Professional Documents
Culture Documents
Reliability Analysis of Settle
Reliability Analysis of Settle
FOUNDATIONS IN BRIDGES
by
A DISSERTATION
Lincoln, Nebraska
July, 2013
UMI Number: 3590300
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
serviceability problems or even failure states in the structural foundations. Therefore, the
which their design needs knowledge about the distribution of settlement under a given
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
To my lovely mother
AKNOWLEDGEMENTS
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
I would like to thank Dr. Anna Rakoczy, Mr Hooman Ghasemi, and Mr.
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
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
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
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 5.56: 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 6.5: Reliability Indices for Total Settlement by all Methods for S5-Site .……...158
Figure 6.6: Reliability Indices for Total Settlement by all Methods for S6-Site .……...159
Figure 6.7: Reliability Indices for Total Settlement by all Methods for S7-Site ………159
Figure 6.8: Reliability Indices for Total Settlement by all Methods for S8-Site ………160
Figure 6.9: Reliability Indices for Total Settlement by all Methods for S9-Site ………160
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 6.13: Reliability Indices for Total Settlement by all Methods for S15-Site ……162
Figure 6.14: Reliability Indices for Total Settlement by all Methods for S16-Site ……163
Figure 6.15: Reliability Indices for Total Settlement by all Methods for S17-Site ……163
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 6.18: Reliability Indices for Total Settlement by all Methods for S22-Site .…...165
Figure 6.19: Reliability Indices for Total Settlement by all Methods for S23-Site .…...165
Figure 6.20: Reliability Indices for Total Settlement by all Methods for S24-Site ……166
Figure 6.21: Reliability Indices for Total Settlement by all Methods for S25-Site ……166
Figure 6.22: Reliability Indices for Total Settlement by all Methods for S26-Site ……167
Figure 6.23: Reliability Indices for Total Settlement by all Methods for S27-Site ……167
Figure 6.24: Reliability Indices for Differential Settlement for S1, S2-Sites in Bridge 1
…………………………………………………………………………………………..170
Figure 6.25: Reliability Indices for Differential Settlement for S3, S5-Sites in Bridge 2
…………………………………………………………………………………………..170
Figure 6.26: Reliability Indices for Differential Settlement for S5, S4-Sites in Bridge 2
…………………………………………………………………………………………..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
Figure 6.29: Reliability Indices for Differential Settlement for S14, S15-Sites in Bridge 4
…………………………………………………………………………………………..172
Figure 6.30: Reliability Indices for Differential Settlement for S16, S17-Sites in Bridge 5
…………………………………………………………………………………….…….173
Figure 6.31: Reliability Indices for Differential Settlement for S21, S22-Sites in Bridge 8
…………………………………………………………………………………………..173
Figure 6.32: Reliability Indices for Differential Settlement for S23, S24-Sites in Bridge 9
………………………………………………………………………………………….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 5.4: 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 6.3: 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
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
One of the most important problems of soil mechanics is stability related to the
lead to serviceability problems or even failure states in the structural foundations. The
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
shallow foundations where applied loads, soil, properties, and resistance parameters are
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
attractive alternative where the subsurface soils are suitable, since they require less time
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.
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
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
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
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,
For each structural and soil condition components, the reliability analysis requires the
following:
4
Determination of the statistical parameters for load and resistance that represent the settlement
Development of the reliability analysis procedure for the most commonly used settlement
expressed in terms of reliability indices, which in this research will be calculated for
The study will address three major tasks, which are stated as:
procedure or procedures.
shallow foundations.
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).
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).
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
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
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
includes the problem statement, research objective, and finally the technical background that
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
CHAPTER 2
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
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
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
Shallow foundations encompass two types: spread footing foundations and mat 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
There are different shapes and sizes of spread footings to accommodate the needs of
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
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
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
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
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
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.
CHAPTER 3
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
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
nonzero probability of failure. Therefore, the convenient way for safety measurement of
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.
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).
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:
i 1
Px (xi ) = 1 (3.2)
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
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
0.8
Probability
0.6
0.4
0.2
0
1 2 3 4 5 6 7 8 9 10
Events
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
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
Since a continuous random variable follows the three axioms of probability, the
P(-∞ ≤ X ≤ +∞ ) = f
x ( x)dx =1 (3.5)
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
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
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
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
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.
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)
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
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
n
1
S2 =
n 1 i1
( xi X ) 2 (3.14)
2
σ = 2 ( x ) f x ( x)dx (3.15)
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
For a discrete and continuous random variable, the standard deviation of a population
n 2
x n( x )
i 1 i
S= (3.16)
n 1
V = │ │ (3.17)
Even though the mean may be a negative value, the coefficient of variation
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
The normal probability distribution, also known as Gaussian distribution, is the most
engineering applications due to its simplicity. The probability density function (PDF) for a
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
x
1 1 x 2
Fx(x) = exp
dx (3.19)
2 2
Fx(x)
μx = 6, σx =3
μx = 3, σx =2
μx = 1, σx =1
Random Variable X
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
1 1
φ(z) = exp z 2 (3.20)
2 2
φ(z)
Random Variable X
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
z = Ф-1(p) (3.22)
normal distribution, where ln(X) is the natural logarithm (to the base e). The probability
where, µY is the mean value of the lognormal distribution, and σY is the standard deviation of
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
fx(x)
Random Variable X
determined based on its relation with the normal distribution. The following formula is used
y y
Fx ( x) FY ( y)
(3.26)
y
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
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
CDF
F(x)
Random Variable X
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
Table 3.1: Examples of Relationship between the Standard Normal Variable and the
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
Standard
Normal
Variable
CDF
F(x)
3
Mean
0
Random Variable X
-1
Standard Deviations
-2
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
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:
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
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
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
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
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
= P(g<0)
If R and Q are statically independent and have probability density functions fR and
Pf FR (li ) f Q (li )dli (3.30)
because the integration requires numerical techniques, and these techniques may not be
other indirect procedures and simulation techniques, such as evaluating the reliability index
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
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
Pf = Φ(-β) (3.34)
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.
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
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-
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
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.
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.
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
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
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
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,
ln R
ln R ln Q
2
Q
2 (3.40)
ln R 2 ln Q 2 VR VQ
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
For the case of a nonlinear state function with uncorrelated random variables,
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
g ( X 1 , X 2 ,..., Xn )
n (3.42)
(a
i 1
i Xi ) 2
48
g
where, ai = are evaluated at µXi.
xi
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.
Simulation is the process of replacing reality with theoretical and experimental models.
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
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
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
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
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
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:
ng ( 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
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
CHAPTER 4
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
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
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.
γ = (W / V) (4.1)
where, W is the weight of soil including moisture, and V is the volume of soil.
γd = (Ws / V) (4.2)
The dry unit weight (γd) can be calculated in terms of the unit weight (γ), as
presented below:
γd = (4.3)
1 w
Typical ranges of unit weight (γ) for various soils are presented in Table (4.1).
53
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
Table 4.2: In-Situ Dry Unit Weight Statistics for Various Cohesionless Soils
(Adapted from Chen, 2004)
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 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
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
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
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
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
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
Mechanism (Em)
(not recommended)
For most correlations, the SPT N-value is corrected for field procedures, and an ER of
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
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)
Over 30 Hard
62
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:
where,(N1)60 is the N60 value corrected to a reference overburden stress of one atmosphere
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,
pa
CN = (4.6)
'z
in which, Pa is the atmospheric stress = (100 kPa or about 2000 Ib/ft2) and σ’z is the
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
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 4.4: 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
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
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.
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).
h
v (4.8)
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)
Clay:
Soft sensitive
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
Medium dense 0.2-0.35
Dense 0.3-0.4
69
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
Axial Strain, ε1
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)
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)
Es (ksi)
Clay:
Loess 2.08-8.33
Silt 0.278-2.7
Fine Sand:
Loose 1.11-1.67
Medium dense 1.67-2.78
Dense 2.78-4.17
Sand:
Loose 1.39-4.17
Medium dense 4.17-6.94
Dense 6.94-11.11
Gravel:
Loose 4.17-11.11
Medium dense 11.11-13.89
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)
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)
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:
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
010 ** Route 84 NA
Manchester, Connecticut
soils and short term settlement was increased, so these sites were excluded from the
comparison.
NA : Not Available.
75
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
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
I- FRA-670-0380 Project
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
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 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.
at the northeast end of Ramp C bridge constructed as part of the massive I-70/I-75
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
The period of construction was almost a year for both piers 18 and 19, which were
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
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
CHAPTER 5
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
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
structures, including shallow foundations, depends on loads, soil properties, dimensions and
practicing engineers, since excessive settlements can lead to serviceability problems or even
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
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
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
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
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
cohesionless soils, which is the instantaneous deformation of the soil mass that occurs after
Allowable total settlement (δa) contains a safety factor and it should be compared to predicted
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
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
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
In the design of a foundation, the allowable differential settlement (δDa) should satisfy
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.
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
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
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
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
Duncan and Tan, (1991) illustrated the subdivision of settlement for bridges into three
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,
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
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
progressively from the abutments of the bridge towards the center of it.
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: Tolerable Movement Criteria of Angular Distortion for Highway Bridges
dimensionless and should be used with the same units for δD and SL.
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
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
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
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.
Legend:
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.
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
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
For most of the methods, the following notations are used, so only the correction factors
Ic = compressibility index
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
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
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
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.
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.
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
Cw = (5.5-a)
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
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)
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
C1 = 1 – 0.5 ( ) (5.7-a)
= effective vertical stress at depth D below the ground surface (Ib/ft2); t = time since
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,
in which, β0, β1 = correlation factors obtained from Table 5.2; OCR= overconsolidation ratio.
Soil Type β0 β1
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.
To compute the magnitude of the strain influence factor for each layer, the peak value,
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,
The strain influence factor at the midpoint of each layer can be calculated from the
in which, zf = depth from bottom of footing to midpoint of layer; Izs = Iz for a square footing;
This method establishes an empirical relationship between average (SPT) blow counts,
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:
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
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
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
Measured Settlements
1.2
1
Settlement (in.)
0.8
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
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,
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.
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
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
Table 5.5: Statistical Parameters of Bias Factor for Settlement (δM /δP) of the
Six Prediction Methods
Where, Max = maximum value; Min = minimum value; STDEV = standard deviation ; COV
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
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
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
Figure 5.14-b: Distribution of Bias Factor for Settlement for Hough Method (1959)
114
Meyerhof Method
1.2
Measured Settlement(in.)
1
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
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 5.16-a: Predicted Versus Measured Settlements for Peck and Bazarra Method
0.25
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
D'Appolonia Method
Measured Settlement(in.) 1.2
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
Predicted Settlement(in.)
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
Measured Settlement(in.) 1.2
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
Predicted Settlement(in.)
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
0.4
0.2
0
0.00 0.50 1.00 1.50 2.00
Predicted Settlement(in.)
0.25
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
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
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
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
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
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.
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
0
0 0.2 0.4 0.6 0.8 1 1.2
-1
-2
-3
Predicted Settlement
y = 1.6099ln(x) + 1.7525
2
Standard Normal Variable
R² = 0.9846
0
0.00 0.20 0.40 0.60 0.80 1.00 1.20
-1
-2
-3
Predicted Settlement
Figure 5.23: CDF of Predicted Settlements by Peck and Bazaraa Method for 23 sites
y = 2.4736ln(x) + 2.3527
2
Standard Normal Variable
R² = 0.9888
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
-1
-2
-3
Predicted Settlement
0
0 0.2 0.4 0.6 0.8 1 1.2
-1
-2
-3
Predicted Settlement
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 5.26: CDF of Predicted Settlements by Burland and Burbidge Method for 23 sites
124
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
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
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
0
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4
-1
Settlement ratio
-2
Normal
Lognormal
-3
0
0.0 1.0 2.0 3.0 4.0 5.0 6.0
-1
Settlement ratio
-2
Normal
Lognormal
-3
3
Standard Normal Variable
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)
3
Standard Normal Variable
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
3
Standard Normal Variable
0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
-1
-2
Settlement ratio
-3 Normal
Lognormal
-4
0
0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0
-1
Settlement ratio
-2 Normal
Lognormal
-3
Settlement Ratio (δM/δP)
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
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.
2
Standard Normal Variable
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
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.
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)
60
50
40
30
20
10
0
0 2 4 6 8 10 12 14 16 18 20 22
B (ft)
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.
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
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.
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
4
Settlementδ (in.)
0
0 1 2 3 4 5 6 7 8 9 10 11
L/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
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
4
Settlementδ (in.)
0
0 0.2 0.4 0.6 0.8 1 1.2
Df/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
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
Foundations on Cohesionless Soils
135
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
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.
0.8
0.6
0.4
0.2
0
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
P (kips)
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
12000
10000
P (kips)
8000
S1-site
6000
S2-Site
4000
2000
0
0 1 2 3 4 5 6
Settlement (in.)
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
0.8
0.6
0.4
0.2
0
0 10 20 30 40 50 60 70
SPT N-Values
4
Settlementδ (in.)
0
0 10 20 30 40 50 60 70 80
SPT N-Values
geotechnical design process. Usually the settlement criterion controls the design process for
shallow foundations on cohesion less soils rather than bearing capacity failure.
bridge, due to the variation of load carried by various shallow foundations and the non-
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
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
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
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
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
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
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
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
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
Figure 5.48: CDF of Actual Differential Settlement for 20 sites of Spread Footings
143
0.25
0.1
0.05
0
0.11
0.22
0.33
0.44
0.055
0.165
0.275
0.385
0.495
0
Differential Settlement
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
Differential Settlement (in.)
0.25
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
Differential Settlement
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
Differential Settlement (in.)
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
Table 5.9: Statistical Parameters of Bias Factor for Differential Settlement (δD/δDP) of
the Two Selected Methods
Figures 5.53 and 5.55 show graphs for the variation involved in the bias factor for
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
0.1
0.05
0
0.43
0.76
1.09
1.42
1.75
2.08
2.41
2.74
0.1
2 y = 1.0532ln(x) - 0.0395
Standard Normal Variable
R² = 0.9502
0
0 0.5 1 1.5 2 2.5 3
-1
-2
-3
Bias Factor of Differential Settlement
Figure 5.54: CDF of Bias Factor for Differential Settlement for 20 sites
by Hough Method (1959)
148
0.35
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
y = 1.0184ln(x) + 0.0493
2
Standard Normal Variable
R² = 0.9409
0
0 0.5 1 1.5 2 2.5 3
-1
-2
-3
Bias Factor of Differential Settlement
Figure 5.56: CDF of Bias Factor for Differential Settlement for 20 sites
by Schmertmann Method (1978)
149
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.
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
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
objective was addressed by the evaluation of six selected prediction methods to estimate the
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
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
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
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.
The limit state function is formulated for the considered structural components and is
This assumption for the limit state function is valid for total and differential settlements.
154
The nominal resistance (R) was assigned as the tolerable settlement (δT), which is
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
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
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
in which, µδP = mean value of predicted settlement, VδP = coefficient of variation for
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
Reliability indices determined for total settlement by the six prediction methods for the
before, are calculated from Equation 6.2. Reliability indices are shown in Figures 6.1 through
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
Figure 6.1: Reliability Indices for Total Settlement by all Methods for S1-Site
157
5.00 Hough
Reliability Index, β
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
Figure 6.2: Reliability Indices for Total Settlement by all Methods for S2-Site
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
Tolerable Settlement (inch)
Figure 6.3: Reliability Indices for Total Settlement by all Methods for S3-Site
158
3.00 Hough
Reliability Index, β
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
Tolerable Settlement (inch)
Figure 6.4: Reliability Indices for Total Settlement by all Methods for S4-Site
4.00
Meyerhof
3.00
2.00 Peck&Bazaraa
1.00
D'Appolonia
0.00
Schmertmann
-1.00
-2.00 Burland&Burbidge
-3.00
0 0.5 1 1.5 2 2.5 3 3.5
Figure 6.5: Reliability Indices for Total Settlement by all Methods for S5-Site
159
5.00 Hough
Reliability Index,β
4.00
Meyerhof
3.00
Peck&Bazaraa
2.00
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
Tolerable Settlement (inch)
Figure 6.6: Reliability Indices for Total Settlement by all Methods for S6-Site
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
Figure 6.7: Reliability Indices for Total Settlement by all Methods for S7-Site
160
5.00
Hough
4.00
Reliability Index, β
Meyerhof
3.00
Peck&Bazaraa
2.00
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
Tolerable Settlement (inch)
Figure 6.8: Reliability Indices for Total Settlement by all Methods for S8-Site
5.00 Hough
4.00
Reliability Index, β
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
Figure 6.9: Reliability Indices for Total Settlement by all Methods for S9-Site
161
5.00 Hough
Reliability Index, β
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
Tolerable Settlement (inch)
Figure 6.10: Reliability Indices for Total Settlement by all Methods for S10-Site
5.00 Hough
Reliability Index, β
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
Figure 6.11: Reliability Indices for Total Settlement by all Methods for S11-Site
162
4.00
Hough
Reliability Index, β
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
Tolerable Settlement (inch)
Figure 6.12: Reliability Indices for Total Settlement by all Methods for S14-Site
3.00
Hough
Reliability Index, β
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
Figure 6.13: Reliability Indices for Total Settlement by all Methods for S15-Site
163
5.00 Hough
Reliability Index, β
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
Figure 6.14: Reliability Indices for Total Settlement by all Methods for S16-Site
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
Figure 6.15: Reliability Indices for Total Settlement by all Methods for S17-Site
164
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
Figure 6.16: Reliability Indices for Total Settlement by all Methods for S20-Site
4.00 Meyerhof
3.00
Peck&Bazaraa
2.00
1.00 D'Appolonia
0.00
Schmertmann
-1.00
-2.00 Burland&Burbidge
-3.00
0 0.5 1 1.5 2 2.5 3 3.5
Figure 6.17: Reliability Indices for Total Settlement by all Methods for S21-Site
165
4.00 Hough
Reliability Index, β
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
Figure 6.18: Reliability Indices for Total Settlement by all Methods for S22-Site
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
Tolerable Settlement (inch)
Figure 6.19: Reliability Indices for Total Settlement by all Methods for S23-Site
166
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
Tolerable Settlement (inch)
Figure 6.20: Reliability Indices for Total Settlement by all Methods for S24-Site
5.00
Meyerhof
4.00
3.00 Peck&Bazaraa
2.00 D'Appolonia
1.00
Schmertmann
0.00
-1.00 Burland&Burbidge
-2.00
0 0.5 1 1.5 2 2.5 3 3.5
Figure 6.21: Reliability Indices for Total Settlement by all Methods for S25-Site
167
4.00 Hough
Reliability Index, β
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
Tolerable Settlement (inch)
Figure 6.22: Reliability Indices for Total Settlement by all Methods for S26-Site
4.00
Hough
Reliability Index, β
3.00
Meyerhof
2.00
Peck&Bazaraa
1.00
D'Appolonia
0.00
Schmertmann
-1.00
-2.00 Burland&Burbidge
0 0.5 1 1.5 2 2.5 3 3.5
Figure 6.23: Reliability Indices for Total Settlement by all Methods for S27-Site
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
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
Table 6.1: Average of Reliability Indices for Total Settlement, Related to Tolerable
Settlements and their Probability of Failures for 23 Sites
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
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
5.000
Reliability Index, β
4.000
3.000
Hough
2.000 Schmertmann
1.000
0.000
0 0.5 1 1.5 2 2.5 3 3.5
Tolerable Settlement (inch)
Figure 6.24: Reliability Indices for Differential Settlement for S1, S2-Sites in Bridge 1
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
Figure 6.25: Reliability Indices for Differential Settlement for S3, S5-Sites in Bridge 2
171
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
Figure 6.26: Reliability Indices for Differential Settlement for S5, S4-Sites in Bridge 2
6.000
5.000
Reliability Index, β
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
7.000
6.000
Reliability Index, β
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
Tolerable Settlement (inch)
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
Tolerable Settlement (inch)
Figure 6.29: Reliability Indices for Differential Settlement for S14, S15-Sites in Bridge 4
173
5.000
Reliability Index, β
4.000
3.000
Hough
Schmertmann
2.000
1.000
0.000
0 0.5 1 1.5 2 2.5 3 3.5
Tolerable Settlement (inch)
Figure 6.30: Reliability Indices for Differential Settlement for S16, S17-Sites in Bridge 5
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
Tolerable Settlement (inch)
Figure 6.31: Reliability Indices for Differential Settlement for S21, S22-Sites in Bridge 8
174
4.000
Reliability Index, β
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
Tolerable Settlement (inch)
Figure 6.32: Reliability Indices for Differential Settlement for S23, S24-Sites in Bridge 9
5.000
Reliability Index, β
4.000
Hough
3.000
Schmertmann
2.000
1.000
0.000
0 0.5 1 1.5 2 2.5 3 3.5
Tolerable Settlement (inch)
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
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
economic analysis, which are the most two important factors. Generally, reliability indices
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,
Reliability index calculations were performed in this study assuming total and
indicated, one for total settlement and the other for differential settlement as shown below:
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
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:
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
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
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
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
CHAPTER 7
7.1 SUMMARY
Safety and reliability of bridge infrastructure is a major concern for many state highway
(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
Even though shallow foundations have demonstrated cost advantages and satisfactory
field performance, deep foundations have remained the principal means for supporting
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
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
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
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
differential settlement, and angular distortion was calculated for the sites that had differential
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:
There are more than a dozen methods that have been examined to estimate the behavior
(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
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
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
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.
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
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,
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
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
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
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
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
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
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
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
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
6. Reliability Analysis
Reliability analysis was performed using the FOSM method and Monte Carlo
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
Reliability indices were calculated from equation 6.2, depending on the assumption of
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
was to design the service limit state for both tolerable and predicted settlements for
The main objective of this study was to develop the reliability-based design
analysis.
The main contribution of this study was formulation of rational acceptability criteria
prediction methods for shallow foundations in highway bridges. The main effort and
Schmertmann methods. This work is underway with the goal of obtaining better
statistical parameters.
interaction problems. The formulation of the limit state function is more difficult for
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
construct, require less time and their performance for highway bridges is determined
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
REFERENCES
AASHTO LRFD Bridge Design Specifications (2012). “American Association of State Highway and
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 (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.
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 II.
Final report FHWA- NHI-06-089, 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
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
APPENDIX B
B-1 Historical Summary of Candidate Bridges in Northeast of
United States
SUBSURFACE INFORMATION
I- Typical soil Profile:
SUBSURFACE INFORMATION
I- Typical soil Profile:
Open to traffic: NA
SUBSURFACE INFORMATION
I- Typical soil Profile:
sides.
Collierville, NY.
SUBSURFACE INFORMATION
I- Typical soil Profile:
MA.
SUBSURFACE INFORMATION
I- Typical soil Profile:
none as built.
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
Geological Setting: 100 ft. high embankment of lacustrine; silt 100 yards
downstream. 20 ft. high outcrop of mica schist 100 yards upstream.
SUBSURFACE INFORMATION
I- Typical soil Profile:
difficult.
216
Settlement points: 24
Tilt/Overturning points: 6
SUBSURFACE INFORMATION
I- Typical soil Profile:
Number of spans: 4
Span lengths: 114 ft. , 132 ft. , 162 ft. , and 174 ft.
Settlement points: 8
Tilt/Overturning points: 3
218
SUBSURFACE INFORMATION
I- Typical soil Profile:
Number of spans: 4
Span lengths: 145 ft. , 220 ft. , 220 ft. , and 175 ft.
Settlement points: 25
Tilt/Overturning points: 8
SUBSURFACE INFORMATION
I- Typical soil Profile:
Settlement points: 20
Tilt/Overturning points: 10
221
SUBSURFACE INFORMATION
I- Typical soil Profile:
Number of spans: 5
Span lengths: 155 ft. , 160 ft. , 171 ft. , 180 ft. , and 195 ft.
Settlement points: 20
(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 :
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=
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
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
Figure C.1 presents the view of the bridge structure taken after the placement of the girder
beams.
SUBSURFACE INFORMATION
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
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
20 The same as above A-2-4 98
25 The same as above A-2-4 61
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
Table C.3: Average of SPT N-Values of the Two Soil Boring Logs
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
5
Depth (ft)
10
15
20
25
30
Figure C.2 : SPT N-Values variations with Depth for Project FRA-670-0380
227
0 10 20 30 40 50 60 70 80
0
5
Depth (ft)
10
15
20
25
30
Project FRA-670-0380
Span lengths of bridge: North span=102.9 ft. and South span=100.2 ft.
central footing.
Tilt/Overturning points: 10
Contact stress: 6
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
I-70
SUBSURFACE INFORMATION
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
Figure C.5: Ramp C Bridge Project Site (MOT I70/I75) -General View
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
10 The same as above A-4a 19 16
15 The same as above A-4a 48 38
20 The same as above A-4a 62 47
25 The same as above A-4a 47 34
30 The same as above A-4a 42 30
35 The same as above A-4a 111 76
40 The same as above A-4a 100 66
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.
15
20
25
30
35
40
45
Figure C.6: SPT- N Values Variations with Depth for C188 Bore Holes (Pier 18)
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
15
20
25
30
35
40
45
Figure C.8: Corrected SPT- N Values Variations with Depth for Pier 18
10
Depth (ft)
15
20
25
30
35
40
Figure C.9: Corrected SPT- N Values Variations with Depth for Pier 19
234
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.
Tilt/Overturning points: 10
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 :
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
Standard Normal Variable
-1
-2
-3
-4
-5
14 15 16 17 18 19 20 21
Footing Width, B(ft)
2
Standard Normal Variable
-1
-2
-3
-4
-5
50 55 60 65 70 75
Footing Length, L(ft)
4
Standard Normal Variable
-2
-4
-6
1.3 1.4 1.5 1.6 1.7 1.8 1.9 2 2.1
Footing Depth, D(ft)
2
Standard Normal Variable
-1
-2
-3
-4
-5
1500 2000 2500 3000 3500 4000 4500 5000
Applied Load, P(kips)
2
Normal Standard Variable
-1
-2
-3
-4
-5
-10 0 10 20 30 40 50 60 70 80 90
SPT-N Values
2
Standard Normal Variable
-1
-2
-3
-4
-5
1.5 2 2.5 3 3.5 4 4.5 5
Bearing Pressure, q(lb/ft2)
2
Normal Standard Variable
-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
2
Normal Standard Variable
-1
-2
-3
-4
0 0.5 1 1.5
Total Settlement (inch)
2
Standard Normal Variable
-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
2
Standard Normal Variable
-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
2
Standard Normal Variable
-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
2
Standard Normal Variable
-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
2
Standard Normal Variable
-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
2
Standard Normal Variable
-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
2
Standard Normal Variable
-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
2
Standard Normal Variable
-1
-2
-3
-4
-5
2 4 6 8 10 12 14 16 18 20
Water Table Correction Factor, Cw
2
Standard Normal Variable
-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
2
Standard Normal Variable
-1
-2
-3
-4
-5
0 0.5 1 1.5 2 2.5 3
Total Settlement (inch)
Figure D.18: CDF of Total Settlement for S1-Site by Peck and Bazaraa Method
245
2
Standard Normal Variable
-1
-2
-3
-4
-5
14 15 16 17 18 19 20 21
Footing Width, B(ft)
2
Standard Normal Variable
-1
-2
-3
-4
-5
50 55 60 65 70 75
Footing Length, L(ft)
4
Standard Normal Variable
-2
-4
-6
1.3 1.4 1.5 1.6 1.7 1.8 1.9 2 2.1
Footing Depth, D(ft)
2
Standard Normal Variable
-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
2
Standard Normal Variable
-1
-2
-3
-4
-5
1500 2000 2500 3000 3500 4000 4500 5000
Applied Load, P(kips)
2
Standard Normal Variable
-1
-2
-3
-4
-5
-60 -40 -20 0 20 40 60 80 100 120
SPT N-Values, N60
2
Standard Normal Variable
-1
-2
-3
-4
-5
1.5 2 2.5 3 3.5 4 4.5 5
Bearing Pressure, q(lb/ft2)
2
Standard Normal Variable
-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
2
Standard Normal Variable
-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
2
Standard Normal Variable
-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
2
Standard Normal Variable
-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
2
Standard Normal Variable
-1
-2
-3
-4
-5
1.5 2 2.5 3 3.5 4 4.5 5
(q-sigmadz )
2
Standard Normal Variable
-1
-2
-3
-4
0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3
Total Settlement (inch)