Evaluation of the TxDOT Texas Cone Penetration Test and ... · Evaluation of the TxDOT Texas Cone...

Evaluation of the TxDOT Texas Cone Penetration Test and Foundation Design

Method Including Correction Factors, Allowable Total Capacity, and Resistance

Factors at Serviceability Limit State

by

Rozbeh B. Moghaddam, BSCE, P.E., M.B.A.

A Dissertation

In

Civil Engineering

Submitted to the Graduate Faculty

of Texas Tech University in

Partial Fulfillment of

the Requirements for

the Degree of

DOCTOR OF PHILOSOPHY

Approved by

William D. Lawson, P.E., Ph.D.

Co-chair of the Committee

Priyantha W. Jayawickrama, Ph.D.

Co-Chair of the Committee

Hoyoung Seo, Ph.D., P.E.

Committee Member

James G. Surles, Ph.D.

Committee Member

Mark Sheridan

Dean of the Graduate School

August 2016

Texas Tech University, Rozbeh B. Moghaddam August 2016

ii

ACKNOWLEDGEMENTS

I would like to thank my family: Akram, Mehdi, Hamzeh and Paola, Yeganeh and

Pouria, and Yekta, for their invaluable support throughout all these years.

I would like to express my deepest appreciation to my research committee members Dr.

Lawson, Dr. Jay, Dr. Seo, and Dr. Surles for their expert advice, guidance, and

understanding during my graduate studies and during the development of this

dissertation. I am grateful for their encouragement to pursue independent thinking.

I sincerely thank the Texas Department of Transportation for sponsoring the TCP

Reliability research study and Terracon, PSI, Rick Coffman (University of Arkansas),

The Arkansas State Highway and Transportation Department, the Missouri Department

of Transportation, the Louisiana Department of Transportation and Development, and

the New Mexico Department of Transportation in their assistance in providing data for

this study.

Last but not least, special thanks to all my friends in Texas, United States, and around

the World for their support during my studies.

Texas Tech University, Rozbeh B. Moghaddam, August 2016

iii

TABLE OF CONTENTS

ACKNOWLEDGMENTS……………………………………………………………………………………………….. ii

ABSTRACT ....................................................................................................................... viii

LIST OF TABLES .................................................................................................................. x

LIST OF FIGURES ............................................................................................................... xi

NOTATIONS ..................................................................................................................... xiii

CHAPTER I: INTRODUCTION .............................................................................................. 1

TCP FIELD TEST ............................................................................................................ 6

TCP FOUNDATION DESIGN CHARTS .............................................................................. 7

ORGANIZATION OF THE DISSERTATION ......................................................................... 8

CHAPTER II: HAMMER EFFICIENCY AND CORRECTION FACTORS FOR THE TXDOT

TEXAS CONE PENETRATION TEST ................................................................................... 11

ABSTRACT ..................................................................................................................... 11

INTRODUCTION ............................................................................................................. 12

THE TXDOT TEXAS CONE PENETRATION TEST ........................................................ 13

Description of the TCP test ....................................................................................... 13

History and development of the TCP test ................................................................. 14

COMPARISON OF SPT VS. TCP TESTS ......................................................................... 18

Correlation of blowcount results ............................................................................... 20

DEVELOPMENT OF CORRECTION FACTORS FOR SPT ................................................. 21

Standardization of the SPT blowcount ..................................................................... 21


iv

Early work by Fletcher and others ............................................................................ 22

Effect of Drilling Rod and Type of Sampler ............................................................ 23

Effect of Driving Technique and Hammer Type ...................................................... 24

Effect of Borehole Diameter ..................................................................................... 26

Development of correction factors for SPT N-values ............................................... 27

The overburden pressure correction for SPT ............................................................ 28

Contemporary practice for correcting SPT blowcount data ..................................... 30

Liquefaction studies .................................................................................................. 32

CORRECTIONS TO TCP BLOWCOUNT .......................................................................... 33

Need for TCP correction factors ............................................................................... 33

TxDOT policy ........................................................................................................... 33

RESEARCH DESIGN AND METHOD ............................................................................... 34

The TCP Reliability research study .......................................................................... 34

Field work and TCP blowcount data sources ........................................................... 34

Hammer energy readings .......................................................................................... 35

The TCP blowcount dataset ...................................................................................... 36

RESULTS FROM TCP BLOWCOUNT DATA ................................................................... 37

Hammer Efficiency for the TCP test (Er-TCP) ............................................................ 38

Development of Rod Length Correction Factors (CR-TCP) ........................................ 39


v

Side by side comparison of TCP and SPT Rod Length correction factors ............... 43

Development of Overburden pressure Correction Factors (CN-TCP) ......................... 44

Overburden correction factor CN-TCP ........................................................................ 49

OTHER FACTORS THAT MAY INFLUENCE TCP TEST HAMMER EFFICIENCY ............... 51

SUMMARY AND CONCLUSIONS ..................................................................................... 52

CHAPTER III: EVALUATION OF PREDICTED VALIDITY OF THE TEXAS CONE PENETRATION

DESIGN CHARTS FOR DEEP FOUNDATIONS BASED ON FULL SCALE LOAD TESTS ......... 54

ABSTRACT ..................................................................................................................... 54

INTRODUCTION ............................................................................................................. 55

THE TXDOT TEXAS CONE PENETRATION TEST ........................................................ 57

Description of the TCP test ....................................................................................... 57

TCP Design Charts ................................................................................................... 58

TCP Design Method for Deep Foundations.............................................................. 62

RESEARCH DESIGN AND METHOD ............................................................................... 63

Development of the Dataset .......................................................................................... 63

Allowable Predicted Total Capacity ............................................................................. 65

Allowable Measured Total Capacities .......................................................................... 67

DATA ANALYSIS............................................................................................................ 72

Qualitative Evaluation .............................................................................................. 75

Statistical Analysis .................................................................................................... 80

RESULTS ....................................................................................................................... 81


vi

Regression Models .................................................................................................... 82

Side-by-Side Comparison of all Measured Capacity Models ................................... 84

Evaluation of TCP charts based on Shaft and Base resistance ................................. 88

SUMMARY AND CONCLUSIONS ..................................................................................... 89

CHAPTER IV: RESISTANCE FACTORS AT SERVICEABILITY LIMIT STATE FOR LRFD OF

DEEP FOUNDATIONS USING THE TEXAS CONE PENETRATION TEST .............................. 93

ABSTRACT ..................................................................................................................... 93

INTRODUCTION ............................................................................................................. 94

TEXAS CONE PENETRATION (TCP) TEST .................................................................... 96

Description of TCP test ............................................................................................. 96

Application and use of TCP blowcount data for foundation design .......................... 97

LOAD AND RESISTANCE FACTOR DESIGN (LRFD) ..................................................... 99

Design approaches and methods................................................................................ 99

Ultimate Limit State (ULS) ....................................................................................... 99

Serviceability Limit State (SLS) .............................................................................. 100

LRFD of Deep Foundations .................................................................................... 101

SERVICEABILITY LIMIT STATE ANALYSIS ................................................................. 103

SLS Performance function ....................................................................................... 103

Tolerable Displacement ........................................................................................... 105

DEVELOPMENT OF SLS RESISTANCE FACTORS ........................................................ 108

Displacement Approach .......................................................................................... 108

Load Approach ........................................................................................................ 110

Calibration of Resistance Factors ............................................................................ 111


vii

RESEARCH DESIGN AND METHOD ............................................................................. 112

Dataset Development ............................................................................................... 112

Load Tests and TCP Borings ................................................................................... 113

Final Compiled Dataset ........................................................................................... 114

Load Corresponding to Tolerable Displacement and Design Load ........................ 114

Tolerable Displacement for the TCP design method ............................................... 117

RESULTS OF ANALYSES .............................................................................................. 119

Bias .......................................................................................................................... 120

Resistance Factors at SLS condition ....................................................................... 123

CONCLUSIONS ............................................................................................................. 128

Summary of Findings .............................................................................................. 128

Limitations/Further Study ............................................................................................... 129

CHAPTER V: SUMMARY AND CONCLUSIONS ................................................................. 131

REFERENCES ................................................................................................................... 140


viii

ABSTRACT

This dissertation explores three different aspects of the Texas Department of

Transportation’s Texas Cone Penetration (TCP) test and its associated foundation

design charts. The first aspect has to do with the reliability of TCP data. The TCP test

hammer efficiency, rod length influence on the hammer efficiency, and effect of

overburden pressure on the TCP test blowcounts (NTCP) are explored. Results are

compared with published correction factors for the Standard Penetration Test (SPT).

The final dataset analyzed for the study, consisted of 293 TCP tests from which 135

tests were instrumented. Analyses showed a statistically significant relationship

between the TCP hammer efficiency and the rod length below ground surface. Rod

length correction factors (CR-TCP) and overburden correction factors (CN-TCP) were

obtained factors for the TCP test.

The second topic involves a quantitative and a qualitative evaluation of the predictive

validity of the Texas Cone Penetration (TCP) foundation design charts where allowable

measured total capacities determined from results of full-scale load tests were compared

to allowable predicted total capacities determined from TCP foundation design charts.

The final dataset analyzed compiled for this study consisted of 60 full scale load testes

comprising 33 driven piles and 27 drilled shafts. Allowable measured capacities were

determined using strength-based and serviceability-based models. Results from

analyses suggested that it is apparent that different performance assessment criteria

leads to different conclusions regarding the predictive validity of the TxDOT TCP


ix

method for deep foundation design. The allowable predicted total capacity determined

from the TCP foundation design charts could be considered to be reasonable based on

a strength-based model, moderately over-conservative, to over-conservative according

to serviceability-based models. Further, it was observed that as the level of tolerable

settlement is increased, the interpretation of the TCP foundation design charts becomes

even more conservative.

The third topic identifies resistance factors for the serviceability limit state (SLS)

condition used in the load and resistance factor design (LRFD) of deep foundations

using the results from the TCP test. The performance function was established based on

load corresponding to tolerable displacement and design load. The compiled dataset

consisted of a total of 60 full scale load test cases comprising 33 driven piles and 27

drilled shafts. The loads corresponding to tolerable displacements were determined

using the load-settlement curves, and these loads were compared with the design load

determined using the TCP test method. Resistance factors for SLS conditions were

obtained for tolerable displacements using both the Monte Carlo simulation (MCS) and

the first order second moment (FOSM) calibration approaches.


x

LIST OF TABLES

2.1. Correlations between NSPT and NTCP………………………………………………………………........

21

2.2. Summary of existing Overburden correction factors (CN)……………………………….......

30

2.3. Field Research Sites and related TCP-Borings ………………………………………………..…..

35

2.4. Rod Length Correction Factors (CR) for TCP and SPT ……………………….........................

44

3.1. Compiled Dataset for Driven Piles in Soils …………………………………………………….......

74

3.2. Compiled Dataset for Drilled Shafts in Soils …………………………………………………….…

75

3.3. Regression line equation with p-value and R2, Driven Piles …………….............................

85

3.4. Regression line equation with p-value and R2, Drilled Shafts ………….............................

85

4.1. Compiled Dataset for Driven Piles ……………………………………………………………………...

116

4.2. Compiled Dataset for Drilled Shafts …………………………………………………………………….

117

4.3. Summary Statistics of Biases for Driven Piles and Drilled Shafts ………………………

122

4.4. Resistance Factors at Serviceability Limit State (SLS) for Foundations in Soils..

126


xi

LIST OF FIGURES

FIGURE CAPTIONS- CHAPTER II

2.1. TCP Test Conical Driving Point and Field Application………………………………… 14

2.2 (a) Allowable Shaft Resistance and (b) Base Resistance vs. TCP blows/12 inches

(TxDOT, 2012)……………………………………………………………………………………….

16

2.3. (a) Allowable Shaft Resistance (b) Allowable Base Resistance vs. TCP

Penetration/100 blows (TxDOT, 2012) )………………………………………………………..

16

2.4. Different Type of Hammers (a) Donut Hammer (b) Safety Hammer (c) Automatic

Hammer (Coduto et al. 2011 used with permission)……………………..................................

25

2.5. Standard Practice of NSPT Correction among State DOTs…………………………….... 31

2.6. Compiled dataset (a) TCP Blowcount Values NTCP (b) Average Hammer

Efficiency……………………………………………………………………………………………....

37

2.7. Scatterplot of Average Efficiency vs Rod length for all Soils………….......................... 42

2.8. N60-TCP and Effective Vertical Stress relationship for each TCP boring……………… 46

2.9. Scatterplot for Normalized TCP Blowcounts (NN-TCP) vs. Normalized Effective

Vertical Stress (’N)………………………………………………………………………………….

48

2.10. Effective Vertical Stress (’v) versus CN for TCP and SPT…………………………… 50

FIGURE CAPTIONS- CHAPTER III

3.1. TCP Test Conical Driving Point (TxDOT, 1999)………………....................................... 59

3.2. Design charts representing (a) allowable unit shaft resistance and (b) allowable

unit base resistance vs. TCP blows/30-cm (12-in), (TxDOT, 2012)………………………..

61

3.3. Illustrative example of determining 𝑃𝑎𝑃 for Case No. 23 of drilled

shafts………………………………………………………………………………………….................

66


xii

3.4. Case No. 23, Davisson’s ultimate Capacity Criterion and Serviceability

Criteria………………………………………………………………………………………………...

72

3.5. The relationship between 𝑃𝑎𝑀 and 𝑃𝑎

𝑃 compared to an equal prediction

line………………………………………………………………………………………………………

78

3.6. Scatterplot and fitted model, Driven Piles………………………………………………… 82

3.7. Scatterplot and fitted model, Drilled Shafts………………………..................................... 84

3.8. Side-by-Side Comparison of all Models analyzed, Driven Piles…................................ 86

3.9. Side-by-Side Comparison of all Models analyzed, Drilled Shafts………………….… 88

FIGURE CAPTIONS- CHAPTER IV

4.1. TCP test conical driving point (From TxDOT, 1999)………………................................ 97

4.2. TCP Foundation Design charts representing (a) allowable unit shaft resistance

and (b) allowable unit base resistance vs. TCP blows/30-cm (12-in), (TxDOT,

2012)……………………………………………………………………................................................

98

4.3. TCP Foundation Design charts representing (a) allowable unit shaft resistance

and (b) allowable unit base resistance vs. TCP penetration/100 blows (TxDOT, 2012) ……………………………………………………………………………...............................................

98

4.4. Loads corresponding to tolerable displacement (Qtol) and displacement from

design loads (Qd)……………………………………………………………………………………..

104

4.5. Tolerable displacement histograms for (a) driven piles and (b) drilled

shafts…………………………………………………………………………………...........................

119

4.6. Histograms and probability plots for driven piles (a and b, respectively) and for

drilled shafts (c and d, respectively)………………………….......................................................

122


xiii

NOTATIONS

The following symbols are used in this Dissertation

a and b Skempton (1986) soil dependent parameters

1 Slope for a linear model equation

o Intercept for a linear model equation

Cb Borehole diameter correction factor for SPT

CI Confidence Intervals

CN Overburden correction factor for SPT

CN-TCP Overburden correction factor for TCP

COV Coefficient of Variation

Cr Rod length correction factor for SPT

CR-TCP Rod length correction factor for TCP

Cs Sampler type correction factor for SPT

D Depth

DBASELINE Depth at which the fitted model has flattened

DR Relative Density

Em Measured hammer energy

Er Hammer Efficiency-SPT

Er-TCP Hammer Efficiency-TCP

Et Theoretical hammer energy

N1-60 SPT Blowcount standardized to 60% energy and corrected for overburden

N1-60-TCP TCP Blowcount standardized to 60% energy and corrected for overburden

N60 SPT Blowcount standardized to 60% energy

N60-TCP TCP Blowcount standardized to 60% energy

NN-TCP

Normalized TCP blowcount to a blowcount corresponding to a reference

stress

NSPT SPT Blowcount

NTCP TCP Blowcount

PI Prediction Intervals

'N Normalized effective vertical stress to a reference stress

SPT Standard Penetration Test

ref Reference Stress (i.e. 100 kPa, 2000 psf)

'v Effective vertical stress

TCP Texas Cone Penetrometer


xiv

g Performance Function

gSLS Serviceability Limit State Performance Function

L Span Length

Nsim Number of Simulations

Pf Probability of Failure

Q Load

Qd Design Load

Qd,TCP Predicted Capacity based on TCP design charts

QDL Dead Load

Qtol Load corresponding to the tolerable displacement

Qi Applied Load

QLL Live Load

Qt Total capacity

R Resistance

RM Measured Resistance

RP Predicted Resistance

SLS Resistance Factor at Serviceability Limit State

d Displacement under Design Load

Angular Distortion

Differential Settlement

t Resistance Factor for total resistance

d Resistance factor for design loads "Greater than 1.0"

trial Initial Resistance Factor for the Monte Carlo Simulation

tol Tolerable displacement

Bias

R-SLS Bias for Resistance at Serviceability Limit State

R Bias for Resistance

LL Bias for Live Load

DL Bias for Dead Load

Load Factor

Reliability Index


1

CHAPTER I

INTRODUCTION

This dissertation presents results of analyses completed for the evaluation of the Texas

Department of Transportation’s (TxDOT) Texas Cone Penetration (TCP) field test and its

associated foundation design method. The TCP field test is studied and correction factors

for the penetration index (NTCP) are developed. Furthermore, the predictive validity of TCP

foundation design charts have been evaluated by comparing the allowable measured

capacity (interpreted from full-scale load tests completed on deep foundations to allowable

predicted capacity (calculated from current published TCP foundation design charts).

Finally, considering the transition from allowable stress design (ASD) method to the Load

and Resistance Factor Design (LRFD) methodology, and the importance of the

serviceability of a structure, resistance factors at serviceability limit state for the TCP

foundation design method have been developed. All three contexts evaluated in this

dissertation share the common core of the TCP test and its foundation design charts with

different focal point. However, the collective outcome resulted from the detail exploration

of each topic will help to improve the TCP-based design of deep foundations.

As would be expected with its Texas roots, the TCP test has seen extensive application

throughout Texas to evaluate subsurface materials ranging from very soft clays, to shales

and fractured limestones, to hard rock. However, the use of TCP test and its associated

design charts are not limited to Texas practice only. Because of its applicability to a wide

range of geomaterials, the TCP test has been used not only throughout the United States

but also for international projects. According to FHWA (2016) Texas currently maintains


2

over 53,000 bridges in the National Bridge Inventory, and most of these bridges plus other

transportation infrastructure in Texas are supported by foundations designed in accordance

with the TCP method. Also, since Oklahoma DOT adopted the TCP test in the 1970s, the

TCP method has been used for the design of the foundations for most of the 23,000 bridges

in Oklahoma’s inventory. Collectively this indicates that over 12% of the nation’s bridges

are associated with the TCP test. Recently, a comprehensive research program in Missouri

evaluated a modified version of the TCP test for bridge foundation design applications

using LRFD concepts (Loehr et al. 2011). The TCP test was introduced to an international

audience during a workshop sponsored by the Port and Airport Research Institute of Japan,

and the application of the method for the design of deep foundations was discussed as part

of Advance in Deep Foundations in Japan, (Vipulanandan, C. 2007). Furthermore, the TCP

test has been used in Korea for field explorations associated with a bridge abutment

overlying intermediate geomaterials and soft rock where cores would have been recovered

in fragments because of joint structures (Nam et al. 2013). In light of the preceding, the

scope of the TCP test and its associated foundation design charts is significant well beyond

its regional origins. Therefore, the presence of the TCP test and its associated foundation

design method in the deep foundations expert community is noteworthy.

Technological advancements have certainly impacted most engineering disciplines

including the geotechnical engineering field. Tools and methods used in the geotechnical

field exploration have been modified and improved throughout the United States. One of

the notorious changes is the introduction and implementation of automatic hammers which

has substituted traditional hammers used in the past during geotechnical drilling and


3

sampling operations. Aside from drilling and sampling, many times, if not at all times, a

penetration index is recorded while performing a dynamic field testing. Throughout many

years, these penetration indices have been correlated to soil strength parameters, and

further, these correlations have been used for the design of many geotechnical elements

including deep foundations.

The TCP test and its associated foundation design charts, are the great example related to

the above stated situation in geotechnical engineering. The TCP method dates back to

1940s when the use of hammers other than an automatic hammer was common. Penetration

indices were correlated to soil strength for the development of TCP foundation design

charts. In current practice, automatic hammers are used to determine penetration indices

but TCP design charts were developed based on a non-automatic hammers and are used for

the design of deep foundations. Therefore, the penetration indices recorded based on

automatic hammers are not the same as those recorded using a none-automatic hammer

which directly impacts the TCP-based design of deep foundation.

Aside from type of hammer, other factors such as type of soil in which the test is being

performed, the diameter of the borehole, existence of fluid in the borehole, among others

may influence the penetration indices and further impact the design parameters determined

from TCP foundation design charts.

Considering previous published work and research studies completed for other

geotechnical dynamic field tests to address similar issues as stated in the preceding


4

paragraphs, it is reasonable to consider the development of correction factors for the TCP

test to account for type of hammer and other influential factors. The first topic presented

in this dissertation explores factors influencing the TCP test blowcounts (NTCP) including

hammer efficiency, rod length, and overburden pressure. Correction factors for the NTCP,

will yield a more refined penetration index that can be used for the determination of

strength parameters from the TCP foundation design charts to further calculate the axial

capacity for a deep foundation element. For this part of the dissertation it was hypothesized

that: (1) There is a relationship between TCP Hammer Efficiency and the length of drilling

rod below ground surface, and (2) There is a relationship between NTCP and the overburden

pressure.

After developing correction factors for the NTCP, another variable that directly influences

the design of deep foundation using the TCP design method, is the predictive validity of

the TCP model itself. Therefore, it is reasonable to consider a detailed analysis and

evaluation of the existing TCP foundation design charts. Another topic thoroughly studied

in this dissertation was related to the evaluation of the predictive validity of the TCP

foundation design charts by comparing allowable measured total capacity to the allowable

predicted total capacity. In engineering design, the allowable load on a structural

component may be determined based on strength-based criteria or serviceability-based (i.e.

settlement-based) criteria. Within the context of deep foundations, the strength-based

allowable load capacity can be defined as ultimate capacity divided by an appropriate factor

of safety. The serviceability-based approach consisted of loads corresponding to vertical

displacements determined from load-settlement curves. In geotechnical engineering,


5

deformations presented at the head of the deep foundation are translated to vertical

settlements noted in the superstructure. When settlements are larger than an established

tolerable settlement, then it is considered that the foundation element has reached a

serviceability limit condition, and the performance of the superstructure is not satisfactory.

For this part of the dissertation it was hypothesized that the predictive validity of the TCP

foundation design charts is within higher confidence that the allowable measured capacity

is equal or greater than the allowable predicted capacity for serviceability-based model

compared to a strength-based model.

Once the TCP field test and foundation design method were evaluated, the application of

this method can be analyzed. In geotechnical engineering and more specifically in the case

of deep foundation design, deformations can be translated to vertical displacement of the

foundation element which causes settlement in the superstructure. Most structures and in

particular, transportation structures such as bridges, suspend service operations as soon as

one of the structure’s members experiences displacements beyond a tolerable

displacement. This could well explain that geotechnical designs are mainly displacement-

based designs rather than strength-based designs. Although, most of geotechnical design

are performed following a strength-based model, the performance of the foundation or the

superstructure is settlement dependent.

Considering that (1) geotechnical engineers working on projects funded by the Federal

Highway Administration (FHWA) have been transitioning from the ASD approach to the

LRFD methodology, (2) the TCP-based designs are mainly used for transportation


6

infrastructures funded by FHWA, and (3) the geotechnical designs are mainly settlement

or serviceability governed designs, then it is an equitable interest to analyze the TCP test

based on the LRFD requirements focused on the serviceability instead of strength. In this

topic, resistance factors at serviceability limit state were developed for the LRFD of deep

foundations using the TCP foundation design charts. For this context it was hypothesized

that as the foundation element is allowed to experience higher settlements, higher

resistance factors at serviceability limit state is determined.

TCP FIELD TEST

The TCP test is a dynamic field penetration test which assesses the consistency of the

material encountered during geotechnical exploration. This test method is documented as

TxDOT Designation Tex-132-E, “Test Procedure for Texas Cone Penetration” (TxDOT

1999). The TCP test uses a 77.0-kg (170-lb) hammer with 60-cm (24-in) drop to force a

7.6-cm (3-in) diameter steel cone into the soil or rock formation. In current practice, the

penetration is to be achieved in three separate increments. The first increment is completed

to ensure proper seating, which consists of driving the cone 12 blows or approximately 15-

cm (6-in), whichever happens first. The TCP blowcount is then determined as the sum of

the number of blows required to achieve second and third 15-cm (6-in) increments of cone

penetration. The total blowcount or NTCP corresponding to 30-cm (12-in) penetration is

used to obtain design parameters. In very hard materials such as rock and intermediate

geomaterials (IGM), after the proper seating process is completed, the cone is driven 100

blows and the penetration value for the first and second 50 blows are recorded.


7

TCP FOUNDATION DESIGN CHARTS

The 1956 edition of the Foundation Exploration and Design Manual provides a series of

correlation curves illustrating unit shear strength estimation based on NTCP. These

correlation curves were based on relationships established between TCP test data and

laboratory-measured shear strength. To obtain soil strength parameters for this purpose,

undisturbed samples were collected and tested using the triaxial test procedure (THD

1956). Based on data obtained from field and laboratory studies, correlation charts were

developed and published in two separate sets of design charts. The first set was developed

for soils with TCP blowcounts less than 100 per 30-cm (12-in); whereas, the second set is

for geomaterials with blowcounts greater than 100 blows per 30-cm (12-in) (i.e. penetration

per 100 blows). The current design charts reflect the allowable stress design philosophy

and present allowable strength values corresponding to a Factor of Safety (FS) of 2.0.

Different literature may refer to the foundation shaft resistance and base resistance using

other terminologies such as skin resistance and toe bearing or skin friction and point

bearing. The TxDOT’s Geotechnical Manual refers to these resistances as allowable skin

friction and allowable point bearing. Furthermore, design charts are categorized based on

soil classification where capacity models for fat clay (CH), Lean Clay (CL), Clayey Sand

(SC), and OTHER soils have been developed. According to the design procedure TxDOT

(2012) any soil calssified as SP, SW, SM, and ML is considered as OTHER category.


8

ORGANIZATION OF THE DISSERTATION

The work completed for a critical evaluation of the TCP field test and its associated

foundation design charts including the development of TCP test blowcount correction

factors, assessment of current TCP foundation design charts based on full-scale load test

results, and the development of resistance factors at the serviceability limit state (SLS) used

in the Load and Resistance Factor Design (LRFD) of Deep Foundations are documented

in this dissertation and presented as follows:

Chapter 1 presents a brief description of the topics explored and developed throughout the

dissertation. The TCP test and its regional, national, and international presence is briefly

discussed followed by a summary of the rationale behind the need for corrections to the

TCP penetration index. Furthermore, a synthesized description of the evaluation of the TCP

foundation design charts based on a comparisons between allowable measured capacities

determined from results of full-scale load tests and allowable predicted capacities

determined using the TCP foundation design charts is presented. Finally, the importance

of settlement governed design of deep foundations and the development of resistance

factors at serviceability limit state for the LRFD of deep foundations using the TCP method

is explained.

Hammer efficiency data, rod length influence on the hammer efficiency, and overburden

pressure correction factors for the Texas Cone Penetration (TCP) test blowcounts (NTCP)

are explored in Chapter 2. Results of analyses are compared to published correction factors

developed for the Standard Penetration Test (SPT). The effort associated with data


9

collection and statistical analyses are described in detail. At the end of the chapter, results

for hammer efficiency, rod length correction factors, and overburden correction factors are

presented.

In chapter 3, presents a detailed evaluation of the TxDOT’s TCP foundation design charts

by comparing the allowable predicted total capacity determined using the TCP foundation

design charts to the allowable measured total capacity obtained from full scale load tests

completed for driven piles and drilled shafts in soils. Two different approaches were used

to determine the allowable measured total capacity from results of full-scale load test: (1)

strength-based method, and (2) serviceability-based approach. After determining the

allowable measured total capacity and the allowable predicted total capacity, these data

were analyzed based on (1) a qualitative evaluation where all data were plotted and

compared to an equal prediction line, and (2) a quantitative evaluation where statistical

analysis and linear regression models were developed to analyze the relationship between

allowable measured total capacity and the allowable predicted total capacity.

Chapter 4 presents the calibration process and the development of resistance factors for the

serviceability limit state (SLS) condition used in the load and resistance factor design

(LRFD) of deep foundations using the results from Texas Cone Penetration (TCP).

Determination of statistical parameters such as mean and coefficient of variation for the

bias is described and resistance factors for SLS conditions were obtained using both the

Monte Carlo simulation (MCS) and the first order second moment (FOSM) calibration

approaches.


10

Chapter 5 presents a summary of the analyses and results obtained for all topics explored

in this dissertation, discusses notable findings, and highlights key contributions. Research

studies and future work recommended for the enhancement of results presented in this

dissertation are presented in this chapter. Furthermore, limitations of compiled datasets for

each topic discussed and analyzed in this dissertation are noted. Finally, based on the

findings of this dissertation, a detailed discussion of the TxDOT TCP test, its foundation

design charts, and the applicability of this method is presented where important conclusions

are highlighted.


Note 1: Moghaddam R.B., Lawson, W.D., Surles, J.G., Seo H., Jayawickrama, P.W. (2016), “Hammer

Efficiency and Correction Factors for the TxDOT Texas Cone Penetration Test”, Journal of Geotechnical

and Geoenvironmental Engineering, Submitted June, 7th, 2016

11

CHAPTER II

HAMMER EFFICIENCY AND CORRECTION FACTORS FOR THE TXDOT TEXAS

CONE PENETRATION TEST 1

ABSTRACT

This study analyzes blowcount fata from instrumented Texas Cone Penetration (TCP) tests.

TCP hammer efficiency, rod length influence on the hammer efficiency, and overburden

pressure correction factors for the TCP blowcounts (NTCP) are explored. Results are

compared to published correction factors for the Standard Penetration Test (SPT). The final

dataset analyzed for this study consisted of 293 TCP tests from which 135 tests were

instrumented. TCP hammer efficiency values for automatic CME hammers ranged from

74% to 101% with an average of 89%. Analyses showed a statistically-significant

relationship between the TCP hammer efficiency and the rod length below ground surface.

Statistical models were developed for undifferentiated soils and rod length correction factor

for the TCP test (CR-TCP) were obtained ranging from 0.90 to 1.00. In a second analysis the

relationship between the overburden pressure and the NTCP was explored and a

mathematical expression for the overburden correction factors for the TCP blowcount

value (CN-TCP) was determined. This work is the first study that has been completed where

corrections to NTCP are explored and the outcome benefits the geotechnical engineering

community using the TCP test and foundation design method.


12

INTRODUCTION

This paper presents correction factors for blowcount values (NTCP) from the Texas Cone

Penetration (TCP) test, a dynamic field penetration test which is similar to yet different

from the Standard Penetration Test (SPT). For over 60 years, the TCP test and its associated

foundation design charts have been used successfully for the design of drilled shaft and

driven pile foundations that support tens of thousands of bridges and other major

transportation infrastructure throughout Texas and parts of Oklahoma. Data for this study

were obtained as part of a program of 293 TCP tests obtained from 21 geotechnical borings

in five states. These data were used to identify the average hammer efficiency for the TCP

test using an automatic hammer, and also to develop rod length correction factors and

overburden pressure correction factors for the NTCP values.

The main purpose of correction factors for field penetration blowcount values is to achieve

consistent and reliable input data for the foundation design procedures associated with the

tests. In the case of SPT, several studies have been completed to address the impact of

different factors on the hammer efficiency which further influences NSPT. In contrast to

SPT, no published work discusses a corrected NTCP nor addresses the influence of different

factors on TCP hammer efficiency. This paper contributes to the geotechnical engineering

community where the TCP design charts are used as the primary method for the design of

deep foundations. The factors presented in this paper help to standardize NTCP and further

obtain accurate design parameters for foundation design based on the TCP design charts.


13

The TCP test and its associated foundation design method are introduced and compared to

the SPT to further highlight similarities and differences. To set the stage for the analyses

presented in this paper, a review of the history of the TCP test and the development of

correction factors for the SPT are described. Tasks associated with the field data collection

for this study followed by statistical analyses and results are discussed in detail. Finally,

hammer efficiency for the TCP test and rod length correction factors for NTCP are

developed and compared to existing factors for the SPT. Furthermore, the effect of the

overburden pressure on the variation of NTCP is analyzed and discussed.

THE TXDOT TEXAS CONE PENETRATION TEST

Description of the TCP test




1999). The TCP test uses a 77.0-kg (170-lb) hammer with 0.61-m (24-in) drop to force a

7.6-cm (3-in) diameter steel cone into the soil or rock formation, Figure 2.1.


14

(a) TCP Conical Driving Point

(TxDOT 1999)

(b) Field Application of the TCP Test

(TxDOT 1999)

Figure 2.1. TCP Test Conical Driving Point and Field Application

In current practice, penetration for the TCP test is performed in three separate increments.

The first is to achieve proper seating, consisting of driving the cone 12 blows or

approximately 15-cm (6-in), whichever comes first. The TCP blowcount is then

determined as the sum of the number of blows required to achieve the second and third

15-cm (6-in) increments. The total blowcount or NTCP corresponding to 30-cm (12-in)

penetration is used to obtain design parameters. In very hard materials such as rock and

intermediate geomaterials (IGM), after the proper seating process, the cone is driven 100

blows and the penetration value for the first and second 50 blows are recorded, with the

NTCP value reported as centimeters (inches) of penetration per 100 blows.

History and development of the TCP test

In the 1940s, the newly-formed Bridge Foundation Soils group of the Texas Highway

Department (THD) Bridge Division identified the need for developing a unified method


15

for the characterization of geomaterials and the design of deep foundations. The TCP test

method was developed by the Bridge Group as an in-situ test method for evaluating the

broad range of geologic materials encountered in foundation construction (TxDOT 2000).

The TCP-based foundation design method was introduced in the 1956 edition of the

Foundation Exploration and Design Manual (THD, 1956). This manual provided a series

of correlation relationships for foundation design which were established based on TCP

test blowcount data (NTCP) and laboratory-measured shear strength using the triaxial test

procedure (THD, 1956). The charts published in 1956 were refined in 1972, 2000, and

2012, and two sets of design charts now exists. The first set was developed for the

prediction of unit shaft resistance (i.e. skin friction) and unit base resistance (i.e. point

bearing) for soils with TCP blowcounts less than 100 blows/30-cm (blows/foot), Figure

2.2. The second set is for geomaterials with blowcounts greater than 100 blows/30-cm

(blows/foot) (i.e. penetration per 100 blows), Figure 2.3. The current design charts reflect

the allowable stress design philosophy and present allowable strength values.


16

(a) (b)

Figure 2.2 (a) Allowable Shaft Resistance and (b) Base Resistance vs. TCP blows/12

inches (TxDOT, 2012)

(a) (b)

Figure 2.3 (a) Allowable Shaft Resistance (b) Allowable Base Resistance vs. TCP

Penetration/100 blows (TxDOT, 2012)


17

The TCP test and its associated foundation design method have been used regionally,

nationally and internationally. As would be expected with its Texas roots, the TCP test

has seen extensive application throughout the state to evaluate subsurface materials

ranging from very soft coastal clays, to shales and weathered/fractured limestones, to hard

and brittle rock. Texas currently maintains over 53,000 bridges in the National Bridge

Inventory (FHWA 2016), and most of these bridges plus other transportation

infrastructure in Texas are supported by foundations designed in accordance with the TCP

method. The Oklahoma DOT adopted the TCP test and foundation design approach in the

1970s, so the foundations for most of the 23,000 bridges in Oklahoma’s inventory were

also designed using the TCP method. Collectively this means over 12% of the nation’s

bridges are associated with the TCP test. The TCP test is discussed in NCHRP Synthesis

360 (Turner 2016), and owing to competitive opportunities for Texas’ major

transportation infrastructure projects, geotechnical practitioners and academics

nationwide have had occasion to learn about and use the TCP test. More recently, a

comprehensive research program in Missouri evaluated a modified version of the TCP

test for bridge foundation design applications using LRFD concepts (Loehr et al. 2011).

The TCP test was introduced to an international audience of deep foundation engineers in

Japan (Vipulanandan, C. 2007), and the TCP test has been used in Korea for field

explorations associated with a bridge abutment overlying intermediate geomaterials and

soft rock where cores would have been recovered in fragments because of joint structures

(Nam et al. 2013). Collectively then, the scope of the TCP test is significant well beyond

its regional origins.


18

COMPARISON OF SPT VS. TCP TESTS

The SPT is an international standard for measuring soil penetration resistance and

obtaining a representative disturbed soil sample for identification purposes (ASTM 2016)

and has adopted a 63-kg (140-lb) hammer mass with a falling height of 76-cm (30-in) to

force a 45-cm (18-in) sampler (i.e. split spoon) with a coupler and driving shoe to penetrate

into soil strata. The conventional SPT driving procedure where blows are recorded for each

of three 15-cm (6-in) increments was introduced in 1954, and the SPT was adopted in 1958

as ASTM Standard D 1586, “Standard Test Method for Standard Penetration Test (SPT)

and Split Barrel Sampling of Soils”.

The form of the TCP test is similar to the SPT in that a steel driving point is advanced into

subsurface material at the bottom of a borehole by hammer strikes, with blowcounts

recorded in three 15-cm (6-in) increments. However, in certain aspects the TCP test differs

from the SPT. The TCP test does not use a split-barrel sampler but rather a solid steel

conical point, Figure 2.1. Hence, the TCP test cannot and does not collect a soil sample.

Also, because of its more robust solid steel design, TCP test refusal is defined as resistance

to penetration greater than 100 blows/30-cm (12-in), so the TCP test is suitable for

evaluating harder geomaterials and rock. In contrast, SPT refusal is customarily achieved

at resistance to penetration greater than 50blows/15-cm (6-in), or when there is no observed

advance of the sampler during the application of 10 successive blows.

Another difference between the SPT and the TCP test is the magnitude of their blowcount

values. The correlation between NTCP and NSPT is not linear, and NSPT values are typically


19

20% to 60% lower than NTCP values, other things being equal (Touma et al. 1972). It is

possible to gain an intuitive sense about this by comparing the hammer energy and the

sampler area for the two tests. The theoretical hammer energy for the tests is roughly

equivalent – 475 N-m (4,200 in-lbf) for the SPT vs. 461 N-m (4,080 in-lbf) for the TCP.

But the cross-sectional area of the samplers is quite different. The nominal cross-sectional

area for an unplugged SPT split spoon is 1,071 mm2 (1.66 in2) compared to the TCP conical

point with cross sectional area of 4,561 mm2 (7.07 in2). Thus, it ought to take a lot less

energy to drive an SPT split barrel sampler than the TCP cone, so SPT blowcounts should

be significantly lower than TCP blowcounts in the same material. This is in fact the case.

Since the 1970s, research has shown that blowcounts obtained from the SPT (NSPT) are

sensitive to factors such as hammer energy, rod length, type of soil, borehole diameter, and

others. In the case of SPT a significant body of literature and research exists to facilitate

direction and guidance for correcting NSPT values and further use a standardized NSPT for

design. In contrast, for the TCP test, published work has not been completed exploring the

effects of these same factors, and perhaps other factors influencing the resistance to

penetration.

It is important to mention that throughout this paper, detailed discussion and mathematical

expressions regarding the SPT and the TCP test are presented. For purposes of clarity and

consistency all factors, equations, and parameters associated with the TCP test will have

the subscript “TCP”. Any other geotechnical field testing parameter presented without the

subscript will be associated with the SPT.


20

Correlation of blowcount results

Another difference between the SPT and the TCP test is the magnitude of their blowcount

values, and it can be stated upfront that NSPT values are typically 20% to 60% lower than

NTCP values, other things being equal. It is possible to gain an intuitive sense about this by

comparing the hammer energy and the sampler area for the two tests. The theoretical

hammer energy for the tests is roughly equivalent – 475 N-m (4,200 in-lbf) for the SPT vs.

461 N-m (4,080 in-lbf) for the TCP. But the cross-sectional area of the samplers is quite

different. The nominal cross-sectional area for an unplugged SPT split spoon is 1,071 mm2

(1.66 in2) compared to the TCP conical point with cross sectional area of 4,561 mm2 (7.07

in2). Thus, it ought to take a lot less energy to drive an SPT split barrel sampler than the

TCP cone, so SPT blowcounts should be significantly lower than TCP blowcounts in the

same material. This is in fact the case.

Burmister (1948) presented an energy-area equation to correct NSPT when compared to

different but similar in-situ penetration tests. Lacroix and Horn (1973) presented a theory

where the resistance to penetration from a non-standard test is compared to a standardized

penetration test such as SPT. These resistances to penetration are correlated based on

parameters creating the driving energy.

TxDOT published correlations based on the research done by Touma and Reese (1972)

which was the first side-by-side correlation between SPT and TCP blowcount values. In

contrast to energy-area equations, the correlations presented by Touma and Reese (1972)


21

identify a distinction between coarse and fine grained soils. Building on this work, Lawson

et al. (2016) developed new correlation factors between NSPT and NTCP based on a larger

side-by-side dataset. Table 2.1 summarizes published correlations between NSPT and NTCP.

Table 2.1. Correlations between NSPT and NTCP

DEVELOPMENT OF CORRECTION FACTORS FOR SPT

Standardization of the SPT blowcount

To reduce the variability of NSPT due to multiple factors associated with the test, several

researchers have recommended the standardization of NSPT to a specific level where the

influence of these factors has been addressed through correction factors. Numerous

technical reports and research studies have been published addressing the factors that may

influence the NSPT including hammer type, sampler, driving mechanism, depth of test,

borehole diameter, and physical aspects of tools used during the test. See Palmer and Stuart

(1957), Fletcher (1965), Ireland (1970), DeMello (1971), Brown (1977), Schmertmann

Reference Type of

Correlation

Correlation Application

Burmister

(1948)

Energy Area NSPT = 0.23NTCP Open Split Spoon

All Soils

Lacroix and

Horn (1973)

Energy Area NSPT = 0.43NTCP Closed Split Spoon

All Soils

Touma and

Reese (1972)

Side-by-Side NSPT = 0.7NTCP

NSPT = 0.5NTCP

Fine Grained Soils

Coarse Grained Soils

Lawson et al.

(2016)

Side-by-Side logN60SPT = 0.1830 + 0.7463 log N60TCP

logN60SPT = 0.7436 + 0.303 log N60TCP

Fine Grained Soils



22

(1979), Uto (1981), Kovacs and Salomone (1982), Seed (1984), Skempton (1986), Daniel

et al. (2005), Odebrecht et al. (2005), and Schnaid (2009). The work completed by these

researchers mainly focused on the determination of a standard NSPT where all the

variabilities associated with the test procedure are accounted for. In later years, this topic

continued to evolve with slightly different focus. Work done by Aggour (2001) and Idriss

and Boulanger (2004, 2012, and 2015) focused on the calibration of NSPT for purposes of

liquefaction and corrections for overburden pressure after the work presented by Skempton

(1986).

Early work by Fletcher and others

Between 1920s and 1930s the SPT was considered as a standard test by introducing the 63-

kg (140-lb) hammer, 76-cm (30-in) drop height, and 5-cm (2-in) outer diameter split

sampler. The influence of human factors and equipment associated with the execution of

the SPT were outlined and examined by Fletcher (1965). The application of the SPT in

both granular and cohesive soils was studied emphasizing limitations and problems

associated with the resistance to penetration including drop height and weight drop

(Fletcher, 1965).

DeMello (1971) presented the first comprehensive state of the art report on SPT where the

sampler penetration and factors impacting the penetration resistance were explored. These

factors were categorized as human factor, energy variation, borehole diameter, use of drill

mud, type of soil, and depth. The human factor was considered unsolvable other than


23

exercising due care or the implementation of statistical analyses where possible, to reduce

systematic errors (DeMello, 1971).

Effect of Drilling Rod and Type of Sampler

Palmer and Stuart (1957) approached the effect of drilling rod size on the energy traveled

through the drilling rod and analyzed this effect using the wave equation. The conclusion

was that the rod size had no major effect on the energy traveled except when the test is

performed in very soft soil.

The influence of the physical aspect of drilling rod on NSPT was evaluated by Brown (1977)

where results from six SPT borings using AWJ and NWJ rods were analyzed. The findings

of the study showed no significant difference between the NSPT obtained from the AWJ rod

and NWJ rod. Uto et al. (1981) compared results of SPT conducted using AWJ rod and JIS

(i.e. the NWJ version of the rod in Japan) in two different soil types. Results obtained from

the study indicated similar NSPT except that a slightly higher N-values were detected when

the test was conducted in hard clay.

Schmertmann (1978) and Schmertmann and Palacios (1979) evaluated the effect of the rod

size on the energy traveled by using strain gauges beneath the anvil and above the split

spoon sampler. The results indicated small variations of NSPT which were mainly related to

rod whip effect and varying slack of the thread joints.

Schmertmann and Palacios (1979) further explored the effect of type of sampler on the

SPT blowcount values, NSPT. The type of sampler refers to whether the split spoon is used


24

with or without a liner. The internal wall friction in the sampler is reduced when a liner is

not used resulting in larger sample recovery (Schmertmann and Palacios, 1979) but in

many cases when the SPT is completed without a liner the NSPT is affected due to plugging.

Effect of Driving Technique and Hammer Type

The donut hammer, safety hammer, and automatic hammer are three commonly-used

hammers in geotechnical exploration for field penetration tests such as the SPT and the

TCP, Figure 2.4. In a donut hammer a circular-shaped mass is lifted by the rope-cathead

arrangement and released by the driller to create the impact at the anvil level. In operations

where a safety hammer is used, the anvil is located inside a 63-kg (140-lb) sleeve (for SPT)

which is lifted and released similar to a donut hammer. In contrast, modern automatic

hammers use a mechanical apparatus to reliably achieve the required height and an

unrestrained hammer drop.


25

(a) (b) (c)

Figure 2.4. Different type of Hammers (a) Donut (b) Safety (c) Automated

(Coduto et al. 2011 used with permission)

A comparison study completed by Ireland et al. (1970) and further discussed by Frydman

(1970) showed that the NSPT obtained from a cat-head hoist was 1.4 times the NSPT obtained

from a self-tripping hammer. These results were confirmed by a further study (Lowther,

1973) indicating that the driving mechanism or release method has a significant impact on

the NSPT.

For a cat-head hoist mechanism (Kovacs et al. 1977) demonstrated that the impact velocity

could be affected by the number of turns of rope around the cat-head, the friction between

the rope and cat-head, and rope deterioration. In a more detailed study Kovacs and

Salomone (1982) completed 17 total field tests using safety hammers for eight tests and


26

donut hammer for the remaining nine tests. Energy ratio (i.e. hammer efficiency) of 50%

and 48% were reported for the safety and donut hammer, respectively.

For the SPT, Nixon (1982) examined the results from studies completed for the European

Standard to the results of studies completed in US and Japan. From the comparison, it was

concluded that before a final standardization of the SPT, it is important to conduct

comprehensive research studies to explore the influence of the type of hammer and sampler

on the NSPT. In another study (Robertson et al. 1983) alternated tests with a donut and safety

hammer. These tests were carried out in the same borehole utilizing the same drill rig and

drill rod with a slip-rope release method. The results from 16 tests showed an average of

62% energy ratio for the safety hammer and 43% for the donut hammer. This finding

followed the same pattern presented in the work completed by Kovacs and Salomone

(1982) where the safety hammer had higher energy ratio compared to the donut hammer.

Effect of Borehole Diameter

At the beginning of its origin, the SPT was completed at the bottom of 2.5-in or 4.0-in.

diameter boreholes. These borehole diameters are still considered common in geotechnical

engineering practice. For example in Japan, diameters such as 2.6-in and 3.4-in are used

(Yoshimi and Tokimatsu, 1983). However, in other places, boreholes with 6.0-in and 8.0-

inch diameters have been reported (Nixon, 1982). In cohesive soils the impact of borehole

diameter on NSPT can be neglected, however, in the case of sandy soils lower NSPT is

obtained with larger diameters (Lake, 1974 and Sanglerat & Sanglerat, 1982). This effect


27

is associated with the disturbance of the soil during testing and sampling procedure.

(Sanglerat and Sanglerat, 1982).

Development of correction factors for SPT N-values

The energy delivered into the drilling rod and the sampler is impacted by the release

method and type of hammer (Skempton, 1986). Based on the results of studies described

above and recommendations presented by Seed et al. (1984), a correction to the NSPT was

proposed by Skempton (1986) based on the hammer energy ratio reported in the United

States. The hammer energy ratio (Er), also termed “hammer efficiency”, refers to the

relationship between measured energy (Em) and theoretical energy (Et), Er = Em/Et *100%.

NSPT values with a known energy ratio can be normalized to 60% of hammer efficiency

using equation (2.1).

N60 = NSPT

Er

60 (2.1)

In equation (2.1), N60 is the normalized NSPT for 60% hammer efficiency, NSPT is the SPT

blowcount for 30-cm (12-in) penetration, and Er is the hammer efficiency for a specific

type of hammer.

In addition to the energy ratio, Skempton (1986) further developed work done by

Schmertmann and Palacios (1979) who determined that the required energy to drive the

sampler is increased as the rod length increases. He also considered research completed by


28

Seed et al. (1984) who showed that approximately 20% more blows corrected for 60% of

hammer efficiency were required to drive as SPT sampler without a liner. From this work,

Skempton (1986) determined correction factors for rod length, sampler type, and borehole

diameter, and modified Equation (2.1) by introducing correction factors:

N60 = NSPT Er

60CrCbCs (2.2)

In Equation (2.2), Cr, Cb, and Cs are correction factors to the SPT blowcounts to account

for hammer efficiency, rod length, borehole diameter, and type of sampler, respectively.

These correction have appeared in standard geotechnical textbooks such as Geotechnical

Engineering: Principles and Practice by Coduto et al. (2011) and Fundamentals of

Geotechnical Engineering by Das et al. (2014).

The overburden pressure correction for SPT

According to Meyerhof (1957) the penetration resistance increases linearly with depth, and

at a constant vertical stress, the resistance to penetration also increases at a rate

approximately equal to the relative density squared, Equation (2.3).

NSPT = 𝐷𝑅2 (a + b𝜎′

𝑣) (2.3)

In Equation (2.3) a and b are soil dependent factors, DR is the relative density, and ’v is

the effective vertical stress at the depth of test.


29

The effect of overburden pressure on the NSPT is further expressed by introducing a

correction factor for overburden pressure (CN). Skempton (1986) presented results of two

trials completed for sand deposits with different particle size. For these trials, a laboratory

model was built where the variation of NSPT with depth was determined while maintaining

a constant relative density (DR) of the sand. During the second trial, the test was carried

out with different relative densities of the sand. The results followed the relationship

presented by Meyerhof (1957) shown in equation (2.3). In addition, results from field

testing at different sites were compared to the laboratory model leading to the development

of an overburden correction factor CN. With the overburden pressure correction factor, the

NSPT and N60 can further be corrected as shown in equations (2.4) and (2.5).

N1−SPT = 𝐶𝑁NSPT (2.4)

N1−60 = Er

60CrCbCs𝐶𝑁NSPT (2.5)

In Equation (2.4) N1-SPT refers to SPT blowcounts corrected for overburden pressure. In

Equation (2.5) N1-60 refers to SPT blowcounts corrected for both 60% hammer efficiency

and overburden pressure. Researchers have developed many expressions for CN where

results have been supported by laboratory models and field test results. Table 2.2 presents

a summary of typical expressions developed for CN.


30

Table 2.2. Summary of existing Overburden correction factors (CN)

Contemporary practice for correcting SPT blowcount data

According to Schnaid (2009), in current practice, the SPT is the most common and popular

in-situ test method to obtain subsurface information and predict soil strength parameters.

As comparator and to assess the standard practice of corrected or normalized NSPT,

geotechnical manuals published by all 50 State DOTs were reviewed. Results show that

72% of the state DOTs specify correction factors shown in Equation (2.2) and further

corrections for effect of the overburden pressure using Equation (2.4) is suggested, Figure

2.5. For the selection of design parameters the N1-60 value is used. Furthermore, in some

Reference CN Observations

Peck et al. (1974) 0.77𝑙𝑜𝑔20

𝜎′𝑣 ’v in tsf

Seed (1976) 1 − 1.25𝑙𝑜𝑔𝜎′𝑣 ’v in tsf

Tokimatsu et al. (1983) 1.7

0.7+𝜎′𝑣 ’v in kg/cm2

Skempton (1986)

200

100+𝜎′𝑣 DR= 40-60% and NC Sand

300

200+𝜎′𝑣 DR= 60-80% and NC Sand

170

70+𝜎′𝑣 for OC Sand

’v in kPa

Liao et al. (1986)

√100

𝜎′𝑣 for NC Sand

[𝜎′𝑅𝐸𝑓

𝜎′𝑣]

𝑘 for k=0.4 to 0.6

’v in kPa

Clayton (1993) 143

43+𝜎′𝑣 for OC Sand

’v in kPa

Robertson et al. (2000) [𝜎′𝑣

𝜎′𝑎𝑡𝑚]

−0.5 for NC Sand

’v in kPa


31

states such as Florida, the specification requires the consultants to report NSPT in the boring

logs but mandates the design of all elements to be based on N1-60. Also, most DOTs require

the consultant to show proof of calibration of the hammer prior to any work completed for

the agency.

Figure 2.5. Standard Practice of NSPT Correction among State DOTs

Utah and Texas do not have requirements for the use of SPT since their local practice is

based on Becker borings in Utah, and the Texas Cone Penetrometer test in Texas,

respectively. According to the TxDOT Geotechnical Manual (2012), the use of SPT for

the design of deep foundations is acceptable by the agency following ASTM D-1586 but

no further guidance or specifications are provided. For the remaining 12 states, the SPT is

State DOTs with specifications

related to NSPT Correction

State DOTs without specifications

related to NSPT Correction

State DOTs with different

in-situ testing


32

described as part of their requirements for soil exploration and characterization but no

guidance or requirements are provided to address corrections for raw blowcounts.

Liquefaction studies

As part of several more recent studies performed for assessment of soil liquefaction

potential, Cetin et al. (2004) and Idriss and Boulanger (2004, 2008, 2012, and 2015)

explored the correction factors for overburden (CN) and rod length (Cr). Cetin et al. (2004)

evaluated the soil liquefaction potential and presented new correlations for the assessment

of the probability of triggering soil liquefaction. All correlation charts presented by Cetin

(2004) either as new development or as reference and comparator, are in function of SPT

blowcounts corrected for both overburden pressure (CN) and typical factors (Cr, Cb, and

Cs). CN and Cr factors were taken slightly different than those presented by Skempton

(1986). The correction factor for rod length follows the model developed by the National

Center for Earthquake Engineering Research (NCEER) work group established in 1997

and the overburden correction factor is taken after the work completed by Liao and

Whitman (1986). Idriss and Boulanger (2012 and 2015) explored the source of differences

in liquefaction triggering correlations. The study was completed by side-by-side

comparison of the correlations developed by Seed et al. (1984), Cetin (2004), and Idriss

and Boulanger (2004, 2008).


33

CORRECTIONS TO TCP BLOWCOUNT

Need for TCP correction factors

The predictive shear strength or allowable foundation capacity models represented in both

SPT and TCP foundation design methods source to empirical data obtained during an era

when the use of conventional donut hammers and safety hammers was a standard practice

for the completion of geotechnical field penetration tests. Because of similarities between

SPT and TCP test procedures where hammer, anvil, drill rod, and blowcounts are common

aspects in both tests, it is reasonable to consider that all uncertainties and variabilities

associated with the SPT could very well be present in the TCP test as well. Therefore,

blowcounts obtained from both SPT and TCP tests using today’s automatic hammers

should be corrected to a standard measurement of blowcounts.

TxDOT policy

TxDOT’s current Geotechnical Manual (2012) presents the guidelines for the use of TCP

as a field test and further refers to TEX 132-E as the approved TCP test method. The use

of an automatic hammer is specified in the Geotechnical Manual, but neither the

Geotechnical Manual nor the test method provides information regarding the need for

correction of NTCP or evaluation of the hammer efficiency. The use of an automatic trip

mechanism is specified to ensure the 0.61-m (24-in) required falling height, and this

requirement is part of TxDOT’s current geotechnical service contracts. Further, TxDOT’s

current geotechnical service contracts require that drilling subcontractors provide annual

certification of TCP hammer efficiency.


34

RESEARCH DESIGN AND METHOD

The TCP Reliability research study

The dataset for this paper was obtained from a research study where the reliability of the

TCP foundation design method was assessed (Seo et al. 2015). Based on the energy data

and soil information collected during field operations for that study, a complementary

research study was designed to further explore the TCP test hammer efficiency (Er-TCP),

rod length (CR-TCP) and overburden pressure correction factors (CN-TCP). A dataset of 293

TCP tests was compiled. These data source to 21 geotechnical borings at 16 different sites

in the states of Louisiana, Arkansas, Missouri, New Mexico and Texas as shown in Table

2.3.

Field work and TCP blowcount data sources

Per TxDOT Geotechnical Manual, TCP tests were performed at 1.5-m (5-ft) intervals

throughout the borings starting at the depth of 1.5-m (5-ft) below the ground surface and

ending at the maximum depth of boring. For identification and classification of the

subsurface materials associated with the TCP test, disturbed samples were collected using

either SPT split-spoon or thin-walled tube samples obtained directly below the TCP test

without having cleaned out the borehole. The sole purpose of the disturbed samples was to

identify and classify the material associated with each TCP test at depth.


35

Throughout the geotechnical drilling and sampling phase, four different drill rigs, each one

equipped with an automatic CME-hammer, were used. For the borings completed in

Missouri and Arkansas Northeast, AWJ rod was used; whereas, the rest of the borings were

carried out using NWJ rod. Prior to drilling operations, the automatic hammer was

disassembled and the hammer mass was weighed to ensure conformance with TCP test

standards (i.e. Tex-132-E).

Table 2.3. Field Research Sites and related TCP-Borings

Hammer energy readings

Force, acceleration, and rod length were measured for all TCP tests at 3.0-m (10-ft)

intervals using the SPT Analyzer. This device includes a subassembly of a 0.61-m (2.0-ft)

length of drill rod that is instrumented with two strain gage bridges and two piezo-resistive

accelerometers. These sensors obtained measured force and velocity signals for each

hammer blow and transferred the data to the SPT-Analyzer console. These data were

State Site Location

Geotechnica

l Borings

TCP

Tests in

Soils

TCP Tests

in

Rock/IGM

Instrumented

TCP Tests in

Soil

Instrumented

TCP Tests in

Rock/IGM

Missouri - Warrensburg 2 2 8 1 4

Missouri - Frankford 2 0 16 0 7

Texas - Houston 1 12 0 7 0

Arkansas - Siloam Springs 2 5 9 2 0

Arkansas - Monticello 2 37 0 20 0

Arkansas - Turrell 2 38 0 20 0

Louisiana -Various 7 102 0 50 0

New Mexico - Various 3 55 9 19 0

Total 21 251 42 119 16


36

collected and processed using the Pile Dynamics. Inc. (PDA 2013) software which

analyzes the force and acceleration recorded by the console for each hammer blow. The

theoretical energy was determined from the product of weight of the hammer and the drop

distance, measured during field operations.

After completion of field operations, data related to type of soil, drill rod size, drilling fluid,

and depth were carefully logged and recorded. Also, necessary data to perform analyses on

hammer efficiency was stored in the SPT-analyzer console. These data were used to

assemble the final dataset analyzed for this study.

The TCP blowcount dataset

The final dataset consisted of 293 TCP tests comprising 251 tests completed in soils and

42 in IGM/Rock. From the 293 TCP tests, 135 tests were instrumented to obtain energy

measurements using the SPT Analyzer, with 119 instrumented tests in soils and 16

instrumented tests in IGM/Rock. Considering that the study presented in this paper focuses

solely on soils, TCP blowcounts corresponding to the IGM/Rock materials have not been

included in the final dataset and as observed in Figure 2.6 (a), maximum TCP blowcounts

reported are less than or equal to 100 blows per 30-cm (12-in) of penetration. For the 119

instrumented TCP tests in soils (i.e. the data used for this study), the hammer energy for

each TCP hammer strike was recorded resulting in 4901 hammer energy measurements,

with the average hammer efficiency for each TCP test plotted in Figure 2.6(b).


37

(a) (b)

Figure 2.6. Compiled dataset (a) TCP Blowcount Values NTCP (b) Average Hammer

Efficiency

RESULTS FROM TCP BLOWCOUNT DATA

A series of factors associated with the completion of the TCP test impacts the energy

transferred to the TCP cone and thereby influences the measured NTCP blowcount values.

Following the customary standardization process for correcting NSPT blowcounts identified

in Equation (2.5), this paper presents TCP hammer efficiency data, correction factors for

rod length, and correction factors for overburden pressure. Correction factors for borehole

diameter and sampler type are not presented however, since for the TCP study the borehole

diameter was constant for all sites, i.e. 10-cm (4 in), and the factor for type of sampler does

not apply since the TCP point is a solid steel cone.

0

5

10

15

20

25

30

35

0 25 50 75 100D

epth

(m

)

TCP Blowcount Values (NTCP)

0

5

10

15

20

25

30

35

0.7 0.8 0.9 1

Dep

th (

m)

Average Hammer Efficiency


38

Hammer Efficiency for the TCP test (Er-TCP)

All TCP tests performed for this study were completed using Central Mine and Equipment

(CME) automatic hammers where the hammer mass is lifted by a hydraulically-operated

chain cam mechanism and the hammer is released using an automatic finger cam, allowing

the hammer to fall and impact an anvil located at the head of the drill string.

Figure 2.6(b) illustrates the TCP hammer efficiencies measured in this study. Using

Minitab 17.3.1 (2016) basic statistical parameters such as mean and coefficient of variation

(COV) for the TCP hammer efficiency were determined. Undifferentiated efficiency values

range from 74% to 101% with mean of 89% and COV of 5.07 associated with TCP

blowcounts ranging from 3 to 100 with an average of 38 blows/30-cm and measured at

depths ranging from 3-m (5-ft) to 30-m (100-ft), average 13-m (43-ft). Each data point

represents the mean hammer efficiency for an individual TCP test. This was determined

from the ratio of measured energy based on the SPT analyzer to the theoretical energy

calculated for the TCP hammer system (and also verified by field measurements) for each

hammer strike associated with a test. TCP hammer efficiency values ranged from 74% to

101% with a mean of 90% and COV of 4.85 for coarse-grained soils, and from 77% to

97% with an average of 88% and COV of 5.13 for fine-grained soils.

As has been noted, the TCP hammer is heavier and the drop is shorter compared to the

SPT. The TCP hammer efficiency values presented herein are similar in range but about

7% higher on average than SPT hammer efficiency values published in the literature. For

example, Honeycutt, et al. (2014) report energy transfer data from a research database for


39

CME automatic hammers and consisting of energy measurements from 17,825 individual

SPT hammer blows (analogous to the 4901 hammer energy measurements obtained for this

TCP study). Their dataset shows an average energy ratio of 82.9%, with a coefficient of

variation of +/-7.4%. In contrast to the SPT where numerous studies exist that identify

hammer efficiency, the data presented herein represent the first published TCP hammer

efficiency values.

Development of Rod Length Correction Factors (CR-TCP)

The relationship between the rod length and hammer efficiency has been explored

considering that the rod length correction for NTCP will follow a similar form as the

correction for NSPT. A linear model was created for all the geomaterials analyzed in this

study based on an equation following the form presented by Equation (2.6):

𝐸𝑟−𝑇𝐶𝑃 = 𝛽0 + 𝛽1 log(𝐷) (2.6)

In Equation (2.6), Er-TCP is the average hammer efficiency for the TCP test, 0 and 1 are

coefficients which will differ based on the soil type, and D is depth, that is length of rod

below ground surface. To develop rod length correction factors, Equation (2.6) is written

for any depth and a baseline depth. In this context, the baseline depth refers to a depth at

which the model line is flattened. At this depth (i.e. baseline depth) the correction factor

would be considered as unity. The subtraction of Equation (2.6) written for any arbitrary

depth from Equation (2.6) written for the baseline depth can be represented by Equation

(2.7):


40

∆ = 𝐸𝑟−𝑇𝐶𝑃 𝑎𝑡 𝐷1 − 𝐸𝑟−𝑇𝐶𝑃 𝑎𝑡 𝐷𝐵𝑎𝑠𝑒𝑙𝑖𝑛𝑒 = 𝛽1(log(𝐷1) − log(𝐷𝐵𝑎𝑠𝑒𝑙𝑖𝑛𝑒)) (2.7)

In Equation (2.7), D1 is the depth of interest and DBaseline is the depth at which D1 is

compared against. Identifying an estimate and confidence interval (CI) for requires

finding an estimate and confidence interval for . For this analysis Minitab 17.3.1 (2016)

was used to determine the value of for each type of soil with a 95% confidence interval.

Substituting lower and upper endpoints of the CI in Equation (2.7) will result in a 95% CI

for . The result of this operation is an estimate and CI for the correction factor for rod

length (CR-TCP) determined by Equation (2.8):

𝐶𝑅−𝑇𝐶𝑃 = log𝐸𝑟−𝑇𝐶𝑃 𝑎𝑡 𝐷1

𝐸𝑟−𝑇𝐶𝑃 𝑎𝑡 𝐷𝐵𝑎𝑠𝑒𝑙𝑖𝑛𝑒 (2.8)

In Equation (2.8), CR-TCP is the correction factor for rod length, and Er-TCP is the average

hammer efficiency for the TCP test at the corresponding depth.

Considering that the TCP design charts were generated during a period of time when the

use of hammers with nominal average efficiency of 60% was considered as the standard

practice, it is reasonable to consider the normalization of the hammer efficiency for the

TCP test by writing Equation (2.1) for the TCP test:

𝑁60−𝑇𝐶𝑃 = 𝑁𝑇𝐶𝑃

𝐸𝑟−𝑇𝐶𝑃

60 (2.9)


41

In Equation (2.9), N60-TCP is the TCP blowcount standardized for 60% of hammer

efficiency, NTCP is the TCP blowcount, and Er-TCP is the hammer efficiency for the TCP

test in percentage.

After standardization of the NTCP for 60% of hammer energy, Equation (2.2) can be

rewritten for the TCP test, equation (2.10).

𝑁60−𝑇𝐶𝑃 = 𝑁𝑇𝐶𝑃

𝐸𝑟−𝑇𝐶𝑃

60𝐶𝑅−𝑇𝐶𝑃 (2.10)

Rod length correction factors (CR-TCP) were developed by analyzing values corresponding

to a baseline depth of 24-m (80-ft) and using the models shown in Equations (2.7) and

(2.8). For this study, the baseline depth was considered as the rod length below ground

surface at which the model flattened and presented a linear tendency. Figure 2.7 presents

average hammer efficiency versus rod length below the ground surface.


42

Figure 2.7. Scatterplot of Average Efficiency vs Rod length for all Soils

Figure 2.7 presents two rod length correction datasets side-by-side, the SPT dataset

(Skempton 1986) and the TCP dataset for this study. For the TCP data, the solid line

represents the predictive model created based on Equation (2.6) with an R-square value of

43%, p-value of 0.000 (p-value<0.0001), and the 95% CI boundaries are represented by

the dashed lines closer to the model line. Results for the TCP test suggest that the lowest

data variability is presented at depths greater than 24-m (80-ft) where the depth correction

factor is unity. In case of the SPT, it is noted that the model line flattens for depths greater

than 10-m (30-ft). Furthermore, observing the depth correction factors for each test at the

0.00

10.00

20.00

30.00

40.00

0 0.2 0.4 0.6 0.8 1 1.2

Ro

d L

engt

h B

elo

w G

rou

nd

Su

rfac

e (m

)

Average Hammer Efficiency

Fine Grained Soils


TCP-Model Line

Confidence Intervals

Prediction Intervals

Skempton (1986) Model

Skempton (1986)

S= 0.0337R2= 43.4%p-value= 0.000


43

same depth, a difference in value is observed. This difference could be primarily associated

with the hammer used for the test and the type of samplers. The dataset analyzed by

Skempton (1986) was based on SPT data obtained using a safety hammer; whereas, the

data analyzed in this study were obtained from TCP tests completed using an automatic

hammer.

Side by side comparison of TCP and SPT Rod Length correction factors

Table 2.4 presents details of the TCP rod length correction factors. As analyzed in this

dataset, coarse-grained soils consisted of poorly graded sands (SP), silty sands (SM), and

clayey sands (SC) with SP being the predominant soil type. For the coarse-grained soils,

rod length correction factors (CR-TCP) were obtained ranging from 0.92 to 1.01. Similarly,

the types of soil described as fine-grained consisted of lean clay (CL), fat clay (CH), and a

low plasticity silt (ML) with CL being the predominant soil. For the fine grained soils, rod

length correction factors (CR-TCP) were obtained ranging from 0.90 and 1.01. A statistical

test established that there are no significant differences between the correction factors for

the coarse and fine soils based on calculated probability p greater than 0.05 (p-value =

0.48).


44

Table 2.4. Rod Length Correction Factors (CR) for TCP and SPT

Depth

(ft)

CR-TCP Depth

(m)

CR-SPT

Coarse

Grained

Fine

Grained

Undifferentiated Recommended Undifferentiated

Soil Type

3 0.92 0.90 0.90 0.90 3-4 0.75

6 0.94 0.93 0.94 0.94 4-6 0.85

9 0.96 0.95 0.95 0.95 6-9 0.95

12 0.97 0.97 0.97 0.97 >9 1.00

15 0.98 0.98 0.98 0.98 >9 1.00

18 0.99 0.99 0.99 0.99 >9 1.00

21 0.99 0.99 0.99 0.99 >9 1.00

24 1.00 1.00 1.00 1.00 >9 1.00

27 1.00 1.01 1.01 1.00 >9 1.00

30 1.01 1.01 1.01 1.00 >9 1.00

Considering the narrow range of rod length correction factors for differentiated soils (i.e.

fine grained and coarse grained soils), the factors recommended correspond to the model

developed for undifferentiated data with values ranging from 0.9 to 1.0 as shown in Table

2.4.

Development of Overburden pressure Correction Factors (CN-TCP)

To develop the overburden correction factor for the TCP test, Equations (2.4) and (2.5) can

be written for the TCP blowcounts (NTCP):

N1−TCP = 𝐶𝑁−𝑇𝐶𝑃NTCP (2.11)

N1−60−TCP = Er

60CR−TCP𝐶𝑁−𝑇𝐶𝑃NTCP (2.12)


45

The relationship between standardized TCP blowcounts (N60-TCP) and TCP test blowcounts

corrected for overburden pressure (N1-60-TCP) can be defined as the overburden correction

factor for the TCP test.

1

𝐶𝑁−𝑇𝐶𝑃=

N60−TCP

N1−60−TCP (13)

In order to develop the correction factor for overburden pressure, a reference stress value

should be defined first. In this study, the atmospheric pressure (= 100 kPa = 2,000 psf) is

taken as the reference stress (ref) following the standard practice for SPT. Therefore, N60-

TCP value obtained at a vertical stress of 100 kPa becomes a reference value (N60-TCP-ref) for

the development of overburden correction factor. Then, three main steps were followed to

develop overburden correction factors for the TCP test (CN-TCP). These steps were: (1)

normalizing the TCP blowcounts with respect to N60-TCP-ref, (2) normalizing the effective

vertical stress with respect to ref, and (3) determining the relationship between steps (1)

and (2). Prior to these steps, all TCP blowcounts (NTCP) were corrected for hammer

efficiency and rod length following Equation (2.10).

Step 1. Normalized TCP Blowcounts (NN-TCP)

For each TCP boring included in the compiled dataset, the relationship between (N60-TCP)

and the effective vertical stress (’v) was analyzed and a linear regression model was

developed as shown in Figure 2.8.


46

Figure 2.8. N60-TCP and Effective Vertical Stress relationship for each TCP boring

From the interception of the horizontal dashed line depicting the reference stress of 100-

kPa (2000-psf) and each solid line depicting the linear model fitted to the data

corresponding to each boring, the N60-TCP-ref corresponding to the reference stress was

determined. At the end of this step, each boring was associated with one TCP blowcount

value (N60-TCP-ref) and this value was used for the normalization of the TCP blowcounts:

𝑁𝑁−𝑇𝐶𝑃 = (𝑁60−𝑇𝐶𝑃

𝑁60−𝑇𝐶𝑃−𝑟𝑒𝑓) (14)

0

50

100

150

200

250

300

350

400

0 50 100 150 200 250 300 350

' v

(kPa

)

N60-TCP

WAR-1 Turrell-1 Turrell-2 Monticello-1

Monticello-2 Crosby-1 Ragley-1 Essen Lane-1

Highlan Park-1 Causeway BLVD-1 Causeway Bridge-1 I49 US71-1

Sunland Park-1 ABQ-1 Ohkay Owingeh-1 I-49 Unknown Pond-1

Reference Stress (Pa), 2000 psf


47

Step 2. Normalized Effective Vertical Stress (’N)

Based on available geotechnical information for each site, unit weights were assigned to

each soil layer. In the case where measured unit weights of soil were not available,

correlations published by TxDOT (2000) and Bowles (1995) were used to determine the

soil unit weight. For all TCP tests, the effective vertical stress was calculated at the depth

of test. Furthermore, the effective vertical stress values determined in this step were

normalized to a reference stress of 100 kPa (2000 psf) using Equation (2.15) below:

𝜎′𝑁 = (𝜎′

𝑣

𝜎𝑟𝑒𝑓) (15)

Each normalized effective vertical stress (’N) was then associated with its corresponding

normalized TCP blowcount.

Step 3. Relationship between NN-TCP and ’N

After completion of Steps 1 and 2, a scatterplot was generated with NN-TCP (= N60-TCP/N60-

TCP-ref) as dependent variable and ’N (=’v/ref) as the independent variable and the

relationship between these variables was further analyzed. Figure 2.9 presents a scatterplot

for NN-TCP versus ’N together with the fitted regression curve from a power model.


48

Figure 2.9. Scatterplot for Normalized TCP Blowcounts (NN-TCP) vs. Normalized

Effective Vertical Stress (’N)

Linear and polynomial regression models were fit to the dataset, and it was observed that

a key assumption of regression was violated; namely that of a constant (stable) variance.

The variance issue was addressed by log10 transformation of the dependent and

independent variables and statistical models with reliable confidence intervals were

determined. After the transformation, a model was generated for the dataset following

Equation (2.16) and converted from log10 scale back to natural scale to obtain Equation

(2.17), which indicates a power function relationship between NN-TCP and ’N. Figure 2.9

0

0.5

1

1.5

2

2.5

3

3.5

4

0 1 2 3 4 5

' v

/' re

f

N60-TCP/N60-TCP-Ref

Coarse Grained Soils Fine Grained Soils

TTU-TCP Model Confidence Intervals

R2= 30% p-value=0.000k = 0.73

N60-TCP/N60-TCP-Ref = ('v/ref)0.73


49

presents the model fitted to the dataset based on Equation (2.16) illustrated by the solid line

including its corresponding 95% confidence intervals (CI) identified by dashed lines.

𝑙𝑜𝑔10(𝑁𝑁−𝑇𝐶𝑃) = 0.73 𝑙𝑜𝑔10(𝜎′𝑁) (16)

(𝑁60−𝑇𝐶𝑃

𝑁60−𝑇𝐶𝑃−𝑅𝑒𝑓) = (

𝜎′𝑣

𝜎𝑟𝑒𝑓)

0.73

(17)

Overburden correction factor CN-TCP

From the combination of Equations (2.13) and (2.17), the overburden correction factor for

the TCP test (CN-TCP) is obtained as follows.

𝐶𝑁−𝑇𝐶𝑃 =1

(𝜎′

𝑣

𝜎𝑟𝑒𝑓)

0.73 (2.18)

Equation (2.18) represents a power function which is the recommended overburden

correction factor for the TCP tests with a power (k) of 0.73. To provide context, the

expression determined for CN-TCP in this study was compared to the models published for

CN for SPT shown in Table 2.2, by plotting the effective vertical stress (’v) versus CN, as

shown in Figure 2.10.

Figure 2.10 illustrates the relationship between the effective vertical stress and the

overburden correction factors published for the SPT compared to the overburden correction

factor for the TCP presented in this paper. The relationship for the TCP test is shown with


50

the solid line with 95% CI dashed lines. From the comparison, it is noted that the

overburden correction factors presented by other researchers for the SPT data vary between

0.35 and 2.1; whereas, the correction factors for the TCP test range from 0.4 to 3.05. It is

inferred that the broad spectrum of the overburden correction factor for the TCP test may

be associated with the fact that the TCP test data analyzed in this study included both fine-

grained and coarse-grained soils whereas the factors developed for the SPT included

coarse-grained soils only.

Figure 2.10. Effective Vertical Stress (’v) versus CN for TCP and SPT

0

50

100

150

200

250

300

350

400

0 0.5 1 1.5 2 2.5 3 3.5

'v

(kP

a)

CN

Peck et. al. (1974)

Seed (1976)

Tokimatsu (1983)

Skempton(1986) DR=40-60%

Skempton(1986) DR=60-80%

Skempton (1986) OC Sand

Liao Whitman (1986)

Clayton (1993)

TTU-CN-TCP

TTU-CN-TCP-Confidence Intervals


51

OTHER FACTORS THAT MAY INFLUENCE TCP TEST HAMMER EFFICIENCY

Analyses of the TCP test data, observations of the TCP field procedure, and comments in

the technical literature suggest that factors other than those corrected for herein could create

variability in the NTCP. Four additional factors – drill rod size, type of soil, drilling fluid,

and depth – were identified and analyzed to explore their influence on TCP hammer

efficiency.

Statistical analyses using ANOVA were performed using SAS9.3 (2015) and Minitab

17.3.1 (2016) to study the effect of drill rod size, type of soil, drilling fluid, and depth, on

the hammer efficiency. For this process, the hammer efficiency was considered as

dependent variable and all other factors as independent variables/factors. This ANOVA

also tested for interactions between these factors. A three-way interaction is observed

between drilling fluid, soil type, and depth. This implies that each of these are associated

with hammer efficiency. Furthermore, drill rod size is involved in two significant two-way

interactions, which implies that it is also associated with hammer efficiency.

Results from these analyses indicate that a statistically significant relationship exists

between the identified factors and hammer efficiency, but because the data were not

obtained from a controlled experiment, the exact nature of these relationships could not be

determined. Hence, to develop correction factors that account for all the variables

influencing TCP hammer efficiency, additional study is required where the research design

allows each factor to be analyzed independently.


52

SUMMARY AND CONCLUSIONS

This paper presents TCP hammer efficiency data, rod length correction factors, and

overburden correction factors for instrumented TCP tests completed in coarse-grained and

fine-grained soils. Other factors that may influence TCP hammer efficiency were also

identified.

A dataset of 293 TCP tests comprising 251 tests completed in soils and 42 in IGM/Rock

was compiled for this study. From the 293 TCP tests, 135 tests were instrumented to obtain

energy measurements using the SPT Analyzer, with 119 instrumented tests in soils and 16

instrumented tests in IGM/Rock. An adjusted dataset of 251 TCP tests were considered for

the development of overburden correction factors for the TCP test in soils. A total of 119

instrumented TCP tests in soils were used for the analysis corresponding to the rod length

correction factors.

Undifferentiated efficiency values range from 74% to 101% with mean of 89% and COV

of 5.07 associated with TCP blowcounts ranging from 3 to 100 with an average of 38

blows/30-cm and measured at depths ranging from 3-m (5-ft) to 30-m (100-ft), average 13-

m (43-ft). TCP hammer efficiency values ranged from 74% to 101% with a mean of 90%

and COV of 4.85 for coarse-grained soils, and from 77% to 97% with an average of 88%

and COV of 5.13 for fine-grained soils. These TCP hammer efficiency values are similar

in range but about 7% higher on average than efficiency values for automatic SPT hammers

published in the literature.


53

Rod length correction factors for the TCP hammer efficiency (CR-TCP) were developed by

creating a non-linear model based on statistical analyses. Rod length correction factors

ranging from 0.90 and 1.00 for undifferentiated data were determined and recommended

for the TCP test.

In a separate set of analyses, the relationship between N60-TCP and soil overburden pressure

was analyzed. Correction factors for the overburden pressure for the TCP test (CN-TCP) were

developed and normalized to a reference stress (100 kPa). The CN-TCP expression

recommended in this paper is a power function with k= 0.73, and this is generally in

agreement with the CN expressions published for the SPT.

Although the TCP test is in some ways similar to the SPT, the two tests are not the same,

even though results from both tests are used to design deep foundations. Therefore, it

makes sense to take steps to ensure reliable data. In case of the SPT, the corrections for

raw blowcount data are well established; whereas, for the TCP test, up until now, no

published corrections exist. The work presented in this paper is intended as an initial step

to further refine the TCP-based design to achieve more reliable design of deep foundations.


Note 3: Moghaddam R.B., Jayawickrama, P.W., Lawson, W.D., Seo H., Surles, J.G. (2016), “Evaluation of

Predicted Validity of the Texas Cone Penetration Design Charts for Deep Foundations based on Full Scale

Load Tests”, Journal of Performance of Constructed Facilities

54

CHAPTER III

EVALUATION OF PREDICTED VALIDITY OF THE TEXAS CONE PENETRATION

DESIGN CHARTS FOR DEEP FOUNDATIONS BASED ON FULL SCALE LOAD

TESTS3

ABSTRACT

This study presents a quantitative and a qualitative evaluation of the predictive validity of

the Texas Cone Penetration (TCP) foundation design charts. Allowable measured total

capacities determined from results of full-scale load tests were compared to allowable

predicted total capacities determined from TCP foundation design charts. The final dataset

compiled for this study consisted of 60 full scale load testes comprising 33 driven piles and

27 drilled shafts. Allowable measured capacities were determined using strength-based and

serviceability-based models. Results from analyses suggested that it is apparent that

different performance assessment criteria leads to different conclusions regarding the

predictive validity of the TxDOT TCP method for deep foundation design. The allowable

predicted total capacity determined from the TCP foundation design charts could be

considered to be reasonable based on a strength-based model, moderately over-

conservative, to over-conservative according to serviceability-based models. Further, it

was observed that as the level of tolerable settlement is increased, the interpretation of the

TCP foundation design charts becomes even more conservative.


55

INTRODUCTION

This paper presents an evaluation of the predictive validity of the Texas Department of

Transportation (TxDOT) Texas Cone Penetration (TCP) foundation design charts by

comparing the allowable predicted total capacity (determined from the TCP foundation

design charts) to the allowable measured total capacity (obtained from full scale load tests)

for driven piles and drilled shafts in soils.

The TCP field test procedure is introduced, and its associated foundation design charts are

presented. The application and use of the TCP charts for the design of deep foundations is

discussed in detail. The effort associated with field data collection including geotechnical

borings using the TCP tests is described, followed by a detailed discussion of the

development of the project dataset. Two different approaches were used to determine the

allowable measured total capacity from results of full-scale load tests. The first approach

was a strength-based method where the total capacities at the ultimate limit state were

determined from the load-settlement curves and then divided by a Factor of Safety (FS) to

determine the allowable measured total capacity. The second approach was serviceability-

based where loads corresponding to tolerable settlements were identified from the load-

settlement curves, and these loads were considered as the allowable measured total

capacity.

After determining the allowable measured total capacity and the allowable predicted total

capacity, these data were analyzed based on (1) a qualitative evaluation where all data were


56

plotted and compared to in reference to the line of equality, and (2) a quantitative evaluation

where statistical analysis and linear regression models were developed to analyze the

relationship between allowable measured total capacity and the allowable predicted total

capacity.

The TCP test and its associated foundation design charts have been used for the design of

driven pile and drilled shaft foundations that support tens of thousands of bridges and other

major transportation infrastructure throughout Texas for over 60 years.

As would be expected with its Texas roots, the TCP test has seen extensive application

throughout Texas to evaluate subsurface materials ranging from very soft clays, to shales

and fractured limestones, to hard rock. However, the use of TCP test and its associated

design charts are not limited to Texas practice only. Because of its applicability to a wide

range of geomaterials, the TCP test has been used not only throughout the United States

but also for international projects. According to FHWA (2016) Texas currently maintains

over 53,000 bridges in the National Bridge Inventory, and most of these bridges plus other

transportation infrastructure in Texas are supported by foundations designed in accordance

with the TCP method. Also, since Oklahoma DOT adopted the TCP test in the 1970s, the

TCP method has been used for the design of the foundations for most of the 23,000 bridges

in Oklahoma’s inventory. Collectively this indicates that over 12% of the nation’s bridges

are associated with the TCP test. The TCP test is discussed in NCHRP Synthesis 360

(Turner 2016), and owing to competitive opportunities for Texas’ major transportation


57

infrastructure projects, geotechnical practitioners and academics nationwide have had

occasion to learn about and use the TCP test. Recently, a comprehensive research program

in Missouri evaluated a modified version of the TCP test for bridge foundation design

applications using LRFD concepts (Loehr et al. 2011). The TCP test was introduced to an

international audience during a workshop sponsored by the Port and Airport Research

Institute of Japan, and the application of the method for the design of deep foundations

was discussed as part of Advance in Deep Foundations in Japan, (Vipulanandan, C. 2007).

Furthermore, the TCP test has been used in Korea for field explorations associated with a

bridge abutment overlying intermediate geomaterials and soft rock where cores would have

been recovered in fragments because of joint structures (Nam et al. 2013). In light of the

preceding, the scope of the TCP test and its associated foundation design charts is

significant well beyond its regional origins.

THE TXDOT TEXAS CONE PENETRATION TEST

Description of the TCP test

The TCP field test is a dynamic penetration test which assesses the strength of the geologic

material encountered during geotechnical exploration. This test method is considered to be

a simple, rugged, yet sufficiently reliable tool for characterization of a broad range of

geologic material that range from clays, sands, intermediate geomaterials and hard rock.

The TCP test method is documented as TxDOT Designation Tex-132-E, “Test Procedure

for Texas Cone Penetration” (TxDOT 1999). It uses a 77.0-kg (170-lb) hammer with 0.61-


58

m (24-in) drop to force a 7.6-cm (3-in) diameter steel cone into the soil or rock formation,

Figure 3.1.

In current practice, penetration for the TCP test is to be achieved in three separate

increments. The first increment is the seating blows and consists of driving the cone 12

blows or approximately 15-cm (6-in), whichever comes first. The TCP blowcount is then

determined as the sum of the number of blows required to achieve the second and third 15-

cm (6-in) increments. The total blowcount or (NTCP) corresponding to 30-cm (12-in)

penetration is used to obtain design parameters. In very hard materials such as rock and

intermediate geomaterials (IGM), after the proper seating process, the cone is driven 100

blows and the penetration values for the first and second 50 blows are recorded, with the

NTCP value reported as centimeters (inches) of penetration per 100 blows.

TCP Design Charts

The axial load capacity of a deep foundation is governed by the shear strength of the

geomaterial supporting the foundation at both shaft and base level. Therefore, reliable

assessment of the shear strength of the geomaterial is necessary for accurate prediction of

the axial load of the foundation.


59

(a) Conical Driving Point (TxDOT 1999) (b) Field Application

Figure 3.1. TCP Test Conical Driving Point (TxDOT, 1999)

One of the unique aspects of the TCP test is that it was developed to evaluate all types of

subsurface materials using a consistent in-situ field test method (TxDOT 2012). Other field

procedures such as the SPT and the Cone Penetration Test are only applicable for a

particular subset of geomaterials, but the TCP test can be used in all materials from soft

clays to dense sands to intermediate geomaterials and rock (TxDOT 2000).

The 1956 edition of the Foundation Exploration and Design Manual provide a series of

correlation curves illustrating shear strength estimation based on NTCP. These correlation

curves were based on relationships established between TCP test data and laboratory-

measured shear strength. To obtain soil strength parameters for this purpose, undisturbed

samples were collected and tested using the triaxial test procedure (THD 1956). Based on

data obtained from field and laboratory studies, correlation charts between NTCP and

foundation capacities were developed and published in two separate sets of design charts.

The first set was developed for soils with TCP blowcounts less than 100 per 30-cm (12-


60

in); whereas, the second set was for geomaterials with blowcounts greater than 100 blows

per 30-cm (12-in) (i.e. penetration per 100 blows). Figure 3.2 presents the current version

of the TCP foundation design charts for soil as published in the 2012 TxDOT Geotechnical

Manual (TxDOT 2012).

Deep foundations such as driven piles and drilled shafts derive their axial load capacity

from both shaft resistance (also referred to as skin resistance or skin friction in literature)

and base resistance (also referred to as toe bearing resistance or point bearing resistance in

literature). The TxDOT’s Geotechnical Manual TxDOT (2012) refers to these resistances

as skin friction and point bearing, respectively. The manual provides two separate sets of

charts, one set to determine allowable skin friction based on measured TCP blow count and

another set to determine allowable point bearing resistance. Figure 3.2 shows the design

charts used for soil materials, i.e. materials in which blowcounts per 300-mm (12-in) is less

than 100. The TCP foundation design charts reflect the allowable stress design (ASD)

philosophy and present allowable strength values corresponding to a Factor of Safety (FS)

of 2.0. Furthermore, it can be seen that these design charts are categorized based on soil

classification where separate correlations are provided for fat clay (CH), Lean Clay (CL),

Clayey Sand (SC), and OTHER soils. According to the design procedure, any soil

calssified as SP, SW, SM, and ML is considered as OTHER category.


61

(a) (b)

Figure 3.2. Design charts representing (a) allowable unit shaft resistance and (b)

allowable unit base resistance vs. TCP blows/30-cm (12-in), (TxDOT, 2012)

From Figures 3.2(a) and 3.2(b) it is noted that the soil constants are reported for materials

softer than 100 blows per 300-mm (12-in) of penetration. These constants represent the

inverse of the slope of the allowable capacity models and the slopes vary based on soil

classification. From the nature of the TCP chart for the shaft resistance it is noted that all

predictive models present a maximum blowcount value beyond which the resistance value

becomes constant. For example, in Figure 3.2(a) the model for fat clays (CH) reports

varying unit shaft resistance NTCP =5 to NTCP =70 blows/ 300-mm (12-in), and constant

shaft resistance of 134-kPa (1.4-tsf) for blowcounts beyond 70 blows.

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

1.1

1.2

1.3

1.4

1.5

0 10 20 30 40 50 60 70 80 90 100

S -

Allo

wab

le S

kin

Fri

ctio

n (

tsf)

wit

h F

.S. =

2.0

Texas Cone Penetrometer (NTCP)- Blows/12-in

CH CLSC

OTHER

CH 50

CL 60

SC 70

OTHER 80

S = N/C

C (Constant)

Note: Apply soil reduction factor for drilled shaft design.

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

5.0

5.5

6.0

6.5

0 10 20 30 40 50 60 70 80 90 100

Pb

-A

llow

able

Po

int

Bea

rin

g (t

sf)

wit

h F

.S. =

2.0

Texas Cone Penetrometer (NTCP)- Blows/12-in

CH

CL

SC

CH 19.5

CL 16.6

SC 16.3

OTHER 18.6

S = N/C

C (Constant)


62

TCP Design Method for Deep Foundations

In the design of deep foundations the total resistance (Qt) is the sum of the shaft (Qs) and

base (Qb) resistances. For the TCP method, the total resistance is expressed in function of

unit shaft (qs) resistance, foundation surface area (As), unit base resistance (qb), and

foundation base area (Ab), Equation (3.1).

𝑄𝑡 = 𝑞𝑠𝐴𝑠 + 𝑞𝑏𝐴𝑏 (3.1)

For a TCP-based design, the unit shaft and unit base resistance values are directly from the

TCP design charts. The main parameters required to determine these resistances are the

TCP test blowcount number (NTCP) and the type of soil described based on Unified Soil

Classification System (USCS) group symbol.

The allowable predicted total capacity for driven piles and drilled shafts compiled for this

study was calculated following the guidelines recommended in the Geotechnical Manual

(TxDOT 2012). Two capacity adjustments are specified for the TCP-based design of deep

foundations. The first adjustment is the disregard depth which applies to all type of

foundations and refers to the portion of shaft capacity neglected during design. According

to the Geotechnical Manual (2012), the shaft resistance from the ground surface to a depth

of 1.5-m (5-ft) to 3.0-m (10-ft) will be disregarded, irrespective of whether the deep

foundation will be installed over non-water crossing or over stream crossing. The second

capacity adjustment is the application of a soil reduction factor of 0.7 to the shaft resistance


63

for the design of drilled shafts. This reduction factor is recommended for use in order to

account for the soil disturbance during the drilling procedure.

The allowable predicted total capacities presented in this study include the soil reduction

factor of 0.7 for the determination of predicted shaft capacity of drilled shafts. The

disregard depth is usually applied to design to account for possible long-term scours. In the

case of this study, predicted capacity is compared to measured capacity from full-scale load

tests where a design assumption such as the disregard depth is not applicable.


Development of the Dataset

The dataset for this study was compiled based on available load test projects from

TxDOT’s historical archive supplemented by projects from neighboring states. TCP boring

data corresponding to full-scale load test projects retrieved from TxDOT’s historical

archive were identified and recorded. For the non-TxDOT projects, new geotechnical

borings using the TCP test method were completed.

TxDOT Historical Archive. Over a period of 50 years, the TxDOT Bridge Division

compiled an archive of load tests completed for driven piles and drilled shafts projects at

various locations in Texas. The TxDOT archive was reviewed in detail and organized

according to (1) availability of TCP data, (2) full scale load test information, and (3)


64

pertinence to this particular study. Based on this organization, a preliminary dataset was

compiled comprising 31 load test projects in soils for driven piles and 15 load test projects

in soils for drilled shafts, including their corresponding TCP data.

Non-TxDOT Full Scale Load Tests. With the objective of supplementing the TxDOT

historical archive, available load test project sites located in Missouri, Arkansas, Louisiana,

and New Mexico were identified and, at each site, new geotechnical borings with TCP tests

were completed in accordance with the guidelines found in the TxDOT Geotechnical

Manual. All TCP tests were carried out at 1.5-m (5-ft) intervals, starting at a depth of 1.5-

m (5-ft) below the ground surface, and ending at the maximum depth of boring. Disturbed

samples were collected using either SPT split-spoon or thin-walled tube samples obtained

directly below the TCP test without having cleaned out the borehole. These samples were

used for identification and classification of the subsurface materials associated with the

TCP test. After field data collection, two load test projects for driven piles and 12 load test

projects for drilled shafts were added to the preliminary dataset.

From this work, the final dataset compiled for this study consisted of 60 full scale load tests

comprising 33 driven piles and 27 drilled shafts in soils. Driven piles selected for this

project were precast square concrete piles ranging in width from 360-mm (14-in) to 510-

mm (20-in) with embedment lengths varying between 5-m (15-ft) and 25-m (84-ft). For

the drilled shafts, the diameters ranged from 460-mm (18-in) to 1830-mm (72-in) with total


65

length varying between 6-m (20-ft) and 30-m (99-ft). The final dataset is summarized in

Table 3.1 for driven piles and Table 3.2 for drilled shafts.

As previously mentioned, the primary objective of this study was to evaluate the predictive

validity of the TCP foundation design method based on a comparison between the

allowable predicted total capacity (𝑃𝑎𝑃) and the allowable measured total capacity (𝑃𝑎

𝑀). To

achieve this evaluation, the compiled dataset was further developed to obtain 𝑃𝑎𝑃 from

results of TCP borings, and 𝑃𝑎𝑀 from results of full scale load tests.

Allowable Predicted Total Capacity (𝑷𝒂𝑷)

Allowable predicted total capacity, 𝑃𝑎𝑃 values for driven piles and drilled shafts were

calculated following the design procedure specified by the TxDOT’s Geotechnical

Manual. For this purpose, a synthesized soil profile showing the stratigraphy and NTCP per

strata was generated from geotechnical TCP borings for each load test project. Figure 3.3

illustrates an example of the synthesized soil profile and the average NTCP profile developed

for Case No. 23 of drilled shafts. Then, with the average NTCP, the type of soil, and using

Figure 3.2, the allowable unit shaft and the allowable unit base resistance are determined.

Further, multiplying these unit resistances by their corresponding area (i.e. shaft area and

base area) and summing the total shaft and total base capacity, the allowable predicted total

capacity (𝑃𝑎𝑃) is determined.


66

Figure 3.3. Illustrative example of determining 𝑃𝑎𝑃 for Case No. 23 of drilled shafts

For the Case No. 23 shown in Figure 3.3, and following the above described process, an

allowable total shaft capacity of 2893-kN (650-Kips) and an allowable total base capacity

of 1073-kN (241-kips) were determined. For the calculation of shaft capacity for the design

of drilled shafts, the TxDOT Geotechnical Manual specifies a soil reduction factor of 0.7.

After applying the reduction factor, a total reduced shaft capacity of 2025-kN (455-Kips)

was obtained. Summing up the total reduced shaft capacity and the total base capacity, an

allowable predicted total capacity of 3000-kN (675-kips) was determined. This process

was completed for all driven piles and drilled shafts included in the compiled dataset. Table

18

16

13

16

22

51

1291

49

79

91

100

0

1.5

3

4.5

6

7.5

9

10.5

12

13.5

15

16.5

18

19.5

21

0 10 20 30 40 50 60 70 80 90 100 110 120 130

Dep

th (

m)

NTCP (Blows/30-cm)

Full

Scal

e L

oad

Te

st-

Dri

lled

Sh

aft,

Dia

me

ter=

16

80-

mm

TCP-Boring

CL

ML (OTHER)

SM (OTHER)

Average NTCP = 17

Average NTCP = 90

SM (OTHER)

NTCP = 18

Average NTCP = 70

NTCP = 51

NTCP = 12 SM (OTHER)

SM (OTHER)


67

3.1 summarizes the results obtained from for driven piles and while Table 3.2 provides the

results obtained for drilled shafts.

In addition to allowable total capacities, the predominant soil for each load test was

identified based on percent of contribution to the total capacity. The predominant soil type

was then categorized as either fine or coarse grained soil. The soil type corresponding to a

percentage greater than 50% of contribution to the total capacity was considered as the

predominant soil. For this selection, first, the predominant source of capacity was

identified. If the predominant source of resistance was the base of the foundation, then the

material below the base was selected as the predominant soil. In cases where the

predominant source of resistance was the shaft, then the highest percentage of contribution

from the soil type was selected as the predominant soil. For example, for the Case No. 23

shown in Table 3.2, the predominant source of total capacity is the shaft with 70% of

contribution to the 𝑃𝑎𝑃 and the predominant soil contributing to the shaft capacity is coarse

grained soil. Therefore, the predominant soil was considered as coarse-grained soil.

Allowable Measured Total Capacities

To investigate the predictive capability of the TCP design charts, the design outcome from

these charts (i.e. allowable predicted total capacity) must be compared to allowable

measured total capacity. For this purpose, the ultimate capacity is first determined from the

load-settlement curves and then divided by an appropriate factor of safety following the

Allowable Stress Design (ASD) methodology. According to TxDOT (2012) Geotechnical


68

Manual, TCP design charts for materials with less than 100 blows per 300-mm (12-in) of

penetration (i.e. soils) incorporate a FOS of 2.0.

Several published methods exist for the determination of the ultimate load capacity (Pult.)

from the load-settlement curve. These include Davisson (1975), 5% relative settlement

criterion (Reese and Wright 1977, Reese and O’Neill, 1988), and 10% relative settlement

(Bruce, 1986).

TxDOT Geotechnical Manual (TxDOT 2012) does not provide any recommendation with

regard to ultimate capacity criteria. In fact, a review of 12 research reports and design

manuals published by state DOTs, showed that 11 of the agencies specifies Davisson’s

criterion for the determination of ultimate capacity for driven piles. For drilled shafts, only

three states specified 5% relative settlement to be used as the ultimate capacity criterion.

Based on these results, and the magnitude of settlement in the load-settlement curves

compiled for this study, Davisson’s ultimate capacity criterion was selected was selected

for this study following AASHTO (2012).

Davisson’s criteria considers the elastic compression of the deep foundation element. After

constructing the elastic line, an offset line parallel to the elastic line is plotted. The ultimate

capacity is the load corresponding to the point of intersection of the offset line and the load-

settlement curve. This process is shown in Figure 3.4 (Case No. 23 of drilled shafts).


69

In engineering design, the allowable load on a structural component may be determined

based on strength-based criteria or serviceability-based (i.e. settlement-based) criteria.

Within the context of deep foundations, the strength-based allowable load capacity can be

defined as ultimate capacity divided by an appropriate factor of safety. Since TxDOT TCP

design charts for soil materials are intended to provide a factor of safety of 2.0, it is

appropriate to use a FS=2.0 in the present analysis. Accordingly, the first set of data was

determined for the allowable measured total capacity, 𝑃𝑎𝑀 by dividing the Davisson’s

ultimate capacity obtained from the load-settlement curves by a FS of 2.0.

The second and third sets of 𝑃𝑎𝑀were serviceability-based, and consisted of loads

corresponding to vertical displacements determined from load-settlement curves. In

geotechnical engineering, deformations presented at the head of the deep foundation are

translated to vertical settlements noted in the superstructure. When settlements are larger

than an established tolerable settlement (tol), then it is considered that the foundation

element has reached a serviceability limit condition, and the performance of the

superstructure is not satisfactory.

The extent of the tolerable displacement may be specified by regulatory agencies such as

AASHTO (2014), it can be determined based on behaviors of similar structures (Roberts

et al. 2010), or it can be calculated based on predictive models. The National Cooperative

Highway Research Program (NCHRP) Report No. 343 (Barker et al. 1991) presents

guidelines and specifications describing allowable total and differential settlement for


70

transportation structures. According to AASHTO (2014), the tolerable differential

settlement for highway bridges can be evaluated in terms of angular distortions ().

Further, for continuous span and single span bridges, an angular distortion of 0.004 and

0.008 are recommended, respectively. However it is also discussed that these large values

“concerns structural designers, who often limit tolerable movement to one-half to one-

quarter, or one less order of magnitude i.e., 0.004 to 0.0004”, (SHRP2-solutions, 2016).

Zhang et al. (2005) presented wok completed for 50 concrete bridge foundations with

angular distortions ranging from 0.001 to 0.08, where 61% of the tolerable cases (i.e.

acceptable settlement) presented angular distortions smaller than or equal to 0.002.

In an effort to establish tolerable settlements for this study, a detailed review of available

technical documents and construction drawings associated with the compiled dataset was

completed. From this review, the average bridge span length corresponding to the projects

associated with the load test cases was determined. A total of 77 multi-span bridges

supported by driven piles with an average span length of 18-m (60-ft), and 28 multi-span

bridges supported by drilled shafts with an average span of 21-m (70-ft) were identified.

Using the angular distortion value of 0.004 recommended by AASHTO (2014), tolerable

settlements of 72-mm (2.83-in) for driven piles and 84-mm (3.31-in) for drilled shafts were

obtained. These values are in accordance with other published tolerable settlement value

of 51-mm (2-in) reported by Coduto et al. (2011).


71

Although the tolerable settlements determined based on angular distortion recommended

by AASHTO (2014) ranged from 72-mm (2.83-in) to 84-mm (3.31-in), load settlement

curves compiled for this study did not reach those values of settlement. Based on the dataset

of this study, the vertical displacement values were selected based on the reasonable

assumption that the allowable vertical displacement for foundations supporting bridge

structures will not exceed 6.35-mm (0.25-in) to 12.7-mm (0.5-in). The final dataset

including all four 𝑃𝑎𝑀, and 𝑃𝑎

𝑃 is presented in Table 3.1 for driven piles and Table 3.2 for

drilled shafts.

Figure 3.5 illustrates the load-settlement curve corresponding to the Case No. 23 of drilled

shafts showing Davisson’s line, and loads corresponding to 6.35-mm (0.25-in) and 12.7-

mm (0.5-in) of settlement. The intercept of Davisson’s line with the load-settlement curve

yielded the Davisson’s ultimate capacity value of 6956-kN (156-kips). This value was

further used to determine 𝑃𝑎𝑀 by dividing the value by 2.0 and 3.0 resulting in 3478-kN

(782-kips) and 2320-kN (521-kips), respectively.

To determine 𝑃𝑎𝑀 based on serviceability-based criteria, horizontal lines corresponding to

settlements of 6.35-mm (0.25-in) and 12.7-mm (0.5-in) were plotted from the vertical axis,

and the intercept of these lines with the load-settlement curve yielded the corresponding

loads of 5518-kN (1240-kips) and 6230-kN (1400-kips), respectively.


72

The process described above for both Davisson’s ultimate capacity criterion and

serviceability criteria, were implemented for all 60 load test projects compiled for the study

presented in this paper.

Figure 3.4. Case No. 23, Davisson’s ultimate Capacity Criterion and Serviceability

Criteria

DATA ANALYSIS

The predictive validity of TCP foundation design charts was evaluated in three separate

analyses: (1) qualitative evaluation of the allowable total capacity of deep foundations

determined from TCP foundation design charts (2) prediction of allowable total capacity

for driven piles, and (3) prediction of allowable total capacity for drilled shafts. A fourth

analysis was attempted where the shaft and base capacity were analyzed separately.

0

10

20

30

40

50

60

70

0 1000 2000 3000 4000 5000 6000 7000 8000

Se

tlle

me

nt -

(

mm

)

Axial load (kN)

(6956, 49)

(6230, 12.7)

(5518, 6.35)

Serviceability Criterion Line, = 12.7-mm

Serviceability Criterion Line, = 6.35-mm


73

The first comparison consisted of preparing scatterplots of allowable total capacities for

driven piles and drilled shafts to observe the overall prediction trends of the TCP

foundation design model. For all four sets of 𝑃𝑎𝑀, individual scatterplots were generated to

observe the behavior of the data compared to line of equality (i.e. 45° line).

The second and third analyses consisted of evaluating the relationship between the three

expressions of 𝑃𝑎𝑀 obtained from full scale load tests and 𝑃𝑎

𝑃 determined from TCP design

charts using linear regression analysis for driven piles and drilled shafts independently.

These analyses were performed using four subsets of data:

Strength-based Models:

Model-1: Allowable load corresponding to Davisson’s ultimate capacity criteria

divided by a FS of 2.0, 𝑃𝑎−𝐷/2𝑀

Serviceability-based (Performance) Models:

Model-2: Allowable load corresponding to loads at 6.35-mm (0.25-in) of settlement

determined from load-settlement curves, 𝑃𝑎−𝛿=6.35

𝑀

Model-3: Allowable load corresponding to loads at 12.7-mm (0.5-in) of settlement

also determined from load-settlement curves. 𝑃𝑎−𝛿=12.35

𝑀

For the second and third analyses, regression models were fitted to the data and compared

among each other, and to the line of equality. All statistical analyses were performed using

Minitab 17.0 (2016).


74

Table 3.1. Compiled Dataset for Driven Piles in Soils

Case

No.

Pile Dimensions

Predominant

Source of Total

Capacity (%)

Soil2

𝑃𝑎𝑃

𝑃𝑢𝑙𝑡3 𝑃𝑎−𝐷/2

𝑀 𝑃𝑎−𝛿=6.35

𝑀 𝑃𝑎−𝛿=12.35

𝑀

Width

(mm)

Length1

(m) Shaft Base (kN) (kN) (kN) (kN) (kN)

1 360 9 96% 4% Fine 431 497 248 490 543

2 360 8 97% 3% Coarse 385 1,654 827 1,513 1,682

3 360 7 91% 9% Fine 437 1,161 581 1,068 1,246

4 360 11 96% 4% Fine 808 866 433 823 894

5 510 21 97% 3% Fine 748 2,901 1,450 1,958 2,670

6 380 11 92% 8% Fine 622 1,574 787 1,157 1,647

7 380 10 94% 6% Coarse 677 1,365 682 1,090 1,424

8 410 15 98% 2% Coarse 552 2,297 1,148 1,308 1,709

9 410 13 99% 1% Coarse 984 1,044 522 935 1,090

10 410 13 96% 4% Coarse 523 1,291 646 1,113 1,308

11 410 14 97% 3% Fine 1,241 1,652 826 1,291 1,669

12 410 5 93% 7% Fine 313 1,047 524 979 1,135

13 410 5 93% 7% Fine 236 1,598 799 1,469 1,691

14 410 14 99% 1% Coarse 910 1,304 652 1,068 1,335

15 410 9 98% 2% Fine 669 1,049 525 957 1,090

16 410 9 96% 4% Fine 441 1,344 672 1,224 1,371

17 410 12 96% 4% Fine 466 579 289 445 734

18 380 9 95% 5% Fine 404 1,255 627 1,046 1,357

19 510 23 98% 2% Fine 1,345 1,117 558 1,024 1,126

20 360 8 94% 6% Fine 1,647 1,570 785 1,580 1,780

21 460 10 93% 7% Fine 526 1,580 790 1,139 1,647

22 460 25 89% 11% Fine 642 1,222 611 979 1,224

23 460 14 97% 3% Fine 877 2,406 1,203 1,037 1,059

24 360 8 96% 4% Fine 530 1,122 561 1,068 1,148

25 460 13 95% 5% Coarse 824 1,722 861 1,335 1,736

26 460 13 95% 5% Coarse 824 1,712 856 1,558 1,736

27 460 12 97% 3% Coarse 763 1,519 760 1,291 1,580

28 460 13 95% 5% Coarse 824 2,668 1,334 1,780 2,603

29 460 17 95% 5% Coarse 1,340 3,132 1,566 2,092 3,004

30 510 22 97% 3% Coarse 2,203 3,031 1,515 2,626 3,026

31 510 22 96% 4% Fine 1,568 2,173 1,086 2,181 2,474

32 360 13 99% 1% Coarse 687 1,096 548 979 1,135

33 360 24 98% 2% Fine 1,837 3,014 1,507 1,424 2,359

Note: 1. Embedment Length, 2. Predominant Soil Type 3. Davisson’s Ultimate Capacity Criterion


75

Table 3.2. Compiled Dataset for Drilled Shafts in Soils

Case

No.

Drilled Shaft

Dimensions

Predominant

Source of Total

Capacity (%) Soil2

𝑃𝑎𝑃 𝑃𝑢𝑙𝑡

3 𝑃𝑎−𝐷/2

𝑀 𝑃𝑎−𝛿=6.35

𝑀 𝑃𝑎−𝛿=12.35

𝑀

Diameter

(cm)

Length1

(m)

Shaft Base

(kN) (kN) (kN)

(kN)

(kN)

1 910 8 70% 30% Fine 2,085 7,958 3,979 3,916 6,008

2 1220 16 76% 24% Coarse 5,461 9,531 4,766 6,853 8,277

3 1220 16 76% 24% Coarse 5,461 9,126 4,563 5,830 7,209

4 1220 16 76% 24% Coarse 5,461 10,633 5,316 6,675 8,240

5 910 18 89% 11% Fine 3,700 7,521 3,761 5,518 6,542

6 610 6 82% 18% Fine 1,009 2,444 1,222 1,780 2,581

7 460 8 84% 16% Fine 402 539 269 534 547

8 460 7 92% 8% Fine 275 491 246 467 512

9 760 7 88% 12% Fine 545 1,157 578 1,024 1,121

10 760 7 88% 12% Fine 545 838 419 757 810

11 760 14 91% 9% Fine 1,114 2,702 1,351 2,314 2,537

12 760 18 90% 10% Coarse 2,280 4,372 2,186 3,783 4,139

13 760 23 96% 4% Fine 1,783 6,058 3,029 5,474 5,919

14 760 14 82% 18% Coarse 3,502 5,344 2,672 3,516 4,406

15 1830 20 76% 24% Coarse 4,423 12,120 6,060 5,474 7,343

16 1220 26 93% 7% Coarse 3,650 10,939 5,470 5,340 7,699

17 1220 27 91% 9% Coarse 3,773 12,019 6,010 5,385 7,521

18 1680 19 84% 16% Fine 3,773 4,931 2,466 3,560 4,228

19 1680 12 85% 15% Fine 2,105 5,120 2,560 3,694 4,094

20 1830 25 87% 13% Fine 4,705 8,405 4,202 6,453 7,120

21 1220 30 92% 8% Fine 4,018 11,176 5,588 5,830 8,277

22 1220 12 23% 77% Fine 1,883 3,756 1,878 2,982 3,427

23 1680 14 70% 30% Coarse 3,000 6,956 3,478 5,518 6,230

24 790 14 89% 11% Coarse 5,067 6,491 3,245 4,450 5,518

25 790 14 89% 11% Coarse 5,108 7,779 3,890 4,984 6,408

26 1220 23 85% 15% Coarse 5,137 8,256 4,128 2,670 4,183

27 810 17 91% 9% Coarse 2,764 6,598 3,299 1,958 3,382

Note: 1. Embedment Length, 2. Predominant Soil Type 3. Davisson’s Ultimate Capacity Criterion

Qualitative Evaluation

To facilitate qualitative evaluation of the predictive capabilities of the TCP foundation

design charts, scatterplots were generated for all the data and compared to line of equality,


76

Figure 3.5. To create these plots, the ultimate 𝑃𝑎𝑀

determined for each model for both driven

piles and drilled shafts were used to generate an undifferentiated scatterplot.

Figures 5(a) correspond to the strength-based model where the ultimate deep foundation

capacity was interpreted using Davisson’s criterion and the allowable measured total

capacities were determined by dividing the ultimate capacity value by a FS of 2.0. Data

points above the equal prediction line (i.e., 1:1 Line) are conservative (𝑃𝑎𝑀 > 𝑃𝑎

𝑃) and data

points below the equal prediction line are unconservative (𝑃𝑎𝑀 < 𝑃𝑎

𝑃). However, this does

not mean that the performance of the foundation is in jeopardy. It is important to note that

the ultimate capacity criterion is not necessarily a physical event but rather a defined

computational condition (Foye et al. 2009). For example, in the case of a full scale load

test completed for a deep foundation, the plunging load can be considered as the physical

event of ultimate limit state but it is different from the load determined from accepted

ultimate criteria (e.g. Davisson’s, 5% or 10% relative settlement) established as the

computational condition. Therefore, unconservative design does not necessarily lead to

poor performance. This is better illustrated with the serviceability-based models in the

following paragraph.

From Figure 3.5(a) it is noted that, even though significant scatter exists, the data points

are approximately evenly split above and below the 1:1 line when a FS of 2.0 is used. In

this plot 34 data points representing slightly more than 50% of the dataset plotted below

the 1:1 line. Based on these findings it is reasonable to conclude that overall, the allowable


77

total capacity predictions made by using the TCP design procedure agree well with

measured values. In other words, the probability that the actual capacity will be higher

than that predicted by the design charts is about the same as the probability that it will be

lower. Accordingly, the design procedure can be expected to yield a bias value close to

unity. However, because of the significant scatter observed in the data, one cannot

conclude that the predictive capabilities of the TCP design procedure as strong. At the

same time, it is important to point out that it would not be reasonable to attribute all of the

scatter observed to inadequacies of the TCP test method. Other factors such as differences

in load testing procedures used (top down versus O-cell, loading rates, time lag between

installation of test foundation and load testing), measurement errors during testing,

differences in foundation construction/installation methods (pile driving, use of slurry

versus dry hole installation used for drilled shafts) would have also contributed to such

scatter.


78

(a) Scatterplot for Model-1: 𝑃𝑎−𝐷/2

𝑀

(b) Scatterplot for Model-2: 𝑃𝑎−𝛿=6.35

𝑀 (c) Scatterplot for Model-3: 𝑃𝑎−𝛿=12.35

𝑀

Figure 3.5. The relationship between 𝑃𝑎𝑀 and 𝑃𝑎

𝑃 compared to an equal prediction line

0

1000

2000

3000

4000

5000

6000

7000

0 1000 2000 3000 4000 5000 6000 7000

Allo

wable

Pre

dic

ted T

ota

l C

apacity (

kN

)

Allowable Predicted Total Capacity (kN)

Model-1

Driven Piles

Drilled Shafts

Conservative

Unconservative

0

1000

2000

3000

4000

5000

6000

7000

0 1000 2000 3000 4000 5000 6000 7000

Allo

wa

ble

Pre

dic

ted

To

tal C

ap

acity (

kN

)


Model-2

Driven Piles

Drilled Shafts

Conservative

Unconservative

0

1000

2000

3000

4000

5000

6000

7000

0 1000 2000 3000 4000 5000 6000 7000

Allo

wa

ble

Pre

dic

ted

To

tal C

ap

acity (k

N)


Model-3

Driven Piles

Drilled Shafts

Conservative

Unconservative


79

For the serviceability-based models, it is observed that for a tolerable settlement of 6.35-

mm (0.25-in), Figure 3.5(b), approximately 7 data points are below the line of equality.

For a tolerable settlement level of 12.7-mm (0.5-in), Figure 3.5(c), only 2 points are below

the equal prediction line (one point for driven piles and one corresponding to drilled shafts).

From these observations, it is apparent that different allowable capacity criteria leads to

different conclusions regarding the validity of the TxDOT TCP method for deep foundation

design. The allowable predicted total capacity determined from the TCP foundation design

charts could be considered to be reasonable based on Model-1, conservative based on

Model-2, and over-conservative for Model-3. In general, this suggests that serviceability-

based interpretation of the TCP foundation design charts (which is closely associated with

observed structure performance) will be conservative as shown by both Models 2 and 3.

As previously described, Coduto et al. (2011) reported a tolerable settlement value of 51-

mm (2-in) for the serviceability conditions. Similarly, angular distortions specified by

AASHTO (2014) yielded tolerable settlements of 72-mm (2.83-in) for driven piles and 84-

mm (3.31-in) for drilled shafts. Collectively, for higher magnitude of tolerable settlements

the TCP foundation design charts will predict the total capacity of the foundation in a more

conservative manner.


80

Statistical Analysis

This section presents the analyses completed for driven piles and drilled shafts where the

predictive validity of the TCP foundation design charts are evaluated quantitatively. For

each subset of data, the relationship between 𝑃𝑎𝑀 and 𝑃𝑎

𝑃 was identified in terms of a

regression model and these models were compared to the line of equality (i.e. 45° line).

Linear regression analyses were performed with the 𝑃𝑎𝑀 considered as the response variable

and the 𝑃𝑎𝑃 as the predictor. Prior to each regression analysis, the data were analyzed to

ensure that all assumptions for a simple linear regression analysis were met (i.e. normality,

homoscedasticity, etc.). Linear models were determined following the form of Equation

(3.2) where o and 1 are the intercept and the slope of the regression model, respectively.

𝑃𝑎−𝑚𝑒𝑎𝑠 = 𝛽0 + 𝛽1𝑃𝑎−𝑝𝑟𝑒𝑑 (3.2)

For each model, the existence of a relationship between 𝑃𝑎𝑀

and 𝑃𝑎𝑃 was confirmed by p-

values and the strength of the relationship was assessed by the coefficient of determination

(R2).

The regression models were compared to the equal prediction line to observe whether the

equal prediction line was completely within the confidence boundaries of the regression

model. The confidence intervals (CI) for the regression model were considered as the set

of all reasonable linear models suggested by the data.


81

RESULTS

Statistical analyses were performed to observe whether the predominant soil type had a

statistically significant impact on the regression models developed for undifferentiated data

by type of soil. For the compiled dataset, the predominant soil was categorized as fine and

coarse grained soils and it was assigned based on percentage of contribution of the soil

type to the predicted allowable total capacity of the foundation.

The compiled dataset showed that the predominant soil for 13 driven piles and 13 drilled

shafts was categorized as coarse grained soil, and the remaining 20 driven piles and 14

drilled shafts had fine grained soils as the predominant soil. An Analysis of Covariance

(ANCOVA) was performed where 𝑃𝑎𝑀 was identified as the response variable, 𝑃𝑎

𝑃 as

predictor (covariate), and the type of soil as the categorical predictor (factor). Results from

these analyses showed that with a p-value greater than 0.05 (i.e. 0.681 for driven piles and

0.223 for drilled shafts) a regression model fitted for fine grained soil did not have a

statistically-significant difference from the model for coarse grained soil. Therefore all

analyses were completed for undifferentiated soil data, but differentiated by foundation

type.


82

Regression Models

Linear regression analyses were completed to further explore the relationship between 𝑃𝑎𝑀

and 𝑃𝑎𝑃 for each type of foundation. Regression models were fitted to the corresponding

data for driven piles, Figure 3.6, and drilled shafts, Figure 3.7, following the form of

Equation (3.2). In these figures, the regression models are shown as solid lines, the 95%

confidence interval (CI) with dashed lines, and the equal prediction line with dash-dot lines.

Figure 3.6. Scatterplot and fitted model, Driven Piles

For driven piles, Figure 3.6, the model shifts from under-predicting to over-predicting at

approximately 800-kN (180-kips). Further, it is noted that only a small portion of the equal


83

prediction line plots within 95% CI boundaries indicating that the line of equality is not a

plausible model that could describe the relationship between 𝑃𝑎𝑀 and 𝑃𝑎

𝑃.

The regression model includes and intercept greater than zero. This is not an indication that

for a 𝑃𝑎𝑃 = 0 the 𝑃𝑎

𝑀 would have a value. This suggests that the regression line is

extrapolating through the region where data did not exist (region comprehended

between 𝑃𝑎𝑃 = 250-kN and 𝑃𝑎

𝑀= 750-kN). This indicates that the true relationship

between 𝑃𝑎𝑀 and 𝑃𝑎

𝑃 is actually nonlinear in that region, but appears linear where data points

exist.

In the case of drilled shafts, linear regression models were fitted to the dataset and it was

observed that the fitted model performs much better, Figure 3.7. The model still slightly

over-predicts for 𝑃𝑎𝑃 greater than 4000-kN (900-kips), but the equal prediction line plots

within the 95% CI boundaries for the regression model indicating an acceptable

comparison between the regression model and the equal prediction line.


84

Figure 3.7. Scatterplot and fitted model, Drilled Shafts

Side-by-Side Comparison of all Measured Capacity Models

As previously described, three models were considered for the performance evaluation of

the TCP foundation design charts. Models 1 was strength-based and Models 2 and 3 were

serviceability-based. For all these models, the relationship between 𝑃𝑎𝑀 and 𝑃𝑎

𝑃 was

investigated, and regression models were developed for driven piles, Figure 3.8, and drilled

shafts, Figure 3.9. These regression lines were plotted side-by-side for comparison

purposes. The existence and the strength of the relationship between 𝑃𝑎𝑀 and 𝑃𝑎

𝑃 were

evaluated based on p-values and R2 determined for driven piles, Table 3.3, and drilled

shafts, Table 3.4.


85

Table 3.3. Regression line equation with p-value and R2, Driven Piles

Model Regression Equation p-value R2 (%)

1 𝑃𝑎−𝐷/2𝑀 = 455+0.431𝑃𝑎

𝑃 0.001 33

2 𝑃𝑎−𝛿=6.35

𝑀 = 777+0.601𝑃𝑎

𝑃 <0.0001 37

3 𝑃𝑎−𝛿=12.35

𝑀 = 935+0.784 𝑃𝑎𝑃 <0.0001 35

Table 3.4. Regression line equation with p-value and R2, Drilled Shafts

Model Regression Equation p-value R2 (%)

1 𝑃𝑎−𝐷/2𝑀

= 623+0.839𝑃𝑎𝑃

<0.0001 65

2 𝑃𝑎−𝛿=6.35

𝑀 = 1205+0.889𝑃𝑎

𝑃 <0.0001 61

3 𝑃𝑎−𝛿=12.35

𝑀 = 1360+0.1.168𝑃𝑎

𝑃 <0.0001 66

From the values shown in Table 3.3 and 3.4, it is noted that with p-values ranging from

0.001 to smaller than 0.0001, a statistically significant relationship exists between 𝑃𝑎𝑀 and

𝑃𝑎𝑃 for both driven piles and drilled shafts. With regard to the strength of the relationship,

in the case of driven piles, R2 ranges between 33% and 37% and this can be categorized as

“weak”, but not atypical for geotechnical engineering data. For drilled shafts, R2 values

ranged between 61% and 66%, which can be considered as a moderate relationship.


86

Figure 3.8. Side-by-Side Comparison of all Models analyzed, Driven Piles

For driven piles the regression lines corresponding to Models 2 and 2 represent a

conservative prediction of the total capacity. In the case of Models 1, except for a cluster

of low magnitude loads, the regression model significantly over-predicts for the allowable

total capacity which is unconservative. It is important to keep in mind that Models 1 is

strength-based models; whereas, Models 2 and 3 are serviceability-based models. Further,

from the comparison it is observed that if the tolerable displacement for a driven pile was

between 6.35-mm (0.25-in) and 12.7-mm (0.5-in), then based on Models 2 and 3, the TCP

0

500

1000

1500

2000

2500

3000

3500

0 1000 2000 3000

Allo

wab

le M

easu

red

To

tal C

apac

ity

(kN

)

Allowable Predicted Capacity (kN)

DRIVEN PILES

Model-1 LineModel-2 LineModel-3 Line

CONSERVATIVE

UNCONSERVATIVE


87

foundation design charts are predicting the allowable total capacity in a conservative

manner which is a basic expectation associated with engineering design. Note that for both

Models 2 and 3 the regression lines are above the equal prediction line with exception of

loads greater than 2000-kN (450-kips) for Model 2.

In contrast to the serviceability-based models, the strength-based models are not fully

indicating a conservative prediction from the TCP foundation design charts. The major

portion of the regression model developed for FS =2.0 is located below the line of equality,

indicating an over-prediction of allowable total capacity for driven piles for loads greater

than 800-kN (180-kips). It is important to emphasize that this over-prediction is from the

design perspective and does not indicate a poor performance of the foundation or the

superstructure.

For drilled shafts, serviceability-based Models 3 and 4 are located completely above the

equal prediction line, indicating a conservative prediction of the allowable total capacity

based on the TCP foundation design charts. For the strength-based models (i.e. Model-1

and 2) it is noted that for Model-1 (FS = 2.0) reasonably compares to the equal prediction

line with slight over-prediction for loads greater than 4000-kN (900-kips). In contrast

Model-2 (FS=3.0) is significantly over-predicting for loads larger than 1000-kN (225-

kips). Further, as with driven piles, the serviceability-based models are predicting the

allowable total capacity of drilled shafts in a more conservative manner.


88

Figure 3.9. Side-by-Side Comparison of all Models analyzed, Drilled Shafts

Evaluation of TCP charts based on Shaft and Base resistance

One of the characteristics of the TCP foundation design method is that the shaft and base

resistances of a foundation element are evaluated separately, and these are summed to

determine the allowable total capacity. Of course, shaft and base capacities can be

measured separately from instrumented load tests; whereas, from non-instrumented load

0

1,000

2,000

3,000

4,000

5,000

6,000

7,000

8,000

9,000

0 1,000 2,000 3,000 4,000 5,000 6,000 7,000 8,000 9,000

Allo

wab

le M

easu

red

To

tal C

apac

ity

(kN

)

Allowable Predicted Capacity (kN)

DRILLED SHAFTS

Model-1 LineModel-2 LineModel-3 Line1:1 Line

CONSERVATIVE

UNCONSERVATIVE


89

tests only total capacity can be obtained. The dataset compiled for this study included 14

instrumented load tests for drilled shafts. Measured shaft and base resistances were

determined from these tests and were compared to the base and shaft resistances obtained

using the TCP foundation design charts. The data were tested and statistical analysis

showed that no statistical-significant relationship existed between measured and predicted

resistances. However, this is not indicating that the measured and predicted resistances

cannot be evaluated. Rather, it is suggesting that the workable dataset is too limited to be

statistically evaluated.


This paper presents an evaluation of the predictive validity of the TxDOT’s foundation

design method using the Texas Cone Penetration (TCP) test. The evaluation was completed

by comparing various expressions of allowable measured total capacity (𝑃𝑎𝑀) to the

allowable predicted total capacity (𝑃𝑎𝑃). These comparisons were first completed in a

qualitative manner, and then quantitatively by establishing statistical regression models for

each type of measured capacities. An effort to evaluate the TCP foundation design charts

based on measured shaft and base resistances, was attempted but not successful due to

limitations of the dataset.

Results from the qualitative evaluation showed that the TCP foundation design method

performs reasonably well for serviceability-based models where out of 60 data points only

10 and 2 points locate below the equal prediction line for a tolerable settlement of 6.35-


90

mm (0.25-in) and 12.7-mm (0.5-in), respectively. This observation showed that as the

tolerable displacement is increased, the data points shift further above the equal prediction

line, indicating higher satisfactory performance of the foundation or the superstructure. For

the strength-based analysis, 34 and 53 points locate below the equal prediction line based

on FS of 2.0 and 3.0, respectively. In both cases, more than half of the data points locate

below the equal prediction line, indicating moderately to severely unconservative design.

Results from quantitative analyses show that for driven piles, R2 values ranging from 33%

to 37%, and p-values varying from 0.001 to values smaller than 0.0001 were determined.

In the case of drilled shafts, R2 values ranging between 61% and 66% and p-values smaller

than 0.0001 were obtained. The comparison of all three models showed that in the case of

driven piles, the serviceability-based regression models locate above the equal prediction

line, and the strength-based models locate below. In the case of drilled shafts, the same

trend is observed, with the difference that the regression line for Model 1 (strength-based

model) is comparable to the equal prediction line.

Collectively, it is reasonably considered that the predictive validity of the TCP foundation

design charts could be categorized as moderately unconservative for Model-1, conservative

for Model-2, and slightly over-conservative for Model-3. Further, in the case of

serviceability-based models when the level of tolerable settlement is increased (as might

be true in the case of serviceability analysis of transportation infrastructure), the


91

interpretation of the TCP foundation design charts becomes even more conservative,

indicating higher performance of the foundation or the superstructure.

TxDOT geotechnical and bridge engineers have designed thousands of bridge foundations

throughout Texas using the TCP test and its associated design methodology. The TCP

design method is a strength-based predictive model and as to date, all research work have

been completed focusing on the strength approach to determine allowable capacities. It is

important to keep in mind that the TCP foundation design charts were developed based on

correlations between TCP blowcounts and shear strength determined from laboratory

testing. The correlation lines of the TCP design charts predict allowable capacity from the

perspective of a strength-based model. This prediction has been strongly discussed among

researchers and the predictive validity of the TCP design charts has been questioned.

However, observing the performance of thousands of bridges designed in accordance to

the TCP design methodology, it is noted that none of these superstructures had shown

unsatisfactory performance.

Based on results of this study, the well performance of the foundations and transportation

infrastructures designed based on the TCP test can be explained from serviceability-based

models and reasonably make the statement that the TCP foundation design charts

moderately over-predict design capacities under strength-based approach, but

conservatively predicts allowable total capacities for serviceability-based approach.


92

The predictive validity of the TCP foundation design charts in soils, is not within high

confidence that the allowable measured total capacity is equal or greater than the allowable

predicted total capacity, which does not necessarily indicates a poor performance of the

foundation or the superstructure. On the other hand, the predictive validity of the TCP

foundation design charts in soils, is within high confidence that the settlement under the

allowable predicted total capacity is smaller than the tolerable settlement indicating a

satisfactory performance of the foundation or superstructure.


Note 2: Moghaddam R.B., Seo H., Lawson, W.D., Surles, J.G., Jayawickrama, P.W. (2016), “Hammer

Efficiency and Correction Factors for the TxDOT Texas Cone Penetration Test”, ASCE-ASME Journal of

Risk and Uncertainty in Engineering System, Part A: Civil Engineering, Submitted June, 20th, 2016

93

CHAPTER IV

RESISTANCE FACTORS AT SERVICEABILITY LIMIT STATE FOR LRFD OF DEEP

FOUNDATIONS USING THE TEXAS CONE PENETRATION TEST 2

ABSTRACT

This study identifies resistance factors for the serviceability limit state (SLS) condition

(SLS) used in the load and resistance factor design (LRFD) of deep foundations using the

results from the Texas Cone Penetration (TCP) test. The performance function was

established based on load corresponding to tolerable displacement (Qtol) and design load

(Qd). A dataset of published full-scale load tests including projects from Texas, Missouri,

Arkansas, Louisiana, and New Mexico was compiled and consisted of 60 load test cases

comprising 33 driven piles and 27 drilled shafts. Resistance factors for SLS conditions

were obtained for tolerable displacements using both the Monte Carlo simulation (MCS)

and the First Order Second Moment (FOSM) calibration approaches. Resistance factors at

SLS conditions were developed ranging from 0.21 to 0.77 using FOSM method and 0.23

to 0.67 using the MCS for driven piles. In the case of drilled shafts, SLS resistance factors

ranged from 0.17 to 0.77 following the FOSM method and 0.18 to 0.86 based on MCS.


94

INTRODUCTION

This paper presents resistance factors for total capacity of deep foundations at the

serviceability limit state condition (SLS) using penetration resistances obtained from the

Texas Cone Penetration (TCP) test. The TCP test method is introduced and the design

procedure associated with the test is discussed. A brief review of the history and

development of the TCP field test, design procedure, and design charts is presented.

Furthermore, the basics of LRFD method are described and the application of this method

to deep foundations at SLS is explored. In addition, the effort associated with the field data

collection followed by statistical analyses and results are discussed in detail. Finally,

resistance factors for total capacity of driven piles and drilled shafts at SLS are developed

and introduced.

Drilled shaft and driven pile foundations are often used for bridges and transportation

structures. Over the past 60 years, the Texas Department of Transportation (TxDOT) has

successfully designed thousands of bridge foundations in every district in Texas using the

TCP test and its associated design method. Aside from the extensive use in Texas, the TCP

test and its design charts have been used in parts of Oklahoma. Due to difficulties

associated with drilling and sampling in shale material using conventional in-situ tests such

as the standard penetration test (SPT) or cone penetration test (CPT), results from the TCP

method have also been used in studies completed in the state of Missouri where the design

of drilled shafts in shales was evaluated.


95

The load and resistance factor design (LRFD) approach has gained wider application in

recent years partly because projects funded by the Federal Highway Administration

(FHWA) are mandated to be designed based on the LRFD approach (FHWA 2011). This

regulation has created grounds for all state DOTs to transition from allowable stress design

(ASD) to the LRFD method.

Ultimate limit state (ULS) and SLS are two limit states often considered in geotechnical

engineering designs that follow the LRFD methodology. Resistance factors have been

developed for different design methods used by state DOTs where the factors are usually

calibrated against ULS criteria. However, most structures and in particular, transportation

structures such as bridges, suspend service operations as soon as one of the structure’s

members experiences displacements beyond a tolerable displacement. In contrast to ULS,

not many published studies discuss the determination of resistance factors at the SLS.

Furthermore, no published work has explored resistance factors at SLS condition for TCP

design method. Results from analyses completed in this paper contribute to the

geotechnical engineering community where the TCP test and its associated design

procedure are used as primary method for the design of deep foundations.

It is important to note that the main thrust of this paper is the development of resistance

factors at SLS conditions for the LRFD of deep foundations using the TCP design

methodology. For purposes of clarity and consistency, the words “settlement” and

“displacement” are used in the same context referring to vertical displacements, and all

analyses and results are presented for vertical displacements only.


96

TEXAS CONE PENETRATION (TCP) TEST

Description of TCP test




1999). The TCP test uses a 77.0-kg (170-lb) hammer with 60-cm (24-in) drop to force a

7.6-cm (3-in) diameter steel cone into the soil or rock formation. Details of cone geometry

and application are presented in Figure 4.1. In current practice, the penetration is to be

achieved in three separate increments. The first increment is completed to ensure proper

seating, which consists of driving the cone 12 blows or approximately 15-cm (6-in),

whichever happens first. The TCP blowcount is then determined as the sum of the number

of blows required to achieve second and third 15-cm (6-in) increments of cone penetration.

The total blowcount or NTCP corresponding to 30-cm (12-in) penetration is used to obtain

design parameters. In very hard materials such as rock and intermediate geomaterials

(IGM), after the proper seating process is completed, the cone is driven 100 blows and the

penetration value for the first and second 50 blows are recorded.


97

(a) Conical driving point (From TxDOT 1999) (b) Field application

Figure 4.1. TCP Test conical driving point

Application and use of TCP blowcount data for foundation design

The TCP test design method was introduced in the 1956 edition of the Foundation

Exploration and Design Manual (THD, 1956). The design manual provided a series of

correlations based on relationships established between TCP test blowcount data and

laboratory-measured shear strength using the triaxial test procedure (THD, 1956). The

charts published in 1956 were refined in 1972, 2000, and 2012, and two sets of design

charts exist today. The first set was developed for the prediction of unit shaft resistance

(i.e. skin friction) and unit base resistance (i.e. point bearing) for soils with TCP

blowcounts less than 100 blows per 30-cm (1-ft), as shown in Figure 4.2(a) and 4.2(b),

respectively. Figure 4.3 presents design charts for geomaterials with blowcounts greater

than 100 blows per 30-cm (1-ft) (i.e. penetration per 100 blows).


98

(a) (b)

Figure 4.2. TCP Foundation Design charts representing (a) allowable unit shaft

resistance and (b) allowable unit base resistance vs. TCP blows/30-cm(12-in)

(TxDOT, 2012)

(a) (b)

Figure 4.3. TCP Foundation Design charts representing (a) allowable unit shaft

resistance and (b) allowable unit base resistance vs. TCP penetration/100 blows

(TxDOT, 2012)


99

LOAD AND RESISTANCE FACTOR DESIGN (LRFD)

Design approaches and methods

Two principal subjects in geotechnical engineering related to stability and elasticity

problems were outlined by Terzaghi (1947). The stability problem was mainly described

as being directly related to the conditions immediately before collapse or total failure;

whereas, problems associated with elasticity were related to soil deformations due to

external loading or the structure’s own weight (Meyerhof 1995). An engineering design

could reach a state where one or more components of the structure fails to meet the

prescribed functions or design criteria. These conditions can be referred to as the ultimate

states (Becker, 1996). The ultimate state can be further classified as ultimate limit state

(ULS) and the serviceability limit state (SLS).

Ultimate Limit State (ULS)

Conceptually, when an applied stress exceeds the strength of a component of a structure

the ULS condition is reached. The ULS can be defined as the condition when a partial or

full collapse of a structure may occur. Extensive deformations and cracks are precedent to

the ULS condition (Salgado 2008). Most engineering design failures are identified before

the structure reaches the ULS conditions, so it would be reasonable to assume that the

probability of occurrence of the ULS conditions is low in comparison to the SLS (Duncan

et al. 1989).


100

It is important to note that the ULS is not necessarily a physical event but rather a defined

computational condition (Foye et al. 2009). For example, in the case of a full scale load

test completed for a deep foundation, the plunging load can be considered as the physical

event of ULS but it is different from the load determined from accepted ultimate criteria

(e.g. Davisson’s, 5% or 10% relative settlement) established as the computational

condition. For purposes of analyses, researchers have selected the load corresponding to

the ultimate capacity criteria (e.g. Davisson’s, 5% or 10% relative settlement) for the

determination of ULS resistance factors (Paikowsky et al. 2006).

Serviceability Limit State (SLS)

An element of a structure can suffer deformations due to applied loads but not to the extent

of reaching deformation levels presented by the ULS conditions. In this case, a collapse is

not likely to occur but the structure or the element of the structure is not serviceable and

cannot meet its prescribed function. According to Allen et al. (2005), the SLS represents

the condition where the function and service requirements of the structure are affected.

In geotechnical engineering and more specifically in the case of deep foundations,

deformations can be translated to vertical displacement of the foundation element which

causes settlement in the superstructure. When the vertical displacement is larger than an

established tolerable displacement, then it is considered that the foundation system has

reached the SLS condition.


101

The ULS condition and its application to geotechnical engineering design has been

reasonably well established and a significant body of literature including technical papers

and research reports exist. In contrast to ULS, the SLS condition and its associated

resistance factors have not been developed with the same emphasis and details as the ULS

(Vu, 2013).

LRFD of Deep Foundations

The primary focus of the LRFD method is to differentiate uncertainties in loading from

those existing in the resistance following a probability theory to assure a safety margin

(Paikowsky 2004). The general form of a performance function (g) can be described as the

difference between two probabilistic parameters such as resistance (R) and load (Q),

Equation (4.1):

𝑔 = 𝑅 − 𝑄 (4.1)

Each of the parameters constituting the performance function g (i.e. R and Q) has its own

probability distribution. A performance function (g) greater than or equal to zero is an

indication of a satisfactory performance whereas a value less than zero is an indication of

unsatisfactory performance (Allen et al. 2005).

In the LRFD of deep foundations, resistance factors are applied to the calculated nominal

capacity resulting in the nominal factored resistance. The deep foundation is considered


102

functional when the nominal factored resistance is greater than the nominal factored design

load. In current practice, deep foundation total capacity is determined from the summation

of the shaft and base resistances which are determined separately.

In cases where instrumented data are not available the design inequality can be written for

total capacity (Qt) where the resistance factor for the total capacity (φt) is determined,

Equation (4.2).

φ𝑡𝑄𝑡 ≥ Σ(𝛾𝐿𝐿𝑄𝐿𝐿 + 𝛾𝐷𝐿𝑄𝐷𝐿) (4.2)

Uncertainties and probabilities of occurrence associated with the loads applied to a

structure have to be accounted for design purposes and are reflected in a load factor ()

greater than unity (Becker 1996; Paikowsky 2006). Depending on the type of the structure

under design, the American Association of State Highway and Transportation Officials

(AASHTO) and the FHWA provide tables containing load and resistance factors.

According to AASHTO (2014) for the design of foundations, load factors of 1.25 for dead

load and 1.75 for live load are recommended.


103

SERVICEABILITY LIMIT STATE ANALYSIS

SLS Performance function

Following the definition of the SLS, a performance function is introduced following the

form of Equation (4.1). Since SLS approach is based on displacements, the terms for

resistance (R) and load (Q) are substituted by tolerable displacement (tol) and displacement

under the design load (d), respectively. A satisfactory performance under SLS conditions

is achieved when the performance function for the SLS condition (gSLS) is greater than or

equal to zero as shown in Equation (4.3). This means that as long as the tolerable

displacement (tol) is greater than the displacement under design load (d), the design

outcome satisfies the SLS condition.

𝑔𝑆𝐿𝑆 = 𝛿𝑡𝑜𝑙 − 𝛿𝑑 (4.3)

The performance function for the SLS condition shown in Equation (4.3) can alternatively

be represented in terms of loads. When tol and d in Equation (4.3) are replaced by a load

corresponding to tolerable displacement (Qtol) and a design load (Qd), respectively, the

SLS performance function can be rewritten as follows:

𝑔𝑆𝐿𝑆 = 𝑄𝛿𝑡𝑜𝑙 − 𝑄𝑑 (4.4)


104

From results of a full scale load test, the relationship between loads applied to the

foundation and their corresponding displacements can be expressed by the load-settlement

curve. From this curve, Qtol and Qd can be determined. Under conditions presented in

Equation (4.4), a satisfactory performance function is obtained when Qtol is greater than

Qd. Figure 4.4 is a graphical representation of the concept described and expressed by

Equation (4.4). Also shown are the loads and displacements associated with the ULS

condition as defined by Davisson’s criterion, and 5% and 10% relative settlement criterion.

Figure 4.4. Loads corresponding to tolerable displacement (Qtol) and displacement from

design loads (Qd)

Se

ttle

me

nt (

)

Load (Q)

Load Corresponding to the Tolerable Displacement

Load-SettlementCurve

Tolerable

Displacement (tol)

tol

d

Design LoadDisplacement under

design load (d)

QdQtol

5% Relative Settlement ULS Criterion

10% Relative Settlement ULS Criterion

Davisson's ULS Criterion (QULS, Dav.)

(QULS, 5%)

(QULS, 10%)


105

Tolerable Displacement (tol)

An excessive amount of settlement will cause a structure to reach the SLS conditions and

if these settlements are further increased, a collapse or failure corresponding to the ULS

conditions may occur. In order to confirm whether an excessive settlement is present or

not, an indicative parameter is required (Salgado 2008). One of the commonly-used

indicative parameters is the angular distortion () which is defined as the ratio of the

differential settlement () to the span length (L):

𝛼 = 𝛿

𝐿 (4.5)

In a study completed by Skempton and McDonald (1956), a total of 98 buildings consisting

of frame buildings and masonry buildings with load bearing walls located in London,

England, were selected and analyzed. For each building, relevant information regarding

displacement was gathered and a database was compiled based on the type of building. In

addition to the available information, laboratory testing on reinforced concrete frames and

brick walls were carried out under different angular distortion until cracking happened.

Results from the combination of the observed data and laboratory testing suggested an

angular distortion of 0.0033 to reach SLS conditions, 0.006 to experience ULS conditions,

and 0.002 for the prevention of both SLS and ULS conditions (Salgado 2008).


106

Burland and Wroth (1974) followed the initial work completed by Skempton and

McDonald (1956) where the excessive settlement was assessed based on visible cracks on

structural elements and determined that a cracking pattern is not always an appropriate

method to further assess and define an excessive settlement. According to Salgado (2008),

visible cracking on structural elements could be an indicator of presence of settlements but

not an indication of ULS or SLS condition. Instead of the assessment of cracking patterns,

Burland and Wroth (1974) introduced the beam analogy where a beam with known span

length, height, and elastic properties underwent different loading patterns until cracking

started at different locations on the beam. Based on the beam analogy, the ratio of the

displacement at the center of the structure to the length of the structure was recommended

as the angular distortion. Results obtained from the study completed by Burland and Wroth

(1974) were compared to the angular distortion proposed by Skempton and McDonald

(1956) and it was concluded that the angular distortion of 0.002 was suitable to prevent

both ULS and SLS condition.

Zhang et al. (2005) obtained detailed information regarding displacement of structures

located in Hong Kong and China and combined the results with the data published by

Skempton and McDonald (1956) and Grant et al. (1974) for buildings, and Moulton (1985)

for bridge foundations. The final dataset associated with the bridge foundations reported

by Zhang (2005) consisted of 50 concrete bridge foundations with angular distortions

ranging from 0.001 to 0.08. According to Zhang (2005), 41 bridges were considered

tolerable cases where the prescribed function of the structure was not affected by angular


107

distortion, and the remaining nine bridges were reported as intolerable cases where the

serviceability of the bridge was significantly affected. Furthermore, 25 out of 41 bridges

representing 61% of the tolerable cases presented angular distortions smaller than or equal

to 0.002. For intolerable cases, 78% of these had angular distortions greater than or equal

to 0.01.

Based on the definition of angular distortion and the work presented by Zhang (2005) and

others, relative to span length (L), the tolerable displacement of a structure can be predicted

using Equation (4.6).

𝛿𝑡𝑜𝑙

𝐿≤ 0.002 (4.6)

The magnitude or extent of the tolerable displacement may be specified by regulatory

agencies such as AASHTO (2014), it can be determined based on behaviors of similar

buildings and structures (Roberts et al. 2010), or it can be determined based on predictive

models. The National Cooperative Highway Research Program (NCHRP) Report No. 343

(Barker et al. 1991) presents guidelines and specifications describing allowable total and

differential settlement for transportation structures.


108

DEVELOPMENT OF SLS RESISTANCE FACTORS

Equations (4.3) and (4.4) represent two alternative forms of performance functions for the

SLS condition. These equations may be identified as the displacement approach and the

load approach, respectively. Both equations have been used for development of resistance

factors in literature. According to Vu (2013), Equations (4.3) and (4.4) have equal

credibility and the choice of which approach to use is directly associated with data

availability.

Displacement Approach

In the geotechnical design process, the resistance is determined based on geotechnical

information obtained from field and laboratory tests with some level of uncertainties. A

design governed by SLS conditions will also be influenced by uncertainties associated with

the prediction of displacements. Therefore, uncertainties can be expressed by a resistance

factor for the SLS condition (SLS) to further reduce the tolerable displacement and

compare the result to the displacement presented under design load. For satisfactory SLS

design, the factored tolerable displacement (i.e., reduced tolerable displacement) must be

greater than the displacement under design load as shown below:

φ𝑆𝐿𝑆𝛿𝑡𝑜𝑙 ≥ 𝛿𝑑 (4.7)


109

Alternatively, Paikowsky et al. (2006) and Honjo et al. (2005) approached the LRFD

methodology at SLS by factoring the displacement presented under design loads (d). In

this case, the factored displacement (dd), i.e. the increased displacement under design

load, must be smaller than the tolerable displacement as follows:

𝛿𝑡𝑜𝑙 ≥ φ𝑑𝛿𝑑 (4.8)

From Equations (4.7) and (4.8), it is noted that in both cases, factors are applied to

displacements. In the case of Equation (4.7), the tolerable displacement is factored and

intuitively the resistance factor (SLS) is less than unity. On the other hand, the factor (d)

is greater than unity to increase the displacements presented due to applied loads, Equation

(4.8).

Wang et al. (2008) explored the development of resistance factors at SLS condition using

Monte Carlo simulation (MCS). The statistics of tolerable displacements reported by

Zhang (2005) were used by Wang (2008) to establish a displacement-based comparison

similar to Equation (4.7) and calibrate resistance factors.

Vu (2013) developed resistance factors at SLS condition for the design of drilled shafts

installed in shales using a factored strength parameter approach for the calibration process

and the use of displacement based method, Equation (4.7), where tolerable displacement

and design displacement were variables of the performance function. An appropriate load


110

transfer model for drilled shafts in shales was identified (Vu 2013) using the “t-z” and “q-

z” analysis and variabilities and uncertainties associated with the model were quantified.

MCS was implemented (Vu 2013) to develop resistance factors.

Load Approach

As previously mentioned, the SLS performance function can be expressed in terms of

loads, as per Equation (4.4). In this case, in order to account for uncertainties carried during

the design process, a resistance factor is applied to the load corresponding to the tolerable

displacement (Qtol). Following the LRFD methodology, the factored Qtol has to be greater

than or equal to the design load (Qd) to ensure satisfactory performance under SLS

conditions as shown in Equation (4.9).

φ𝑆𝐿𝑆𝑄𝛿𝑡𝑜𝑙 ≥ 𝑄𝑑 (4.9)

The load-based approach for the calibration of resistance factors for the SLS condition has

been used by several researchers. Each study has contemplated a unique case and dataset

to achieve resistance factors at SLS condition.

Phoon and Kulhawy (2003) explored the design of drilled shafts in medium, stiff and very

stiff clays, and developed resistance factors for the SLS condition. A performance function

was proposed and analyzed following the form shown in Equation (4.9) where the load


111

corresponding to tolerable displacement was factorized and compared to the nominal

design load.

Similarly, Misra and Roberts (2009) developed a series of research studies focusing on the

behavior of drilled shafts at SLS. Roberts et al. (2008 and 2010) developed resistance

factors for the SLS conditions using MCS method and following the form of Equation

(4.9). Roberts et al. (2010) explored the uncertainties associated with the determination of

shaft and base resistances. Using the "t-z” and “q-z" model, the percentage of the design

load carried by shaft and base at any load cycle was determined. Based on this approach,

resistance factors were developed for allowable total displacements using the First Order

Reliability Method (FORM).

The load-based approach was selected for this present study, and the calibration process

for the development of resistance factors at SLS condition follows the expression shown

in Equation (4.9).

Calibration of Resistance Factors

In the LRFD methodology, several methods can be used to calibrate resistance factors.

Two of the widely-used methods are the First Order Second Moment (FOSM) and the

Monte Carlo Simulation (MCS), see Baecher and Christian (2005) and Fenton and Griffiths

(2007). According to Allen et al. (2005), prior to the determination of resistance factors


112

(), it is important to understand the statistical distribution of the bias () defined as the

ratio of measured resistance (RM) and predicted resistance (RP), Equation (4.10).

𝜆𝑅 = 𝑅𝑀

𝑅𝑃 (4.10)

In the case of deep foundations and calibration process of the resistance factor, the

measured resistance (RM) can be determined from results of a full scale load test; whereas,

the predicted resistance (RP) is obtained from results of predictive models. The statistical

distribution of the bias will aid to select the appropriate method for the calculation of

resistance factors.

For calibration process of SLS resistance factors (SLS), the measured resistance RM is

replaced by the load corresponding to tolerable displacement Qtol which is determined

from the load-settlement curves, and the predicted resistance Rp is replaced by design load

Qd predicted from the TCP design method.


Dataset Development

The calibration process of resistance factors is directly a function of measured and

predicted capacities. Therefore, predictive models used for the design of deep foundations

should be paired with the corresponding resistance factors calibrated using the same


113

predictive model. For example, if resistance factors are calibrated based on predicted

capacities obtained from a SPT-based method, then those resistance factors are only

applied to SPT-based design method. In a recent TxDOT-sponsored research project

assessing the reliability of the TCP foundation design method, Seo, et al. (2015a, 2015b,

2015c) presented the development of resistance factors at the ULS conditions. The dataset

from that study was used for the SLS analyses presented in this paper.

Load Tests and TCP Borings

The Seo et al. (2015) dataset consisted of 60 load test cases with corresponding TCP

borings. These data source to full scale load tests from TxDOT archive projects,

supplemented by other full-scale load test projects from neighboring states including

Louisiana, Arkansas, Missouri, and New Mexico (Seo et al. 2015a, 2015b, 2015c).

For the non-TxDOT projects, 21 TCP borings were completed at 16 sites following

TxDOT’s Geotechnical Manual specifications. For identification and classification of the

subsurface materials associated with the TCP test, disturbed samples were collected using

split-spoons and Shelby tubes. Due to insufficient information corresponding to full scale

load tests, data corresponding to IGM and rock material were not considered for the

analyses presented in this study.


114

Final Compiled Dataset

The final dataset for this study consisted of 60 full scale load tests comprising 33 driven

piles and 27 drilled shafts in soils. Driven piles selected for this project are precast square

concrete piles. The width of the pile ranged from 36-cm (14-in) to 51-cm (20-in) with

embedment lengths varying between 5-m (15-ft) and 25-m (84-ft). For the drilled shafts,

the diameters ranged from 46-cm (18-in) to 183-cm (72-in) with total length varying

between 6-m (20-ft) and 30-m (99-ft). Additionally, each foundation element included in

the final dataset was associated with a TCP boring carried out within an average distance

of 30-m (100-ft) from the location of the load test. The data extracted from the borings

included predominant soils and the TCP test blowcounts for each strata. Results from full

scale load test projects and TCP borings are presented in Table 4.1 for driven piles and

Table 4.2 for drilled shafts.

Load Corresponding to Tolerable Displacement (Qtol) and Design Load (Qd)

Since the load-based approach was used to calibrate SLS resistance factors in this study,

the determination of the load corresponding to tolerable displacement (Qtol) and the design

load (Qd) are main requirements for the calibration process. Using load-settlement curves

from full-scale load tests; e.g. Figure 4.4, the load corresponding to the tolerable

displacement (Qtol) was determined following two main steps: (1) identify the tolerable


115

displacement on the vertical axis, and (2) read the corresponding load from the horizontal

axis.

Predicted allowable capacity for driven piles and drilled shafts was calculated following

the design procedure specified by the TxDOT’s Geotechnical Manual. The design soil

profile associated with each foundation element was identified and the average TCP

blowcounts per stratum was determined. Unit shaft and unit base resistances were

determined using the average NTCP per soil stratum and design charts shown in Figure

4.2(a) and 4.2(b). The allowable shaft and base resistances were calculated for each driven

pile and drilled shaft using the dimensions shown in the load test report. The total allowable

capacity, which has been considered as design load (Qd) in this study, was determined by

summing up the allowable shaft and base capacities.

From the above described process, Qtol corresponding to different levels of tolerable

displacement and Qd based on results from TCP borings were determined for driven piles,

Table 4.1, and for drilled shafts, Table 4.2.

It is important to emphasize that the term allowable refers to the maximum recommended

load that a foundation element can support. It is reasonable to assume that design loads are

always equal or smaller than allowable loads. For this study, information regarding design

loads were not available and the allowable load calculated from the TCP design charts was

considered as the design load.


116

Table 4.1. Compiled Dataset for Driven Piles

Case

No.

Pile Dimensions Measured Capacity Corresponding to

Tolerable Displacement, Qtol (kN) Predominant

Soil

Predicted

Design Load,

Qd (kN) Width

(cm)

Length

(cm) 2.54

(mm)

5.08

(mm)

6.35

(mm)

7.62

(mm)

12.70

(mm)

19.05

(mm)

25.40

(mm)

1 36 914 401 481 490 498 543 561 570 Fine 431

2 36 777 935 1335 1513 1558 1682 1802 1851 Fine 385

3 36 671 757 979 1068 1113 1246 1304 1308 Fine 437

4 36 1052 623 779 823 846 894 912 935 Coarse 808

5 51 2112 1068 2047 1958 2092 2670 3026 3249 Fine 748

6 38 1068 712 1024 1157 1291 1647 1731 1851 Coarse 622

7 38 1005 668 1046 1090 1246 1424 1491 1647 Coarse 677

8 41 1501 801 1170 1308 1420 1709 1905 2020 Fine 552

9 41 1298 623 890 935 979 1090 1157 1193 Fine 984

10 41 1341 712 979 1113 1157 1308 1393 1437 Fine 523

11 41 1425 801 1157 1291 1380 1669 1847 1958 Fine 1241

12 41 497 579 868 979 1024 1135 1215 1273 Coarse 313

13 41 457 801 1313 1469 1535 1691 1780 1936 Coarse 236

14 41 1425 534 935 1068 1157 1335 1433 1491 Fine 910

15 41 914 712 890 957 1001 1090 1135 1139 Fine 669

16 41 945 801 1113 1224 1268 1371 1424 1437 Coarse 441

17 41 1219 223 378 445 534 734 948 1157 Fine 466

18 38 945 579 935 1046 1135 1357 1513 1602 Fine 404

19 51 2347 823 979 1024 1068 1126 1157 1179 Fine 1345

20 36 808 1246 1558 1580 1602 1780 1882 1936 Fine 1647

21 46 1006 712 890 1139 1273 1647 1936 2118 Fine 526

22 46 2545 645 890 979 1046 1224 1313 1380 Coarse 642

23 46 1353 979 1024 1037 1041 1059 1068 1072 Fine 877

24 36 785 801 1024 1068 1113 1148 1215 1237 Fine 530

25 46 1280 668 1202 1335 1513 1736 1780 1825 Coarse 824

26 46 1280 846 1424 1558 1602 1736 1758 1780 Coarse 824

27 46 1219 734 1113 1291 1335 1580 1713 1749 Coarse 763

28 46 1280 890 1513 1780 2003 2603 3093 3422 Coarse 824

29 46 1676 890 1691 2092 2537 3004 3160 3204 Fine 1340

30 51 2195 1647 2359 2626 2759 3026 3257 3422 Fine 2203

31 51 2195 1647 2092 2181 2225 2474 2568 2617 Fine 1568

32 36 1311 534 890 979 1068 1135 1193 1202 Coarse 687

33 36 2438 668 1246 1424 1691 2359 2982 3427 Fine 1837


117

Table 4.2. Compiled Dataset for Drilled Shafts

Case

No.

Drilled Shafts

Dimensions

Measured Capacity Corresponding to

Tolerable Displacement, Qtol (kN) Predominant

Soil

Predicted

Design Load,

Qd (kN) Diam.

(cm)

Length

(cm) 2.54

(mm)

5.08

(mm)

6.35

(mm)

7.62

(mm)

12.70

(mm)

19.05

(mm)

25.40

(mm)

1 91 762 2225 3338 3916 4005 6008 7209 7743 Fine 2085

2 122 1554 4450 6230 6853 7343 8277 8811 9123 Coarse 5461

3 122 1554 3115 4895 5830 6230 7209 8010 8411 Coarse 5461

4 122 1554 3338 5785 6675 8900 14240 17800 20025 Coarse 5461

5 91 1829 3560 4984 5518 5741 6542 6964 7254 Fine 3700

6 61 610 1157 1602 1780 2003 2581 3115 3471 Coarse 1009

7 46 805 512 525 534 534 547 556 556 Fine 402

8 46 701 436 458 467 490 512 534 543 Fine 275

9 76 701 868 979 1024 1046 1121 1157 1179 Fine 545

10 76 701 668 725 757 792 810 837 859 Fine 545

11 76 1372 1869 2225 2314 2359 2537 2670 2723 Fine 1114

12 76 1798 2804 3560 3783 3872 4139 4272 4361 Coarse 2280

13 76 2347 2804 4895 5474 5607 5919 6030 6052 Coarse 1783

14 76 1372 2403 3338 3516 3738 4406 4895 5296 Coarse 3502

15 183 1996 3738 4806 5474 5785 7343 7877 8856 Coarse 4423

16 122 2627 2670 4139 5340 5652 7699 8856 9568 Coarse 3650

17 122 2652 2670 4228 5385 5563 7521 8589 9879 Coarse 3773

18 168 1895 2003 3071 3560 3738 4228 4495 4673 Fine 3773

19 168 1237 2848 3516 3694 3783 4094 4317 4539 Fine 2105

20 183 2496 4005 6141 6453 6675 7120 7432 7743 Fine 4705

21 122 3018 3071 5073 5830 6675 8277 9167 9968 Fine 4018

22 122 1204 1958 2715 2982 3115 3427 3560 3649 Fine 1883

23 168 1433 3471 5251 5518 5741 6230 6453 6675 Coarse 3059

24 79 1433 2804 3961 4450 4673 5518 5963 6453 Coarse 5067

25 79 1433 2314 4450 4984 5385 6408 7031 7476 Coarse 5108

26 122 2271 1335 1914 2670 2937 4183 5207 6186 Coarse 5137

27 81 1673 890 1558 1958 2314 3382 4450 5340 Coarse 2764

Tolerable Displacement for the TCP design method

Displacements corresponding to the predicted design load (i.e. allowable capacity)

determined from TCP design charts were obtained from the load-settlement curves


118

corresponding to each drilled shaft or driven pile analyzed in this study. Histograms

showing the distribution of displacements under design load Qd were generated and

presented in Figure 4.5. From the histograms and the cumulative distribution function, it is

noted that from a total dataset of 33 driven piles, 20 and 28 piles experienced displacements

less than 2.54-mm (0.1-in) and 6.35-mm (0.25-in), respectively, under design load. In the

case of drilled shafts, a total of 16 and 21 out of 27 drilled shafts experienced displacements

less than 2.54-mm (0.1-in) and 6.35-mm (0.25-in), respectively, under design load. This

indicates that when the foundation was sized in such a way that the allowable capacity was

about the same as the design load, which is typical design practice, the majority of the deep

foundations designed following the TCP method did not experience more than 6.35-mm

(0.25-in) of settlement. Furthermore, none of the driven piles exceeded 12.7-mm (0.5-in)

of settlement under the allowable capacity, and only one case from drilled shafts slightly

exceeded 12.7-mm (0.5-in).

In a separate process, a detailed review of available technical documents and construction

drawings associated with the compiled dataset presented in this study was completed. From

this review, the average bridge span length corresponding to the projects associated with

the load test cases was determined. From the review, a total of 77 bridge spans for driven

piles and 28 bridge spans for drilled shafts were identified. With an average bridge span of

18-m (60-ft) for bridges supported by driven piles and 21-m (70-ft) for bridges supported

by drilled shafts, average tolerable displacements were calculated using Equation (4.6),

resulting in 36-mm (1.4-in) for driven piles and 42-mm (1.7-in) for drilled shafts.


119

Based on results obtained from the histograms generated for this study, Figure 4.5, and the

outcome of Equation (4.6) for the average bridge spans, the tolerable displacements for

driven piles and drilled shafts analyzed in this study range from 2.54-mm (0.1-in) to 31.75-

mm (1.25-in). However, load-settlement curves associated with each foundation did not

surpass a maximum displacement of 25.4-mm (1.0-in). Therefore, in order to avoid any

uncertainties associated with extrapolations of load-settlement curves beyond 25.4-mm (1-

in) settlement, all analyses were completed for tolerable displacements ranging from 2.54-

mm (0.1-in) to 25.4-mm (1.0-in).

(a) (b)

Figure 4.5. Tolerable displacement histograms for (a) driven piles and (b) drilled

shafts

RESULTS OF ANALYSES

Statistical analyses were performed to determine resistance factors for the TCP design

method at the SLS condition. Based on predicted design load (Qd) and measured load

corresponding to tolerable displacement (Qtol), the bias factor was calculated followed by

10987654321

14

12

10

8

6

4

2

0

Displacement (mm)

Fre

qu

en

cy

1

43

11

14

Histogram of DisplacementsDriven Piles

2019181716151413121110987654321

18

16

14

12

10

8

6

4

2

0

Displacement (mm)

Fre

qu

en

cy

1

54

16 Histogram of Displacements Drilled Shafts

1

2.54-mm

6.35-mm

2.54-mm

6.35-mm


120

a careful analysis of the bias distribution. After selecting the proper distribution of bias,

FOSM and MCS were used to determine resistance factors.

Bias

As previously described, the bias is considered as the ratio of the measured and predicted

capacity following the form presented in Equation (4.10). For the SLS analyses presented

in this paper, the measured capacity RM in Equation (4.10) is substituted by the load

corresponding to the tolerable displacement (Qtol) and the predicted capacity RP is the

design load (i.e. total allowable capacity) for the deep foundation calculated using the TCP

predictive design model (Qd,TCP). Accordingly, Equation (4.10) is rewritten as follows:

𝜆𝑅−𝑆𝐿𝑆 = 𝑄𝛿𝑡𝑜𝑙

𝑄𝑑,𝑇𝐶𝑃 (4.11)

Similarly, the design inequality equation to be used for calibration of resistance factor is

written as follows:

φ𝑆𝐿𝑆𝑄𝛿𝑡𝑜𝑙 ≥ 𝑄𝑑,𝑇𝐶𝑃 (4.12)

For each load test analyzed in this study, the values of Qtol were obtained from the load-

settlement curves and Qd, TCP was obtained using the TCP design charts. Then the biases

were obtained using Equation (4.11) for various values of tolerable displacement. To


121

establish the statistical distribution of the biases used in this study, an Anderson-Darling

test was performed using Minitab 17.3.1 (2016) to evaluate the null hypothesis Ho: “Data

distribution for the biases follows a lognormal distribution.” For this test, it was determined

that with a p-value greater than 0.05 there is not sufficient evidence that suggests a data

distribution different from lognormal. Therefore, for all statistical analyses, a lognormal

distribution of the biases was considered. Figure 4.6 illustrates one of the representative

cases of bias histograms for driven piles and drilled shafts including their corresponding

probability plot for a tolerable displacement of 6.35-mm (0.25-in).

For each tolerable displacement, the mean and variance for biases were estimated using the

Uniformly Minimum Variance Unbiased Estimators (UMVUEs). Surles et al. (2016)

explored the use of UMVUEs for lognormal distributions following the work presented by

Finney (1941). An examination and comparison of Maximum Likelihood Estimators

(MLE) versus the UMVUE method, resulted with the conclusion that for a lognormal

distribution, the UMVUE approach is more appropriate as it has higher precision and less

data variability. Table 4.3 presents the summary statistics for all tolerable displacements

selected for this study using the UMVUE approach.


122

(a)

(b)

(c)

(d)

Figure 4.6. Histograms and probability plots for driven piles (a and b, respectively)

and for drilled shafts (c and d, respectively)

Table 4.3. Summary Statistics of Biases for Driven Piles and Drilled Shafts in Soil

Tolerable Displacement

tol(mm)

Driven Piles Drilled Shafts

R-SLS COV R-SLS COV

2.54 (0.1 in) 1.146 0.491 0.957 0.505

5.08 (0.2 in) 1.675 0.488 1.31 0.419

6.35 (0.25 in) 1.851 0.497 1.445 0.368

7.62 (0.3 in) 1.979 0.492 1.53 0.349

12.7 (0.5 in) 2.282 0.501 1.806 0.331

19.05 (0.75 in) 2.472 0.513 1.982 0.333

25.4 (1 in) 2.603 0.527 2.112 0.333

From the summary statistics presented in Table 4.3 for both driven piles and drilled shafts,

it is noted that the bias (R-SLS) is approximately equal to 1.0 for a tolerable displacement

6.44.83.21.60.0

16

14

12

10

8

6

4

2

0

( )

Fre

qu

en

cy

Histogram of Bias for Driven Piles

= 6.35-mm (0.25-in)

Lognormal Fit

101

99

95

90

80

7060504030

20

10

5

1

N 33

AD 0.233

P-Value 0.781

( )

Perc

en

t

Probability Plot of Bias, Driven PilesLognormal - 95% Confidence Interval (CI)

= 6.35-mm (0.25-in)

3.02.41.81.20.6

8

7

6

5

4

3

2

1

0

Fre

qu

en

cy

Lognormal Fit

( )

Histogram of Bias For Drilled Shafts

= 6.35-mm (0.25-in)

1.51.00.50.0-0.5-1.0

99

95

90

80

7060504030

20

10

5

1

N 27

AD 0.436

P-Value 0.277

( )

Perc

en

t

Probability Plot of Bias, Drilled ShaftsLognormal-95% Confidence Inteval (CI)

=6.35-mm (0.25-in)


123

of 2.57-mm (0.1-in). Further, as the tolerable displacement is increased, the bias factor is

increased as well. This trend is expected. Intuitively, when the foundation element is

allowed to experience larger tolerable displacement, larger load corresponding to the

tolerable displacement (Qtol) is obtained from the load-settlement curve. Recalling the bias

(R-SLS = Qtol/Qd, TCP) defined by Equation (4.11), for a constant denominator (Qd, TCP) and

an increasing numerator (Qtol), the ratio (i.e., the bias factor) increases which explains the

trend observed in the results presented in Table 4.3.

Furthermore, from Table 4.3 it is observed that the coefficient of variation (COV) slightly

increases for driven piles and decreases in the case of drilled shafts as the tolerable

displacement increases. The COV is a measure of spread that indicates the degree of

variability of the data relative to the mean. Although a trend is observed in the COV for

both driven piles and drilled shafts, the variance (i.e., standard deviation) is increasing for

both type of foundations following the same trend as the bias.

Resistance Factors at SLS condition SLS

For target reliability indices () of 2.33 and 3.00, dead load to live load ratio of two (i.e.

DL/LL = 2.0), and tolerable displacements ranging from 2.54-mm (0.1-in) to 25.4-mm

(1.0-in), resistance factors at SLS condition (SLS) were determined for driven piles and

drilled shafts. For analyses purposes, all load factors corresponding to dead load (DL) and

live load (LL) were considered as unity following the specifications marked by AASHTO


124

(2014). From the histograms created for the biases, it was determined that the statistical

distribution of the bias for the dataset compiled for this study follows a lognormal

distribution. Therefore, based on the FOSM calibration method with a lognormal

distribution and the expression presented by Barker et al. (1991), resistance factors were

developed. In addition to the FOSM and as comparator, the MCS method was used to

calibrate resistance factors.

In the Monte Carlo simulation, the zeroth step is to select all parameters that are going to

remain fixed throughout the simulation. One such parameter is the simulation size. In this

study, the simulation size was chosen to be Nsim = 1,000,000. Furthermore, reliability

indices () of approximately 2.33 and 3.00 corresponding to target probabilities of failure

of 0.01 and 0.001, respectively, will remain fixed. Other parameters required are the

nominal live load (LL), along with its associated coefficient of variation (COVLL), bias

(λLL), and load factor (γLL). Additionally required are the nominal dead load (DL), along

with its associated coefficient of variation (COVDL), bias (λDL), and load factor (γDL). Note

that the nominal live/dead load and their associated coefficient of variations are in the

original scale. In the study presented in this paper, LL= 1.15, COVLL = 0.2, DL= 1.05, and

COVDL= 0.1, were used following suggestions by NCHRP (Paikowsky 2004). The

following algorithm outlines the calculation of the LRFD resistance factors following the

MCS:


125

0. Set the values of all parameters as described above. The mean bias for resistance

(R-SLS) and coefficient of variation for the bias (COVR-SLS) are taken to be the

weighted UMVUE sample estimates suggested by the sample. Take an initial guess

at to be that suggested by FOSM, trial.

1. Generate a lognormal random sample of Number of simulations (Nsim) live loads

with their corresponding mean and COV to determine the ith random live load

(LLRND (i)).

2. Generate a lognormal random sample of Nsim dead loads with their corresponding

mean and COV to determine the ith random dead load (DLRND (i)).

3. Generate a lognormal random sample of Nsim resistances with their corresponding

mean and COV to determine the ith random resistance (RRND (i)).

4. Let G (i) = RRND (i) – LLRND (i) – DLRND (i). The estimated probability of failure is

the number of G (i) <0 divided by Nsim. Depending on the result the trial can be

increased or decreased. If the result is less than the target probability, then trial can

be increased, otherwise it can be decreased.


126

The stopping criteria was to stop when the value of trial (to the nearest 0.005) was found

which gave the estimated probability of failure nearest to the target probability of failure.

Software R, version 3.3.0 (2013) was used for the simulations.

Table 4.4 summarizes resistance factors determined for the SLS conditions based on FOSM

and MCS for driven piles and drilled shafts and indicates that SLS increases with increasing

tol. This is intuitive in the sense that more robust foundation design would be needed when

the superstructure can tolerate only very small displacement. On the other hand, when tol

is large, only small reduction of Qtol is required (e.g. SLS for 25.4 mm of tol obtained

from MCS is 0.86 for drilled shaft with = 2.33). In fact, since a load factor of unity is

used for SLS design, a SLS of 1.0 would be essentially the same as ASD for the TCP design

method indicating that the TCP design charts accurately predict total capacity of deep

foundations at tolerable displacements corresponding to a SLS = 1.0.

Table 4.4. Resistance Factors at Serviceability Limit State (SLS) for Foundations in Soils

tol

(mm)

Driven Piles Drilled Shafts

=2.33 =3.00 =2.33 =3.00

FOSM MCS FOSM MCS FOSM MCS FOSM MCS

2.54 (0.1 in) 0.30 0.32 0.21 0.23 0.24 0.26 0.17 0.18

5.08 (0.2 in) 0.44 0.47 0.31 0.34 0.40 0.43 0.29 0.32

6.35 (0.25 in) 0.47 0.51 0.33 0.37 0.49 0.54 0.37 0.41

7.62 (0.3 in) 0.51 0.55 0.36 0.40 0.54 0.60 0.41 0.47

12.7 (0.5 in) 0.58 0.62 0.41 0.45 0.66 0.74 0.51 0.59

19.05 (0.75 in) 0.61 0.66 0.43 0.45 0.72 0.80 0.55 0.64

25.4 (1.0 in) 0.62 0.67 0.43 0.47 0.77 0.86 0.59 0.68


127

As previously described, many researchers have reported values corresponding to tolerable

displacements for the SLS condition greater than 25.4-mm (1.0 in). It is reasonable to

consider the same condition for the study presented in this paper. However, the load-

settlement curves compiled for this study did not extend beyond 25.4 mm (1.0 in)

settlement which precluded the development of resistance factors for displacements of

more than 25.4 mm (1.0 in).

Although the concept of angular distortion previously presented and discussed is in

function of differential settlement, the tolerable displacements shown in Table 4.4 are total

vertical displacements and these were considered as settlements obtained from results of

full scale load tests completed for the driven piles and drilled shafts included in the dataset.

This further highlights that more work remains to be done for foundations where the

tolerable displacement is greater than 25.4-mm (1-in).

Load factors of unity were used in this study, but the variability in loads considered for a

ULS condition and addressed with load factors could affect the SLS condition as well. It

is reasonable to consider a separate research study to further explore the impact of load

variability on the SLS approach in the LRFD methodology.


128

CONCLUSIONS

Summary of Findings

Although most geotechnical engineering designs are based on ultimate strength,

displacements experienced by the superstructure govern the functionality and

serviceability of the superstructure. Resistance factors at serviceability limit state (SLS)

for the load and resistance factor design of deep foundations using TxDOT’s TCP design

method have been developed and presented in this paper.

The final dataset compiled for the study consisted of 60 full scale load tests, carried out for

33 driven piles and 27 drilled shafts in soils. Design load (Qd) and loads corresponding to

various values of tolerable settlement (Qtol) were determined for the entire dataset using

TxDOT’s TCP design procedure and load-settlement curves, respectively. From these

values, bias factors were determined and their corresponding statistical distributions were

analyzed. For each tolerable displacement, resistance factors for SLS condition (SLS) were

developed using the FOSM and MCS approaches with a dead load to live load ratio of 2.0.

Analysis results indicated that the majority of the foundations designed following

TxDOT’s current allowable stress design method would not yield more than 12.7-mm (0.5-

in) of head settlement.

For driven piles and a target reliability index of 2.33, SLS resistance factors ranged from

0.30 to 0.77 using FOSM method and 0.32 to 0.67 using the MCS approach for


129

corresponding tolerable settlement of 2.54-mm (0.1-in) and 25.4-mm (1.0-in), respectively.

Similar statistical analyses were completed for a target reliability index of 3.00 and SLS

resistance factors varied from 0.21 to 0.43 under FOSM method and 0.23 to 0.47 based on

MCS.

In the case of drilled shafts, for a target reliability index of 2.33, SLS resistance factors

were obtained ranging from 0.24 to 0.77 following the FOSM method and 0.26 to 0.86

based on MCS for corresponding tolerable settlement of 2.54-mm (0.1-in) and 25.4-mm

(1.0-in), respectively. Similarly, for target reliability index of 3.00, SLS resistance factors

varied from 0.17 to 0.59 under FOSM analysis whereas the outcome of MCS ranged

between 0.18 and 0.68.

Limitations/Further Study

A range of SLS resistance factors for driven piles and drilled shafts corresponding to

different levels of tolerable displacements was developed based on a relatively large dataset

compared to previous work related to serviceability limit state.

One of the limitations of this study is that the tolerable displacements (tol) were restricted

to 25.4-mm (1.0-in). The load-settlement data were obtained from projects where because

of specifications or needs, loads applied to the tested foundation did not reach levels that

would have created settlements beyond 25.4-mm (1.0-in). Load test data for foundations

with a wider range of tolerable displacements would address this gap.


130

Further, the load test dataset analyzed herein depicts foundations in soil materials only.

Separate studies are needed to develop resistance factors at serviceability limit state for

foundations bearing in intermediate geomaterials and rock, both materials being capable of

evaluation using the TCP test. With separate studies, the range of tolerable displacement

can be assigned based on expected displacements. Smaller tolerable displacements can be

evaluated for intermediate geomaterials and rock; whereas, larger tolerable displacements

can be analyzed for soil materials.


131

CHAPTER V


The effort associated with the development of this dissertation is summarized in this

chapter. Notable findings, recommendations, and key contributions are briefly discussed

and highlighted. Further, limitations to the dataset and challenges associated with the

development of datasets noted. Finally, based on the findings of this dissertation, a

discussion of the TxDOT TCP test, its foundation design charts, and considerations are

presented.

For the study of hammer efficiency, rod length correction factors, and overburden

correction factors, a dataset of 293 TCP tests including 251 tests completed in soils and 42

in IGM/Rock was compiled. From the 293 TCP tests, 135 tests were instrumented to obtain

energy measurements using the SPT Analyzer, with 119 instrumented tests in soils and 16

instrumented tests in IGM/Rock. An adjusted dataset of 251 TCP tests were considered for

the development of overburden correction factors for the TCP test in soils. A total of 119

instrumented TCP tests in soils were used for the analysis corresponding to the rod length

correction factors. From measured and theoretical hammer energy values, the average TCP

hammer efficiency was obtained to be 89%, 90%, and 88% for undifferentiated data,

coarse-grained soils, and fine-grained soils, respectively. Results from analyses showed

that the TCP hammer efficiency is similar in range to the published values for the SPT but

about 7% higher than the SPT hammer efficiency.


132

Based on statistical analyses, rod length correction factors and overburden correction

factors for the TCP test Blowcount values (NTCP) were developed. Correction factors for

the SPT have been published and are well established in the literature. These values were

compared to the TCP correction factors developed in this dissertation, and it appears that

the TCP correction factors for both rod length and overburden pressure are in similar with

higher correction factors for the TCP test. One of the main factors that could explain this

difference, is the type of sampler where in the case of the SPT a hollow split spoon is used,

and in the case of the TCP test a solid steel cone tip is driven into the subsurface material.

Although not as influential as the case of the samplers, the difference between the TCP and

SPT hammer weight, and hammer drop distance could influence the correction factors as

well. The determination of the TCP hammer efficiency, and the development of rod length

and overburden correction factors for the TCP test are considered as initial steps towards

the improvement and refinement of the TCP test.

Admittedly, even though correction factors were to be applied to the NTCP, if the predictive

validity of the TCP foundation is within low confidence, then the design charts cannot be

used reliably for the design of deep foundations. Hence, an evaluation of the predictive

validity of the TCP design charts were completed to explore this scenario. For this

evaluation a dataset of 60 full-scale load tests comprising 33 driven piles and 27 drilled

shafts was used where the allowable measured total capacity was compared to the

allowable predicted total capacity. This comparison was completed based on a qualitative

analysis where the data points were compared to a line of equality, and a quantitative

analysis where for the data corresponding to driven piles and drilled shafts regression


133

models were developed and further were compared to the line of equality. For the allowable

measured total capacity strength-based model and settlement-based models were used.

Results from the qualitative analyses suggested that different allowable capacity criteria

leads to different conclusions regarding the predictive validity of the TxDOT TCP method

for deep foundation design. The allowable predicted total capacity determined from the

TCP foundation design charts could be considered to be reasonable based on a strength-

based model, and conservative to over-conservative for settlement-based models. In

general, this suggests that serviceability-based interpretation of the TCP foundation design

charts (which is closely associated with observed structure performance) will be on the

conservative side where the performance of the foundation or the superstructure is

satisfactory. From the quantitative analyses, similar trend was observed where the

serviceability-based models were conservative and the strength-based models were

unconservative. For the quantitative analysis the predictive validity of the TCP design

charts were explored based on foundation type. From the analyses it was evident that the

predictive validity of the TCP foundation design charts for driven piles was not as robust

as it was in the case of drilled shafts. The predictive validity of the TCP test and its

foundation design charts has been questioned by practitioners in general and the main

uncertainty has been that if the TCP design chart is a strength-based model, and most of

the transportation structures are design using strength-based models, then why the

foundations or the superstructures are performing well? The answer is because of the

serviceability-based models. Both qualitative and quantitative analyses showed that models

based on tolerable settlements are conservative, meaning that the performance of the


134

foundation or the superstructure will not be jeopardized, even though the strength-based

model is unconservative. So, the predictive validity of the TCP foundation design charts is

“alright” in case of strength-based models, and “very good” in case of serviceability-based

models.

Although most geotechnical engineering designs are based on ultimate strength,

displacements experienced by the superstructure govern the functionality and

serviceability of the superstructure. The TCP design method is still considered as an

Allowable Stress Design (ASD) method, and the geotechnical community involved with

federally funded projects are transitioning to the Load and Resistance Factor Design

(LRFD) methodology. Considering that the foundation system for many transportation

infrastructures are designed based on the TCP test, resistance factors for the serviceability

limit state condition were developed for seven tolerable settlements and presented in this

dissertation. The final dataset compiled for the study consisted of 60 full scale load tests,

carried out for 33 driven piles and 27 drilled shafts in soils. For driven piles and a target

reliability index of 2.33, resistance factors ranged from 0.30 to 0.77 using First Order

Second Moment (FOSM) method and 0.32 to 0.67 using the Monte Carlo Simulation

(MCS) approach for corresponding tolerable settlements of 2.54-mm (0.1-in) and 25.4-mm

(1.0-in), respectively. In the case of drilled shafts, for a target reliability index of 2.33, SLS

resistance factors were obtained ranging from 0.24 to 0.77 following the FOSM method

and 0.26 to 0.86 based on MCS for corresponding tolerable settlements of 2.54-mm (0.1-

in) and 25.4-mm (1.0-in), respectively. Results from analyses showed that as the

foundation element is allowed to experience higher vertical displacements, larger


135

resistance factors are determined. Further, it was observed that in the case of serviceability

higher probability of failure such as (1/75 or 1/50) can be used instead of the rigorous

values used for an ultimate limit state such as (1/100 or 1/1000).

NOTABLE FINDINGS

From the above discussion and the content of this dissertation, notable findings are

summarized:

Average TCP hammer efficiency was determined as 90% for coarse-grained soils,

88% for fine-grained soils, and 89% for undifferentiated data

Average TCP hammer efficiency is about 7% higher than the average hammer

efficiency for the SPT

Rod-length correction factors are slightly higher for the TCP test compared to the

SPT

The mathematical expression developed for the overburden pressure correction

factor for the TCP closely matches expressions developed for the SPT

The predictive validity of the TCP foundation design charts is within higher

confidence for serviceability-based models compared to strength-based models

Although TCP is a strength-based design method, foundations and superstructures

designed based on the TCP method perform satisfactory because the predictive

validity is within high confidence for serviceability-based models

The predictive validity of the TCP design charts is within higher confidence for

drilled shafts which is not true in the case of driven piles


136

As the foundation element is allowed to experience larger settlements, higher

resistance factors for serviceability limit state is determined

According to literature and the trend observed in the serviceability limit state

analysis, the tolerable displacements for bridges can be as large as 51-mm (2-in)

For various level of tolerable displacement an increasing trend in the resistance

factors at serviceability limit state was observed. At 25.4-mm (1.0-in) a resistance

factor of 0.67 for driven piles and 0.86 for drilled shafts were determined. This

suggests, that as the tolerable displacement approaches values beyond 25.4-mm

(1.0-in) the resistance factors at serviceability limit state will reach 1.0.

KEY CONTRIBUTIONS

Although the concept of the TCP test is not new, the hypotheses, analyses, in depth study,

and results obtained in this dissertation, represents significant contribution to the

geotechnical engineering community where the TCP test and its associated charts are their

primary method for the design of deep foundations. A series of key contributions are

highlighted in this section.

One of the limitations in geotechnical engineering and in particular foundation

design, is the sample size. The dataset compiled for measured and predicted

capacity, instrumented TCP test, and TCP data could strongly contribute to any

future research related to the deep foundation design.

The datasets compiled for serviceability limit state where loads corresponding to

seven levels of settlement are presented, contributes to the development of the


137

LRFD methodology for serviceability limit state criterion applied to deep

foundations

The work completed for the development of TCP correction factors, helps to

determine a more reliable penetration index and further contributes to a better and

safer design of deep foundations using the TCP chart

It was determined that there are factors other than rod-length and overburden

pressure that have statistically-significant influence on the average hammer

efficiency, and to further explore those factors, a rigorous and controlled research

is recommended to further study these factors

12% of bridges maintained din the National Bridge Inventory, are from Texas and

Oklahoma, and designed in accordance to the TCP test. Evaluating the predictive

validity of the TCP foundation design charts further contributes to the development

of a predictive model which leads to a better and safer design

Serviceability limit state analyses contributes to the development of settlement-

based criteria in geotechnical engineering which is intently discussed among

scholars and practitioners

Development of resistance factors at serviceability limit state sets the ground for

future research work on LRFD and the TCP test

Limitations

Major limitations for this dissertation was presented during the development of resistance

factors at serviceability conditions. The tolerable displacements had to be restricted to 25.4-

mm (1.0-in), mainly because load-settlement curves compiled for this dissertation, did not


138

present settlement levels greater than 51-mm (25.4-in). Load test data for foundations with

a wider range of displacements would address this gap. Further, the load test dataset

analyzed herein depicts foundations in soil materials only. Separate studies are

recommended to develop resistance factors at serviceability limit state for foundations

bearing in intermediate geomaterials and rock, both materials being capable of evaluation

using the TCP test.

The work presented in this dissertation is intended as an initial step to further refine the

TCP-based design to achieve more reliable design of deep foundations. The TCP design

method is a strength-based predictive model and as to date, all research work have been

completed focusing on the strength approach to determine allowable capacities. It is

important to keep in mind that the TCP foundation design charts were developed based on

correlations between TCP blowcounts and shear strength determined from laboratory

testing. The correlation lines of the TCP design charts predict allowable capacity from the

perspective of a strength-based model. This prediction has been strongly discussed among

researchers and the predictive validity of the TCP design charts has been questioned.

However, observing the performance of thousands of bridges designed in accordance to

the TCP design methodology, it is noted that none of these superstructures had shown

unsatisfactory performance.

Based on results of this study, the well performance of the foundations and transportation

infrastructures designed based on the TCP test can be explained from serviceability-based

models and reasonably make the statement that the TCP foundation design charts


139

moderately over-predict design capacities under strength-based approach, but

conservatively predicts allowable total capacities for serviceability-based approach.

Collectively, considering limitations to the dataset, scatter in the data, unavailability of

instrumented load tests, and wide range of geomaterials, more rigorous and well planned

research could enhance the data presented in this dissertation and further study the TCP

test and its associated design charts.


140

REFERENCES

CHAPTER 1 AND 2

Aggour, M. S., & Radding, W. R. (2001). Standard penetration test (SPT)

correction, Report No. MD02-007B48, Maryland State Highway Administration,

Baltimore.

ASTM International (2016). Standard Test Method for Standard Penetration Test (SPT)

and Split-Barrel Sampling of Soils, Test Method D1586-11. ASTM International: West

Conshohocken, PA.

Boulanger, R. W., & Idriss, I. M. (2004). Evaluating the potential for liquefaction or cyclic failure

of silts and clays (p. 131). Center for Geotechnical Modeling.

Boulanger, R. W., and Idriss, I. M. (2012). “Probabilistic Standard Penetration Test–Based

Liquefaction–Triggering Procedure.” Journal of Geotechnical and Geoenvironmental

Engineering, 138(10), 1185-1195.

Boulanger, R. W., and Idriss, I. M. (2015). “CPT-Based Liquefaction Triggering

Procedure.” Journal of Geotechnical and Geoenvironmental Engineering, 10.1061.

Brown, R. E. (1977). “Drill rod influence on standard penetration test.” Journal of the

Geotechnical Engineering Division, 103(11), 1332-1336.


141

Burmister, D. M. (1948). “The Importance and Practical Use of Relative Density in Soil

Mechanics.” Proceedings ASTM. Vol. 48, p. 1249.

Coduto, D. P., Yeung, M. C. R., & Kitch, W. A. (2011). Geotechnical engineering:

principles and practices 2nd Ed., Prentice Hall. New Jersey

Das, B.M., Sobhan, K. (2014). Principles of Geotechnical Engineering, 8th Ed., Cengage

Learning, Connecticut, USA

DeMello, V. F. (1971, June). “The standard penetration test.” In Proceedings of the 4th

Pan-American Conference on Soil Mechanics and Foundation Engineering. San Juan,

PR Vol. 1, pp. 1-86.

FHWA (2016). “National Bridge Inventory.” Bridges and Structures, Safety, Bridge

Inspection. U.S. Department of Transportation, Federal Highway Administration:

Washington, DC.

Fletcher, G. F. (1965). “Standard Penetration Test Its uses and Abuses.” Journal of Soil

Mechanics and Foundations SM4, Paper 4395.

Honeycutt, J.N., Kiser, S.E. and Anderson, J.B. (2014). “Database Evaluation of Energy

Transfer for Central Mine Equipment Automatic Hammer Standard Penetration Tests”.

ASCE Journal of Geotechnical and Geoenvironmental Engineering, 140(1): 194-200).


142

Idriss, I. M., & Boulanger, R. W. (2012). “Examination of SPT-based liquefaction

triggering correlations.” Earthquake Spectra, 28(3), 989-1018.

Ireland, H. O., Moretto, O., & Vargas, M. (1970). “Discussion: The dynamic penetration test: a

standard that is not standardized.” Geotechnique, 20(4), 452-456.

Kovacs W.D. and Salomone L.A. (1982) “SPT Hammer Energy Measurement.” Journal

of Geotechnical Engineering, 110(4), 562-564

Lacroix, Y. and Horn, H. (1973). “Direct Determination and Indirect Evaluation of

Relative Density and Its Use on Earthwork Construction Projects.” In Evaluation of

Relative Density and Its Role in Geotechnical Projects Involving Cohesionless Soils.

ASTM Special Technical Publication 523, pp. 251–280

Liao, S. S., & Whitman, R. V. (1986). “Overburden correction factors for SPT in

sand.” Journal of geotechnical engineering, 112(3) 373-377

Loehr, J.E., J.J. Bowders, L. Ge, W.J. Likos, R. Luna, N. Maerz, B.L. Rosenblad, and R.W.

Stephenson (2011). Engineering Policy Guidelines for Design of Drilled Shafts. Missouri

Department of Transportation, Final Report for Project TRyy0922, Report cmr12003, 75

pp.


143

Meyerhof, G. G. (1957). “Discussion on research on determining the density of sands.”

In Proc., 4th Int. Conf. of Soil Mechanics and Foundation Engineering.

Nam, M.S., Y.H. Park and Y.S. Park (2013). “Prediction of End Bearing Capacity for Pre-

Bored Steel Pipe Piles Using Instrumented Spt Rods”. Journal of the Korean Geotechnical

Society. Vol. 29, No. 12, December 2013.pp. 105-111.

Odebrecht, E., Schnaid, F., Rocha, M. M., & de Paula Bernardes, G. (2005). “Energy

efficiency for standard penetration tests.” Journal of Geotechnical and Geoenvironmental

Engineering. 131(10), 1252-1263

Palmer, D. J., & Stuart, J. G. (1957). “Some observations on the Standard Penetration Test

and a Correlation of the Test with a new Penetrometer.” In Proceedings of 4th Int. Conf.

on Soil Mechanics and Foundation Eng. p. 231.

Robertson, P. K., Campanella, R. G., & Wightman, A. (1983). “SPT-CPT Correlations.”

Journal of geotechnical engineering. 109(11), 1449-1459.

Schmertmann, J. H., & Palacios, A. (1979). “Energy dynamics of SPT.” Journal of the

Geotechnical Engineering Division, 105(8), 909-926.

Schnaid, F. (2009). In situ Testing in Geomechanics: the main tests. 1st Ed. Taylor &

Francis, London and New York


144

Seed, H. B., Tokimatsu, K., Harder, L., & Chung, R. M. (1984). The influence of SPT

procedures in evaluating soil liquefaction resistance. Report, University of California

Berkley, Earthquake Engineering Research Center, 84-15

Seo H., R. B. Moghaddam, J. G. Surles, P. W. Jayawickrama, T. A. Wood, and W. D.

Lawson (2015). Reliability Based Deep Foundation Design Using Texas Cone

Penetrometer (TCP) Test: VOLUME 1, Development of Load Test Dataset. DRAFT

Report. FHWA/TX-14-0-6788-1-Vol. 1. Texas Tech University Center for

Multidisciplinary Research in Transportation: Lubbock, TX.

Skempton, A. W. (1986). “Standard penetration test procedures and the effects in sands of

overburden pressure, relative density, particle size, ageing and over-consolidation.”

Geotechnique. 36(3), 425-447

Texas Department of Transportation (2012) Geotechnical Manual. Texas Department of

Transportation

Touma, F.T. and Reese L.C. (1972). The behavior of Axially Loaded Drilled Shafts in Sand,

Report No. CFHR 3-5-72-176-1 University of Texas Center for Highway Research

Turner, J. (2006). Rock-Socketed Shafts for Highway Structure Foundations. NCHRP

Synthesis 360. A Synthesis of Highway Practice. Transportation Research Board:

Washington DC.


145

Vipulanandan, C. (2007). “Recent advances in designing, monitoring, modeling and testing

deep foundations in North America.” Advances in Deep Foundations: Proceedings of the

International Workshop on Recent Advances of Deep Foundations (IWDPF07), Port and

Airport Research Institute, Yokosuka, Japan, 1-2 February, 2007. Y. Kikuchi, J. Otani, M.

Kimura, and Y. Morikawa (eds). Taylor and Francis Group: London.

CHAPTER 3

AASHTO (2014). AASHTO LRFD Bridge Design Specifications, Customary U.S. Units,

6th Edition, American Association of State Highway and Transportation Officials

Cavusoglu, E., Nam, M. S., O’Neill, M. W., & McClelland, M. P. (2004). Multi-method

strength characterization for soft cretaceous rocks in Texas. Geo-Support 2004: Innovation

and Cooperation in the Geo-Industry. pp. 199-210.

Coduto, D. P., Yeung, M. C. R., & Kitch, W. A. (2011). Geotechnical engineering:

principles and practices 2nd Ed., Prentice Hall. New Jersey

FHWA (2016). “National Bridge Inventory.” Bridges and Structures, Safety, Bridge

Inspection. U.S. Department of Transportation, Federal Highway Administration:

Washington, DC.


146

Hamoudi, M., Coyle, H., & Bartoskewitz, R. T. (1974). Correlation of the Texas Highway

Department Cone Penetrometer Test with Unconsolidated-Undrained Shear Strength of

Cohesive Soils. Texas Transportation Institute (TTI). Texas A&M University.

Lawson, W. D., Mikulec, C. M., McClelland, M. P., Irby, E. F., Jayawickrama, P. W., &

Levi, L. (2009). “Rock Strength by Texas Cone Penetrometer and Uniaxial Compression

Tests.” Geo-Institue of ASCE, GeoFlorida 2009. Palm Beach, Fl, 79-86.

MINITAB version 17.3.1 (2016). Computer Software, Minitab Inc.

Nam, M. S. (2004). Improved Design for Drilled Shafts in Rocks. University of Houston,

Houston, Tx

Nam, M.S., Y.H. Park and Y.S. Park (2013). “Prediction of End Bearing Capacity for Pre-

Bored Steel Pipe Piles Using Instrumented SPT Rods”. Journal of the Korean Geotechnical

Society. Vol. 29, No. 12, December 2013.pp. 105-111.







147


Transportation

Texas Department of Transportation Construction Division, (1999) Test Procedure for

Texas Cone Penetration TxDOT Designation: Tex-132-E Texas Department of

Transportation, Austin, TX

Texas Highway Department (1956). Foundation Exploration and Design Manual, Bridge

Division, Texas Highway Department

Turner, J. (2006). Rock-Socketed Shafts for Highway Structure Foundations. Report

NCHRP Synthesis 360. A Synthesis of Highway Practice. Transportation Research Board:

Washington DC.

Valluru, S., Vennalaganty, K. M., Endly, S. N., & Goin, W. (2007). “Soil characterization

of State Highway 130 (SH130) alignment.” Texas Section ASCE. Austin, Tx: ASCE

Varathungarajan, D. A., Garfield, S. M., & Wright, S. G. (2009). Characterization of

Undrained Shear Strength Profiles for Soft Clays at Six Sites in Texas. Report. Center for

Transportation Research (CTR). University of Texas at Austin


148

Vijayvergiya, V. N., Hudson, W. R., & Reese, L. C. (1969). Load Distribution for a Drilled

Shaft in Clay Shale. Report. Center for Transportation Research (CTR). University of

Texas at Austin.

Vipulanandan, C., Puppala, A. J., Jao, M., Kim, M. S., Vasudevan, H., Kumar, P., & Mo,

Y. L. (2008). Correlation of Texas Cone Penetrometer Test Values and Shear Strength of

Texas Soils. Report. University of Houston, Houston, TX

Vipulanandan, C. (2007). “Recent advances in designing, monitoring, modeling and testing

deep foundations in North America.” Advances in Deep Foundations: Proceedings of the

International Workshop on Recent Advances of Deep Foundations (IWDPF07), Port and

Airport Research Institute, Yokosuka, Japan, 1-2 February, 2007. Y. Kikuchi, J. Otani, M.

Kimura, and Y. Morikawa (eds). Taylor and Francis Group: London.

Wright, S. G., Garfield, S. M., & Varathungarajan, D. A. (2008). Correlation of Shallow,

Low Blow Count Texas Cone Penetrometer Values and Shear Strength of Texas Soils.

Report. Center for Transportation Research (CTR). University of Texas at Austin

Zhang, L. M., & Chu, L. F. (2005). “Calibration of methods for designing large-diameter

bored Piles: Serviceability Limit State”. Soils and foundations, 49(6), 897-908.


149

CHAPTER 4

AASHTO (2014). AASHTO LRFD Bridge Design Specifications, Customary U.S. Units,

6th Edition, American Association of State Highway and Transportation Officials

Allen, T. M., Nowak, A. S., and Bathurst, R. J. (2005). “Calibration to Determine Load

and Resistance Factors for Geotechnical and Structural Design.” Transportation Research

Circular E-C079 Transportation Research Board, Washington, D.C.

Barker, R. M., Duncan, J. M., Rojiani, K. B., Ooi, P. S. K., Tan, C. K., and Kim, S. G.

(1991), NCHRP Report No. 343 Manuals for the Design of Bridge Foundations.

Washington, D.C., Transportation Research Board, National Research Council.

Baecher, G. B., & Christian, J. T. (2005). Reliability and Statistics in Geotechnical

Engineering John Wiley & Sons, New Jersey

Becker, D. E. “Eighteenth Canadian Geotechnical Colloquium: Limit States Design for

Foundations. Parts I and II. An Overview of the Foundation Design Process.” Canadian

Geotechnical Journal, Vol. 33, 1996, pp. 956-1007.

Burland, J. B., & Wroth, C. P. (1974, April). “Allowable and differential settlement of

structures, including damage and soil structure interaction”. Settlement of Structures,

Proceedings of the Conference of the British Geotechnical Society, Cambridge. Pentech

Press, London, UK (pp. 611-764).


150

Duncan, J. M. (2000). “Factors of Safety and Reliability in Geotechnical Engineering”.

Journal of Geotechnical and Geoenvironmental Engineering, 126(4), pp. 307-316.

FHWA. (2011). “Load and Resistance Factor Design (LRFD) and Rating of Structures”.

(F. H. Administration, Producer) Retrieved March 19, 2012, from

http://www.fhwa.dot.gov/research/deployment/lrfd.cfm

Fenton, G. A., & Griffiths, D. V. (2007). “Reliability-based deep foundation design.

Probabilistic” Applications in Geotechnical Engineering, 170, 1-12.

Finny D.J. (1941) “On the Distribution of a Variate whose Logarithm is Normally

Distributed”. Journal of the Royal Statistics Society, Series B-7 PP. 155-161

Foye, K. C., Abou-Jaoude, G., Prezzi, M., & Salgado, R. (2009). “Resistance Factors for

use in Load and Resistance Factor Design of Driven Pipe Piles in Sands”. Journal of

Geotechnical and Geoenvironmental Engineering, 135(1), 1-13.

Grant, R., Christian, J. T. & Vanmarcke, E. H. (1974). “Differential settlement of buildings.

Journal of Geotechnical Engineering”. ASCE 100, No. 9, pp. 973–991.


151

Hamilton, J., and Murff, J. (1992). “Selection of LRFD Resistance Factors for Pile

Foundation Design.” Proceedings Structures Congress American Society of Civil

Engineers Structures Congress, San Antonio, Texas,

Honjo, Y., Yoshida, I., Suzuki, S., Hyodo, J., & Paikowsky, S. (2005). “Reliability

Analysis for Serviceability and the Problems of Code Calibration”. In

ERTC10/GeoTechNet WP2 Workshop, Evaluation of Eurocode7, Trinity College, Dublin.

Meyerhof, G. G. (1995). “Development of geotechnical limit state design”. Canadian

Geotechnical Journal, 32(1), pp. 128-136.

Misra, A., & Roberts, L. A. (2009). Service limit state resistance factors for drilled shafts.

Géotechnique 59.1 53-61. http://dx.doi.org/10.1680/geot.2008.3605

Moulton, L. K. (1985). Tolerable Movement Criteria for Highway Bridges, Report No.

FHWA/RD-85/107, Federal Highway Administration, Washington, DC

Paikowsky S.G and Lu, Y. (2006) “Establishing Serviceability Limit State in the Design

of Bridge Foundations”. Proceedings of Foundations Analysis and Design pp. 49-58

Paikowsky, S. G. (2004). NCHRP Report 507 Load and Resistance Factor Design (LRFD)

for Deep Foundations. Washington, DC: Transportation Research Board.


152

Phoon, K. and Kulhawy, F. (2003). “Development of a Reliability-Based Design

Framework for Transmission Line Structure Foundations”. Journal of Geotechnical and

Geoenvironmental Engineering 129.9 (2003): 798-806.

Roberts, L. A., Gardner, B. S., & Misra, A. (2008) “Multiple Resistance Factor

Methodology for Service Limit State Design of Deep Foundations using a ‘t-z’ Model

Approach”. Proceedings of the Geo-Congress 2008

Roberts, L., & Misra, A. (2010). “Performance-Based Design of Deep Foundation Systems

in Load and Resistance Factor Design Framework”. Transportation Research Record:

Journal of the Transportation Research Board, (2186), 29-37.

Salgado, R. (2008). The engineering of foundations. McGraw Hill, New York






Skempton, A. W., & MacDonald, D. H. (1956). “The allowable settlements of buildings”.

Proceedings of the Institution of Civil Engineers, 5(6), 727-768


153

Surles, J.G., H. Seo, W.D. Lawson, P.W. Jayawickrama, R.B. Moghaddam, J. Doerzbacher

(2016). “Sample Size Determination for Resistance Factors in LRFD.” Recommended for

publication 06/07/2016. International Journal of Geomechanics. ASCE.

Terzaghi K. (1947). Theoretical soil mechanics, John Wiley and Sons, Inc. New York,

London, Sydney

Texas Highway Department (1956). Foundation Exploration and Design Manual, Bridge

Division, Texas Highway Department


Transportation

Vu T.T. (2013) “Load and Resistance Factor Design of Drilled Shafts at the Service Limit

State”. Doctoral Dissertation, University of Missouri-Columbia

Wang, Yu, and Fred H. Kulhawy (2008). "Reliability index for serviceability limit state of

building foundations." Journal of Geotechnical and Geoenvironmental Engineering

134.11, 1587-1594.

Zhang, L. M., & Chu, L. F. (2005). “Calibration of methods for designing large-diameter

bored Piles: Serviceability Limit State”. Soils and foundations, 49(6), 897-908.

Evaluation of the TxDOT Texas Cone Penetration Test and ... · Evaluation of the TxDOT Texas Cone...

Documents

Transcript of Evaluation of the TxDOT Texas Cone Penetration Test and ... · Evaluation of the TxDOT Texas Cone...