Reliability Analysis of Settle

RELIABILITY ANALYSIS OF SETTLEMENT FOR SHALLOW
FOUNDATIONS IN BRIDGES
by
Aseel Yassen Ahmed
A DISSERTATION
Presented to the Faculty of
The Graduate College at the University of Nebraska
In Partial Fulfillment of Requirements
For the Degree of Doctor of Philosophy
Major: Civil Engineering
Under the Supervision of Professors Andrzej S. Nowak and Maria M. Szerszen
Lincoln, Nebraska
July, 2013
UMI Number: 3590300
All rights reserved
INFORMATION TO ALL USERS

The quality of this reproduction is dependent upon the quality of the copy submitted.
In the unlikely event that the author did not send a complete manuscript
and there are missing pages, these will be noted. Also, if material had to be removed,
a note will indicate the deletion.
UMI 3590300
Published by ProQuest LLC (2013). Copyright in the Dissertation held by the Author.
Microform Edition © ProQuest LLC.
All rights reserved. This work is protected against
unauthorized copying under Title 17, United States Code
ProQuest LLC.
789 East Eisenhower Parkway
P.O. Box 1346
Ann Arbor, MI 48106 - 1346
RELIABILITY ANALYSIS OF SETTLEMENT FOR SHALLOW
FOUNDATIONS IN BRIDGES
Aseel Yassen Ahmed, Ph.D.
University of Nebraska, 2013
Advisers: Andrzej S. Nowak and Maria M. Szerszen
The design of shallow foundations is becoming an attractive choice for supporting
highway bridge structures. The settlement of structures built on soil is a subject of
considerable interest to practicing engineers since excessive settlements can lead to
serviceability problems or even failure states in the structural foundations. Therefore, the
design of shallow foundations is most often governed by settlement requirements.
A reliability-based approach can be applied for designing shallow foundations for
which their design needs knowledge about the distribution of settlement under a given
footing. The performance of geotechnical structures, including shallow foundations depends
mostly on loads and soil properties, which are the random variables described by statistical
parameters. However, most of the current research efforts have been focused on the
reliability models for the superstructure components with less attention paid to the
substructure and foundations. The random variables and limit state functions for shallow
foundations are very different from those for superstructures such as bridges or buildings.
The objective of this study is to develop the methodology for risk analysis procedures
of shallow foundations. This requires development of the reliability models that can be applied to
shallow foundations and involves the establishment of serviceability limit state functions.
Reliability, as it can be expressed as a measure of structural performance, is expressed in
terms of reliability indices, which in this research are calculated for total and differential
settlement of shallow bridge foundations. The most important part of this study is the
development of statistical models for total and differential settlements of shallow foundations on
cohesionless soils by Monte Carlo simulations technique. This study verifies six settlement
prediction methods for spread footing foundations by using compiled field data from research
projects based on vertical settlements measured on 23 sites of footings for 12 highway bridges
selected in US.
i
DEDICATION
BY THE NAME OF GOD
This dissertation is dedicated to
The soul of the greatest woman in the world
To my lovely mother
My devoted beloved husband Mohammed for his
unconditioned love and support
My three beautiful angels Zainab, Ahmed, and Alhussain
My lovely father Dr. Yassen
My lovely brothers Rafed and Mustafa

ii
AKNOWLEDGEMENTS
I wish to express my sincere thanks to Professor Andrzej S. Nowak, my academic
adviser, for his kindness, guidance, and encouragement throughout this study. Very
special thanks are due to Dr. Maria M. Szerszen, my co-adviser, for her unconditional
support, whose valuable comments during this research have made an appreciable
difference.
I would like to thank Professor Raymond Moore, Professor Eddy Rojas, and
Professor Laurence Rilett, members of my doctoral committee, for their reviews and
valuable suggestions.
Very special thanks are due to Mr. Stephen H. Nickel for his unconditioned support
and help during the final preparations of this study.
I would like to thank Dr. Anna Rakoczy, Mr Hooman Ghasemi, and Mr.
Mohammed Al Bardan, my colleagues at the University of Nebraska-Lincoln, for their
friendship, help, and cooperation during the period of this study. Very special thanks are
due to a very special person, my best friend, Ann Whipple and her family for their
friendship, kindness, and support during living in Nebraska.
The Civil Engineering Department at University of Nebraska-Lincoln is
acknowledged for the study, guidance, and support during this study. Also
acknowledgements are due to the Ministry of Higher Education in Iraq and the Iraqi
Cultural Office in Washington DC for their financial support, kindness, and guidance
through this study.

iii
Especially, I wish to express my sincere thanks and love to my kindly mother
whose encouragement and valuable love made me complete this study, whom I wish she
was alive for this moment. I wish to offer my deepest gratitude and love to my husband
Mohammed whose patient love and greatest support enabled me to complete this work.
Hugs and kisses are to the leaves of my heart, my daughter Zainab, my son Ahmed, and
my baby boy Alhussain, whom was born during this study and was the blessing for it.
Last but not least, very special thanks and love are due to my father Dr. Yassen and my
two brothers Rafed and Mustafa for their encouragement and belief in me.
iv
CONTENTS
LIST OF FIGURES ........................................................................................ix

LIST OF TABLES ...................................................................................... xvii
Chapter 1. INTRODUCTION ..................................................................... 1
1.1 Problem Statement.................................................................................. 2
1.2 Research Objective ................................................................................ 4
1.3 Technical Background ........................................................................... 5
1.4 Dissertation Layout ............................................................................... 7
Chapter 2. TYPES OF SHALLOW FOUNDATONS ................................. 9

2.1 Introduction ............................................................................................ 9
2.2 Types of Shallow Foundations ........................................................... 11
2.2.1 Spread Footing Foundations ........................................................... 11
2.2.1.1 Square Spread Footings ............................................................... 11
2.2.1.2 Rectangular Spread Footings ....................................................... 12
2.2.1.3 Continuous Spread Footings ........................................................ 13
2.2.1.4 Combined Footings ...................................................................... 14
2.2.2 Mat Foundations ............................................................................. 16
Chapter 3. STRUCTURAL RELIABILITY THEORY .............................. 18

3.1 Introduction ......................................................................................... 18
3.2 Random Variables ............................................................................... 19
3.2.1 Probability Distributions of Random Variables .............................. 20
3.2.1.1 Discrete Random Variable .......................................................... 20
3.2.1.2 Continuous Random Variables .................................................... 22
3.2.2 Parameters of Random Variables ................................................... 25
v
3.2.2.1 Mean Value .................................................................................. 26

3.2.2.2 Variance ........................................................................................ 26
3.2.2.3 Standard Deviation ....................................................................... 27
3.2.2.4 Coefficient of Variation ............................................................... 27
3.3 Common Continuous Probability Distributions ................................. 28
3.3.1 Normal Distribution ........................................................................ 29
3.3.2 Lognormal Distribution ................................................................... 32
3.4 Normal Probability Paper .................................................................. 34
3.5 Limit State Functions .......................................................................... 38
3.6 Probability of Failure and Reliability Index ....................................... 40
3.7 Reliability Index Evaluation .............................................................. 44
3.7.1 The First-Order Second-Moment Reliability Method (FOSM) ..... 45
3.7.2 Monte Carlo Simulation Techniques .............................................. 48
CHAPTER 4. EVALUATION OF SOIL PARAMETERS FOR

SHALLOW FOUNDATONS IN COHESIONLESS SOILS ....................... 51
4.1 Introduction ......................................................................................... 51
4.2 Evaluation of Soil Parameters ............................................................. 51
4.2.1 Unit Weight ..................................................................................... 52
4.2.2 Soil Properties .................................................................................. 54
4.2.2.1 Standard Penetration Test for Field Procedures ........................... 55
4.2.2.2 Case History of Automatic Hammer Efficiency ......................... 63
4.2.3 Overconsolidation Ratio ................................................................... 66
4.2.4 Elastic Soil Parameters ..................................................................... 67
4.2.4.1 Poisson’s Ratio ............................................................................. 67

4.2.4.2 Soil Modulus of Elasticity ............................................................ 69
vi
4.3 Description and Background Information of the Highway Bridges

Selected in this Study ........................................................................ 73
4.3.1 Highway Bridges in Northeast United States .................................... 73
4.3.2 Highway Bridges Selected in Ohio .................................................. 76
CHAPTER 5. SETTLEMENT ANALYSIS FOR SHALLOW

FOUNDATONS IN COHESIONLESS SOILS ............................................ 79
5.1 Introduction ......................................................................................... 79
5.2 Types of Settlement ............................................................................ 81
5.2.1 Total Settlement ................................................................................ 82
5.2.2 Differential Settlement ..................................................................... 83
5.3 Evaluation of Settlement in the Context of Bridge Design ................ 84
5.4 Load – Settlement Behavior ............................................................... 90
5.5 Settlement Prediction Methods Selected in this Study ....................... 93
5.5.1 Method of Hough (1959) ................................................................. 95
5.5.2 Method of Meyerhof (1965) ............................................................ 96
5.5.3 Method of Peck and Bazaraa (1969) ............................................... 97
5.5.4 Method of D’ Appolonia (1970) ...................................................... 98
5.5.5 Method of Method of Schmertmann (1978) ................................. 100
5.5.6 Method of Burland and Burbidge (1985) ...................................... 104
5.6 Data Base of Measured Settlements Compared With Predicted
Settlements ........................................................................................ 106
5.7 Distribution of Settlement ................................................................ 119
5.7.1 Distribution of Measured and Predicted Settlements .................... 119
5.7.2 Distribution of Measured to Predicted Settlement Ratios ............. 124
5.8 Influence of Various Factors on Settlement ..................................... 129
5.8.1 Influence of Footing Size (B) ........................................................ 129
vii
5.8.2 Influence of Length to Width Ratio (L/B) ..................................... 131
5.8.3 Influence of Footing Depth Ratio (Df/B) ....................................... 132
5.8.4 Influence of Depth of Water Table (Dw/B) ................................... 134
5.8.5 Influence of Axial Load (P) ........................................................... 135
5.8.6 Influence of SPT- N Values .......................................................... 136
5.9 Estimation of Differential Settlement and Angular Distortion of

Shallow Foundations on Cohesionless Soils ................................... 138
5.9.1 Differential Settlement ................................................................... 138
5.9.2 Angular Distortion ......................................................................... 147
5.10 Tolerable Settlement Criteria ......................................................... 151
CHAPTER 6. RELIABILITY ANALYSIS OF SETTLEMENT FOR

SHALLOW FOUNDATONS IN COHESIONLESS SOILS ..................... 152
6.1 Introduction ....................................................................................... 152
6.2 Reliability Analysis Procedure ......................................................... 152
6.3 Limit State Function ......................................................................... 153
6.4 Nominal Resistance and Load Model .............................................. 154
6.5 Reliability Index Calculations .......................................................... 154
6.5.1 Reliability Indices for Total Settlement ........................................ 156
6.5.2 Reliability Indices for Differential Settlement .............................. 169
6.6 Target Reliability Levels .................................................................. 176
6.6.1 Target Reliability Level for Total Settlement ............................... 176
6.6.2 Target Reliability Level for Differential Settlement .................... 178
CHAPTER 7. CONCLUSIONS AND RECOMMENDATIONS ........... 180

7.1 Summary ................................................................................................................... 180
viii
7.2 Conclusions ............................................................................................................ 182
7.3 Recommendations for Future Work ............................................................. 191
7.4 Implementations of this Study ........................................................................ 192
REFERENCES ............................................................................................ 193

APPENDIX A .............................................................................................. 200
APPENDIX B .............................................................................................. 206
APPENDIX C .............................................................................................. 224
APPENDIX D .............................................................................................. 236
ix
LIST OF FIGURES
Figure 2.1: Typical Spread Footing with ………...………………………….…………..10

Figure 2.2: Square Spread ………..………………….……………………..…………....12
Figure 2.3: Rectangular Spread ..……………………………..…….…………………....13
Figure 2.4: Continuous Spread ….……………………………………………………....14
Figure 2.5: Combined ….………………………………………………………………..15
Figure 2.6: Spill-through Abutments on Combination Strip Footing (FHWA, 2002c) ....16
Figure 2.7: A Mat Foundation ………………………………………………………….17
Figure 3.1: Schematic Representation of a Random Variable ……………………….….19
Figure 3.2: A Probability Mass Function for a Discrete Random Variable ………….…21
Figure 3.3: A Cumulative Distribution Function for a Discrete Random Variable .…… 22
Figure 3.4: Probability Density Function (PDF) for a Continuous Random Variable ….24
Figure 3.5: Cumulative Distribution Function (CDF) for a Continuous Random
Variable…………………………………………………………………………………..26
Figure 3.6: PDF’s of Normal Distribution with Different Distribution Parameters ….…29
Figure 3.7: CDF’s of Normal Distribution with Different Distribution Parameters ..…...30
Figure 3.8: PDF of the Standard Normal Distribution ………..………………………...31
Figure 3.9: Probability Density Function of Lognormal Distribution ………….…….…33
Figure 3.10: The S-shaped CDF for a Normal Random Variable …………….…….…..35
Figure 3.11: Normal Distribution Function on the Normal Probability Paper ……….....37
Figure 3.12: PDF’s of Resistance, Load, and Safety Margin ………………….…….… 39
Figure 3.13: Reliability Index defined as the Shortest Distance in the Space of Reduced
Variables ………………………………………………………………………………...41
Figure 3.14: Simulated Values of the Limit State Function Generated by Monte Carlo
Method ..………………………………………………………………………………....50
Figure 4.1: Sequences of driving during the Standard Penetration Test (FHWA 2006,
Volume I) ……………………………………………………………………………..…56
Figure 4.2: SPT Hammer Types: (a) Donut, (b) Safety, (c) Automatic (FHWA 2006,
Volume I) ………………..………………………………………………….……..….…58
Figure 4.3: Automatic Hammer Data in Hancock, IA ……………………………..……64
x
Figure 4.4: CDF of SPT N-Values for Logs B-1 and B-2 in Hancock, IA..…………......65
Figure 4.5: Strength Measured by in-situ tests at Peak of Stress-Strain Curve .……..…69
Figure 5.1: Total Settlement in a Spread Footing Foundation ………………….……….82
Figure 5.2: Superstructure and Substructure Parts of the Bridge …………….…..……...85
Figure 5.3: Components of Settlement and Angular Distortion in Bridges (Adapted from
Duncan and Tan 1991) ..………………………………….………….………………….86
Figure 5.4: Schematic Concept of both Total and Differential Settlement, and Angular
Distortion in Bridges, (Adapted from FHWA 2010) ………………….………………...88
Figure 5.5: Typical Axial Load-Displacement Curves, (Adapted from Hirany and
Kulhawy 1988) ………………………………………………………………………….90
Figure 5.6-a: Identification of Critical Construction Points ……….…………….……...91
Figure 5.6-b: Conceptual Load-Displacement Pattern for a given footing width ……....92
Figure 5.7: Bearing Capacity Index Versus Corrected SPT (Adapted from Cheney
Chassie, 2000 after Hough 1959) ………………………………………….……………95
Figure 5.8: Correlation between Modulus of Compressibility and SPT Blow Counts
(Adapted by D’Appolonia 1970) .…..………………….………………………………..98
Figure 5.9: Values of Embedment Correction Factor used for D’Appolonia Method .…99
Figure 5.10: Value of Thickness Correction Factor used for D’Appolonia Method …...99
Figure 5.11: Variation of Iz with Depth (Adapted by Schmertmann, 1978) ……….….102
Figure 5.12: Measured Settlements for Study Bridge Elements ……………….……....106
Figure 5.13: Comparison of Measured Settlements with Predicted Settlements for all Six
Methods based on Field Data……………………………………………..…….……....109
Figure 5.14-a: Predicted Versus Measured Settlements for Hough Method (1959) .......113
Figure 5.14-b: Distribution of Bias Factor for Settlement for Hough Method (1959) ...113
Figure 5.15-a: Predicted Versus Measured Settlements for Meyerhof Method (1965)...114
Figure 5.15-b: Distribution of Bias Factor for Settlement for Meyerhof Method (1965)
………………………………………………………………………………………….114
Figure 5.16-a: Predicted Versus Measured Settlements for Peck and Bazarra Method..115
Figure 5.16-b: Distribution of Bias Factor for Settlement for Peck and Bazarra Method
(1969) …………………………………………………………………………..……....115
xi
Figure 5.17-a: Predicted Versus Measured Settlements for D’Appolonia Method (1970)
…………………………………………………………………………………………..116
Figure 5.17-b: Distribution of Bias Factor for Settlement for D’Appolonia Method (1970)
…………………………………………………………………………………………..116
Figure 5.18-a: Predicted Versus Measured Settlements for Schmertmann Method (1978)
…………………………………………………………………………………………..117
Figure 5.18-b: Distribution of Bias Factor for Settlement for Schmertman Method (1978)
……………………………………………………………………………….…….........117
Figure 5.19-a: Predicted Versus Measured Settlements for Burland and Burbidge Method
…………………………………………………………………………..………..……..118
Figure 5.19-b: Distribution of Bias Factor for Settlement for Burland and Burbidge
Method (1985) ..……………………………………………………………………..…118
Figure 5.20: CDF of Measured Settlements for 23 Sites ………………………..……..120
Figure 5.21: CDF of Predicted Settlements by Hough Method for 23 sites …………...121
Figure 5.22: CDF of Predicted Settlements by Meyerhof Method for 23 sites ………..121
Figure 5.23: CDF of Predicted Settlements by Peck and Bazaraa Method for 23 sites..122
Figure 5.24: CDF of Predicted Settlements by D’Appolonia Method for 23 sites……..122
Figure 5.25: CDF of Predicted Settlements by Schmertmann Method for 23 sites….…123
Figure 5.26: CDF of Predicted Settlements by Burland and Burbidge Method for 23 sites
…………………………………………………………………………………………..123
Figure 5.27: CDF of (δM/δP) for 23 sites by Hough Method (1959) ……………….....125
Figure 5.28: CDF of (δM/δP) for 23 sites by Meyerhof Method (1965) ……….……...126
Figure 5.30: CDF of (δM/δP) for 23 sites by D’Appolonia Method (1970) .…………..126
Figure 5.31: CDF of (δM/δP) for 23 sites by Schmertmann Method (1978)….…….…127
Figure 5.32: CDF of (δM/δP) for 23 sites by Burland and Burbidge Method (1985) ....127
Figure 5.33: CDF of (δM/δP) for 23 sites by all Six Prediction Methods .…………….128
Figure 5.34: Influence of Footing width (B) on Settlement on Settlement of Shallow
Foundations on Cohesionless Soils …………………………..………………………...129
Figure 5.35: Effect of Increasing Footing width (B) on Settlement …………….......…130
Figure 5.36: Influence of Square Root of Footing width (B) on Settlement on Settlement
……………………………………………………………………………………….….130
xii
Figure 5.37: Influence of Length to Width Ratio (L/B) on Settlement of Shallow

Foundations on Cohesionless Soils ..……………………………………..…………....131
Figure 5.38: Effect of Increasing Length to Width Ratio (L/B) on Settlement ………..132
Figure 5.39: Influence of Depth Ratio (Df /B) on Settlement of Shallow Foundations on
Cohesionless Soils …………………………………………………………………..…133
Figure 5.40: Effect of Increasing Depth Ratio (Df/B) on Settlement ……………….....133
Figure 5.41: Influence of Water Table Depth Ratio (Dw /B) on Settlement of Shallow
Foundations on Cohesionless Soils ………………………………………………….…134
Figure 5.42: Effect of Increasing the Water Table Depth Ratio (Dw/B) on Settlement
…………………………………………………………………………………………..135
Figure 5.43: Influence of Applied Loads (P) on Settlement of Shallow Foundations on
Cohesionless Soils ……………………………………………………………………..135
Figure 5.44: Effect of Increasing the Axial Load (P) on Settlement for Two Different
Sites …………………………………………………………………………….………136
Figure 5.45: Influence of Corrected SPT-N Values on Settlement of Shallow Foundations
on Cohesionless Soils.…..…………………………………………………….…….….137
Figure 5.46: Effect of Increasing the SPT-N Values (N60) on Settlement ……………137
Figure 5.47: Distribution of Actual Differential Settlement for 20 sites of Spread Footings
………..………………………………………………………………………..………..142
Figure 5.48: CDF of Actual Differential Settlement for 20 sites of Spread Footings.…142
Figure 5.49: Distribution of Differential Settlement by Hough Method (1959)………..143
Figure 5.50: CDF of Differential Settlements by Hough Method for 20 sites…...…….143
Figure 5.51: Distribution of Differential Settlement by Schmertman Method (1978)…144
Figure 5.52: CDF of Differential Settlements by Schmertmann Method for 20 sites….144
Figure 5.53: Distribution of Bias Factor for Differential Settlement by Hough Method
(1959) ….……………………………………………………………………………….147
Figure 5.54: CDF of Bias Factor for Differential Settlement for 20 sites Hough Method
(1959) ………………………………………………………………………………….147
Figure 5.55: Distribution of Bias Factor for Differential by Schmertmann Method (1978)
…………………………………………………………………………………………..148
xiii
Figure 5.56: CDF of Bias Factor for Differential Settlement for 20 sites by Schmertmann
Method (1978)..………………………………………………..……………………… 148
Figure 6.1: Reliability Indices for Total Settlement by all Methods for S1-Site ..….….156
Figure 6.2: Reliability Indices for Total Settlement by all Methods for S2-Site ..….….157
Figure 6.3: Reliability Indices for Total Settlement by all Methods for S3-Site .……...157
Figure 6.4: Reliability Indices for Total Settlement by all Methods for S4-Site ………158
Figure 6.10: Reliability Indices for Total Settlement by all Methods for S10-Site ..…..161
Figure 6.11: Reliability Indices for Total Settlement by all Methods for S11-Site...…..161
Figure 6.12: Reliability Indices for Total Settlement by all Methods for S14-Site ……162
Figure 6.16: Reliability Indices for Total Settlement by all Methods for S20-Site ........164
Figure 6.17: Reliability Indices for Total Settlement by all Methods for S21-Site .…...164
Figure 6.24: Reliability Indices for Differential Settlement for S1, S2-Sites in Bridge 1
…………………………………………………………………………………………..170
…………………………………………………………………………………………..170
…………………………………………………………………………………………..171
xiv
Figure 6.27: Reliability Indices for Differential Settlement for S6, S8,9-Sites in Bridge 3
…......................................................................................................................................171
Figure 6.28: Reliability Indices for Differential Settlement for S8, 9 and S10,11-Sites in
Bridge 3 ..………………………………………………………………………………172
…………………………………………………………………………………………..172
…………………………………………………………………………………….…….173
…………………………………………………………………………………………..173
………………………………………………………………………………………….174
Figure 6.33: Reliability Indices for Differential Settlement for S26,S27-Sites in Bridge 12
………………………………………………………………………………………….174
Figure A-1: Logs of Boring B-1 in Hancock, IA .………………………………….......200
Figure A-2: Logs of Boring B-2 in Hancock, IA .………………………………….......203
Figure C.1: Bridge Structure After Placement of Beams (Sargand and Masada 2006,
FHWA) ..……………………………………………………………………………….224
Figure C.2: SPT N-Values variations with Depth for Project FRA-670-0380 .………..226
Figure C.3: Corrected SPT N-Values Variations with Depth for Project FRA-670-0380
………………………………………………………………………………………….227
Figure C.4: Old and New Interchange Designs (MOT-70/75), (Sargand and Masada 2006,
FHWA) ………………………………………………………………………………...229
Figure C.5: Ramp C Bridge Project Site (MOT I70/I75) -General View ( Sargand and
Masada 2006, FHWA) …..……………………………………………………………..230
Figure C.6: SPT- N Values Variations with Depth for C188 Bore Holes (Pier 18) …...232
Figure C.7: SPT- N Values Variations with Depth for C187 Bore Holes (Pier 19) …...232
Figure C.8: Corrected SPT- N Values Variations with Depth for Pier 18 ……………..233
Figure C.9: Corrected SPT- N Values Variations with Depth for Pier 19 ……………..233
Figure D.1: CDF of Footing Width, B for S1-Site by Meyerhof Method ……………..236
Figure D.2: CDF of Footing Length, L for S1-Site by Meyerhof Method …………….236
xv
Figure D.3: CDF of Footing Depth, D for S1-Site by Meyerhof Method ……………..237
Figure D.4: CDF of Load Applied, P for S1-Site by Meyerhof Method …………........237
Figure D.5: CDF of SPT N-Values for S1-Site by Meyerhof Method ………………...238
Figure D.6: CDF of Bearing Pressure, q for S1-Site by Meyerhof Method……………238
Figure D.7: CDF of Depth Correction factor, Fd for S1-Site by Meyerhof Method …..239
Figure D.8: CDF of Total Settlement for S1-Site by Meyerhof Method …………........239
Figure D.9: CDF of Footing Width, B for S1-Site by Peck and Bazaraa Method …….240
Figure D.10: CDF of Footing Length, L for S1-Site by Peck and Bazaraa Method …..240
Figure D.11: CDF of Footing Depth, D for S1-Site by Peck and Bazaraa Method ……241
Figure D.12: CDF of Applied Load, P for S1-Site by Peck and Bazaraa Method …….241
Figure D.13: CDF of Unit weight of Soil for S1-Site by Peck and Bazaraa Method .…242
Figure D.14: CDF of SPT N-Values, (N1)60 for S1-Site by Peck and Bazaraa Method
…………………………………………………………………………………………..242
Figure D.15: CDF of Bearing Pressure, q for S1-Site by Peck and Bazaraa Method
…………………………………………………………………………………………..243
Figure D.16: CDF of Water Table Correction Factor, Cw for S1-Site by Peck and
Bazaraa Method ..…………………………………………………………………........243
Figure D.17: CDF of Depth Embedment Correction Factor, CD for S1-Site by Peck and
Bazaraa Method ..……………………………………………………….……………...244
Figure D.18: CDF of Total Settlement for S1-Site by Peck and Bazaraa Method .........244
Figure D.19: CDF of Footing Width, B for S1-Site by Schmertmann Method ………..245
Figure D.20: CDF of Footing Length, L for S1-Site by Schmertmann Method ……….245
Figure D.21: CDF of Footing Depth, D for S1-Site by Schmertmann Method ………..246
Figure D.22: CDF of Unit Weight of Soil, Gamma for S1-Site by Schmertmann Method
…………………………………………………………………………………………..246
Figure D.23: CDF of Applied Load, P for S1-Site by Schmertmann Method …………247
Figure D.24: CDF of SPT N-Values for S1-Site by Schmertmann Method …………...247
Figure D.25: CDF of Bearing Pressure, q for S1-Site by Schmertmann Method ……...248
Figure D.26: CDF of Modulus of Elasticity, Es for S1-Site by Schmertmann Method
…………………………………………………………………………………………..248
xvi
Figure D.27: CDF of Effective Vertical Stress for S1-Site by Schmertmann Method
…………………………………………………………………………………………..249
Figure D.28: CDF of Shape Correction Factor, C3 for S1-Site by Schmertmann
Method………………………………………………………………………………….249
Figure D.29: CDF of Depth Correction Factor, C1 for S1-Site by Schmertmann
Method………………………………………………………………………………….250
Figure D.30: CDF (q-sigma-dz) for S1-Site by Schmertmann Method.…………….....250
Figure D.31: CDF of Total Settlement for S1-Site by Schmertmann Method ………...251
xvii
LIST OF TABLES
Table 3.1: Examples of Relationship between the Standard Normal Variable and the
Corresponding Probability ………………………………………………………………36
Table 3.2: Reliability Index β and Probability of Failure Pf (Adapted from Nowak and
Collins, 2013) ……………………………………………………………………………43
Table 4.1: Typical Unit Weights (Adapted from Coduto 2000) ………………………...53
Table 4.2: In-Situ Dry Unit Weight Statistics for Various Cohesionless Soils (Adapted
from Chen 2004)…………………………………………………………………………54
Table 4.3: SPT Hammer Efficiencies (Adapted from Clayton 1990) .………………......60
Table (4.5): Soil Properties Correlated with Standard Penetration Test Values (Adapted
from Peck et al. 1974)……………………………………………………………………61
Table 4.6: Statistical Parameters for SPT N-Values for Automatic Hammer for Logs B-1
and B-2 of Soil Borings in Hancock, IA .……………………………………………….66
Table 4.7: Typical Ranges of Poisson’s Ratio for Various Types of Soils (Adapted from
AASHTO 2012) …………………………………………………………………………68
Table 4.8: Estimations of Soil Modulus of Elasticity for Various Types of Soils (Adapted
from Bowels 1988) …………………………………………………………...................70
Table 4.9: Typical Ranges of Young’s Modulus of Elasticity for Various Types of Soils
(Adapted from AASHTO, 2012) ………………………………………………………..71
Table 4.10: Equivalent Elastic Modulus for Various Types of Soils Based on SPT- N
Values (Adapted from AASHTO, 2012) ………………………………………………..72
Table 4.11: Typical Ranges of Drained Modulus for Sandy Soils (Adapted from
Kulhawy and Mayne 1990) .……………………………………………………………..72
Table 4.12: Classification and Location of Bridges Data Field used in this Study ……..74
Table 4.13: Classification and Location of Bridges in Ohio used in this Study ….........77
Table 5.1: Tolerable Movement Criteria of Angular Distortion for Highway Bridges
(Adapted from AASHTO 2012) …………………………………………………….......88
Table 5.2: Correlation Factors (Adapted from Coduto, 2000) …………………………101
Table 5.3: Comparison of Measured and Predicted Values of Settlement .…….……...107

xviii
Table 5.4: Bias Factor for Settlement (δM/δP) of the Field Data ……………………..110
Table 5.5: Statistical Parameters of Bias Factor for Settlement (δM /δP) of theSix
Prediction Methods ..…………………………………………………………………...111
Table 5.6: Comparison of Actual and Calculated Values of Differential Settlement ….141
Table 5.7: Statistical Parameters of Differential Settlement for the Six Prediction
Methods ………………………………………………………………………………..142
Table 5.8: Bias Factor for Differential Settlement (δD/δDP) of the Field Data ...........145
Table 5.9: Statistical Parameters of Bias Factor for Differential Settlement (δD/δDP) of
the Two Selected Methods …………………………………………………..................146
Table 5.10: Comparison of Actual and Calculated Values of Angular Distortion …….149
Table 5.11: Statistical Parameters of Angular Distortion for the Six Prediction Methods
..........................................................................................................................................150
Table 6.1: Average of Reliability Indices for Total Settlement, Related to Tolerable
Settlements and their Probability of Failures for 23 Sites ……….…………………….168
Table 6.2: Average of Reliability Indices for Differential Settlement, Related to

Tolerable Settlements and their Probability of Failures ……...……..…………………175
Table 6.3: Target Reliability Levels for Total Settlement, Related to Tolerable
Settlements…….………………………………………………………………………..177
Table 6.4: Target Reliability Levels for Differential Settlement, Related to Tolerable
Settlements ……………………………………………………………………………..179
Table B.1: Data Established for the Ten Highway Bridges Adapted from (FHWA, 1987
and 2010) .……………………………………………………………………….……..222
Table C.1: Soil Boring Logs for Bore Hole RB-9 ……………………………….…….225
Table C.2: Soil Boring Logs for Bore Hole RB-11 …………………………….……...225
Table C.3: Average of SPT N-Values of the Two Soil Boring Logs ………………….226
Table C.4: Soil Boring Logs C188 (Located Pier 18) …………………….…………...231
Table C.5: Soil Boring Logs C187 (Located Pier 19) ……………………………........231
Table C.6: Data Established for the Two Highway Bridges Adapted from (Sargand and
Masada 2006, FHWA) …………………………………………………………………235
1
CHAPTER 1
INTRODUCTION
Soil as an engineering material differs greatly from any other structural materials
such as steel or concrete because soils cannot be ordered to suit the designer’s needs.
Therefore, in most circumstances the foundation is designed to match the specific types of
soils and loads.
A foundation design specialist has many duties; with one of the most important,
of establishing an economical design that satisfies the prescribed structural criteria safely
and properly accounts for the intended function of the structure.
One of the most important problems of soil mechanics is stability related to the
condition of excessive settlements. The settlement of structures supported on soils is a
subject of considerable interest to practicing engineers, since excessive settlements can
lead to serviceability problems or even failure states in the structural foundations. The
design of shallow foundations is most often governed by settlement requirements. This
means a reliability-based approach can be applied for designing shallow foundations,
providing design data about the distribution of settlement under a given footing. Since the
development of new generation of design codes, reliability has been accepted as a rational
measure of structural performance as the performance of geotechnical structures including

2
shallow foundations where applied loads, soil, properties, and resistance parameters are
subjected to random variation.
The selection of a foundation system for highway bridges involves the
consideration of performance and cost. For an adequate performance, the foundation must
satisfy the bearing capacity requirements to support the piers, abutments, and
superstructure, and must satisfy the settlement limits. Only few states such as New York,
Ohio, and District of Columbia, Shad et al. (1999) used shallow foundations for highway
bridges instead of piles. However, currently, shallow foundations are becoming an
attractive alternative where the subsurface soils are suitable, since they require less time
to construct and cost less compared to deep foundations.
However, most of the current research efforts have been focused on the reliability
models for the superstructure components with less attention paid to the substructure and
foundations. Meanwhile, the random variables and limit state functions for shallow
foundations are very different from those for superstructures such as buildings and
bridges.
1.1 PROBLEM STATEMENT
Deep foundations are generally considered to be more preferable than shallow
foundations for bridges due to the lack of tolerable vertical movement guidelines for
spread footings. It has been a common belief of spread footings on soil settle much more
than deep foundations and bring higher maintenance costs and uncertainties of soil
properties and of the accuracy of settlement prediction methods. Most bridge spread
3
footings on soils are supported in cohesionless soils, which does not exhibit secondary
compression and which consolidates quickly. This study will concentrate on footings
supported in cohesionless soils.
The development of a new generation of reliability- based design code for shallow
foundations requires an effort to harmonize with structural codes. For this reason,
geotechnical design codes around the world are beginning to move towards some form of
reliability-based design (RBD). This requires definition of critical failure states in
conjunction with load and resistance factors that are calibrated to achieve the target
reliabilities associated with the various limit states. These load and resistance factors
would then be incorporated into Load and Resistance Factor Design (LRFD).
Both AASHTO and FHWA had agreed that all state DOT’s would use LRFD for
design of structures as of October 2007. Since AASHTO-LRFD is calibrated only for the
strength limit state at this time, the load and resistance factors for service limit state are
still not calibrated. The goal of this research is to develop the methodology to start the
process. The calculation of reliability indices for settlement of shallow foundations in
bridges can be considered as a service limit state (SLS).
Limit states are boundaries between safety and failure and for shallow foundations,
limit states can be grouped into substructural limit states for foundation movement
(vertical settlement) which need to be estimated and compared with tolerable movement
criteria. LRFD specifications must be considered for service limit state of substructures,
especially shallow foundations.
For each structural and soil condition components, the reliability analysis requires the
following:
4
 Formulation of the limit state function.
 Determination of the statistical parameters for load and resistance that represent the settlement
needs in cohesionless soils deposits.
 Development of the reliability analysis procedure for the most commonly used settlement
prediction methods for shallow foundations.
1.2 RESEARCH OBJECTIVE
The main objective of this research is to extend the reliability-based design
methodology of shallow foundations on cohesionless soils in bridges. This objective can be
implemented by a critical evaluation of current prediction methods to estimate the settlement
of shallow bridge foundations on cohesionless soils under a sensitivity analysis.
Reliability, as it can be expressed as a measure of structural performance, will be
expressed in terms of reliability indices, which in this research will be calculated for
settlement of shallow bridge foundations.
The study will address three major tasks, which are stated as:
 Development of a critical review of the available settlement prediction methods for
shallow foundations on cohesionless soils, and selection of the most efficient
procedure or procedures.
 Development of an efficient reliability analysis procedure for settlement limit states of
shallow foundations.
 Development of selection criteria for the target reliability levels.

5
1.3 TECHNICAL BACKGROUND
The performance of geotechnical structures, including shallow foundations depends
on loads, soil properties, and resistance parameters that are subjected to random variables.
These variables are associated with previous studies by Filippas et al. (1988), Fenton and
Griffiths (1996, 1999, 2002, 2003, and 2005), Phoon et al. (1995 and 1999), Fenton et al.
(2003), Zekkos et al. (2004), Phoon (2006), Akbas (2007), Akbas and Kulhawy (2005 and
2009).
Significant steps in the development of a new generation of reliability- based
design codes for shallow foundations can be found in the Canadian Foundation Engineering
Manual(1985, 1992, and 2006), Euro code 7 (1993), Australian Standard AS 2159 (1995),
Australian Standard AS 4678 (2002), Australian Standard AS 5100.3 (2004), and NCHRP
Report 507 (2004). These RBD provisions are most presented in the form of a limit state
design (LSD), while failure states can be defined as various service limit states (SLS) with the
use of load and resistance factors referred to as Load and Resistance Factor Design (LRFD).
In order to control serviceability problems arising from excessive settlement of
shallow foundations, geotechnical design codes generally include specifications regarding
maximum settlement, which often govern the foundation design, However, once the
foundation has been designed and constructed, the actual settlement experiences on a real
three-dimensional soil mass that can be quite different than expected, due to soil’s spatial
variability. Therefore, many methods have been developed for the assessment of tolerable and
predicted settlement for shallow foundations on cohesionless soils. These methods are
6
presented in many textbooks, but they have unknown reliability. The predictions of settlement
are based on different in situ tests: standard penetration test (SPT), cone penetration test
(CPT), dilatometer test (DMT), or pressuremeter test (PMT) for cohesionless soils.
The most popular methods used for predicting immediate settlements for
cohesionless soils are those of Terzaghi and Peck (1948 and 1967), Gibbs and Holtz (1957),
Hough(1959), Menard and Rousseau (1962), Alpan (1964), Meyerhof (1965 and 1982), Peck
and Bazaraa (1969), D’Appolonia et al. (1970), Benjamin and Cornell (1970),Parry (1971),
Schultze and Sherif (1973), Schmertmann et al. (1978), Burland and Burbidge (1985), Bowles
(1988), Anagnostopoulos (1991), Berardi and Lancellotta (1991) , Coduto (1994), Samuel et
al. (1994), John Christian (1994), Terzaghi et al. (1996), Mayne and Poulos (1999), Salgado
(2001), Sargand et al. (2003), Salgado et al. (2004), Foye et al(2006), Braja and Nagaratnam
(2007), and Rasin (2009).
By referring to several case studies, comparisons between predicted and measured
settlements of spread footings on cohesionless soils were presented by Keene (1978),
Gifford et al. (1978), Jeyapalan and Boehm (1986), Tan and Duncan (1991), Baus (1992),
Papadopoulos (1992), Briaud and Gibbens (1997), Sivakugan et al. (1998), and Sargand et al.
(1999). Also, the probabilistic approach proposed by Sivakugan and Johnson (2004) had an
effective way of quantifying the risk associated with the settlement prediction methods.
Lee and Salgado (2002) worked on estimation of footing settlement on sands using
the finite element method while Anderson et al. (2007) examined different insitu test methods
for predicting settlement of shallow foundations and compared them with the finite element
7
analysis. Finally, Das and Sivakugan (2007) compared traditional settlement prediction
methods with recently developed ones.
In recent studies in the literature for the reliability analysis, the target probability of
failure (Pf)T for service limit state (SLS) of footings varies considerably, Zekkos et al. (2004)
worked on reliability analysis of shallow foundations and got values of (Pf)T as high as 30%,
Fenton et al. (2005) evaluated (Pf)T as 5 %. Finally, Akbas (2007) used a probabilistic
approach and found (Pf)T to be about 1.1%. Related to all of these studies, the probability of
failure is still high.
1.3 DISSERTATION LAYOUT
Chapter 1 of this dissertation serves as an introduction to this study, which
includes the problem statement, research objective, and finally the technical background that
presents a comprehensive literature review on several approaches for predicting settlement of
shallow foundations on cohesionless soils. In addition, it shows the target probability of
failure for service limit state (SLS) of footings from previous studies.
In Chapter 2, types of shallow foundations and basic definitions of these types are
presented. Chapter 3 summarizes basic concepts and principles of the reliability theory.
Definitions of random variables, probability distributions, and limit sate functions are
introduced. Methods for calculating reliability index and simulation techniques are described.
Chapter 4 presents soil parameters used in the sensitivity analysis. In addition, the
chapter contains a description of the 12 highway bridges selected in this study. Chapter 5
includes definitions of total and differential settlements with angular distortion. This chapter
8
describes all the six prediction methods used for calculating total and differential settlements
for shallow foundations on cohesionless soils, and then it provides the statistical parameters
for settlement presented in tables as well as the distribution of settlement for each method. It
also includes the influence of various factors on settlement. Chapter 6 describes the reliability
analysis procedure for total and differential settlement, and the model considered in this study.
Finally, Chapter 7 presents the summary and conclusions of this research and the
recommendations for future work including implementations of this study.

9
CHAPTER 2
TYPES OF SHALLOW FOUNDATONS
2.1 INTRODUCTION
The process of foundation design for any structure should insure that the
structural loads are transmitted to the subsoil surface and the loads will not cause shear failure
or deformation settlements. Foundations can be categorized as “shallow” or “deep”,
depending on the depth at which the loads are transferred to the supporting soil. This chapter
will introduce several types of shallow foundations with their basic definitions.
Shallow foundations are the lowest part of a structure that transmits the applied
structural loads to the nearest subsurface soils; generally where the foundation depth (Df) is
less than either the width of the footing (B) and the length of the footing (L), and is less than
(3m) as shown in Figure 2.1. Shallow foundations are used when the surface soils are
sufficiently strong and stiff to support the imposed loads. They are generally unsuitable in
weak or highly compressible soils, such as poorly-compacted fill, peat, and alluvial deposits.
They are most often used in small- to medium- sized structures with moderate to good soil
conditions. Sometimes they can be used on some large structures when they are underlain by
an exceptionally good soil or bedrock.

10
Figure 2.1: Typical Spread Footing with Axial Loading
Shallow foundations are preferable to deep foundations where the subsurface soils
are suitable, since shallow foundations require less construction time, are of low cost, and
demonstrate ease of construction compared to deep foundations. Shallow foundations may be
built in different shapes and sizes to accommodate individual needs.

11
2.2 TYPES OF SHALLOW FOUNDATIONS
Shallow foundations encompass two types: spread footing foundations and mat foundations.
2.2.1 SPREAD FOOTING FOUNDATIONS
Spread footings are the most common type of foundation. They are used to support
individual loads, such as those for a column or a bearing wall. Spread footings are usually not
connected, which means a structure may include many individual footings.
There are different shapes and sizes of spread footings to accommodate the needs of
the design construction, and these include the following:
2.2.1.1 Square Spread Footings
They can simply be called square footings, with plan dimensions of BxB, with B
being the width of the footing , while the depth from the ground surface to the base of the
footing is Df and the thickness of the footing is T, as shown in Figure 2.2. Square footings
usually support a single central column and, for bridge columns, the footings are typically
greater than 10 x10 ft (3x3 m). The size of the footing is a function of the distributed loads
carried by the supported columns and the strength with the compressibility characteristics of
the materials beneath the footings.

12
Figure 2.2: Square Spread Footing
2.2.1.2 Rectangular Spread Footings

Rectangular footings have with plan dimensions of BxL, where B is the width and L
is the length of the footing, as shown in Figure 2.3. Rectangular footings are used when there
is consideration of large moment loads or there are construction requirements that prevent
using square footings due to width restrictions.

13
Figure 2.3: Rectangular Spread Footing
2.2.1.3 Continuous Spread Footings
Continuous spread footings are also be known as strip footings or wall footings. Strip
footings are considered when the ratio L/B ≥ 10, as shown in Figure 2.4.They are the most
commonly types of foundations used for buildings. Strip footings are used to support line
loads, either due to a load-bearing wall, or if a line of column positions are so close that
individual footings would be inappropriate.

14
Figure 2.4: Continuous Spread Footing
2.2.1.4 Combined Footings
Combined footings are similar to square or rectangular spread footings except that
they are used to support more than one column and are rectangular or trapezoidal in shape as
shown in Figure 2.5. They are useful when columns are located too close to each other to have
individual spread footings.

15
Figure 2.5: Combined Footing
An example of a combined footing used in bridge abutments, is called the “spill-
through” type abutment, as shown in Figure 2.6. This configuration was used during some of
the initial constructions of the Interstate Highway System on new alignments where spread
footings could be founded on competent native soils, FHWA (2006).

16
Figure 2.6: Spill-through Abutments on Combination Strip Footing (FHWA, 2002c)
2.2.2. MAT FOUNDATIONS
Mat foundations are the second type of shallow foundations. They are also known as
raft foundations and are shown in Figure 2.7. Raft foundations are used to spread the load
from a structure over a large area, normally the entire area of the structure. They are used
17
when column loads or other structural loads are close together and an individual spread
footing would interact or if the foundation is large, e.g., as with the support of a tank.
A mat foundation normally consists of a reinforced concrete slab, which extends over
the entire loaded area to support a number of loads from columns and wall. Usually mat
foundations have the advantage of reducing differential settlements as the concrete slab resists
differential movements between loading positions. They are often selected on soft or loose
soils with low bearing capacity, as they can spread the loads over a larger area than individual
spread footings.
Figure 2.7: A Mat Foundation

18
CHAPTER 3
STRUCTURAL RELIABILITY THEORY
3.1 INTRODUCTION
With respect structures, reliability should be considered to mean the probability, that a
structure will not reach each specified limit state either, the ultimate limit state (ULS) or
serviceability limit state (SLS) during a specified period of time. Each limit state or failure
mode should be treated separately and explicitly. However, most structures have a number of
possible failure modes, and this should be taken into account when determining the reliability
of the structure. Even though it is impossible to examine all failure modes for structures,
representative failure scenarios should be chosen so that the estimation of structural reliability
would be related to specified failure modes.
Civil engineers use a probabilistic evaluation for reliability in designing structures
that is intended to assure their desirable performance under applied loads during construction
and service. The concept of structural reliability encompasses real input observations and a
mathematical model of phenomena, which will be defined as the reliability index.
It is impossible to have an absolutely safe structure because each structure has a
nonzero probability of failure. Therefore, the convenient way for safety measurement of
structures would be in terms of reliability index rather than probability of failure.

19
3.2 RANDOM VARIABLES
A random variable is a variable whose value is nondeterministic. In mathematical
terms, a random variable is defined as a function that maps events onto intervals on the axis of
real numbers, Nowak and Collins (2013). This is schematically shown in Figure 3.1. A
random variable could be either a continuous random variable or a discrete random variable.
A discrete random variable may only take integer or discrete values, while a continuous
random variable may take a continuous range of values within an interval of real numbers.
Figure 3.1: Schematic Representation of a Random Variable

20
3.2.1 PROBABILITY DISTRIBUTIONS OF RANDOM VARIABLES
Each random variable has a distribution function that completes the description of its
probabilistic characters, such as the mean, variance, standard deviation, and coefficient of
variation, which will be defined later in this chapter. Different types of random variables have
probability functions and distributions, which can be described in more details in Thoft-
Christensen and Baker (1982), Ayyub (1997), and Nowak and Collins (2013).
3.2.1.1 Discrete Random Variable
A discrete random variable may only take on discrete value such as an integer value.
The probability of a discrete random variable is given by the probability mass function (PMF),
Px (xi ), which is defined as the probability of occurrence for each event. The PMF specifies
the probability that a discrete random variable X would be equal to a specific value of x, and
is denoted by:
Px (xi ) = P(X = xi ) (3.1)
The sum of all possible probabilities must be equal to one as:

i 1
Px (xi ) = 1 (3.2)
in which, N is the total number of outcomes.

21
For a hypothetical set of values, the PMF for a discrete random variable is shown in
Figure 3.2.
0.3
0.25
0.2
Probability
0.15
0.1
0.05
0
1 2 3 4 5 6 7 8 9 10
Events
Figure 3.2: A Probability Mass Function for a Discrete Random Variable
The cumulative distribution function (CDF), Fx (xi ), for a discrete random variable
is used to indicate that the probability of a random variable X is less than or equal to xi , as
given by the following:
Fx (xi ) = P(X ≤ xi ) (3.3)

22
An example of CDF for a discrete random variable is shown in Figure 3.3.
0.8
Probability
0.6
0.4
0.2
0
1 2 3 4 5 6 7 8 9 10
Events
Figure 3.3: A Cumulative Distribution Function for a Discrete Random Variable
3.2.1.2 Continuous Random Variable
A continuous random variable takes values within an interval of real numbers that
any possible event can be as an outcome. Each event for a continuous random variable is
different and has the same probability of occurrence.

23
For a continuous random variable, the probability of occurrence is defined as the
probability density function (PDF). Specifically, the probability for a continuous random
variable X that lies within the interval from x1 to x2 can be given by:
x2
P(x1 ≤ X ≤ x2 ) = Fx (x2 ) - Fx (x1 ) =  f ( x)dx

x1
x (3.4)
where fx (x) is the probability density function.
Since a continuous random variable follows the three axioms of probability, the
integral of the PDF from -∞ to +∞ would be equal to 1, i.e.,

P(-∞ ≤ X ≤ +∞ ) = f

x ( x)dx =1 (3.5)
The cumulative distribution function (CDF) for a continuous random variable
describes the probability distribution of a random variable X that will be found at a value less
than or equal to x. The CDF is a positive non decreasing function whose value is between 0
and 1, can be written in the form of:

Fx (x0 ) = P(X ≤ x0 ) = f
x0
x ( x)dx (3.6)
24
The probability density function (PDF) is defined only for continuous random
variables as the first derivative of the cumulative distribution function (CDF). The PDF is
denoted as fx (x), and the CDF is denoted as Fx (x ), and are related as follows:
d
fx (x) = Fx (x ) (3.7)
dx
x
Fx (x ) = f

x ( x)dx (3.8)
Both PDF and CDF functions are shown in Figures (3.4) and (3.5), respectively for a
continuous random variable X.

Probability
Random Variable X
Figure 3.4: Probability Density Function (PDF) for a Continuous Random Variable
Probability 25
Random Variable X
Figure 3.5: Cumulative Distribution Function (CDF) for a Continuous Random Variable
3.2.2 PARAMETERS OF RANDOM VARIABLES
Although the value of a random variable is uncertain, there are certain parameters that
can help mathematically to describe the properties of that variable. Starting with the mean
value, variance, standard deviation, and finally the coefficient of variation, are parameters that
can be used to describe the distributions of a random variable.

26
3.2.2.1 Mean Value
The mean (expected value) can be defined as the first moment measured about the
origin. It is also the average of all observations on a random variable X. The population mean
is indicated as µ.
For a discrete random variable, the mean µ, is given by:
µ=  x P (x )
i 1
i x i (3.10)
For a discrete random variable, if all n observations are given equal weights of
[ Px (xi) = 1/n ] , then the sample mean X would be the average of the observed values as
given by:
1 n
X =  xi
n i 1 (3.11)
For a continuous random variable, the mean is computed as:

µ=  xf

x ( x)dx (3.12)
3.2.2.2 Variance
The variance is the second moment about the mean. The variance of a population is
denoted by σ2, while the variance of a sample is denoted by S2.

27
For a discrete random variable, the variance is computed by:
n
2
σ =  (x
i 1
i   ) 2 Px ( xi ) (3.13)
When all n observations in a sample are given equal weights, [ Px (xi) = 1/n ], then
the variance for a discrete random variable is as follows:
n
1
S2 = 
n  1 i1
( xi  X ) 2 (3.14)
For a continuous random variable, the variance is computed as follows:

   2
σ = 2 ( x ) f x ( x)dx (3.15)

3.2.2.3 Standard Deviation
The standard deviation is the square root of the variance. It has the same units as the
random variable and the mean. Therefore, it is a better descriptor of the spread of the data or
the distribution than the variance.
For a discrete and continuous random variable, the standard deviation of a population
is denoted by σ and is computed by the square root of the variance.

28
For a sample, the standard deviation is denoted by S and defined as:
 n 2

 x   n( x )
 i 1 i 
 
S= (3.16)
n 1
3.2.2.4 Coefficient of Variation
The coefficient of variation is a dimensionless quantity denoted by V, and is defined
as the absolute value of standard deviation divided by the mean as follows:
V = │ │ (3.17)

Even though the mean may be a negative value, the coefficient of variation
parameter is always considered to be a positive value.
3.3 COMMON CONTINUOUS PROBABILITY DISTRIBUTIONS
Several probability distributions have been involved to the common practical use of
engineers. They are classified as discrete and continuous distributions. The most commonly
used ones are the continuous distributions, which are classified as: uniform, normal,
lognormal, gamma, exponential, and extreme type I, II, and III. This study only summarizes
the most two common continuous distributions used for structural engineering research work,
the normal distribution and the lognormal distribution. Further details about distribution types
can be found in Nowak and Collins (2013).

29
3.3.1 NORMAL DISTRIBUTION
The normal probability distribution, also known as Gaussian distribution, is the most
important distribution in structural reliability analysis. It is the most widely used in
engineering applications due to its simplicity. The probability density function (PDF) for a
normal distribution for a domain of - x + is given by:
1  1  x   2 
fx(x)= exp     (3.18)
 2  2    
where μ, is the mean value and, σ, is the standard deviation, are the parameters of the
distribution. Examples of PDF with different μ and σ are presented in Figure 3.6.
fx(x)
μx = 1, σx =1
μx = 3, σx =2
μx = 6, σx =3
Random Variable X
Figure 3.6: PDF’s of Normal Distribution with Different Distribution Parameters

30
The cumulative distribution function of the normal distribution is given by:
x
1  1  x   2 
Fx(x) =  exp  

  dx (3.19)
  2 2
   
Examples of CDF with different μ and σ are presented in Figure 3.7.
Fx(x)
μx = 6, σx =3
μx = 3, σx =2
μx = 1, σx =1
Random Variable X
Figure 3.7: CDF’s of Normal Distribution with Different Distribution Parameters

31
The standard normal distribution is a special case of the normal distribution with
distribution parameters of μx = 0 and σx =1.0. The probability density function of the standard
normal distribution is denoted by φ(z) , which is shown in Figure 3.8, is determined as : Ф
1  1 
φ(z) = exp   z 2  (3.20)
2  2 
φ(z)
Random Variable X
Figure 3.8: PDF of the Standard Normal Distribution

32
The cumulative distribution function of the standard normal distribution is denoted by
Ф(z) is determined as :
z
1  1 
Ф(z) =

 2
exp   z 2 dz
 2 
(3.21)
The cumulative probability “p” for the standard normal variable z is formulated as the
inverse of the standard normal cumulative distribution function as:
z = Ф-1(p) (3.22)
3.3.2 LOGNORMAL DISTRIBUTION
A random variable X is considered to have a lognormal distribution if Y = ln(X) has a
normal distribution, where ln(X) is the natural logarithm (to the base e). The probability
density function (PDF) of the lognormal distribution is given by:
1  ln( x)  ln X )  1  1  ln( x)   

2

f x ( x)  
   exp   Y
  (3.23)
x ln( X )   ln( X )  Y
 x Y 2  2   
where, µY is the mean value of the lognormal distribution, and σY is the standard deviation of
the lognormal distribution. These parameters are defined as follows:

33
  
2

Y 2
  ln( X )
2
 ln 1   X  2
  ln(1  VX ) (3.24)
   X  
1
Y  ln(X )  ln(  X )   X 2 (3.25)
2
An example for the probability density function is shown in Figure 3.9.
fx(x)
Random Variable X
Figure 3.9: Probability Density Function of Lognormal Distribution

34
The cumulative distribution function (CDF) of the lognormal distribution can be
determined based on its relation with the normal distribution. The following formula is used
for describing the cumulative distribution function for lognormal distribution:
 y  y 
Fx ( x)  FY ( y)    
   (3.26)
 y 
3.4 NORMAL PROBABILITY PAPER
Probability paper is used to graphically determine the type of the probability distribution
for a set of experimental data. The normal probability distribution is the most commonly used
as an efficient interpretation of statistical data.
Normal probability paper is constructed to conveniently present cumulative distribution
functions (CDF) for the normal distribution, which is an increasing “S-shape”. Normal
probability paper modifies the vertical scale so that the normal CDF can be plotted as a
straight line as shown in Figure 3.10. For normal probability paper, a straight line represents a
normal distribution. The statistical parameters such as the mean and standard deviation, as
well as the type of distribution, can be evaluated directly from the graph. Many textbooks
include more information about normal probability paper such as Benjamin and Cornell
(1970), and Nowak and Collins (2013).

35
CDF
F(x)
Random Variable X
Figure 3.10: The S-shaped CDF for a Normal Random Variable
The random variable considered as the test data, is presented on the horizontal
axis, while the standard normal variable is presented on the vertical axis. The vertical axis
represents the mean value in terms of standard deviations. The vertical coordinate can also be
considered as the probability of exceeding the corresponding value of the variable. For any
value of the random variable (horizontal axis), the vertical coordinate of CDF corresponds to a
certain probability of being exceeded. For example, value of 1 on the vertical scale
corresponds to (0.841) times the probability that the value of the random variable will be
36
exceeded. More examples of the relationship between the standard normal variable and
probability are given in Table 3.1.
Table 3.1: Examples of Relationship between the Standard Normal Variable and the
Corresponding Probability
Standard Normal Variable Corresponding Probability
4 0.9999683
3 0.99865
2 0.9772
1 0.841
0 0.5
-1 0.159
-2 0.0228
-3 0.00135
-4 0.0000317
The shape of the resulting curve representing CDF allows for analysis of the test data
plotted on the normal probability paper. The indication of degree for variation of these data is
presented by the slope of the straight line as shown in Figure 3.11.

37
Standard
Normal
Variable
CDF
F(x)
3
Mean
0
Random Variable X
-1
Standard Deviations
-2
Figure 3.11: Normal Distribution Function on the Normal Probability Paper

38
3.5 LIMIT STATE FUNCTIONS
The boundary between the safe “acceptable” and failure “unacceptable” domain is
represented by the limit state function or performance functions. Structural design is based on
the concept of limit state functions in most design codes. In order to evaluate the structural
reliability, it is common to define the limit state function that describes the performance of the
structural components. The limit state function is defined as a function of capacity and
demand, it is denoted as g and expressed as:
g(R,Q) = R-Q (3.27)
where R, is the structural resistance or capacity of the structural component, Q is the load
effect or demand of the structural component with the same units as the resistance.
The performance function g(X) is a function of capacity and demand variables (X1,
X2,…, Xn which are basic random variables for both R and Q) such that:
g(X) = g(X1, X2, …, Xn) (3.28)
g(X1,…,Xn) ≥ 0 for a safe realization (3.28 –a)
g(X1,…,Xn)< 0 for failure (3.28 –b)
Both R and Q are continuous random variables having their own probability distribution
function (PDF). The quantity (R-Q) is defined as a safety margin and is also a continuous
random variable with its own PDF as shown in Figure 3.12.

39
PDF Q, Load R, Resistance
R-Q, Safety margin
Failure Probability
= Shaded area
μR-Q μQ μR
Random Variable X
Figure 3.12: PDF’s of Resistance, Load, and Safety Margin
A limit state violation does not necessarily correspond to the collapse of the
structure. Moreover, for structural engineering, three types of limit states can be used with
reference to structural reliability analyses which are investigated as:
1) Ultimate limit states (ULS) are usually related to the loss of structural capacity. These
limit states have been evaluated for the new generation of design codes.
2) Serviceability limit states (SLS) represent the failure in service condition, but do not
necessarily mean structural failure. There are several examples of SLSs such as
unacceptable deformations (vertical and horizontal movements), excessive vibrations,
and local damage (flexural and corrosion cracking). These limit states have not been
thoroughly investigated for the development of the new generation of design codes.
3) Fatigue limit states (FLS) represent the loss of strength for a structural component
under repeated loading. Fatigue can lower the level of acceptable loads.
40
3.6 PROBABILITY OF FAILURE AND RELIABILITY INDEX
The performance function formulated in Eq.(3.27) depends on the consequence of
setting the boundary g(R,Q) =0 between safe and unsafe performance while the limit state
function g(R,Q) > 0 represents safe performance, and when g(R,Q) < 0 represents failure or
unsafe performance. Following the definition of structural reliability, the probability of failure
can be expressed as:
Pf = P((R-Q)<0) = P(R<Q) (3.29)
= P(g<0)
= P[g(X1, X2, …, Xn)]
where Pf , is the probability of failure and R,Q are random variables.
If R and Q are statically independent and have probability density functions fR and
fQ , and the cumulative distribution function of R and Q will be represented as FR and FQ
respectively, then the probability of failure can be expressed as :

Pf   FR (li ) f Q (li )dli (3.30)

Although the above equation seems to be straightforward, it is not easy to solve,
because the integration requires numerical techniques, and these techniques may not be
accurate or adequate. Therefore, the probability of failure will be calculated depending on

41
other indirect procedures and simulation techniques, such as evaluating the reliability index
that will be discussed below.
The probability of failure, Pf , can be related to an indicator called the reliability
index, β . Probabilistic methods based on reliability index are used in structural design. In fact,
the reliability index is the shortest distance from the origin of the space of reduced variables
ZR and ZQ to the line of g(ZR , ZQ) = 0. This concept is clarified in Figure 3.13.
Figure 3.13: Reliability Index defined as the Shortest Distance in the Space of
Reduced Variables
42
The reduced variables are defined as:
R  R
ZR = R (3.31-a)
Q  Q
ZQ = Q (3.31-b)
In terms of reduced variables, the resistance (R) and the load (Q) can be expressed as:
R   R  Z R . R (3.32-a)
Q  Q  ZQ . Q (3.32-b)
Therefore, the limit state function g(R,Q) R-Q in terms of reduced variables can be
written as :
g (Z R , Z Q )   R  Z R . R  Q  Z Q . Q (3.33)
Assuming that the limit state function is normally distributed and the random variables
are uncorrelated, the reliability index is related to probability of failure as:
Pf = Φ(-β) (3.34)
Where, Φ is the standard normal distribution function.

43
Table 3.2 provides an overview of the reliability index evolution, with the function of
probability of failure based on Equation (3.34). It can be noticed from the following table,
the smaller the probability of failure, the higher the reliability index.
Table 3.2: Reliability Index β and Probability of Failure Pf

(Adapted from Nowak and Collins, 2013)
Probability of Failure (Pf) Reliability Index (β)
0.1 1.28
0.01 2.33
0.001 3.09
0.0001 3.71
0.00001 4.26
0.000001 4.75
0.0000001 5.19
0.00000001 5.62
0.000000001 5.99
The reliability index, β, depends on the limit state function, which describes the
failure of the structural element. The resistance, R, and load effects, Q, might have various
types of distributions for which each one of them is either a product or sum of basic random
variables with various distribution parameters as well.

44
3.7 RELIABILITY INDEX EVALUATION
The equations for various limit states usually define the variables for the load, Q, and
resistance, R. The simplest case is by assuming the limit state function normally distributed,
so the reliability index related to the probability can be calculated directly from the following
equation as:
β = - Φ-1(Pf) (3.35)
-1
where - Φ is the inverse standard normal distribution function .
Most limit states include several variables for both load and resistance. Therefore, it
would be very complicated for solving structural problems and obtaining the equivalent
parameters for one of the components of the limit state function. Analytical and simulation
methods have been developed to solve complicated problems with more or less precision. The
analytical methods such as the First-Order Reliability Method (FORM), First-Order Second-
Moment Reliability Method (FOSM), Hasofer-Lind Reliability Index, and Rackwitz-Fiessler
Procedure, among others, are examples of more advanced reliability index evaluation
methods. The (FOSM) will be presented in the next section while details on other methods
can be found in the literature, Nowak and Collins (2013).
Simulation techniques represent the alternative to obtain reliability indices when it is
too complex or not possible to use the simplified methods. Several techniques may be used to
solve structural reliability problems such as Monte Carlo Simulation, Latin Hypercube
Sampling, and Rosenblueth’s 2k+1 Point Estimate Method. More details on these
simulations can be found in Nowak and Collins (2013). In this study, the simulation
45
procedure will be performed using the Monte Carlo method, which has become very popular
with the increase in computer power and will be presented and discussed in a later section.
3.7.1 THE FIRST-ORDER SECOND-MOMENT RELIABILITY
METHOD (FOSM)
The simplest popular method to calculate the reliability index is the First-Order
Second-Moment method. This method takes into consideration the linear limit state functions
or their linear approximation using Taylor series. First order means that only the first Taylor
derivative is used in calculations and Second-Moment refers to the second moment of the
random variable, Ditlevsen and Madsen (1996). First moment is the expected value E(X) and
the second moment E(X2) is a measure of the dispersion in other words variance.
I. Linear Limit State Function:
Linear limit state function can be written in the form of :
n
g ( X 1 , X 2 ,..., X n )  ao  a1 X 1  a2 X 2  ...  an X n  ao   ai X i (3.36)
i 1
Where, ai terms (i=0,1,2,…,n) are constants and Xi terms are uncorrelated random
variables.
The reliability index for a linear limit state function with uncorrelated random
variables and, which depends only on the means and standard deviation of the random
variables can then be calculated from the following expression:

46
n
ao   ai  Xi
  n
i 1
(3.37)
 (a 
i 1
i Xi ) 2
Therefore, the reliability index calculated from Equation (3.37) can be called a second-
moment of structural safety because only the first two moments including the mean and
variance are required for calculating the reliability index.
If the resistance R and the load Q are normally distributed with uncorrelated random
variables, the limit state function g(R,Q) would have the mean value of g, and the standard
deviation of g, as in the following equations respectively:
µ g = µR- µQ (3.38-a)
 g   R2  Q2 (3.38-b)
Then, the reliability index can be calculated from Cornell (1967, 1969):
g  R  Q
  (3.39)
g  R2   Q2
In the other case when R and Q are lognormally distributed with uncorrelated
random variables, the reliability index can be calculated by the following formula,
Rosenblueth (1975, 1981):

47
 
ln  R 
 ln R   ln Q  
 2
Q 
 2 (3.40)
 ln R 2   ln Q 2 VR  VQ
The implementation of the First-Order Second-Moment (FOSM) method is simple.
Since calculations can be performed only for normal distributions, the reliability index for
distributions other than normal includes considerable level of error, Thoft-Christensen and
Baker (1982), and Nowak and Collins (2013).
II. Nonlinear Limit State Function:
For the case of a nonlinear state function with uncorrelated random variables,
an approximation can be obtained by linearizing the nonlinear function using Taylor
series expansion as:
n
g
g ( X 1 , X 2 ,..., X n )  g (  X 1 ,  X 2 ,...,  Xn )   ( X i   Xi ) (3.41)
i 1 X i
g
where are evaluated at µXi.
xi
As Xi, are statically independent random variables, the reliability index for the linear
function can be calculated from the following equation as:
g (  X 1 ,  X 2 ,...,  Xn )

n (3.42)
 (a 
i 1
i Xi ) 2
48
g
where, ai = are evaluated at µXi.
xi
The reliability index calculated in Equation (3.32) is called a first-order second-
moment mean value reliability index because first-order terms are used in Taylor series
expansion and second-moment due to the means and variances being the only ones needed,
and finally mean value is due to using Taylor series expansion about the mean values.
3.7.2 MONTE CARLO SIMULATION TECHNIQUES
Simulation is the process of replacing reality with theoretical and experimental models.
Theoretical simulation is also called numerical or computer experimentation. It is a practical
tool that allows obtaining data, either instead of, or in addition to real-world data. Simulating
a phenomenon numerically assumes events occurring a finite number of times. The frequency
of occurrence of an event in the entire set of simulations approximates its probability of
occurrence. This relatively straightforward concept often requires complex procedures. The
most commonly used simulation technique is the Monte Carlo Method, Thoft-Christensen and
Baker (1982).
The Monte Carlo method is a computational algorithm for simulating the behavior of
various physical or mathematical systems. It is distinguished among other methods by using
nondeterministic random numbers, which allow results to be generated numerically without
doing any actual testing. It is mostly used to solve complex problems for which closed-form
solutions are either impossible or extremely difficult to obtain. Moreover, Monte Carlo
simulation provides an efficient way to determine the probability of structural failure, and
then indirectly, the reliability index for both linear and nonlinear limit state functions.
49
The Monte Carlo simulation method considers the following steps:
1- Generation of uniformly distributed random variables u1, u2,….,un, which are between
0 and 1.
2- Calculation of the standard normal values using generated numbers, including the
types of distributions with their statistical parameters (mean and standard deviation
values) for each design variable.
3- Calculation of the standard random number (zi) from the following equation as :
zi = Φ-1(ui) (3.43)
where, Φ-1 is the inverse of the standard normal cumulative distribution function.
4- Using standard random values (mean and standard deviation values), generate the
values of sample random numbers for the random normal variable (x) or the random
lognormal variable [ln(x)], depending on the distribution of the statistical parameters.
5- Since all random variables are defined, Monte Carlo simulations can be used to
calculate the whole limit state. Therefore, the probability of failure (serviceability or
ultimate) which is the probability of exceeding the allowable limit state can be
described as:
ng ( x)  0
Pf =
N g ( x)
(3.44)
where, n[g(x)] is the number of simulations when the limit state is not satisfied, and
N is the total number of simulations for the limit state.

50
It is important to simulate an efficient number of sets, such that the variation of the
design parameters in a single simulation will not influence the solution of the entire process of
simulations. Moreover, the accuracy of the method depends on the number of simulations.
The reliability index can be calculated by Monte Carlo simulations following the
previous procedure of generating the random variables with their proper distribution. Then,
the simulated values of the limit state function will be plotted on normal probability paper.
The probability of failure can be found at the location where the plotted data curve intersects a
vertical line passing through the origin. Finally, β can be considered as the value of the
standard normal variable at the intersection point. The plotted curve can be extrapolated if the
plotted curve does not intersect the vertical axis as shown in an example presented in Figure
3.14.
3
Standard Normal Variable
0
0 10 20 30 40 50 60
-1
g(x)
-2
-3
-4
-5 β = 4.5
Figure 3.14: Example for Simulated Values of the Limit State Function Generated by
Monte Carlo Method

51
CHAPTER 4
EVALUATION OF SOIL PARAMETERS FOR

SHALLOW FOUNDATONS IN
COHESIONLESS SOILS
4.1 INTRODUCTION
This chapter contains two parts. The first part describes the procedures for evaluating the
soil parameters that are required for the settlement prediction of shallow foundations on
cohesionless soils. The second part of this chapter presents comprehensive description of the
12 highway bridges selected in this study.
4.2 EVALUATION OF SOIL PARAMETERS
This section reviews the evaluation of the soil parameters used in the settlement
prediction analysis for shallow foundations in cohesionless soils. Direct measurements are
preferred more than estimations using well established correlations. This section will be
limited to soil parameters that are utilized in the analyses presented in this study.
52
4.2.1 UNIT WEIGHT
The soil unit weight (γ) of undisturbed samples can be estimated using the soil
description and can be determined by measuring their physical dimensions and weights.
Most weight-volume parameters should be evaluated either directly or indirectly from
laboratory tests on soil samples.
The total unit weight (γ) is presented by:
γ = (W / V) (4.1)
where, W is the weight of soil including moisture, and V is the volume of soil.
The dry unit weight (γd) is presented by:
γd = (Ws / V) (4.2)
where, Ws is the weight of the solids of soil.
The dry unit weight (γd) can be calculated in terms of the unit weight (γ), as
presented below:

γd = (4.3)
1 w
where, w is the moisture content (%) which is detailed in ASTM D2216.
Typical ranges of unit weight (γ) for various soils are presented in Table (4.1).
53
Table 4.1: Typical Unit Weights (Adapted from Coduto, 2000)
Soil Type and Unified Soil Typical Unit Weight, (γ)

Classification
Above Groundwater Table Below Groundwater Table
(lb/ft3) (KN/m3) (lb/ft3) (KN/m3)
GP-Poorly-graded gravel 110-130 17.5-20.5 125-140 19.5-22.0
GW-Well-graded gravel 110-140 17.5-22.0 125-150 19.5-23.5
GM-Silty gravel 100-130 16.0-20.5 125-140 19.5-22.0
GC-Clayey gravel 100-130 16.0-20.5 125-140 19.5-22.0
SP-Poorly-graded sand 95-125 15.0-19.5 120-135 19.0-21.0
SW-Well-graded sand 95-135 15.0-21.0 120-145 19.0-23.0
SM-Silty sand 80-135 12.5-21.0 110-140 17.5-22.0
SC-Clayey sand 85-130 13.5-20.5 110-135 17.5-21.0
ML-Low plasticity silt 75-110 11.5-17.5 80-130 12.5-20.5
MH-High plasticity silt 75-110 11.5-17.5 75-130 11.5-20.5
CL-Low plasticity clay 80-110 12.5-17.5 75-130 11.5-20.5
CL-High plasticity clay 80-110 12.5-17.5 70-125 11.0-19.5
Guidelines given by Chen (2004) to estimate the dry unit weight (γd) for a wide range of
cohesionless soils are shown in Table (4.2), which gives the uncertainty involved in the
estimation of dry unit weights in terms of mean value, standard deviation, and coefficient of
variation (COV) referred to each soil type.

54
Table 4.2: In-Situ Dry Unit Weight Statistics for Various Cohesionless Soils
(Adapted from Chen, 2004)
Mean (γd) Standard COV

Soil Type (KN/m3) Deviation (%)
Quaternary uniform sand
(fine to coarse) 14.0 0.9 6.3
Pre-Quaternary fine sand 17.0 0.2 1.4
Gap-graded gravelly sand 17.8 0.6 3.1
Quaternary sandy gravel 19.1 0.9 4.5
Gravelly cobbles 20.8 1.2 6.0
Lateritic gravel & cobbles 18.2 1.3 6.9
Volcanic sand 10.3 1.6 15.4
4.2.2 SOIL PROPERTIES
In-situ methods of determining soil properties are applicable for sandy soils,
mostly due to the practical difficulties of obtaining undisturbed samples, which will not be
suitable for clayey soils. The in-situ methods are based on measuring the corresponding
response of the soil to evaluate the properties of soil such as strength or stiffness, or both.
The most common in-situ tests are the Standard Penetration Test (SPT), the Cone
Penetration Test (CPT), the Dilatometer Test (DMT), the Pressuremeter Test (PMT), and
the Vane Shear Test (VST).

55
4.2.2.1 STANDARD PENETRATION TEST FOR FIELD PROCEDURES
The soil exploration method that will be considered in this research is the SPT,
which is the in-situ test that has been used most extensively. Schmertmann (1978) and
Kovacs and Salomone (1982) have shown that the standard penetration resistance is measured
using different energy delivery system hammers with various degrees of efficiency, different
size boreholes, different borehole fluids, and different kinds of sampling tubes. The SPT blow
counts N (number of blows per 12 inches of penetration) obtained in a given sand deposit at a
particular effective overburden stress level and constant relative density can still have a wide
variation even under good site control procedures. Since cohesionless soils have excellent
drainability, no excess pore water pressure will be developed. One of the procedures for the
SPT is detailed in ASTM D1586, as shown in Figure 4.1 and the other one (essentially
identical) in AASHTO T206.
The SPT has advantages of being less expensive than other in-situ tests, so that
greater numbers of tests can be performed and the results can be available immediately.
However, it has disadvantages, such as the soil classification being very difficult and the
engineer has less control over confining stresses and drainage.
Although the SPT is a standardized test, it has been reported that the blow counts can
vary by as much as a factor of three, Kovacs (1994). This level of uncertainty includes any
consistent correlation between the N-values and various geotechnical parameters, such as
relative density, bearing capacity, liquefaction potential, and settlement, the geotechnical
parameter considered in this study.

56
Figure 4.1: Sequences of driving during the Standard Penetration Test

(FHWA 2006, Volume I)
Because of the variability of results, it has been found necessary to correct
measured N-values need to a consistent point of reference. There are two types of
corrections in the standardization of the SPT N-values that will be described in the
following two sections.

57
 Correction of SPT N-Values for Field Procedures
The theoretical free-fall energy for the SPT is 0.623 kN (140 lb) times 0.76 m
(30 in) or 0.475 kN.m (4200 lb.in). It was shown by Kulhawy and Mayne (1990) that the
actual energy delivered to the drill rods during the SPT, expressed in terms of rod energy
ratio (ER), can vary in different parts of the world by between 40% and 90% of the
theoretical free-fall energy used be delivered by the falling hammer. The actual energy
may be less due to a number of factors including friction losses and eccentric loading,
which are specifically related to the hammer drop. The recommended value cited in many
references for the energy ratio (ER) is 60% of the free-fall energy of the standard hammer
weight and drop Skempton (1986), which is recognized internationally. The most common
procedure used for raising and releasing the SPT hammer in the US, as with most of the
world, is the rope and cathead technique, where the cathead is a powered pulley controlled
by a number of turns of rope, which slips on the pulley. An operator provides tension,
allowing the hammer to be lift up while the rope still slips. A release of tension allows the
hammer to drop.
There are typically three types of SPT drop hammers i.e., the safety hammer, the
donut hammer, and the automatic hammer, as shown in Figure 4.2. The two most common
types of hammers are the safety and donut hammers.

58
Figure 4.2: SPT Hammer Types: (a) Donut, (b) Safety, (c) Automatic
(FHWA 2006, Volume I)
Typical hammer efficiencies are listed in Table 4.3. Other parameters found to
affect N-values include borehole diameter, and rod length. Approximate correction
factors of borehole, sampler, and rod to the measured N-values for the field procedures are
given in Table 4.4.

59
Table 4.3: SPT Hammer Efficiencies (Adapted from Clayton, 1990)
Country Hammer Type Hammer Release Hammer Efficiency,
Mechanism (Em)
Argentina Donut Cathead 0.45
Brazil Pin weight Hand dropped 0.72
China Automatic Trip 0.6
Donut Hand dropped 0.55
Donut Cathead 0.5
Colombia Donut Cathead 0.5
Japan Donut Tombi trigger 0.78-0.85
Donut Cathead 2 turns 0.65-0.67
UK Automatic Trip 0.73
US Safety 2 turns on Cathead 0.55-0.60
Donut 2 turns on Cathead 0.45
Venezuela Donut Cathead 0.43

60
Table 4.4: SPT Correction Factors for Field Procedures

(Adapted from Skempton, 1986)
Factor Equipment Variables Correction Value
Borehole Diameter 65-115 mm (2.5-4.5 in) 1.00
Factor, CB 150 mm (6 in) 1.05
200 mm (8 in) 1.15
Sampling Method Standard Sampler 1,00
Factor, CS Sampler without liner 1.20
(not recommended)
Rod Length Factor, CR 3-4 m (10-13 ft) 0.75
4-6 m (13-20 ft) 0.85
6-10 m (20-30 ft) 0.95
>10 m (> 30 ft) 1.00
For most correlations, the SPT N-value is corrected for field procedures, and an ER of
60% by converting the measured N to N60 , as described below:
EmCBCS CR N
N60  (4.4)
0.60
61
where, N is the measured blow count, N60 is the SPT N-value corrected for field procedures,
and Em is the hammer efficiency (from Table 4.3). The correction factors are CB, the borehole
diameter correction factor; CS, the sampler correction factor; and CR, the rod length correction
factor. The correction factor values are listed in Table 4.4
The SPT N-values or N60 can be given as an indication for the relative density of
cohesionless soils and the consistency of cohesive soils. Even though N-values can be
correlated to the relative density of sands and the consistency of fine-grained soils,
correlations are unreliable for gravels, silts and clays as shown in Table 4.5. Soil type,
density, and overburden pressure are the most significant factors affecting SPT N-values.
Table (4.5): Soil Properties Correlated with Standard Penetration Test Values
(Adapted from Peck et al., 1974)
Sands (Reliable) Silts and Clays ( Unreliable)
N60 Relative Density N60 Consistency
0-4 Very Loose Below 2 Very Soft
5-10 Loose 2-4 Soft
11-30 Medium Dense 5-8 Medium
31-50 Dense 9-15 Stiff
Over 50 Very Dense 16-30 Very stiff
Over 30 Hard
62
 Normalization of SPT N-Values for Overburden Stress
The standard penetration test N-values depend not only on the relative density as
field procedure corrections are applied, but also depend on the effective overburden stress
at the depth of measurement. This dependence was first discovered in the laboratory by
Gibbs and Holtz, (1957) and then it was demonstrated in the field by Skempton, (1986).
The corrected N-values (N60) can be normalized to an over burden stress of 1-atmosphere
(100 kPa or about 2000 lb/ft2) because the N-values increases with increasing effective
overburden stress up to about 1-atmosphere. The N-values corrected for overburden stress
are related to the measured value corrected for field procedures as follows:
( N1)60 = CN N60 (4.5)
where,(N1)60 is the N60 value corrected to a reference overburden stress of one atmosphere
and CN is the correction factor for overburden stress.
The correction factor has been proposed by Kulhawy and Mayne, (1990) from field
data and by Skempton, (1986) from laboratory test data. Among these relationships, the
simplest equation for calculating the correction factor, and which is the most widely used,
is given by Liao and Whitman (1986) as:
pa
CN = (4.6)
 'z
in which, Pa is the atmospheric stress = (100 kPa or about 2000 Ib/ft2) and σ’z is the
vertical effective stress at the test location.

63
It can be noticed from the above equation, high values of correction factor (CN) can
be obtained from low values of vertical effective stress, which is typically obvious for the
case of shallow foundations. Even though Liao and Whitman did not place any restriction
limits on the correction factor, it is probably the best to consider ( N1)60 ≤ 2 N60 , which
was considered in this study for normalizing SPT values for overburden stress with the
methods of analysis that considered this correction factor.
4.2.2.2 Case History of Automatic Hammer Efficiency
In this section, a case history is presented that illustrates the standard penetration
test N-values using automatic hammers. The site investigation involved two borings: B-1,
consisting of 12 SPTs, and B-2, consisting of 13 SPTs. Logs of both borings, performed
by Terracon Company in Hancock, IA on October 28, 2009, are shown in Figure 4.3. The
approximate surface elevation for both logs of borings is 95.0 ft. Numerical data for each
boring log are presented with more detail in Figures A-1 and A-2 in Appendix A.
The SPT N-values using automatic hammers are analyzed separately within each
stratum for the logs of soil borings shown in Figure 4.3 in terms of SPT N- values with the
depth.
Moreover, the CDF of the SPT N-values for each of the borings is plotted on the
normal probability paper as shown in Figure 4.4, to obtain the statistical parameters for the
borings.
64
Figure 4.3: Automatic Hammer Data in Hancock, IA

65
Figure 4.4: CDF of SPT N-Values for Logs B-1 and B-2 in Hancock, IA
66
The statistical parameters obtained from Figure 4.4 for the logs B-1 and B-2 of soil
borings in Hancock, IA are summarized in Table 4.6.
Table 4.6: Statistical Parameters for SPT N-Values for Automatic Hammer for Logs
B-1 and B-2 of Soil Borings in Hancock, IA
Logs of Borings Mean of N- Values Standard Deviation COV of N-Values

of N-Values
B-1 24.47 11.9 0.48
B-2 21.83 10.1 0.46
4.2.3 OVERCONSOLIDATION RATIO
The overconsolidation ratio (OCR) is an important parameter that has significant

influence on the deformation characteristics of all soils beside it has an effect on the
behavior of soils. It is the ratio of the highest previous effective overburden stress to the
existing effective overburden stress. The value of OCR depends on the stress history of the
soil deposit. Moreover, the value of OCR is not constant with the depth on a given soil
stratum.
There was no information available about the stress history of all the case histories
used in this study. Even though it is difficult to estimate the in-situ OCR for cohesionless
soils, the value of OCR had to be estimated in this study.
If where the geologic history cannot be obtained and an estimate the OCR value
is required, the following equation can be used to estimate the OCR:
67
 c
'
OCR = (4.7)
 ' zo
where, σ’c is the preconsolidation stress determined from a one-dimensional
consolidation test, and σ’zo is the vertical effective stress.
Due to the assumptions involved in this study, any additional information about the
geological history of the case history sites was taken into consideration for selecting the
appropriate value of the overconsolidation ratio (OCR).
Most analyses assume normally consolidated soils, which uses OCR = 1, unless
there is a clear evidence of overconsolidation.
4.2.4 ELASTIC SOIL PARAMETERS
This section will introduce the two main parameters that are used to represent the
deformation characteristics of soils, which are Poisson’s ratio (ν) and soil modulus or
Young’s modulus (Es).
4.2.4.1 Poisson’s Ratio
Poisson’s ratio (ν) is a non-dimensional value that represents a relationship

between the normal strains of a laterally unconfined soil element in uniaxial incremental
stresses and can be defined as follow:
h

v (4.8)
where, εh is horizontal normal strain, and εv is vertical normal strain.

68
For most designs in foundation engineering, Poisson’s ratio has a small variation in
values and is rarely measured. Table 4.7 shows typical values for Poisson’s ratio with
respect to soil type. These values can be used in analysis of deformation with little error.
Table 4.7: Typical Ranges of Poisson’s Ratio for Various Types of Soils
(Adapted from AASHTO, 2012)
Soil Type Poisson’s Ratio, ν
Clay:
Soft sensitive
Medium stiff to stiff 0.4-0.5

Undrained
Very stiff
Loess 0.1-0.3
Silt 0.3-0.35
Fine Sand:
Loose
Medium dense 0.25
Dense
Sand:
Loose 0.2-0.36
Medium dense 0.2-0.36
Dense 0.3-0.4
Gravel:
Loose 0.2-0.35
Dense 0.3-0.4
69
4.2.4.2 Soil Modulus of Elasticity
The stress-strain behavior of soil properties is nonlinear and much more complex
than other structural engineering materials, such as steel or concrete, because soil is a
particulate material. However, the elastic modulus or modulus of elasticity (Es) is widely
used to define the compressibility of soil.
The elastic soil modulus (Es), is defined as the deviator stress (σ1 – σ3) divided by
the axial strain (ε1) of a soil sample in axial compression. These are obtained from the
results of triaxial compression tests. However, for a deformation analysis, the soil modulus
is situated close to the initial stress of the material, while the various in-situ penetration
tests produce failure and, therefore, refer to the peak of the stress-strain curve, as
illustrated in Figure 4.5.
Point Measured by
Strength Penetration tests
( σ1 – σ3)max
Deviator Stress, (σ1 – σ3)
Es
Region Corresponding to Most

Geotechnical Deformation Problems
qo = σvo (1-Ko) = initial stress state
Axial Strain, ε1
Figure 4.5: Strength Measured by in-situ tests at Peak of Stress-Strain Curve

70
Correlations between SPT-N60 test results and soil modulus of elasticity based on
soil types have been summarized by Bowles (1988). The expressions are indicated in
Table 4.8.
Table 4.8: Estimations of Soil Modulus of Elasticity for Various Types of Soils
(Adapted from Bowles, 1988)
Soil Type Soil Elastic Modulus , Es (kPa)
Sand (normally consolidated) 500 * ( N60 + 15)
Sand (saturated) 250 * ( N60 + 15)
Sand (overconsolidated) 18000 + (750 * N60 )@
Gravelly sandy and gravel 600 * ( N60 + 6) for N60 ≤ 15

600 * ( N60 + 6) +2000 for N60 > 15
Clayey sand 320 * ( N60 + 15)
Silty sand 300 * ( N60 + 6)
Notes:
@
: Bowles’ equation from the plot of D’Appolonia et al. (1970).
For deformation analyses, several direct correlations are presented in Tables 4.9 and
4.10 respectively. These correlations, based on soil types for the SPT test values, have
been developed by AASHTO code (2012) to obtain the soil modulus of elasticity.
71
Table 4.9: Typical Ranges of Young’s Modulus of Elasticity for Various Types of
Soils (Adapted from AASHTO, 2012)
Soil Type Young’s Modulus of Elasticity Values,
Es (ksi)
Clay:
Soft sensitive 0.347-2.08
Medium stiff to stiff 2.08-6.94
Very stiff 6.94-13.89
Loess 2.08-8.33
Silt 0.278-2.7
Fine Sand:
Loose 1.11-1.67
Dense 2.78-4.17
Sand:
Loose 1.39-4.17
Dense 6.94-11.11
Gravel:
Loose 4.17-11.11
Dense 13.89-27.78
72
Table 4.10: Equivalent Elastic Modulus for Various Types of Soils Based on
SPT N-Values (Adapted from AASHTO, 2012)
Soil Type Equivalent Elastic Modulus Values,

Es (ksi)
Silts, Sandy silts, slightly cohesive
mixtures
0.056 ( N1)60
Clean fine to medium sands and slightly

silty sands 0.097 ( N1)60
Coarse sands and sands with little

gravel 0.139 ( N1)60
Sandy gravel and gravels

0.167 ( N1)60
Emperical correlations between the drained soil modulus (Ed) and the in-situ tests
measured values are widely used in geotechnical engineering practice due to the effects of
sampling disturbance. Typical values for the normalized drained elastic modulus (Ed / pa)
are presented in Table 4.11.
Table 4.11: Typical Ranges of Drained Modulus for Sandy Soils

(Adapted from Kulhawy and Mayne, 1990)
Consistency Typical Normalized Elastic Modulus, (Ed / pa)
Loose 100 to 200
Medium 200 to 500
Dense 500 to 1000

73
4.3 DESCRIPTION AND BACKGROUND INFORMATION OF

THE HIGHWAY BRIDGES SELECTED IN THIS STUDY
This section presents the background information for the spread footing sites
selected in this study. These data include historical construction summary, subsurface
conditions, design characteristics of the bridge structures, and field instrumentation plans.
Several case studies are selected in this study, and are presented as follows:
4.3.1 Highway Bridges in Northeast United States
In this study, the performance prediction methods for calculating immediate
settlement of spread footing foundations on cohesionless soils are verified by using
compiled field data based on measured settlements for a research project on 20 sites of
spread footings for 10 highway bridges in Northeast of US (Gifford et al., 1987, and
Samatani et al., 2010). The ten highway bridges were located in New York, Connecticut,
Massachusetts, Rhode Island, and Vermont. One of these bridges was a 5-span structure,
two were double-span, three were 4-span bridges, and four were single-span structures.
Nine of the bridges were designed to carry highway traffic, while one instrumented bridge
consisted of a 4-span railroad bridge across an interstate highway. Five of the bridges were
simple-span structures, while the other five were continuous-beam structures. Details of
the bridges, locations, and structural elements evaluated, are shown in Table 4.12.
74
Table 4.12: Classification and Location of Bridges Data Field used in this Study
Bridge No. Bridge Location Structural Element Element Designation

Evaluated
001 Highway VT127 Abutment 1 S1
Burlington, Vermont Abutment 2 S2
002 Dicker Rd. Abutment 1 S3
Cheshire, Connecticut Abutment 2 S4
Center Pier S5
003 Branch Avenue West Abutment S6
Providence, Rhode Island East Abutment S7
Pier 1 North S8
Pier 1 South S9
Pier 2 North S10
Pier 2 South S11
Pier 3 North S12*
Pier 3 South S13*
004 Route 28 South Abutment S14
Collierville, New York North Abutment S15
005 Route 146 North Abutment S16
Uxbridge, Massachusetts South Abutment S17
006 VT Route 11 Abutment 1 S18*
Chester, Vermont Abutment 2 S19*
007 Conrail over I-86 Abutment 2 S20
Manchester, Connecticut
008 Tolland Turnpike Abutment 1 S21
Manchester, Connecticut Abutment 2 S22
009 Route 84 Abutment 1 S23
Manchester, Connecticut Abutment 2 S24
010 ** Route 84 NA
Manchester, Connecticut
Notes: * Construction problems at these footings resulted in disturbance to the subgrade
soils and short term settlement was increased, so these sites were excluded from the
comparison.
** Total settlement was not measured.
NA : Not Available.
75
During the period of construction, which was approximately three years, 24
foundations were monitored from the start of construction, through completion of
construction, to actual use. The post-deck settlement average was approximately 0.25 inches
(6.35 mm) in these bridges. Settlement performance data on the initial through the post-
construction stages were obtained for nine of the ten selected highway bridges. Only 20 sites
were used, due to construction problems with four other sites. The actual settlement of the
sites could be compared with the predicted settlement using the appropriate prediction
methods.
The overall settlements of the spread footing foundations at the 20 sites ranged from
0.02 to 2.72 inches (0.5 to 69 mm), with an average of 0.61 inches (15 mm). Typically about
two thirds (2/3) of the total settlement took place prior to the deck construction.
The thickness of the subsurface bearing soils beneath the spread footings at each of
these bridges ranged from 20 ft to 90 ft (6.1 to 27.5 m) and they were covered with
compacted granular fill (sand and gravel) with a thickness range of 4 ft to 28 ft ( 1.22 to 8.5
m). Two bridge sites contained compressible silt strata beneathe the granular foundation
bearing soils. Prior to the construction of these bridges, placement of embankment preloads
was required as a part of foundation preparation. More information and details about the case
history of each bridge can be found in Appendix B.

76
4.3.2 Highway Bridges Selected in Ohio
This section presents the field performance of the three spread footings at two
highway bridge sites, FRA-670-0380 and MOT-70/75, in Ohio (Sargand and Masada,
2006). These three spread footings were instrumented and monitored during different
phases of construction for the performance prediction methods of calculating immediate
settlement on cohesionless soils.
I- FRA-670-0380 Project
This project was identified by Ohio Department Of Transportation (ODOT). The
bridge is a two-span structure in the city of Colmbus, that allows crossing of High Street
over I-670 . The superstructure of the bridge is a composite consisting of a concrete deck
supported by steel girder beams. Compiled field data based on measured settlements was
available for one footing of this bridge. More details about this bridge, location, and site
are shown in Table 4.13.
The instrumented foundation was monitored for approximately 5 months from the
start of construction through initiation of use. The post-deck settlement average was
approximately 0.13 inches (3.3 mm) in this bridge, which had the largest increase in the
settlement. The value of settlement for the central pier footing of this project varied
between 0.01 to 0.26 inches (0.25 to 6.6 mm), with an average of 0.2 inches (5.1 mm) at
the end of construction.
The soil boring indicated that the soil beneath the footing was stiff, silty sand, and
gravel. Two soil borings, RB-9 and RB-11, were placed in the area of the central pier
foundation.
77
More details about the subsurface conditions and soil boring logs values and averages for
this bridge, construction history and bridge structure with contact pressure are presented in
Appendix C.
II- MOT-70/75 Project
This project was also identified by Ohio Department Of Transportation (ODOT)
at the northeast end of Ramp C bridge constructed as part of the massive I-70/I-75
interchange reconstruction project near Dayton, in Montgomery County. The ramp is a
continuous bridge with 20 spans of steel girders with reinforced concrete piers and
decking. Two locations, pier 18 and pier 19, were investigated to determine how their
foundations reacted to the load generated in each construction stage. These locations
were considered by using compiled field data based on measured settlements. More details
about this bridge, location, and sites are shown in Table 4.13.
Table 4.13: Classification and Location of Bridges in Ohio used in this Study
Bridge No. Bridge Location Structural Element Element Designation

Evaluated
11 High Street Over I-670 Central Pier S25
Columbus, Ohio
12 The Interchange I-70/I-75 Pier 18 (East) S26
Montgomery, Ohio Pier 19 (West) S27
The period of construction was almost a year for both piers 18 and 19, which were
monitored from the start of construction through the initiation of use.

78
The post- deck settlements were small for the two piers: 0.06 inches (1.5 mm) for
pier 18 footing, and 0.13 inches (3.3 mm) for pier 19 footing. The total settlement for the
pier 18 footing varied between 0.36 and 1.20 inches (9.1 and 30.5 mm), with an average of
0.7 inches (17.8 mm). For the pier 19 footing the settlement varied between 0.84 and 1.14
inches (21.3 and 29.0 mm), with an average of 0.96 inches (24.4 mm) at the end of
construction. The largest increase in settlements for both piers 18 and 19 footings were
induced during the placement of girder beams.
The soil boring indicated that the soil beneath the two piers footing was gray sandy
silt with some clay and trace amounts of gravel. More details about the subsurface
conditions and soil boring logs for these two footing locations, construction history, and
bridge structure with contact pressure are presented in Appendix C.

79
CHAPTER 5
SETTLEMENT ANALYSIS FOR SHALLOW

FOUNDATONS IN COHESIONLESS SOILS
5.1 INTRODUCTION
Shallow foundations have been utilized to support various types of structures over the
years. However, one type of structure for which shallow foundations have not seen much
widespread application, both in Nebraska and across the US, is highway bridges. The design
of shallow foundations is most often governed by settlement requirements, and there is a need
to develop rational design criteria for this limit state. A settlement criterion generally controls
the design of shallow foundations on cohesionless soils, and is more critical than safety
against bearing capacity failure. Settlement criteria should ensure that total and differential
settlement is within the tolerable limits. Therefore, the task of estimating total and differential
settlement is an important part for the design process of shallow foundations on cohesionless
soils.
Spread footings should not be designed on soft or weak soil if the design loads are to be
supported without excessive movements. Therefore, it should be taken into consideration for
unsuitable materials to be removed and replaced by suitable and properly to compacted fill
materials. Moreover, spread footings should be designed and proportioned, such that the
supporting soil, provides adequate resistance considered as potential for both bearing capacity
and settlement.
80
Traditionally, bridge engineers take a rather conservative approach by specifying deep
foundations rather than shallow foundations to support the bridge superstructure weight and
live loads. This is because their confidence level in shallow foundations is less than that in
deep foundations when it comes to providing long-term support for highway bridges, and it is
also due to the lack of tolerable vertical movement guidelines for spread footings. It is a
common belief that spread footings settle much more than deep footings, in addition to having
a higher maintenance cost and uncertainty in the soil properties, which considerably affects
accuracy of the settlement prediction methods or models. The performance of geotechnical
structures, including shallow foundations, depends on loads, soil properties, dimensions and
depth of bearing, among other factors.
The settlement of structures built on soil is a subject of considerable interest to
practicing engineers, since excessive settlements can lead to serviceability problems or even
failure states in the structural foundations.
This chapter presents the performance prediction methods for spread footing foundations
by using 23 sites of footings for ten bridges in Northeast US and two bridges in Ohio. Six
settlement prediction methods are selected for calculating immediate settlement of spread
footings on cohesionless soils for each site, with the predictions compared to measured values
of settlement. Differential settlements are calculated between different sites of these bridges.
Statistics of all these six settlement prediction methods are evaluated and summarized.
Due to the complex nature of the compressibility of cohesionless soils and the practical
difficulties in obtaining undisturbed samples, it is impossible to effectively simulate the soil
conditions in the laboratory. Therefore, most of methods considered for calculating
immediate settlement under shallow foundations cohesionless soils depend on empirical

81
correlation between in-situ tests and soil deformation properties. This study looks at the
immediate settlement for shallow foundations placed in sandy soils by using prediction
methods depending on soil stiffness, which is quantified indirectly through the standard
penetration test (SPT) and the blow counts values.
5.2 TYPES OF SETTLEMENT
The vertical downward load is often the greatest load applying to foundations.
Therefore, the resulting downward movement is often the largest and the most effective and
important movement which can be defined as “settlement”.
A question that faces foundation engineers is not whether the foundation would settle,
but rather, how much settlement will occur, with concern that it should not exceed the
tolerable limit.
There are many different ways that structures can settle, which means there are many
types of settlement. Sometimes structures move down as a unit, so the settlement is uniform,
causing no tilt or distortion to the structure, but there might be problems with other close
structures or with the structure’s interface with the adjacent ground. Other settlement
possibilities would be if the structure settles with a linear variation, causing the structure to tilt
with no distortion. Finally, the most dangerous one, which has the greatest source of
problems, is when the structure settles irregularly causing distortion to the structure. It is
obvious, that there are different modes of settlement; so to simplify this problem; settlement
82
will be described by using only two parameters as “total settlement” and “differential
settlement”, both of which will be defined in this section.
5.2.1 Total Settlement
Total settlement (δ) can be defined as the change in foundation elevation from the
unloading position to the final loading position as shown in Figure 5.1.Total settlement can be
considered as elastic or immediate settlement (short-term settlement) during construction in
cohesionless soils, which is the instantaneous deformation of the soil mass that occurs after
loading. Consolidation settlement (long-term settlement) could be counted as part of total
settlement, but it is calculated only for cohesive soils.
Figure 5.1: Total Settlement in a Spread Footing Foundation

83
Allowable total settlement (δa) contains a safety factor and it should be compared to predicted
values of settlement, as shown in the following equation below:
δ ≤ δa (5.1)
Typical values of allowable total settlement for bridges are about 2.0 inches (50 mm) by
Coduto, (2000). However, Zekkos, (2004) considered a total foundation settlement of 1.0 inch
(25mm) as a common design value resulting in acceptable differential settlements.
5.2.2 Differential Settlement
Differential settlement (δD) is the difference in total settlement between two
foundations or between two sites on the same foundation. Differential settlement often causes
angular distortion, which is both more dangerous and troublesome than total settlement
because it can cause distortions to the structure. More details will be discussed in the
evaluation of settlement in the context of bridges in this chapter.
Geotechnical engineers usually design the foundations for a structure such that all of
footings would have the same computed total settlement, which means the structure will settle
uniformly. Meanwhile, the actual performance of the footings would usually not be the same
as predicted ones having some of them settling more than expected and others less than
expected, which in other words would be the actual differential settlement.
In the design of a foundation, the allowable differential settlement (δDa) should satisfy
the following criteria:

84
δD ≤ δDa (5.2)
For tolerable movement of highway bridges, AASHTO (2012) suggests for a 100-ft
(30-m) span, the acceptable differential settlement would be 4.8 inches ( 122 mm) for
continuous spans and would be 9.6 inches (244 mm) for simple spans.
5.3 EVALUATION OF SETTLEMENT IN THE CONTEXT OF

BRIDGE DESIGN
Bridges are structures with spans larger than 20 ft (6 m) and for long span bridges
it can be over 300 feet (100 m). There are 584,000 bridges in the USA. Interstate highway
network consists of 73,200 km of highways and 54,800 bridges.
Bridges contains three parts as: superstructure, substructure, and bearings.
Superstructure is over the bearings and it includes the girders and slab of the bridge. While,
substructure is below the bearings, and it includes the foundations, piers, and abutments of the
bridge. These parts are shown in Figure 5.2.
Bridge foundations should be designed so that their deformations (immediate and
differential settlements) will not cause any damage to the bridge structure. Unequal
displacements of bridge abutments and pier foundations can affect the safety of the traveling
public and the structural integrity of the bridge.

85
Superstructure Over the bearings
Substructures Below the bearings
Figure 5.2: Superstructure and Substructure Parts of the Bridge
Tolerable movements of bridges were discussed by FHWA, (1987) and Duncan and
Tan , (1991) through several studies which led them to found that “foundation movements
would become intolerable for some other reason before reaching a magnitude that would
create intolerable rider discomfort.” The “other” reasons might include reduction of clearance
at overpasses and drainage considerations. The severity of the consequences of uneven
movements of bridge structures, superstructure as well as substructure, increases with the
magnitude of the settlements and lateral movements.

86
Duncan and Tan, (1991) illustrated the subdivision of settlement for bridges into three
components as shown in Figure 5.3 below:
Figure 5.3: Components of Settlement and Angular Distortion in Bridges
(Adapted from Duncan and Tan, 1991)

87
The three components of settlement subdivision are clarified as:
1. Uniform Settlement: This case describes equal settlement in all bridge support
elements. It can cause differential settlement even though the bridge support
elements settle equally, due to the approach embankment, approach slabs, and
utilities that are located across the end-spans of the bridges. Several problems
could be created because of differential settlement in this case such as: reduction of
the clearance of the overpass, a bump could be created at the end of the bridge,
drainage of drainage problems, and destruction of underground utilities at the
bridge approaches.
2. Tilt or Rotation: This case occurs mostly in single span bridges and is associated
with stiff superstructures. Tilt or rotation many not cause harm to the
superstructure, but the associated differential settlement may cause problems
similar to those indicated in uniform settlement as discussed above.
3. Differential Settlement: This case is the most dangerous one, and it directly
results in distortion of the superstructure of the bridge. There are two different
patterns of differential settlement, as described below:
a. Regular Pattern: This case considers increasing the settlement
progressively from the abutments of the bridge towards the center of it.
b. Irregular Pattern: This case considers variation of the settlement at each
support along the length of the bridge.

88
Both of the above patterns of differential settlement lead to angular distortion as
clarified in Figure 5.3. The angular distortion can be defined as the difference between two
points of immediate settlement, divided by the distance between these two points or the
span length. Tolerable movement criteria for angular distortion for bridges are shown in
Table 5.1 below:
Table 5.1: Tolerable Movement Criteria of Angular Distortion for Highway Bridges
(Adapted from AASHTO, 2012)
Limiting Angular distortion, (δD/SL) Type of Bridge Span
0.004 Continuous-Span Bridges
0.008 Simple-Span Bridges
Note: δD is differential settlement, SL is the span length. The quantity (δD/SL) is
dimensionless and should be used with the same units for δD and SL.
A hypothetical case can be used to demonstrate the interrelationship of total and
differential settlement and angular distortion, as shown in Figure 5.4. For a 4-span bridge
structure with 5 support elements, including 2 abutments and 3 piers, it can be seen that the
total settlement (δ) is different at each support, resulting in differential settlement, and then
angular distortions that that have the potential to cause structural damage to the bridge.
89
Figure 5.4: Schematic Concept of both Total and Differential Settlement and Angular
Distortion in Bridges, (Adapted from FHWA, 2010)
Even though the calculation of total or immediate settlement is an important step in
evaluating the methodology of using shallow foundations on cohesionless soils for bridge
foundations, still the differential settlement and angular distortion are the most effective
parameters for the evaluation of damaged bridge facilities. The issue of using a specific
method for calculating immediate settlement combined with issues related to accuracy and
reliability are all effectively combined, when settlements are expressed in terms of differential
settlements and angular distortions.

90
5.4 LOAD – SETTLEMENT BEHAVIOR
Load test results can provide any important information to verify or modify an existing
foundation design. Load-settlement curves for axial compression tests for shallow foundations
can normal take one of the three curves shown in Figure 5.5 (Hirany and Kulhawy 1988) .
Curve A reaches a peak ultimate capacity of the foundation, after which the load decreases
with additional downward movement for the foundation. Curve B reaches the ultimate
capacity of the foundation without decreasing the load later. Moreover, Curve C does not
reach any peak load, therefore not defining the maximum capacity or resistance. The majority
of load-settlement curves take the form of Curve C. It is demonstrated below that the load-
settlement curves for sites selected in this study most closely resemble Curve C.
Load
Displacement
Figure 5.5: Typical Axial Load-Displacement Curves,

(Adapted from Hirany and Kulhawy 1988)
91
The criterion of tolerable movements is used by most geotechnical designers, who
normally consider all the loads to be applied at the same time to the bridge structure, which
would not cause a large effect on the superstructure. However, in reality, the loads are applied
gradually to the structure, and settlement will occur gradually as construction proceeds. There
are points during construction that are more critical than other stages of construction. Figure
5.6 (Adapted from FHWA, 2010) shows a schematic concept of the critical construction
points for a bridge-pier. Moreover, Figure 5.6-a shows the critical construction points and
their associated loads. Figure 5.6-b shows the load-settlement behavior. It can be seen that
settlements between points of loads X and Z are the most critical and relevant ones.
Figure 5.6-a: Identification of Critical Construction Points

92
Figure 5.6-b: Conceptual Load-Displacement Pattern for a given footing width
Legend:
W Load after footing construction

δW Displacement under load W
X Load after pier column/wall construction
δX Displacement under load X
Y Load after superstructure construction
δY Displacement under load Y
Z Load after wearing surface construction
δZ Displacement under load Z
93
5.5 SETTLEMENT PREDICTION METHODS SELECTED IN THIS

STUDY
The most popular methods used for predicting immediate settlements are presented in
text books. There are numerous publications on various aspects of spread footings and
settlement calculations based on either standard penetration tests (SPT) or cone penetration
tests (CPT) for cohesionless soils. In this study, settlement calculations are based on (SPT)
only.
For shallow foundations in cohesionless soils, Douglas (1986) mentioned the
availability of 40 methods for predicting settlement that consider settlement as a function of
the soil stiffness, foundation width (B), foundation shape (L/B), embedment depth (D/B),
depth of ground water table (Dw/B), and the applied pressure (q).
Many settlement prediction methods have been widely used over the past two decades.
Six of these methods, are selected for this study for calculation of the immediate settlement of
spread footings on cohesionless soils. These methods were selected for their simplicity,
popularity, and wide acceptance in practice. These methods and their corresponding equations
are discussed below.
The general forms for the selected six methods are given in this section. More detailed
procedures for each method can be found in the corresponding references. Each method is
applied in conformance with the procedures presented by the respective authors.
For most of the methods, the following notations are used, so only the correction factors
applied in each method need to be defined separately for each method.

94
B= width of footing (ft)
Df = depth of footing below ground surface (ft)
Dw = depth of ground water below base of foundation (ft)
Es = soil modulus of elasticity (lb/ft2)
γt = total unit weight of soil (lb/ft3)
γ’ = effective unit weight of soil (lb/ft3)
H= thickness of soil layer below the footing (ft)
Ic = compressibility index
= influence factor at midpoint of soil layer
L= length of footing (ft)
M = modulus of compressibility (tsf) depending on (SPT) blow counts
N= uncorrected number of SPT blows per foot.
N60 = N corrected for field procedures to 60% energy ratio
(N1)60 = N corrected for overburden stress
po = effective overburden pressure at the mid-height of subdivided layer (lb/ft2)
= increase in overburden pressure at the mid-height of subdivided layer (lb/ft2)
qnet = net allowable bearing pressure ( kips/ft2)
S = total settlement (inch)
σo = total overburden pressure (lb/ft2)
σo’= effective overburden pressure (lb/ft2)
z = thickness of subdividual soil layer (ft)

95
5.5.1 Method of Hough (1959)
This is a method that assumes one-dimensional compression. This method has been
recommended by AASHTO code (2012). It has several advantages when compared with the
other methods that are used to estimate immediate settlement on cohesionless soils, because it
considers soil layering and the zone of stress influence beneath the footing. For analysis, the
subsurface soil profile is subdivided into layers with a depth of about three times the footing
width. The change in effective vertical stress at the mid-height of each layer, as a result of
applied loads is estimated using elastic theory relationship.
The dimensionless bearing capacity index (C’) is calculated from an empirical chart
related to the (SPT) blow counts corrected for overburden pressure (N1)60 as shown in Figure
5.7. Then C’ is used for calculating the immediate settlement.
Figure 5.7: Bearing Capacity Index Versus Corrected SPT

(Adapted from Cheney and Chassie, 2000 after Hough 1959)
96
The governing equation is used for calculating settlement for each subdivided layer is
∆z = z log10 [ ] (5.3-a)
Then, the total settlement is calculated for all subdivided layers as:
S=∑ (5.3-b)
in which, n is number of soil layers within the zone of stress influence of the footing and ∆z
is the settlement of the subdividual layer (ft).
5.5.2 Method of Meyerhof (1965)
This is an empirical method modified from Terzaghi and Peck (1948). In this
approach, Meyerhof proposed a correlation of the net bearing pressure with the corrected
standard penetration resistance (N60). He suggested that the net allowable bearing pressure
should be increased about 50%. He also suggested that the ground water effects would be
implicitly incorporated into the SPT results, so no water table correction was recommended.
Finally, he suggested the use of a depth factor, due to the effectiveness of both the depth and
width of the footing on settlement calculations. Hence, the final equation considered for
calculation of settlement is expressed in Equations (5.4-a and 5.4-b) depending on the width
of the footing.
S= for (B ≤ 4 ft) (5.4-a)

97
and S=( )( ) for ( B > 4 ft) (5.4-b)
in which Fd is depth factor = 1 + 0.33(Df / B)
5.5.3 Method of Peck and Bazaraa (1969)
This method is an empirical method with modification from the method of Terzaghi
and Peck (1948), and Gibbs and Holtz (1957) for predicting settlement of spread footings on
cohesionless soils. The modifications were based on replacing the N60 with (N1)60, which is
the blow count of (SPT) corrected for overburden stress. In addition, the settlement equation
was multiplied by both a water table correction factor and an embedment correction factor.
The final equation used for calculating settlement is expressed in as follows:
S = Cw Cd ( )( ) (5.5)
In which, Cw is the water table correction factor. It is calculated as the ratio of vertical
total overburden stress at a depth of 0.5 B, assuming that no ground water exists, to the
effective overburden stress at a depth of 0.5 B. The other correction factor is Cd, which is the
embedment or depth correction factor. The two correction factors can be determined from
Equations (5.5-a and 5.5-b), respectively as follows:

98
Cw = (5.5-a)
Cd = 1.0 – 0.4 ( )0.5 (5.5-b)
5.5.4 Method of D’ Appolonia (1970)
This method has a basis in elastic theory, and it utilizes SPT blow counts to estimate
the soil modulus of compressibility (M), as shown in Figure 5.8. This method also uses
correction factors for footing embedment and layer thickness. Those correction factors can be
obtained from the charts obtained by Christian and Carrier (1978), which are given in Figure
5.9 and Figure 5.10.
Figure 5.8: Correlation between Modulus of Compressibility and SPT Blow Counts
(Adapted by D’Appolonia, 1970)
99
Figure 5.9: Values of Embedment Correction Factor used for D’Appolonia Method
Figure 5.10: Value of Thickness Correction Factor used for D’Appolonia Method
100
The equation used for calculation settlement on spread footings of cohesionless soils is
presented as follows:
S = μo μ1 (5.6)
in which, μo = embedment correction factor depending on (Df/B); μ1 = thickness correction factor
of sand layer depending on (H/B).
5.5.5 Method of Schmertmann (1978)
This is a semi-empirical method based on elastic theory. This method is based on the cone
penetration test (CPT) data, but for the case where values of CPT are not available, the (SPT) data
can be converted and used for calculation of modulus of elasticity (Es) as specified in Coduto,
(2000). This method uses a simplified distribution of vertical strain under the center of a shallow
footing, expressed in terms of a strain influence factor, Iz to calculate settlement. The Federal High
Way Administration FHWA, (2006 and 2010) recommends using this method, since it is a rational
method that considers the elastic properties of the soil even when the soils are layered. In addition, it
addresses applied stresses and their associated strain influence distribution with the depth below
various footing shapes. In addition, this method has empirical corrections for depth of embedment,
secondary creep in the soil, and footing shape. The formula used for computing settlement is
presented as follows:
S = C1 C2 C3 qnet ∑ (5.7)
101
in which, C1= depth factor; C2 = secondary creep factor; C3 = shape factor = (1.0 for square and
circular foundations); n= number of soil layers considered in the calculations; Izi= value of strain
influence factor, Iz at midpoint of soil layer i.
The correction factors can be determined from the following equations:
C1 = 1 – 0.5 ( ) (5.7-a)
C2 = 1 + 0.2 log ( ) (5.7-b)
C3 = 1.03 – 0.03 (L/B ) (5.7-c)
= effective vertical stress at depth D below the ground surface (Ib/ft2); t = time since
application of load (yr).
Several direct correlations between Es and N60 have been developed by Kulhawy and Mayne
(1990) that developed a relationship which should produce approximate, if somewhat conservative,
values of Es as shown in the following equation:
E s = β0 OCR + β1 N60 (5.8)
in which, β0, β1 = correlation factors obtained from Table 5.2; OCR= overconsolidation ratio.
Table 5.2: Correlation Factors (Coduto, 2000)
Soil Type β0 β1
(lb/ft2) (kPa) (lb/ft2) (kPa)

Clean Sands (SW and SP) 100,000 5,000 24,000 1,200
Silty Sand and Clayey Sands

(SM and SC) 50,000 2,500 12,000 600
102
The relative contribution of each layer to the total settlement is taken into account using
strain influence factors that were derived from load model tests and finite element analyses. The
variation of Iz with depth is given in Figure 5.11 for a square and continuous footing separately.
Figure 5.11: Variation of Iz with Depth ( Adapted by Schmertmann, 1978)

103
To compute the magnitude of the strain influence factor for each layer, the peak value,
Izp , should be determined first from the following equation:
Izp = 0.5+ 0.1 ( ) (5.9)
in which, = initial effective vertical stress at depth of peak strain influence factor,
computed at depth of (D+B/2) below the ground surface for square and circular foundations,
and at depth of (D+B) for continuous footings.
The strain influence factor at the midpoint of each layer can be calculated from the
following equations as:
1) Square and circular footings ( L/B =1):
Iz = 0.1 + [(zf / B) (2 Izp -0.2)] for zf = 0 to B/2 (5.10)
Iz = 0.67 Izp ( 2- zf / B) for zf = B/2 to 2B (5.11)
2) Continuous footings ( L/B ≥10):
Iz = 0.2 + [(zf / B) ( Izp -0.2)] for zf = 0 to B (5.12)
Iz = 0.33 Izp ( 4- zf / B) for zf = B to 4B (5.13)

104
3) Rectangular footings ( 1< L/B < 10) :
Iz = Izs + 0.11[( Izc - Izs) (L/B -1)] (5.14)
in which, zf = depth from bottom of footing to midpoint of layer; Izs = Iz for a square footing;
Izc = Iz for a continuous footing.
5.5.6 Method of Burland and Burbidge (1985)
This method establishes an empirical relationship between average (SPT) blow counts,
foundation width, and foundation subgrade compressibility. It is based, on a statistical
analysis for several case studies of settlement data for shallow foundations on cohesionless
soils. Blow counts are not corrected for overburden pressure, but they are corrected for
subgrade layers, with three correction factors being used for calculating settlement. The
equation used for computing settlement for normally consolidated soils is presented as:
S = qnet B0.7 Ic fs fl ft (5.15)
in which, Ic = compressibility index = ;
N60’ = is the average value of N60 within the influence depth; fs = shape correction factor; fl :=
correction factor for thickness of soil layer; ft = time factor , used if t ≥ 3 yrs.
105
The average SPT-N value corrected for hammer efficiency must be at least 15 and is adjusted
by Equation (5.16) as shown below:
N60’ = 1.25 N60 for N60 ≤ 15 (5.16-a)
N60’ = 15 + 0.5 (N60 -15) for N60 > 15 (5.16-a)
The correction factors can be determined from the following equations:
fs = ( ) (5.17-a)
fl = ( ) (5.17-b)
ft = 1 + R3 + Rt log(t’/3) (5.17-c)
in which, z1 = influence depth (m); R3 = time dependent settlement during the first three years
of loading = (0.3- 0.7); Rt = time-dependent settlement that takes place after the first three
years at a slower rate = (0.2 – 0.8); t’ = time at the end of construction (yr) .
106
5.6 DATA BASE OF MEASURED SETTLEMENTS COMPARED

WITH PREDICTED SETTLEMENTS
In this study, field data based on measured settlements of a total of 23 spread footings
are selected from 12 highway bridges. Ten of the bridges are in the Northeast US and are
described by, Gifford et al. (1987) and Samtani et al. (2010). Twenty individual sites of spread
footings have been selected from the ten bridges. All information and details about location
and sites of these 10 highway bridges are presented in Chapter 4, section 4.3.1, and Appendix
B . The other two highway bridges are located in Ohio, ODOT projects FRA-670-0380 and
MOT-70/75, as presented in Sargand and Masada (2006). Three sites of spread footings have
been selected from these two bridges. All information and details about location and sites of
these two highway bridges are presented in Chapter 4, section 4.3.2, and Appendix C.
The values of measured settlements (δM) are shown in Figure (5.12) for the 23 sites of
spread footings for the 12 highway bridges.
Measured Settlements
1.2
1
Settlement (in.)
0.8
0.6 Mean = 0.5

0.4
0.2
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Sites of Spread Footings
Figure 5.12: Measured Settlements for Study Bridge Elements

107
The values of measured settlements (δM) for the 23 sites of spread footings are
compared with predicted values of settlement (δP), calculated by the six prediction methods,
are listed in Table 5.3.
The spread of data obtained from Table 5. 3 are shown in Figure 5.12, which presents
the accuracy of all the six methods such that the more accurate method has more data points
closer to the diagonal line 1:1, demonstrating that some methods are more accurate than
others.
Table 5.3: Comparison of Measured and Predicted Values of Settlement
Settlement Values (inch)

Site Measured Predicted (δP)
#
(δM) Hough Meyerhof Peck D’Appolonia Schmertmann Burland
and and
Bazaraa Burbidge
S1 0.35 0.83 0.25 0.32 0.50 0.66 0.23
S2 0.67 0.63 0.15 0.19 0.34 0.52 0.13
S3 0.94 0.82 0.19 0.25 0.20 0.85 0.16
S4 0.76 1.43 0.42 0.39 0.35 0.66 0.54
S5 0.61 1.12 0.45 0.29 0.25 0.34 0.63
S6 0.42 0.87 0.18 0.54 0.33 0.28 0.14
S7 0.61 1.51 0.36 0.39 0.65 0.37 0.49
S8 0.28 0.72 0.38 0.32 0.45 0.31 0.42

108
S9 0.26 0.89 0.30 0.24 0.41 0.29 0.39
S10 0.29 0.64 0.39 0.50 0.55 0.30 0.43
S11 0.25 0.56 0.35 0.46 0.46 0.32 0.34
S14 0.46 1.12 0.41 0.85 0.57 0.48 0.42
S15 0.34 1.59 1.07 0.97 0.74 0.92 1.64
S16 0.23 0.75 0.18 0.17 0.26 0.26 0.19
S17 0.44 0.86 0.31 0.58 0.35 0.41 0.39
S20 0.64 0.71 0.34 0.25 0.44 0.97 0.38
S21 0.46 1.20 0.35 0.21 0.58 0.18 0.62
S22 0.66 0.96 0.28 0.53 0.48 0.62 0.33
S23 0.61 1.10 0.33 0.53 0.42 0.85 0.33
S24 0.28 0.66 0.25 0.44 0.29 0.63 0.19
S25 0.20 0.44 0.29 0.11 0.19 0.42 0.21
S26 0.70 0.86 0.31 0.26 0.36 0.63 0.35
S27 0.96 0.74 0.20 0.13 0.27 0.82 0.22

109
1.00
0.80
Measured Settlement(inch)
Hough
0.60 Meyerhof
Peck&Bazarra
0.40
D'Appolonia
0.20
Schmertmann
Burland&Burbidge
0.00
0.00 0.50 1.00 1.50 2.00
Predicted Settlement (inch)
Figure 5.13: Comparison of Measured Settlements with Predicted Settlements for all Six
Methods based on Field Data.
In the context of Load and Resistance Factor Design (LRFD), accuracy refers to
bias factor, which is defined as the ratio of the mean value of the measured settlement to the
value of predicted settlement (δM/ δP). In Table 5.4 , values of the bias factor for settlement
of the six prediction methods are listed. Then, statistical parameters including mean value (µ),
standards deviation (STDEV), and coefficient of variation (COV) of the bias factor of
settlement for the six prediction methods are presented in Table 5.5.
110
Table 5.4: Bias Factor for Settlement (δM/δP) of the Field Data
Site Bias Factor (δM /δP)
# Hough Meyerhof Peck and D’Appolonia Schmertmann Burland

Bazaraa and
Burbidge
S1 0.42 1.39 1.11 0.69 0.53 1.53
S2 1.06 4.33 3.49 1.97 1.29 5.30
S3 1.15 5.05 3.69 4.65 1.11 5.77
S4 0.53 1.80 1.95 2.17 1.15 1.41
S5 0.54 1.34 2.09 2.49 1.79 0.97
S6 0.49 2.36 0.78 1.28 1.50 2.91
S7 0.40 1.71 1.57 0.94 1.65 1.25
S8 0.39 0.74 0.89 0.62 0.90 0.66
S9 0.29 0.86 1.06 0.64 0.90 0.67
S10 0.46 0.75 0.58 0.53 0.97 0.67
S11 0.45 0.71 0.55 0.54 0.78 0.74
S14 0.41 1.13 0.54 0.81 0.96 1.08
S15 0.21 0.32 0.35 0.46 0.37 0.21
S16 0.30 1.28 1.32 0.88 0.88 1.23
S17 0.51 1.41 0.75 1.27 1.07 1.14
S20 0.90 1.87 2.58 1.44 0.66 1.68

Table 5.4, Continued 111
S21 0.38 1.32 2.16 0.79 2.56 0.75
S22 0.69 2.37 1.24 1.37 1.06 2.03
S23 0.55 1.87 1.14 1.44 0.72 1.87
S24 0.43 1.14 0.63 0.95 0.44 1.45
S25 0.49 0.69 1.82 1.05 0.48 0.95
S26 0.81 2.26 2.69 1.94 1.11 2.00
S27 1.30 4.80 7.38 3.56 1.17 4.36
Table 5.5: Statistical Parameters of Bias Factor for Settlement (δM /δP) of the
Six Prediction Methods
Statistical Hough Meyerhof Peck and D’Appoloina Schmertmann Burland

Parameters Bazaraa and
Burbidge
Max 1.30 5.05 7.38 4.65 2.56 5.77
Min 0.21 0.32 0.35 0.46 0.37 0.21
Mean 0.57 1.80 1.76 1.41 1.05 1.77
STDEV 0.28 1.29 1.54 1.03 0.49 1.47
COV 0.50 0.71 0.88 0.73 0.47 0.83
Where, Max = maximum value; Min = minimum value; STDEV = standard deviation ; COV
= coefficient of variation = STDEV/Mean ; Count =23.

112
Even though AASHTO (2012) recommended the use of Hough method (1959) and
FHWA (2006, 2010) recommended the use of Schmertman method (1978), which are both
used for calculating immediate settlement on cohesionless soils, other methods were selected
in this study to check which method would be the most accurate one for calculating settlement
of spread footings on cohesionless soils. It can be concluded from Table 5.5, which presents
the statistical parameters of settlement for the six prediction methods, that the Hough (1959)
and Schmertmann (1978) methods gave the lowest values for COV of 0.5 and 0.47,
respectively, and the Schmertmann method resulted in a value for bias factor closest to 1.0.
Both the Hough and Schmertmann methods seem to produce better results than other
methods, but the Schmertmann method appears to be the best for predicting immediate
settlement on cohesionless soils, based on bias factors.
Figures 5.14 through 5.19 show graphs for each of the prediction methods for calculated
settlements versus measured settlements for all the 23 sites and graphs showing the variation
involved in the bias factor for settlement in terms of frequency histograms. These are
presented to develop a better understanding of the corresponding variation among the six
prediction methods.
It turns out that the Schmertmann method (1978) has the highest coefficient of
correlation (R2), is a statistical measure of two or more random variables, which by its value
indicates how much of a change in one variable is explained by a change in the other.
A value of 1.0 indicates perfect correlation between the two variables; the method with the
highest coefficient correlation is more accurate than the others for calculating immediate
settlement for shallow foundations on cohesionless soils.

113
Hough Method
Measured Settlement(in.) 1.2
1
y = 0.1854x + 0.3274
R² = 0.0632
0.8
0.6
0.4
0.2
0
0.00 0.50 1.00 1.50 2.00
Predictedd Settlement(in.)
Figure 5.14-a: Predicted Versus Measured Settlements for Hough Method (1959)
0.3
Hough Method
0.25
Mean = 0.57
0.2 STDEV = 0.28
COV = 0.5
Frequency
n = 23
0.15
0.1
0.05
0
0.21
0.07
0.14
0.28
0.35
0.42
0.49
0.56
0.63
0.77
0.84
0.91
0.98
1.05
1.12
1.19
1.26
1.33
1.47
1.54
0
0.7
1.4
Bias Factor of Settlement (δM/δP)
Figure 5.14-b: Distribution of Bias Factor for Settlement for Hough Method (1959)
114
Meyerhof Method
1.2
Measured Settlement(in.)
1
0.8 y = -0.2395x + 0.577

R² = 0.0368
0.6
0.4
0.2
0
0.00 0.20 0.40 0.60 0.80 1.00 1.20
Predicted Settlement(in.)
Figure 5.15-a: Predicted Versus Measured Settlements for Meyerhof Method (1965)
0.25
Meyerhof Method
0.2 Mean = 1.80
STDEV = 1.29
COV = 0.71
Frequency
0.15 n = 23
0.1
0.05
0
3.75
0.25
0.75
1.25
1.75
2.25
2.75
3.25
4.25
4.75
5.25
2
0
5.5
0.5
1.5
2.5
3.5
4.5
Figure 5.15-b: Distribution of Bias Factor for Settlement for Meyerhof Method (1965)
115
Peck and Bazarra Method

1.2
Measured Settlement(in.)
1
y = -0.1987x + 0.5737
0.8 R² = 0.0367
0.6
0.4
0.2
0
0.00 0.20 0.40 0.60 0.80 1.00 1.20
Figure 5.16-a: Predicted Versus Measured Settlements for Peck and Bazarra Method
0.25
Peck and Bazarra Method

0.2
Mean = 1.76
STDEV = 1.54
0.15 COV = 0.88
Frequency
n = 23
0.1
0.05
0
3.15
0.35
1.05
1.75
2.45
3.85
4.55
5.25
5.95
6.65
7.35
7
0
0.7
1.4
2.1
2.8
3.5
4.2
4.9
5.6
6.3
7.7
Figure 5.16-b: Distribution of Bias Factor for Settlement for

Peck and Bazarra Method (1969)
116
D'Appolonia Method
1
y = -0.3429x + 0.6373
0.8 R² = 0.0482
0.6
0.4
0.2
0
0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80
Figure 5.17-a: Predicted Versus Measured Settlements for D’Appolonia Method (1970)
0.25
D'Appolonia Method
0.2
Mean = 1.41
STDEV = 1.03
0.15 COV = 0.73
Frequency
n = 23
0.1
0.05
0
0.25
0.75
1.25
1.75
2.25
2.75
3.25
3.75
4.25
4.75
5.25
0
5
0.5
1.5
2.5
3.5
4.5
5.5
Figure 5.17-b:Distribution of Bias Factor for Settlement for D’Appolonia Method (1970)
117
Schmertmann Method
1 y = 0.5265x + 0.2193
R² = 0.3088
0.8
0.6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2
Figure 5.18-a: Predicted Versus Measured Settlements for Schmertmann Method (1978)
0.3
Schmertmann Method
0.25
Mean = 1.05
0.2 STDEV = 0.49
COV = 0.47
Frequency
n = 23
0.15
0.1
0.05
0
1.35
0.15
0.45
0.75
1.05
1.65
1.95
2.25
2.55
2.85
3.15
0
3
0.3
0.6
0.9
1.2
1.5
1.8
2.1
2.4
2.7
3.3
Figure 5.18-b:Distribution of Bias Factor for Settlement for Schmertman Method (1978)
118
Burland and Burbidge Method

0.8 y = -0.0917x + 0.533

R² = 0.0156
0.6
0.4
0.2
0
0.00 0.50 1.00 1.50 2.00
Figure5.19-a:Predicted Versus Measured Settlements for Burland and Burbidge Method
0.25
Burland and Burbidge Method

0.2
Mean = 1.77
STDEV = 1.49
COV = 0.83
Frequency
0.15
n = 23
0.1
0.05
0
1.8
5.1
0
6
0.3
0.6
0.9
1.2
1.5
2.1
2.4
2.7
3.3
3.6
3.9
4.2
4.5
4.8
5.4
5.7
6.3
6.6
Figure 5.19-b: Distribution of Bias Factor for Settlement for

Burland and Burbidge Method (1985)
119
5.7 DISTRIBUTION OF SETTLEMENT
To incorporate model uncertainties in the reliability analyses, the assumption of
lognormality distribution to the measured or predicted settlements, or to bias factor of
settlement, must be established.
5.7.1 Distribution of Measured and Predicted Settlements
This section will present the type of distribution for measured settlements as well as
predicted settlements by the six selected methods. It can be seen from the plot of the measured
values of settlement for the 23 sites on the normal probability paper, as shown in Figure 5.20,
that the distribution looks more lognormal rather than normal, since its coefficient of
correlation (R2) is closer to 1. The construction and use of the normal probability paper is
described in Chapter 3, section 3.4. It is a convenient way to present cumulative distribution
functions (CDF), as it allows for an easy evaluation of the most important statistical
parameters as well as type of distribution function. The horizontal axis represents the basic
variable. In the case of the considered test data, it is the value of either measured or predicted
settlement. Vertical axis is the inverse normal probability scale, and it represents the distance
from the mean value in terms of standard deviations. The vertical coordinate can also be
considered as the probability of exceeding the corresponding value of the variable. For any
value of settlement (horizontal axis), the vertical coordinate of CDF corresponds to a certain
probability of being exceeded. For example, value of -1 on the vertical scale corresponds to a
probability of 0.159 that the indicated settlement will be exceeded.

120
The shape of CDF is an indication of the type of distribution. A straight line means that
the distribution is normal, and a curved line indicates another distribution, that maybe
lognormal or other types of distributions as mentioned in Chapter 3, section 3.2.
CDF of Measured Settlements

for 23 sites
3
y = 1.8876ln(x) + 1.5138
2
R² = 0.9655
0
0 0.2 0.4 0.6 0.8 1 1.2
-1
-2
-3
Measured Settlement
Figure 5.20: CDF of Measured Settlements for 23 Sites
The CDF’s presented in Figures 5.21 through 5.26 for predicted values of settlement of
the six prediction methods for all the 23 sites of spread footings also appear to have a
lognormal distribution rather than normal. The coefficient of correlation for Hough, Peck and
Bazarra, D’Appolonia, and Shmertmann methods had a value greater than 0.95. The
Meyerhof method had a value of 0.88 and Burland and Burbidge method had a value of 0.93.
All produced good correlation with lognormal distribution for settlement.

121
CDF of Predicted Settlements

by Hough Method
3
y = 2.702ln(x) + 0.3881
2 R² = 0.9712
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8
-1
-2
-3
Predicted Settlement
Figure 5.21: CDF of Predicted Settlements by Hough Method for 23 sites

by Meyerhof Method
3
y = 2.0825ln(x) + 2.4592
2 R² = 0.8846
0
0 0.2 0.4 0.6 0.8 1 1.2
-1
-2
-3
Figure 5.22: CDF of Predicted Settlements by Meyerhof Method for 23 sites

122

by Peck and Bazarra Method
y = 1.6099ln(x) + 1.7525
2
R² = 0.9846
0
0.00 0.20 0.40 0.60 0.80 1.00 1.20
-1
-2
-3
Figure 5.23: CDF of Predicted Settlements by Peck and Bazaraa Method for 23 sites

by D'Appolonia Method
3
y = 2.4736ln(x) + 2.3527
2
R² = 0.9888
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
-1
-2
-3
Figure 5.24: CDF of Predicted Settlements by D’Appolonia Method for 23 sites

123

by Schmertmann Method
3
y = 1.9259ln(x) + 1.4407
R² = 0.9723
0
0 0.2 0.4 0.6 0.8 1 1.2
-1
-2
-3
Figure 5.25: CDF of Predicted Settlements by Schmertmann Method for 23 sites

by Burland and Burbidge Method
3
y = 1.5153ln(x) + 1.665
R² = 0.9334
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8
-1
-2
-3
Figure 5.26: CDF of Predicted Settlements by Burland and Burbidge Method for 23 sites
124
5.7.2 Distribution of Measured to Predicted Settlement Ratios
As mentioned above, the assumption of lognormality of the ratio of measured to
predicted settlements or for bias factor for settlement should be established for the reliability
analyses. However, it should be noted that the professional factor which is another name for
bias factor of settlement would have statistics that are joined and not only include the model
uncertainties but also other uncertainties, which are assumed to be the minimum values of
measurement error and inherent variability.
Moreover, even if the professional factor would have deterministic parameters, which
means, mean = 1.0 and standard deviation = 0.0, the predicted settlements would still have
some uncertainty and inherent variability. In other words, by considering any settlement
prediction method that correlates an in-situ test results with settlements and assuming that the
predicted method calculates the settlement perfectly, i.e., (δM /δP) = 1.0, even though the
measured settlements would still be different from the predicted settlements due to inherent
variability of the soil and the measurement error. The in-situ test results used in predicted
settlement are represented in SPT N-values or Es, which is unlikely to represent the soil
conditions of the whole site. The critical situations are when the measured values significantly
exceed analytically the predicted values.
For the six prediction methods of settlement, the settlement ratio (δM /δP) is plotted on
the normal probability paper with its both normal and lognormal distributions, as shown in
Figures 5.27 through 5.32. From inspection of these plots, it can be noticed that the lognormal
distribution is a reasonable assumption for the distribution of the bias factor or settlement ratio
(δM /δP).
125
CDF of (δM/δP) by Hough Method

3
0
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4
-1
Settlement ratio
-2
Normal
Lognormal
-3
Settlement Ratio (δM/δP)
Figure 5.27: CDF of (δM/δP) for 23 sites by Hough Method (1959)
CDF of (δM/δP) by Meyerhof Method

3
0
0.0 1.0 2.0 3.0 4.0 5.0 6.0
-1
Settlement ratio
-2
Normal
Lognormal
-3
Figure 5.28: CDF of (δM/δP) for 23 sites by Meyerhof Method (1965)

126
CDF of (δM/δP) by Peck & Bazaraa Method

4
3
0
0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0
-1
-2
Settlement ratio
-3 Normal
Lognormal
-4
Figure 5.29: CDF of (δM/δP) for 23 sites by Peck and Bazaraa Method (1969)
CDF of (δM/δP) by D'Appolonia Method

4
3
0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
-1
-2
Settlement ratio
-3 Normal
Lognormal
-4
Figure 5.30: CDF of (δM/δP) for 23 sites by D’Appolonia Method (1970)

127
CDF of (δM/δP) by Schmertmann Method

4
3
0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
-1
-2
Settlement ratio
-3 Normal
Lognormal
-4
Figure 5.31: CDF of (δM/δP) for 23 sites by Schmertmann Method (1978)
CDF of (δM/δP) by Burland & Burbidge Method

3
0
0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0
-1
Settlement ratio
-2 Normal
Lognormal
-3
Figure 5.32: CDF of (δM/δP) for 23 sites by Burland and Burbidge Method (1985)
128
It can be seen from the above figures for Settlement Ratio (δM/δP) that the Hough,
Meyerhof, and Burland and Burbidge methods seem not to have a perfect lognormal
distribution, but it fits better than normal distribution. While Peck and Bazaraa, D’Appolonia,
and Schmertmann methods seem to fit a lognormal distribution rather than a normal
distribution. Therefore, the lognormal distribution would be a reasonable assumption for these
methods rather than normal assumption.
The CDF’s in Figure 5.33 for the settlement ratio of all the six prediction methods are
not normal but rather lognormal. The critical situations are when the measured values
significantly exceed analytically predicted values. In minimizing critical situations, the Hough
and Schmertmann methods seem to produce better results than other methods. In addition, it
appears that the CDF’s indicate that none of them can be considered as a normal distribution.
CDF of Settlement Ratio (δM/δP)

3
2
0
0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0
Hough
-1
Meyerhof
Peck&Bazaraa
-2 D'Appolonia
Schmertmann
Burland&Burbidge
-3
Figure 5.33: CDF of (δM/δP) for 23 sites by all Six Prediction Methods
129
5.8 INFLUENCE OF VARIOUS FACTORS ON SETTLEMENT
In this section, the influences of foundation size, foundation length-to-width ratio,

foundation depth, and the depth of water table on the settlement prediction are investigated
statistically using the data base of all the 23 spread footing sites. Also, using the Schmertmann
method for prediction settlement, a random site was selected, and checked for the effect of
predicted settlement of increasing these factors.
5.8.1 Influence of Footing Size (B)
Figure 5.34 shows the footing width (B) and the related settlements for all 23 sites.
There is a large scatter in the data and there is a slight trend for an increase of settlement with
footing width, although the coefficient of correlation is small because of all other factors
related to settlement. The influence of footing size on predicted settlement (Schmertmann
method) of one footing is shown in Figure 5.35 by increasing the width of the footing for one
spread footing (site S1). The predicted settlement decreased with increasing width up to 16
feet. Above 16 feet, the settlement decrease was not significant.
Influence of Footing Width (B) on Settlement

1.2
1 y = 0.0164x + 0.2193
R² = 0.137
n = 23
Settlement (in.)
0.8
0.6
0.4
0.2
0
0 5 10 15 20 25 30
B (ft)
Figure 5.34: Influence of Footing width (B) on Settlement on Settlement of Shallow

Foundations on Cohesionless Soils
130
Effect of Increasing Footing Width on

Settlement
100
90
80
70
Settlementδ (in.)
60
50
40
30
20
10
0
0 2 4 6 8 10 12 14 16 18 20 22
B (ft)
Figure 5.35: Effect of Increasing Footing width (B) on Settlement
Another factor to be considered related to footing width is the square root of footing width,
as Meyerhof (1965) suggested to take into consideration. Figure 5.36 shows that the square
root of footing width had a similar trend to that of footing width with settlement, which was to
be expected.
Influence of Square Root of Footing Width on Settlement

1.2
1 y = 0.1271x - 0.0202
R² = 0.1337
n = 23
Settlement (in.)
0.8
0.6
0.4
0.2
0
0 1 2 3 4 5 6
B ^ 0.5 (ft)
Figure 5.36: Influence of Square Root of Footing width (B) on Settlement on Settlement
131
5.8.2 Influence of Length-to-Width Ratio (L/B)
Figure 5.37 shows the length-to-width ratio (L/B) for all the 23 sites and their
related settlements. There is a large scatter in the data, and no definite trend is noticed. There
is a very small correlation coefficient due to variability of all other factors related to
settlement. The influence of L/B on predicted settlement (Schmertmann method) is
demonstrated in Figure 5.38 by increasing the ratio (L/B) of the footing for site S1. The
settlement decreased up to a ratio to 8, after which it had no significant effect on the
settlement.
Influence of L/B on Settlement

1.2
1 y = -0.0105x + 0.5328
R² = 0.006
n = 23
Settlement (in.)
0.8
0.6
0.4
0.2
0
0 1 2 3 4 5 6 7 8
L/B
Figure 5.37: Influence of Length to Width Ratio (L/B) on Settlement of Shallow

132
Effect of Increasing L/B on Settlement

5
4
Settlementδ (in.)
0
0 1 2 3 4 5 6 7 8 9 10 11
L/B
Figure 5.38: Effect of Increasing Length to Width Ratio (L/B) on Settlement
5.8.3 Influence of Footing Depth Ratio (Df/B)
It can be noticed in Figure 5.39 that, for all the 23 sites, there is a slight trend of
decreasing in settlement with an increase in the ratio (Df/B). A very low correlation
coefficient is due to the variability with other factors related to settlement. However, the
results are not convincing enough to suggest a firm correlation between the depth ratio (Df/B)
and the settlement.
A second plot is shown in Figure 5.40 to evaluate whether increasing the depth ratio
would improve the results and decrease the predicted settlement (Schmertmann method). It
can be seen that by increasing the depth ratio, there is not a very significant decrease in the
predicted settlement, especially after the depth ratio of more than 0.6.
133
Influence of Df/B on Settlement

1.2
y = -0.1603x + 0.5437
1
R² = 0.0383
Settlement (in.)
n = 23
0.8
0.6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1 1.2
Df/B
Figure 5.39: Influence of Depth Ratio (Df /B) on Settlement of Shallow Foundations on
Cohesionless Soils
Effect of Increasing Df/B on Settlement

5
4
Settlementδ (in.)
0
0 0.2 0.4 0.6 0.8 1 1.2
Df/B
Figure 5.40: Effect of Increasing Depth Ratio (Df/B) on Settlement

134
5.8.4 Influence of Depth of Water Table (Dw/B)
Some methods take into account the depth of water table level, specifically; such as
Terzaghi and Peck (1948), and Peck and Bazarra (1969), while, others, such as Meyerhof
(1965) and D’Appolonia (1970), suggest the effect of water table is already included in the
SPT N-values. Obviously, there is a disagreement about the effect of water table depth on
settlement. Using the compiled data base for the 23 spread footings, the influence of water
table depth on settlement is shown in Figure 5.41. It can be seen, that there is a large scatter
in the data, and no definite trend is noticed with, evidenced by the very small coefficient of
correlation. In Figure 5.42, it can be seen, that by increasing the water table level, there is no
significant decrease in the settlement. Therefore, it can be concluded that, considering the high
variability, the level of water table has no significant effect on settlement. This means the
effect of the water table is already reflected in the SPT N-values. Moreover, this conclusion is
also supported by the study results of Burland and Burbidge (1985).
Influence of Dw/B on Settlement

1.2
1 y = -0.0251x + 0.5187
R² = 0.0118
Settlement (in.)
0.8 n = 23
0.6
0.4
0.2
0
0 0.5 1 1.5 2 2.5 3 3.5 4
Dw/B
Figure 5.41: Influence of Water Table Depth Ratio (Dw /B) on Settlement of Shallow
135
Effect of Increasing Dw/B on Settlement

5
3
settlementδ (in.)
0
0 0.5 1 1.5 2 2.5 3 3.5
Dw/B
Figure 5.42: Effect of Increasing the Water Table Depth Ratio (Dw/B) on Settlement
5.8.5 Influence of Axial Load (P)
Using the compiled data base for the 23 spread footings, the influence of applied loads
on settlement is shown in Figure 5.43. It can be seen that even with the scatter in the data,
there is a correlation coefficient larger than that for the factors considered above.
Influence of Load (P) on Settlement

1.2
y = 0.1565ln(x) - 0.6414
1
R² = 0.2122
n = 23
Settlement (in.)
0.8
0.6
0.4
0.2
0
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
P (kips)
Figure 5.43: Influence of Applied Loads (P) on Settlement of Shallow Foundations on

Cohesionless Soils
136
Figure 5.44 shows the effect on predicted settlement (Schmertmann method) of increasing
the axial load. The plots in the figure show that increasing the load for two randomly selected
sites (S1 and S2) will increase the settlement. All the graphs for the selected sites take the
form of Curve C, explained previously in section 5.4.
Effect of Increasing Load on Settlement

14000
12000
10000
P (kips)
8000
S1-site
6000
S2-Site
4000
2000
0
0 1 2 3 4 5 6
Settlement (in.)
Figure 5.44: Effect of Increasing the Axial Load (P) on Settlement

for Two Different Sites
5.8.6 Influence of SPT N-Values
The influence of the soil properties quantified in the SPT N-values on settlement for
the 23 sites is shown in Figure 5.45. There is a slight trend for an increase in settlement with
N-value although the correlation coefficient is small. It can be seen in Figure 5.46 that with
increasing the SPT N-values the predicted settlement decreases (Schmertmann method), but
above N60 = 50, there is no significant decrease in settlement.
137
Influence of SPT N-Values on Settlement

1.2
y = 0.0038x + 0.3796
1 R² = 0.0547
n = 23
Settlement (in.)
0.8
0.6
0.4
0.2
0
0 10 20 30 40 50 60 70
SPT N-Values
Figure 5.45: Influence of Corrected SPT N-Values on Settlement of Shallow

Effect of Increasing SPT N-Values on

Settlement
5
4
Settlementδ (in.)
0
0 10 20 30 40 50 60 70 80
SPT N-Values
Figure 5.46: Effect of Increasing the SPT N-Values (N60) on Settlement

138
5.9 ESTIMATION OF DIFFERENTIAL SETTLEMENT AND ANGULAR

DISTORTION OF SHALLOW FOUNDATIONS ON COHESIONLESS SOILS
5.9.1 Differential Settlement
Both total and differential settlement evaluation is an essential part of the
geotechnical design process. Usually the settlement criterion controls the design process for
shallow foundations on cohesion less soils rather than bearing capacity failure.
It is reasonable to expect variation of settlement among different footings of a
bridge, due to the variation of load carried by various shallow foundations and the non-
homogeneity of the soil under the footing.
Although equally loaded footings for a single bridge may be designed for equal
settlement, the cohesionless foundation soil may have variation and uncertainties in
geotechnical properties across the site that will cause differential settlements. In addition, for
settlement prediction, the methods that are used are empirical, based on correlations between
the observed settlements and in-situ test results, which, in themselves, could cause
uncertainties in the settlement estimations and measurement errors. Therefore, predicted
settlements can be expected to differ from measured settlements, and, more importantly, from
differential settlements. Differential settlements are more critical than uniform settlement,
because differential settlements can cause the distortion and damage to the superstructure. All
definitions about different types of settlement are outlined in sections 5.2 and 5.3 with more
details.
139
Foundation movements must be tolerated by the superstructure, so that
performance of the structure is not affected. This requires an accurate estimation for the
expected and allowable values of total and differential settlements for a particular site and
structure. The allowable differential settlement value is a function of many factors, such as
the material properties, soil conditions, structural connections, foundation shape and size, and
and consideration of the size of the superstructure. Consequently, the value of differential
settlement cannot be easily determined.
In this section, differential settlements are calculated for the 12 highway bridges
selected in this study, but there are some sites excluded due to technical problems or lack of
multiple sites for a single bridge, so only 20 sites were used for calculating differential
settlement. The determined differential settlement is based on the procedure described in
section 5.3, with allowable values for differential settlement. These values are presented in
Table 5.6, which contains actual differential settlements, based on measured settlements and
on differential settlements calculated by the six methods. Statistics of differential settlement
are provided in Table 5.7. It can be concluded from this table that the Hough and
Schmertmann methods produced a value of average differential settlement (0.25 in.) closest to
the actual differential settlement (0.23 in.). Even though the D’Appolonia method resulted in
the lowest value for COV (0.36), its average predicted differential settlement was very low
(0.11 in.) compared to the actual value. Therefore, the Hough and Schmertmann methods are
the two prediction methods recommended for calculating differential settlement on
cohesionless soils for shallow foundations.

140
Table 5.6: Comparison of Actual and Calculated Values of Differential Settlement
Differential Settlement Values (inch)

Location Actual Calculated (δDP)
(δD) Hough Meyerhof Peck and D’Appolonia Schmertman Burland
Bazaraa and
Burbidge
Bridge1- S1 0.32 0.20 0.10 0.12 0.16 0.14 0.10
Bridge 1-S2
Bridge2- S3 0.33 0.30 0.27 0.04 0.05 0.51 0.47
Bridge2- S5 0.15 0.31 0.03 0.10 0.10 0.32 0.09
Bridge2- S4
Bridge3-S6 0.15 0.06 0.16 0.26 0.10 0.06 0.32
Bridge3-S8,S9 0.0 0.21 0.03 0.20 0.08 0.04 0.05
Bridge3-S10,S11
Bridge 4-S14 0.12 0.47 0.66 0.12 0.17 0.44 1.22
Bridge 4-S15
Bridge 5-S16 0.21 0.11 0.13 0.41 0.09 0.15 0.2
Bridge 5-S17
Bridge 8-S21 0.20 0.25 0.07 0.32 0.10 0.42 0.29
Bridge 8-S22
Bridge 9-S23 0.33 0.44 0.08 0.09 0.13 0.22 0.14
Bridge 9-S24
Bridge 12-S26 0.26 0.12 0.11 0.13 0.09 0.19 0.13
Bridge 12-S27
141
Table 5.7: Statistical Parameters of Differential Settlement for the

Mean Standard Coefficient

Method Deviation of Variation
(STDEV) (COV)
Hough (1959) 0.25 0.14 0.56
Meyerhof(1965) 0.16 0.19 1.14
Peck & Bazaraa(1969) 0.18 0.12 0.65
D'Appolonia(1970) 0.11 0.04 0.36
Schmertmann(1978) 0.25 0.16 0.66
Burland & Burbidge(1985) 0.30 0.35 1.16
Actual Differential Settlement 0.23 0.08 0.36
Average for all Methods 0.21 0.13 0.75
Therefore, the Hough and Schmertmann methods will be selected in this study to check
the distribution of differential settlement and to calculate reliability indices. The distribution
of the actual differential settlement is presented in Figure 5.47. The CDF of the actual
differential settlement is presented in Figure 5.48, in which the distribution looks lognormal
with a high correlation coefficient (R2 =0.95). Figures 5.49 and 5.51 show the distribution of
differential settlement calculated using the two prediction methods selected in this section.
The CDF’s of differential settlement by the Hough and Schmertmann methods are presented
in Figures 5.50 and 5.52 respectively. Also, it can be concluded from the figures that the
distribution is lognormal with high coefficient of correlations R2 =0.95 for differential
settlement by the Hough method and R2 =0.94 for the Schmertmann Method.
142
0.35
0.3
Mean = 0.23
STDEV = 0.08
0.25
COV = 0.36
Frequency
0.2
0.15
0.1
0.05
0.24
0.17
0.31
0.38
0.135
0.205
0.275
0.345
0.1
Differential Settlement
Figure 5.47:Distribution of Actual Differential Settlement for 20 sites of Spread Footings
CDF of Actual Differential Settlement

1.5
y = 2.065ln(x) + 3.1568
1
R² = 0.945
0.5
0
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40
-0.5
-1
-1.5
Differential Settlement (in.)
Figure 5.48: CDF of Actual Differential Settlement for 20 sites of Spread Footings
143
0.25
0.2 Mean = 0.25

STDEV = 0.136
COV = 0.55
0.15
Frequency
0.1
0.05
0
0.11
0.22
0.33
0.44
0.055
0.165
0.275
0.385
0.495
0
Figure 5.49: Distribution of Differential Settlement by Hough Method (1959)
CDF by Hough Method

1.5
1 y = 1.2415ln(x) + 1.9402
R² = 0.9528
0.5
0
0.00 0.10 0.20 0.30 0.40 0.50 0.60
-0.5
-1
-1.5
-2
Figure 5.50: CDF of Differential Settlements by Hough Method for 20 sites

144
0.25
0.20 Mean = 0.25

STDEV = 0.165
COV = 0.66
0.15
Frequency
0.10
0.05
0.00
0.06
0.12
0.18
0.24
0.36
0.42
0.48
0.54
0
0.3
Figure 5.51: Distribution of Differential Settlement by Schmertman Method (1978)
CDF by Schmertmann Method

1.5
1 y = 0.9353ln(x) + 1.5487
R² = 0.9445
0.5
0
0.00 0.10 0.20 0.30 0.40 0.50 0.60
-0.5
-1
-1.5
-2
Figure 5.52: CDF of Differential Settlements by Schmertmann Method for 20 sites

145
As mentioned previously for the context of Load and Resistance Factor Design (LRFD),
accuracy refers to bias factor, defined as the ratio of the mean value of the actual differential
settlement to the value of calculated differential settlement (δD/ δDp). Values of bias factor
for differential settlement are presented in Table 5.8 for the Hough and Schmertmann
methods, since theses will be the two methods used for calculating differential settlement. In
addition, Table (5.9) contains the statistical parameters including mean value (µ), standard
deviation (STDEV), and coefficient of variation (COV) of the bias factor for differential
settlement of these two selected methods.
Table 5.8: Bias Factor for Differential Settlement (δD/δDP) of the Field Data
Location Bias Factor (δD/δDP)

Hough Schmertmann
Method Method
Bridge1- S1 1.63 2.29
Bridge 1-S2
Bridge2- S3 1.08 0.65
Bridge2- S5 0.49 0.47
Bridge2- S4
Bridge3-S6 2.5 2.5
Bridge3-S8,S9 0.0 0.0
Bridge3-S10,S11
Bridge 4-S14 0.26 0.27
Bridge 4-S15
Bridge 5-S16 1.94 1.40
Bridge 5-S17
Bridge 8-S21 0.81 0.48
Bridge 8-S22
Table 5.9, Continued 146
Bridge 9-S23 0.74 1.5

Bridge 9-S24
Bridge 12-S26 2.17 1.37
Bridge 12-S27
Table 5.9: Statistical Parameters of Bias Factor for Differential Settlement (δD/δDP) of
the Two Selected Methods
Statistical Hough Schmertmann

Parameters Method Method
Max 2.5 2.5
Min 0.0 0.0
Mean 1.16 1.09
STDEV 0.85 0.85
COV 0.73 0.78
Figures 5.53 and 5.55 show graphs for the variation involved in the bias factor for
differential settlement in terms of frequency histograms to develop a better understanding of
the corresponding variation for both the Hough and Schmertmann methods, respectively, for
the 20. Figures 5.54 and 5.56 present the CDF’s of bias factor of differential settlement
predicted by the two selected prediction methods. It can be concluded for both methods that
bias factor for differential settlement follows a lognormal distribution rather than a normal
distribution.
147
0.25
Hough Method
0.2
Mean = 1.16
STDEV = 0.85
Frequency
0.15 COV = 0.73
0.1
0.05
0
0.43
0.76
1.09
1.42
1.75
2.08
2.41
2.74
0.1
Bias Factor of Differential Settlement
Figure 5.53: Distribution of Bias Factor for Differential Settlement by

Hough Method (1959)
CDF of (δD/δDP) by Hough Method

3
2 y = 1.0532ln(x) - 0.0395
R² = 0.9502
0
0 0.5 1 1.5 2 2.5 3
-1
-2
-3
Figure 5.54: CDF of Bias Factor for Differential Settlement for 20 sites
by Hough Method (1959)
148
0.35
0.30 Schmertmann Method

Mean = 1.09
0.25 STDEV = 0.85
COV = 0.78
Frequency
0.20
0.15
0.10
0.05
0.00
0.31
0.62
0.93
1.24
1.55
1.86
2.17
2.48
2.79
0
Figure 5.55: Distribution of Bias Factor for Differential Settlement by

Schmertmann Method (1978)
CDF of (δD/δDP) by Schmertmann Method

3
y = 1.0184ln(x) + 0.0493
2
R² = 0.9409
0
0 0.5 1 1.5 2 2.5 3
-1
-2
-3
Figure 5.56: CDF of Bias Factor for Differential Settlement for 20 sites
by Schmertmann Method (1978)
149
5.9.2 Angular Distortion
Angular distortion can be defined as the differential settlement between two neighboring
footings, divided by the distance between the two sites. More details about the concept and
components of angular distortion in bridges and their tolerable movement criteria are
reviewed in section 5.3.
The maximum angular distortion for a regular settlement pattern would be the same for
every pair of footings. However, the maximum angular distortion for an irregular settlement
pattern can be found anywhere through the foundation system.
In this section, angular distortions are calculated for determined differential settlements
for the 20 sites in this study. These values are presented in Table 5.10, which contains actual
angular distortion (AD) calculated based on actual differential settlements and predicted
settlements by the methods. Statistical parameters of angular distortion values are provided in
Table 5.11.
Table 5.10: Comparison of Actual and Calculated Values of Angular Distortion
Angular Distortion Values

Location Actual Calculated (ADP)
(AD) Hough Meyerhof Peck and D’Appolonia Schmertman Burland
Bazaraa and
Burbidge
Bridge1- S1 0.00019 0.00012 5.73E-05 7.35E-05 9.79E-05 8.33E-05 6.11E-05
Bridge 1-S2
Bridge2- S3 0.00025 0.00023 0.00020 2.84E-05 3.62E-05 0.00034 0.00035
Bridge2- S5 0.00011 0.00023 2.44E-05 7.43E-05 7.58E-05 0.00024 6.82E-05
Bridge2- S4
Bridge3-S6 0.00009 0.00004 0.00010 0.00016 6.22E-05 3.7E-05 0.00019
Bridge3-S8,S9 0.0 0.00014 2.08E-05 0.00014 5.24E-05 2.78E-05 3.47E-05
Bridge3-S10,S11
Table 5.10, Continued
150
Bridge 4-S14 0.00005 0.00019 0.00028 5.14E-05 7.26E-05 0.00019 0.00052

Bridge 4-S15
Bridge 5-S16 0.00013 0.00007 8.11E-05 0.00025 5.56E-05 9.26E-05 0.00012
Bridge 5-S17
Bridge 8-S21 0.00008 0.00011 3.03E-05 0.00014 4.27E-05 0.00018 0.00013
Bridge 8-S22
Bridge 9-S23 0.00020 0.00026 4.82E-05 5.47E-05 7.71E-05 0.00013 8.33E-05
Bridge 9-S24
Bridge 12-S26 0.00022 0.00008 7.64E-05 9.03E-05 6.25E-05 0.00013 9.03E-05
Bridge 12-S27
Table 5.11: Statistical Parameters of Angular Distortion for the

Mean Standard Coefficient

Method Deviation of Variation
(STDEV) (COV)
Hough (1959) 0.000148 7.8E-05 0.53
Meyerhof(1965) 9.24E-05 8.5E-05 0.92
Peck&Bazaraa(1969) 0.000106 6.7E-05 0.63
D'Appolonia(1970) 6.35E-05 1.8E-05 0.29
Schmertmann(1978) 0.000150 0.00011 0.71
Burland&Burbidge(1985) 0.000166 0.00015 0.93
Actual Angular Distortion 0.000147 7E-05 0.47
Average for all Methods 0.000121 4.5E-05 0.37
151
5.10 TOLERABLE SETTLEMENT CRITERIA
The task of determining the allowable displacement limit for a structure is very
complex and difficult. These limits are usually determined depending on serviceability
requirements rather than structural ones. This is because these displacements can affect the
structure appearance or performance, and can cause damage that may occur before the
ultimate limit state (ULS) is reached. Determining the serviceability limit state (SLS) is more
subjective than the determining the ultimate limit state (ULS), due to the dependence of many
factors for the allowable settlement, such as the structure function and type, the consequences
of damages, and the influence of all these factors by personal opinion.
Various allowable settlement criteria that have been established, are functions of total
or differential settlement. Typically, spread footings have a common value for allowable
settlement as 1 inch (25 mm), reported by Zekkos (2004). A typical tolerable settlement value
for highway bridges is considered to be 2 inches (50 mm), which was reported by Coduto
(2000). However, all of these limits are decreased frequently due to the possibility of
differential settlements.
The ratio of differential settlement to total settlement is often much higher in sands
than in clays because of the high inherent variability in the compressibility characteristics in
sand deposits, this fact was indicated by the results of Bjerrum’s study (1963).
Considering the many sources of uncertainties, a common empirical rule was proposed
by Holtz (1991), which stated that the differential settlement is equal to about one-half of the
predicted total settlement. However, Terzaghi et al. (1996) suggested for footings on sand,
that the differential settlement would not exceed the 75 percent of the estimated total
settlement.
152
CHAPTER 6
RELIABILITY ANALYSIS OF SETTLEMENT FOR

SHALLOW FOUNDATONS IN COHESIONLESS SOILS
6.1 INTRODUCTION
This chapter will present a reliability analysis for settlement of shallow foundations for
highway bridges on cohesionless soils. The main objective of this study was to develop the
methodology of reliability-based design of shallow foundations on cohesionless soils. The
objective was addressed by the evaluation of six selected prediction methods to estimate the
total and differential settlements through a sensitivity analysis.
6.2 RELIABILITY ANALYSIS PROCEDURE
A reliability analysis was performed to verify total and differential settlements of shallow
foundations for highway bridges on cohesionless soils. The reliability procedure includes the
following steps:
1. Representative highway bridges were selected. Data were obtained, including soil
parameters, applied loads, and measured settlements. Twelve highway bridges were
selected in this study. Details and information about the bridges are presented in
chapter 4, and Appendices B and C.

153
2. The limit state function was defined as the tolerable settlement, and was larger than the
predicting settlement.
3. The nominal resistance (R) was assigned as the tolerable settlement (δT), which was
treated as a deterministic value.
4. The load model (Q) was based on the prediction settlement (δP), evaluated by the six
prediction methods selected in this study. Both Excel and MATLAB software were
used for the prediction analysis, and then statistical parameters for settlement were
obtained. A Monte Carlo simulation method was used for predicting settlement, and
identifying the distribution of settlement. The settlement was determined to be
lognormally distributed.
5. Reliability indices were calculated to assess how they function in the selected
prediction methods of settlement. The reliability analysis was performed for different
sites of spread footings with different applied loads and soil parameters.
6.3 LIMIT STATE FUNCTION
The limit state function is formulated for the considered structural components and is
given by the following equation:
g(δT,δP) = (δT/ δP) ≥ 1.0 (6.1)
in which, δT = tolerable settlement (inch), δP = predicted settlement (inch).
This assumption for the limit state function is valid for total and differential settlements.
154
6.4 NOMINAL RESISTANCE AND LOAD MODEL
The nominal resistance (R) was assigned as the tolerable settlement (δT), which is
assumed to be treated as a deterministic value with λ = 1.0 and COV = 0.0.
The model load (Q) was assigned to each of the prediction methods to obtain the
predicted settlement (δP) for each method. The Monte Carlo method was used by MATLAB
code with 100,000 simulations for each of the 23 sites of foundations. For each site, the
simulation resulted in many figures describing the geometry, applied load, soil parameters,
and total settlement determined. More details about Monte Carlo method are presented in
Chapter 3, section 3.7.2. In addition, Appendix D presents graphs for CDF’s for site-S1 for
the Schmertmann method as the best considered method, the Meyerhof method, and the Peck
and Bazarra method. It can be concluded from the CDF’s of settlement that the distribution of
settlement looks lognormal rather than normal.
6.5 RELIABILITY INDEX CALCULATIONS
In this study, the reliability analysis is based on the calculations of the reliability index, β.
Structural performance is measured in terms of the reliability index, β, which is defined as the
function of the probability of failure, Pf = P(g<1.0).
The equation to calculate, β, can be defined as follows depending on the assumption of
lognormal distribution:
155
ln(T )  ln(P )
 (6.2)
 2ln(T )   2ln(P )
Since δT, is treated as deterministic µln( δT) = ln (δTnominal) , and σ2 ln( δT) =0.0
σ2 ln( δP) = ln ( 1 + VδP2 ) (6.3)
µln( δP) = ln (µ δP ) – 1/2 σ2 ln( δP) (6.4)
in which, µδP = mean value of predicted settlement, VδP = coefficient of variation for
predicted settlement, δTnominal = nominal value of tolerable settlement.
Tolerable settlement values will be considered in the range between 0.25 and 3.0
inches. These will be considered as nominal values, and will be treated as deterministic
values. However, mean values of predicted settlement would depend on nominal values of
predicted settlement (calculated by the methods), and bias factor of settlement, depending on
the predicted method with its related coefficient of variation. These values can be obtained
from Table 5.5 in chapter 5.

156
6.5.1 Reliability Indices for Total Settlement
Reliability indices determined for total settlement by the six prediction methods for the
23 sites of footings, depending on the range of tolerable settlement values as mentioned
before, are calculated from Equation 6.2. Reliability indices are shown in Figures 6.1 through
6.23 for all of the selected sites.
Reliability Index for Total Settlement for S1-Site
5.000
4.000 Hough
3.000
Reliability Index, β
Meyerhof
2.000
Peck&Bazaraa
1.000
0.000 D'Appolonia
-1.000
Schmertmann
-2.000
Burland&Burbidge
-3.000
0 0.5 1 1.5 2 2.5 3 3.5
Tolerable Settlement (inch)
Figure 6.1: Reliability Indices for Total Settlement by all Methods for S1-Site
157

6.00
5.00 Hough
4.00 Meyerhof
3.00
Peck&Bazaraa
2.00
D'Appolonia
1.00
0.00 Schmertmann
-1.00
Burland&Burbidge
-2.00
0 0.5 1 1.5 2 2.5 3 3.5

5.00
4.00 Hough
Reliabilty Index, β
3.00 Meyerhof
2.00 Peck&Bazaraa
1.00 D'Appolonia
0.00 Schmertmann
-1.00 Burland&Burbidge
-2.00
0 0.5 1 1.5 2 2.5 3 3.5
158

4.00
3.00 Hough
2.00 Meyerhof
1.00 Peck&Bazaraa
0.00 D'Appolonia
-1.00 Schmertmann
-3.00
0 0.5 1 1.5 2 2.5 3 3.5

6.00
5.00 Hough
4.00
Meyerhof
3.00
2.00 Peck&Bazaraa
1.00
D'Appolonia
0.00
Schmertmann
-1.00
-3.00
0 0.5 1 1.5 2 2.5 3 3.5
159

6.00
5.00 Hough
Reliability Index,β
4.00
Meyerhof
3.00
Peck&Bazaraa
2.00
1.00 D'Appolonia
0.00 Schmertmann
-2.00
0 0.5 1 1.5 2 2.5 3 3.5

6.00
5.00
Hough
4.00
3.00 Meyerhof
2.00 Peck&Bazaraa
1.00
D'Appolonia
0.00
-1.00 Schmertmann
-2.00
Burland&Burbidge
-3.00
0 0.5 1 1.5 2 2.5 3 3.5
160

6.00
5.00
Hough
4.00
Meyerhof
3.00
Peck&Bazaraa
2.00
1.00 D'Appolonia
0.00 Schmertmann
-2.00
0 0.5 1 1.5 2 2.5 3 3.5

6.00
5.00 Hough
4.00
Meyerhof
3.00
Peck&Bazaraa
2.00
D'Appolonia
1.00
0.00 Schmertmann
-2.00
0 0.5 1 1.5 2 2.5 3 3.5
161

6.00
5.00 Hough
4.00 Meyerhof
3.00
Peck&Bazaraa
2.00
D'Appolonia
1.00
0.00 Schmertmann
-2.00
0 0.5 1 1.5 2 2.5 3 3.5

6.00
5.00 Hough
4.00 Meyerhof
3.00
Peck&Bazaraa
2.00
D'Appolonia
1.00
0.00 Schmertmann
-2.00
0 0.5 1 1.5 2 2.5 3 3.5
162

5.00
4.00
Hough
3.00
Meyerhof
2.00
1.00 Peck&Bazaraa
0.00 D'Appolonia
-1.00 Schmertmann
-2.00
Burland&Burbidge
-3.00
0 0.5 1 1.5 2 2.5 3 3.5

4.00
3.00
Hough
2.00
Meyerhof
1.00
Peck&Bazaraa
0.00
-1.00 D'Appolonia
-2.00 Schmertmann
-3.00
Burland&Burbidge
-4.00
0 0.5 1 1.5 2 2.5 3 3.5
163
Reliability Index with Total Settlement for S16-Site

6.00
5.00 Hough
4.00 Meyerhof
3.00
Peck&Bazaraa
2.00
D'Appolonia
1.00
0.00 Schmertmann
-2.00
0 0.5 1 1.5 2 2.5 3 3.5

5.00
4.00 Hough
3.00 Meyerhof
2.00 Peck&Bazaraa
1.00 D'Appolonia
0.00 Schmertmann
-1.00
Burland&Burbidge
-2.00
0 0.5 1 1.5 2 2.5 3 3.5
164

5.00
4.00 Hough
3.00 Meyerhof
2.00
Peck&Bazaraa
1.00
D'Appolonia
0.00
Schmertmann
-1.00
Burland&Burbidge
-2.00
0 0.5 1 1.5 2 2.5 3 3.5

7.00
6.00
Hough
5.00
4.00 Meyerhof
3.00
Peck&Bazaraa
2.00
1.00 D'Appolonia
0.00
Schmertmann
-1.00
-3.00
0 0.5 1 1.5 2 2.5 3 3.5
165

5.00
4.00 Hough
3.00
Meyerhof
2.00
Peck&Bazaraa
1.00
0.00 D'Appolonia
-1.00 Schmertmann
-2.00
Burland&Burbidge
-3.00
0 0.5 1 1.5 2 2.5 3 3.5

4.00
Hough
3.00
2.00 Meyerhof
1.00 Peck&Bazaraa
0.00 D'Appolonia
-1.00 Schmertmann
-2.00
Burland&Burbidge
-3.00
0 0.5 1 1.5 2 2.5 3 3.5
166

5.00
4.00 Hough
Reliability index, β
3.00 Meyerhof
2.00 Peck&Bazaraa
1.00 D'Appolonia
0.00 Schmertmann
-1.00
Burland&Burbidge
-2.00
0 0.5 1 1.5 2 2.5 3 3.5

7.00
6.00 Hough
5.00
Meyerhof
4.00
3.00 Peck&Bazaraa
2.00 D'Appolonia
1.00
Schmertmann
0.00
-2.00
0 0.5 1 1.5 2 2.5 3 3.5
167

5.00
4.00 Hough
3.00
Meyerhof
2.00
Peck&Bazaraa
1.00
D'Appolonia
0.00
-1.00 Schmertmann
-3.00
0 0.5 1 1.5 2 2.5 3 3.5

5.00
4.00
Hough
3.00
Meyerhof
2.00
Peck&Bazaraa
1.00
D'Appolonia
0.00
Schmertmann
-1.00
0 0.5 1 1.5 2 2.5 3 3.5
168
It can be concluded, from the above figures that the Hough and Schmertmann methods
produced better results for reliability indices than other methods. In addition, the
Schmertmann method produced the best results among all the methods, in which it had the
highest values of reliability indices in 15 sites, while the Hough method had the highest values
of reliability indices in the oher 8 sites of footings.
To make it more clear, the average values of reliabilty indices for total settlement with
their related probability of failures for the 23 sites for the Hough and Schmertmann methods
are presented in Table 6.1 below:
Table 6.1: Average of Reliability Indices for Total Settlement, Related to Tolerable
Settlements and their Probability of Failures for 23 Sites
Tolerable Reliability Indices, β Probability of Failure, Pf

Settlement
(inch)
Hough Method Schmertmann Hough Method Schmertmann
Method Method
0.25 -1.23 0.48 0.8910 0.3160
0.5 0.25 2.03 0.4010 0.0212
0.75 1.12 2.94 0.1310 0.00164
1.0 1.73 3.58 0.0418 0.000172
1.25 2.21 4.08 0.0139 0.0000225
1.5 2.60 4.49 0.00466 0.00000356
1.75 2.90 4.83 0.00187 0.000000683

169
Table 6.1, Continued
2.0 3.21 5.13 0.0006640 0.000000145
2.25 3.46 5.40 0.000270 3.33E-08
2.5 3.68 5.63 0.000117 9.01E-09
2.75 3.89 5.84 0.0000501 2.61E-09
3.0 4.07 6.04 0.0000235 7.71E-10
6.5.2 Reliability Indices for Differential Settlement
Since the Hough and Schmertmann methods produced the best results for differential
settlement compared with actual differential settlement values, for the selected foundations,
these will be the methods considered for reliability indices calculations for differential
settlements.
Reliability indices determined for differential settlement by the two selected prediction
methods for the 20 selected sites of footings (yielding 10 measurements of differential
settlement), for the range of tolerable settlement values mentioned before, are calculated from
Equation 6.2. Reliability indices are shown in Figures 6.24 through 6.33 for the selected sites.
It can be concluded from the figures, that the Schmertmann method produced better
results for reliability indices than the Hough method for the selected foundations.
Furthermore, the Schmertmann method had the highest values of reliability indices for seven
of the locations, while the Hough method had the highest values of reliability indices in the
other three locations.

170
Reliability Index for Differential Settlement

6.000
5.000
4.000
3.000
Hough
2.000 Schmertmann
1.000
0.000
0 0.5 1 1.5 2 2.5 3 3.5

4.000
3.500
3.000
2.500
2.000
Hough
1.500
Schmertmann
1.000
0.500
0.000
-0.500
-1.000
0 0.5 1 1.5 2 2.5 3 3.5
171

4.500
4.000
3.500
3.000
2.500
2.000 Hough
1.500 Schmertmann
1.000
0.500
0.000
-0.500
0 0.5 1 1.5 2 2.5 3 3.5

7.000
6.000
5.000
4.000
Hough
3.000 Schmertmann
2.000
1.000
0.000
0 0.5 1 1.5 2 2.5 3 3.5
Figure 6.27: Reliability Indices for Differential Settlement for S6, S8,9-Sites in Bridge 3
172

8.000
7.000
6.000
5.000
Hough
4.000
Schmertmann
3.000
2.000
1.000
0.000
0 0.5 1 1.5 2 2.5 3 3.5
Figure 6.28: Reliability Indices for Differential Settlement for

S8, 9 and S10,11-Sites in Bridge 3

4.000
3.500
3.000
2.500
2.000
1.500 Hough
1.000 Schmertmann
0.500
0.000
-0.500
-1.000
-1.500
0 0.5 1 1.5 2 2.5 3 3.5
173

6.000
5.000
4.000
3.000
Hough
Schmertmann
2.000
1.000
0.000
0 0.5 1 1.5 2 2.5 3 3.5

4.500
4.000
3.500
3.000
2.500
Hough
2.000
Schmertmann
1.500
1.000
0.500
0.000
-0.500
0 0.5 1 1.5 2 2.5 3 3.5
174

5.000
4.000
3.000
2.000 Hough
1.000 Schmertmann
0.000
-1.000
-2.000
0 0.5 1 1.5 2 2.5 3 3.5

6.000
5.000
4.000
Hough
3.000
Schmertmann
2.000
1.000
0.000
0 0.5 1 1.5 2 2.5 3 3.5
Figure 6.33:Reliability Indices for Differential Settlement for S26,S27-Sites in Bridge 12

175
To more clearly demonstrate the reliability indices of differential settlement for the
foundation sites, the average values of reliabilty indices with their related probability of
failures for the selected sites for the Hough and Schmertmann methods are shown in Table 6.2
Table 6.2: Average of Reliability Indices for Differential Settlement, Related to

Tolerable Settlements and their Probability of Failures
Tolerable Reliability Indices, β Probability of Failure, Pf

Settlement
(inch)
Hough Method Schmertmann Hough Method Schmertmann
Method Method
0.25 0.38 1.09 0.352 0.138
0.5 1.44 2.09 0.0749 0.0183
0.75 2.05 2.68 0.0202 0.00368
1.0 2.49 3.1 0.00639 0.000968
1.25 2.83 3.42 0.00233 0.000313
1.5 3.11 3.68 0.000935 0.000117
1.75 3.34 3.91 0.000434 0.0000461
2.0 3.55 4.10 0.000193 0.0000207
2.25 3.73 4.27 0.0000957 9.77E-06
2.5 3.89 4.42 0.0000501 4.94E-06
2.75 4.03 4.56 0.0000279 2.56E-06
3.0 4.17 4.690 0.0000152 1.37E-06

176
6.6 TARGET RELIABILITY LEVELS
The acceptable safety levels can be expressed in terms of target reliability indices, in
which it should be established for various design requirements. The selection of the target
reliability level is a multi-specialty task involving a structural safety analysis, and an
economic analysis, which are the most two important factors. Generally, reliability indices
below the target reliability value, βT, are not acceptable.
Target reliability indices calculated for settlement of shallow foundations can
vary with the suggested values of tolerable total and differential settlements. In other words,
for each assumption of a tolerable settlement value for either total or differential settlements,
different values of target reliability indices can be obtained.
Reliability index calculations were performed in this study assuming total and
differential settlements as lognormally distributed. Two target reliability indices can be
indicated, one for total settlement and the other for differential settlement as shown below:
6.6.1 Target Reliability Level for Total Settlement
Reliability indices were calculated for total settlement of shallow foundations for the 23
sites in this study. The reliability indices were determined depending on the six selected
prediction methods. The Hough and Schmertmann methods were found to give best values for
reliability indices. The averages of the indices calculated for the 23 sites of these methods are
shown in Table 6.1. The selection of a target reliability index for total settlement depends on
the assumption of tolerable values of total settlement. AASHTO-LRFD code (2012) does not
assume a specific value for allowed total settlement. Zekkos (2004) reported a typical value
177
of allowable total settlement of 1 inch (25 mm) for buildings, and Coduto (2000) suggested
allowable total settlement of 2.0 inches (50 mm) for bridges. In addition, both the Ohio
Department Of Transportation (ODOT), and the Nebraska Department Of Roads (NDOR)
suggest allowable total settlement for bridges as 1.5 inches. The actual values of total
settlement for the bridges selected did not exceed 0.96 inch (24 mm). Therefore, the suggested
tolerable value for total settlement in this study is 1.5 inch (37.5 mm). As a result, a target
reliability index can be calculated, based on this assumption, by averaging the reliability
indices of the Hough and Schmertmann methods, as shown in Table 6.3 below:
Table 6.3: Target Reliability Levels for Total Settlement,

Related to Tolerable Settlements
Tolerable Reliability Indices, β Target Level

Settlement
(inch)
Hough Method Schmertmann Reliability Probability of
Method Index, βT Failure, (Pf)T
1.5 2.60 4.49 3.54 0.00020
Therefore, as a conclusion of this study, the target reliability index would be estimated,
rounded to the nearest 0.05 to be (βT =3.5) for total settlement with probability of exceeding
the limit around 0.02%, and related with allowable suggested value for total settlement of 1.5
inch.
178
6.6.2 Target Reliability Level for Differential Settlement
Reliability indices were calculated for differential settlement of shallow
foundations for the 20 sites of foundations (10 locations) on the 8 highway bridges selected in
this study. The reliability indices were determined only for the Hough and Schmertmann
methods, because the two methods produced to give values of calculated differential
settlement closest to the actual ones. The averages of the indices were calculated for these two
methods, as shown in Table 6.2. The selection of target reliability index for differential
settlement depends on the assumption of tolerable values of differential settlement. Even
though AASHTO code (2012) suggested allowable values of differential settlement for 100–
foot-long spans for bridges as 4.8 inch for continuous spans, and 9.6 inch for simple spans,
These are considered to be very high values. The actual values of differential settlement in this
study were estimated to be 0.23 inches, which is a small value. A common empirical rule,
reported by Holtz (1991), is that the differential settlement is equal to about one-half of the
predicted total settlement. Therefore, this indicates that the allowable differential settlement
can be indicated as 0.75 inch (18,75 mm), depending on the suggested value of allowable total
settlement of 1.5 inches. As a result, the target reliability index can be calculated depending
on this assumption by the average of the reliability indices of the Hough and Schmertmann
methods, as shown in the Table 6.4 below:

179
Table 6.4: Target Reliability Levels for Differential Settlement,

Related to Tolerable Settlements
Tolerable Reliability Indices, β Target Level

Settlement
(inch)
Hough Method Schmertmann Reliability Probability of
Method Index, βT Failure, (Pf)T
0.75 2.05 2.68 2.37 0.00889
Therefore, as a conclusion of this study, the target reliability index would be estimated,
rounded to the nearest 0.05 to be (βT =2.4) for differential settlement, with the probability of
exceeding the limit around 0.8 %, related to the assumption of the selected allowable
differential settlement of 0.75 inch.

180
CHAPTER 7
CONCLUSIONS AND RECOMMENDATIONS
7.1 SUMMARY
Safety and reliability of bridge infrastructure is a major concern for many state highway
agencies. The American Association of State Highway and Transportation Officials
(AASHTO) has implemented the new approach for Load and Resistance Factor Design
(LRFD) on the basis of recent developments in structural design and analysis, and statistical
methods. Sections 10.5 and 10.6 of the AASHTO are relevant to the analysis and design of
bridge spread footing foundations. At this time, AASHTO-LRFD, is only calibrated for
strength limit states and is still not calibrated for service limit states (SLS). Calculations of
reliability indices for both total and differential settlements of spread footing foundations in
bridges can be considered as service limit states.
Even though shallow foundations have demonstrated cost advantages and satisfactory
field performance, deep foundations have remained the principal means for supporting
highway bridge structures in most states in US.

181
The goal of this study was to evaluate the reliability and accuracy of six prediction
settlement methods for calculating immediate settlement for spread footings on cohesionless
soils in highway bridges. Especially, the settlement prediction method outlined in AASHTO –
LRFD Bridge Design Specifications (2012), was selected in this study. Twelve highway
bridge structures were examined. Ten of the bridges were in Northeast US, and two bridges
were in Ohio. A total 23 spread footings were selected from these bridges to establish
statistical parameters for settlement and determine the statistical distribution of settlement as
well. Moreover, reliability indices were calculated for both total and differential settlements
for the best-selected prediction methods related to allowable values of settlement.
The selected 12 highway bridges had been instrumented with sensors and reference
points. The field performance of each instrumented spread footing site was monitored from
the start of construction until the bridge open to traffic. Measured settlement data for each site
were available at the end of construction, including post-deck settlement data.
The settlement prediction methods for shallow foundations on cohesionless soils,
selected for this study included the methods proposed by Hough (1959), which was the
method outlined in AASHTO-LRFD (2012), Meyerhof (1965), Peck and Bazarra (1969),
D’Appolonia (1970), Schmertmann (1978), and finally Burland and Burbidge (1985). For
these selected methods, statistical parameters for total and differential settlements were
determined related to the methods. The statistical distribution of actual settlement, as well as
that of settlements predicted by the methods was verified by using the Monte Carlo
simulation method. The assembled spread footing data base was then examined from different
views to establish correlations between the measured and predicted values of settlement.
182
Then, correlations between settlement and various factors were established to develop
guidelines for using spread footings in highway bridge structures.
Finally, reliability indices were calculated for total settlement by all the six prediction
methods and target reliability level was determined according to the best two methods related
to allowable suggested values of total settlement. Also, reliability indices were calculated for
differential settlement predicted by the best two prediction methods, and target reliability level
was determined for differential settlement according to allowable suggested values of
differential settlement, and angular distortion was calculated for the sites that had differential
settlements and compared with allowable values.
7.2 CONCLUSIONS
Extensive conclusions were obtained during the current study due to its multi-phased
nature. Relating to the objectives of the study, the following conclusions are presented:
1. Settlement Prediction Methods
There are more than a dozen methods that have been examined to estimate the behavior
of settlement in shallow foundations on cohesionless soils. Results of a study by Gifford et al.
(1978) indicated that the methods proposed by D’Appolonia (1970) and Burland and
Burbidge (1985) were the best settlement prediction methods. Briaud and Gibbens (1997)
concluded the methods proposed by Schmertmann (1970), Peck and Bazarrra (1969), and
183
Burland and Burbidge (1985) were the more accurate methods. Sargand (1997) stated that
Hough (1959) method was the most reliable settlement prediction method. Later, Sargand and
Masada (2006), concluded that the methods proposed by Hough (1959), Terzaghi and Peck
(1948), Schmertmann (1970), and Burland and Burbidge (1985) were the accurate prediction
methods selected among 14 prediction methods.
New techniques and models are available in addition to the traditional approaches for
analyzing the settlement behavior of spread footings on sands. These methods, including the
finite element analysis and neural networks, can be developed with computer programming
and efficient computational tools to be a more powerful tool for gaining further sights into
shallow foundations behavior. These analyses were not considered in this study because they
require many more parameters and much more data that lead to a higher level of uncertainty.
Based on the information available, the six prediction methods were selected for
calculating immediate settlement of shallow foundations on cohesionless. The objective was
to examine the accuracy of these methods, with accuracy evaluated by comparing predicted
and measured settlements of 23 spread footings on cohesionless soils. The range of measured
settlements was from 23 to 0.96 inches (5.75-24 mm) and was within the acceptable limits.
Statistical parameters for settlement were evaluated for each method and presented in Table
5.5. Methods proposed by Hough and Schmertmann indicated the best statistical parameters
for settlement. Schmertmann’s method resulted in bias factor closer to 1.0 (bias = 1.05), and
had the lowest value of coefficient of variation (COV =0.47). In addition, Schmertmann’s
method had the highest correlation coefficient (R2) between measured and predicted
settlements, which is an indication that this method is more accurate than the others. Even
though the Hough method had the lowest value of (COV= 0.5) among the other methods and
184
the bias factor was less than 1.0 (bias = 0.57), most of the predicted settlement values are
about twice the actual settlements. Therefore, the recommendation of this study is to use
Schmertmann’s method for calculating immediate settlement of shallow foundations on
cohesionless soils. The Hough method can be recommended, but predictions will be
overconsevative. However, the bias factors corresponding to these two methods can be similar
if the nominal values of settlement predicted by Hough’s method are divided by about 2.0.
2. Distribution Function of Settlement
To develop a better understanding of the variation involved in the bias factor for
settlement corresponding to the six prediction methods, frequency histograms were plotted as
shown in Figures 5.14 through 5.19.
The cumulative distribution function (CDF) of the measured settlement is plotted on the
normal probability paper as shown in Figure 5.20. The construction and use of the normal
probability paper is described in section 3.4. It is a convenient way to allow for an evaluation
of the most important statistical parameters as well as the type of distribution function. The
horizontal axis represents the basic variable. In case of considered test data, it is the ratio of
the settlement. The vertical axis is the inverse normal probability scale, and it represents the
distance from the mean value in terms of standard deviations. The vertical coordinate can also
be considered as the probability of exceeding the corresponding value of the variable. For any
value of the bias factor (horizontal axis), the vertical coordinate of CDF corresponds to a
certain probability of being exceeded. For example, value of -1 on the vertical scale
corresponds to a probability of 0.159 that the value of bias factor will be exceeded. The shape
of CDF is an indication of the type of distribution. A straight line means that the distribution
185
is normal. The CDF in Figure 5.20 for measured settlement is not for a normal distribution,
but rather for a lognormal distribution.
The CDF’s presented in Figures 5.21 through 5.26 for predicted values of settlement of
the six prediction methods for all the 23 sites of spread footings also appear to have a
lognormal distribution rather than normal. The correlation coefficient for all methods are
higher than 0.9. The Meyerhof method resulted in R2 =0.88, which is also acceptable.
Therefore, it seems all methods produced good results with lognormal distribution for
settlement. Additional evidence for treating settlement as lognormal distribution was
demonstrated through Monte Carlo method for 100,000 simulations, as shown in Appendix D.
All the settlement curves had a lognormal distribution rather than normal for the selected
methods in the Appendix.
In addition, the CDF’s in Figure 5.33 for the bias factor of settlement or settlement ratio
(δM/δP) for 23 sites by all six prediction methods seem to be lognormal rather than normal.
The critical situations are when the measured values significantly exceed analytically
predicted values. Hough and Schmertmann methods seem to produce better results than the
other methods. Therefore, the CDF’s indicate that none of the predicted methods can be
considered to have a normal distribution.
3. Influence of Various Factors on Settlement
Various factors have influence on settlement, especially geometry factors, as the
width, length, depth of embedment, and depth of water table. Other factors as the load and
soil properties quantified by the SPT N-values have influence on the settlement.
186
It can be noticed from Figures 5.34 and 5.35 that the width of footing, (B) has a
slight influence on settlement, that by increasing the width of footing, the settlement will
decrease. Meyerhof (1965), suggested checking check the influence of square root of
width of footing (B^ 0.5) on settlement. It turned out with the present data that the square
root of footing width has the same influence as the width of footing (B), as shown in
Figure (5.36). As, expected the settlement decreased as the footing width increased.
Another factor that has influence on the settlement is the length ratio (L/B), as
shown in Figure 5.37. It can be noticed there is a large scatter of data and no significant
trend can be seen. However, with increasing the L/B ratio, the settlement decreased as
expected, as shown in Figure 5.38.
The influence of the ratio of embedment depth to footing width (Df/B) had an
influence on settlement, with a slight trend of decreasing settlement value with increasing
the ratio (Df/B), as shown in Figures 5.39 and 5.40. It can be noticed that this is not
significant.
There is a large scatter in the data and no definite trend is considered between the level
of water table and the settlement as shown in Figure 5.41. As shown in Figure 5.42, it can be
concluded the level of water table has no significant effect on settlement, even if it is
increased. Due to the high variability, the effect of water table level is already reflected in the
SPT-N values, supported by the study results of Burland and Burbidge (1985).
The axial load (P) has an influence on settlement as shown in Figure 5.43. With an
increase in the axial load, there is a significant increase in the settlement, as shown in Figure
187
5.44 and the settlement behavior of the curves is close to Curve C, mentioned in Figure 5.5
in Section 5.4.
Finally, the last factor that has influence on settlement is the corrected SPT-N values
with settlement, as shown in Figure 5.45, which shows a small trend even if the correlation
coefficient is small. It can be noticed in Figure 5.46 that with increasing SPT-N values the
settlement decreases, but after reaching value of N60 = 50, there is no significant decrease in
settlement.
4. Differential Settlement
Differential settlement is the most troublesome type of settlement, especially in highway
bridges. The serviceability limit state (SLS) will be exceeding when the differential settlement
is higher than the allowable limits, causing the foundation design to be unsatisfactory. Various
sources suggest that footing differential settlements can be estimated as a percentage of the
total settlements.
In this study, differential settlements were calculated for the 12 highway bridges, but
only for 20 footing sites. The values are presented in Table 5.6, which contains actual
measured differential settlements and calculated differential settlements predicted by the
identified methods. Statistics of differential settlement are presented in Table 5.7. It can be
concluded that Hough and Schmertmann methods produced results closest to the average
calculated differential settlement (0.25 in.), compared with the actual differential settlement of
(0.23 in.). Therefore, the Hough and Schmertmann methods are the two prediction methods
recommended for calculating differential settlement on cohesionless soils for shallow
foundations.
188
The distribution of the actual differential settlement is shown in Figures 5.47 and 5.48. It
seems to have a lognormal distribution rather than a normal distribution. The CDF’s of
differential settlement for Hough and Schmertmann methods are shown in Figures 5.50 and
5.52, respectively. It can be concluded that the distribution is lognormal with high coefficients
of correlation: R2 =0.95 for differential settlement by Hough method and R2 =0.94 for the
Schmertmann Method. Therefore, the lognormal distribution will be recommended for
differential settlement as well.
5. Angular Distortion
Angular distortion is usually results from differential settlement. Angular distortions are
evaluated depending on estimated differential settlements for the 20 footing sites in this study,
were differential settlement was reported. Actual and Predicted angular distortions are
presented in Table 5.10. It can be concluded that all the actual values of angular distortions
and the values predicted by the selected methods are within the range of limiting angular
distortions as given in AASHTO (2012), as of 0.004 for continuous span bridges and 0.008
for single span bridges. Table 5.11 contains statistical parameters for angular distortion for all
the six selected methods. It can be noticed that the Hough and Schmertmann methods had
results closest to the actual angular distortion values.

189
6. Reliability Analysis
Reliability analysis was performed using the FOSM method and Monte Carlo
simulations, Nowak and Collins, (2013). The reliability, as a measure of structural
performance, was expressed in terms of reliability indices. In the present study, reliability
indices were calculated for both total and differential settlements of shallow foundations in
highway bridges with the objective developing a risk analysis procedure specifically for
prediction of settlement for shallow foundations lying on cohesionless soils.
Reliability indices were calculated from equation 6.2, depending on the assumption of
treating settlement as lognormally distributed. Tolerable settlement was considered in the
range between 0.25 and 3.0 inches, and was treated as deterministic values. Reliability indices
determined for total settlement for the 23 sites of foundations, are shown in Figures 6.1
through 6.23. It can be concluded that the Hough and Schmertmann methods had the best
reliability indices among the six methods and that Schmertmann was the best of all. For
differential settlement, the same procedure was used, but only the Hough and Schmertmann
methods were considered, since they predicted the best results. Reliability indices for
differential settlement are shown in Figures 6.24 through 6.33. It can be concluded that the
Schmertmann method was better than the Hough method for indicating reliability.
The target reliability level is the acceptable safety level that should be established for
various design requirements. The allowable total settlement was suggested in this study to be
1.5 inch (37.5 mm) as NDOR and ODOT suggest 1.5 inch for tolerable settlements, and our
actual values did not exceed 0.96 inches (24 mm). Therefore, the target reliability level was
190
obtained using 1.5 inch as tolerable suggested settlement based on the average of the Hough
and Schmertmann methods. Therefore, the target reliability index was determined to be βT
=3.5 for total settlement, with probability of failure or probability of exceeding the limit of
0.02% and with allowable selected value of total settlement of 1.5 inch. For differential
settlement, the same procedure was used for obtaining the target reliability level. The
allowable differential settlement was considered half the total settlement suggested by Holtz
(1991). Therefore, the allowable differential settlement would be considered to be 0.75 inch
(18.75 mm). Consequently, the target reliability index was determined to be βT =2.4 for
differential settlement, with the probability of failure or probability of exceeding the limit of
(0.8%) and with allowable suggested differential settlement of 0.75 inch. It can be concluded
that the target probability of failure obtained for both total and differential settlements was
less than recent studies for service limit states, as Zekkos et al. (2004) got values as high as
30%, Fenton et al. (2005) evaluated (Pf)T as 5 %., and Akbas (2007) used a probabilistic
approach and found (Pf)T to be about 1.1%.
Finally, main conclusions can be outlined in this study as:
 LRFD specifications are intended to include considerations of service limit states of
substructures, especially shallow foundations. The procedure evaluated in this study
was to design the service limit state for both tolerable and predicted settlements for
shallow foundations, recognizing the excessive settlements can lead to serviceability
or even failure states in the bridge substructure.
 The main objective of this study was to develop the reliability-based design
methodology of shallow foundations on cohesionless soils in bridges. This objective
was implemented by a critical evaluation of current prediction methods to estimate the

191
settlement of shallow foundations on cohesionless soils in bridges under a sensitivity
analysis.
 The main contribution of this study was formulation of rational acceptability criteria
for tolerable settlement and development of statistical parameters for settlement
prediction methods for shallow foundations in highway bridges. The main effort and
contribution was concentrated on the development of reliability analysis procedure and
statistical parameters for total and differential settlements.
7.3 RECOMMENDATIONS FOR FUTURE WORK
1. Perform a reliability analysis of settlement for shallow foundations on cohesive soils.
Consolidation settlement should be evaluated and added to immediate settlement for
obtaining the total settlement.
2. Develop a settlement prediction method combining features of the Hough and
Schmertmann methods. This work is underway with the goal of obtaining better
statistical parameters.
3. Begin to consider system reliability, rather than component reliability, in soil-structure
interaction problems. The formulation of the limit state function is more difficult for
substructures, particularly shallow foundations, than for superstructures, not only
because individual failure modes require more complicated analyses, but also because
of system behavior.
4. Comparing measured settlements from Europe with United States, and estimating
statistical parameters and reliability indices.

192
5. Develop research in areas of serviceability of substructures.
7.4 IMPLEMENTATIONS OF THIS STUDY
 Compared to other foundations, shallow foundations are less expensive, easier to
construct, require less time and their performance for highway bridges is determined
by the deformation criteria, not the bearing failure of soil.
 The formulation of SLS for shallow foundations requires the best fit to the target
reliability index, and would provide a significant way to use shallow foundations to
support bridges superstructure weight and live loads.

193
REFERENCES
AASHTO LRFD Bridge Design Specifications (2012). “American Association of State Highway and
Transportation Officials”, Washington, D.C.
Akbas, S. O. (2007). “Deterministic and Probabilistic Assessment of Settlements of Shallow

Foundations in Cohesionless Soils.” Ph.D. Dissertation, Cornell University, N.Y.
Allen, T.M., Nowak, A.S. and R.J. Bathurst (2005) . “Calibration to Determine Load and Resistance
Factors for Geotechnical and Structural Design”. TRB Circular E-C079, Transportation Research
Board, Washington, D.C.
Anderson B. J., Townsend F. C. and Rahelison L. (2007).”Load Testing and Settlement Prediction of
Shallow Foundation.” J. Geotech. Geoenv. Eng., vol. 133, No.12, 1494-1502.
Ang, A.H-S. and Tang W.H., (2007). “Probability Concepts in Engineering”, John Wiley & Sons, Inc.
Ayyub, B.M. and McCuen, R.H., (1997). “Probability, Statistics and Reliability for Engineers”, CRC
Press, New York.
Benjamin, J. R., and Cornell, C. A. (1970). “Probability, Decision, and Statistics for Civil Engineers”,
McGraw-Hill, New York.
Berardi R., Jamiolkowski M., and R. Lancellotta (1991). “Settlement of Shallow Foundations in Sand:
Selection of Stiffness of the Basis of Penetration Resistance”. Geotechnical Engineering Congress, No.
27, Geotechnical Special Publications, ASCE, pp.185-200.
Bjerrum L., and Eggestad A. (1963). “Interpretation of a Loading Test on Sand”, Proceedings 1st
European Conference on Soil Mechanics and Foundation Engineering, Wiesbaden, pp. 199-204.
194
Bowles, J.E. (1988). “ Foundation Analysis and Design”, 4th edition, McGraw Hill Book Company,
New York.
Braja M. Das (2000). “Fundamentals of Geotechnical Engineering”. Brooks/Cole.
Braja M. Das (2007) . “Principles of Foundation Engineering”. Cengage Learning, Stamford, CT.
Braja M. Das, and Nagaratnam Sivakugan (2007). “Settlement of Shallow Foundations on Granular
Soils - an overview”. International Journal. of Geotechnical Engineering. No. 1, pp. 19-29.
Brian J. Anderson , Townsend F.C., and Rahelison (2007). “Load Testing and Settlement Prediction of
Shallow Foundation”. Journal of Geotechnical and Geoenviromental. Engineering. Vol. 133, No. 12 ,
pp.1494-1502.
Burland, J.B., and M. C Burbidge (1985). “Settlement of Foundations on Sand and Gravel “. Institution
of civil engineers, Glasgow and West of Scotland Association, London, Vol. 78, No. 1, pp. 1325-1381.
Chen, J.R. (2004). “Axial Behavior in Drilled Shafts in Gravelly Soils”. Ph.D. Thesis, Cornell
University, Ithaca.
Cheney R. and Chassie R. (2000). “ Soils and Foundations Workshop Reference Manual”, Final
Report FHWA/ NHI–00–045, US Department of Transportation , Federal Highway Administration.
Clayton, C.R.I. (1990). “SPT Energy Transmission: Theory, Measurement and Significance”, Ground
Engineering. Vol. 23, No. 10, pp. 33-42.
Coduto P. Donald (2001). Foundation Design. Principles and Practices. Prentice –Hall New Jersey.
Cornell C.A. (1967). “Bounds on the Reliability Structural Systems”, Journal of the Structural Division,
ASCE. Vol. 93, pp. 193-200.
195
Cornell C.A. (1969). “A Probability Based Structural Code”, ACI Journal , Vol. 66, pp. 974-985.
Christian J.T. and Carrier W.D. (1978). “Janbu, Bjerrum, and Kaernsli’s Chart Reinterpreted”, Canadian
Geotechnical Journal. Vol. 15, No. 1, pp. 123-128.
D’ Appolonia, D.J., D’Appolonia E. D., and R. F. Brissette (1970). “Closure : Settlement of Spread
Footings on Sand”. Journal of Soil Mechanics and Foundation Division. Vol. 96, No. 2, pp. 754-762.
D’ Appolonia, D.J., and T.W. Lambe (1970). “Method for Predicting Initial Settlement”. Journal of
Soil Mechanics and Foundation Division. ASCE, Vol. 96, SM2, pp. 523-543.
Ditlevsen O. and Madsen H.O. (1996). “Structural Reliability Methods”, John Wiley & Sons Inc., New
York.
Douglas D.J. (1986). “State –of-the –art”. Ground Engineering. Vol. 19, No. 2, pp. 2-6 .
Fenton, G.A. and Griffiths, D.V. (2005). “Resistance Factors for Settlement Design.”. Canadian .
Geotechnical. Journall., Vol.(42), 1422-1436.
Fillippas, O. B., Kulhawy, F. H., and Grigoriu, M. D.(1988). “Reliability Based Foundation Design for
r Transmission Line Structures: Uncertainties in soil property measurement.” Report. No. EL-5507(3), El
r Research Institute, Palo Alto, California.
Gifford, D.C., Wheeler, J. R. , Kraemer, S. R., and A. F. McKown (1987). “Spread Footings for
Highway Bridges”. Final report FHWA/ RD-86 /185, US Department of Transportation , Federal
Highway Administration.
Gibbs, H.J. and Holtz, W.H. (1957). “Research on Determining the Density of Sands by Spoon
Penetration Testing”, Proceedings 4th International Conference on Soil Mechanics and Foundation
Engineering, London, pp. 35-39.
196
Hirany A. Kulhawy, F.H. (1988). “ Conduct and Interpretation of Load Tests on Drilled Shaft
Foundations”, Report EL-5915, Electric Power Research Institute, Palo Alto.
Holtz R.D. and Allen T. M. (1991). “Design of Retaining Walls Reinforced with Geosynthetics”.
Hough B. K. (1959).” Compressibility as the Basis for Soil Bearing Value”. Journal of Soil Mechanics
and Foundation Division. Vol. 85, No. 4, pp. 11-39.
Junhwan Lee, and Rodrigo Salgado (2002).” Estimation of Footing Settlement in Sand”. The
International Journal of Geomechanics, Vol. 2, No.1, pp. 1-28.
Kovacs W.D. (1994). “Effects of SPT Equipment and Procedures on the Design of Shallow Foundations
on Sand”, Proceedings, Vertical and Horizontal Deformations of Foundations and Embankments
(GSP40), Ed. A. T. Yenug and G. Y. Felio, ASCE New York.
Kovacs W.D and Salmone L.A. (1986). “ SPT Hammer Energy Measurement”, Journal of the
Geotechnical Engineering Division, ASCE. Vol. 108, No. 4, pp. 599-620.
Kulhawy, F.H. and Mayne, P.W. (1990). “Manual on Estimating Soil Properties for Foundation
Design”, Report EL-6800, Electric Power Research Institute, Palo Alto.
Liao, S.S. and Whitman, R.V. (1986). “Overburden Correction Factors for SPT in Sand”, ”. Journal of
Geotechnical Engineering, ASCE. Vol. 112, No. 3 , pp.373-377.
Lutenegger J. Alan and DeGroot J. Dan (1995). “Settlement of Shallow Foundations on Granular
Soils”. Final Report of research Conducted For Massachusetts Highway Department.
Meyerhof G.G. (1956). “Penetration Tests and Bearing Capacity of Cohesionless Soils”. Journal of Soil
Mechanics and Foundation Division. ASCE, Vol. 82, No. 1, pp. 1-19.
197
Meyerhof G.G. (1965). “Shallow Foundations”. Journal of Soil Mechanics and Foundation Division.
ASCE, Vol. 91, SM2, pp. 21-31.
Nowak, A.S., (1995). "Calibration of LRFD Bridge Code", ASCE Journal of Structural Engineering,
Vol. 121, No. 8, pp. 1245-1251.
Nowak, A.S., (1999). "Calibration of LRFD Bridge Design Code", NCHRP Report 368, Transportation
Research Board, Washington, D.C.
Nowak, A.S. and Collins, K.R. (2013). “Reliability of Structures”, McGraw-Hill, New York.
Nowak, A.S. and Szerszen, M.M. (2003). “Calibration of Design Code for Buildings (ACI 318) Part 1:
Statistical Models for Resistance”, ACI Structural Journal, Vol. 100, No. 3, pp. 377-382.
Peck, R.B., and A.R.S.S. Bazaraa (1969). “Discussion of Settlement of Spread Footings on Sand”.
Journal of Soil Mechanics and Foundation Division. ASCE, Vol. 95, SM3, pp. 305-309.
Peck, R.B., Hanson W. E., and Thornburn T.H.(1974).”Foundation Engineering”, 2nd Edition, Wiley,
New York.
Phoon, K. K. (2006). “Modeling and Simulation of Stochastic Data.” Geotechnical Engineering in the
Information Technology Age ,CD-ROM.
Rasin Duzceer (2009). “ Observed and Predicted settlement of Foundation.”2nd International

Conference on New Development in soil Mechanics and Geotechnical Engineering, Near East
University, Nicosia, North Cyprus , 28-30 May.
Rosenblueth E. (1975). “Point Estimates for Probability Moments”, Proceedings of the Nature academy
of Science. Vol. 72, No. 10.
Rosenblueth E. (1981). “ Two Point Estimate in Probabilities”, applied Mathematical Modeling Journal.
198
Samtani N. C. and Nowatzki E A. (2006). “Soils and Foundations”. Reference Manual-Volume I.

Final report FHWA- NHI-06-088, US Department of Transportation , Federal Highway Administration.
Samtani N. C. and Nowatzki E A. (2006). “Soils and Foundations”. Reference Manual-Volume II.
Final report FHWA- NHI-06-089, US Department of Transportation , Federal Highway Administration.
Samtani N. C. , Nowatzki E A. , and D. R. Mertz (2010). “Selection of Spread Footings to Support

Highway Bridge Structures”. Final report FHWA-RC/TD-10-001, US Department of Transportation ,
Federal Highway Administration.
Schmertmann. J.H (1970). “Static Cone to Compute Static Settlement Over Sand”. Journal of Soil
Mechanics and Foundation Division. ASCE, Vol. 96, No. 3, pp. 1011-1043.
Schmertmann, J.H. , Hartman J. P., and P.R. Brown (1978). “Improved Strain Influence Factor
Diagrams”. Journal of Soil Mechanics and Foundation Division. ASCE, Vol. 104, No. 8,, pp.1131-
1135.
Shcultz E. and G. Sherif (1973). “ Prediction of Settlements from Evaluated Settlement Observations for
Sand”. Proc. , 8th International. Conference on Soil Mechanics and Foundation Engineering. Moscow.
Sargand M. Shad, Masada Teruhisa, and Richard Engle (1999). “Spread Footing Foundation for
Highway Bridge Application”. Journal of Geotechnical. and Geoenviromental Engineering. Vol. 125,
No. 5, pp. 373-382.
Sagrand, S.M., Masada T., and Abdalla B. (2003).” Evaluation of Cone Penetration Test based on
Settlement Prediction Methods for Shallow Foundations on Cohesionless soils at Highway Bridge
Construction Sites.” Journal of Geotechnical. and Geoenviromental Engineering of ASCE, Vol.(129),
pp. 900-908.
Sargand M. Shad and Masada Teruhisa. (2006). “ Further Use of Spread Footing Foundation for
199
Highway Bridges” Final Report FHWA-OH-2006/8, Ohio Department of Transportation and US

Department of Transportation , Federal Highway Administration.
Skempton A.W. (1986). “ Standard Penetration Test Procedures and the Effects in Sands Overburden
Pressure, Relative Density, Particle Size, Ageing and Overconsolidation”, Geotechnique. Vol. 36, No. 3,
pp. 425-470.
Sivakugan N., and K. Johnson (2004). “Settlement Predictions in Granular Soils : A Probabilistic
Approach”. Journal of Geotechnique. Vol. 54, No. 7, pp. 499-502.
Tan, C.K. and J.M. Duncan (1991). “Settlement of Footings on Sands: Accuracy and Reliability”.
Proceedings of Geotechnical Engineering Congress. Colorado, No. 1, pp. 446-455.
Terzaghi, K., and Peck R.B (1948). “ Soil mechanics in Engineering Practice”, Wiley, New York.
Terzaghi, K., Peck, R. B., and Mesri, G.(1996). “Soil Mechanics in Engineering Practice”, 3rd Ed.,
Wiley, New York.
Thoft-Christensen, P. and Baker, M.J. (1982). “Structural Reliability Theory and Its Applications”.
Springer-Verlag.
Zekkos, D. P., Bray, J. D., and Der Kiureghian, A. (2004). “Reliability of Shallow Foundation Design
Using the Standard Penetration Test.” Proc., 2nd International. Conference on Site Characterization
(2), Millpress, Rotterdam, 1575–1582.
200
APPENDIX A
Logs of Borings No. B-1, No.B-2 in Hancock, IA
Figure A-1: Logs of Boring B-1 in Hancock, IA

201
Figure A-1: Logs of Boring B-1 in Hancock, IA, continued

202

203
Figure A-2: Logs of Boring B-2 in Hancock, IA

204

205

206
APPENDIX B
B-1 Historical Summary of Candidate Bridges in Northeast of
United States
1) Candidate Bridge Number : M5000

 Location: North Avenue Sideline over VT 127,
Burlington, Vermont.
 Geological Setting: Lacustrine deposits from glacial Lake Champlain, some
verve indicated in borings, fine sands interlayered with silt.
 SUBSURFACE INFORMATION
I- Typical soil Profile:
Depth Below Generalized Soil Standard Penetration Resistance

Footing Description
(ft) Range Average
0-12 Medium –Fine Sand 28-92 56
12-22 Sandy Silt A-4 6-50 25
22-37 Fine Sand A-3 5-22 12
37-57 Sandy Silt A-4 5-8 6
57-77 Fine Sand A-3 8-30 22
77-129 Sandy Silt A-4 14-79 41
129 Bedrock
II- Other Subsurface Information:

 Number of borings: 8
 Thickness of natural granular soil beneath footing: 10-25 ft.

207
 Thickness of structural fill beneath footing: 12 ft.
 Remarks: Compressible Silt Strata beneath granular soils.
 BRIDGE DESIGN DATA
Number of spans: 1 Span length: 128 ft.
Design bearing pressure: 4 ksf.
Special features: surcharge embankment
Remarks: 30 day surcharge embankment. 10-20 ft. height over silt.
Bottom of embankment is 12 ft. below footing.
 BRIDGE CONSTRUCTION DATA
Surcharge: 11/1983-11/1984 Begin construction: 04/1985
Footings completed: A1 - 05/1985 , A2 - 05/1985
Abutment wall completion: 06/1985 Wing wall completion: 06/1985
Backfill complete: 06/1985 Structure complete: 08/1985
Pavement complete: 09/1985 Open to traffic: 09/1985
 INSTRUMENTATION SYSTEM: Full Scale
Settlement points: 16 Settlement platforms: 14
Tilt/Overturning points: 8 Deep Settlement/Strain: 1
Remote settlement/Profile: 3 Contact stress: 10
Applied loading: yes

208
2) Candidate Bridge Number : 131-132-11

 Location: I-691 under relocated Dickerman Road Southington
Cheshire, Connecticut.
 Geological Setting: Glacial outwash sands over Arkosic sandstone.

Footing Description
(ft) Range Average
0-35 Fine Sand and Silt 7-24 16
35 + Arkosic Sandstone

 Thickness of natural granular soil beneath footing: 35 ft.
 Thickness of structural fill beneath footing: 0.5-6.0 ft.
Number of spans: 2 Span lengths: 112 ft.
Design bearing pressure: 5.0 ksf.
Special features: Prestressed/Pretentioned girders.

209
Surcharge: 10/1984 Begin construction: 04/1985
Abutment wall completion: A1 - 03/1985 , A2 - 12/1984
Pier completion: 12/1984
Backfill complete: A1 - 11/1985 , A2 - 03/1985
Structure complete: 11/1985 Pavement complete: NA
Open to traffic: NA
Remarks: Not complete as of 06/1986
*NA: Not applicable
Settlement points: 13 Settlement platforms: NA
Tilt/Overturning points: 10 Deep Settlement/Strain: NA
*NA: Not applicable
3) Candidate Bridge Number : 931

 Location: Branch Avenue, Northeast Corridor
Providence, Rhode Island.
 Historical Background: Replacement of 1910 vintage steel bridge across
Amtrak Northeast Corridor.
 Geological Setting: Glacial outwash plain.

210

Footing Description
(ft) Range Average
0-15 Medium Fine Sand 16-38 24
15-140 Fine Sand (Outwash) 9-88 48
140-155 Glacial Till 100 +
155 Bedrock

 Thickness of granular soil beneath footing: 20- 30 ft.
 Thickness of structural fill beneath footing: 0 as built-18 ft.
 Remarks: Preconsolidated sand.
Number of spans: 42 Span lengths: 120 ft. ±
Special features: 74 inch sewer under centerline. Bridge footings on both
sides.
Remarks: Bridge across railroad

211
Begin construction: 10/1983
Wing wall completion: A1 - 12/1983 , A2 - 06/1984
Backfill complete: A1 - 01/1984 , A2 - 07/1984
Structure complete: 08/1984 Pavement complete: 11/1984
Open to traffic: 11/1984
Applied loading: yes , two places
*NA: Not applicable

 Location: Relocated Gersoni Rd. – Route 28 over D&H RR and Route 7,
Collierville, NY.
 Geological Setting: Quaternary Alluvium over silt.

212

Footing Description
(ft) Range Average
0-20 Coarse –Fine Sand/ Silty 2-37 17
Sand
20-45 Silty Fine Sand 10-44 24
45-165+ Silt 3-13 6

 Thickness of natural granular soil beneath footing: 20-45 ft.
 Thickness of structural fill beneath footing: 25 ft.
 Remarks: Compressible Silt strata beneath granular soils.
Number of spans: 2 Span lengths: 112 ft.
Special features: Surcharge north abutment
Surcharge: 10/1983 Begin construction: 03/1984
Footings completed: N - 04/1984 , S - 04/1984
Abutment wall completion: : N - 05/1984 , S - 04/1984

213
Wing wall completion: 05/1984 Pier completion: 06/1984
Settlement points: 18 Settlement platforms: 3
Tilt/Overturning points: 15 Deep Settlement/Strain: 1
5) Candidate Bridge Number : U-2-59

 Location: Route 146 Southbound over relocated Lackey Dam Road, Uxbridge,
MA.
 Historical Background: Route 146 widened to 4 lanes. History of many
accidents at former 2 lane bridge.
 Geological Setting: Glacial outwash plain.

Footing Description
(ft) Range Average
0-55 Coarse to Fine Sand 10-47 27
55 + Glacial Till 100
214

 Thickness of natural granular soil beneath footing: 45-55ft.
 Thickness of structural fill beneath footing:4.0-5.5 ft. Proposed,
none as built.
Design bearing pressure: 5.0 ksf. Max toe pressure.
Footings completed: N - 11/1983 , S - 12/1983
Abutment wall completion: N - 01/1984 , S - 01/1984
Wing wall completion: 01/1984
Backfill complete: N - 05/1984 , S - 01/1984
*NA: Not applicable

215

 Location: Vermont Route 11 over the middle branch – Williams River.
Chester, Vermont.
 Historical Background: River noted for rapid rise and fall off water level.
Fourth bridge at this location; second bridge destroyed in 1927 flood; third
bridge built in 1928 was obsolete.
 Geological Setting: 100 ft. high embankment of lacustrine; silt 100 yards
downstream. 20 ft. high outcrop of mica schist 100 yards upstream.

Footing Description
(ft) Range Average
0-20 Silty Sand / Silt 18-171 95

 Thickness of natural granular soil beneath footing: 20 ft. ±
 Thickness of structural fill beneath footing: 3.0 ft.
 Remarks: Boulders and cobbles made driving of sheet piles
difficult.
216
Wing wall completion: : A1 - 08/1983 , A2 - 09/1983
 INSTRUMENTATION SYSTEM: Partial
Settlement points: 24
Tilt/Overturning points: 6

 Location: I-86, Manchester, Connecticut.
 Geological Setting: Verve Glacial outwash sands over Arkosic sandstone.

Footing Description
(ft) Range Average
0-90 + Coarse to Fine Sand 8-59 29
217

 Thickness of structural fill beneath footing: A-2 - 2.0 ft.
Number of spans: 4
Span lengths: 114 ft. , 132 ft. , 162 ft. , and 174 ft.
Remarks: Railroad bridge crossing I-86
Wing wall completion: 12/1983 Pier completion: 03/1984
218

 Location: I-86 and CD Roadway under Tolland Turnpike
Manchester, Connecticut.
 Geological Setting: Verve Glacial outwash sands.

Footing Description
(ft) Range Average
5-45 Medium Sand 9-62 38

 Thickness of natural granular soil beneath footing: 45 ft. +
 Thickness of structural fill beneath footing: 5.0 ft.
Number of spans: 4
Span lengths: 145 ft. , 220 ft. , 220 ft. , and 175 ft.
Remarks: Superstructure designed for Post-Construction differential
settlement between adjacent superstructure units.

219
Wing wall completion: A1 - 02/1984 , A2 - 07/1983
Pier completion: 02/1984 Backfill complete: 07/1984

 Location: CD-WB Roadway and Ramp 1 over Buckland ST. (I-86)

Footing Description
(ft) Range Average
some Silt
35 + Decomposed Rock 100 +
220

 Thickness of structural fill beneath footing: 5.0-10.0 ft.
Begin construction: 08/1983 Footings completed: 08/1983
Abutment wall completion: 09/1983 Wing wall completion: 10/1983
Pavement complete: Prior to 11/1985
Open to traffic: Prior to 11/1985
221

 Location: I-86 EB and WB and Ramps A, J, and P under middle Turnpike

Footing Description
(ft) Range Average
0-10 Gravely Sand 8-55 24
10-30 Fine Sand and Silt 3-44 30
30 + Arkosic Sandstone

 Thickness of structural fill beneath footing: 2.0 ft. typical
Number of spans: 5
Span lengths: 155 ft. , 160 ft. , 171 ft. , 180 ft. , and 195 ft.
Design bearing pressure: 6.0 – 8.0 ksf.
Remarks: Superstructure designed for Post-Construction differential
settlement between abutment # 2 and pier # 4.
Begin construction: ND Footings completed: ND

222
Abutment wall completion: ND Wing wall completion: ND
Backfill complete: ND Structure complete: ND
Pavement complete: ND Open to traffic: 11/1985
*ND: Not determined
Remarks: Post construction settlement only.
B-2 Summary of the Relevant Data of the Ten Highway

Bridges in Northeast of United States
This summary includes all the information data of the footings of the ten
highway bridges in northeast of US. It includes the dimensions of the footings
(B,L,Df), the maximum contact pressure (qmax), the corrected SPT N-values (Ncorr),
and measured settlement (Sm) for each site of footing as shown in the following Table
B.1 below :
Table B.1: Data Established for the Ten Highway Bridges

Adapted from (FHWA, 1987 and 2010)
Site Structure B L/Df Df/B N qmax SM SM/B

# (ft) (dim) (dim) (dim) (ksf) (inch) (%)
1 Bridge 1- Abutment 1 17.0 3.7 0.1 44 3.2 0.35 0.17
2 Bridge 1- Abutment 2 17.0 3.7 0.1 58 2.6 0.67 0.33
3 Bridge 2- Abutment 1 15.3 3.4 0.1 43 2.4 0.94 0.51
4 Bridge 2- Abutment 1 16.8 3.1 0.2 19 2.4 0.76 0.38
5 Bridge 2 - Pier 12.5 3.3 0.4 12 1.8 0.61 0.41
Continued, Table B.1
223
6 Bridge 3- W. Abutment 11.0 6.8 0.0 34 1.8 0.42 0.32

7 Bridge 3- E. Abutment 18.5 4.3 0.3 22 2.4 0.61 0.27
8 Bridge 3-Pier 1 North 21.0 1.0 0.2 18 2.0 0.28 0.11
9 Bridge 3-Pier 1 South 21.0 1.5 0.2 18 1.6 0.26 0.10
14 Bridge 4- S. Abutment 8.1 5.3 0.8 21 3.4 0.46 0.47
15 Bridge 4- N. Abutment 8.1 5.3 0.8 8 3.4 0.34 0.35
16 Bridge 5- N. Abutment 16.8 4.6 0.4 42 2.4 0.23 0.11
17 Bridge 5- S. Abutment 15.3 5.0 0.4 24 2.4 0.44 0.24
18 Bridge 6- Abutment 1 15.3 4.0 0.6 55 1.8 2.26 1.23
19 Bridge 6- Abutment 2 15.3 4.4 0.6 39 1.8 0.83 0.45
20 Bridge 7- Abutment 2 28.0 1.0 0.0 24 2.2 0.64 0.19
21 Bridge 8- Abutment 1 20.0 5.0 1.1 23 3.0 0.46 0.19
22 Bridge 8- Abutment 2 20.0 5.0 0.3 38 3.2 0.66 0.28
23 Bridge 9- Abutment 1 21.8 2.0 0.1 39 3.6 0.61 0.23
24 Bridge 9- Abutment 2 16.0 2.8 0.0 49 3.4 0.28 0.15
Notes: dim- dimensionless, NA- Not Available.
where, B = width of footing (ft), L= length of footing (ft), D = depth of footing below
ground surface (ft), Ncorr = Corrected SPT-N values with 60 % hammer efficiency, qmax=
bearing pressure (ksf), SM: measured settlement (inch).
The total soil unit weight (γ) was assumed to be between (115-120) kcf. for all the 24 sites.
The depth of water (Dw) ranged between (0-44) ft. for all the 24 sites.
The thickness of soil layer below footing (H) ranged between (10-197) ft. for all 24 sites.
xial
oadin
224
APPENDIX C
C-1 Historical Summary of Candidate Bridges in Ohio
11) Candidate Bridge Number : FRA-670-0380

 Location: High Street over I-670 highway between Goodale Street and Popular
Avenue in Columbus, Franklin County, Ohio.
 Historical Background : The bridge is located between the city’s convention and a
shopping district, therefore the deck has a total width of 78 ft (23.8 m), which
would not only support traffic in two lanes, but also pedestrians on walkways and
shops of both sides.
Figure C.1 presents the view of the bridge structure taken after the placement of the girder
beams.
Figure C.1: Bridge Structure After Placement of Beams

( Sargand and Masada 2006, FHWA)
225
The bridge had two soil bore holes (RB-9 and RB-11) which were placed in the
vicinity of the central pier foundation. No groundwater table was detected close to the
bottom of the footing in each of the bore holes. The boring log records for bore holes
RB-9 and RB-11 are listed in Table C.1 and Table C.2 respectively. The averages of
these borings are presented in Table C.3 and the corrected hammer efficiency of %60
is also presented in this table.
Table C.1: Soil Boring Logs for Bore Hole RB-9
Depth Below
Footing Generalized Soil Description SPT N-Value
(ft)
0 Gray Sandy Silt, Trace Gravel, Trace 32
Cobbles A-4a
5 The same as above A-4a 65
10 Gray Gravel, Little Sand, Little Silt, Trace 73
Cobbles< Trace Shale A-2-4
15 The same as above A-2-4 83
Table C.2: Soil Boring Logs for Bore Hole RB-11
Depth Below
Footing Generalized Soil Description SPT N-Value
(ft)
0 Gray Sandy Silt, Trace Gravel, Trace 69
Cobbles A-4a
15 The same as above A-4a 43
20 Gray Gravel, Little Sand, Little Silt, 70

Trace Cobbles< Trace Shale A-2-4
226
Table C.3: Average of SPT N-Values of the Two Soil Boring Logs
Depth Below Corrected

Footing SPT N-Values (blows/ft) SPT N-Values
(ft)
RB-9 RB-11 Average (N1)60
0 32 69 51 44
5 65 ---- 65 52
10 73 ---- 73 54
15 83 43 63 44
20 98 70 84 56
25 61 69 70 44
The average SPT N-values for the two bore holes which are analyzed within each
stratum for the logs of soil borings are shown in Figure C.2 in terms of SPT N- values
with the depth. The corrected SPT N-Values with depth are shown in Figure C.3
SPT N- Values (blows/ft)

0 20 40 60 80 100
0
5
Depth (ft)
10
15
20
25
30
Figure C.2 : SPT N-Values variations with Depth for Project FRA-670-0380
227
Corrected SPT N- Values
0 10 20 30 40 50 60 70 80
0
5
Depth (ft)
10
15
20
25
30
Figure C.3 : Corrected SPT N-Values Variations with Depth for
Project FRA-670-0380

 Number of spans: 2
 Span lengths of bridge: North span=102.9 ft. and South span=100.2 ft.
 Average contact pressure: 2.85 tsf.
 Special features: The deck construction produced the largest contact
pressure increase among all the construction stages.
 Remarks: The superstructure is supported by a massive concrete

abutment wall at each end and a spread footing near the mid-span.

 Begin construction: 02/2003, when the ground was excavated for the
central footing.
 Footing and pier column construction: 03/2003.

228
 Soil backfill and pier cap construction: 04/2003.
 Barrier wall construction: End of April 2003.
 girder beams placement: 05/2003
 Deck construction: 06/2003.
 Bridge open to traffic: 07/29/2003.
 Last visit to bridge site was made on 09/16/2003.

 Settlement monitoring points: 5
 Tilt/Overturning points: 10
 Contact stress: 6
 Applied loading: yes
12) Candidate Bridge Number : MOT – 70/75 Interchange

 Location: I-70/75 Interchange in Montgomery County, Ohio.
 Historical Background: The goal of this project was to replace the 1950’s
interchange deign with a more modern one and a higher capacity efficient ramp
design. The new design is supposed to reduce the accident rates and handle more
traffic. The old and new designs are shown in Figures C.4 ,(a) as the old design
and (b) as the new design.

229
(a) Old Interchange Design
I-70
Ramp C Bridge I-75 Piers 18 & 19
(b) New Interchange Design
Figure C.4: Old and New Interchange Designs (MOT-70/75)

230
The bridge had two soil bore holes (C188 and C187) which were placed in the
Pier 18 and Pier 19 respectively which are shown as a general view in Figure C.5. The
groundwater table was found at about 9.3 ft (2.85 m) above the bottom of footing for
Pier 18 while for Pier 19 it was about 7.7 ft (2.35 m) below the bottom of the footing.
The boring log records for bore holes C188 and C187 are listed in Table C.4 and Table
C.5 respectively. The corrected hammer efficiency of %60 of SPT-N values (N1)60 are
presented in these tables also.
Figure C.5: Ramp C Bridge Project Site (MOT I70/I75) -General View

231
Table C.4: Soil Boring Logs C188 (Located Pier 18)
Depth Corrected
Below Generalized Soil Description SPT N- SPT N-
Footing Values Values
(ft) (blows/ft)
(N1)60
0 Gray Sandy Silt, Some Clay, Trace 22 20
Gravel, A-4a
5 The same as above A-4a 15 13
Table C.5: Soil Boring Logs C187 (Located Pier 19)
Depth Below Corrected

Footing Generalized Soil Description SPT N- SPT N-
(ft) Values Values
(blows/ft)
(N1)60
0 Gray Sandy Silt, Some Clay, Trace 22 17
Gravel, A-4a
20 The same as above A-4a 100+ 68
35 The same as above A-4a 100+ 62
232
The SPT N-values for the two bore holes which are analyzed within each stratum for
the logs of soil borings for Pier 18 and Pier 19 are shown in Figures C.6 and C.7.

0 20 40 60 80 100 120
0
5
10
Depth (ft)
15
20
25
30
35
40
45
Figure C.6: SPT- N Values Variations with Depth for C188 Bore Holes (Pier 18)

0 20 40 60 80 100 120
0
10
Depth (ft)
15
20
25
30
35
40
Figure C.7: SPT- N Values Variations with Depth for C187 Bore Holes (Pier 19)
233
The corrected SPT N-Values (N1)60 with depth for both Pier 18 and Pier 19 are shown in
Figures C.8 and C.9 respectively.
Corrected SPT N- Values (blows/ft)

0 20 40 60 80
0
5
10
Depth (ft)
15
20
25
30
35
40
45
Figure C.8: Corrected SPT- N Values Variations with Depth for Pier 18
Corrected SPT N- Values (blows/ft)

0 10 20 30 40 50 60 70 80
0
10
Depth (ft)
15
20
25
30
35
40
Figure C.9: Corrected SPT- N Values Variations with Depth for Pier 19
234

 Number of spans: Continuous span bridge with 20 spans.
 Total length of the ramp bridge: 2377 ft (724.5 m).
 Total width of the ramp bridge: 45.6 ft (13.9 m).
 Average contact pressure: 1.39 tsf for Pier 18

2.01 tsf for Pier 19

 Begin construction: 08/2004 for both Piers 18 and 19.
 Footing construction: 09/2004 for Pier 18 and 08/2004 for Pier 19.
 Pier wall construction:10/2004 for Pier 18 and 09/2004 for Pier 19.
 Soil backfill: 10/2004 for Pier 18 and 09/2004 for Pier 19.
 Girder beams placement: 11/2004 for both Piers 18 and 19.
 Deck construction: 05/2005 for both Piers 18 and 19.
 Bridge open to traffic: 08/2005 for both Piers 18 and 19.

 Settlement monitoring points: 5
 Tilt/Overturning points: 10
 Contact stress: 7 for Pier 18 and 5 for Pier 19
 Applied loading: yes

235
C-2 Summary of the Relevant Data of the Two Highway

Bridges in Ohio
This summary includes all the information data of the footings of the two
highway bridges in Ohio state. It includes the dimensions of the footings (B,L,Df), the
maximum contact pressure (qmax), the corrected SPT N-values (Ncorr), and measured
settlement (Sm) for each site of footing as shown in Table C.6 below :
Table C.6: Data Established for the Two Highway Bridges

Adapted from (Sargand and Masada 2006, FHWA)
Site Structure B L/B Df/B Ncorr qmax Sm Sm/B

# (ft) (dim) (dim) (dim) (tsf) (inch) (%)
25 Bridge 11- Central Pier 8.0 5.03 0.275 49 2.85 0.20 0.025
26 Bridge 12- Pier 18 21.0 2.45 0.21 38 1.39 0.70 0.033
27 Bridge 12- Pier 19 24.0 2.05 0.183 51 2.0.1 0.96 0.04
The definitions of all the variables used are the same as used in Table B.1.
The total soil unit weight (γ) was assumed to be (120) kcf. for all the three sites.
The depth of water (Dw) ranged between (0-9.3) ft. for all the three sites.
236
APPENDIX D
D-1 MONTE CARLO Method for 100,000 simulations by Meyerhof
Method for Total Settlement for S1-Site.
CDF of B for S1-Site
5
2
-1
-2
-3
-4
-5
14 15 16 17 18 19 20 21
Footing Width, B(ft)
Figure D.1: CDF of Footing Width, B for S1-Site by Meyerhof Method
CDF of L for S1-Site

5
2
-1
-2
-3
-4
-5
50 55 60 65 70 75
Footing Length, L(ft)
Figure D.2: CDF of Footing Length, L for S1-Site by Meyerhof Method

237
CDF of D for S1-Site

6
4
-2
-4
-6
1.3 1.4 1.5 1.6 1.7 1.8 1.9 2 2.1
Footing Depth, D(ft)
Figure D.3: CDF of Footing Depth, D for S1-Site by Meyerhof Method
CDF of P for S1-Site

5
2
-1
-2
-3
-4
-5
1500 2000 2500 3000 3500 4000 4500 5000
Applied Load, P(kips)
Figure D.4: CDF of Load Applied, P for S1-Site by Meyerhof Method

238
CDF for N60 for S1-Site

4
2
Normal Standard Variable
-1
-2
-3
-4
-5
-10 0 10 20 30 40 50 60 70 80 90
SPT-N Values
Figure D.5: CDF of SPT N-Values for S1-Site by Meyerhof Method
CDF of q for S1-Site

5
2
-1
-2
-3
-4
-5
1.5 2 2.5 3 3.5 4 4.5 5
Bearing Pressure, q(lb/ft2)
Figure D.6: CDF of Bearing Pressure, q for S1-Site by Meyerhof Method

239
CDF for Fd for S1-Site

4
2
-1
-2
-3
-4
1.026 1.028 1.03 1.032 1.034 1.036 1.038 1.04 1.042 1.044
Correction Depth Factor, Fd
Figure D.7: CDF of Depth Correction factor, Fd for S1-Site by Meyerhof Method
CDF for Settlement for S1-Site, Meyerhof Method

4
2
-1
-2
-3
-4
0 0.5 1 1.5
Total Settlement (inch)
Figure D.8: CDF of Total Settlement for S1-Site by Meyerhof Method

240
D-2 MONTE CARLO Method for 100,000 Simulations by Peck and

Bazaraa Method for Settlement for S1-Site.
CDF of Footing Width, B for S1-Site
5
2
-1
-2
-3
-4
-5
4 4.5 5 5.5 6 6.5
Footing Width, B(m)
Figure D.9: CDF of Footing Width, B for S1-Site by Peck and Bazaraa Method
CDF of Footing Length, L for S1-Site

5
2
-1
-2
-3
-4
-5
15 16 17 18 19 20 21 22 23
Footing Length, L(m)
Figure D.10: CDF of Footing Length, L for S1-Site by Peck and Bazaraa Method
241
CDF of Footing Depth, D for S1-Site

5
2
-1
-2
-3
-4
-5
0.42 0.44 0.46 0.48 0.5 0.52 0.54 0.56 0.58 0.6
Footing Depth,D(m)
Figure D.11: CDF of Footing Depth, D for S1-Site by Peck and Bazaraa Method
CDF of Applied Load, P for S1-Site

4
2
-1
-2
-3
-4
-5
0.8 1 1.2 1.4 1.6 1.8 2 2.2
Applied Load , P(kN) 4
x 10
Figure D.12: CDF of Applied Load, P for S1-Site by Peck and Bazaraa Method
242
CDF of Gamma, Unit Weight for S1-Site

5
2
-1
-2
-3
-4
-5
15 16 17 18 19 20 21 22 23 24 25
Gamma,(Kg/m3)
Figure D.13: CDF of Unit weight of Soil for S1-Site by Peck and Bazaraa Method
CDF of SPT N-Values, (N1)60 for S1-Site

5
2
-1
-2
-3
-4
-5
-50 0 50 100 150 200 250
SPT N-Values, (N1)60
Figure D.14: CDF of SPT N-Values, (N1)60 for S1-Site by Peck and Bazaraa Method
243
CDF of Bearing Pressure, q for S1-Site

5
2
-1
-2
-3
-4
-5
80 100 120 140 160 180 200 220 240
Bearing Pressure, q (kN/m2)
Figure D.15: CDF of Bearing Pressure, q for S1-Site by Peck and Bazaraa Method
CDF of Water Table Correction Factor, Cw for S1-Site

5
2
-1
-2
-3
-4
-5
2 4 6 8 10 12 14 16 18 20
Water Table Correction Factor, Cw
Figure D.16: CDF of Water Table Correction Factor, Cw for S1-Site

by Peck and Bazaraa Method
244
CDF of Depth Correction Factor, CD for S1-Site

5
2
-1
-2
-3
-4
-5
0.86 0.87 0.88 0.89 0.9 0.91 0.92 0.93
Depth Correction Factor, CD
Figure D.17: CDF of Depth Embedment Correction Factor, CD for S1-Site

by Peck and Bazaraa Method
CDF of Total Settlement for S1-Site by Peck&Bazarra Method

5
2
-1
-2
-3
-4
-5
0 0.5 1 1.5 2 2.5 3
Figure D.18: CDF of Total Settlement for S1-Site by Peck and Bazaraa Method
245
D-3 MONTE CARLO Method for 100,000 Simulations by Schmertmann

Method for Total Settlement for S1-Site.
CDF of B for S1-Site
5
2
-1
-2
-3
-4
-5
14 15 16 17 18 19 20 21
Footing Width, B(ft)
Figure D.19: CDF of Footing Width, B for S1-Site by Schmertmann Method
CDF of L for S1-Site

5
2
-1
-2
-3
-4
-5
50 55 60 65 70 75
Footing Length, L(ft)
Figure D.20: CDF of Footing Length, L for S1-Site by Schmertmann Method

246
CDF of D for S1-Site

6
4
-2
-4
-6
1.3 1.4 1.5 1.6 1.7 1.8 1.9 2 2.1
Footing Depth, D(ft)
Figure D.21: CDF of Footing Depth, D for S1-Site by Schmertmann Method
CDF of Gamma for S1-Site

5
2
-1
-2
-3
-4
-5
0.09 0.1 0.11 0.12 0.13 0.14 0.15
Unit Weight of soil ,Gamma(lb/ft3)
Figure D.22: CDF of Unit Weight of Soil, Gamma for S1-Site by Schmertmann Method
247
CDF of P for S1-Site

5
2
-1
-2
-3
-4
-5
1500 2000 2500 3000 3500 4000 4500 5000
Applied Load, P(kips)
Figure D.23: CDF of Applied Load, P for S1-Site by Schmertmann Method
CDF of N60 for S1-Site

5
2
-1
-2
-3
-4
-5
-60 -40 -20 0 20 40 60 80 100 120
SPT N-Values, N60
Figure D.24: CDF of SPT N-Values for S1-Site by Schmertmann Method

248
CDF of q for S1-Site

5
2
-1
-2
-3
-4
-5
1.5 2 2.5 3 3.5 4 4.5 5
Bearing Pressure, q(lb/ft2)
Figure D.25: CDF of Bearing Pressure, q for S1-Site by Schmertmann Method
CDF of Es for S1-Site

5
2
-1
-2
-3
-4
-5
-1 -0.5 0 0.5 1 1.5 2 2.5 3
Modulus of Elasticity, Es(lb/ft2) 6
x 10
Figure D.26: CDF of Modulus of Elasticity, Es for S1-Site by Schmertmann Method

249
CDF of Effective Stress for S1-Site

5
2
-1
-2
-3
-4
-5
0.16 0.18 0.2 0.22 0.24 0.26 0.28
Effective Vertical Stress,sigmadz (lb/ft2)
Figure D.27: CDF of Effective Vertical Stress for S1-Site by Schmertmann Method
CDF of C3 for S1-Site

5
2
-1
-2
-3
-4
-5
0.88 0.89 0.9 0.91 0.92 0.93 0.94 0.95
Shape Correction Factor, C3
Figure D.28: CDF of Shape Correction Factor, C3 for S1-Site by Schmertmann Method
250
CDF of C1 for S1-Site

5
2
-1
-2
-3
-4
-5
0.958 0.96 0.962 0.964 0.966 0.968 0.97 0.972 0.974
Depth Correction Factor, C1
Figure D.29: CDF of Depth Correction Factor, C1 for S1-Site by Schmertmann Method
CDF of(q-sigmadz ) for S1-Site

5
2
-1
-2
-3
-4
-5
1.5 2 2.5 3 3.5 4 4.5 5
(q-sigmadz )
Figure D.30: CDF (q-sigmadz) for S1-Site by Schmertmann Method

251
CDF of Total Settlement for S1-Site

4
2
-1
-2
-3
-4
0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3
Figure D.31: CDF of Total Settlement for S1-Site by Schmertmann Method

Reliability Analysis of Settle

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Reliability Analysis of Settle

Uploaded by

Copyright:

Available Formats

RELIABILITY ANALYSIS OF SETTLEMENT FOR SHALLOW

Aseel Yassen Ahmed

Presented to the Faculty of

The Graduate College at the University of Nebraska

In Partial Fulfillment of Requirements

For the Degree of Doctor of Philosophy

Major: Civil Engineering

Under the Supervision of Professors Andrzej S. Nowak and Maria M. Szerszen

All rights reserved

INFORMATION TO ALL USERS

Aseel Yassen Ahmed, Ph.D.

University of Nebraska, 2013

Advisers: Andrzej S. Nowak and Maria M. Szerszen

The design of shallow foundations is becoming an attractive choice for supporting

highway bridge structures. The settlement of structures built on soil is a subject of

considerable interest to practicing engineers since excessive settlements can lead to

design of shallow foundations is most often governed by settlement requirements.

A reliability-based approach can be applied for designing shallow foundations for

footing. The performance of geotechnical structures, including shallow foundations depends

BY THE NAME OF GOD

This dissertation is dedicated to

The soul of the greatest woman in the world

My devoted beloved husband Mohammed for his

unconditioned love and support

My three beautiful angels Zainab, Ahmed, and Alhussain

My lovely father Dr. Yassen

My lovely brothers Rafed and Mustafa

I wish to express my sincere thanks to Professor Andrzej S. Nowak, my academic

and help during the final preparations of this study.

Mohammed Al Bardan, my colleagues at the University of Nebraska-Lincoln, for their

friendship, kindness, and support during living in Nebraska.

The Civil Engineering Department at University of Nebraska-Lincoln is

through this study.

Especially, I wish to express my sincere thanks and love to my kindly mother

LIST OF FIGURES ........................................................................................ix

Chapter 2. TYPES OF SHALLOW FOUNDATONS ................................. 9

Chapter 3. STRUCTURAL RELIABILITY THEORY .............................. 18

3.2.2.1 Mean Value .................................................................................. 26

CHAPTER 4. EVALUATION OF SOIL PARAMETERS FOR

4.2.4 Elastic Soil Parameters ..................................................................... 67

4.2.4.1 Poisson’s Ratio ............................................................................. 67

4.3 Description and Background Information of the Highway Bridges

4.3.2 Highway Bridges Selected in Ohio .................................................. 76

CHAPTER 5. SETTLEMENT ANALYSIS FOR SHALLOW

5.8.2 Influence of Length to Width Ratio (L/B) ..................................... 131

5.8.3 Influence of Footing Depth Ratio (Df/B) ....................................... 132

5.8.4 Influence of Depth of Water Table (Dw/B) ................................... 134

5.8.5 Influence of Axial Load (P) ........................................................... 135

5.8.6 Influence of SPT- N Values .......................................................... 136

5.9 Estimation of Differential Settlement and Angular Distortion of

CHAPTER 6. RELIABILITY ANALYSIS OF SETTLEMENT FOR

CHAPTER 7. CONCLUSIONS AND RECOMMENDATIONS ........... 180

7.2 Conclusions ............................................................................................................ 182

7.3 Recommendations for Future Work ............................................................. 191

7.4 Implementations of this Study ........................................................................ 192

REFERENCES ............................................................................................ 193

Figure 2.1: Typical Spread Footing with ………...………………………….…………..10

Figure 5.37: Influence of Length to Width Ratio (L/B) on Settlement of Shallow

Table 5.2: Correlation Factors (Adapted from Coduto, 2000) …………………………101

Table 5.3: Comparison of Measured and Predicted Values of Settlement .…….……...107

Table 6.2: Average of Reliability Indices for Differential Settlement, Related to

soils and loads.

and properly accounts for the intended function of the structure.