of Caltrans engineers, were deisgned to simulate
seat-type abutment behavior, in which the backwall is sheared off from the stemwall,
using existing test setup. Phase II-System Test was conducted on a seat-type abutment
which had a backwall separated from the seat and wingwalls.

For Phase I-Test 1, a specimen was built to model of diaphragm abutment in

which translational and rotational movement is restricted. The Caltrans method
underestimates both the maximum capacity and stiffness as shown in Figure 5.4 and
Figure 5-9 respectively.

For Phase I-Test 2, a specimen was built to model seat-type abutment behavior
using exising diaphgram wall test setup from Test 1. The wingwalls and backwall were
built monolithically but vertical movement was allowed to simulate the condition in
which the backwall is sheared off from the stemwall in seat-type abutments using exising
test setup from Test 1. Stiffness prediction using Caltrans method shows good agreement
with the measured secant stiffness at displacement of larger than 1 in. as shown in Figure
5-10. However, this approach underestimated the secant stiffness at smaller displacement.
As shown in Table 5-4, all methods except Rankine overpredict the maximum capacity of
the abutment. Due to an existing slope cut, it is believed that the abutment in Test 2
developed less passive resistance and resulted in overestimating the capacity by
theoretical methods (i.e., Log Spiral and Coulomb Methods) as well as Caltrans method.

For Phase I-Test 3, a specimen was built similar to Test 2 but with greater height
and larger excavated area prior to backfill to prevent the effect of pre-existing weak
plane. Stiffness prediction using Caltrans method shows good agreement with the
measured secant stiffness at displacement of larger than 2 in. as shown in Figure 5-11.
However, this approach underestimated the secant stiffness at smaller displacement. The
predicted maximum capacity using Caltrans method is greater than the measured result.

130
 For Phase I-Test 4, a specimen was built similar to Test 2 but with larger
excavated area prior to structure backfill to prevent the effect of pre-existing weak plane.
Stiffness prediction using Caltrans approach shows good agreement with the measured
secant stiffness at displacement of larger than 1.5 inches. However, similar to Test 2 and
Test 3, this approach underestimated the secant stiffness lower displacement as shown in
Figure 5-12. The predicted maximum capacity using Caltrans method show good
agreement with the measured result as shown in Figure 5-7.

For Phase II-System Test, a specimen was built to simulate seat-type abutment
which had a backwall separated from the seat and wingwalls. Stiffness prediction using
Caltrans approach shows good agreement with measured secant stiffness at displacement
of larger than 1.2 inches as shown in Figure 5-13. Similar to Test 2, 3 and 4, this
approach underestimated the secant stiffness at lower displacement. The predicted
maximum capacity using Caltrans method shows reasonable agreement with the
measured result as shown in Figure 5-8.

The Caltrans approach was also evaluated against the experimental results of
Maroney (1995) and Rollins and Cole (2006) as shown in Figures 5-15 through 5-19.
While the Caltrans method matches the ultimate capacity reasonably well for Maroney
(1995) the test on which the procedure is based, as well as the Rollins and Cole (2006)
results for clean sand and fine gravel, it significantly underestimates the ultimate capacity
for silty sand and coarse gravel (Rollins and Cole, 2006).

In summary, the current Caltrans procedure at times works rather well, but can
often under- or over-estimate the ultimate capacity of a longitudinally loaded abutment
depending on the abutment geometry, backfill soil, and construction practice. As can be
seen from the results from this study as well as others, the Log Spiral method generally
provides a much better agreement with measured ultimate capacity. In addition, it would
be a relatively simple procedure to impact into practice that accounts for site specific soil
type and abutment geometry. Regarding the stiffness, the Caltrans procedure significantly

131
underestimates secant stiffness at displacement of less than 2 inches in nearly all cases
considered. However, the secant stiffness calculated using the Caltrans bi-linear
relationship shows good agreement with the measured secant stiffness at displacement
larger than 2 inches, as shown in several of the comparisons. Therefore, for estimated
displacements of larger than 2-inches, the current Caltrans procedure provides a
reasonable estimate of the stiffness and is reasonably easy to implement in practice. If it
is desired to model the post-peak degradation behavior of the abutment system, then the
analytical procedure developed as part of this study is recommended.

5.8 PERFORMANCE-BASED LIMIT STATES

An important advancement in structural/bridge engineering in recent years has
been the development of performance-based seismic design. The emphasis of this
methodology is on characterizing the uncertainties and variabilities inherent in the
prediction of structure performance. This framework involves global and component
damage measures (DM), engineering demand parameters (EDP), and seismic hazard
intensity measures (IM). In order to have proper EDPs and proper model for DMs, it is
necessary to have accurate characterization and prediction of the structure at the system
and component level.

For the failure mechanism developed during the field tests, three limit-states or
damage-states are defined using the proposed force-displacement model (see Figure 5­
20). These performance-based limit states are characterized as follows: Limit-State I­
where the stiffness degradation starts. It was observed that the stiffness degradation starts
at half of the maximum passive resistance and it was consistent in all tests. Up to Limit-
State I, the stiffness is constant and equal to the initial stiffness calculated based on
elasticity theory (Appendix E). The displacement associated with this force level is very
small and no repair is required at this limit-state. Limit State II- where the passive failure
wedge is formed and the maximum passive resistance is reached. It was observed from
the test results that the ypmax is ranging from 2% to 3% of the backwall height which is
consistent with the commonly adopted for the required displacement to fully mobilize the

132
passive earth pressure. A repair need associated with this limit-state is replacement of the
backfill material and relocation of the backwall to its original position. Limit-State III-
where softening or degradation in passive resistance occurs and the residual passive
resistance is reached. It was found that the residual strength of 60% to 70% of the
maximum strength is reached at 3ypmax displacement (yresidual = 3ypmax). This limit state is
associated with a repair need to replacement of cracked or damage components, structure
backfill material, and perhaps some portion of embankment material, if the failure wedge
is extended to the embankment area. Tables 5-9 to 5-11 show the computed force and
displacement at each limit-state using the proposed model for Test 1, 2, 3, 4, and the
System Test.

133
 Table 5-1. Typical range of soil adhesion factor, β (after Mokwa 1999)

Interface soil Cohesion Adhesion Factor

c (psf) (β)
Very Soft Cohesive Soil 0-250 1.0
Soft Cohesive Soil 250-500 1.0
Medium Stiff Cohesive Soil 500-1000 1.0 to 0.75
Stiff Cohesive Soil 1000-2000 0.75 to 0.5
Very Stiff Cohesive Soil 2000-4000 0.5 to 0.3

Table 5-2. Computed value of δmob for each test

Test number Mobilized interface friction angle

δmob (degrees)
Test 1 21 (=δmax)
Test 2 7.0 (=δmob)
Test 3 4.7 (=δmob)
Test 4 7.0 (=δmob)
System Test 5.1 (=δmob)

134
 Table 5-3. Input parameters for calculating the force-displacement curves of bridge
abutments

System
Variable Test 1 Test 2 Test 3 Test 4
Test
Height H (ft) 5.5 5.5 7.5 5.5 5.0
Length b (ft) 15.5 15.5 15.5 15.5 18.0
Soil unit weight γ (pcf) 120a 121a 121 121 121
Soil friction angle φ (degrees) 28 34.5 34.5 34.5 34.5
Cohesion c (psf) 1000 600 600 600 600
Wall friction δ (degrees) 21 δmobb δmob δmob δmob
Poisson’s ration ν 0.35 0.3 0.3 0.3 0.3
Initial soil
Ei (kip/ft2) 588 714 714 714 714
modulus
Adhesion factor β 0.85 0.0c 0.0 0.0 0.0
ymax/H -------- 0.04 0.02 0.02 0.02 0.03
Weight of the
w (kips) 50.5 45.5 45.5 45.5 31.5
wall+actuators
Bottom+side
f (kips) 86 0.0 0.0 0.0 48.0
friction force
3D effect factor D3D 1.1d 1.1 1.1 1.1 1.0e
Residual strength
R 0.6 0.6 0.6 0.6 0.65
factor
yresidual/ymax ------- 3.0 3.0 3.0 3.0 3.0
a
The average value from field density measurements
b
Determined by repeated trials
c
The adhesion force is negligible when the wall is free to move vertically
d
3D effect factor was assumed equal to 1.1 based on the observation during the test
e
No 3D effect was observed during the test

135
 Table 5-4. Summary of computed and measured passive resistance

Phase I Phase II
Method
Test 1* Test 2 Test 3 Test 4 System Test*
(kips) (kips) (kips) (kips) (kips)
Rankine 374.8 296.9 455.7 296.9 351.6
Coulomb 622.5 350.3 511.3 350.3 389.3
Log Spiral, without 3D 675.6 360.5 523.4 360.5 400.1
correction
Log Spiral, with 3D correction 734.5 396.5 575.7 396.5 400.1
Caltrans 512.25 426.25 792.6 426.25 457.1
Measured 788.1 334.7 576.5 381.8 415.7
*
The friction force was added to the computed passive resistance values for consistency with the
measured test data.

Table 5-5. Computed the initial stiffness of the abutment wall by Caltrans method

Phase I Phase II
Parameter
Test 1 Test 2 Test 3 Test 4 System Test
KCaltrans (kip/in/ft) 20 20 27.3 20 18.2

136
Table 5-6. Summary of Yolo Loam soil properties used for computing passive resistance
of Maroney’s Test.

West Abutment
Variable
Test
Height h(ft) 5.5
Length b (ft) 15.5
Soil unit weight γ (pcf) 117
Soil friction angle φ (degrees) 0.0
Cohesion c (psf) 2000
Wall friction δ (degrees) 20a
Poisson’s ration ν 0.30
Initial soil
Ei (kip/ft2) 450
modulus
Adhesion factor β 0.85
ymax/h -------- 0.08
3D effect factor D3D 1.1
Residual strength
Rresidual 0.65
factor
yresidual/ymax ------- 3.0
a
The structural backfill material, consisting of the unwashed sand
was at the wall-soil interface.

137
Table 5-7. Summary of soil properties used for computing passive resistance of Rollin
and Cole’s Tests.

Clean Fine Coarse Silty

Variable
Sand Gravel Gravel Sand
Height h (ft) 3.67 3.67 3.67 3.67
Length b (ft) 17 17 17 17
Soil unit weight γ (pcf) 117 132 147 121
Soil friction φ
39 34 40 27
angle (degrees)
Cohesion c (psf) 0.0 80 150 570
δ
Wall friction 30 26 30 20
(degrees)
Poisson’s ration ν 0.27 0.31 0.26 0.35
Initial soil Ei
775 706 814.5 800
modulus (kip/ft2)
Adhesion factor β ------ 0.85 0.85 0.85
ymax/h -------- 0.03 0.03 0.03 0.05
3D effect factor D3D 1.36 1.26 1.38 1.18
Residual
Rresidual 0.6 0.6 0.6 0.6
strength factor
yresidual/ymax ------- 3.0 3.0 3.0 3.0

138
Table 5-8. Interface friction angles, δ, between different backfill materials and concrete
(after U.S. Navy 1986)

Interface Friction angle

Interface material
δ (degrees)
Formed concrete against the following soils:
Clean gravel, gravel-sand mixture, well-graded rock 22 to 26
fill with spalls
Clean sand, silty sand-gravel mixture, single size hard 17 to 22
rock fill
Silty sand, gravel or sand mixed with silt or clay 17
Fine sandy silt, nonplastic silt 14

139
 Table 5-9. Computed force and displacement at Limit-State I
Test Number kmax pmax / 2 y @ pmax / 2
(kips/in) (kips) (in)
Test 1 1197 367 0.44
Test 2 1437 198 0.20
Test 3 1641 288 0.25
Test 4 1437 198 0.20
System Test 1515 200 0.21

Table 5-10. Computed force and displacement at Limit-State II

Test Number pmax y @ pmax
(kips) (in)
Test 1 734 2.6
Test 2 396 1.3
Test 3 576 1.8
Test 4 396 1.3
System Test 400 1.8

Table 5-11. Computed force and displacement at Limit-State III

Test Number Presidual y = 3y @ pmax
(kips) (in)
Test 1 440 7.9
Test 2 238 4.0
Test 3 346 5.4
Test 4 238 4.0
System Test 260 5.4

140
 O
θ A C α = 45 − φ/2
α α E

H δ HR
PP
B αw D
Rankine zone
φ 90−φ

Log Spiral
Surface

Figure 5-1. Shape of failure surface-composed of a Log Spiral and a straight line

b
x

y
x+dx w x

H E+dE E

L
S
R
α φ
N

Figure 5-2. Forces on slice element (after, Shields et al. 1973)

141
 P

Pmax

Pres=R.Pmax m
1

kmax
1
1 xinf =2 3 x=Δ/Δmax

Figure 5-3. Force-displacement relationship- sheared abutment backwall

142
Figure 5-4. Computed and measured force-displacement curves for the abutment wall in
Phase I-Test 1

143
Figure 5-5. Computed and measured force-displacement curves for the abutment wall in
Phase I-Test 2

144
Figure 5-6. Computed and measured force-displacement curves for the abutment wall in
Phase I-Test 3

145
Figure 5-7. Computed and measured force-displacement curves for the abutment wall in
Phase I-Test 4

146
Figure 5-8. Computed and measured force-displacement curves for the abutment wall in
Phase II-System Test

147
Figure 5-9. Computed and measured stiffness-displacement curves for the abutment wall
in Phase I-Test 1

148
Figure 5-10. Computed and measured stiffness-displacement curves for the abutment
wall in Phase I-Test 2

149
Figure 5-11. Computed and measured stiffness-displacement curves for the abutment
wall in Phase I-Test 3

150
Figure 5-12. Computed and measured stiffness-displacement curves for the abutment
wall in Phase I-Test 4

151
Figure 5-13. Computed and measured stiffness-displacement curves for the abutment
wall in Phase II-System Test

152
 Test 1 12.
12.4’ Test 2 14.
14.4’
11.7’ 11.6’

5.5’ 5.5’
Rankine Zone Rankine Zone

Log Spiral Prediction

Log Spiral Prediction
Observed
Obs
Obseerv
rveed

Test 3 Test 4
18.
18.5’ 12
12..3’
15.
15.3’ 11
11..6’

7.5’
7.5’ 5.
5.55’
Rankine Zone Rankine Zone

Log Spiral Pre

Preddiction
Log Spir
Spiral Predi
Predicction
Observed
Observed
Obs
Obseerved
rved

System Test
12.0’
10.2’

5.0’ Rankine Zone

Log Spiral Prediction

Observed

Figure 5-14. Computed and observed passive failure plane

153
Figure 5-15. Computed and measured force-displacement curves for the west abutment
wall in Maroney’s Test

154
 Clean sand

Figure 5-16. Computed and measured force-displacement curves of clean sand, Cole and
Rollin’s Tests

155
 Fine gravel

Figure 5-17. Computed and measured force-displacement curves of fine gravel, Cole and
Rollin’s Tests

156
 Coarse gravel

Figure 5-18. Computed and measured force-displacement curves of coarse gravel, Cole
and Rollin’s Tests

157
 Silty sand

Figure 5-19. Computed and measured force-displacement curves of silty sand, Cole and
Rollin’s Tests

158
 Pmax Limit State II

Presidual Limit State III

Horizontal Force

Pmax/2 Limit State I

y pmax/2 y pmax 3yPmax

Horizontal Displacement

Figure 5-20. Performance-based limit-states of abutment backwall

159
6 SUMMARY AND CONCLUSIONS

An experimental and analytical research program was conducted to investigate

abutment capacity and stiffness from field tests. The key variables in the experimental
study were structure backfill soil type, area of structure backfill, backfill height, and
vertical movement of the wall. It is believed that each of these variables has an important
influence on the response of bridge abutments. The analytical part of this research
program was focused on developing a model to predict the force-displacement
relationship of bridge abutments accurately, considering these important factors.

The work accomplished in this study includes developing “typical” soil

parameters for structure backfill materials based on a statewide field investigation,
conducting five large-scale abutment tests loaded longitudinally, performing laboratory
tests on imported structure backfill materials, and developing an analytical model to
predict the force-displacement relationship of a bridge abutment as a function of soil
type.

6.1 EXPERIMENTAL INVESTIGATION

Initially, a field investigation was conducted throughout the state of California.

The objective of this investigation was to have a proper characterization of the soil types
used for structure backfill behind abutments and its potential variation in the field. The
results of the field investigation helped determine the type of structure backfill soils to be
used for the tests. The results from the field investigation showed a quite wide range of
soil types have been used as structure backfill materials in bridge abutments in state of
California. Based on field investigation, soil types were sorted into three soil categories
(sands, silty/clayey sands, and lean clay) with representative geotechnical properties.
Table 3-3 defines these three categories.

160
 A series of large-scale field tests were then performed at UCSD on bridge
abutments loaded in the longitudinal direction with key variables of soil type, area of
structure backfill, backfill height, the vertical movement of the wall. The experimental
program included two phases. In the first phase of the experiment, a diaphragm abutment
(without a foundation) was built at full height-scale and 50% width-scale of a prototype
diaphragm abutment, to study the longitudinal stiffness and capacity of the abutment.
Four tests were carried out in Phase I. Except for the first test, the abutment wall was
allowed to move in the vertical direction to resemble the response of a sacrificial
backwall. It should be noted that the wall was restrained from the rotational movement
about three directions in all the tests, Phase I and II. In Test 1, the wall was restrained in
the vertical direction to study the case where the backwall is not free to move vertically.
The abutment backwall is restrained from vertical movement: 1) if the backwall does not
shear off from seat during the earthquake due to poor detailing of the construction joint at
the seat-backwall interface, 2) if the friction force at the superstructure-backwall interface
is large enough to resist the wall vertical movement, 3) if the abutment wall does not
shear off from its foundation in case of diaphragm abutments.

The second phase of this research program investigated the stiffness and strength
of a seat-type abutment. In this test, it was assumed that the backwall was already sheared
off from the seat (stemwall) and wingwalls and the only force-resistance system of the
abutment was provided by backwall-structure backfill interaction.

Results from testing program support the main thesis of the research, that the
response of bridge abutment is nonlinear with regard to to longitudinal loading.
Furthermore, the capacity and stiffness of bridge abutments depend on many factors
which were studied during this research program to evaluate their influence. These
factors are:

1. Soil properties: The amount of passive resistance that develops behind a bridge
abutment depends on the soil shear strength which can be determined by a set of

161
 laboratory tests. To evaluate the passive earth pressure over a range of wall
movement, both soil shear strength and stiffness should be considered.

2. Vertical wall movement: If the wall is free to move in the vertical direction (the
backwall is sheared off from the stemwall in seat-type abutment and there is no
vertical restraining force in the backwall), small relative displacement across the
interface occurs which results in only partial mobilization of the interface friction.
In this case, the shape of the developed passive failure surface is approximately
linear, which is consitent with log-spiral theory for low wall friction..
Additionally, abutment walls free to displace in the vertical direction develop less
passive earth resistance than the abutment walls restrained in vertical movement.
Degradation in post-peak behavior of force-displacement relationship should be
considered for the wall with no restriction in vertical movement.

3. Height of the backfill: Experimental results indicate that the current Caltrans
procedure, using linear proportionality relationship (i.e., Eq. 2.9), to estimate
maximum capacity of bridge abutments may not be a good predictor of actual
capacity. However, this is based on these limited experimental results. Further
testing of seat-type abutments of various dimensions is recommended to
investigate the validity of this method.

4. Area of structure backfill: It was intended to follow the same construction

phases of real bridge abutments in this project to study the influence of
construction method on bridge abutment behavior. Based on common practice,
the structure backfill material is placed at a very limited area restrained between
the wingwalls for diaphragm-type abutments. There are two key issues related to
having a small structure backfill area. First, it is important to ensure that the
passive failure wedge occurs within the structure backfill area and it doesn’t
extend into the generally weaker embankment material. Secondly, as occurred in
Test 1 and Test 2, the excavated slope prior to backfilling introduces a weak plane

162
 of failure to the system and sliding occurs at that excavated surface instead of
developing passive failure wedge within the structure backfill material. Thus, it
develops less earth pressure resistance and changes the failure zone. Therefore, it
is important to ensure that the excavated surface is beyond the passive failure
wedge surface.

In conclusion, the load tests performed in this study indicate that the response of
bridge abutments in longitudinal direction is nonlinear and function of several influential
factors which need to be considered in prediction. The major resistance of bridge
abutments in longitudinal direction is provided by backwall-soil interaction, and the
passive earth resistance behind the bridge abutments. The passive resistance of the
backwall is controlled by the soil shear strength and interface friction angle. The vertical
movement of the wall has a significant effect on post-peak behavior of abutments.

6.2 ANALYTICAL STUDY

An analytical model was developed for evaluating the response of bridge

abutments loaded longitudinally. The approach involves calculating maximum passive
resistance of the structure backfill material, and creating p-y curves to predict force­
displacement relationship of longitudinally loaded bridge abutments.

The Log Spiral theory was used to estimate the passive earth pressure developed
back of the abutment wall. Based on Terzaghi’s (1943) suggestion the failure surface
composed of a logarithmic spiral and a straight line. In order to calculate the total passive
earth pressure, the Method of Slices was used. The method of calculating the maximum
passive force has been programmed in Matlab.

In order to evaluate the capacity of bridge abutments over a range of expected

movements during earthquakes, a model was developed to predict the response of
abutments during the loading and after reaching the peak capacity. Since the resistance of
bridge abutments in the longitudinal direction is mainly provided by backwall-soil

163
interaction and the passive earth pressure, the proposed approach models the passive
earth resistance developed behind the backwall. The relationship between passive earth
force and the backwall displacement is characterized by p-y curves using rational
function. The rational function formulation in conjunction with the exponential
formulation is capable of predicting the nonlinear force-displacement relationship of the
abutment wall with the post-peak capacity degradation. Note that the force-displacement
relationship of the abutment wall which is not expected to shear off during the earthquake
can be approximated using just the ration function formulation. The comparison of
calculated and measured force-displacement values fall within an acceptable range.
Overall, the developed model represents a significant improvement over the current
design practice.

6.3 SIMPLIFIED CALTRANS BI-LINEAR MODEL

The current Caltrans procedure at times works rather well, but can often under- or
over-estimate the ultimate capacity of a longitudinally loaded abutment depending on the
abutment geometry, backfill soil, and construction practice. As can be seen from the
results from this study as well as others, the Log Spiral method generally provides a
much better agreement with measured ultimate capacity. In addition, it would be a
relatively simple procedure to impact into practice that accounts for site specific soil type
and abutment geometry. Regarding the stiffness, the Caltrans procedure significantly
underestimates secant stiffness at displacement of less than 2 inches in nearly all cases
considered. However, the secant stiffness calculated using the Caltrans bi-linear
relationship shows good agreement with the measured secant stiffness at displacement
larger than 2 inches, as shown in several of the comparisons. Therefore, for estimated
displacements of larger than 2-inches, the current Caltrans procedure provides a
reasonable estimate of the stiffness and is reasonably easy to implement in practice. If it
is desired to model the post-peak degradation behavior of the abutment system, then the
analytical procedure developed as part of this study is recommended.

164
6.4 RECOMMENDATIONS FOR FUTURE RESEARCH

The response of bridge abutments subjected to longitudinal loading was

investigated in this research project. The results of the experimental studies showed that
the response of bridge abutments in longitudinal direction is function of several factors
which need to be considered in prediction. The analytical approach and developed model
to predict the response of bridge abutments subjected to longitudinal loading are expected
to be useful to improve the design method for bridge abutments. The following
recommendations for future research include additional experimental tests, as well as
numerical studies.

1. Additional experimental studies are needed to explore the effect of the approach
slab, and the interface friction between the backwall and superstructure after gap
closure. These studies should include the large-scale seat-type abutment
considering the approach slab on top of the structure backfill. The backwall
should be loaded by means of the deck which is resting on bearings on the
stemwall. The experimental results would provide the information which defines
the failure mode and the amount of vertical wall movement. The experimental
studies would be enhanced by performing finite element analysis to verify the
interpretation based on experimental observation.

2. The performance of bridge abutments during dynamic loading suffers from a

considerable level of uncertainties. Research, including large-scale experimental
studies, is needed in this area to evaluate the capacity and stiffness of bridge
abutments subjected to seismic loading. This study would be enhanced by
implement a finite element model of an entire bridge which explicitly
incorporates the nonlinear soil response behind the abutment wall.

3. Research is needed to investigate the response of skewed bridge abutments during

cyclic and seismic loading. Large-scale experimental studies should be included

165
 in a well planned test program to explore the unknown variables related to skewed
abutment such as tensional stiffness.

4. A close look should be taken at the design of the seat-type abutments in order to
develop a design that will allow the backwall to shear off from seat and wingwalls
during major seismic events and plow into the backfill.

5. Testing of seat-type abutments of various dimensions is recommended to

investigate the validity of the current method (SDC, 2006) using linear
proportionality constant to estimate stiffness and maximum capacity of bridge
abutments.

166
REFERENCES

AASHTO (2004). “LRFD Bridge Design Specifications.” American Association of State

Highway and Transportation Officials, Washington, D.C.

Applied Technology Council (ATC). (1996). Improved seismic design criteria for
California bridges: Provisional recommendations.

ASTM, D 1140, “Standard Test Method for Amount of Material in Soils Finer Than the
No. 200 (75-μm) Sieve”, American Society for Testing and Materials, West
Conshohocken, Pennsylvania, USA.

ASTM, D 1557, “Standard Test Method for Laboratory Compaction Characteristics of

Soil Using Modified Effort”, American Society for Testing and Materials, West
Conshohocken, Pennsylvania, USA.

ASTM, D 2419, “Standard Test Method for Soils for Sand Equivalent Value of Soils and
Fine Aggregate”, American Society for Testing and Materials, West Conshohocken,
Pennsylvania, USA.

ASTM, D 2487, “Standard Classification of Soils for Engineering Purposes (Unified Soil
Classification System”, American Society for Testing and Materials, West
Conshohocken, Pennsylvania, USA.

ASTM, D 4318, “Standard Test Method for Liquid Limit, Plastic Limit, and Plasticity
Index of Soils”, American Society for Testing and Materials, West Conshohocken,
Pennsylvania, USA.

167
Bowls, J., (1988). Foundation Analysis and Design (4th edition), McGraw-Hill, Inc., New
York.

Bozorgzadeh, A., Megally, S., Restrepo, J. I., Ashford, S. A. (2006). “Capacity

evaluation of exterior sacrificial shear keys of bridge abutments.” J. Bridge Engrg.,
11(5), 555-565.

Briaud, J. L., Lim, Y. (1997). “Soil-nailed wall under piled bridge abutment: simulation
and guidelines.” J. Geotech. and Geoenvir. Engrg., ASCE, 123(11), 1043-1050.

California Test Method (CTM) 217, November 1999.

Caltrans. (2002). Caltrans Standard Specifications.

Clough, G. W., Duncan, J. M. (1971). “Finite element analysis of retaining wall

behavior.” Journal of Soil Mechanics and Foundations Division, ASCE, 97(SM12),
December, 1657-1673.

Cole, R. T., Rollins, K. M. (2006). “Passive earth pressure mobilization during cyclic
loading.” J. Geotech. and Geoenvir. Engrg., ASCE, 132(9), September, 1154-1164.

Das, B. M., (2005). “Fundamentals of Geotechnical Engineering”, 2nd edition, Thomson

Douglas, D. J., Davis, E. H. (1964).” The movement of buried footings due to moment
and horizontal load and the movement of anchor plates.” Geotechnique, 14(2), 115-132.

Duncan, J. M., and Mokwa, R. L. (2001). “Passive earth pressure: theories and tests.” J.
Geotech. and Geoenvir. Engrg., ASCE, 127(3), 248-257.

Duncan, J. M., and Chang, C. Y. (1970). “Nonlinear analysis of stress and strain in soils.”
J. Soil Mech. And Found. Div., ASCE, 96(5), 1629-1653.

168
Earth Mechanics, Inc. (2005). Field Investigation Report for Abutment Backfill
Characterization, January 2005.

Eid H. T., Stark, T. D., Doerfler, C. K. (1999). “Effect of shear displacement rate on
internal shear strength of a reinforced geosynthetic clay liner.” Geosynthetics
International, 6(3), 219-239.

Fang, Y-S., Chen, T-J., Wu, B-F. (1994). “Passive earth pressures with various wall
movements.” J. Geotech. and Geoenvir. Engrg., ASCE, 120(8), August, 1307-1323.

Fang, Y-S., Chen, J. M., Chen, C. Y., (1997). “Earth pressures with sloping backfill.” J.
Geotech. and Geoenvir. Engrg., ASCE, 123(3), 250-259.

Fang, Y-S., Chen, C. Y., Chen, J. T., (2002). “Passive earth pressures with critical
concept.” J. Geotech. and Geoenvir. Engrg., ASCE, 128(8), 651-659.

Fang, Y-S., Chen, T-J., (2004). “Reduction of boundary friction in model tests.”
Geotechnical Testing Journal, 27(1), 3-12.

Faraji, S., Ting, J. M., Crovo, D. S., Ernst, H., (2001). “Nonlinear analysis of integral
bridges: finite-element model.” J. Geotech. and Geoenvir. Engrg., ASCE, 127(5), 454­
461.

Gadre, A. D., Dobry, R. (1998). “Lateral cyclic loading centrifuge tests on square
embedded footing.” J. Geotech. and Geoenvir. Engrg., ASCE, 124(11), November, 1128­
1138.

Goel, R. K., Chopra, A. K. (1997). “Evaluation of bridge abutment capacity and stiffness
during earthquakes.” Earthquake Spectra, 13(1), February, 1-23.

169
James, R. G., Bransby, P. L. (1970). “Experimental and theoretical investigations of a
passive pressure problem.” Geotechnique, 20 (1), 17-37.

Maroney, B. H. (1995). “Large scale bridge abutment tests to determine stiffness and
ultimate strength under seismic loading.” PhD thesis, Dept. of Civil Engineering,
University of California Davis, CA.

Martin, G. R., Yan, L.-P., and Lam, I. P. (1997). “Development and implementation of
improved seismic design and retrofit procedures for bridge abutments.” Final Report on a
Research Project Funded by the California Dept. of Transportation.

Mokwa, R. L. (1999). “Investigation of the resistance of pile caps to lateral loading.”

PhD thesis, Dept. of Civil Engineering, Virginia Polytechnic Institute and State
University, Blacksburg, Virginia.

Narain, J., Saran, S., Nandakumaran, P. (1969). “Model study of passive pressure in
sand.” J. of Soil Mech. and found. Engrg. Div., ASCE, 95(4), 969-983.

Naval Facilities Engineering Command, “Foundations and Earth Structures, Design

Manual 7.02” ,Alexandria VA, September, 1986.

OpenSees, The Open System for Earthquake Engineering Simulation.,

http://opensees.berkeley.edu/index.php, 1998-2007.

Plaxis (2006). Finite Element Code for Soil and Rock Analysis. 2D-Version 8,
www.plaxis.nl.

Potyondy, J. G. (1961). “Skin friction between various souls and construction materials.”
Geotechnique, 11(1), 339-353.

170
Rollins, K. M., Cole, R. T. (2006). “Cyclic lateral load behavior of a pile cap and
backfill.” J. Geotech. and Geoenvir. Engrg., ASCE, 132(9), September, 1143-1153.

Rollins, K. M., Sparks, A., E. (2002). ”Lateral resistance of full-scale pile cap with gravel
backfill.” J. Geotech. and Geoenvir. Engrg, ASCE, 128(9), 711-723.

Romstad, K., Ketter, b., Maroney, B. H. (1995). “Large scale bridge abutment tests to
determine stiffness and ultimate strength under seismic loading.” PhD thesis, Dept. of
Civil Engineering, University of California Davis, CA.

Rowe, P. W., Peaker, K., (1965). “Passive earth pressure measurements.” Geotechnique,
Londong, England, 15(1), 57-78.

Salgado, R., Bandini, P., Karim, A. (2000). “Shear strength and stiffness of silty sand.” J.
Geotech. and Geoenvir. Engrg., 126(5), 451-462.

Sargin, M., Handa, V. K. (1968). “Structural concrete and some numerical solutions.”
Proceeding of the 1968 23rd ACM National Conference, ASCE, 99(SM12), August 27­
29, 563-574C.

SDC, Caltrans. (2006). Seismic design of abutments for ordinary standard bridges.

Shamsabadi, A., Ashur, M., and Norris, G. (2005). “Bridge abutment nonlinear force­
displacement-capacity prediction of seismic design.” J. Geotech. and Geoenvir. Engrg.,
ASCE, 131(2), 151-161.

Shamsabadi, A., Nordal, S., (2006). “Modeling passive earth pressures on bridge
abutments for nonlinear seismic soil-structure interaction using Plaxis.” Plaxis Bulletin,
issue 20, October, 8-15.

171
Shields, D. H., Tolunay, A. Z. (1973). “Passive pressure coefficients by method of
slices.” Journal of Soil Mechanics and Foundations Division, ASCE, 99(SM12),
December, 1043-1053.

Shing, B. P., Spencer, B. (2001). “Modeling of shear behavior of RC bridge structure.”

Modeling of Inelastic Behavior of RC structures under Seismic Loads, ASCE, 315-333.

Siddaharthan, R., El-Gamal, M., and Maragakis, E. A. (1995). “Influence of free-field

strains on nonlinear lateral abutment stiffness.” Proceedings of 7th Canadian Conference
on Earthquake Engineering, Montreal, Canada, 739-346.

Stark, T. D., Eid, H. T. (1994). “Drained residual strength of cohesive soils.” J. Geotech.
Engrg., ASCE, 120(5), 856-871.

Terzaghi, K. (1943). Theoretical soil mechanics, Wiley, New York.

Terzaghi, K., Peck, R. B., and Mezri, G., (1996). Soil mechanics in Engineering Practice,
3rd Ed., Wiley, New York.

Wilson, j. C. (1988). “Stiffness of non-skew monolithic bridge abutments for seismic

analysis.” Earthquake Engineering and Structural Dynamics, 14, 339-354.

Yang, Z., Boris, J., (2003). ”Numerical study of group effects for pile groups in sands.”

International Journal for Numerical and Analytical Methods in Geomechanics, 2003, 27,

1255-1276.

172
173
APPENDIX A - CU TRIAXIAL TEST RESULTS

174
Triaxial Test Results, Clayey Sand:

175
176
177
178
179
180
181
182
Triaxial Test Results, Silty Sand:

183
184
185
186
187
188
189
190
APPENDIX B - REINFORCING DETAILS OF THE TEST
BASE

191
 Figure B-1 shows the reinforced details of the reaction block which was used as a
test base in Phase I and II abutment tests.

Figure B-1. Reinforcing details of the reaction block

192
 Figure B-2 shows the reinforcing detail of the foundation block. The abutment test
unit in Phase I was built on top of the foundation block with no structural join between
the foundation block and test unit.

Figure B-2. Reinforcing details of the foundation block

193
APPENDIX C - MAXIMUM DENSITY AND OPTIMUM
MOISTURE

194
Maximum Density and Optimum Moisture, Clayey Sand:

195
Maximum Density and Optimum Moisture, Silty Sand:

196
APPENDIX D - EQUATIONS FOR DEFINING THE
CENTER OF LOG SPIRAL PASSIVE WEDGE

197
Calculate the center of the logarithmic spiral of the passive failure wedge:

This Appendix describes the approach to calculate the center of the Log Spiral of
the passive failure wedge. Shields et al. (1973) developed a simplified slice method to
calculate the passive earth pressure. This approach was used to calculate the center of
Log Spiral.

Mokwa et al. (1999) developed a method to determine the location of the center
of the log spiral by iterating on the values of r (radius of the log spiral), x0 (the horizontal
distance of the log spiral center to the wall) until the dimensions of the failure surface are
consistent. Unlike other methods, the equations presented in this appendix, can determine
the geometry and sizes of the failure zone without iterating procedure. The equations
used to determine the geometry of the log spiral are given below. The dimensions that are
used in this analysis are shown in Figure D-1.
x0 w

O
y0 α
θ A C α = 45 − φ/2
E
λ α α
r0
r hR
h

B D
αw Rankine zone
φ 90−φ

Log Spiral
surface
Figure D-1. Log Spiral failure surface

The angle between the horizontal line and the failure surface at the wall is
determined by αw, which is function of φ and δ. The angle αw is considered to be positive
when it is above the horizontal and negative when is below the horizontal. The angle

198
between the horizontal and line OB at the center, O, and at the wall corner, B, are
identical:

φ
90 − φ − α w = θ max + 45 − (D.1)
2

φ
θ max = 45 − α w − (D.2)
2

The angle between vertical and line OB, at point O, is equal to:

φ φ φ
λ = 90 − (θ max + 45 − ) = 90 − 45 + α w + − 45 + (D.3)
2 2 2

Therefore λ = α w + φ (D.4)
From trigonometry:

⎛ x0 ⎞
λ = tan −1 ⎜⎜ ⎟⎟ (D.5)
⎝ y 0 + h ⎠

Therefore x0 = ( y 0 + h ) tan λ (D.6)

On the other hand:

y0 ⎛ φ⎞
= tan ⎜ 45 − ⎟ (D.7)
x0 ⎝ 2⎠

⎛ φ⎞
Therefore y 0 = x 0 tan⎜ 45 − ⎟ (D.8)
⎝ 2⎠

199
By substituting Eqs. (D.8) and (D.4) into Eq. (D.6):

⎛ ⎛ φ⎞ ⎞
x0 = ⎜⎜ x0 tan⎜ 45 − ⎟ + h ⎟⎟ tan (α w + φ ) (D.9)
⎝ ⎝ 2⎠ ⎠

Solving for x0:

h tan (α w + φ )
x0 = (D.10)
⎛ φ⎞
1 − tan (α w + φ ) tan ⎜ 45 − ⎟
⎝ 2⎠

Then by substituting Eq. (D.10) into Eq. (D.8):

h tan (α w + φ )
y0 = (D.11)
1 − tan (α w + φ )

Therefore, knowing the location of the center of log spiral, the failure surface can be
established. The equation defining the log spiral surface is:

r = r0 e (θ tan φ ) (D.12)

where r is the radius of the log spiral at the angle of θ (the angle between r and r0), φ is
the soil friction angle, and r0 is the starting radius and equal to:

r0 = x02 + ( y 0 + h )
2
(D.13)

Therefore, the depth of Rankine zone, hR, can be calculated by:

200
 φ
hR = rmax sin(45 − ) − y 0 (D.14)
2

where rmax is the maximum log spiral radius and can be calculated by substituting θ=θmax
in Eq. (D.12). The maximum angle, θmax, can be calculated by Eq. (D.2). Therefore:

hR
w= (D.15)
⎛ φ⎞
tan⎜ 45 − ⎟
⎝ 2⎠

201
APPENDIX E - COMPUTING THE ABUTMENT WALL
INITIAL ELASTIC STIFFNESS

Abutment wall initial stiffness:

The initial stiffness of the abutment backwall which is corresponding to the initial
slope of the load-displacement curve can be approximated by using elasticity theory.
Douglas and Davis (1964) developed the elastic solution for calculating the horizontal
displacement of a vertical rectangle in an elastic half-space. Duncan and Mokwa (2001)
used this method to estimate the pile cap stiffness. The displacement of the flexible
rectangular structure is taken as the average displacement of the corners. The parameters
needed to estimate kmax include Poisson’s ratio (ν), the initial tangent modulus of the soil
(Ei), and the dimensions of the backwall. The values of the Poisson’s ratio for soils can
be approximated by the following empirical equation:

1 − sin φ
ν= (E.1)
2 − sin φ

where ν is the Poisson’s ratio; and φ is the internal friction angle (degrees). Based on Eq.
(E.1) a Poisson’s ratio of 0.3 was estimated for the silty sand and a value of 0.35 was
calculated for the clayey sand. The analysis is not sensitive to ν, and Eq. (E.1) estimates
the reasonable values based on internal friction angle of soils. Values of the initial
tangent modulus, Ei, were estimated using the hyperbolic formulation described by
Duncan and Change (1970). In order to determine the Ei value, the stress-strain data
(from triaxial tests) are plotted on transformed axes as shown in Figure E-1 and Figure E­
2.The estimated values of Ei are 714 ksf and 588 ksf for silty sand and clayey sand,
respectively.

202
 The equations used to calculate kmax are given below. The dimensions that are
used in this analysis are shown in Figure E-3.

Surface, z=0
A D

H
B C
b
Figure E-3. Schematic of a vertical rectangle in an elastic half-space

25
Stress Difference-( σ1- σ3), ksf

Difference- ε /(σ1-σ3), 1/ ksf

Asymptote=(σ1-σ3)ult=1/b 0.018
20
0.015
Axial Strain/Stress
Slope: b= 0.0439 ksf-1
Ei=1/a
15 0.012 Intercept: a=14 x10-4 ksf-1
1
Lab data point 0.009
10 b
Hyperbolic curve fit 0.006
1
5 Lab data point
0.003
a
0 0
0 0.1 0.2 0.3 0 0.1 0.2 0.3
Axial Strain- ε Axial Strain- ε

a) Hyperbolic stress-strain curve-silty sand b) Transformed stress-strain curve-silty

sand

Figure E-1. Stress-strain curves from triaxial testing-Silty Sand

12 Asymptote=(σ1-σ3)ult=1/b
Stress Difference-( σ1-σ3), ksf

0.03
Difference- ε/( σ1- σ3), 1/ ksf

10
Ei=1/a 0.025 Slope: b= 0.0861 ksf-1
Axial Strain/Sress

8 Intercept: a=17 x10-4 ksf-1

1 0.02
Lab data point
6
lab data point
Hyperbolic curve fit 0.015 b
4 0.01 1

2 0.005
0 a
0
0 0.1 0.2 0.3
0 0.1 0.2 0.3
Axial Strain-ε Axial Strain-ε

a) Hyperbolic stress-strain curve-clayey d) Transformed stress-strain curve-clayey

sand sand

Figure E-2. Stress-strain curves from triaxial testing-Clayey Sand

203
For a uniform pressure, p, applied horizontally, the displacement, y1 at corner A and B is:

pb(1 − ν )
y1 = {(3 − 4ν )F1 + F4 + 4(1 − 2ν )(1 − ν )F5 } (E.2)
16πEi (1 − ν )

The displacement at corner of the bottom of the rectangular area, y2, is:

pb(1 − ν )
y2 = {(3 − 4ν )F1 + F2 + 4(1 − 2ν )(1 − ν )F3 } (E.3)
16πEi (1 − ν )

where p is the uniform pressure applied to the rectangular area ABCD horizontally, ν is
the Poisson’s ratio, Ei is the initial soil tangent modulus, b is the wall length, F1 through
F5 are the function factors described below.

⎛ K1 ⎞ ⎛ 2 ⎞
F1 = −K 1 ln⎜ ⎟ − 2 ln⎜ ⎟ (E.4)
⎜2+ 4+ K2 ⎟ ⎜K + 4+ K2 ⎟
⎝ 1 ⎠ ⎝ 1 1 ⎠

F2 = 2 ln ⎜ (
⎛ 2 K + 1+ K 2
1 1 )⎞⎟ + K ln⎛⎜ 2 + 4 + K 12 ⎞⎟
− K1
⎛ 4+ K2
2⎜ 1
−
1 + K 12 ⎞
⎟ (E.5)
⎜ ⎟ 1
⎜ K1 ⎟ ⎜ K1 K1 ⎟
⎝ K1 + 4 + K1
2
⎠ ⎝ ⎠ ⎝ ⎠

⎛ ⎞ ⎛ ⎞ ⎛
( 4+ K )
2 ⎞

F3 = −2K1 ln⎜
K1 ⎟ + K ln⎜ K1 ⎟ − ln⎜ K1 + 4 + K1 ⎟ + K1
( ) − K1
2
⎜ 1+ 1+ K 2 ⎟ 1
⎜ 2+ 4+ K2 ⎟ ⎜ 2 K + 1+ K 2 ⎟ 4 1
⎝ 1 ⎠ ⎝ 1 ⎠ ⎝ 1 1 ⎠

(
− K1 1+ K12 − K1 ) (E.6)

⎛ ⎞ ⎛ ⎞
⎟ + K ln⎜ 2 + 4 + K 1
2
2
F4 = −2 ln⎜ ⎟ (E.7)
⎜ ⎟ 1
⎜ K1 ⎟
⎝ K1 + 4 + K1
2
⎠ ⎝ ⎠

204
 ⎛ ⎞ ⎛
⎟ + ln⎜ K 1 + 4 + K 1
⎞ K
( 4+ K )
2
K1
F5 = −K 1 ln⎜ ⎟− 1 2
− K1 (E.8)
⎜2+ 4+ K2 ⎟ ⎜ 2 ⎟ 4 1
⎝ 1 ⎠ ⎝ ⎠
where K1=2H/b. the load-displacement relationship using Douglas and Davis’s (1964)
equations is linear elastic. Thus, kmax is the slope of the load-displacement curve. The
applied, P, is given by:

P = pbH (E.9)

where p is the applied horizontal pressure, b is the length of the wall and H is the height
of the wall. The average displacement, yave, is :
y1 + y 2
y ave = (E.10)
2

Therefore, the initial elastic stiffness, kmax, is:

P
k max = [unit: force per length] (E.11)
y ave

205
APPENDIX F - DISTRIBUTION OF SOIL PRESSURE AT
WALL SURFACE VS. DEPTH OF THE WALL

206
 0 20 40 60 80 100
0
-10 PSW4
-20

Depth, inch
-30 PSW3 100 kips
200 kips
-40 300 kips
PSW2 400 kips
-50 500 kips
600 kips
-60 PSW1 700 kips
-70
Soil pressure at w all surface, psi

Figure F. 1. Distribution of soil pressure at west side of the wall surface vs. depth, Test 1

0 20 40 60 80 100 120
0
PSC4
-10
-20
Depth, inch

-30 PSC3 100 kips

200 kips
-40 PSC2 300 kips
400 kips
-50 500 kips
600 kips
-60 PSC1 700 kips
-70
Soil pressure at w all surface, psi

Figure F. 2. Distribution of soil pressure at center of the wall surface vs. depth, Test 1

- 207 ­
 0 10 20 30 40 50 60
0
-10
PSW4
-20

Depth, inch
PSW3
-30
100 kips
-40
PSW2 200 kips
-50
300 kips
-60
PSW1
-70
Soil pressure at wall surface, psi

Figure F. 3. Distribution of soil pressure at west side of the wall surface vs. depth, Test 2

0 10 20 30 40 50
0
-10 PSC4
-20
Depth, inch

-30
PSC3
-40
100 kips
-50 PSC2
200 kips
-60 300 kips
PSC1
-70
Soil pressure at wall surface,psi

Figure H. 4. Distribution of soil pressure at center of the wall surface vs. depth, Test 2

- 208 ­
 0 10 20 30 40 50 60
0
-10 PSW5
-20
-30 PSW4

Depth, inch
-40
100 kips
-50 PSW3 200 kips
-60
PSW2 300 kips
-70 400 kips
-80 500 kips
PSW1
-90
Soil pressure at wall surface, psi

Figure F. 5. Distribution of soil pressure at west side of the wall surface vs. depth, Test 3

0 20 40 60 80
0
-10
PSC5
-20
-30
Depth, inch

-40 PSC4 100 kips

-50 PSC3 200 kips
-60 300 kips
-70 PSC2 400 kips
-80 PSC1 500 kips
-90
Soil pressure at wall surface, psi

Figure F. 6. Distribution of soil pressure at center of the wall surface vs. depth, Test 3

- 209 ­
 0 10 20 30 40 50
0
-10 PSW4
-20

Depth, inch
PSW3
-30
-40 PSW2
-50 100 kips
200 kips
-60 PSW1 300 kips
-70
Soil pressure at wall surface, psi

Figure F. 7. Distribution of soil pressure at west side of the wall surface vs. depth, Test 4

0 10 20 30 40 50
0
-10 PSC4
-20
Depth, inch

PSC3
-30
100 kips
-40
PSC2 200 kips
-50 300 kips
-60 PSC1
-70
Soil pressure at wall surface, psi

Figure F. 8. Distribution of soil pressure at center of the wall surface vs. depth, Test 4

- 210 ­
 0 10 20 30 40 50 60
0
-10 PSW4

Depth, inch
-20
PSW3
-30
100 kips
-40 PSW2 200 kips
-50 300 kips
PSW1 400 kips
-60
Soil pressure at wall surface, psi

Figure F. 9. Distribution of soil pressure at west side of the wall surface vs. depth,
System Test

0 10 20 30 40 50 60
0
100 kips
-10 PSC4 200 kips
300 kips
Depth, inch

-20 400 kips

-30 PSC3
-40 PSC2
PSC1
-50
-60
Soil pressure at wall surface, psi

Figure F. 10. Distribution of soil pressure at center of the wall surface vs. depth,
System Test

- 211 ­