You are on page 1of 267

Joint

Transportation

Research

Program

JTRP
FHWA/IN/JTRP-99/8

Final Report

PILE DESIGN BASED

ON CONE PENETRATION

TEST RESULTS

Rodrigo Salgado
Junhwan Lee

October 1999

Indiana

Department
of Transportation

Purdue
University

Final

Report

FHWA/IN/JTRP-99/8
PILE DESIGN BASED

ON CONE PENETRATION TEST RESULTS


Rodrigo Salgado
Principal Investigator

Junhwan Lee
Research Assistant

and
School of Civil Engineering

Purdue University
Joint Transportation

Project
File

Research Program

Number: C-36-450
Number: 6-18-14

Conducted in Cooperation with the


Indiana Department of Transportation
and the
Federal

The contents of this

Highway Administration

report reflect the views of the authors,

who

are responsible for

The contents do not


Federal Highway Administration and

the facts and the accuracy of the data presented herein.


necessarily reflect the views or policies of the

the Indiana Department of Transportation.

This report does not constitute a

standard, specification, or regulation.

Purdue University

West

Lafayette,

IN 47907

October 1999

Digitized by the Internet Archive


in

2011 with funding from

LYRASIS members and Sloan Foundation; Indiana Department

of Transportation

http://www.archive.org/details/piledesignbasedoOOsalg

TECHNICAL REPORT STANDARD TITLE PAGE


Report No.

1.

2.

Government Accession No.

3. Recipient's

Catalog No.

FHWA/LN/JTRP-99/8
4. Title

and

Subtitle

Report Date

5.

October 1999
Pile

7.

Designs Based on Cone Penetration Test Results

Authors)

6.

Performing Organization Code

8.

Pel-forming Organization Report No.

Rodrigo Salgado and Junhwan Lee

FHWA^N/JTRP-99/8
9.

Name and Address


Transportation Research Program

Performing Organization

Joint

10.

Work Unit No.

1284 Civil Engineering Building

Purdue University
West Lafayette, Indiana 47907-1284
Contract or Grant No.

11.

SPR-2142
12.

Name and Address

Sponsoring Agency

Type of Report and Period Covered

13.

Indiana Department of Transportation


Final

State Office Building

Report

100 North Senate Avenue


Indianapolis,

IN 46204
Sponsoring Agency Code

14.

15.

Supplementary Notes

Prepared in cooperation with the Indiana Department of Transportation and Federal


16.

Highway Administration.

Abstract

The bearing capacity of piles

consists

of both base resistance and side resistance.

maximum base resistance is reached. As the side resistance

is

The

side resistance of piles

is

in

most cases fully mobilized well before the


is a key element of

mobilized early in the loading process, the determination of pile base resistance

pile design.

cone penetration is well related to the pile loading process, since it is performed quasi -statically and resembles a scaled-down pile load test In order
advantage of the CPT for pile design, load-settlement curves of axially loaded piles bearing in sand were developed in terms of normalized base resistance
Although the limit state design concept for pile design has been used mostly with respect to either s/B = 5% or s/B =
(qv'qc) versus relative settlement (s/B).
10%, the normalized load-settlement curves obtained in this study allow determination of pile base resistance at any relative settlement level within the - 20%
Static

to take

range.

The normalized base resistance

for both non-displacement

and displacement

In order to obtain the pile base load-settlement relationship, a

3-D

piles

were addressed.

non-linear elastic-plastic constitutive

model was used

in finite element analyses.

The

3-

D non-linear elastic-plastic constitutive model takes advantage of the intrinsic and state soil variables that can be uniquely determined for a given soil type and
condition.

model.

A series of calibration chamber tests were modeled and analyzed using the finite element approach with the 3-D non-linear elastic-plastic stress-strain

The predicted

load-settlement curves

showed good agreement with measured load-settlement

curves.

Calibration chamber size effects were also

investigated for different relative densities and boundary conditions using the finite element analysis.

The value of the normalized base resistance q> q was not a constant, varying as a function of the relative density, the confining stress, and the coefficient of
pressure at rest The effect of relative density on the normalized base resistance qt/q c was most significant, while that of the confining stress at the
pile base level was small. At higher relative densities, the value of qb/q t was smaller (qtAfc = 0. 12 -0.13 for Dr = 90%) than at lower relative densities (qt/q< =
0. 19 - 0.2 for L\ = 30%).
The values of the normalized base resistance qt/q c for displacement piles are higher than those for non-displacement piles, being
typically in the 0.15 - 0.25 range for s/B = 5% and in the 0.22 - 0.35 range for s/B = 10%
(

lateral earth

The values of the normalized base resistance qjq, for silty sands are in the 0. 12 - 0. 17 range, depending on the relative density and the confining stress at
The confining stress is another important factor that influences the value of qi/q, for silty sands. For lower relative density, the value of q^qj

the pile base level.

decreases as the pile length increases while that for higher relative density increases.

For effective use of CPT-based pile design methods in practice, the method proposed in this study and some other existing methods reviewed in this study
were coded in a FORTRAN DLL with a window-based interface. This program can be used in practice to estimate pile load capacity for a variety of pile and
soil conditions with relatively easy input and output of desired data.
17.

Keywords

piles,

18. Distribution

sands,

constitutive

cone

model,

design, calibration

penetration
finite

element

bearing

analysis,

capacity,

No restrictions.

This document

is

available to the public through the

National Technical Information Service, Springfield,


limit

states

VA 22161

chamber test.

19. Security Classif. (of this report)

Unclassified

Form DOT F 1700.7 (8-69)

test,

Statement

20. Security Classif. (ofthis page)

Unclassified

21. No. of Pages

249

22. Price

TABLE OF CONTENTS

Page

TABLE OF CONTENTS

LIST

OF TABLES

LIST

OF FIGURES

vii

IMPLEMENTATION REPORT

CHAPTER

INTRODUCTION

Background

1.2

Statement of Problem

1.3

Objective and Scope

Report Outline

1.1

1.4

CHAPTER

2 PILE

DESIGN BASED ON IN-SITU TEST RESULTS

2.1

Introduction

2.2

Estimation of Pile Load Capacity Based on

2.3

Meyerhofs method

10

Aoki and Velloso's method

11

2.2.3

Reese and O'Neill's method

12

2.2.4

Briaud and Tucker's method

14

2.2.5

Neely's method

Estimation of Pile Load Capacity Based on


2.3.1
The Dutch method
Schmertmann's method
2.3.2
Aoki and Velloso's method
2.3.3

15

CPT Results

LCPC method

Summary
3

17

18

20
22
22
27

METHODS OF INTERPRETATION OF LOAD-SETTLEMENT


CURVES

3.2

Results

2.2.2

CHAPTER

3.1

SPT

2.2.1

2.3.4

2.4

28

Introduction
Interpretation

28

Methods

29

3.3

3.4

3.5

3.2.1

90% and 80% methods

29

3.2.2

Butler and Hoy's method

31

3.2.3

Chin's method

31

3.2.4

Davisson's method

33

3.2.5

De

35

3.2.6

Permanent

Beer's method
set

method

35

Limit States Design

37

3.3.1

Limit states design in Eurocode 7

3.3.2

Limit states design for pile foundations

37

39

Tolerable Settlements for Buildings and Bridge Foundations


3.4.1

Buildings

3.4.2

Bridges

43

43
47

Summary

51

CHAPTER 4 MECHANICAL BEHAVIOR OF SAND

53

4.1

Introduction

53

4.2

Stress Tensor and Invariants

54

4.3

Elastic Stress-Strain Relationship

4.4

Elastic Behavior of Soil

60
67
67

4.5

4.6

modulus

4.4.1

Initial elastic

4.4.2

Hyperbolic stress-strain relationship

71

4.4.3

Degradation of Elastic Modulus

75

at

small strain

Failure Criterion and Soil Plasticity

76
76

4.5.1

Failure criterion

4.5.2

Flow

4.5.3

Soil dilatancy

rule

and

stress

and

hardening

critical state

79
of sand

82

Summary

CHPATER

3-D

85

NON-LINEAR ELASTIC-PLASTIC STRESS-STRArN MODEL

87

5.1

Introduction

87

5.2

Intrinsic

and State Soil Variables


Modified Hyperbolic Model for Non-linear Elasticity
Non-Linear Elastic Model for Three Dimensions

87

5.3

5.4

5.5

5.6

5.4.1

Modified hyperbolic

5.4.2

Variation of bulk modulus and Poisson's ratio

5.4.3

Determination of the parameters f and g

stress-strain relationship for three

Plastic Stress-Strain Relationship for

Three Dimensions

90
95
dimensions

..

95

99
101

118

5.5.1

Drucker-Prager failure criterion

5.5.2

Non-linear failure surface and flow rule

120

5.5.3

Incremental stress-strain relationship

121

Summary

118

126

Ill

CHAPTER

NUMERICAL ANALYSIS AND EXPERIMENTAL


INVESTIGATION OF CALIBRATION

6.1

6.2

CHAMBER TESTS

Introduction
Calibration

129

Chamber

Plate

Load Tests

129

6.2.1

Description of test and experimental procedures

6.2.2

Test material and boundary conditions for calibration chamber

129

plate load tests

6.3

6.4

6.5

131

Numerical Modeling of Plate Load Tests

in Calibration

Chambers

ABAQUS

6.3.1

Program

6.3.2

Finite element

Calibration

Chamber

136
136

modeling of plate load test


Predicted and measured plate resistance

6.3.3

129

Size Effects on Plate

137
139

Load Test Results

155

6.4.1

Definition of size effect

155

6.4.2

Investigation of size effects for different boundary conditions

155

Summary

161

CHAPTER 7 DETERMINATION OF PILE BASE RESISTANCE

163

7.1

Introduction

163

7.2

Methods for Investigating Load-Settlement Response


Finite Element Modeling of Pile Load Test
Cone Penetration Resistance from Cavity Expansion Analysis
Determination of Base Resistance for Non-Displacement Piles

7.3

7.4
7.5

7.5.1

Load-settlement response for various

7.5.2

Normalized base resistance for non-displacement

7.5.3

The

soil

conditions

Kq

effect of initial stress ratio

7.6

Determination of Base Resistance for Displacement Piles

7.7

Normalized Base Resistance for

7.8

Summary

CHAPTER

ASSESSMENT OF PROPOSED NORMALIZED BASE


RESISTANCE VALUES BASED ON CASE HISTORIES

Introduction

8.2

Non-Displacement Piles

8.4

Sands

174
181

187

189
191

198

198

8.2.1

Georgia Tech load

8.2.2

Sao Paulo load

8.2.3

Simonini's results

8.2.4

Calibration

200
200

test

200

test

chamber

201
plate load tests

Displacement Piles
8.3.1

Purdue university load

8.3.2

NGI

Summary

167

174

196

8.1

8.3

Silty

piles

164
165

load tests

201

202
test

202
204
205

IV

CHAPTER 9

PILE DESIGN USING CPT RESULTS

9.1

Introduction

9.2

Determination of Base and Shaft Resistance


9.2.1

Base resistance

9.2.2

Shaft resistance

9.2.3

Factor of safety

9.4

Use of SPT Blow Counts


Program CONEPILE

9.5

Summary

9.3

CHAPTER
10.1

10

Summary
Recommendations

206
206
206
209
211
214
219
223

CPT-based Method

SUMMARY, CONCLUSIONS AND RECOMMENDATIONS

10.2 Conclusions
10.3

in

206

..

224
224
226
228

LIST OF REFERENCES

229

APPENDIX

242

LIST

OF TABLES

Page

Table

K and a for different soil

2.1

Values of

2.2

Values of F] and F 2 for different

2.3

Values of correlation factor

2.4

Values of the factor

types

13

pile types

for the

13

Dutch method

18

by Schmertmann (1978)
different soil and pile types

20

and

25

c sf

2.5

Values of k for

2.6

Values of

3.1

Relationship between angular distortion and total settlement (after

3.2

Skepmton and MacDonald 1956)


Tolerable movement for buildings

3.3

Settlement criteria for bridges expressed in terms of settlement magnitude

44
45
49

3.4

Tolerable angular distortion for bridge by Moulton

50

3.5

Data used by Moulton

kc for different soil

et al.

24

pile types

(after

Eurocode

1)

et al.

(1985)

(1985) to establish criteria for angular

50

distortion
4.1

Relationship between different elastic modulus

4.2

Values of
Salgado

Cg

66

e g and n g for different sand type (after Salgado 1993,


,

72

al 1999). ...~

et

5.1

Basic properties of Ticino sand (after Ghionna

5.2

Values of

5.3

Values of

6.1

6.6

Soil and stress conditions in calibration chamber tests


Boundary conditions in calibration chamber tests
Size effect in calibration chamber test for BC1 condition
Size effect in calibration chamber test for BC2 condition
Size effect in calibration chamber test for BC3 condition
Size effect in calibration chamber test for BC4 condition

7.1

Basic

7.2

Pile

7.3

Values of qt/q c according to several authors


Values of qt>/q c at s/B = 5% and 10%

6.2
6.3

6.4
6.5

7.4

1994)

and g from triaxial test results


and g for different relative densities

soil properties

geometry and

7.6

Base resistance
Values of qt/q c

7.7

Values of

7.5

et al.

used

soil

in finite

ratio for

117
1

element analysis

conditions used in

FEM

analysis

133

134
157
157
158
158
175
184

184

displacement and non-displacement piles

parameters with different

17

166

190
190

for displacement piles

soil intrinsic

102

silt

contents (after Salgado

VI

et al.

7.8

192

1999, Bandini 1999)

Values of

R with

friction angle

different

silt

7.9

7.10

Values of

qt>/q c

contents (after

and g used

Values of

<(>

and dilatancy parameters Q and


Salgado et al. 1999, Bandini 1999)

at critical state

in finite

element analyses for

silty

sands

for silty sands with different relative densities

9.

193

and
194

pile lengths
8.

192

Values of qb/q c from load


displacement piles

tests

on non-displacement and
199

Resistance modification factor fp and factor of safety for different


field tests (after

Canadian Geotechnical Society 1992)

9.2

Partial factor of safety for the base resistance

9.3

Correlation between

9.4

Correlation parameters for estimation of relative density

CPT

and

SPT

212
214
218
221

Vll

LIST

OF FIGURES

Page

Figure
1.1

Research scope and process

2.1

2.4

Examples of methods for estimation of pile bearing capacity


Dutch method for determination of base resistance
Reduction factor in Schmertmann's method (1978)
Equivalent cone resistance q ca for LCPC method

3.1

Definition of failure load in

3.2

Brinch Hansen's

3.3

Definition of failure load in Butler and Hoy's criterion

32

3.4

Chin's criterion for definition of failure load

32

3.5

Definition of failure load in Davisson's criterion

34

2.2
2.3

80%

90%

19

21

26
30

criterion

30

criterion

DeBeer's criterion

36

3.6

Definition of failure load in

3.7

Definition of failure load in permanent set criterion

36

3.8

Load

38

3.9

Differential settlements for (a) smaller-diameter and


(b) larger-diameter piles

41

3.10
3.11

Load-settlement curves with load versus s R (after Franke 1991)


Settlement criteria (after Wahls 1994)

42
46

3.12

Components of

levels at ultimate

and serviceability limit

settlement and angular distortion in bridge for (a) uniform,

settlement, (b) uniform

and

(d)

states

nonuniform

tilt

or rotation, (c) nonuniform regular settlement,

irregular settlement (after

4.1

Nine components of

4.2

Definition of mechanical behavior of a body

4.3

Non-linear stress-strain behavior of

stress tensor in a soil

Duncan and Tan 1991)

element
(after

48
55

Chen and Han 1988)

..

61

68

soil

69

4.4

Seismic cone penetration

4.5

Hyperbolic model

4.6

Stress states for elastic-plastic material

4.7

Stress-strain behavior for hardening, perfectly plastic

4.8

Different behavior of dense and loose sand (after

5.1

Secant modulus for non-linear stress-strain behavior

5.2

Modulus degradation relationship for normally consolidated sand

test

(a) stress-strain

.......

curve and (b) linear representation

73

78
and
81

softening material

Lambe and Whitmann 1986)83


89

vin

(after

Teachavorasinskun

91

1991)

et al.

5.3

Modulus degradation curve

5.4

Definition of x

5.5

Modulus degradation curves


f = 0.97 and g = 0.18

for

5.6

Modulus degradation curves


f = 0.97 and g = 0.15

for

D R = 48.8%

and o 3 = 200 kPa with

5.7

Modulus degradation curves


f = 0.97 and g = 0.18

for

D R = 48.2%

and o 3 = 500 kPa with

5.8

Modulus degradation curves


f = 0.97 and g = 0.20

for

DR =

Modulus degradation curves


f = 0.93 and g = 0.20

for

5.10

Modulus degradation curves


f = 0.95 and g = 0.25

for

D R = 82.3% and G

5.11

Modulus degradation curves


f = 0.95 and g = 0.20

for

DR =

5.12

Modulus degradation curves


f = 0.95 and g = 0.20

for

Dr = 91.1% and a 3 = 150 kPa with

Modulus degradation curves


f = 0.95 and g = 0.25

for

DR

Modulus degradation curves


f = 0.95 andg = 0.27

for

5.15

Modulus degradation curves


f = 0.94 and g = 0.32

for

DR

5.16

Modulus degradation curves


f = 0.94 and g = 0.28

for

D R = 100%

for

DR

5.9

5.13

5.14

5.17

x,

for different values of f

and Xmax for

Modulus degradation curves


f = 0.94 and g = 0.20

(a)

and g

93

constant and (b) varying confinement

D R = 51.5% and

rj 3

= 400 kPa

with

104
105

106

50.8% and

o~ 3

10 kPa with

107

D R = 84.6% and a

= 650 kPa with


108

= 100 kPa with


109

88.9% and o 3 = 200 kPa with


110

Ill

= 100% and G 3 = 200 kPa with


1

D R = 100%

and a 3 = 400 kPa with

= 100% and o 3 = 600 kPa with


and

= 98.6% and

5.19

Plastic strain in

ct 3

(a) in

15

119
Drucker-Prager

failure criterion with associated

5.20

Non-linear failure surface with non-associated flow rule


Plate load test in calibration

6.2

Types of boundary conditions in calibration chamber test


Finite element model for calibration chamber plate load test

6.7

h--sjJ\ plane and (b) in principal

6.1

6.6

14

= 100 kPa with

Stress plane

6.5

a 3 = 800 kPa with

116

Drucker-Prager failure surface

6.4

12

113

5.18

6.3

96

flow rule

chamber

Deformed finite element mesh of calibration chamber plate load test


with D R = 55.2%, o' v = 62.0 kPa, and rj' h = 24.4 kPa at s/B = 10%
Vertical stress distribution in calibration chamber plate load test
with D R = 55.2%, o' v = 62.0 kPa, and a' h = 24.4 kPa at s/B = 10%
Vertical displacement distribution in calibration chamber plate load test
with D R = 55.2%, a' v = 62.0 kPa, and & h = 24.4 kPa at s/B = 10%
Horizontal displacement distribution in calibration chamber plate load test

122
122

130
135

138

141

142
143

IX

with

D R = 55.2%,

a' v

= 62.0 kPa, and

a' h

= 24.4 kPa

at

s/B =

10%

6.8

Variation of shear modulus

6.9

Load-settlement curves for calibration chamber plate load

6.10

Load-settlement curves for calibration chamber plate load

6.

Load-settlement curves for calibration chamber plate load tests

144
145

tests

(Test No. 300, 301, and 302)

146
tests

(Test No. 303, 304, and 306)


1

147

(Test No. 307, 308, and 309)

148

6.12

Load-settlement curves for calibration chamber plate load

6.13

Load-settlement curves for calibration chamber plate load

6.14

Load-settlement curves for calibration chamber plate load

tests

(Test No. 310, 311, and 312)

149
tests

(Test No. 313, 314, and 317)

150
tests

(Test No. 321, 322, and 323)

6.15

151

Load-settlement curves for calibration chamber plate load tests


(Test No. 324, 325, and 326)

6.16

152

Load-settlement curves for calibration chamber plate load tests


(Test No. 327, 328, and 329)

153

Measured and predicted plate unit loads in calibration chamber tests


Comparison of pile base unit load with plate unit load in calibration
chamber plate load tests

154

6.18

7.1

Load-settlement curves for pile load

168

7.2

Different failure

7.3

Slip pattern under cone penetrometer (after Salgado 1993)

171

Stress rotation

between different zones (after Salgado 1993)


model for 5-m pile
Finite element model for 10-m pile
Finite element model for 20-m pile
Base load-settlement curves for (a) 5-m, (b) 10-m, and (c) 20-m
Normalized load-settlement curves for (a) 5-m, (b) 10-m, and

173

Finite element

176

6.17

7.4
7.5

7.6
7.7
7.8

7.9

for deep penetration

177

178
piles

base level and (b) relative d ensity (D R )

Ko on normalized base

7.11

Effect of

7.12

Values of qb/q c for

8.

Values of qb/q c
(b) s/B

silty

179

182

186
188

resistance qb/q c

sand

in calibration

159
170

piles in

at the pile

Georgia tech

terms of qt/q c and s/B


Normalized base resistance qb/q c with (a) mean effective stress (a' m )
(c)

7.10

20-m

mechanisms

test at

195

chamber

plate tests for (a) s/B

5%

and

= 10%

203

9.1

Estimation of pile base resistance using different methods

208

9.2

Estimation of pile shaft resistance using different methods

210

9.3

CPT-SPT

9.4
9.5

correlation with the

mean

grain size (after Robertson and

Campanella 1983)

215

Cone resistance q c and SPT blow count N with depth


Estimation of the base resistance for a given soil condition

217
222

IMPLEMENTATION REPORT

In the present study, in order to take advantage of the cone penetration test for pile
design, load-settlement curves in terms of normalized base resistance (q b/q c ) versus
relative settlement (s/B)

where q c = cone

diameter were developed.

resistance, s

Although the limit

been used mostly with respect

to either s/B

pile

state design

5%

or s/B

B =

base settlement,

pile

concept for pile design has

= 10%,

the normalized load-

settlement curves obtained in this study allow determination of pile base resistance for

any relative settlement

in the

of specific

consideration

components of the

facility,

- 20%

project

features,

by selecting a

This

range.
related

is

important, as

the

to

it

permits
or

superstructure

other

specific value of tolerable settlement for use in

design.

The value of

the normalized base resistance

is

qb /q c

not a constant, varying as a

function of the relative density, the confining stress, and the coefficient of lateral earth

pressure at

The

rest.

significant, while that

effect

of relative density on the normalized base resistance

of the confining

relative densities, the value

stress at the pile

DR =

obtain undisturbed granular soil samples.

estimation of the relative density be

displacement

5% and
CPT
if

pile design

therefore,

The evaluation of

90%) than

at

usually very difficult to

recommended

that the

reasonable correlations based on in-

typically in the 0.15

= 10%, depending on

used for a given

0.25 range for s/B

the value of relative density.

methods can be used with the SPT blow count


is

DR =

At higher

resistance q b /q c proposed in this study can also be used for

0.35 range for s/B

a proper value of q c/N

piles

is,

It is

small.

test.

The values of q b/q c were

piles.

in the 0.22

It

30%).

made through

such as the cone penetration

The normalized base

is

of q b/q c was smaller (q b /q c = 0.12-0.13 for

lower relative densities (q b/q c = 0.19-0.2 for

situ test results

base level

is

for practical purposes,

soil condition.

the relative settlement associated with the limit states design of

should be done with consideration of the type, functionality,

location,

and

XI

The

importance of the superstructure.

relative

settlement s/B of piles leading to

serviceability or ultimate limit states is usually in excess of

For implementation, use of the method

recommended.
participation of

This

may be done

INDOT

with the cone penetration


pile load test data

future

It is

on both driven

strongly

and shaft resistance developed

piles

and

drilled shafts

separate

at the interface

and cone penetration


located.

tests

also

projects

is

is

number of

be performed along

The

fully instrumented

of the pile with different

recommended

be such

that a

measurement of the base and the

separation of load capacity in base and shaft capacities


It is

piling

recommended

test at representative sites in Indiana.

would allow

of the design methods.

INDOT

as part of an implementation project with extensive

design personnel.

fully instrumented load tests

in

10%.

shaft resistance,

soil layers.

This

essential for further validation

that the selection

that sites with a variety

of

sites for pile load

of Indiana

soil

types be

Results for such sites would be useful for further validation of the CPT-based

pile design procedure for a

wide range of soils.

CHAPTER 1 INTRODUCTION

1.1

Background

With

the rapid growth of metropolitan areas, and fast industrialization resulting

from the fast-paced economic globalization, there has been a need


taller structures

on marginal

sites,

where surface

are usually not the best design solution.

soils are

to build heavier

weak and shallow foundations

At the same time, advances

in piling

technology

permit the installation of several types of piles, particularly non-displacement

lower costs than was possible


further

improvements

in

pile

and

piles, at

This in turn generates the motivation for

in the past.

design capability.

Additionally, there

is

growing

realization in the foundation engineering industry that certain types of deep foundation

(such as large-diameter drilled shafts) are conservatively designed (Harrop-Williams


1989, Hirany and
pile

Kulhawy 1989, De Mello and Aoki

1993).

In this context,

advances

in

design methods can have significant economic impact and should be actively

pursued.

Based on

the

method of

installation,

displacement or non-displacement

displacement
displacement

piles,

and

piles.

piles.

pile

foundations are classified as either

Driven piles are the most

drilled shafts (bored piles) are the

most

common

common

type of

type of non-

The load carrying capacity of both displacement and non-

displacement piles consists of two components: base resistance and side resistance.
side resistance of piles

resistance

is

is

in

most cases

reached (Franke 1993).

increment of axial load

is

fully

After

mobilized well before the


full

maximum

The
base

mobilization of side resistance, any

transferred fully to the base.

As

the side resistance

is

mobilized early

element

1.2

loading process, the determination of base resistance

in the

is

a key

in pile design.

Statement of Problem

Although
solely

friction piles are

on side resistance

to

sometimes used,

it

usually desirable to avoid relying

is

develop the needed pile load capacity.

placing the pile base on a bearing layer.

Physically,

axial load is the base resistance

layer, since the side resistance is fully

mobilized early

which

is

in the

number of

soil

immediately beneath the

different

developed

pile

in this

Plunging

loading process.

lateral

confinement imposed

to assess

deep bearing capacity

pile base.

methods have been proposed

based on the load-settlement curve obtained from pile load

Beer 1967; Chin 1970; Davisson 1972).


the limit state concept.

done by

the base unit load overcomes the limit base resistance q b i_,

dependent not only on the density but also on the

by the surrounding

when

is

what keeps a well designed

from plunging when acted upon by an

will take place only

This

One modern

tests

(Brinch Hansen 1963;

pile design

philosophy

is

De

based on

limit state is evaluated with respect to either a loss of

functionality or collapse of the superstructure and/or foundation (Franke 1990; Salgado

1995).

According

greater than
situations,

10%

to

Franke (1991), drilled shafts must typically undergo settlements

of the base diameter before a limit state

however, settlements

less

than

10%

it is

reached.

of the pile diameter

foundations or the supported structure to reach a limit


value of settlement leading to a limit state,

is

state.

In

some design

may

cause the

Irrespective of the specific

clearly necessary that a

methodology be

available to calculate the settlements caused by a given load and vice-versa in order to

design piles within the limit-state framework.

As

shaft resistance is mobilized early in

the loading process, the determination of the load-settlement relationship for the pile base
is

a key element in pile design and

is

the focus of this study.

Pile design in sands

have been mostly based on results of the

today widely recognized to have numerous limitations (Seed

serious limitation

is that its

is

significantly, with the operator

alternative for pile design is to base

it

to drive a

is

test

not

is

will also vary,

A much

and operating procedures.

on data from a cone penetration

is

Skemptonl986).

The SPT blow counts

a cylindrical penetrometer with a conical point

statically,

which

obtained based on a dynamic process, which

well related to the quasi-static pile loading process.

test,

et al.1985,

test,

main measurement, the number of blows required

sampler one foot into the ground,

sometimes

SPT

(CPT).

better
In this

pushed into the ground quasi-

and a number of measurements are made.

The CPT was invented

in

northern Europe precisely for the purpose of pile

foundation design (the test can be seen as a scaled-down load

been increasingly used

in

test

on a

pile),

and has since

Europe, and to a lesser extent in the Americas and Asia, for pile

design and other purposes.

Although there have been several proposed methods of

design based on cone penetration test results, an example of which

is

the

pile

LCPC method

(Bustamante and Gianeselli 1982), only recently has a method based on establishing the
relationship between cone tip resistance and the load-settlement relationship of a pile

been proposed (Ghionna

et al.

1993,1994; Salgado 1995).

If

such a relationship can be

established reliably for a variety of soil conditions, significant economies in materials

volume and

pile installation charges will

become

possible with respect to current design

methods.

1.3

Objective and Scope

The main

objective of this research

is to

develop the methodology to determine pile

load-carrying capacity based on the results of cone penetration

test.

The focus

will be

on

the design of piles used to support typical transportation structures, with focus on the

response of piles bearing on sandy soils to vertical loading.

In order to
finite

develop the load-settlement curves for a variety of

element modeling

strongly non-linear and

modeling of

soil

may

3-D

In general, the pile response to an external load is

used.

is

soil conditions,

involve large irreversible deformations.

For more

realistic

behavior around axially loaded piles, a non-linear elastic-plastic stress-

strain relationship will

be used

in the finite

The

element analyses.

calculated load-

settlement curves are normalized with the cone resistance q c and the pile diameter
the base resistance qb

and the settlement

s,

respectively.

The

fully

for

developed load-

settlement curves in terms of qb/q c versus s/B can be used to determine the normalized
pile

base resistance

analyses are

qt>/q c

compared with
Figure

field pile load tests.

1.4

for

any settlement-based design


results

1.1

criterion.

from calibration chamber

shows the general scope of

The

results of the

plate load tests

and

free

this study.

Report Outline

This Report consists often chapters, including this introduction.

Chapter 2 reviews the pile design methods based on in-situ

methods presented

in

test results.

The

Chapter 2 are based on the Standard Penetration Test (SPT) and

Cone Penetration Test (CPT).


Chapter 3 describes the methodology for defining a "failure" load for a pile from
load-settlement curve.

It

also introduces the limit state design concept with

its

main focus

on ultimate and serviceability limit states that are important in geotechnical engineering.
Tolerable settlements for different types of structures, including buildings and bridges are
discussed as well.

Chapter 4 covers the conceptual framework for describing the mechanical behavior
of

soils,

including

relationships,

stress

tensors,

and the concept of

are all discussed.

invariants,

plasticity.

linear

and

non-linear

Critical state, dilatancy

stress-strain

and shear strength

Constitutive

Modeling

Finite

Element

Analysis

Calibration

<C

Chamber Test

C>

Pile

Load Test

Investigation of

Size Effects

Pile

Load-

Determination of

Settlement Curves

Cone Resistance

Normalized LoadSettlement Relationship

Assessment of Results
Usins Field Test Data

Limit States

Design of Piles

Pile

Base

Pile Shaft

Resistance

Resistance

Determination of Pile

Load Capacity

Figure

Research scope and process

Chapter 5 presents the non-linear


concepts of intrinsic and state

elastic-plastic soil

model used

soil variables are explained.

in this study.

The

Values of the parameters

required for the non-linear soil model are presented based on experimental test results.

The Drucker-Prager

plastic

model for the

definition of post-failure behavior of the soil

is

described with incremental stress-strain relationship.

Chapter 6 presents the

chamber

plate load tests.

finite

The

element modeling and analysis of the calibration

analytical results are

of plate resistance in calibration chamber plate load

compared with the measured values

tests.

This chapter aims to verify the

accuracy of plate resistance predictions and assess the existence of chamber size effects

on plate resistance values.

Chapter 7 presents the determination of pile base resistance based on the normalized
load-settlement curves fully developed for axially loaded piles bearing in sand for a
variety of soil
lateral earth

and

stress conditions.

The

effects of relative density

and the coefficient of

pressure on the pile base resistance are explained.

Chapter 8 presents case histories for the validation of the results obtained
study.

The case

histories include both non-displacement

Chapter 9 discusses
Chapters

1-8.

correlations

In

pile design using

CPT

and displacement

in this

piles.

results in the light of the results of

order to present a more complete discussion on the subjects,

between SPT and

CPT

are also addressed.

for estimating pile load capacity in practice

is

computer program developed

briefly also introduced.

Chapter 10 presents the conclusions drawn from

this study.

CHAPTER 2 PILE DESIGN BASED ON IN-SITU TEST RESULTS

2.1

Introduction

In general the application of in-situ tests to pile design is

(1) Indirect

done through:

Methods

or
(2) Direct

Indirect

Methods

methods require the evaluation of the

internal friction angle

4>

soil characteristic

and the undrained shear strength

su,

parameters, such as the

from

in-situ test results.

This requires consideration of complicated boundary-value problems (Campanella


1989).
situ

test

On

the other hand, with direct methods, one can

the results

from

in-

measurements for the analysis and the design of foundations without the

The

evaluation of any soil characteristic parameter.


analysis and the design of foundations

empirical relationships.
indirect

make use of

et al.

and

is,

application of direct

methods

to the

however, usually based on empirical or semi-

Figure 2.1 shows

some examples of

the

methods available

for

direct approaches in different applications.

Indirect

methods for

pile design include

Vesic (1977), Coyle and Castello (1981),

and P method (Burland 1973) for cohesionless

method (Tomlinson 1971),


Focht 1972) for cohesive

(3

soil,

and

su

method (Burland 1973), and

soil.

Most

indirect pile design

A.

method (Bowles 1988), a


method (Vijayvergiya and

methods define the correlation

Estimation of Pile Bearing Capacity

Using In-Situ Tests

Indirect

Sandy

method

Qayey

soil

Direct

Base

Shaft

Base

SPT

Rock

soil

method

CPT

J-9su
|

a method

(3

method

A.

method

Vesic

Coyle and Castello

Shaft

(3

Figure

2.

method

Coyle and Castello

Examples of methods

for estimation of pile bearing capacity.

between the

factor

stress state

and base or

shaft resistance

based on the soil-strength

parameters.
Direct methods used for pile design have been mainly based on the standard

Although the

penetration test (SPT) and the cone penetration test (CPT).

used more extensively,


al.

it is

widely recognized that

Skempton 1986).

1985,

blow count)

is

purposes.

loading

if it

is

that its

not well related to the pile loading process.

In this test, a cylindrical

ground as

has a

serious limitation

vary depending on operation procedures.

were a scaled

number of

it

The CPT

is

has been

limitations (Seed et
(the

SPT

The SPT blow count can

also

main measurement

a superior test for pile design

penetrometer with a conical

pile load test.

SPT

tip is

In addition to the similarity

pushed into the

between the

pile

and cone penetration testing mechanisms, the possibility of simultaneous

measurement of shear wave velocities makes


subsurface

soils,

which may improve

possible to estimate elastic properties of

it

the quality of the design with

more accurate

in-situ

soil properties.

The main focus of


direct methods, the

study

is

cone penetration

the estimation of pile bearing resistance based


test in particular.

will

be reviewed.

Estimation of Pile Load Capacity Based on

SPT

Results

In

count

for pile design using the

most SPT methods, the

pile load capacities are defined in terms of the

and the correlation parameters.

These relationships are

SPT blow

typically of the

form

(Bandini and Salgado 1998):

qb

=n N b

qs=%n N
si

where

q b = base resistance

(2.1)

si

on

In this chapter, the existing

SPT and CPT,

methods

2.2

this

(2.2)

10

n b = factor to convert

SPT blow

count to base resistance

N b = representative N S pt value at the pile base level


q s = shaft resistance
n si = factor to convert

Sj

SPT blow

count to shaft resistance

representative Nspt value along the pile shaft in layer

For the computation of base resistance,

it

is

recommended

should represent the condition near the pile base.


representative

value have been proposed.

They

will

i.

that the

SPT

value

Different ways to define the

be discussed as the methods are

presented.

Meyerhofs method

2.2.1

Meyerhof (1976, 1983) proposed


based on

SPT

the following expressions for the base resistance

results:

for sands and gravels

qb
for nonplastic

=0AN1M jPa <4.0N^Pa

silts

^=0.4N160 ^Po
where

(2.3)

<3.0iV li60 P

(2.4)

fl

qb = base resistance
Ni,6o

= SPT

N value corrected for field procedures and overburden

stress

D = pile embedment depth


B = pile
Pa =

The upper
case of

diameter

reference stress

= 100 kPa =

0.1

MPa =

limits of base resistance given in (2.3)

D/B > 10

for sands

1 tsf.

and

(2.4) are always applied in

and gravels, and D/B > 7.5 for nonplastic

silts.

For pile

11

diameters within the range of 0.5

cm ~ 40

in.

= 3.28

ft.,

qb

is

< B/B R <

2,

where B R = reference length =

reduced using the factor

(B + 0.5B R Y
IB

where n =

1,2,

r b as

= 100

follows:

(2.5)

or 3 for loose, medium, or dense sand, respectively.

Meyerhof (1976,

1983) also proposed the expression of shaft resistance for small- and large-displacement
piles:

for small-displacement piles in cohesionless soil

N
<Is=
100

(2-6)

60

for large-displacement pile in cohesionless soil

.-*.

where

P.7)

q s = shaft resistance

N 6 o = SPT N value corrected for field procedures only.

2.2.2

Aoki and Velloso's method

Aoki and Velloso (1975) proposed the following formula for

qb

=^rN P
b

different soil

and

pile types:

(2.8)

12

q si

where

=^N

si

(2.9)

K = empirical factor in function of soil type


F] and

F2 =

empirical factors in function of pile type

a = shaft resistance

factor depending

on

soil

type

N b = average of the three N SPT values close to the pile base


N

si

= average

of Nsft values along the pile shaft in layer

Nb

those used to calculate

Pa =

The values of K,

reference stress

a, and Fi,

F2

i,

excluding

= 100 kPa =

0.1

MPa ~

and

are given in Table 2.1

1 tsf.

2.2, respectively.

Reese and O'Neill's method

2.2.3

Based on the observation of 41 loading

tests,

Reese and O'Neill (1989) proposed the

following SPT-based relationship for the base resistance of drilled shafts

embedded

in

sand:

qb

=0.6NP

where q b = base resistance; Pa = reference


value given in (2.10)

was

beyond

any of the loading

that value for

stress

selected because

< 45Pa

= 100 kPa =

(2.10)

0.1

MPa =

tsf.

The

limit

no ultimate bearing pressure was observed

test results.

In (2.10), the

SPT N

value should

be mean uncorrected value within a distance of two times the base diameter (B b ) below
the base of the drilled shaft.
In order to restrict the settlement of large-diameter shafts, they also suggested to use

a reduced value of the base resistance (q b

,r

as follows:

13

Table

Type of

2.

Values of

K and a for different soil

Soil

Sand
Silty

sand

Clayey

silty

sand

Clayey sand
Silty clayey

sand

Silt

Sandy

types.

(%)

10.0

1.4

8.0

2.0

7.0

2.4

6.0

3.0

5.0

2.8

4.0

3.0

5.5

2.2

4.5

2.8

2.3

3.4

2.5

3.0

Clay

2.0

6.0

Sandy clay
Sandy silty clay

3.5

2.4

3.0

2.8

Silty clay

2.2

4.0

3.3

3.0

silt

Clayey sandy
Clayey

silt

Sandy clayey

Silty

silt

silt

sandy clay

Table 2.2

Values of Fi and

F2

for different pile types.

F,

F2

Franki Piles

2.50

5.0

Steel Piles

1.75

3.5

Type of

Pile

Precast Concrete Piles

Bored

Piles

1.75

3.5

3.0-3.50

6.0-7.0

14

^=1-25
v

where q b r = reduced base


.

resistance;

for

qb

5*

BR =

According to Reese and O'Neill (1989),

Bb >

reference length
it

is

not

expressions for drilled shafts with a depth less than 15

24

in

(610 mm).

use of the

(3

1.25

For the shaft resistance of

BR

(2.11)

recommended
ft

(4.6

m)

~ 40
to

in.

= 3.28

ft.

use the above

or a diameter less than

drilled shaft in sands, they

recommend

the

method.

Briaud and Tucker's method

2.2.4

Based on the review and parametric study of 33 instrumented

pile load tests,

and Tucker (1984) developed a method for determining base (q b ) and


as a function of pile settlement

modeled
driving

as hyperbolic curves.

was

In this

(s).

method, both q b versus

Briaud

shaft (q s ) resistance

and q s versus

are

Residual stress resulting from rebounding after pile

also considered in that hyperbolic formula.

The hyperbolic equations

for

both base and shaft resistance are given by:

+q

Qs

Ts.max

where

tf=18684(tf)

-<3s.>

00065

(2.12)

(2.13)

"s.r

(2.14)
i

BR

15

q tm =19.75(N pl
Z

Ires

(2.16)

(2.17)

021
S!de )

m = 40

in.;

SPT blow

pile base;

and

Ap

r- Pa

and

A
s

(2.19)

reference stress

L, P, Ep, and

area, respectively;

average

Pa =

5irle

/A <

q s ,es=<lresAp

(2.18)

BJ
K

^.^=0.224(N

(2.20),

(2.15)

\K T P

K =200(N

Pa

= 5.57 LSlP
a

&=

In (2.12)

)- 36

(2.20)

= 100 kPa =

0.1

MPa ~

are the pile length, perimeter,

ltsf;

BR =

reference length

modulus and cross

are the pile base and shaft areas;

pt

is

section

the uncorrected

count within the zone from 4 diameters above to 4 diameters below the

side

is

the uncorrected average

SPT blow count

within the layer where

shaft resistance is considered.

Neely's method

2.2.5

Neely (1990, 1991) suggested new empirical relationships between SPT

and base resistance for expanded-base and auger-cast piles

in sands.

value

For expanded-base

piles such as a Franki pile, he pointed out that the ultimate base resistance of twice the

value for conventional driven piles, as suggested by Meyerhof (1956), would result in

overestimated value of base resistance, based on the observation of load tests on 93

expanded-base

piles.

It

was explained

that

the

overestimated base resistance for

expanded-base piles by Meyerhof s (1956) suggestion


effect.

The

piles

is

due to the ignorance of casing

having uncased and compacted concrete shaft create very large

16

pressures between the shaft and surrounding soil and

comparable

piles

having cased

show

greater load capacity than

shaft.

According to Neely (1990), the ultimate base resistance of expanded-base


sands can be given

piles in

as:

q b =0.2S-N

< 2.8-N-Pa

(2.21)

q b = base resistance

where

D = embedment depth of the maximum cross section of the base resistance


as the

sum

of the driven length and on-half the base diameter.

D b = diameter of expanded base


N = SPT N value
Pa = reference

The

limit value of

stress

= 100 kPa =

2.8N applies whenever the

0. 1

MPa ~

ratio of

1 tsf.

D/D b

is

For

greater than 10.

augered, cast-in-place (auger-cast) piles, the base resistance was suggested as follows

(Neely 1991):

qb

The

=\.9N-Pa

(2.22)

auger-cast piles are different from the conventional drilled shafts in terms of the

installation process.

The

auger-cast piles are installed supporting the side of augered

hole by the soil-filled auger without use of temporary casing or bentonite slurry.

From
shown

that

the

comparison with the

(2.22)

results

reasonable

resistance of auger-cast piles.

grain size

(D 50 )

increases,

results of loading test

It

was

on auger-cast

piles,

it

was

agreement with the measured value of base

also observed that the ratio qt/N increases as

and decreases as fines content increases (Neely 1989).

mean

17

2.3

CPT Results

Estimation of Pile Load Capacity Based on

Similarly to what
capacity based on

is

done

in the case of the

CPT results can

be expressed

qb

pile load

as:

=c b q

1s=Zc

SPT, the determination of

si

2 23 )
-

q ci

(2.24)

q b = base resistance

where

= empirical parameter

cb

q c = cone resistance

at

to convert q c to base resistance

the pile base level

q s = shaft resistance
c si

= empirical parameter

to convert

q c to shaft resistance

q C = representative cone resistance for layer


i

Values of c b and c si have been proposed mostly based on empirical correlations developed

between

and

pile load test results

different values of c b

and c si

CPT

Because different authors proposed

results.

the use of such parameters should be applied under

conditions similar to those under which they were determined (Bandini and Salgado
1998).

Although most expressions were based on cone resistance q c some authors (Price
,

and Wardle 1982, Schmertmann 1978) suggested the use of cone sleeve

friction fs for the

estimation of shaft resistance with the following general expression:

1s=c f
sfi

where

c s f,

is

a empirical parameter to convert cone sleeve friction to shaft resistance and

fS is a representative
i

(2.25)

si

cone sleeve

friction for layer

for the determination of pile load capacity using

i.

In this section,

CPT results

some of the methods

are described.

18

The Dutch method

2.3.1

In the

cohesionless

Dutch method (DeRuiter and Beringen 1979),


soil is

computed from the average cone

8B above and 4B below

a pile base,

where

Figure 2.2, the average cone resistance q c

determined

along

the path 'abed',

is

for

the pile diameter.

layer

'x'

is

Similarly, the average cone resistance q c2 for the layer

along the path 'efgh'.

The base

resistance q b

is

base resistance in

resistance q c between the depth of

the

which

in

pile

below

As can be seen

in

base

is

the

selected so as to

above the

pile

pile

minimize q c i.

base

is

calculated

then obtained from the average of q c] and

q c2 as follows:

2*L5*2 <l50P

(2.26)

q b = base resistance

where

w = correlation factor
q c = average cone resistance for the layer below pile base
i

q c2 = average cone resistance for the layer above pile base

Pa =

The values

reference stress

of correlation factor

= 100 kPa =

0.1

MPa ~

for the

1.0

Very gravelly coarse sand; sand with


Fine gravel; sand with

Dutch method.
Values of

Soil Condition

OCR =

tsf

for several soil conditions are given in Table 2.3.

Table 2.3 Values of correlation factor

Sand with

OCR = 6 to

OCR = 2 to 4

10

0.67

0.50

19

Pile-

D
8B

1'

B
4B

_L

Depth

Figure 2.2

Dutch method

for determination of base resistance.

20

Schmertmann's method

2.3.2

For the estimation of

base resistance in

pile

cohesive

stiff

soil,

Schmertmann (1978)

proposed the use of an average cone resistance with multiplying the reducing factor

shown

The average cone

in Figure 2.3.

8B above and 0.7B

to

described previously.

4B below

He

also

from the Dutch method by


For shaft resistance

resistance

is

a pile base in the

same way

recommended reducing

60%

in case of

in sand, the

calculated within a depth between


as in the

Dutch method

the base resistance that

is

using mechanical cone for a cohesive

obtained

soil.

following values of the shaft resistance factor c s of

(2.24) were proposed for different pile types:

cs

= 0.008

for

cs

= 0.012

for precast concrete and steel displacement piles

cs

= 0.018

open-end

for vibro

steel tube piles

and cast-in-place displacement

piles with steel

driving tube removed, and timber piles

According

to

Schmertmann (1978),

shaft resistance in cohesive soil.


friction to shaft resistance are

the cone sleeve friction fs can also be used to estimate

The values of

c s f of (2.25) that relates

given by Table 2.4 for displacement

cone sleeve

piles.

Table 2.4 Values of the factor c sf by Schmertmann (1978).

Values of c S f
fs/Pa

Concrete and Timber piles

Steel Piles

Pa =

0.25

0.97

0.97

0.50

0.70

0.76

0.75

0.48

0.58

0.88

0.40

0.52

1.00

0.36

0.47

1.50

0.27

0.43

2.00

0.20

0.40

reference stress

= 100 kPa =

0.1

MPa

tsf

21

Undrained shear strength (kPa)


50

100

150

200
i_

1000

2000

3000

Undrained shear strength

4000
(lb/ftA 2)

Figure 2.3 Reduction factor in Schmertmann's method (1978).

22

Aoki and Velloso's method

2.3.3

Based on the load

tests

and

CPT

results,

Aoki and Velloso (1975) proposed the

following relationship for both shaft and base resistance in terms of cone resistance qc

q s =--q c

where a, Fi and F 2 are the same empirical parameters

(2.28)

as

shown

in

Table 2.1 and

2.2.

LCPC method

2.3.4

From

number of load

tests

and

CPT

Bustamante and Gianeselli (1982) presented a


both pile and soil types.

method.

The

results for several pile

pile design

The method presented by them

basic formula for the

LCPC method can

method using
is

and

factors related to

often referred to as the

be written

as:

(2- 29 >

q,=-qe

(2.30)

kc

= base

resistance factor;

q ca = equivalent cone resistance

ks =

LCPC

K^

q>

where

types,

soil

at pile

base level;

shaft resistance factor;

q c = representative cone resistance for the corresponding layer

23

The

values of kc and k s depend on the nature of soil and

its

degree of compaction as

Tables 2.5 and 2.6 show the values of k s and kc with

well as the pile installation method.

different soil and pile types, respectively.

According to Bustamante and Gianeselli

(1982), the values of kc for driven piles cannot be directly applied to H-piles and tubular
piles with an

The

open base without proper investigation of

equivalent cone resistance q ca used in (2.29) represents an arithmetical

measured along the distance equal

the cone resistance

where B =

base,

The curve of

to 1.5B

above and below the

for determining

is

smoothened

pile

q ca consists of the

in order to eliminate local

raw curve.

Beginning with the smoothened curve, the mean cone resistance q cm of smoothened
resistance between the distance equal to 1.5B above and

(3)

mean of

2.4):

the cone resistance q c

irregularities of the

(2)

The procedure

pile diameter.

following steps (see also Figure

(1)

full-scale load tests.

The

equivalent cone resistance q ca

smoothened curve

at

pile base is obtained.

calculated as the average after clipping the

is

0.7q cm to 1.3q cm

below

This clipping

is

carried out for the values

higher than 1.3 q cm below the pile base, and the values higher than 1.3 q cm and

lower than 0.7q cm above the

In the
resistance.

LCPC

pile base.

method, separate factors of safety are applied

factor of safety equal to 2 for shaft resistance

were considered, so

Qw

Q w = allowable load
S

limit shaft load

b
QL =

limit base load

QL

and base

and 3 for base resistance

that the carrying load is given by:

where

to the shaft

Ob
(2 31)
'

24

Table 2.5

Values of k s for different

soil

and

pile types.

Maximum q

Value of k s
Nature of Soil

Soft clay and

q c /Pa

mud

Moderately compact

Silt

and loose sand

Compact

to stiff clay

mA mB

iia

iib

<10

30

30

30

30

0.15

0.15 0.15 0.15 0.35

10 to

40

80

40

80

0.35

0.35 0.35 0.35

0.8

<1.2

(0.8) (0.8) (0.8)

<50

60

150

60

120

0.35

0.35 0.35 0.35

0.8

>50

60

120

60

120

0.35

0.35 0.35 0.35

0.8

<50

100 120 100 120

0.35

50 to
120

100 200 100 200

0.8

sand and gravel

Weathered

>50

60

Moderately compact

<2.0

0.35 0.35 0.35

0.8

0.35

0.8

1.2

<2.0

1.2

1.5

<2.0

1.2

1.5

<2.0

0.8

(1.2) (0.8) (1.2)

60

80

80

fragmented chalk
to very

HA HB

Pa

(0.8) (0.8) (0.8)

Soft chalk

Compact

IB

and compact chalk

to

IA

50

clay

Type
IA
IB

1.2

0.8

1.2

(1.5) (1.2) (1.5)

>120

150 300 150 200

compact sand and

1.2

0.8

1.2

(1.5) (1.2) (1.5)

gravel

Pa =

reference stress

Type

= 100 kPa =

IA: Plain bored piles,

0.

mud

MPa =

bored

tsf

piles,

hollow auger bored

piles, cast

barrettes,

and micropiles installed with low injection pressure.

Type

IB:

Bored

Type

IIA: Driven or jacked precast piles and prestressed concrete piles.

Type

IIB: Driven or jacked steel piles.

Type

IIIA:

Type

MB:

piles with steel casing

and driven cast

Driven grouted piles and driven rammed

High pressure grouted

with high injection pressure.

piles with

screwed

piles, piers,

piles.

piles.

diameter greater than 250

mm

and micropiles

installed

25

Table 2.6

Values of

kc for different soil

and

pile types.

Value of
Nature of Soil

Group
Soft clay and

mud

Moderately compact clay


Silt

to stiff clay

and compact

0.50

10 to 50

0.35

0.45

< 50

0.40

0.50

> 50

0.45

0.55

< 50

0.20

0.30

0.40

0.50

> 50

0.20

0.40

120

0.30

0.40

Soft chalk

Weathered

Compact

Pa =

Group

to

to very

= 100 kPa =

0.

Plain bored piles, cased

barrettes, micropiles installed with

Group

II:

to

compact sand and gravel

reference stress
I:

50

fragmented chalk

MPa =

bored

Group

0.40

silt

Moderately compact sand and gravel

< 10

and loose sand

Compact

kc

/Pa

120

1 tsf.

piles,

mud

bored

piles,

hollow auger bored

low injectionpressure.

Driven cast-in-place piles and piles

II

in

Type

IIA, ID3, IIIA, and IIIB of Table 2.5.

piles, piers,

26

= 1.5B

Depth
ica

Figure 2.4 Equivalent cone resistance q ca for

LCPC

method.

27

2.4

Summary

Pile design

indirect

methods using

in-situ test results

and direct methods.

can be classified

methods

two categories,

methods require the evaluation of the

Indirect

On

characteristic parameters in the estimation of pile bearing capacity.

the direct

in

soil

the other hand,

utilize the in-situ test results directly in the estimation of pile

bearing

capacity.

Direct methods have been based mainly on the standard penetration test (SPT) and
the cone penetration test (CPT).
pile bearing capacity to the

In

SPT methods, most proposed

SPT blow

count

expressions relate the

and correlation parameters.

These

methods include Meyerhof (1976, 1983), Aoki-Velloso (1975), Reese and O'Neil (1989),
and Neely (1990, 1991), while Briaud and Tucker (1984) presented a hyperbolic formula
for the base

and shaft resistance as a function of

evaluation of equation parameters.

It

pile settlement with

should be noticed

been developed under different conditions, including

soil

SPT N

that, since

and

value in the

every method has

pile type, the consideration

of such factors must be taken into account for the selection of pile design methods.

The cone penetration

test is

regarded as a better alternative to the

reflects well the vertical pile loading

the

mechanism.

SPT because

it

The widely used CPT methods include

Dutch method (DeRuiter and Beringen 1979), Schmertmann (1978), Aoki-Velloso

(1975), and the

LCPC method

(Bustamante and Gianeselli 1982).

Most CPT methods

cone penetration resistance q c using empirical

relate the base

and

shaft resistance to the

parameters.

The

empirical parameters relating pile resistance to q c are given as a

function of soil and pile type.

The

LCPC method

(Bustamante and Gianeselli 1982)

provides relatively detailed information regarding soil and pile types.

Some

authors

propose the use of cone sleeve friction fs for the estimation of shaft resistance, while
others propose that

it

be done on the cone penetration resistance q c

28

CHAPTER 3 METHODS OF INTERPRETATION OF


LOAD-SETTLEMENT CURVES

3.1

Introduction

The

pile load-settlement curve obtained

indication of pile load-carrying capacity.


criterion that

In general,

test

provides an important

however, there

can clearly define a "failure load" or "bearing capacity" of a

load-settlement curve.
pile

from a load

is

pile

et al. 1981).

selected for interpreting a load-settlement curve should account for the

characteristics of the load-settlement curve

For geotechnical structures


satisfy certain

based on a

Although several methods for evaluating the "failure" load of a

have been proposed, they produce a very wide range of results (Horvitz

The approach

no unique

to

and the

perform "properly",

fundamental requirements.

Whenever

thereof fails to satisfy a performance criterion,


Potentially, the

number of

soil condition.

it is

it is

necessary that the structure

a geotechnical structure or a part

said to have reached a "limit state".

limit state events is infinite.

It is

therefore necessary that the

consideration of limit state events be reduced to a relatively small

number of

events in order to achieve a balance between the needs for safety and

economy

in

critical

design

(Bolton 1989).
In this chapter,

some of

the methods proposed for interpretation of pile load-

settlement curves will be reviewed.

Additionally two important types of limit states in

geotechnical engineering and tolerable settlement for different types of structures are
discussed.

29

3.2

Interpretation

3.2.

90%

and

Methods

80% methods

The 90% method was proposed by Brinch Hansen

90

(1963). In the

% method, the

failure point in the load-settlement curve is defined as the load that causes a settlement

twice as large as that caused by

of this method

is to

90%

The main goal

of that same load (see Figure 3.1).

define the failure load as the point from which significant change in

the rate of displacement to load increment occurs.

For the application

Hansen

also suggested

to both the quick

80%

method.

and slow maintained

In this

method, the failure load

80%

load that produces four times the strain caused by

load (Q f) and corresponding settlement

(s f) in the

Q=

load.

From

is

80% method

plotted in terms of yfs

the relationships between

is

Q versus

s,

Brinch

defined as the

The

failure

are defined based

on the

of the same load.

hyperbolic relationship of a transformed load-settlement curve.


the load-settlement curve

pile load tests,

As shown
where

0.8Q f versus 0.25s f and

in

Figure 3.2,

settlement and

versus

sf

through a

hyperbolic equation in Figure 3.2, the failure load (Q f) and corresponding settlement

can be obtained

as:

Qf

=^=

(3.1)

=^

(3.2)

in

which

respectively.

(Sf)

and

C2

are the slope and the intersection point of the curve in Figure 3.2,

30

2S

So.9

Figure 3.1

Definition of failure load in

90%

.9

method.

JL
Q

Js/Q = C s + C2
1

Figure 3.2

Brinch Hansen's

80%

method.

31

3.2.2

Butler and Hoy's method

Butler and
point of

two

Hoy

(1977) considered the failure load as the load that

lines tangent to the load-settlement

line is the initial straight line that

other line

is

curve, where

ton

9.8

curve

the intersection

One

at different points.

can be thought of as an elastic compression

tangent

line.

The

tangent to a point having a slope of 0.00125B R /Q r in the load-settlement

BR =

reference length

m = 40 in.

= 3.28

ft

and

Q =
r

reference load

kN.

Usually, the rebound portion of the load-settlement curve


the true elastic line.

rebound

is

Based on

this observation, Fellenius

compression

line as an elastic

is

more or

less parallel to

(1980) suggested the use of a

line instead of an initial straight line in Figure 3.3

for determining a "failure" load.

3.2.3

Chin's method

Based on the assumption

that the pile

load-settlement curve

is

approximately

hyperbolic, Chin (1970) proposed the following (Figure 3.4):

(1)

The

load-settlement curve

(2)

The

failure load

is

It

drawn

in

(Q f) or ultimate load

Chin's failure method


tests.

is

is

test.

(Quit) is

s;

defined as Qf

1/Cj.

applicable to both the quick and slow maintained load

may, however, provide an

not used in the pile load

terms of s/Q versus

unrealistic "failure load" if a constant time increment

Extrapolation for hyperbolic load-settlement curve also

requires that the load test be extended sufficiently far.

s
32

,,--

1/(0.00 125B R

1
1

IT
t

+*

B R = reference

length

reference load

f
Figure 3.3

Definition of failure load in Butler and Hoy's method.

s/Q
C, = 1/Qf

rf^
x^

s/Q = CjS +

C2

Qf=l/Ci
*-

Figure 3.4

Chin's method for definition of failure load.

33

method

3.2.4 Davisson's

In Davisson's

method (Davisson 1972),

the "failure" load

is

defined as the load

leading to a deformation equal to the summation of the pile elastic compression and a

This relationship

deformation equal to a percentage of the pile diameter.

where

Sf

ff

AE
L

0.0038lB R +
+ nnnoo,z>
,

is

given by:

^_

3.05

(3.3)

fl c

settlement at failure condition

Q = applied load
L = pile
E=

length

Young's modulus of pile

A = cross sectional
BR =

B=

As can be seen

area of pile

reference length

m = 40 in.

= 3.28

ft

pile diameter.

in Figure 3.5, the elastic

compression line of the pile can be obtained from

column which

the elastic deformation equation of a

is

given by a equation of

s d asuc

QUAE.
Since
data

this

method

is,

in general,

from quick maintained load

regarded conservative,

tests.

Due

to the

dynamic

it

appears to

effect, loads

the quick maintained load tests tend to be higher than loads obtained

maintained load

tests,

sometimes

(Coduto 1994).

best with

obtained from

from the slow

significantly so in clayey material (Fellenius 1975).

may, therefore, lead to overly conservative


maintained load

work

results

tests resulting in considerable

when

It

applied to data from the slow

underestimation of a pile failure load

34

Of

0.00381B r + B/(3.05Br)

Figure 3.5

Definition of failure load in Davisson's method.

35

De

3.2.5

Beer's method

De Beer

maximum
curve

is

(1967) defined the "failure load" as the load corresponding to the point of

curvature on the load-settlement curve.

shown

plotted using log-log scale as

in

In this

method, the load-settlement

Figure 3.6.

determined as the load corresponding to the point

at

The

which two

failure load is then

straight lines intersect.

This method was, however, originally proposed for the slow maintained pile load

3.2.6

Permanent

In the

specified

set

test.

method

permanent

set

method (Horvitz

amount of permanent deformation

et al. 1981), failure is defined

by a certain

that occurs after full load removal.

To

determine the failure load with this method, the value of the permanent settlement should

be predefined prior to performing a load

As shown

test.

in Figure 3.7, the failure load

corresponding to a specified permanent

settlement can be determined by conducting the load-rebound for each applied load.

permanent settlement appearing as a


plastic soil deformation.
this

The

result of unloading is generally associated with

This method does not provide one "failure" load, as "failure"

in

case depends on what level of permanent settlement the user associates with

"failure".

36

logQ M

-^~
logs

Figure 3.6

Definition of failure load in

De

Beer's method.

Specified settlement

Figure 3.7

Definition of failure load in permanent set method.

37

3.3

Limit States Design

In general, there are

and serviceability

when

of

loss

two types of

limit states

components of the

An

(Ovesen and Orr 1991).


severe

equilibrium,

static

limit states in civil engineering design: ultimate

structure occurs.

On

structural

is

reached

damage, or rupture of

critical

ultimate limit state

the other hand, a serviceability limit state

is

associated with loss of functionality of the structure, typically related to settlement,

deformation,

appearance, and comfort.

utility,

Figure 3.8 illustrates the characteristic

difference of the load levels between ultimate and serviceability limit states on the load-

deformation curve.
In practice,

design.

It is,

it is

usually difficult to determine which type of limit states governs

therefore, required that both the ultimate

and serviceability

limit states be

investigated.

3.3.1

Limit states design in Eurocode 7

The Eurocodes were


communities.

Design

Geotechnical design

is

problems are defined

established for

criteria in the

addressed

in

use of design codes in European

Eurocodes are based on the

Eurocode 7 (1993).

in order to establish

quality of geotechnical investigation

common

minimum

limit states concept.

Three categories of design

requirements for the extent and

and calculations as well as construction control-

checks (Franke 1990, Ovesen and Orr 1991).

The

factors taken into consideration for determination of the geotechnical categories

in each particular design situation are as follows:

(1)

Nature and size of the structure;

(2) Conditions related to the location of the structures (neighboring structures,


utilities,

vegetation, etc.);

38

Load

*
Ultimate limit state

Deformation

Figure 3.8

Load

levels at ultimate

and serviceability

limit states.

39

(3)

Ground conditions;

(4)

Groundwater conditions;

(5) Regional seismicity;


(6) Influence of the

environment (hydrology, surface water, subsidence, seasonal

changes of moisture,

Geotechnical Category

etc.)

This includes small and relative simple structures for which


that the

performance

criteria will

it is

possible to ensure

be satisfied on the basis of experience and qualitative

geotechnical investigations with no risk for property and

life.

Geotechnical Category 2

This category includes structures for which quantitative geotechnical data and
analysis are necessary to ensure that the performance criteria will be satisfied, but for

which conventional procedures of design and construction may be used.

These

necessitate the involvement of qualified engineers with relevant experience.

Geotechnical Category 3

This category includes

abnormal

very large or unusual

risks or unusual or exceptionally difficult

structures,

structures

involving

ground or loading conditions, or

structures in highly seismic areas.

3.3.2

Limit states design for pile foundations

The
Eurocode

limit states for a pile foundation


7, are

under axial loading condition, according to

defined as follows (Salgado 1995):

40

Bearing capacity

(IA)

failure of single pile,

which may correspond to either

(IA-1) a condition of large settlement for unchanged load on the pile;

(IA-2) the crushing or other

damage

to the pile

element

itself;

(EB)

Collapse or severe damage to the superstructure due to foundation movement;

(II)

Loss of functionality or serviceability of the superstructure due

to displacement of

the foundations;

Overall stability failure, consisting of the development of a failure mechanism

(III)

involving the pile foundation or a part thereof.

Limit state (HI)

a possibility for the case where the structure

is

is

located close to a

topographic discontinuity area such as a slope, river and retaining wall.

In

most

pile

design situations under axial loading condition, either limit state (U) or limit state (IB)

governs design.

This

is

so because either the serviceability or the stability of the

supported structure would be jeopardized before a given pile developed a classical


bearing capacity failure.
than that for limit state
It is

and

II

The settlement required

for the limit state (IB) is usually greater

(II).

necessary to establish tolerable settlements

(Franke 1991, 1993).

The

sib

tolerable settlements Sib

corresponding differential settlements Asib and Asuas the ratio of differential settlement

between them,

is

and

Sn

according to limit state IB

and

su

The angular

can be related to the

distortion (p), defined

between two adjacent columns

to the distance

often used to determine tolerable differential settlements.

Figure 3.9 shows the settlements

at the

foundation level with smaller- and larger-

As can be seen

diameter piles, caused by the superstructure.

in Figure 3.9, a s

and

ai

represent the distances between two adjacent piles, and As s ,max and Asimax are the

maximum
It

differential settlements for the smaller-

must be realized

that the distance

ai is

and larger-diameter

piles, respectively.

generally greater than a s because larger-

diameter piles are used to carry heavier axial loads with larger spans.
angular distortion

(3 max

for each case can be defined as:

The maximum

41

As.s. max

1.

(a)

Figure 3.9

Differential settlements for (a) smaller-diameter

max

(b)

and

(b) larger-diameter piles.

42

for smaller-diameter piles

"max

As max
^^s,
(3.4)
a.

As /.max
for larger-diameter piles

(3.5)

Irrespective of pile size, the


(3.5)

maximum

angular distortion

fimax

given in (3.4) and


This implies

cannot be larger than the value of the tolerable angular distortion.

that the tolerable differential settlement

As usually increases with the span.

tolerable settlement accordingly, also increases with span.

also increase with span, a

common way

Given

to define limit states for piles is to establish a


s to

the pile

According to Franke (1991), the relative settlement required to cause either

loss of functionality or collapse of the superstructure is larger than s R

Figure 3.10, by the time


mobilized.

total

that the pile diameters

value of tolerable relative settlement sr as a ratio of the pile settlement

diameter B.

The

sR

0.

As shown

reaches 0.1, the shaft resistance has already been fully

This implies that the evaluation of the base resistance

is

a key element in the

limit states design of piles.

Load

0.01-0.02

SR "

Figure 3.10

in

Load-settlement curves with load versus

s R (after

Franke 1991).

43

Tolerable Settlements for Buildings and Bridge Foundations

3.4

Buildings

3.4.1

Tolerable settlements for buildings have been extensively studied by several authors

(Skempton and MacDonald 1956, Polshin and Tokar 1957, Burland and Wroth 1974,

Wahls

1981).

movement

Skempton and MacDonald (1956) presented

for building structures based

on the observed settlements and damages of 98

These values of tolerable movement

buildings.

the values of tolerable

for building structures are

still

widely

Because most of the observed damage appeared

accepted as a satisfactory criterion.

to

be related to distortional deformation, the angular distortion (P) was selected as the
critical

to

(3

From

index of settlement.

= 1/300 was suggested

the field data, a limit value of angular distortion equal

for the

corresponding differential settlement for a typical span equal to 20


limit value of angular distortion that

be

p=

caused structural damage

histories,

suggested the correlation between the

settlement.

settlement

As would be

about 3/4

in.

The

frames was observed to

was smaller

maximum

angular distortion

for a raft than for isolated foundations.

maximum

These relationships can be given by

suggested by Skempton and

Pmax

Skempton and MacDonald (1956)

expected, the angular distortion for a given

that the angular distortion for a given

a sand.

in

ft is

1/150.

Based on the examination of the case


also

The

condition of cracking in panel walls.

MacDonald

= 255 R

25B r

Pmax

= 18.755 R

"

for a clay than for

(3.9) with the numerical factors

(1956):

for sand with isolated foundations

(3.7)

P^max

for clav with raft foundations

(3.8)

)3 max

for sand with isolated foundations

(3.9)

31

as well

(3.6)

Pv = 155 P /L av
=

(3.6)

maximum

was observed

was smaller

total

for clay with isolated foundations

/3 max

Pmax

settlement

It

and

44

p^ = the maximum settlement;

in

which

reference length

= 40

total

settlement suggested by

the

maximum

angular distortion; and

summarizes the

3.1

BR

limit values of total

representing the ratio of angular distortion to the

Skempton and MacDonald

were based on the tolerable angular


foundations and

ma x

Table

in.

settlement, and the correlation factor

(3

distortion of

The values of Table

(1956).

3.1

p = 1/300 for the different types of

soils.

Polshin and Tokar (1957) proposed separate tolerable settlement criteria for framed
structures

and load bearing walls.

distortion are similar to those

For framed

terms of deflection ratio A/L, where


points and
3.1

L=

shows

More
tolerable

the

of angular

by Skempton and MacDonald (1956), ranging from 1/500

For load bearing walls, the

to 1/200.

structures, the limit values

were suggested

limit values of angular distortion

A = maximum

from two reference

relative settlement

distance between them, depending on the length to height ratio.

common

criteria in the

found

The

Eurocodes.

in

Eurocodes are very similar to those by Skempton and

MacDonald (1956) and Polshin and Tokar


Eurocode

angular distortion in the Eurocode

tolerable settlements in

Figure

settlement criteria including the deflection ratio.

recently, guidelines for tolerable settlement can be

movement

(1957).

(1993).

It

Table 3.2 shows the values of

can be seen that the values of tolerable

appear to be more restrictive than those by Skempton

and Macdonald (1956) and Polshin and Tokar (1957).

Table 3.1 Relationship between angular distortion and


(after

Skempton and MacDonald

Pmax

BR =

reference length

1/31.25

0.075B R

Pmax

0.05B R

1956).

1/25

1/15

settlement

Raft foundation

Sand

total

Isolated foundation

Clay

in

m = 40 in. 3.28

ft.

0.075B R to0.125B R

1/18.75
a

0.05B R

to

0.075B R

45

Table 3.2

Tolerable

movement

for buildings (after

Eurocode

1).

Total settlement

mm

Isolated foundations

25

Raft foundations

50mm

Differential settlement

between adjacent columns

Open frame

Frames with flexible cladding or finishing

Frames with

Angular distortion

mm
10 mm
5 mm

20

rigid cladding or finishing

1/500

46

E
iHf

Pab

Pmax

AsabT

"^J!^^
(a) settlement

without

(b) settlement with

As AB =

differential settlement

tilt

tilt

between

A and B

(3 A b = angular distortion between A and B


Pmax = total settlement
L = distance between two reference points (A and E)
0) = tilt = rigid body rotation
A = relative deflection
= maximum displacement from a straight line

connecting two reference points (A and E)

A/L = deflection

Figure 3.11

ratio

Settlement criteria (after Wahls 1994).

47

Bridges

3.4.2

According to Walkinshaw (1978), the

tolerability of the

movement should be

assessed qualitatively by the agency responsible for each bridge using the following

"Movement

definition:

repairs

not tolerable

is

damage

if

and a more expensive construction

Although said of bridges,

As shown

this is a

Although
it

would have been preferable".

concept of general applicability.

tilt

(or rotation),

settlement represents a condition where

the bridge,

to avoid this

Figure 3.12, the settlement of a bridge can be divided into three

in

components: uniform settlement,

amount.

requires costly maintenance and/or

it

may

all

and nonuniform settlement.

foundation elements

settle

cause such problems as drainage and clearance

Uniform

by the same
at the

end of

does not cause significant distortion of the bridge superstructure (Stermac

1978, Yokel 1990,

Wahls

angular distortion.

This pattern of settlement

bridges with very

Uniform

1990).

stiff superstructures.

tilt

is

or rotation corresponds to a uniform

possible only for single-span bridges or

This can also cause some problems with the

approach slab, with drainage, and with clearance, while distortion effects are largely
absent in the superstructure.

Nonuniform

settlement

can

be

by

categorized

two

representative

types:

regular nonuniform [see Figure 3.12 (c)] and irregular [see Figure 3.12 (d)] nonuniform
If the

settlement.

cause

more

same amount of total settlement

is

considered, irregular settlement will

significant distortion of the superstructure than regular settlement,

to the greater differential settlement

the differential settlement

(Xanthakos 1995).

and angular distortion not

to

mainly due

Foundation design must control

exceed the limit

state.

Limit values of tolerable settlements for bridge have been proposed by several
investigators

(Walkinshaw 1978, Bozozuk 1978, Grover 1978, Wahls 1990).

measure of settlement or

damage, due

to the

distortion,

No

single

however, can be regarded as a sole indicator of bridge

complexity of settlement patterns.

tolerable bridge settlements in terms of the

Table 3.3 shows criteria of

magnitude of the settlement.

Although the

48

^
(a)

(b)

Angular distortion =

Figure 3.12 Components of settlement and angular distortion in bridge for (a) uniform
settlement, (b) uniform

and

(d)

nonuniform

tilt

or rotation, (c) nonuniform regular settlement,

irregular settlement (after

Duncan and Tan

1991).

49

values and degree of

damage were suggested by

different authors,

appears that they are

it

reasonably consistent.

As shown

in

mm was suggested
The upper limit value of approximately 100 mm was

Table 3.3, the smallest value of approximately 50

by Bozozuk (1978)

as not harmful.

suggested as the settlement that


findings by the Federal

may

cause some damage, yet remaining tolerable.

Highway Administration (FHWA) (Moulton

et al.

Later

1985, Moulton

1986) turned out to be reasonably consistent with the tolerable values given in Table 3.3.

They found
which the

269

vertical

50

less than

that the

mm,

movement could be regarded

movement was

100

less than

as tolerable for

mm

90%

of the cases for

and the horizontal movement was

based on the measurements of bridge movement for 439 abutment and

piers.

According to Moulton
establishing tolerable

et al.

movement magnitude

distortion for bridges, suggested


criteria

given

in

(1985), the angular distortion gives a rational basis for


for bridges.

by Moulton

The values of

et al. (1985), are

shown

in

tolerable angular

Table

3.4.

The

Table 3.4 were based on the observation for 56 simple span and 119

continuous span bridges.

Table 3.3 Settlement

criteria for bridges

expressed in terms of settlement magnitude.

Settlement
Basis for recommendation

Recommended by

51

Not harmful

Bozozuk (1978)

63

Ride quality

Walkinshaw (1978)

Structural distress

Walkinshaw (1978)

102

Ride quality and structural distress

Graver (1978)

102

Harmful but tolerable

Bozozuk (1978)

Usually intolerable

Wahls (1990)

Magnitude (mm)

>63

> 102

50

The

details considered in

are less

Table 3.4 are summarized

complex than buildings, the values shown

than those for buildings.

The

in

in

Table

Although most bridges

3.5.

Table 3.4 are just slightly higher

criteria for the tolerable angular distortion in

(1994) were also established based on the recommendations of Moulton et

Duncan and Tan

(1985).

(1991), however, indicated that the tolerable angular distortion for

single-span bridges, equal to

of data used by Moulton

(3

et al.

1/200,

may

be overly conservative based on the review

They recommended

(1985).

angular distortion for single-span bridge as

(3

the use of the tolerable

1/125.

Table 3.4 Tolerable angular distortion for bridges (Moulton

et al. 1985).

Angular distortion

Basis for recommendation

1/250

Tolerable for multi-span bridge

1/200

Tolerable for single-span bridge

Table 3.5

al.

AASHTO

Data used by Moulton

et al.

(1985) to establish criteria for angular distortion.

% of 199 continuous bridges


Angular distortion

which

amount of
angular distortion was
for

this

considered to be tolerable

% of 56 single-span bridges for


which this amount of angular
distortion

was considered

be tolerable

0.000-0.001

100%

98% (100%)

0.001-0.002

97%

98% (100%)

0.002 - 0.003

97%

98%(100%)

0.003 - 0.004

96%

98% (100%)

0.004 - 0.005

92%

98% (100%)

0.005 - 0.006

88%

96% (98%)

0.006 - 0.008

85%

93% (95%)

to

51

Summary

3.5

The concept of
Most

a "failure load" of an axially loaded pile

proposed

criteria

based on

to define a failure load

is

potentially controversial.

pile load-settlement

curves

associate failure with relatively dramatic changes in settlement increment for a given load

increment.

These methods include the

Butler and

Hoy method

method (Horvitz

90%

method (Brinch Hansen 1963),

criteria

(1977), the Davisson

method (1972) and

The 80% method (Brinch Hansen

et al. 1981).

the

the permanent set

1963), Chin's

method

(1970), and DeBeer's method (1967), on the other hand, define the failure conditions

from the transformed load-settlement curve using


Because

scale.

different

the selection of the

methods may produce widely

method

the

Whenever
of

soil condition.

modem

of

quality

foundations

and

criteria, it is said to

serviceability limit state is reached


limit state is reached

value of

superstructures

have reached a "limit

limit states in civil engineering design: ultimate

damage.

design approach, in which


is

considered.

a geotechnical structure or part of a geotechnical structure fails to satisfy one

performance

its

two

technical

and

been proposed as a

limit state concept has

adequate

different values of failure load,

for defining the failure condition should account for the

characteristic shape of load-settlement curve

The

either hyperbolic relationship or log-log

when

there

when
is

In general, there are

serviceability limit states.

a structure loses functionality, while an ultimate

a loss of static equilibrium or severe structural

Franke (1991) suggested that

10%

and

state".

if relative

settlement

s R is

kept below a limit

neither loss of functionality nor collapse of the superstructure will take

place.

The value
states

of tolerable settlement associated with serviceability and ultimate limit

may differ depending on

For building

the type, functionality and importance of a given structure.

structures, values of tolerable settlement

have been proposed mainly based

on the results proposed by Skempton and MacDonald (1956) and Polshin and Tokar
(1957).

The

results suggested

that the angular distortion varies

by these two authors are


from 1/150

to 1/500.

relatively consistent,

showing

52

For bridge structures, several authors proposed limit values for tolerable settlement.

No

single

measure of settlement or angular

indicator of bridge

damage due

distortion,

however, can be regarded as a sole

complex settlement

to the

patterns.

suggested the limit value of settlement equal to approximately 100


settlement for bridges.

This value

Administration (Moulton 1986).

is

Bozozuk (1978)

mm

as a tolerable

similar to that proposed by Federal

Highway

53

CHAPTER 4 MECHANICAL BEHAVIOR OF SAND

4.1

Introduction

The simplest way

to describe

mechanical behavior of a

In linear elasticity, the stress-strain relationship is represented

Although

of the definition of a failure condition.

convenient to use,
problems.

it is

may

soil

by

be linear

elasticity.

linearity in the

absence

linear elasticity is simple,

hence

usually not suitable for typical stress-strain ranges in geotechnical

Linear elastic behavior can be observed in

condition up to strains of the order of 10"

soil

from the

initially

unstrained

After this strain range, the soil usually shows

highly non-linear behavior.


In order to represent the non-linear stress-strain behavior of a soil before the failure

condition takes place, several soil models have been proposed (Kondner 1963,

Duncan

and Chang 1970, Hardin and Dmevich 1972, Fahey and Carter 1993, Tatsuoka
1993).

The hyperbolic

types of soil models are

representing the non-linear behavior of a

soil.

among

the

most popular

These types of

soil

soil

models

et al.

models for

are relatively

simple to use, and are reasonably accurate.

Fundamentally, the hyperbolic types of

models are based on the quasi-linear

model assuming piece wise

for each increment of stress


Stability

where large

and

elastic

from modest

stress increases.

usually associated with plastic response.

two

basic concepts,

i.e.

linear behavior

strain.

problems are concerned mostly with the portion of the

strains result

soil

stress-strain curve

These larae deformations are

Description of the plastic condition requires

definition of a failure criterion

and flow

rule.

defines a stress state in a soil that leads to a failure condition.

failure criterion

After the stress state

54

satisfies a failure criterion, the stress-strain relationship

The flow

linear elasticity or non-linear elasticity.

and

stress

strain in the plastic state

potential function is the

associated.

Otherwise,

the failure criterion, the soil

This chapter will

through the plastic potential function.

referred to as a non-associated flow rule.

may

start

rule defines the relationship

as the failure criterion, the flow rule

same

it is

can no longer be defined by either

4.2

Stress Tensor

As shown

in

is

After satisfying

stress tensor.

The

in the

Indicial notation, originally

mathematical treatment of these concepts.

and Invariants

Figure 4.1, there are nine components of stress

Cartesian coordinate system.

and

referred to as

stress-strain relationship of soil, including linear elastic,

non-linear elastic and plastic responses are then discussed.

adopted by Einstein, will be used

the plastic

harden, soften, or remain without any change in stress.

by describing the basic concept of the

fundamental ways to describe the

is

If

between

denoted by Oy.

The

The

stress tensor is a set

stress tensor Ojj

in a soil

element

in a

of these nine components of stress

can be expressed

in

matrix notation as

follows:

au

an

o\,

o\,

0\,

o.

o\,

a 33

Imposing moment equilibrium on the

stress tensor

13

(4.1)

a shown
X)

in (4.1),

it

follows that

it

symmetric:

u=j,
or
<7

U =

(7 21

(4.2)

is

55

2,y

'

o?->

l,x

Figure

4.

Nine components of

stress tensor in a soil element.

56

CT

(7

The

stress tensor

CTij

= (T 31

13

23

(4.3)

"32

of (4.1) can also be written, using other symbols,

as:

i}

"tfll

<^12

CT .3

^2.

<T 22

"23

CT 31

<T 32

^33.

=
_

The diagonal

stress

components of

Txy

T*

^.vx

CT.vv

CTv

T >,

T v:

o\.

cr..,

a.

a.

(4.4) are

(4.4)

normal to the planes xz, xy, yz while the

other stress terms are tangential to these planes and are referred to as shear stresses.
stress tensor

given by (4.4)

is

The

a second-order symmetric tensor.

According to the

The

invariants of a tensor

characteristics of a tensor, stresses

have the invariant property.

represent the quantities that are constant irrespective of the rotation of the coordinate
axes.

The

invariants of a tensor can be obtained

from the characteristic equation of the

square matrix given by:

<j-

-I o~ +IjO-Iz =0

(4.5)

From

the characteristic equation of (4.5) and the stress tensor matrix of (4.4), the

quantities of

Ii, I2,

and

I3

are obtained as follows:

/,

=<y
tl

=a u +a

22

+<7 33

<J,

<*n

CT 22

\2

11

<7

/,

23

a 23
a 33

(4.6)

+
\i

a 33

a,

.v:

+
a.

=a +a +o_

<7

(4.7)

...

57

= determinant of o v - \<J..

/,

The

invariants of the stress tensor

shown

in (4.6)

(4.8)

(4.8) are often expressed using

different formulation:

=/ =

=C"ii +cr 22 +cr 33

CT

(4.9)

(4.10)

(4.11)

where

/,

I2

and

73

are the

first,

respectively while

some authors

invariants

stress

of the

symmetric tensor

is

second, and third invariants of the stress tensor,

refer

It

I2 ,

and

I3

to as the first, second,

and

third

Another important property of the second-order

tensor.

the existence of principal directions related to invariants.

For a

given stress tensor, this implies that a set of planes, for which only the normal stresses are
non-zero, can be found.

The

directions of these normal stresses acting on such planes are

referred to as the principal directions, and the corresponding normal stresses are the
principal stresses.

In principal planes, the shear stresses are

always equal to zero.

Using

the principal stresses, the stress tensor of (4.1) can be rewritten as:

0"

"<7.

lj

c\

(4.12)
CT

where

0\,

a 2 and
,

03 are the principal stresses.

3.

In general, the largest principal stress is

referred to as the major principal stress while the smallest


stress.

The

is

called the

minor principal

third principal stress is referred to as the intermediate principal stress.

The

58

principal stresses can be determined

given in (4.5).

Three

from the

characteristic equation of the stress tensor

equation correspond to the major, minor, and

real roots of the

intermediate principal stresses.


It is

possible to

decompose

the stress tensor of (4.4) into

parts;

one called the

and the other one called the deviatoric

spherical or the hydrostatic stress tensor,


tensor.

two

stress

In matrix form, the hydrostatic tensor is given by:

0"

(4.13)

p
p_

where

l
,

(4.14)

and
o"

**

(4.15)

The

deviatoric stress tensor S

l}

is

the original stress tensor of (4.4).

defined by subtracting the hydrostatic stress tensor from

Then

the deviatoric stress tensor

=<7

'j

with the corresponding matrix form given

is

written as:

(4.16)

P* 9

as:

~s u

12

sa

s*

S 22

s 23

s i2

5 33.

n-P

<^I2

Vv

a 22 - p

"31

CT

32

^,3
CT

^33

23

"P

(4.17)

59

The

deviatoric stress tensor represents physically the shear deviatoric state of stress

excluding the hydrostatic

state

of

Because the deviatoric

stress.

stress tensor is also a

second-order symmetric tensor, the invariants of the deviatoric stress tensor can be
obtained from the characteristic equation.
the deviatoric stress tensor

should be noticed that the

It

invariant of

first

zero, as:

is

S = 5n

+S :: +S 33
1

-a mm

= an
.

The second

(4.18)

invariant J 2 of the deviatoric stress tensor

is

given by:

h-\s*

v]]

= i[((T n -

pf

(<7 22

^22

~'~^'>t

13

12

*^23

33

32

31

- p) 2 + 7 33 - p) 2 + 2S 12 2 + 2S 23 2 + 25 13

(4.19)

Since S12

(J12,

S23

/,

CT23,

and S13 =

(o-jj

G\3, (4.19)

can be rewritten

+ a 12 + a 33 + 2a n +

2ct 23

as:

2<7 13

- 3p 2 )

(4.20)

or

t[('ii

Using the principal

-ct.O +(o- 2: -o- 3 ,) +(o- n

stresses,

-ovJ

2
]

o-,

+ o\ 3 + ct 13

(4.21)

'

60

J 2 =-[(<7\-<y 2Y +(0- 2 -<T 3 )- +(C, -^3)"]

(4.22)

The

third invariant of the deviatoric stress tensor can be expressed as:

= I ((7 _/l5
o

7,3

'

^
)(cr
y

"

_Zl<5
jm

--/,/,. +
1

where

4.3

I,

1^

and

are the

first,

) (ct
ft*1

*}

tm

-!l$ mi )
o

(4.23)

Z,1

2?

second and third invariants of the

stress tensor.

Elastic Stress-Strain Relationship

The complete
relationships:
relationship.

definition of mechanical behavior of a

equilibrium

condition,

Figure 4.2 illustrates

The equilibrium condition defines

how

compatibility

body requires
and

condition

three basic
stress-strain

these relationships are connected with each other.

the relationship between the internal stresses

external forces including surface tractions

condition defines the relationship between the displacement u and the strain

because

the

The compatibility

and body forces F.

stress-strain relationship is also called constitutive equation

a and

it

e.

The

reflects the internal

constitution of the material.

The
elasticity

linear relationship

between

and can be defined by

stress

o~jj

the generalized

and

strain

sk

is

referred to as linear

Hooke's law as follows:

61

External force

Displacement

F,T

Equilibrium

>-

Compatibility

Stress

a
\1

Stress-strain relationship

(Constitutive law)

Figure 4.2

Definition of mechanical behavior of a

body

(after

Chen and Han

1988).

62

where

i, j,

k,

and

1, 2,

and

3;

= c eu
m

and Qjh

is

(4.24)

the fourth-order elastic moduli tensor.

(4.24) implies that the strain generated by the stress

condition of eliminating or decreasing the stress.


tensors,

Opi

also symmetric.

is

is

Eq.

always recoverable for the

Since both

Gjj

and Eh are the symmetric

Using a number of modulus symmetries the number of

The

independent constants can be reduced to 21.

final

expression for

Cjjki

of (4.24)

is

then written in matrix form as:

c mi

*"1122

'-'1133

&2Z

'-'2211

>-")!-))

c
c
'-'3322
c 1222
c 2322
c
'-'1322

c
c
'"3333
c 1233
c 2333
c
'"1333

^33

__

*12

*-"121I

c 23
c

<*23

.<V

where En,

22,

and

833

material, the general

*3311

1 1

131

A,

u.,

and

*'33

axial strains; Y12, Y23,

a are

and

=M 8 U +1*60,6
V

scalar constants.

c
'-'ins
c 2223
c
'-'3323
c 1223
c 2323
c
*"1323

c
'-'3312
c 1212
c 23
c
'"1312

form of the fourth-order tensor

Cm

where

">

c
*"1112
c2212

fi

Y13

c
c
c"3313
v
c 1213
c 23
c
*"1313
'-'1113

''213

= shear

ljk i

11

-)->

33

(4.25)

Ya
Y23

Ya

For an isotropic

strains.

can be given by:

+Ct8 a 8 jk

Because Cyw

is

(4.26)

symmetric tensor, the following

relationship should be satisfied:

=C

(4.27)

This leads to

X8 6 +Li6 ik 8 +aS8 Jk =X8 6 +n8 jk 8 a +a6 8


ij

kl

jt

JI

kl

Jl

ik

(4-28)

63

From

(4.27) and (4.28),

Taking (4.29)

it

into (4.24),

Then we

can be seen that n = a.

we

ij

are left with:

get:

M 8u e
u

H($ik s ji

+ 5 S jt
.!

) u

4 -30)

or

G,
J

The two independent


constants.

=AS

iJ

E kk+ 2fl

n and X

material constants

(4.31)
IJ

in (4.31) are referred to as

Lame's

Following Hooke's law for the three dimensional and isotropic element, the

stress-strain relationships

can be expressed as follows:

u = [^ii -v(ff a + V,)1

(4-32)

22

= [a
E

22

e33=^33- V ( CT

ll

7,2

7 23

G
=

+a

-v((Tu

C,
G

33 )]

(4.33)

+Cr 22)]

(4-34)

(4.35)

(4.36)

7,3

where

E and v = Young's modulus and

(4.35)

(4.32)

(4.37),

<5\z,

(4.37)

Poisson's

ratio;

G:3 and O13 represent the shear stresses.

(4.37) can be rewritten as:

elastic shear

Using

modulus.

In

indicial notation,

64

or

o
"

Comparing

(4.39)

and

(4.31),

=
\

+v

E
e

"

vE_

(l

;:., c,g,
+ v)(l-2v)

Lame's constants

u.

(4.39)

and X can be obtained

as:

H=G=

(4.40)
2(l

+ v)

vE

A=

(4.41)

(l+v)(l-2v)

There

is

another elastic parameter, called the bulk modulus K, which defines the

relationship between the hydrostatic stress

of hydrostatic compression, (4.31)

is

and the volumetric

expressed

the relationship of

p-

<7

= o\ 2 =

<7

32

For the condition

as:

a kk =3ke kk +2 tie

From

strain.

(4.42)

kk

p = (X + -fi)s kk

where e kk = 8n +

22

33

= volumetric

strain.

Then

K = -?- = A + -fi
u

(4.43)

the bulk

modulus

K is

defined

as:

(4.44)

65

Using (4.40) and

(4.41), the bulk

modulus

K is rewritten

as:

(4.45)
3(1

-2v)

Table 4.1 summarizes the relationships between the different


described so

far.

Using these

elastic parameters, the elastic

parameters

elastic

moduli tensor

ijk i

shown

in

(4.25) can be written in a matrix form as:

(1-v)

(1-v)

(1-v)
(1

[C]

-2v)
(4.46)

(l

+ v)(l-2v)

(1

-2v)
2

d-2v)

or

K + -G K--G K--G
3

3
9

K--G K + -G K--G
[c]=

3
?

3
?

K--G K--G

(4.47)

G
G

66

Table 4.1 Relationship between different elastic modulus.


Basic Pair

Constant
X,\i

X
(l

K, n

vE

3K-2n

+ v)(l-2v)

2(1

3A +

3(1

fxOk +

A+

lfx)

-2v)

//

2(A +

K
9KpL

3K +

IJ.

3K-2H

A
V

+v)

2a*

E,v

V
jU)

6K + 2/A

67

4.4

Elastic

4.4.1

Behavior of Soil

Initial elastic

known

well

It is

modulus

is in

upon unloading.

is,

There

10

modulus

The

for sands.

modulus

elastic

initial

however, a certain

strain

i.e.,

elastic

modulus for

small strains.

whether the loading

is

et al.

behaves

As can

usually very narrow with an upper

this strain

range

referred to as the

is

monotonic or cyclic (Shibuya

in-situ tests, laboratory tests,

and Richart 1963,


Salgado

is

has been observed that the

It

There are a number of ways to evaluate the

Those include

soil

shear

initial

a constant, for a given soil condition, regardless of the nature of the

is

loading type,

at

range within which the

stress-strain relationship is fully recoverable.

Figure 4.3, this linear-elastic strain range

in

The

part associated with strains that cannot be fully recovered

and the

as a linear elastic material

limit of

small strain

that the stress-strain behavior of soil is highly non-linear.

non-linear soil response

be seen

at

Yu and

1997c).

most

field

shear modulus for a given

soil.

and empirical equations (Janbu 1963, Hardin

Richart 1984, Baldi et

In

initial

et al. 1992).

al

1989, Viggiani and Atkinson 1995,

and laboratory

tests, the initial

shear modulus

is

determined by measuring the shear wave velocity based on the following relationship:

G =p(V

where
shear

wave

is

transmitted; and

analysis of surface

examples of

is

initial

hole test using the

shows

wave

in-situ

penetration test

obtaining the

shear modulus; p

initial

test

tests

(4.48)

= mass density of

shear

(SASW)

wave

velocity.

the

medium through which

The cross-hole

and the seismic cone penetration

for evaluating initial

test, the spectral

test are the typical

The seismic cone

shear modulus.

regarded as one of the most economical and rapid in-situ


shear modulus.

The

wave generated

at the

test is

performed

in a similar

surface (Robertson et

the schematic illustration of the seismic

cone penetration

al.

test.

the

manner

1985).

tests for

to a

down-

Figure 4.4

68

Plastic ranse

Non-linear elastic range

Linear elastic range

Figure 4.3

Non-linear stress-strain behavior of

soil.

69

Oscilloscope
1

oo

zf
Source

Seismic Cone
Penetrometer

Figure 4.4

Seismic cone penetration

test.

70

For laboratory

The bender element

often used.

used increasingly since.


in a

The

conventional triaxial

elements) attached
the shear

column

testing, the resonant

at the

wave do not

test

test

test,

and the bender element

test

test are

has been developed relatively recently, but has been

can be easily performed using the same

soil

sample as

with the wave generator and receiver (called bender

end caps of a

disturb the

triaxial

initial soil

Because the

sample.

strains generated

by

condition, the conventional triaxial test can

be performed immediately after the bender element

test for the initially

assumed

soil

condition.

The empirical equations


form

for the initial shear

modulus

are generally expressed as a

of:

-2- = CF(.

(4.49)

P.

where

initial

shear modulus;

Pa =

reference pressure used for normalization;

non-dimensional material constant; F(e) = function of the void


constant; p'

= mean

dependence of the

effective stress in the

initial

units as

Pa

Eq

n = material

(4.49) indicates the

shear modulus on the degree of compactness of the soil and the

magnitude of the confinement.


initial

same

ratio;

C =

shear modulus of sand

is

One

of the

commonly used

empirical equations for the

the one suggested by Hardin and Black (1966) and written

as follows:

where

Cg

initial

void

e g and n 2

effective stress in the


the values of

Ca

material constants that depend only on the nature of the soil; e

Pa =

ratio;

e,

reference pressure

same
and n

unit as

Pa

= 100 kPa =

In the original

were suggested

for

kgf/cm

2
;

and o' m =

initial

mean

work by Hardin and Black (1966),

round grains (Ottawa sand) and angular

71

grains (crushed Quartz).

values of

Cg

eg

After Hardin and Black (1966), several authors presented the

Table 4.2 shows the values of C, ng

and n g for other kinds of sands.

and e for different sands.

Hyperbolic stress-strain relationship

4.4.2

As discussed

in the previous section, the stress-strain

highly non-linear relationship from the early stage of loading.


soil

models has been widely used

wide range of

strains.

behavior of a

shows

soil

The hyperbolic family of

to represent such non-linear behavior of a soil over a

Since Kondner (1963)

first

proposed the original hyperbolic

equation for the stress-strain relationship, several modifications have been suggested

(Duncan and Chang 1970, Hardin and Dmevich 1972, Fahey and Carter 1993, Purzin and
Burland 1996).

The hyperbolic soil-models

strain relationship

assuming piece-wise

are based

on the quasi-linear

elastic stress-

linear behavior with respect to a stress

and

strain

level.

The conventional hyperbolic equation

for stress-strain curve

by Kondner (1963)

is

written for triaxial or plane-strain condition as:

where

CTi

and

O;,

are the

-<7 3 =

<7 1

(4.51)

a + be

major and minor principal

stresses;

is

the axial strain; and a and

b are the material constants that characterize the feature of stress-strain curve.
seen in Figure 4.5, the constants a and b

Kondner (1963) correspond

the conventional hyperbolic equation

to the value of the reciprocals of the initial elastic

and of the asymptotic value

strain curve.

in

As can be

(Gi

03) U

by

modulus

of deviatoric stress in the hyperbolic stress-

72

Table 4.2

Values of
(after

Cg

eg

and n g for different sand type

Salgado 1993, Salgado

et al.

1999)

Cg

eg

ng

Ottawa

612

2.17

0.44

Ticino

647

2.27

0.43

Toyoura

900

2.17

0.40

Hokksund

942

1.96

0.46

Monterey No.O

326

2.97

0.50

Sand type

73

cti

-cr 3

Asymptote =

(o"i -

a3 ) =

1/b

(a)

(b)

Figure 4.5

Hyperbolic model

(a) stress-strain

curve and (b) linear representation.

74

From

(4.51) and Figure 4.5,

easily seen that an infinite

is

it

amount of

required for the stress-strain curve to reach the ultimate stress level (O]

implies that the ultimate stress (ai


relationship

is

a3 ) u

a 3 ) ult

strain is

This also

in the conventional hyperbolic stress-strain

it

greater than the actual soil strength, which

is

taken as the

maximum

stress

at failure.

For the purpose of better

hyperbolic relationship to a real soil stress-strain

fitting the

Duncan and Chang (1970) modified

curve,

The

material constant Rf into the equation.

relating the ultimate deviatoric stress (Oi

curve to the actual deviatoric stress of

(<J 1

-o

The modified hyperbolic equation

then written

~i

stress at failure (C\

is

referred to as the failure ratio,

o 3 )f

(4.52)

3 ) uh

as:

nr

by introducing the

of the original hyperbolic stress-strain

=Rf (o -a

eo

where the deviatoric

a 3 ) ul

Rf

factor

soil at failure (o,

3 )f

is

<'.-">

the hyperbolic equation

(4 53)
'

y,-<7 3 ) /

a3 ) f

determined from the Mohr-Coulomb

is

failure criterion:

(a,

- <r,) =

2c-cos</>
-

'

where
than

stress

1,

and

(j)

are the

(CT] -

o3 )f

at failure.

(o"i -

WW

-sin</>

(A

l-sin0

Mohr-Coulomb shear

the asymptotic value

+ 2o\-

c3 ) u

is

strength parameters.

If

the factor Rf

4n

is less

1/R f times greater than the actual deviatoric

The hyperbolic model

describes the non-linearity of the stress-

75

strain relationship

without recourse to plasticity concepts.

As

a result,

it

can be readily

incorporated into numerical analyses using an incremental elastic implementation.

Degradation of Elastic Modulus

4.4.3

The hyperbolic equation given by


terms of shear stress and

(4.53) can be rewritten so as to be expressed in

strain:

7
(4.55)
1

where

shear stress

initial

Y-Rf

shear modulus in the very small strain range; and

at failure.

Using the relationship y = x

x^

G, (4.55) can be written

= maximum
as:

(4.56)

*/

WG

Thus,

- = l-Rf

where

G =

secant shear modulus.

modulus from

its initial

maximum

Duncan and Chang (1970)

(4.57)

Eq. (4.57) describes the degradation of the shear


value

according to the magnitude of shear

implicitly described the degradation ratio

stiffness as varying linearly with the stress level x/Xma*

G/G

according to (4.57).

stress.

of elastic

The degraded

76

magnitude of the

elastic

modulus

determined by the value of Rf

at failure is

According to Duncan and Chang (1970), the value of Rf typically

lies

in (4.57).

between 0.75 and

1.0.

4.5

Failure Criterion and Soil Plasticity

Failure criterion

4.5.1

Although the non-linear behavior of


there

When

increment, can be observed.

and behaves as a

is

is in this

condition,

failure criterion

plastic behavior.

is

In

For a given temperature, the

is

no longer

failure criterion

said to have "failed"

used to define the limit stress

most

the material has reached a failure condition

behavior, Hooke's law, given by (4.24),

classical plastic theories,


is

defined by a failure

and started exhibiting

plastic

valid.

F can be expressed

as a function of

components as follows:

F = F(a u = F(o n ,a 22 ,o 33 ,o v M
)

where F =

failure criterion;

ay =

six stress

isotropic, the soil properties are the

same

as

not affected by coordinate transformations.

given

it is

considered as elastic below the failure surface which

Once

criterion.

the soil

The

plastic material.

which the material exhibits

the material

stress

from the very early stages of loading,

a condition for which relatively dramatic changes of stress, for a given strain

is

state at

soil starts

in

follows:

components.
all

,,o n )

If

the soil

(4.58)

is

assumed

to

be

directions, thus the material constants are

It

follows that the failure criterion can be

terms of principal stresses or stress invariants, for isotropic conditions, as

77

F = F(<r lf <Tj,cr 3 )

(4.59)

F=

(4.60)

or

where

o~i,

o 2 and
,

03 = major, intermediate, and minor principal stresses;

invariant of stress tensor; and J 2

and

J3

Ii

the

first

the second and third invariants of deviatoric

For certain materials, experimental observations show that the influence of

stress tensor.

hydrostatic

F(I ,J 2 ,J i )

pressure on

failure

observations, the failure criterion

of the material
is

is

minimal.

Based on

these

the

expressed for these materials in a more simplified

form as follows:

F = F(J 2 ,J i )

In

some

elastic-plastic

before the failure criterion


plastic range.

models
is

in soil

(4.61)

mechanics, a

soil

remains

in the elastic

reached and deforms under a constant shear stress

For such materials, referred to as perfectly

F<

plastic, the

range
in the

following holds:

for the elastic range

(4.62)

and

F=

For a perfectly

plastic material, the condition

illustrates the elastic

Ojj

and

the

at failure

and

state

and the

takes place inside the

is

remains on the failure surface,

stress

failure

corresponds to the elastic condition of (4.62).


state

F>

(4.63)

Figure 4.6

physically impossible.

plastic stress states for the elastic-plastic material.

doij represent the current stress

stress

and beyond

increment

In the figure

at failure, respectively.

surface given by (4.59) or (4.60),

As shown

in

this is referred to as the

Figure 4.6,

when

it

the stress

On

loading condition.

other hand, the case where the stress state drops below the failure surface

unloading condition.

If

is

the

called

78

Loading

dGjj,

loading

Failure surface, F(aij)

Figure 4.6

Stress states for elastic-perfectly plastic material.

79

Considering a small stress increment dOjj from the current stress


for loading

state CTy, the

conditions

and unloading are given by:

/(<7)

/(<T

df=-l- do,.

and

(4.64)

fl

for loading

and

and

df=^-da<0

(4.65)

for unloading.

4.5.2

Flow

When

rule

and

stress

hardening

the stress state of a elastic-plastic material reaches a failure surface, the

The

material starts exhibiting plastic behavior.


state

can be given by the

sum

The flow
state.

total strain

can be determined.

=de g '+de/

increment; ds,/ and

rule defines the relationship

Based on the flow

rule, the

The flow

increment tensor

at plastic

of the elastic and plastic strain increment:

de 9

where dey =

total strain

rule

de^ =

between the

(4.66)

elastic

and

plastic strain increment.

plastic strain

magnitude and direction of the


is

and the current

plastic strain

stress

increment

given by:

dB,'-kl0<7

(4.67)

80

a positive scalar factor of proportionality and Q.

where

A. is

When

the plastic potential function Q.

is

is

same

the

Otherwise

called an associated flow rule.

it

is

a plastic potential function.

as the failure criterion F, the flow rule

referred to as a non-associated flow

is

Eq. (4.67) also indicates that the direction of the plastic strain vector

rule.

to the plastic potential surface defined

space.

This condition

For the perfectly


strain

increment

de,j in

is

positive

condition

is

is

normal

plastic potential function Q. in the stress

referred to as the normality rule.

plastic material, the

work done by

the stress increment da,j and

the plastic state should be equal to zero since

no further increment

Some

materials, however,

of stress, after the failure condition

show

by a

d|j

is

reached,

amount of work done by

referred to as

is

possible.

additional stress and strain increments.

work hardening

or a stable material condition.

This

According

to

Drucker, the following two conditions should be satisfied for the hardening material:

(1)

During the application of the added

set of forces, the

force on the changes in displacements


(2)

Over

produces

first

produces

is

the cycle of application and removal of the

work performed by

The

it

is

the external

work done by

the external

positive.

added

set of forces, the

agency on the change

in

new

displacements

non-negative.

condition can be given by:

do

>0

-de
:j

(4.68)

:J

or

da (d/+de ")>0
iJ

The second condition implies

(4.69)

lJ

that:

dcr

-de/>0

(4.70)

it

81

Hardening, stable

Figure 4.7 Stress-strain behavior for hardening, perfectly plastic and softening material.

82

work done by

If the

referred to as

additional stress

work softening or

hardening, perfectly plastic, and

An

increment

negative, the material

is

is

Figure 4.7 shows the conditions for the work

unstable.

work softening

materials.

important property that characterizes the state of cohesionless granular materials

The

the relative density.

soil

strain

Soil dilatancy and critical state of sand

4.5.3

is

and

with respect to

its

DR

relative density is defined as the degree of

most and

e,mx

max

where

DR

initial

void

least

dry unit weight.

states.

xl00% =

'

and emm =

and Yd.mm =

relative density

Yd
'

rf.max

Yd ma

written as:

xl00%

'

D R is

(4.71)

rf.min

maximum and minimum

maximum

Relative densities equal to

The

7rimax

nun

relative density; e max

ratio; Yd.ma*

dense

compactness of the

void ratios; e

and minimum dry unit weight;

DR =

100% and

0%

Yd

initial

represent the densest and

loosest conditions of sand, respectively.

The

stress-strain

behavior and volumetric changes of sand during shearing differ

significantly depending on the level of relative density.

looser sand

shows

a stress-strain

showing hardening behavior.

On

As can be

seen in Figure 4.8, the

curve with gradually decreasing curvature

until failure,

the other hand, the denser sand has a stress-strain curve

with a clear peak.

With respect

to the volumetric strain of dense sand,

negative volumetric

strains,

is

usually observed that

representing increases of volume, occur following the

contractive behavior initially observed.


referred to as dilatancy.

it

This phenomenon observed

in

dense sand

is

For the same amount of confinement, consequently, the denser

the sand, the greater the shear strength,

and the higher the peak

friction angle.

The

other

83

O'l

-C,
1

Dense sand

n.

^w

Critical
state

s^~^

Loose sand

taxial
1

Dense sand

^_^

taxial
|

+
Loose sand

Loose sand

\.

^^-"""^

Dense sand

'

^^axial

Figure 4.8

Different behavior of dense and loose sand


(after

Lambe and Whitman

1986).

84

factor that affects the dilatancy of sand

the confinement.

is

lower confinement produces higher dilatancy

There

is

in sand.

a condition far beyond the peak point of the stress-strain curve, for

no volume change

is

observed.

This stage

in Figure 4.8, the volumetric strain vol


critical state

has been observed that

It

remain constant.

The

is

referred to as the critical state.

deviatoric stress o'i

g'
3

friction angle at this stage is

and void

which

As shown

ratio e at the

regarded as a material

property that depends only on the nature of the sand, and not on either the

initial

density

or confining stress.

The

friction angle at the critical state of

lr

sin0 f =

where

(J)

sand

is

(4.72)

friction angle at the critical state;

given by:

and

0"'
]c

and o' 3c = major and minor

principal stresses at the critical state.


In order to quantify the dilatancy of sand, Bolton (1986)

proposed the following

relationship based on the experimental test results:

0,=0

in

which

<j)

= peak

dilatancy angle.

The

friction angle;

dilatancy angle

(j>c

\|/

= 6.251 R

iff

+O.8vr,

friction angle at the critical state;

is

given by:

for plane-strain conditions

(4.73)

and

\|/

(4.74)

and

iff

- 3.751 R

for triaxial conditions

(4.75)

85

The

dilatancy index I R in (4.74) and (4.75)

IR

where I D =

relative density as a

lkgf/cnr; P' p

= mean

is

given by:

=I D [Q + M-)]-l

number between

effective stress at

and

1;

peak strength

(4.76)

Pa =

in the

reference stress

same

material constant approximately equal to 10 for clean silica sand.


(4.76),

it

is

seen that the dilatancy angle

confinement.

The higher

vj/

(4.73) through

the relative density, the higher the dilatancy angle, whereas the

(1986) limits the values of the dilatancy angle

From

P a and

depends on both relative density and

higher the confinement, the lower the dialtancy angle.

(p

units as

= lOOkPa =

- <p c < 20

For

practical purposes, Bolton

at:

for plane-strain conditions

(4.77)

for triaxial conditions

(4.78)

and
<j>

4.6

-<p c

<\2

Summary

In this chapter, the

mechanical behavior of

relationship between stress Oy

and

strain

soil

was discussed.

The

linear elastic

E^ can be defined by the generalized Hooke's

law as following equation:

ay = C

where the
the bulk

elastic

modulus

moduli tensor Cpi

is

M ,,
kl

ijkl

expressed

in

terms of two elastic constants, either

K and shear modulus G or Young's

modulus E and Poisson's

ratio v.

86

Hyperbolic stress-strain models have been used to represent the non-linear behavior

beyond the very small


Hyperbolic

soil

strains

for

which

soil

models are based on quasi-linear

behaves as a linear
elasticity

and can be expressed

of the degraded elastic modulus as a function of the stress or strain level.


elastic

modulus can be obtained from the

stress level

and the

initial elastic

in

terms

The degraded

ratio of the current stress level to the

modulus, which

material.

elastic

maximum

given by the confinement and the

is

relative density for sands.

At

large strains, soil exhibits plastic behavior.

behavior of

soil,

two conditions

are required: a failure criterion

failure criterion defines the stress limit

After this stress limit,

soil

For the description of

under which

no longer follows the

soil

and a flow

The

rule.

remains as an elastic material.

The

elastic stress-strain relationship.

flow rule represents the relationship between the stress and strain rate

The magnitude and

plastic

in the plastic range.

direction of the plastic strain increment can be determined based

on

the flow rule.

The peak

friction angle of

the critical state

sands can be expressed in terms of the friction angle

The

and the dilatancy angle.

friction angle at the critical state is a

constant for a given sand, and depends only on the nature of the sands.
angle,

on the other hand,

is

a function of confinement

(1986) proposed the following relationship

where

fyp

= peak

friction angle;

in

and

linear surface.

is

The dilatancy

relative density.

Bolton

order to quantify the dilatancy of sand:

c = friction angle

at critical state;

determined by the peak

and

\|/

= dilatancy

As

a result, the failure

friction angle, is

regarded as a non-

angle determined as a function of density and confining stress.


surface of sands, which

at

87

CHAPTER 5

5.1

3-D

NON-LINEAR ELASTIC-PLASTIC STRESS-STRAIN MODEL

Introduction

The

non-linear elastic models mentioned in the previous chapter were developed on

the basis of a plane-strain condition.


circular plate within a calibration

three-dimensional modeling.

Vertical loading of either a pile in the free field or a

chamber

is

an axi-symmetric problem, thus requiring

In order to obtain

more

realistic results for

such problems,

therefore, the full description of the stress-strain relationship including non-linear elastic

and

plastic behavior in three

In this chapter,

we

dimensions

discuss

first

is

necessary.

the characteristics of the intrinsic

variables that will be used in the stress-strain model.


soil

5.2

model

for three dimensions

Intrinsic

is

The

full

and

state soil

non-linear elastic-plastic

then presented.

and State Soil Variables

Soil variables used for the description of soil behavior can be classified as either
intrinsic or state variables.

only a function of

Salgado

et al.

Intrinsic soil variables

soil particle,

1997a).

do not change with

soil state

mineralogy, shape, and size distribution (Been

This implies that the intrinsic

uniquely determined irrespective of the stress

soil variables for a

state, history

variables include the friction angle at the critical state

given

et al.
soil

or initial condition.

(<j) c ),

and are

specific gravity

1991,

can be

These

(G s ), and

88

maximum and minimum

void ratios (e^x,

(4.50) for the initial shear

modulus

The parameters C g eg and n g used

and

in

are also regarded as intrinsic soil variables.

state represents the physical condition


(e Q ) or relative density (Dr),

on the other hand, are determined by the

Soil state variables,

most important

&nun)-

under which the

in-situ vertical

state variables of sands,

The

soil exists.

and horizontal

The

soil state.
initial

stresses (o' v

soil

void ratio

o'h) are the

and control the behavior of the sand (Been

et al.

In contrast to the intrinsic variables, the determination of state variables (for

1986).

example by using laboratory


significant

difficulties

cohesionless

soils.

for the determination

The

intrinsic

and

complex

soil

test,

model should be able

to relate the

models have come

may be

parameters.

the

The procedures

experimental efforts.

stress-strain

as the cone

behavior to

With the rapid growth of computing power,

state variables.

results

particularly for

variables,

however, offers more effective, indirect ways

to

be practical for more

In the applications of such soil models, the

problems.
accurate

state

under in-situ conditions.

to estimate soil state variables

useful soil

of soil

recent use of in-situ testing methods, such

penetration test and the pressuremeter

This presents

undisturbed soil samples.

tests) requires

to

the use of

soil

more

realistic analyses of geotechnical

most

critical factor for

reasonably accurate determination of

soil

obtaining

and model

determine these parameters in general require significant

The more complicated

the soil

models

are, the

more laborious

the

determination of soil and model parameters.

As

a simplified approach, the secant modulus

is

sometimes used

However, the non-

non-linear behavior of soil before a failure condition (Figure 5.1).


linearity that soil

shows

level, stress state,

and displacement

is

to represent the

quite complex, varying significantly with the initial density


level of interest.

Consequently, the selection of the

value of the secant modulus depends on the soil condition and the characteristics of the

geotechnical structure being analyzed.


secant

modulus

It is

that will

It

also implies that there

produce acceptable results for

possible to back-estimate the secant

function of displacement level.

all

possible

is

no

initial

single value of

conditions.

modulus using experimental

This approach, however,

is

results as a

not a fundamental solution,

89

Figure

5. 1

Secant modulus for non-linear stress-strain behavior.

90

and may
state

result in an inaccurate relationship

between modulus, density level and

due to the limited number of experimental

soil variables in

results.

The use of the

intrinsic

stress

and

state

a non-linear elastic-plastic soil model combined with suitable numerical

techniques allows a more fundamental solution to geotechnical problems.

5.3

Modified Hyperbolic Model for Non-Linear

The conventional hyperbolic equation of


elastic

modulus with

respect to stress level

curves of real soils under

static

Elasticity

(4.57) implies that the degradation of the

is linear.

However, the measured degradation

or quasi-static loading can be quite different from linear.

Figure 5.2 shows measured modulus degradation curves of normally consolidated sand

and the degradation


figure

was

line

by the conventional hyperbolic equation plotted together.

plotted in terms of normalized shear

stress T/x max

where

and Tmax represent the

shear stress at failure, respectively.


cyclic loading

As can be

The

modulus G/G Q and normalized shear


shear modulus and the

initial

seen, the

maximum

modulus degradation curve

for

On

the

shows good agreement with the conventional hyperbolic equation.

other hand, the modulus degradation curve for monotonic loading reveals significantly
different response

from

modulus degradation
quite rapid.

The

for

that of the conventional hyperbolic equation.

monotonic loading from the

initial

The measured

shear modulus

rate of the degradation then drops as the

is initially

normalized shear

stress

increases.
In order to account for the characteristics of the
real

soils,

modulus degradation observed

in

Fahey and Carter (1993) proposed a modification of the conventional

hyperbolic model.

The modification

consists of the introduction of a parameter g into

(4.57):

|*- =

l-/()*

(5.1)

91

1.0

Cyclic

0.8

NC

sand

</T

3
3

"\
0.6

0.4

Hyperbolic

equation

o
%,

JZ

V)

"

>v

0)

>*

,N

/*'*

15

E
o 0.2

N*
"*

Monotonic
NC sand

'

.
*--.

0.0
0.0

0.2

0.4

0.6

Normalized Shear Stress,

Figure 5.2

Modulus degradation
(after

0.8

\
1.0

x/T max

relationship for normally consolidated sand

Teachavorasinskun

et al. 1991).

92

The parameter

f in (5.1) has the

and g are both

If f

level.

original hyperbolic equation

role as Rf in

Duncan and Chang's hyperbolic

The parameter g determines

equation of (4.53) and (4.55).

curve in terms of the stress

same

by Kondner (1963).

set

the shape of the degradation

equal to

1,

(5.1)

becomes

the

equal to 0, (5.1) represent the

If f is

modulus

linear elastic relationship with a constant value of shear

irrespective of stress

level as:

-^-

=1

G=G

or

(5.2)

If

to be equal to

is set

with a certain value of

and Carter (1993) suggested

and

the

Duncan and Chang hyperbolic

Figure 5.3 shows typical modulus degradation

relationship given by (4.57) results.

curves for different values of

f,

For a

g.

of

that the value

typical normally consolidated sand,


f

= 0.98 and g = 0.25 lead

Fahey

to reasonable

agreement with observed modulus degradation curve.


In numerical analysis of soil behavior using non-linear stress-strain models, a

tangent modulus with successive incremental procedure rather than a secant modulus

with successive iterative procedure


Christian

1977).

is

Chang 1970, Desai and

According to Duncan and Chang (1970), the advantage of the

incremental procedure

which

often used (Duncan and

is

the initial stress state can readily be taken into account,

is that

very important in geotechnical problems.

The accuracy of

the incremental

procedure can significantly be improved by combining the iterative method for each
increment.
(5.1).

The tangent shear modulus

Using the relationship x = G-y,

can be obtained from the differentiation of

(5.1) is rewritten as:

J- = !-/(__)*
G

(5.3)

T,

or

-=G
V
/

r-

-G

f{
T
max

)'(t)'-

(5.4)

93

1.0

G/Go

0.0

Figure 5.3

X/Xn

Modulus degradation curve

1.0

for different values of f

and

g.

94

Replacing 1/y by

,:

7 =

<T

(5.5)

Then,

dr

dx

-Vl-Gj- -GJ(g -1)( )' (t)'"


1

2
]

T max

[G t-

+G /U-1)(
r

y
)

(r)^

2
]

max

Gj- [i-f(i-g)(y]
T max
(5.6)

Yiry[GX'-GJi
T

Thus,

dY

i-/d- g )(-L-)

dT

Because

dr

dy

(5.7)

^max

G [l-f() s

(5.7)
2
]

becomes:

i-fa-g)(y

T max

(5.8)

'

or

S
G

)
G

i-fa-gxy

(5.9)

95

Using

(5.9), the tangent

the current secant shear

5.4

modulus G, corresponding

modulus

to the current stress level x/x max

and

G can be obtained.

Non-Linear Elastic Model for Three Dimensions

5.4.1

Modified hyperbolic

dimensions

stress-strain relationship for three

Application of the hyperbolic model to the analysis of the loading of a natural soil
deposit requires the resolution of

what shear

two

Firstly, there is the issue

issues.

stresses to use in the stress-strain model.

stress-strain response,

to define

Ko = g\/o\.

which

Most

is

determined by a

laboratory studies on

however, have been done on isotropically consolidated samples.

In order for the analysis to be


soil

how

In naturally deposited soil, the

stress state is anisotropic with an existing initial shear stress

coefficient of lateral earth pressure at-rest,

of

more

realistic, the initial

deposit should be taken into account.

the non-linear elastic models,

shear stress existing in a natural

second issue

which have been based on

is

related to the extension of

common

laboratory tests and are

two-dimensional, to three dimensions.


In the

modified hyperbolic model by Fahey and Carter (1993), the degradation of

shear modulus
f

and

g.

To

is

expressed in terms of shear stress level x/Xma* with the model parameters

take account of the

initial

shear stress due to

initial stress

anisotropy in the

hyperbolic model, the following formulation should be used instead of (5.1):

^- = l-/(

where x =
t

initial

x and Tmax

shear stress due to the

were defined

initial

in the present study.

Ko

condition.

(5.10)

Figure 5.4 illustrates

In Figure 5.4 (a),

it

how

should be noticed

96

/
B

^^

Ko

^^

Tmax

A
X
To

(a-,

g' )/2
3

a
(a)

y Kf

^
AE^^

d/|/
/I
T

'

\
'

Tmax

/ ^S\ c

(b)

Figure 5.4 Definition of x

t and ima* for (a) constant and (b) varying confinement.

97

that

x and Xma X are

The

a' 3 )/2.

all

stress path

defined

AB

The

anisotropic stress condition.

more representative of
where

condition,
stresses.

is

at

the

same confining

only possible in a simple shear


stress path

CF

Figure 5.4

in

stress path

maximum

CF, the

on

stress

line

onto the Kf line) represents the

CF.

As an

maximum

current shear stress t and confining stress a' m

initially

the other hand,

is

Ko

and confining

shear stress x^,* varies along the line DF,

5.4 (b) representing the current shear stress x, point

on

(a'i

the axial loading of a foundation for soils with an initial

depending on the current shear

point

(b),

with

test

failure is reached as a result of increases in both shear

For the

by

stress as represented

illustration, for point

Figure

in

(obtained by vertically projecting


shear stress corresponding to the

Full description of the stress state under various types of external loading requires a

specification of the intermediate principal stress as well as the


stresses.

Three-dimensional

stress invariants.

The

first

stress conditions

invariant

I]

may

major and minor principal

also be represented

of the stress tensor given by (4.9)

by the use of

is

a measure of

confining stress, and the square root of the second invariant J 2 of the deviatoric stress
tensor given

by (4.22)

invariants,

the

dimensions

as:

is

a measure of shear stress in three dimensions.

hyperbolic

relationship

S-i-/(

where ^/77 V-^o" anc* V^2max


.

are

t 'ie

may now be

of (5.10)

f~*

ff

Using the

rewritten

for

in (5.10).

necessary for J2max to be defined using a three-dimensional failure criterion.

Prager failure criterion

is

this

three

(5.ii)

^-D equivalents of xQ x and Tmax

Drucker-Prager failure criterion was selected in

stress

study for that purpose.

It is

The

The Drucker-

given by:

F = //7-ctf,-/c =

(5.12)

98

where

a and K are

related to the

Mohr-Coulomb

strength parameters c

and

<j)

through:

., ...

2sin</>

a = j=

(5.13)

V3(3-sin0)
and

K=

* CCOS *

(5.14)

V3(3-sin0)

In (5.13)

and

(5.14),

sands, c and hence

obtained

(j)

k=

and
0.

c represent the friction angle

From

(5.12), J 2ma x for sands at a given

of

a result, the

For

confinement would be

as:

7 2max

As

and cohesion, respectively.

versus

x,

3-D

stress plane of

Ii

=a

versus .^77

(5.15)

is

used, instead of the

to obtain the stress level associated with (5.11) in a

2-D

manner

stress plane

similar to that

of Figure 5.4.

Considering the stress path

CH in Figure

5.4 (b), the shear

modulus should increase

due to the increase of confinement, while the magnitude of shear


In order to account for the influence of confinement

stress is kept constant.

on shear modulus during loading,

(5.11) is modified as follows:

Sl = [l- /(

where

Ii

and

I ]0

f f
2

y ](^)

n<

(5.16)

are the first invariants of the stress tensor at the current

The parameter n g

is

the

same

as appears in (4.50).

dependence of shear modulus on confinement.


that the degradation of shear

In both equations,

Use of

and

initial states.

n g represents the

(5.16), rather than (5.1 1), permits

modulus be properly expressed

in

terms of both shear stress

99

and confining
Ii

=I ]o

at

any

stress levels. Eq. (5.16) reverts to (5.11) for the

As

which

stress level.

discussed in chapter 4, the stress-strain response of an elastic material

described by two constants; the bulk modulus

and the shear modulus

elastic stress-strain relationship using the bulk

can be written

in

modulus

O-n
CT 22

^33

K + 4G

A
K-lGy/

and the shear modulus

K .2 G/

K+

K -2G

22

A
4Gy/

/:

K -2G/ K _2G/ k+ 4Gj

CT
.

G
G

723

13.

of (5.17) also change.

7.3.

As described

earlier, the

shear modulus

is

strain relationship as a function of current stress state.

modulus depends mainly on the magnitude of

The K-G model

is

al.

modulus

G and bulk

given by the non-linear stress-

The magnitude of

the confining stress (Naylor et

one of the ways for accounting for the nonlinear

before yield (Naylor et

Thus, the complete description of a non-linear elastic

relationship requires proper representation of variations in both shear

model

(5.17)
Yl2

the stress state changes in a non-linear elastic model, the elastic parameters

modulus K.

^33
<

21

are often used.

K _2G

On

As

is

matrix form as follows:

and

in

Variation of bulk modulus and Poisson's ratio

5.4.2

The

simple shear case

1981, Salgado 1993).

The

the bulk
al.

1981).

elastic characteristics

basic considerations of the

K-G

are:

1.

The magnitude of the shear modulus

2.

The magnitude of the shear modulus decreases with

increases with increasing confining stress.

increasing shear stress.

100

The magnitude of the bulk modulus increases with increasing confining

3.

The

first

two considerations have already been included

in the

stress.

development of the non-

linear elastic relationship of (5.16).

Based on the discussion on the

K-G model by

Naylor

et al. (1981), the

tangent bulk

modulus K, can be represented by the following equation:

=D

K,

where P a = reference

same

units as

Pa

stress

= 100 kPa =

and the

initial

rCP

<1

kgf/cm

-"<

(5.18)

n )

a' m

= mean

effective stress in the

material constant that can be calculated from the initial values of

bulk modulus and confining

The values of the

-7'
5

initial

stress;

and n k can be taken

bulk modulus

may

as 0.5 with reasonable accuracy.

be obtained from the

initial

shear modulus

Poisson's ratio v D which can, in most cases, be taken in the 0.1


,

0.15

range.

The expression of Poisson's

ratio

obtained from the relationships of Table


the tangent

Young's modulus E,

in terms of the current loading state

4.

Based on these

to the initial

can be

relationships, the ratio of

Young's modulus

ED

can be written as

either:

5.. G '" + V)

(5.19)

G.a+v.)

E
or
E,

__

K,(l-2v)
(5.20)

K (l-2v o
{

in

which v =

initial

Poisson's ratio; Ki,

moduli, respectively.
state

can be given by:

From

(5.19)

and

K G
t,

and

initial

and current bulk and shear

(5.20), the Poisson's ratio

at the current stress

101

a+vjS-(i-2vj
a
v,=4
2-Ml+v + (l-2vj
g

(5.2D

a:

In (5.21),

is

it

seen that the Poisson's ratio approaches 0.5 as the shear modulus

approaches zero, as expected.

Determination of the parameters

5.4.3

The parameters
elastic

modulus.

and g

in (5.16)

and g

determine the characteristics of degradation of the

and

In order to evaluate the values of f

g, a set

of triaxial

test results

performed by Giuseppe (1991) and Vecchia (1991) for Ticino sand was analyzed.
sand has been studied extensively (Salgado 1993, Bellotti

number of

When
results,

it

The

the

properties of Ticino sand are

should be pointed out that the

first

soil

an isotropic

stress

a' m

(o"'i+2a'3)/3

considered to undergo changes

in

is to

Table

The conventional

If

confinement
a's, a

5.1.
triaxial test

triaxial loading is subjected


triaxial test is

performed by

is

defined through the

triaxial

soil

stress.

is

mean

specimen can be

As

a result,

it is

not

shear stress for a triaxial loading condition,

due to continuously varying confinement, as discussed


possible to identify a different value of

1994, Salgado

then a deviatoric axial stress, which

rather than

maximum

al.

be determined from

both confinement and shear

possible to define a single value of

is

in

specimen under

pressure a' 3

cell

increased (or decreased) until failure.


effective

shown

modulus degradation relationship

to continuously increasing confinement.

applying

1996) and involved in a

et al.

laboratory plate load and cone penetration tests (Ghionna et

et al. 1997a).

Ticino

maximum

in the previous section.

Instead,

it

shear stress corresponding to each

value of current confinement a' m as illustrated in Figure 5.4(b).


,

The use of Young's modulus


relationship

from

triaxial tests.

is

more

suitable for obtaining the

Young's modulus

modulus degradation

in a triaxial test is calculated

from the

102

Table 5.1

Basic properties of Ticino sand (after Ghionna

D 10 (mm)

0.36

D 50 (mm)

0.54

(G s )

Specific gravity

2.623

Coefficient of uniformity (U)


Friction angle at critical state

34.8

^max

0.922

Cmin

0.573

Ymax (kN/m
Ymin

(<J> C )

1.5

(kN/m

16.68

13.65

cg

647

ng

0.44

eB

2.27

et al. 1994).

103

applied axial stress (a'i

The

constant.

&?,)

stress level

and the axial

strain E^ai

with the lateral stress

a^ remaining

can be determined from the applied axial stress

maximum

normalized with respect to the unique value of

(a'i

a' 3 )

axial stress at failure,

(o~'i

o-' 3 ) f .

Figures 5.5 through 5.17 show the measured and calculated modulus degradation

curves obtained from


(i.e.

triaxial tests

medium dense sand

on samples with two different

with relative density equal to approximately 50%, and dense

As can be

sand with relative density equal to or higher than 90%).

the degradation curves were plotted in terms of normalized

versus deviatoric axial stress level (a'i


secant and

Young's modulus

initial

calculated from the

initial

relative density levels

a'3)/(o"

a'3)f,

The

respectively.

seen in the figures,

Young's modulus E s/E

where E s and E represent

the

Young's modulus E D was

initial

shear modulus given by (4.50) and the

Poisson's ratio

initial

at

small strain taken as 0.15.

The values of
found to be

in the

dense sand were

summarizes the
samples used

and g for the medium dense sand shown

0.96

in the

0.97 and 0.15


0.93

initial elastic

in Figures 5.5

that the values of f

From

Figures 5.5

and g vary according to the

for the

and g for the

5.17 and Table 5.2,

relative density level.

were

medium

triaxial soil

it is

As

observed

the relative

This result

density increases, the value of f decreases while the value of g increases.

modulus

5.8

Table 5.2

0.32 range, respectively.

modulus and the values of

5.17.

Those

0.20 range, respectively.

0.95 and 0.2

in Figures 5.5

value

is

higher for

denser than for looser sand, and the rate of degradation of elastic modulus

is

higher for

indicates that the ratio of the elastic

looser than for denser sand.

Because the
for only

two

The

triaxial tests

at failure to its initial

results for both cases are in

directly applied to other relative density levels.

and g shown

in

Table 5.2 cannot be

In order to determine the parameters f

relative density levels, the values of f

and g for

DR

= 30% and 70% were

extrapolated and interpolated, respectively, from those corresponding to

90% which

intuition.

by Giuseppe (1991) and Vecchia (1991) were performed

relative density levels, the values of f

and g for other

agreement with

were obtained based on the measured modulus degradation

DR

= 50% and

relationship.

104

1.0

measured
0.8

calculated

0.6

0.4

o
UJ
UJ

0.2

0.0
0.0

0.2

0.4

(a'i

Figure 5.5

a' 3 )/(a'i

Modulus degradation curves


with

= 0.97 and ? =

0.18.

0.8

0.6

for

1.0

a' 3 )f

D R = 51.5%

and

a 3 = 400 kPa

105

1.0

measured

0.8

calculated

0.6

o
LU

LU

0.4

0.2

0.0

0.0

0.2

0.4

0.8

0.6

{0\-& 3 )/(0' -0 3

1.0

Figure 5.6

Modulus degradation curves


with f = 0.97 and g = 0.15.

for

)i

D R = 48.8%

and 03 = 200 kPa

106

1.0 *-

measured

0.8

calculated

Figure 5.7

Modulus degradation curves


with

f=

0.97 and g

0.18.

for

D R = 48.2%

and a 3 = 500 kPa

107

1.0

measured

0.8

calculated

0.6

o
HI

ai
0.4

0.2

0.0

0-0

0.2

0.4

0.6

(a\ - a' 3 )/(a\ -

Figure 5.8

Modulus degradation curves


with f =0.97 and g = 0.20.

for

0.8
<j'

1.0

3 )f

D R = 50.8% and a

10 kPa

108

1.0

measured
0.8

calculated

0.6
o
UJ
Jo

UJ

0.4

0.2

0.0
0.0

(ct'j

Figure 5.9

a' 3 )/(a' 1

Modulus degradation curves


with f = 0.93 and g

0.20.

0.8

0.6

0.4

0.2

for

1.0

a' 3 ) f

D R = 84.6% and o 3 = 650 kPa

109

1.0

measured
calculated

0.0

0.2

0.4

0.6

0.8

1.0

(CT'j-a^/Ca'j-aS),

Figure 5.10

Modulus degradation curves


with f =0.95 and g

0.25.

for

D R = 82.3% and o

= 100 kPa

110

1.0

measured
0.8

calculated

0.6
o
LU
CO

0.4

0.2

0.0

(o'i

Figure

5. 1 1

o' 3 )/(a\

Modulus degradation curves


with

=0.95 and g =

0.8

0.6

0.4

0.2

0.0

0.20.

1.0

a' 3 ) {

for

DR =

88.9% and o 3 = 200 kPa

Ill

1.0

measured

0.8

calculated

0.6
o
UJ
us
0.4

0.2

0.0

0.0

0.2

0.4

(o'i

Figure 5.12

0.6

- o\)l{o\ -

Modulus degradation curves


with

= 0.95 and g =

0.20.

0.8

1.0

o' 3 ) f

for

D R = 91.1% and a

= 150 kPa

112

1.0

measured

'

0.8

calculated

0.6

o
UJ
-a,

LU

0.4

0.2

0.0

0.2

0.0

0.4

0.6

(a\ - G' 3 )/(a\ -

Figure 5.13

Modulus degradation curves


with

= 0.95 and g =

0.25.

0.8

1.0

a' 3 ) f

for

DR =

100% and a 3 = 200 kPa

113

1.0

measured

0.8

calculated

0.6

UJ
HI

0.4

0.2

0.0
0.0

0.2

0.4

0.6

0.8

1.0

(a\ - a\)/(o\ - a' 3 ) f

Figure 5.14

Modulus degradation curves


with

= 0.95 and g =

0.27.

for

D R = 100%

and

o 3 = 400 kPa

114

1.0

measured

0.8

calculated

0.6

o
UJ

*"

UJ
0.4

0.2

0.0

0.0

Figure 5.15

0.2

0.4

Modulus degradation curves


with

= 0.94 and g =

0.8

0.6

0.32.

for

DR =

100% and

1.0

a?,

= 600 kPa

115

1.0

measured

0.8

calculated

0.6
o
UJ

UJ

0.4

0.2

0.0

0.0

0.4

0.2
(a'j

Figure 5.16

0.6

&

3 )/(o\

Modulus degradation curves


with

= 0.94 and g =

0.28.

0.8

1.0

o' 3 ) f

for

DR =

100% and c 3 = 800 kPa

116

1.0

measured

0.8

calculated

0.6
o
LU
111

0.4

0.2

0.0
0.0

Figure 5.17

0.2

0.4

0.6

Modulus degradation curves


with

= 0.94 and g =

0.20.

for

0.8

DR =

1.0

98.6% and a 3 = 100 kPa

117

Table 5.3 shows the values of f and g for


will

DR

= 30%, 50%, 70% and 90%.

be used for the analysis of calibration chamber plate load

tests

and

These values

pile load tests in

the next chapters.

Table 5.2

DR

Values of

and 2 from

triaxial test results.

o' 3 (kPa)

E (MPa)

51.5

400

367.1

0.97

0.18

48.8

200

266.0

0.97

0.15

48.2

500

396.6

0.97

0.18

50.8

110

207.2

0.97

0.20

84.6

650

563.6

0.93

0.20

82.3

100

243.7

0.95

0.25

88.9

200

344.2

0.95

0.20

91.1

150

307.8

0.95

0.20

100

200

369.6

0.95

0.25

100

400

501.4

0.95

0.27

100

600

599.3

0.94

0.32

100

800

680.2

0.94

0.28

98.6

100

270.5

0.94

0.20

(%)

Table 5.3

Values of

and g for different relative

densities.

Dr (%)

30

0.98

0.17

50

0.97

0.20

70

0.96

0.23

90

0.95

0.26

118

Plastic Stress-Strain Relationship for

5.5

5.5.

Three Dimensions

Drucker-Prager failure criterion

Mohr-Coulomb

In geotechnical engineering, the

used to describe plastic


failure criterion

is

The

soil behavior.

basic assumption in the

stresses are considered.

directly applied to the analysis of three-dimensional

account.

Mohr-Coulomb

problems

It

cannot, therefore, be

which the major and

in

stresses as well as the intermediate principal stress should be taken into

and post-failure

In order to describe failure

dimensional stress

As can be

state, the

soil

response for the three-

Drucker-Prager plastic model was adopted

seen in (5.12), the Drucker-Prager failure criterion

of the stress invariants

J2

and

given by (4.22) and

Ii

(4.9).

the failure surface in the Drucker-Prager plastic model,

it

When

is

in this study.

expressed in terms

the stress state reaches

should satisfy the condition

F = VA-(<tf +*0 =

where the parameters

a and K

strength parameter c

strain conditions, the

are given

and

parameters

<j>

by (5.13) and (5.14)

in

terms of the Mohr-

obtained from the triaxial conditions.

a and K are expressed

of:

(5.22)

Coulomb

been often

the plane-strain condition assuming two-dimensional stress state in

which only the major and minor principal

minor principal

failure criterion has

For plane-

as follows:

tan</>

a=

(9

+ 12tan

K=

-r
2

-r
2

(9

(5.23)

</>/

+ 12tan 0) /2

(5.24)

119

Failure criterion

JJl-(aI

+k:)

(a)

Mohr-Coulomb

Drucker-Prager

failure surface

failure surface

(b)

Figure 5.18

Drucker-Prager failure surface


(b) in principal stress plane.

(a) in l\-^[j 2

plane and

120

Since the stress invariants shown in (5.22) include


stresses, (5.22)

all

components of the three principal

can be used for the description of a failure condition under the three-

dimensional stress

states.

Figure 5.18 shows the failure surface defined by the Drucker-Prager failure criterion
in

both

h-^J^

plane and principal stress plane.

As shown

in Figure 5.18, the

Drucker-

Prager failure surface appears as a straight line in l\-*JJ 2 plane, and a smooth circle in

As

plane.

a result, the Drucker-Prager failure criterion

numerical procedure while the

computation due

to the corners of the

hexagon

readily incorporated into a

readily be used in numerical

in Figure 5.18 (b)

Non-linear failure surface and flow rule

5.5.2

The

failure surface given

as a straight line as

a.

Mohr-Coulomb cannot

is

71-

As discussed

shown

in

by the original Drucker-Prager

Figure 5.18 (a) with the Drucker-Prager friction parameter

peak

in chapter 4, the

with relative density and confining


(4.76).

Because the

failure criterion is defined

parameter

friction angle

stress.

in the

<t>

for sand

is

not constant, varying

This relationship was expressed in (4.73) -

Drucker-Prager failure criterion

is

obtained from

the peak friction angle using (5.13), the envelope of the Drucker-Prager failure surface

also non-linear.

As a

result, the failure surface

becomes steeper

is

as the level of confining

stress decreases.
It

has been widely recognized that the original Drucker-Prager plastic model with an

associated flow rule always causes a large negative volumetric increment,


dilational behavior (Desai
illustrated in Figure 5.19.

increment vector
5.19.

The

de/ be

i.e.,

and Siriwarden 1984, Chen and Baladi 1985).

The associated flow

and horizontal (de/") components

as:

This

is

rule requires that the plastic strain

perpendicular to the failure surface, such as

plastic strain increment vector

excessive

de/ can be decomposed

at

point

in Figure
5

into the vertical (de,/

121

de,'=de, m +de,"

The

vertical

component

component
pv

dij

dE\}

ps

represents the

(5.25)

plastic

The

shear-strain.

horizontal

represents the plastic volumetric-strain that has a negative direction for

an associated flow

This indicates that the plastic flow

rule.

Prager plastic model

is

always accompanied by an increase

in

in the original

Drucker-

volume.

In order to suppress unrealistic dilation in the plastic state, a non-associated

rule with the

von Mises

von Mises

plastic potential function

plastic potential function

Q. is the

von Mises

deviatoric stress tensor.

et al. 1989).

The

can be given by:

Q.

where

was adopted (Borja

flow

= t]J 2 -k

(5.26)

plastic potential function

and

Jt is the

second invariant of the

Figure 5.20 shows the non-linear failure surface used in this

study and plastic strain increment with the non-associated flow rule.

Incremental stress-strain relationship

5.5.3

The most common approach


the

incremental

According

to the

method

for applying plasticity theory to numerical analysis is

calculating

the

tangent

stiffness

for

plastic

condition.

concept of perfect plasticity theory, the increment of plastic strain

cannot be uniquely determined from the current stress state ay and stress increment do^.

The

stress

plastic

increment

strain

do;j,

however, can be obtained from a current

increment d\f.

This relationship

is

stress

ct,j

and a given

refereed to as the consistency

condition, which forces the stress state to remain on the failure surface, and given by:

dF^^da
3ov

t]

=0

(5.27)

122

JTl
d lJ

pv

Drucker-Prager
failure surface

Figure 5.19

Plastic strain in Drucker-Prager failure criterion with associated flow rule.

Non-linear
failure surface

I.

Figure 5.20

Non-linear failure surface with non-associated flow

rule.

123

The consistency condition of

(5.27) has already been discussed in Chapter 4 and

From

the conditions required for perfectly plastic behavior.

and

(4.67), the stress

increment

d(jjj

is

one of

the flow rule given by (4.66)

with Hooke's law can be written

as:

d0,=Cm (de H -d H ')

(5.28)

Thus,

d0 9

in

which

and

Q,

Qjio

elastic

= Von-Mises

=Cm dB H -M:m l-

moduli matrix;

dki

and

plastic potential function.

dki

(5.29)

total

and

plastic strain increment;

Plugging (5.29) into (5.27), X

is

obtained

as:

A=

(530)

j/ C jQ
drstu
*\

<7

Equation (5.30) indicates

that, for a

increment dy, the factor

A.

do,..

given material with a failure surface

can be uniquely determined through (5.30).

(5.30) into (5.29), the incremental stress-strain relationship

dQ.
" mn

do, = [Cm

da

mn
5-=

dF

is

F and

strain

Substituting

expressed as a form

of:

dF

da pq
fe
dh

pqkl

]de u

(5.31)

All indicies in (5.31) are based on index notation as used in Chapter 4.


relationship in a plastic state can then be defined using (5.31).
(5.31) represents the elastic-plastic tensor of tangent modulus:

The

The

stress-strain

coefficient tensor of


124

dF

dQ.

=C -

C*

C.

Since the function

and F are expressed

(532)

fe

dF

terms of the stress invariants, the

in

following relationships can be obtained:

dQ
d mn

dl

37,

#1

9o- mn

dF

dF

3/,

dQ

37,

37 2

3o- mn

_3F

37 2

37 2

do pq

dQ.

37,

37,

(5.33)

3a

and

do pq

in

which

Ii

the

first

37, 3o-

+
.

p9

dF
^^- +
.

L.

(5.34)

37 3 3<x M

invariant of stress tensor; and J 2

invariants of the deviatoric stress tensor.

37,

In (5.33)

and

and

J3

the second and third

(5.34), the derivatives of the

stress invariants are written as:

37,

=s.

37 2

3a

37,
do,,

=s

-s

(5.35)

(5.36)
h

ik

s*-

(5.37)

125

where

5jj

= Kronecker

delta;

and Sy = deviatoric

the derivatives of the functions Q.

and F with the

stress tensor.

From

stress invariants,

(5.22)

and

which appear

(5.26),

in (5.33)

and (5.34), can be obtained as follows:

AQ,

(5.38)

61,

do.

dJ 2

2^/77

(5.39)

a,Q

(5.40)

6J 3
and

dF
:

-a

(5.41)

d/,

dF

dJ 2

2^/77

(5.42)

dF

(5.43)

dJ 3

Then, (5.33) and (5.34) are rewritten

as:

dQ.

da mn

2^77

S mn

(5.44)

and

afriir
Now

the elastic-plastic tangent

"-^

modulus matrix of (5.32) can be given more

(5 - 45)

explicitly by:

126

As

Jh
2

tl

C kl " = Ca

(5.46)

a result, the incremental stress-strain relationship for the plastic state

da,} =

And

the plastic strain

is

obtained

C"ds u

component can be expressed

as:

(5.47)

as:

dE /=U-^=-ccS)

(5.48)

iJJl

in

which

5.6

A. is

given by (5.30).

Summary

The conventional hyperbolic


analyses which

may be used

stress-strain

relationship

for two-dimensional stress states.

a pile in the free field or a circular plate within a calibration

problem.

is

Although the analysis of such problem

three-dimensional problems,

it

still

is

much

to treatment using

is

This

is

an axi-symmetric

simpler than that of ordinary

requires the definition of stress states in

certainly different

two-dimensional stress-strain relationships and failure

soil

model takes advantage of

model

three

In that sense, an axi-

from the plane-strain condition, which

In this chapter, the non-linear elastic-plastic soil

presented.

Vertical loading of either

chamber

dimensions, including vertical, radial, and tangential stresses.

symmetric problem

based on plane-strain

is

suitable

criteria.

for three dimensions

was

the intrinsic and state soil variables that

127

can be uniquely determined for a given


soil

behavior before a failure condition

elastic

and condition.

stress-strain relationship

This non-linear elastic model represents changes of elastic

parameters (the shear modulus


stress invariants

For the description of

reached, the three-dimensional non-linear

is

model was developed based on the modified hyperbolic

by Fahey and Carter (1993).

The

soil type

were used

and the bulk modulus K) according

to the stress level.

The

to represent the three-dimensional stress state.

expressions for the variation of shear modulus and bulk modulus according to the stress
level are given by:

f*'^

^- = [!-/(

)"'

)'](f-

and

K,=D

The parameters

and g

in (5.49)

-(C7' m

r(P f-

nt>

were determined from

measured modulus degradation curves,

it

according to the relative density

As

level.

is

triaxial test results.

observed that the values of

From

the

and g vary

the relative density increases, the value of f

decreases while the value of g increases.


In order to describe failure
stress state, the
this study.

is

soil

response for the three-dimensional

Drucker-Prager plastic model with non-associate flow rule was adopted in

The Drucker-Prager

where the parameters


(which

and post-failure

failure criterion is given by:

and k are related to the Mohr-Coulomb strength parameter c

zero for cohesionless soils) and

Because the peak

friction angle for

sand

is

obtained from the triaxial conditions.

not constant due to the dilatancy, the friction

128

parameter

a was

critical state.

defined in terms of the dilatancy friction angle and the friction angle

at

129

CHAPTER 6 NUMERICAL ANALYSIS AND EXPERIMENTAL


INVESTIGATION OF CALIBRATION CHAMBER TESTS

6.1

Introduction

chamber

Calibration

tests

have been used to investigate both the load-settlement

response of the base of non-displacement piles and cone penetration resistance under a
variety of conditions (Parkin 1991,

be performed

at

Ghionna

any desired values of

et al. 1994).

relative density

and

Calibration chamber tests can


vertical

and horizontal stresses

under controlled conditions.

Based on the 3-D non-linear

elastic-plastic stress-strain

model presented

in

the

previous chapter, calibration chamber tests are modeled and analyzed using the finite

The

element approach.

analytical results will be

plate resistance in calibration

verify the accuracy of the

chamber

6.2

size effects

Calibration

Chamber

calibration

non-displacement

plate load tests.

model predictions for

the measured values of

The objective of this chapter

plate resistance

is to

and assess calibration

on plate resistance values.

Description of

6.2.1

chamber

compared with

test

Plate

and experimental procedures

chamber

pile.

Load Tests

plate load test can be used to simulate the axial loading of a

In such a test, a cylindrical

sand specimen

is

carefully prepared,

LVDT
Top

Plate

Rubber

Membrane

Rod
Steel

Casing

Load

Cell

Rigid Plate

Base

Membrane
Water-Filled

Base Cushion

Displacement
Transducer

Figure 6.1

Plate load test in calibration chamber.

131

consolidated to a desired stress


1999a).

order

In

and tested (Salgado

state,

to simulate

displacement pile under a variety of conditions, plate load


flexible calibration
to 1.5

m.

The

chamber (Figure

series of

30

tests

al.

1998a,b, Lee and Salgado

response

load-settlement

the

et

tests

of the base of a non-

were performed within a

6.1) with a diameter equal to 1.2

was previously described by Ghionna

and height equal

et al. (1994),

where

details regarding the experimental procedures are discussed at length.

The

samples were formed by pluviation, which was interrupted when the

test

surface of the sample reached mid-height, so that a rigid circular plate having a diameter

of 104

mm

inner rod

After the plate connected to an

could be positioned on top of the sample.

was

same diameter

positioned, an outer casing with the

sample prepared

conditions near the base of a non-displacement pile.

from the

soil

by the outer casing, there

is

was

Sample pluviation was

positioned above the plate and fixed to the calibration chamber.


then continued outside the outer casing.

as the plate

in this

manner simulates

Since the loading rod

is

the

separated

no side resistance along the push rod and

the

vertical load-displacement response of the plate simulates the load-displacement response

of the base of a non-displacement


After preparation

The

pile.

completed, the sample

is

relative density of each

consolidated under Ko-conditions.

sample was controlled by the intensity of the sand flow

during pluviation, and was accurately determined

was disassembled.

is

at the

test

when

the sample

tested.

The

test results

end of each

Both dense and medium-dense samples were

are presented later.

6.2.2

Test material and boundary conditions for calibration chamber plate load tests

The sand used


sand,

whose

in the calibration

properties are

are divided typically into

shown

two

in

chamber

Table

plate load tests

5.1.

different levels,

The

was Ticino

sand, a silica

relative densities used in the tests

medium dense and

dense.

The medium

dense and dense samples represent the relative-density levels equal to around

DR

= 50%

132

and

D R = 90%, respectively.

Both normally- and over-consolidated conditions were used

For over-consolidated

in the tests.

within the 2.73

soil

(OCR) were

desired confining stress levels were established by

The

7.61 range.

samples, the over-consolidation ratio

applying separately the vertical and lateral boundary stresses on the sample surfaces.

The

This allows different Ko values.

boundary

vertical

stresses

were

in the

range while the lateral boundary stresses were in the 25 - 235 kPa range.
the value of

used

in the tests

and stress conditions of the

test

chamber

In calibration

were

in the 0.34

samples used

- 0.97

chamber

four different types of boundary conditions can be

tests,

according to the types of boundary conditions imposed on the

chamber
used

Table 6.2 and Figure 6.2

surfaces.

in the calibration

from each

other.

chamber

BC4

lateral

lateral,

illustrate the types

conditions

top and bottom

of boundary conditions

and how these boundary conditions

plate load tests

Based on the

soil

plate load tests.

These boundary conditions include BC1, BC2, BC3, and

used.

Accordingly,

Table 6.1 shows the

range.

in calibration

62-513 kPa

boundary conditions

that has

differ

indeed significant

influence on the plate resistance or cone penetration resistance, the boundary conditions

can be classified into two categories, constant-stress and fixed boundary conditions.
constant-stress boundary conditions include

BC3

conditions while

BC2

and

conditions are categorized as the fixed boundary conditions.


It

should be noticed that none of these boundary conditions perfectly reproduces the

boundary conditions corresponding


calibration

chamber used

in a test

calibration

chamber

and

infinite sizes

useful as

tests

were used.

BC1 and BC4

has a limited

free-field tests

The other

tests

et al.

under

conditions (Salgado et

(1994) include 26

BC3

is

tests

conditions, and

used in each calibration chamber

test

test are

BC2

and

calibration

under

condition,

BC4

shown

in

results

6.1.

is

is

between

chambers with

conditions are not as

because achieving a

very difficult due to the

chamber

conditions.

Table

BC3

This

sample boundary

BC1

so because the

exist if calibration

1998b).

al.

The 30
under

that

is

The difference of

size.

would not

limitation

characteristics of flexible chambers.

Ghionna

This

to real field situations.

true no-displacement condition for the lateral

BC1 and BC4

The

test

plate load test

under

BC2

by

conditions,

The boundary conditions

133

Table 6.1

and

stress conditions in calibration

Boundary

DR

Condition

(%)

(kPa)

(kPa)

51.0

115.0

92.0
50.4

327

BC1
BC1
BC1
BC1
BC1
BC1
BC1
BC1
BC1
BC1
BC1
BC1
BC1
BC1
BC1
BC1
BC1
BC1
BC1
BC1
BC1
BC1
BC1
BC1
BC1
BC3
BC3
BC2

328
329

Test No.

300
301

302
303
!

Soil

304
305

306
307

308
309

310
311

312
313

314
315

316
317
318

319
320
321

322
323

324
325

326

chamber

tests.

Ko

OCR

51.0

0.443

1.00

115.0

40.0

0.347

1.00

113.0

51.0

0.451

2.73

55.2

313.0

140.0

0.447

1.00

55.2

214.0

92.0

0.429

1.00

58.4

512.0

223.0

0.435

1.00

49.4

116.0

76.0

0.655

2.70

92.8

314.0

123.0

0.391

1.00

92.5

216.0

85.0

0.393

1.00

92.5

115.0

77.0

0.669

2.73

90.9

66.0

25.0

0.378

1.00

91.7

63.0

25.0

0.396

1.00

55.5

513.0

235.0

0.458

1.00

51.7

62.0

26.0

0.419

1.00

49.1

54.0

52.0

0.962

7.61

43.5

410.0

189.0

0.460

1.00

56.8

512.0

184.0

0.359

1.00

55.2

62.0

24.4

0.393

1.00

59.0

216.0

92.0

0.425

1.00

58.0

65.0

53.0

0.815

6.34

59.3

63.0

50.0

0.793

6.54

60.2

412.0

177.0

0.429

1.00

92.2

314.0

129.0

0.410

1.00

92.0

66.0

64.0

0.969

6.27

91.2

66.0

27.0

0.409

1.00

91.4

116.0

47.0

0.405

1.00

91.2

65.0

25.0

0.384

1.00

90.9

65.0

26.0

0.400

1.00

BC4

90.9

65.0

26.0

0.400

1.00

BC1

90.6

65.0

26.0

0.400

1.00

Table 6.2

Boundary

Boundary conditions

Constant stress

= constant)

(o v = constant)

No

displacement

No

0)

constant)

Uh represents horizontal displacement

u v represents vertical displacement

displacement
(Uv

Constant stress

No

displacement

(0h

Top/Bottom Boundary

Constant stress

(u h

BC4

tests.

Condition

(u h

BC3

chamber

Condition

(o" h

BC2

Boundary

Lateral

Condition

BC1

in calibration

b
)

Constant stress
(o v = constant)

No

displacment
(u v

0)

135

ah

ah

-<y h

rrm
(a)

BC1

(b)

BC4

(d)

BC2

o,

LLLLA

IMI!
a,

(c)

Figure 6.2

BC3

Types of boundary conditions

in calibration

chamber

test.

136

6.3

Numerical Modeling of Plate Load Tests

Program

6.3.1

in Calibration

Chambers

ABAQUS

Finite-element analysis of calibration chamber plate load tests or pile load tests in
the free field requires an accurate

model of the

stress-strain relationship that

the complicated behavior of the soil around the rigid plate or pile base.

can represent

The

description

of a stress-strain relationship for soils, therefore, must take into account the non-linear,
stress-dependent stress-strain response before failure as well as the post-failure soil

behavior with non-linear strength envelope.

The commercial
Inc.,

Pawtucket,

R.I.)

finite-element program

was used

to

ABAQUS

model both the calibration chamber

the axial loading of non-displacement piles.

many

analyze

engineering problems.

It

The program

two

parts (Lee et

al.

(2)

implementation of history data.

Model information includes element

ABAQUS

The

has been used to

analysis procedure

in the analysis.

types, material definitions,

is

divided

The

and boundary conditions,

history data consist of analysis type and any

control parameter necessary for a non-linear solution procedure.


initial

plate load tests and

model information;

interpretation of

problems, the

Sorensen,

1999):

(1)

which are required

&

provides a set of material models available for

geotechnical problems and several types of elements.


in

(Hibbitt, Karlsson

condition

is

defined as a geostatic equilibrium

In

most geotechnical

state.

Instead of using the optional material model provided by the original program, a

subroutine

was

written for the

previous chapter.

3-D non-linear

stress-strain relationship described in the

137

Finite element

6.3.2

modeling of plate load

The load-displacement curve

test

chamber

for each calibration

was

plate load test

predicted numerically using the finite element method with the non-linear elastic plastic

model discussed

Figure 6.3 shows a typical finite element

earlier.

the plate load tests in calibration chamber.

This

finite

mesh used

model

to

element mesh was constructed

using the same dimensions as those of the real calibration chamber having a diameter
equal to 1.2

and height equal

to 1.5

The elements were eight-noded axisymmetric

m.

elements, necessary for modeling the three-dimensional loading conditions, with four
internal integration points.

In

ABAQUS,

the finite element

elements can be plotted using any visual angle desired.


in

The

model with axi-symmetric


finite

element mesh shown

Figure 6.3 was plotted with a visual angle equal to 180 for better visualization.

The

finite

element model for the calibration chamber plate load

6.3 consists of

two

different element groups,

i.e.

soil

test

shown

in

Figure

The

and casing elements.

elements were modeled using the 3-D non-linear stress-strain relationship written
specific subroutine.

Because the

steel

casing

is

relatively rigid

compared

casing

Since the axial load

how

is

the actual tests were performed.

applied only on the circular rigid-plate located

at the

of the calibration chamber, no significant friction between the soil and casing

However,

realistically,

elements.
relative

in order to simulate the calibration

thin-layer

The

The

element model was set to be fixed allowing no displacement

in the finite

throughout the analysis, as that was

to occur.

in a

to the soil, the

casing elements were modeled as a linear elastic material with very high stiffness.
steel

soil

interface

elements

chamber

is

middle

expected

plate load test

were used between the

soil

more

and casing

thin-layer interface elements have zero initial thickness, and allow the

movement of

Coulomb

friction

exceeds a

critical

the soil and casing,

i.e.

slippage.

The

interface elements follow a

mechanism, where slippage takes place when

the tangential stress

shear stress defined by a friction angle and normal stress acting on the

surface between the casing and soil.

138

Figure 6.3

Finite element

model

for calibration

chamber

plate load test.

139

As

discussed in the previous section, four different types of boundary conditions

were used

chamber

in the calibration

boundary conditions imposed on the

plate load tests.

finite

In the

numerical modeling, the

element mesh were those actually imposed on

the samples in the actual tests.

The values of
modulus, were

= 90%

the parameters f and g, which define the degradation of the elastic

= 0.97 and g = 0.20

level, selected

for

D R = 50%

level,

and

Tables 5.1 and 6.1 (C g n g e g


,

After defining
analysis

is

<j) c ,

e max e^n, e
,

DR

The only

other soil

which

are given

state variables,

o' v g\),

the required geometry information and material properties, the

all

performed

0.95 and g = 0.26 for

with basis on the values given in Table 5.3.

parameters required in the analyses were the intrinsic and


in

first

by checking the

given boundary condition and stress

initial

geostatic equilibrium condition for a

state.

Predicted and measured plate resistance

6.3.3

Figures 6.4
calibration

chamber

6.7

show

the graphical results of the numerical analyses for the

plate load tests.

All these results were obtained at the relative

settlement (defined as the ratio of the vertical settlement to the plate or pile base

diameter) equal to s/B

shown

in

= 10%.

The displacement of

the

deformed

finite

element mesh

Figure 6.4 was exaggerated by a magnification factor equal to 5 for better

visualization.

Figures 6.5 - 6.7 represent the vertical stress, vertical displacement, and lateral

displacement distribution, respectively.


it

can be seen that significant

near the plate edge.

From

the distribution of stress

stress concentration

Consequently,

it is

and shear

and displacement,

stress increases are

also expected that the

most

observed

significant reduction

of shear modulus would occur near the plate edge.

Figure 6.8 shows the variation of secant shear modulus with horizontal and vertical
distances from the plate for three different settlement levels.

As can be

seen in Figure

140

6.8

(a),

the reduction of secant shear

modulus

at the level

of the plate base

is

most severe

The shear modulus of

near the plate edge due to high shear stresses there.

the soil

immediately below the center of the plate undergoes rather slow degradation, as the
state is

dominated by the increase of confinement rather than the increase of shear

Such a slow reduction of shear modulus underneath

stress
stress.

the plate is in agreement with the

observation of the formation of an "elastic core" beneath the base of axially loaded piles

(BCP

1971; Salgado et

Figure 6.8 (b) shows the degradation of the secant

1997a).

al.

shear modulus with depth along the plate axis.


1),

the shear

modulus

initially

At

the early stages of loading (e.g. curve

down

decreases with depth

to a depth of around 10

(equal to the plate diameter), and then increases again.

down

reductions can be seen


increase of shear

zones (curve

modulus

is

to the depth of 10

Figures 6.9 - 6.16

approximately 10

plate, but this

time no

Similar observations of modulus reduction with depth from the

2, 3).

chamber

below the

loads, similar

observed below that depth, due to the expansion of the shear

analyses of footings were noted by Fahey et

calibration

cm

For higher

cm

show

plate

mm,

(1994).

measured and predicted load-settlement curves of the

the

load

al.

The curves extend up

tests.

to

a settlement

of

corresponding to a relative settlement level equal to s/B = 10%.

The predicted load-settlement curves obtained by

the finite

element analyses compare

favorably with the measured responses in most of the cases.

Figure 6.17 shows measured versus predicted plate


settlement levels equal to s/B

agreement

is

5 and

10%

very satisfactory, showing a

for

calibration

all

maximum

dilational response of

and g were obtained) and plate load

sand under
test

triaxial

This

load

at

the

chamber

tests.

20%.

There

error of about

underestimation of plate resistance for dense sand.

between the

unit

may

relative

Overall
is

a slight

be due to differences

loading conditions (from which f

loading conditions.

141

Figure 6.4

Deformed

DR

finite

= 55.2%

g'v

element mesh of calibration chamber plate load

62.0 kPa, and

&

= 24 .4 kPa

at

s/B

= 10%.

test

with

142

Figure 6.5

Vertical stress distribution in calibration

DR =

55.2%

g' v

= 62.0 kPa, and

o' h

chamber

= 24.4 kPa

plate load test with

at s/B

= 10%.

143

Figure 6.6

Vertical displacement distribution in calibration

chamber

D R = 55.2%

at

o'v

= 62.0 kPa, and

a' h

= 24.4 kPa

s/B

plate load test with

= 10%.

144

Figure 6.7

Horizontal displacement distribution in calibration chamber plate load

with

DR =

55.2%

o'v

= 62.0 kPa, and

a' h

= 24.4 kPa

at

s/B

= 10%.

test

145

35000

at plate

^__i

base level
"""""^

30000
co

0.
V)

25000

''

+~

.-'*

"'
_'

O 20000

.c

15000

//

A'*,,.

10000

."

/
/

m
c
CO
o

*'
*"

Sy

^;_.^-

..A*

.*

03

5000

,"'

/
CO
0)

."

- s/B

1%

-s/B

2%
3%

"

'

- .

s/B =
..

'

50

A-

100

150

Horizontal Distance

From

(a) at the pile

200

Plate Center,

250

mm

base level

Secant Shear Modulus, kPa


10000

20000

30000

along plate axis ,-v


50
TO

a.

100

E
2
u.
<u

150

o
c
CO

to

5
b

200

CO

>

2%
= 3%
= 1%

s/B =

250

s/B
s/B

300

(b) along the plate axis

Figure 6.8

Variation of shear modulus.

40000

'

146

2500

Test No.

2000

300
..

1500
o
c 1000
3

Q)
ffl

500

a.

r
'

fill
-

10

12

10

12

Settlement (mm)
3UUU

load

Test No. 301

4000
Oo o

unit

Oo o

J5
a.

1000

^^\.

'
-

n
4

Settlement (mm)

2500
A3

a.

Test No. 302

2000

..

o
ro

1500

E 1000
3
JO
0.

500

"

^fT

r
i

10

12

Settlement (mm)

-Measured

Predicted

Figure 6.9 Load-settlement curves for calibration chamber plate load tests
(Test No. 300, 301 and 302).

147

Test No. 303

to

0.
2C

load

unit

4000
ooo

.'

ooo

0)

to

1000

Q.
J.

..

10

12

Settlement (mm)

Test No. 304


!*

3000

2000

(0

^^

'E

^'

1000

<0

Q.

_i

10

12

10

12

Settlement (mm)
3000

2500

Q.
T3

Test No.

306

2000

<0

1500

'E

3 1000
y<f

0)

500

Settlement (mm)

Measured

Predicted

Figure 6.10 Load-settlement curves for calibration chamber plate load tests
(Test No. 303, 304, and 306).

'

'

'

148

OUUU
Test No. 307

T
X.
^-

^^^

6000

^^^"^

""

^^\
^^^

TJ

TO

^^*

4000

,*>r.

^%'

c
3

y/f'
2000

sT

jt

CO

Q.

10

12

Settlement (mm)

euuu
Test No. 308

"to

6000

a
TO

"

4000

^--*^*V--

*"
'

'E

~TO

2000

0.

''
i

10

12

Settlement (mm)
DUUU
"to"

5000

4000

Test No. 309

TO

3000

'E

3 2000
a

*-*

TO
jj;

1000

10

Settlement (mm)

Measured

Predicted

Figure 6.11 Load-settlement curves for calibration chamber plate load tests
(Test No. 307,

308 and 309).

149

4000
Test No. 310

1?
(k

ooo

load

oo o

yT

unit

.-

ooo
Plate

Settlement

10

12

10

12

(mm)

4000
Test No. 311

CO

a.

3000

a
CO

2000

c
3
0)

1000

Settlement

(mm)

Test No. 312

(0

6000

TJ
CO

..--''
4000

'E

2000

(0

0.

1.

10

Settlement

Measured

12

(mm)

Predicted

Figure 6.12 Load-settlement curves for calibration chamber plate load tests
(Test No. 310,311 and 312).

'

'

150

2000
Test No. 313

CO

^
D

1500

CO

1000

c
3
CD

S
CO

500

10

12

Settlement (mm)

Test No. 314

CD

Q.

..-

1500

T3

<*"^

CO

1000

500

^*"T^
/

a.

10

12

Settlement (mm)

Test No. 317

1500

(0

1000

^s^\

'

f^-

3
J2

500

(0

a.

10

12

Settlement (mm)

Measured

Predicted

Figure 6.13 Load-settlement curves for calibration chamber plate load tests
(Test No. 313,

314 and 317).

"

151

5000

4000

ro

3000

2000

2
n

Test No. 321

1000

'

_i__

10

Settlement (mm)
8000
TO

a.

6000

(0

4000

'E

2000

10

12

Settlement (mm)

a.

D
TO

Test No. 323

4000
3000

/ '

2000
TO

.-

1000

a.

1_

10

12

Settlement (mm)

Measured

Predicted

Figure 6.14 Load-settlement curves for calibration chamber plate load tests
(Test No. 321, 322 and 323).

'

'

152

Test No. 324

"JsT

a
o
o

3000

CO

^^-

2000

'E

'

1000

JO

_^

...

10

12

Settlement (mm)

a.
ji
ro

4000

3000

Test No. 325

^s* ^-

/^.-''

_o

C 2000
3

jr

0>

1000
Q.

10

12

10

12

Settlement (mm)
4000

326

Test No.

CO

0-

3000

cc

"

2000
'c

3
0)

s
CO

1000
jr

'

Settlement (mm)

Measured

Figure

6.

Predicted

15 Load-settlement curves for calibration

(Test No. 324, 325 and 326).

chamber

plate load tests

'
'

153

a!
J
re

Test No. 327

4000
3000

^^^

C 2000
3
1000
Q.

S/

10

12

Settlement (mm)

Test No. 328

"re"

^
a

3000

re

*-*

-^\7--

2000

'

'E

1000

re

a.

/'

10

12

Settlement (mm)

Test No. 329

re

S
^^
D
o

+*

3000
^_

re

^s^

"

.*

^ ,
-

^,0^

>^

^*"****'^**^

J
j*"**"^

C
3
Sj

'

^t^* "^

2000

1000

(0

J^

10

12

Settlement (mm)

Measured

Predicted

Figure 6.16 Load-settlement curves for calibration chamber plate load tests
(Test No. 327,

328 and 329).

154

ouuu

D R = 90%

TO

0.

5000

D R = 50%

73

re

o 4000

.-"

.*

"

.0

o 3000

o.

re

a.

2000

.-o

0)

0*

"

"-

o-o

V 1000

.no

s/B

5%

'

'

2000

1000

3000

4000

5000

Measured Plate

(a) s/B

Unit Load,

6000

kPa

5%

8000

DR

re

0.
JC

= 90%
*

6000

O D R = 50%

re

,'

_l
-

'c
13

o 4000

.-&

4"1

'

re

Q.
.s>'*

TJ

2000
o

o. ymj
o

o*

a-'
.

s/B

a.

= 10%

4000

2000

6000

8000

Measured Plate Unit Load, kPa


(b)

Figure 6.17

Measured and predicted

s/B

= 10%

plate unit loads in calibration

chamber

tests.

155

6.4

Calibration

6.4.1

Chamber

Size Effects on Plate

Load Test Results

Definition of size effect

Because of
field conditions.

their finite size, calibration

As

chambers do not perfectly reproduce

measurements made

a result,

specimen are different from what would be observed


density and stress state.

chamber; hence the term


Calibration

cone penetration
1991;

These

size effects

where

Mayne and Kulhawy

testing.

It

chamber sand

in the field for the

same

relative

be more pronounced, the smaller the

have been extensively studied

large deformations take place (e.g.,

1991; Salgado et

al.

in

connection with

Schnaid and Houlsby

1998a), but not with respect to plate load

has been determined that size effects in calibration chambers are more

significant in dense sands than in loose sands.

induced

a calibration

size effect.

chamber
testing,

difficulties tend to

in

free-

in the

sand specimen are

much

In plate load testing, the displacements

less than those

caused by cone penetration.

the other hand, the plate diameter used in the study (100

mm)

is

much

On

larger than typical

cone diameters (most commonly 35.7 mm).


For the

chamber

results of calibration

tests to

be effectively applied to pile design,

the possible size effect should be properly addressed.


calibration

chambers can be used

If size effects are

well understood,

to experimentally assess pile base resistance

under

controlled conditions.

6.4.2

Investigation of size effects for different boundary conditions

In order to investigate size effects in calibration

chamber

plate load tests, numerical

analyses of both full-scale non-displacement piles and calibration chamber plate tests

were performed.

For studying the loading of non-displacement

diameter piles with lengths

L=

m, 10

and 20

were analyzed.

piles,

three

60-cm

The corresponding

156

ratios

IVB of

the length

The

respectively.

values of

DR

piles

to the diameter

B were

equal to approximately

were positioned within a granular

30, 50, 70 and

90%.

The

initial vertical

and

soil

8,

16 and 33,

deposit with

assumed

lateral effective stresses at

the pile base level for each pile length are as follows:

g'

= 100 kPa and

o' h

= 43 kPa

for the

5-m

(2)

g'v

= 200 kPa and

a' h

= 86 kPa

for the

10-m

(3)

a' v

= 400 kPa and

o' h

= 172 kPa

(1)

for the

pile
pile

20-m

pile.

All of these states correspond to a normally consolidated condition.


calibration

and

chamber were

stresses.

also simulated numerically for these

Comparison of

unit load in the calibration

Plate load tests in a

same values of densities

the base unit load in the full-scale pile load tests with plate

chamber

same

tests for the

the basis for conclusions regarding size effects.

relative settlement level provides

The sand adopted was Ticino

sand,

whose properties were discussed previously.

The comparison of

pile load tests

and chamber

tests

conditions that were defined in Table 6.2 and Figure 6.2.

show

were made for

Tables 6.3

pile base unit loads versus plate unit loads in calibration

BC2, BC3, and


effect in

BC4

Tables 6.3

boundary

all

6.6 and Fig. 6.18

chamber

tests

with BC1,

conditions for the relative settlement s/B equal to 10%.


6.6

was defined

as a ratio of the calibration

chamber

The

size

plate resistance

to the free-field pile base resistance.


In Figure 6.18, the different points for the
relative densities
results for

BC3

BC4

DR

same

= 30%, 50%, 70%, and 90%.

conditions were similar to those for

conditions were similar to those for

BC2.

This

pile length represent the results for

As can be

BC1

seen in Figure 6.18, the

conditions, and the results for

illustrates that the lateral

boundary

condition has an overwhelming influence on plate unit load measured in a calibration

chamber.

157

Table 6.3
Pile

S ze effect

in calibration

Load Test

Calibration

Pile length

DR

(m)

(%)

30

1516

50

2066

70

2794

90

3631

30

2152

10

20

q b (kPa)
(s/B

= 10%)

condit on.

Chamber Test

Size

C7' v

DR

(kPa)

(%)

100

30

1286

84.8

50

1762

85.3

70

2361

84.5

90

3087

85.

30

2018

93.8

50

2699

92.0

200

q b (kPa)
(s/B

= 10%)

Effect

(%)

2933

3869

70

3528

91.2

90

4959

90

4497

90.7

30

3112
4162

30

3152

100.1

50

50

4120

98.9

70

5404

70

5262

97.4

90

6836

90

6623

96.7

400

Size effect in calibration chamber

Load Test

Calibration

Dr

(m)

(%)

30

1516

20

BC1

50

Pile length

10

test for

70

Table 6.4
Pile

chamber

= 10%)

BC2

condition.

Chamber Test

DR

q b (kPa)
(s/B

test for

q b (kPa)

= 10%)

Size
Effect

(%)

(kPa)

(%)

100

30

1555

102.6

50

2085

100.9

99.2

(s/B

50

2066

70

2794

70

2772

90

3631

90

3600

99.1

30

2152

30

2364

109.8

50

2933

50

3108

105.9

70

3869

70

4017

103.8

90

4959

90

5119

103.2

30

3112

30

3591

115.3

50

4625

111.1

200

400

50

4162

70

5404

70

5854

108.3

90

6836

90

7338

107.3

158

Table 6.5 Size effect


Pile

Load Test

Pile length

DR

(m)

(%)
30

10

20

= 10%)

(s/B

1516

(kPa)

100

BC3

condition.

Chamber Test

(%)
30

q b (kPa)

= 10%)

(s/B

Size
Effect

(%)

1534

101.1

50

2053

99.3

50

2066

70

2794

70

2717

97.2

90

3631

90

3521

96.9

30

2152

30

2333

108.4

50

2933

50

3063

104.4

70

3869
4959

70

3945

101.9

90

90

5017

101.1

30

3112

30

3549

114.0

200

400

50

4162

50

4564

109.7

70

5404

70

5774

106.8

90

6836

90

7213

105.5

in calibration

Load Test

chamber

Calibration

DR

(m)

(%)

30

1516

50

2066

20

test for

Dr

q b (kPa)

Pile length

10

chamber

Calibration

Table 6.6 Size effect


Pi e

in calibration

q b (kPa)
(s/B

= 10%)

test for

BC4

condition.

Chamber Test

Size
Effect

a' v

DR

(kPa)

(%)

100

30

1348

88.9

50

1828

88.4

q b (kPa)
(s/B

= 10%)

(%)

70

2794

70

2422

86.7

90

3631

90

3156

86.9

30

2152

30

2106

97.8

50

2933

50

2789

95.1

70

3869

70

3630

93.8

90

4959

90

4602

92.8

30

3112

30

3272

105.1

50

4162

50

4249

102.1

70

5404

70

5402

99.9

90

6836

90

6778

99.1

200

400

159

ouuu
..-'
.*

CO

6000

.-"

*-'

a
.*

CO

o
c 4000
3

7**

0)
(0
CO

A
t

43

=
Q.

*'

5m pile

2000

10

m pile

20

m pile

-"

.-''

2000

4000

6000

8000

Plate unit load (kPa)


(a)

BC1

conditions

8000
.'

TO

Q.

6000
.-**

TO

_.-

**

'E

4000
#

s
0)

2000

jt*

5m pile
10

m pile

20

m pile

2000

4000

6000

8000

Plate unit load (kPa)


(b)

BC2

conditions

Figure 6.18 Comparison of pile base unit load with plate unit load

chamber

plate load tests.

in calibration

8000

re

0.

6000

a
re

"E

4000

-i

05

5m pile
2000

10 mpile

20

4000

2000

m pile

6000

8000

Plate unit load (kPa)


(c)

BC3

conditions

8000

re

Q.

6000

T3
re

O
3

4000

5m pile

re

2000

10

m pile

20

m pile

a.

4000

2000

6000

8000

Plate unit load (kPa)


(d)

Figure 6.18

BC4 conditions

Comparison of pile base

chamber

unit load with plate unit load in calibration

plate load tests (continued).

161

Differences between pile base unit load and plate unit load were small for
relative density levels.

It

is

observed, however, that the plate unit loads under

all

BC1

conditions were smaller than the pile base unit loads, and more substantially so in the

case of short piles, for which confinement

low

is

differences between pile base unit load and plate unit load were around 14 to
5

m, 6

to

The

BC1

9%

L=

for

results for

10 m, and

BC2

to

3%

for

conditions, as

15%

for

L=

L = 20 m.

shown

in

Figure 6.18

(b), differ

from those for

conditions in that the plate unit loads from the calibration chamber plate load tests

were found

to be greater than pile base unit load values

The

differences were around

L =

20 m.

to

3%

for

L=

m,

3 to

under corresponding conditions.

9%

for

L=

10 m, and 7 to

This indicates that the differences are more significant for long

which confinement

6.5

The

level of the pile base.

at the

is

high

at

15%

for

piles, for

the pile base level.

Summary

Calibration

chamber

plate load tests

have been used

settlement relationship of non-displacement piles.

chamber

plate load tests

were analyzed through the

In

finite

dimensional non-linear elastic-plastic model presented

to investigate the base loadthis

chapter, the calibration

element method using the three-

in the previous chapter.


al.

compared with

The

predicted load-settlement

curves showed good agreement with measured load-settlement curves.

between computed and measured plate unit loads for the

series

(1994) were modeled and

of calibration chamber tests performed by Ghionna et


the results of finite element analyses.

The comparison

relative settlement of s/B

and 10% was also made for both dense (D R = 90%) and medium dense (D R = 50%) sand.
Predicted plate unit loads were in good agreement with measured results, showing
relative errors not larger than about

Calibration

chamber

20%

at

s/B

= 10%.

size effects, resulting

from the

also investigated for different relative densities and

finite size of the

chamber, were

boundary conditions using

finite

162

element analysis.

The comparison was made between

pile

base unit load for piles

loaded under field conditions and plate unit loads from the calibration chamber

The

piles

were modeled with three

densities.

Plate unit loads in calibration

and higher

(for

BC2, BC3) than

pile

chamber

size effects

longer piles.

under

BC2

at

simulate pile base loading.

(1998b) regarding
or

BC4

results

be used

show

CPT

The

stress level at the

size effects

under

to shorter piles, while

size effect is small, validating the use of

In

chamber

tests to

keeping with the recommendation of Salgado

chamber

that, for practical

chamber

BC4)

high confinement, corresponding to

testing in calibration chambers,

in calibration

The confining

calibration

at

different relative

tend to be lower (for BC1,

low confinement, corresponding

were more pronounced

The magnitude of

tests

base unit loads.

pile base level also influences size effect.

BC1 were more pronounced

and four

different pile lengths

tests.

it is

plate load testing.

purposes, no correction

is

recommended
If this is

et al.

BC1

here that

done, our numerical

needed to the measured

plate

unit load in estimating pile base unit load, unless very short piles are being simulated.

163

CHAPTER 7 DETERMINATION OF PILE BASE RESISTANCE

7.1

Introduction

The standard penetration


most popular methods
1998).

test

(SPT) and the cone penetration

to obtain

SPT blow

quasi-static pile loading process, that of the static

pile load test.

Ghionna

et al.

(CPT)

for pile design using in-situ test results (Bandini

While the process followed

pile loading process.

test

The

test is

counts

is

are the

and Salgado

not well related to the

cone penetration

is

better related to the

performed quasi-statically and resembles a scaled-down

According to many authors

(e.g.

De Beer

1984; Franke 1989,1993;

1993, 1994; Jamiolkowski and Lancellotta 1988; Fioravante et

cone penetration resistance

two

may be

used as a proxy for limit base resistance in

al.

1995),

piles.

In this chapter, the base load-settlement curves of axially loaded piles bearing in

sand are obtained for different stresses and densities using the
a non-linear elastic-plastic stress-strain model.

same

soil

Cone

finite

resistance q c

element analysis with


is

determined for the

conditions from the penetration resistance analysis of Salgado et

using the program

CONPOINT

(Salgado 1993; Salgado

1998a); q c values determined in this


settlement curves.

The

fully

relative settlement s/B can be

manner

et al.

to

(1997a)

1997a, b; Salgado et

al.

are used to normalize the pile base load-

developed load-settlement curves


used

al.

in

terms of qb/q c versus

determine the normalized pile base resistance qt/q c

for any settlement-based design criterion.

164

7.2

Methods

If a pile

for Investigating Load-Settlement

Response

were continuously pushed down into a homogenous granular

would eventually reach a condition of penetration


cone penetrometer does
interest in pile design,
test

it

in a

cone penetration

at

test

a constant load in the

(CPT).

This

is

soil

mass,

same way

it

as a

a condition of limited

and nearly always impossible to establish with conventional load

procedures and equipment, unless the pile

is

relatively small in diameter

and length,

not bearing in a very strong soil layer, and the engineer specifically requires a

is

plunging load

Typical load-settlement curves for piles embedded in sand show

test.

gradually increasing curvature rather than a clear peak load.

Three approaches are possible

Those include

a pile.

modeling.

to obtain the vertical load-settlement relationship for

chamber

full-scale pile load tests, calibration

tests

and numerical

Full-scale pile load tests are the best option to investigate the load-settlement

relationship for a specific site and pile, but cannot typically be used to obtain an accurate

between base resistance and the

correlation

Only

soil state.

fully instrumented load tests

with significant efforts to characterize the soil around the pile could potentially be used
for that purpose.
relative density

Difficulties that

and

would

lateral stress are

still

need

known

not

to be addressed include the fact that

in the field,

and

that typical natural

granular deposits tend to be variable (Lee and Salgado 1999b).


Calibration
testing.

chamber

testing

and numerical analysis are more flexible than

variety of stress states, densities

Calibration

chamber

and boundary conditions can be considered.

plate load tests, as discussed in the previous chapter, have been used

to investigate the load-settlement response of

1994, Lee and Salgado 1999a).


for the
finite

pile load

non-displacement piles (Ghionna

number of numerical techniques have been

same purpose (Desai and Christian 1977, Lee

element method

is

among

the

most popular, as

et al.

it

et al.

identified

1989, Poulos 1989).

The

allows modeling of complicated

non-linear soil behavior and various interface conditions, with different geometries and
soil

conditions (Lee and Salgado 1999c).

the use of a relevant constitutive model,


soil

behavior.

key element

in a finite

element analysis

is

which should model the strongly non-linear

In the present study, the non-linear elastic-plastic stress-strain presented in

165

chapter 5

is

used in a

finite

element analysis to determine the load-settlement response of

vertically loaded piles.

7.3

Finite

Element Modeling of Pile Load Test

The commercial
loaded

The

piles.

finite

ABAQUS

element modeling of pile load

curves was done in the same

chamber

element program

finite

way

was used

to

model

vertically

tests for obtaining load-settlement

as used in the investigation of size effects in calibration

Instead of using one of the material models available in the program, a

tests.

subroutine was written for the non-linear elastic-plastic model described previously.

Because the

made of

compared

pile stiffness is very large

to soil stiffness, the pile

was assumed

Eight-noded axisymmetric

linear elastic material throughout the analysis.

elements with four internal integration points were used to model both the
pile.

Thin layer interface elements with zero

used between the pile and the

soil.

The

showed

the calibration

Harris 1993)

compared.

The

chamber

al.

(1991).

plate load tests in the previous chapter

In order to further assess the

element analysis, a pile load

(Mayne and

19.5

thickness, allowing slippage, were

the validity of the proposed finite element analysis of the pile base load-

settlement response in sand.


finite

and the

necessity of interface elements in the analysis of

axially loaded piles is discussed in Trochanis et

The analyses of

initial

soil

test site

test

performed by the Georgia

was modeled, and

the numerical

has a layer of residual,

underlain by partially-weathered rock

bedrock.

performance of the proposed

silty silica

down

Grain size distribution analysis showed the

sand, with the clay fraction under 10%.

to

Institute of

and experimental
sand extending

20 - 24.8

soil to

soil

results

down

to 15.8

be composed of about

series of laboratory tests

on the

property profile with depth and layers used in the analysis.

were

and then sound

collected at several depths were performed to obtain basic soil properties.

shows the

Technology

soil

70%

samples

Table 7.1

166

Table 7.1

Basic

soi 1 properties

used

in finite

element analysis.

Layer No.

Depth (m)

Ko

a' V o(kPa)

eD

0-1.82

34

0.44

18.2

0.70

23642

1.82-3.93

34

0.44

53.8

0.70

43838

3.93-5.93

37

0.40

87.6

0.73

54056

5.93-7.93

33

0.46

119.9

0.73

67081

7.93

- 9.93

32

0.47

151.8

0.74

76051

32

0.47

185.6

0.72

87849

9.93-11.93

(kPa)

11.93-13.93

36

0.41

219.9

0.70

96102

13.93 - 14.93

38

0.38

244.8

0.67

100472

14.93

- 16.76

36

0.41

268.6

0.67

112540

10

16.76-18.28

36

0.41

296.8

0.67

119145

167

The bottom boundary of


from the
18.28

surface, at

the finite element

mesh was

which the bedrock was encountered.

located at a depth of 21.93

The

last layer

of the

mesh from

m to 21.93 m was the partially weathered rock layer, which was assumed to behave
Values of

as an elastic material.

shear modulus

initial

for each soil layer

calculated based on Hardin and Black's equation for angular sand.

Hardin and Black equation and the observation made by Salgado

changes

in stiffness as a function of fines content, the values

wereC =214,

According to the

et al.

(1999) regarding

adopted for

ec

the

model parameters

and g of the non-linear

elastic plastic

were taken as 0.98 and 0.05 respectively, based on the observed non-linear
properties of silty sand (Salgado et

76

cm

were

and n

e =2.97 and n =0.5.

The values of

to

and a length equal

al.

to 16.8

1999).

m.

The

test drilled shaft

model
elastic

had a diameter equal

Measured and predicted base load-settlement

curves were plotted together in Figure 7.1.

Overall, agreement

is

observed to be

satisfactory.

7.4

Cone

Penetration Resistance from Cavity Expansion Analysis

number of methods have been proposed

analytically

and experimentally to

correlate the cone penetration resistance q c to stress state and soil conditions (Terzaghi

1943, Vesic 1972, Durgunoglu and Mitchell 1975, Baligh 1985,

Salgado 1993, Salgado

et al.

are generally classified into

1997a).

For undrained clay and

(Yu and Mitchell 1998):

(1)

Bearing capacity theory

(2)

Cavity expansion theory

(3)

Steady

(4)

Incremental finite-element analysis

state

(5) Calibration

deformation

chamber

testing.

Yu

and Houlby 1991,

fully drained sand, those

168

measured
....... calculated
TO

0.

o
(0

10

20

15

Base Settlement (cm)

Figure 7.1 Load-settlement curves for pile load

test at

Georgia Tech.

169

useful penetration resistance theory should be able to relate values of cone

penetration resistance q c to the intrinsic and state soil variables (Salgado et

The

cavity expansion solution by Salgado (1993)

is

1997a).

al.

one of the well-validated theories for

obtaining the cone penetration resistance q c and takes advantage of utilizing the intrinsic
,

and

state soil variables.

According

to

Salgado

et al. (1997a), for clean,

uncemented

soils,

the cone penetration resistance q c can be expressed as follows:

qc

in

which q c = function containing

o' v and a' h

pushed into

soil,

to zero.

The

in-situ vertical
it

creates

=q c {DR ,cr:,(j'h )

(7.1)

DR

intrinsic variables;

and horizontal effective

relative density of sand;

stress.

and expands a cylindrical cavity

When

and

a penetrometer

is

that has an initial radius equal

cavity expansion analysis by Salgado (1993) allows the calculation of cone

penetration resistance q c based on the cavity expansion resistance required to form such a
cylindrical cavity.

Figure 7.2 shows different types of assumed failure mechanisms proposed for deep
penetration.

It

should be noticed that the mechanism shown in Figure 7.2

(a), (b)

and

(d)

are theoretically impossible for soils with the usual value of the coefficient of lateral earth

pressure ratio

Ko

that lies in the 0.3

violate the path of least work.

1.0 range.

This

is

so because those mechanisms

In other words, the slip lines should

be directed towards

the side since the values of lateral stresses are smaller for these values of
vertical stresses.

The mechanism of Figure 7.2

Based on the observation of


Salgado

(1993)

proposed

the

slip

pattern

on the

expansion pressure p L

slip line

that of

kinematically possible.

the displacement field under the pile base

axisymmetirc condition of cone penetration.


principal stress

(c) is

Ko than

shown

in

Figure

In Figure 7.3, 0\

under cone penetrometer and

7.3
P

is

(BCP

1971),

considering

the

represents horizontal
related to the cavity

170

(a)

(b)

Terzaghi (1943)

Oe Beer (1948)
Hu (1966)

ill

(O
Berezantzev
Vesic (1963)

Figure 7.2

et ai. (1961)

Different failure

Biarezetal. (1961)

Hu

mechanisms

(1966)

for

deep penetration.

171

d.

Figure 7.3

Slip pattern under cone penetrometer (after Salgado 1993).

172

The

calculation of cavity expansion pressure p L requires the numerical sequence

considering plastic, non-linear elastic and elastic stress region around the cavity.

The

relationship between cavity expansion pressure

p L and cone penetration resistance q c was

established based on stress rotation analysis.

Following Bolton (1979), the major

principal stresses in

angle

two

different zones

shown

Figure 7.4 are related for the rotation

in

Avj/ as:

a? = afe 2 * v *a *

in

which 0\

and Gi

rotation angle
angle.

For the

= major

principal stresses in zones

between different principal stresses


stresses in Figure 7.3, (7.2)

<7

P
o"i

is

in

=cr

zones

and

in Figure 7.4;

and B; and =

Ay = Till

Man *

representative friction angle in the transition zone


respect to the projected area of cone

tip,

stress

Ay =

friction

as:

(7.3)

related to the cavity expansion pressure

determined numerically for a given

can be written with

where

(7.2)

pL

Q
;

0"i

is

related to q c

Figure 7.3.

in

and

tyr is

Integrating G\ Q with

the cone penetration resistance q c can be


state

and

soil

More

conditions.

details

regarding the procedure to compute q c using cavity expansion analysis can be found in

Salgadoetal. (1997a).

The cone

penetration resistance q c obtained from this procedure will be used to

normalize the pile base resistance using the program


procedure for cavity expansion solution (Salgado

et al.

CONPOINT

containing the

1997a, b, Salgado et

al.

1998a,

b).

173

Zone B

Zone

Figure 7.4 Stress rotation between different zones (after Salgado 1993).

174

7.5

7.5.1

Determination of Base Resistance for Non-Displacement Piles

Load-settlement response for various

The

soil

conditions

pile base load-settlement response for various soil

and

can be

stress conditions

obtained by analyzing axially loaded piles with different pile lengths and relative

The

densities.

pile

dimensions and

soil

conditions used for obtaining load-settlement

curves in this section are the same as used in the investigation of size effect in calibration

chamber

Chapter

test in

equal to 60

cm were

6.

used.

Three

unit

pile base level (corresponding to 5

400 kPa,

weight = 20

m, 10

kN/m\

initial vertical stresses at the

and 20 m) are a' vo = 100 kPa, 200 kPa and

were positioned within a granular

consolidated Ticino sand with K

0.43.

soil deposit

selected because

and has been used

(Ghionna

in

et al.

calculations were

it

assumed

as

normally

K value of 0.43 was selected because

near the center of the typical range of 0.39 to 0.47 observed for

tests

with a diameter

respectively.

All the piles

was

Different pile lengths imply different confining stress levels

For an assumed

at the pile base.

m, 10 m, and 20

pile lengths of 5

K,, in

sands.

it is

Ticino sand

has been studied extensively (Salgado 1993, Bellotti et

al.

1996)

hundreds of calibration chamber plate load and cone penetration

1994, Salgado

et al.

Values of relative density

1997a).

30%, 50%, 70% and 90%

DR

used

for each of the pile lengths assumed.

basic pile geometry and soil conditions are illustrated in Table 7.2.

In Table 7.2, a' v

in

The
and

cf h represent the in-situ vertical and horizontal effective stresses at the pile base level,
respectively.

The

finite

element meshes for each pile length are shown

bottom boundaries of the meshes were located

at

to or larger than

the pile lengths.

in this study are sufficiently large to eliminate

showed

The widths of

that the

7.7.

two times
the

Finite element analyses

separately with infinite elements at the lateral boundary

used

Figures 7.5

a depth larger than

corresponding pile length measured from the ground surface.

were equal

in

The
the

meshes

performed

mesh dimensions

geometric boundary effects.

175

Figure 7.8 shows a set of pile base load-settlement curves obtained from

element analyses done for the different

and

pile lengths

relative densities.

finite

The curves

in

Figure 7.8 extend up to a base settlement equal to 12 cm, corresponding to a relative


settlement s/B of
densities

show

20%.

stiffer

It is

observed

responses than those for lower relative densities under the same

The

confinement level or pile length.


density

is

The

also higher, and the rate of

was about

difference

The

2.2

MPa

7.2.

Pile

modulus degradation

Relative

(m)

density (%)

DR

is

lower.

= 30% and 90% appear more


For a 5-m

pile, the

settlement equal to 6 cm, which corresponds to s/B

pile

was about

geometry and

Pile Length

20

shear modulus of soil with higher relative

level or pile length increases.

at a

20-m

difference for a

Table

10

initial

differences of pile base unit load between

pronounced as the confinement

10%.

that the load-settlement curves for higher relative

soil

3.0

MPa

at the

conditions used in

same

relative settlement.

FEM analyses.

Type of Sand

a' v (kPa)

o' h (kPa)

30
50
70
90

Ticino

100

43

Ticino

100

43

Ticino

100

43

Ticino

100

43

30
50
70

Ticino

86

90

Ticino

200
200
200
200

30
50
70
90

Ticino

400
400
400
400

172

Ticino
Ticino

Ticino
Ticino

Ticino

86
86
86

172

172
172

176

20

Figure 7.5

Finite element

model

for

5-m

pile.

177

20

Figure 7.6

Finite element

model for 10-m

pile.

178

40

Figure 7.7

Finite

element model for 20-m

pile.

"

179

12

re

D R = 30%
D R = 50%
D R = 70%
D R = 90%

10

Q.

12

15

12

15

Base Settlement (cm)


(a)

12

15"

10

Q.

2
a
>

D R = 30%
D R = 50%
D R = 70%
D R = 90%

re

o
_i

4->

C
Z>

0)
(0
re

Base Settlement (cm)


(b)

Figure 7.8

Base load-settlement curves for

(a)

5-m, (b) 10-m, and

(c)

20-m

piles.

180

12

15

Base Settlement (cm)


(C)

Figure 7.8

Base load-settlement curves for


(continued).

(a)

5-m,

(b)

10-m, and

(c)

20-m

piles

181

Normalized base resistance for non-displacement

7.5.2

In order to relate pile

piles

base resistance to cone penetration resistance, the load-

settlement curves obtained previously were normalized as follows:

resistance q c for the

The cone

al.

same

and

(1997a), contained in the program

et al.

1997a, b; Salgado et

al.

1998a,

was divided by

the cone

the pile diameter B.

was determined using

resistance q c

stress state

soil conditions;

was divided by

the settlement

(2)

Salgado

DR

the base resistance q b at a given

(1)

the penetration resistance analysis of

CONPOINT

(Salgado 1993; Salgado

et

Figure 7.9 shows the fully developed load-

b).

settlement curves in terms of qb/q c and s/B for each of the pile lengths and relative
densities.

A common
the load

is less

design approach for non-displacement piles

is to

stipulate that so long as

than the load required for the pile settlement to reach a certain percentage

of the shaft diameter B, serviceability and ultimate limit states are not reached.

modest safety factor can also be used.


critical

load corresponds to s/B

load corresponding to s/B


0.1 criterion.

It

may be

requirements, that
values of qt/q c

some

0.05.

Franke (1993), for example, proposed that the

0.1, while

The

British

analyses.

5%

and
It

code for pile design

on

other values of s/B not be exceeded.

is

critical

based on the s/B =

structural or architectural

Table 7.3 shows different

several authors.

Table 7.4 shows the values of qb/q c


densities

Reese and O'Neill (1988) define a

stipulated in a given project, based

recommended by

at

s/B

= 5% and 10%

pile lengths obtained using the finite

can be seen that values of qt/q c

and the 0.12-0.21 range for s/B = 10%.

fall

for different relative

element and penetration resistance

within the 0.07 - 0.13 range for s/B =

182

0.5

SI

o:

0.4

0)

o
c
(0

to

0.3

re

-Q

0.2

u
N

re

0.1

0.0

0.00

0.10

0.05

0.15

0.20

0.25

0.20

0.25

Relative settlement, s/B

(a)

0.5

J3

cr
aT

0.4

o
c

re
55

3S

re

.o

0.3

0.2

v
N
re

E
o

0.1

0.0

0.00

0.10

0.05

0.15

Relative settlement, s/B

(b)

Figure 7.9

Normalized load-settlement curves for


in

terms of qb/q c and s/B.

(a)

5-m, (b) 10-m, and

(c)

20-m

piles

183

0.5

o
o
c

0.4

TO

C5
5

0.3

v
to

.Q

0.2

V
N

15

E
o

0.1

0.0

0.00

0.05

0.10

0.15

0.20

0.25

Relative settlement, s/B

(C)

Figure 7.9

Normalized load-settlement curves for


in

terms of qt/q c and s/B (continued).

(a)

5-m, (b) 10-m, and

(c)

20-m

piles

184

Table 7.3

Values of qt/q c according to several authors.


qb/q c

qb/q c

= 5%)

(s/B

(s/B

= 10%)

German Specification
(DIN 4014, Franke 1993)

N/A

0.2

Franke (1989)

N/A

0.2

Jamiolkowski and

0.2

Lancellotta (1988)

Ghionna

et al.

(for

(1994)

0.09

0.02

Table 7.4

Values of qb/q c

Length

DR

(m)

(%)

(kPa)

(kPa)

30

939

1517

50
70
90

1303

2238

2062
2789
3630

30

1343

50

1817

70
90

2409
3054

30

1933

50
70

2590
3357
4289

10

20

90

0.13 0.02

N/A

Salgado(1995)

Pile

N/A

B < 60 cm)

0.15

at s/B

q b (s/B = 5%) q b (s/B = 10%)

1726

2158
2915
3871
4970

3106
4158
5401

6845

5%

qc
(kPa)

7157
12052
19562
30121

and 10%.

qt/qc

(s/B

= 5%)

qi/qc

(s/B

= 10%)

0.13

0.21

0.10

0.17

0.09

0.14

0.07

0.12

10922

0.12

0.20

17544
26644
38816

0.10

0.16

0.09

0.14

0.08

0.13

16716

0.11

0.19

25694
36718
50524

0.10

0.16

0.09

0.15

0.08

0.13

'

185

Figure 7.10 illustrates the influence of pile length


level)

and

relative density

(i.e.,

on the normalized base resistance

qt/qc-

Figure 7.10 correspond to the relative settlement level of s/B


(a),

it

is

observed that the effect of pile length on qb/q c

is

confinement

The values of

= 10%.

Only a

slight decrease of q\Jq c

For dense sand,

loose sand as the pile length increases.

base

qb/q c in

From Figure 7.10

not significant.

the pile-base resistance q b and the cone penetration resistance q c

confining stress in a similar way.

at pile

This

is

because

depend on

initial

can be observed for

pile length has essentially

no

influence on q\Jq c for the range of lengths investigated.

As can be seen
base resistance

is

in

Figure 7.10(b), the influence of relative density on the normalized

The value of

relative density increases.

whereas

it

is

The normalized base

substantial.

0.12 to 0.13 for

qb/q c at s/B

Dr = 90%.

These

resistance qt/q c decreases as the

= 10%

is

0.19 to 0.2 for

DR

= 30%,

results indicate that larger settlements

are required for soils with higher relative densities to reach a base resistance equal to a set

percentage of cone penetration resistance q c

The

results also offer

some

insight into

q b by multiplying q c by a certain constant


to

Franke 1989) also place an upper

MPa

(e.g.,

limit,

range, on possible values of q b

why most
0.2 for

which

When

is

pile design

10%

methods

that calculate

relative settlement, according

usually taken as a value in the 4.5

piles are

embedded

layers, the results of the present analysis indicate that, as an

in very

dense sand

example, q b = 0.2 q c [the

value proposed by Franke (1989) irrespective of relative density] would be too high.

Following the results given


appropriate.

However,

if

in

Table

7.4, the

Placing a limit on q b (of say 5


the qt/q c values of Table 7.4

for setting an upper limit

on q b

value of qjqc

MPa)

= 0.12 would be more

serves a purpose in that case.

and Figure 7.9 are used, there may not be a need

186

0.35

n
U"

0.30

O
o
c

0.25

0.20

2
to

-*- Dr
-- Dr
-^- Dr
_*_ D R

"35

Qi
i-

=
=
=
=

30%
50%
70%
90%

o
</>

(0

0.15

Si
a

2
X

_-*

0.10

.N
15

0.05

0.00

50

Mean

100

150

stress at pile base

200

250

level, rj' m

300

(kPa)

(a)

0.35
.cr

n
CT

0.30

L =

6
o
c

0.25

ro

5m

L=10m
L = 20 m

'35

0.20

2>

a
v>
<o

0.15

XI

"a

0.10

"re

0.05

0.00

20

40

60

Relative density,

80

100

D R (%)

(b)

Figure 7.10 Normalized base resistance qb/q c with (a) mean effective stress (o' m )
pile base level

and

(b) relative density

(D R ).

at

the

187

The

7.5.3

effect of initial stress ratio

Ko

Normally consolidated sand deposits have


earth pressure at rest (Ko) in the 0.39

Ko

initial

0.5 range.

values of the coefficient of lateral

Overconsolidated sand deposits have

values typically higher than that of normally consolidated deposits.

investigate the effects of

and

1.0)

were assumed

Ko on normalized base
in a series of finite

resistance qb/q c three


,

may

Ko values

(0.4, 0.7

The value of Ko =

element analyses.

corresponding to an isotropic stress condition,

In order to

1.0,

be regarded, for practical purposes, as

the upper limit on Ko, observed for highly overconsolidated sand deposits.

Ko on

Figure 7.11 represents the effects of

the normalized base resistance qb/q c

The value of normalized base

resistance qb/q c in Figure 7.11 were determined for the

relative settlement equal to s/B

= 10%.

for different relative densities

10m and 20

5 m,
It

is

The curves shown

in

Figure 7.11 were plotted

(D R = 30%, 50%, 70% and 90%) and three

pile lengths

m).

observed that qb/q c tends to decrease as the

initial

Ko

increases, although

DR

and

deviation from this trend can be seen in the case for the highest relative density of

90%. This trend was most obvious

1.0

m,

the difference

was equal

and 20 m)
level,

in

which

between the normalized base resistance qt/q c

DR

= 90%

Figure 7.11,
is

relative density

These

at

it

for

For

for the lowest relative density.

to approximately 0.05.

was found

qt/q c

(L

No

Ko = 0.4

at

D R = 30%

Ko = 0.4 and

at

Ko =

difference in the normalized base resistance

to 1.0.

Comparing

the three cases (L

can also be observed that the confinement

determined by the pile length, does not have as

at

5,

10

the pile base

much

influence as the

Ko

value, the values

on the relationship between qb/q c and Ko.

results suggest that,

when

the soil

is

loose and has a high

of normalized base resistance qb/q c given in Table 7.4 and Figure 7.9 need to be modified
considering the variations in the values of qb/q c shown in Figure 7.11.
soil,

no such modification of the value of qt/q c

is

necessary.

For very dense

188

0.30

L=

0.25

5m

0.20
o

0.15
0.10
0.05
0.00

_i

0.0

0.2

0.4

i_

0.6

0.8

1.0

1.2

K
0.30

L=

0.25

10

0.20

XI

0.15
0.10

0.05

0.00
0.0

0.2

0.4

0.6

0.8

1.0

1.2

K
L = 20

0.25
0.20

""---..

0.15

0.10
0.05

00
0.0

0.2

0.4

0.6

0.8

1.0

1.2

K
D R = 30%

Figure 7.1

D R = 50%

Effect of

D R = 70%

Kq on normalized base

D R = 90%

resistance qb/q c

189

Determination of Base Resistance for Displacement Piles

7.6

Values and analyses presented previously were developed for non-displacement

The

piles.

settlement

limit base resistance q b L, as

and

levels

is

mentioned

conceptually

for

identical

displacement piles (De Beer 1984, 1988, Ghionna

The normalized base

earlier, is

et al.

both

mobilized

at

very large

and

non-displacement

1993, Salgado et

al.

1997a).

resistance values of Figure 7.9 for non-displacement piles, however,

cannot be directly applied to displacement piles, because of the very different loadsettlement response of displacement and non-displacement piles for low to moderate

settlement levels.

De Beer

(1984, 1988) has

by a displacement

shown

that,

under the same conditions, the loads carried

and a geometrically identical non-displacement pile

pile

differ

Non-

significantly for values of relative settlement s/B of interest in pile design.

displacement piles

settle

more than displacement

piles for the

same applied

load.

mainly due to the different installation processes for these two types of
installation of

around the
of the

the

is

The

piles.

displacement piles usually causes considerable densification of the

soil

In terms of base resistance, this process could be seen as a preloading

pile.

soil in the

immediate neighborhood of the

when compared with non-displacement


by

This

two types of

settlement increases.

pile for the

De

The

piles.

same settlement

pile base,

hence the suffer response

difference between the loads carried


level

becomes

less

pronounced as

Beer's observations were later confirmed by other authors

Jamiolkowski and Lancellotta 1988, Ghionna

(e.g.,

et al. 1993).

Table 7.5 shows the typical ratio of base resistance of displacement piles to base
resistance of non-displacement piles for different values of relative settlement but
initial soil

conditions.

As can be

seen, the ratio of q b of a non-displacement pile to q b of

a geometrically identical displacement pile is


settlements, but approaches
that

may be used

same

much

smaller than

as the settlement approaches infinity.

to determine qb/q c values as a function of s/B

at

small relative

simple approach

for displacement piles is

the application of the ratios of Table 7.5 to the results of Table 7.4 for non-displacement

190

Such an approach can provide useful design

piles.

more

criteria, but is naturally subject to

uncertainties than the values proposed for non-displacement piles in Table 7.4.

Table 7.6 provides the normalized base resistance qt/q c for displacement piles obtained
using this approach.

Table 7.5

Base resistance

ratio for

displacement and non-displacement


_

Relative Settlement

9,b>ND

2.5%

0.482

5%

0.517

0.15-0.21

10%

0.587

0.3-0.5

25%

0.715

0.3-0.7

>

->

oo

Table 7.6

Length (m)

Dr
(%)

10

20

,b.D

De Beer (1988)

Ghionnaetal. (1993)

-> 1.0

1.0

= base resistance for non-displacement


= base resistance for displacement pile

Pile

'

(s/B)

pile

qt>,ND

q b ,D

piles.

Values of qb/q c for displacement


qb/q c
(s/B

= 5%)

piles.

qt/q c
(s/B

= 10%)

30

0.25

0.35

50

0.19

0.29

70
90

0.17

0.24

0.14

0.20

30

0.23

0.34

50

0.19

0.27

70

0.17

0.24

90

0.15

0.22

30

0.21

0.32

50

0.19

0.27

70

0.17

0.26

90

0.15

0.22

191

7.7

Normalized Base Resistance for

Silty

Sands

Natural sand deposits often contain a certain amount of fines.

These sand deposits

are also bearing layers for piles, if the fines contents of soils are not high and soils are
sufficiently strong to support axial loads.

In order to investigate the

normalized base

resistance qt/q c for soils containing fines, the load-settlement response of piles in silty

sands was analyzed.

The mechanical properties of

They

(1999) and Bandini (1999).


stiffness

silty

sands were extensively studied by Salgado

investigated the behavior of silty sands focusing on

The

and strength characteristics.

contents equal to 5%, 10%, 15%, and


intrinsic parameters of silty

the tables are the

same

20%

stress increases.

in weight.

On

the other hand, as

(J)

silt

As can be

shown

content.

The

in

modulus and

Cg

Tables 7.7 and

in the finite

low

7.8, the higher the silt content, the

dilatancy parameters

in higher

and

best-fit regression of test results,

R=

element analysis of pile load

into granular soil deposit with

considered

The parameters shown

This also resulted

Bolton (1986) correlation was based on fixing

in

soil

seen in Table 7.7, the parameter

Table

in

at critical state.

Table 7.8 were obtained from the

parameters given

Tables 7.7 and 7.8 show the

content increases, while the exponent parameter n g for the confining

silt

angles with increasing

finite

silt

as used in (4.50) and (4.76) for the initial shear

higher the friction angle

The

was Ottawa sand with

material used

sands with different fines contents.

dilatancy angle of sand, respectively.

decreases as the

et al.

7.8.

Since

fines content,

element analyses.

friction

for silty sands in

while the original

1.

tests in silty

it is

peak

sands was done using the

usually desirable for piles to be placed

silt

contents equal to

The values of

5%

and 10% were

the parameters f and g for the

non-linear elastic-plastic stress-strain model were determined based on the stress-strain


relationship of silty sands

parameters

by Bandini (1999).

Table 7.9 shows the values of the

and g for different relative densities and

silt

contents.

192

Table 7.7

Values of
(after

Silt

Salgado

et al.

parameters with different

silt

contents

1999, Bandini 1999).

cg

eg

ng

^min

^max

612

2.17

0.439

0.48

0.78

454

2.17

0.459

0.42

0.70

10

357

2.17

0.592

0.36

0.65

content (%)

Table 7.8

Values of

friction angle

with different
Silt

soil intrinsic

silt

at critical state

and dilatancy parameters

contents (after Salgado et

al.

1999, Bandini 1999).

<1>C

29.0

9.0

0.49

30.5

9.0

-0.50

10

32.0

8.3

-0.69

content (%)

and

193

Table 7.9

Values of

and g used

5%

in finite

element analyses for

silty sands.

10%

silt

silt

Dr (%)
f

o
5

a
&

30

0.98

0.15

0.98

0.13

50

0.97

0.18

0.97

0.16

70

0.96

0.21

0.96

0.19

90

0.95

0.24

0.95

0.22

meshes and

Finite element

The cone

for clean sands previously presented.

obtained using the program

used

stress states

CONPOINT

intrinsic parameters given in Tables 7.7

in the analyses

resistance q c at the pile base level

(Salgado

and

were the same as those

et al.

1997a, b, 1998a, b) with soil

7.8.

Table 7.10 and Figure 7.12 show the normalized base resistance q\Jq c for
with different relative densities and pile lengths.

most values of qb/q c were

in the

0.12

- 0.17

For both

range.

For

in

On

at

the

silt

was small

10%

stress

silt

content.

pile length increases.

The value of

qt/q c of the

DR

20-m

to the increasing influence of the confining stress with increasing

As shown

in (4.50)

and

was even

(4.76), the

greater than that for

= 30%.

pile for

These
silt

DR

results

content.

parameter n g in Table 7.7 represents the effect of the

confining stress on the elastic modulus.

The higher

influence of the confining stress on the value of qt/q c

the value of n g , the greater the

As

qb/q c for the silty sand between different pile lengths were
for the clean sand.

was more

This observation

may be due

content

of

irrespective of the

the

silt

contents,

earlier, the influence

= 90% with

10%

sands

the other hand, the value of qt/q c for higher

(D R = 70%, 90%) increases as the

was more pronounced

silty

Table 7.10, qb/q c for lower relative density (D R = 30%)

decreases with increase in pile length.


relative density

and 10%

however, the influence of the confining

silty sands,

As can be seen

significant.

5%

As discussed

the confining stress on the value of qt>/q c for clean sands


relative density.

was

a result, the differences of

more pronounced than those

194

Table 7.10

Values of q\Jq c for

and
Silt

Pile

content

length

5m

DR
(%)

sands with different relative densities

silty

pile lengths.

q b (kPa)

q b (kPa)
(s/B

5%

(s/B

= 10%)

qb/q c

(kPa)

(s/B

= 5%)

qb/q c
(s/B

= 10%)

30

828

1336

8012

0.103

0.167

50

1103

1760

11911

0.093

0.148

70

1405

2239

16960

0.083

0.132
|

5%

10

20

10%

20

1749

2839

22488

0.078

0.126

30

1194

1909

12509

0.095

0.153

50

1551

2486

17391

0.089

0.143

70

1961

3144

23321

0.084

0.135

90

2425

3902

29800

0.081

0.131

30

1668

2736

19705

0.085

0.139

50

2191

3556

25596

0.086

0.139

70

2763

4460

32390

0.085

0.138

10

90

90

3411

5469

39808

0.086

0.137

30

755

1253

8789

0.086

0.143

50

1001

1641

12186

0.082

0.135

70

1283

2100

16231

0.079

0.129

90

1630

2662

20716

0.079

0.128

30

1129

1865

14186

0.080

0.131

50

1505

2440

18329

0.082

0.133

70

1903

3076

22976

0.083

0.134

90

2341

3815

27963

0.084

0.136

30

1682

2809

22953

0.073

0.122

50

2224

3674

27783

0.080

0.132

70

2813

4629

32863

0.086

0.141

90

3456

5596

38029

0.091

0.147

195

0.25

0.20

0.15

.o-D

o-

0.10

5-m

0.05

pile

10-m

pile

20-m

pile

0.00

20

40

60

80

100

Relative density (%)


(a)

5%

content

silt

0.25

0.20

0.15

0.10

5-m
0.05

pile

10-m

pile

20-m

pile

0.00

20

40

60

80

100

Relative density (%)


(b)

Figure 7.12

10%

silt

Values of

content

qt>/q c

for silty sand.

196

7.8

Summary

The cone

penetration test (CPT) resembles the vertical loading process on a pile.

advantage of the

this chapter, in order to take

in terms of

for pile design, load-settlement curves

normalized base resistance (q\Jq c ) versus relative settlement (s/B) were

Although the

developed.

CPT

In

limit state design

5%

with respect to either s/B =

or s/B

concept for pile design has been used mostly

= 10%,

the normalized load-settlement curves

obtained in this study allow determination of pile base resistance


settlement level within the

- 20%

range.

This

is

important, as

it

at

any relative

permits consideration

of specific project features, related to the superstructure or other components of the


facility.

In order to obtain the pile base load-settlement relationship, finite

were performed with a 3-D non-linear


diameter piles with lengths of 5 m, 10

were positioned within a granular

elastic-plastic constitutive

m and 20 m were
with

soil deposit

the soil conditions around the piles were

assumed

DR

to

pile installation, the results obtained represent those


piles.

The cone

Salgado

et al.

resistance q c

was calculated from

used

element analyses

model.

Three 60-cm

in the analyses.

30, 50,

The

70 and 90%.

piles

Because

be the same as those existing before


corresponding to non-displacement

the penetration resistance analysis of

(1997a) and used to normalize the load-settlement curves in order to

express them in terms of qt/q c and s/B.

Most
analyses

10%.

qb/q c values obtained

fall

The

within the 0.07

effect of relative density

densities, the value of qb/q c


relative densities (qt/q c

pressure

decrease as the value of


densities,

= 0.19

Ko

at the pile

0.2 for

Ko was

increases.

element and penetration resistance

5%

and the 0.10

0.20 range for s/B =

on the normalized base resistance was significant,

was smaller

at rest

finite

0.13 range for s/B =

while that of the confining stress

lateral earth

from the

base level was small.

(qb/q c

DR

= 0.12 -0.13

= 30%).

also investigated.

for

The

DR

At higher

= 90%) than

at

lower

effect of the coefficient of

The value of

This trend was more pronounced

and negligible for very dense sand.

relative

at

qb/q c tends to

lower relative

197

Based on the
piles

results

by De Beer (1984, 1988), the values of qt/q c for displacement

for s/B

= 5% and

in the 0.22

The normalized base

5%

The values of qt/q c were

were obtained as well.

and 10%

silt

contents,

0.35 range for s/B

typically in the 0.15

0.25 range

= 10%.

resistance qb/q c for silty sands

most values of qt/q c were

was

in the

also investigated.

0.12

- 0.17

range.

For both

These

values are typically smaller than those for clean sands, which ranged from 0.12 to 0.21.

The confining

stress

was another important

factor for the value of qb/q c of silty sand.

For

lower relative density (D R = 30%), the value of qb/q c decreases as the pile length increases
while that for higher relative density increases.

198

CHAPTER 8 ASSESSMENT OF PROPOSED NORMALIZED BASE


RESISTANCE VALUES BASED ON CASE HISTORIES

8.1

Introduction

In

this

chapter,

presented previously.

we reexamine

We

pile load tests, calibration

do so

values of normalized base resistance qt/q c

the

in the context of a

chamber

few case

histories involving actual

and other numerical analyses.

data,

Both non-

displacement and displacement piles are addressed.

tests

The case

histories for non-displacement piles include

on drilled

shafts, a series of calibration

numerical results of Simonini (1996).

on the Georgia Tech campus

Mayne 1994) and

For displacement

et al.

section

NGI

al.

(Gregersen

load

test

tests

(Ghionna

et al. 1994),

on a drilled shaft performed

and the
at

a site

Georgia (Mayne and Harris 1993, Harris and

at the

University of Sao Paulo (USP), Sao Carlos,

load test on a steel H-pile performed

1972) and the load

et al.

tests

1973) were used.

at the

qt/q c for each case are

summarized

Purdue

on the precast concrete

piles

All the load tests used in this

were instrumented, so base load-settlement responses were recorded

The observed values of

pile load

1995) are analyzed.

piles, the

University campus (Goble et


carried out by

in Atlanta,

another performed

experimental field (Albiero

chamber

two instrumented

in

Table

8.1.

separately.

199

Table 8.1
Pile

Values of qt/q c from load

Type

Test

tests

on non-displacement and displacement

Name

qt/q c
(s/B

= 5%)

qt/qc

(s/B

= 10%)

Non-Displacement

Georgia Tech Test

0.18

0.26

Pile (Drilled shaft)

USP Test

0.09

0.20

Simonini's Analysis

N/A

0.17

Purdue Test

0.27

0.37

0.32

0.36

NGITest B

0.32

0.36

NGITest C

0.30

0.39

0.43

0.47

Displacement
Pile (Driven pile)

NGITest

NGITest

piles.

200

Non-Displacement

8.2

Piles

Georgia Tech load

8.2.1

test

The Georgia Tech

was described

load test

the

in

previous

The

section.

representative value of q c for base resistance calculation, an average value between the
level of the pile base

base,
are

is

6.5

and a

one-and-a half to two pile diameters below the pile

level about

MPa. The measured base

580 and 810 kN, corresponding

to qt/q c values of 0.18

the relative density of the soil around the pile base

between q c and the

The estimated

DR

relative density

relative density of the soil

about 20 - 30%, representing a loose


qb/q c at s/B

10%

loads for the shaft at 5 and

and 0.26 respectively.

was not determined,

suggested by Salgado et

al.

Since

the relationship

(1997b) was used.

around the pile base with a Ko value of 0.41 was

state.

and 10% can be found

relative settlement

Using Table

as 0.12

7.4, the

- 0.13 and 0.20 -

corresponding values of
0.21, respectively.

The

difference between measured and predicted values of q b in this example appears to be

small (20

8.2.2

- 30%

underprediction).

Sao Paulo load

test

The USP, Sao Paulo,

test site is

located near

downtown Sao

geology consists of two clayey

Sao Paulo, Brazil (Teixeria and Albiero 1994).

The

sand layers separated by a thin layer of pebbles.

The upper

site

layer

Cenozoic age, collapsible upon inundation (Vilar 1979), with void


lower layer

is

at

is

relative

typically

a depth of 10

loads corresponding to s/B

10%

is

a reddish material of

ratios close to

1.

a brownish residual soil from a sandstone of the Bauru formation.

water table oscillates, but

was placed

Carlos, in the state of

settlement

m
5

around 10 meters deep.

The base of

and the shaft diameter was 50 cm.


and

10%

are 50

The

The

the test shaft

The measured base

and 114 kN, respectively.

was obtained from a quick maintained load

The load
test

for

(QMLT)

201

slow maintained load

performed

after a

value of q c

was obtained from

measured qb/q c values for

this case

namely qb/q c = 0.09 - 0.10

medium dense

to

sounding provided

5%

The

relative settlement.

in Albiero et al.

(1995) as 2.8

were also favorably compared with the results

= 5% and

for s/B

qb/q c

0.14 - 0.17 for s/B

in

The
Table

= 10%

for

sand.

Simonini's results

8.2.3

Simonini (1996) carried out


non-displacement piles

soil unit

to s/B

weight Ym = 16.5

= 10% was found

The

resistance

al.

1997a,

b),

were the same

was modeled as having a diameter equal

were a dry sand deposit with

The value of

cone resistance q c

was

fy c

al.

at the pile

D R = 90%

base level, the penetration

(1997a, b), available in the program

used.

The

7.4, equal to

is

CONPOINT

properties used to obtain the cone

soil

base resistance value with q c the value of qb/q c

Table

33,

the base resistance q b corresponding

as those used in the finite element analyses.

slightly higher than that of

to

MPa.

resistance analysis of Salgado et


et

pile

soils

kN/m J

to be 8.6

In order to obtain the

(Salgado

element analyses to obtain the base resistance of

finite

The

in sands.

m and a length equal to 30 m.


and

(SMLT) up

This yields qb/q c values of 0.09 and 0.20 for s/B = 5 and 10%, respectively.

MPa.

7.4,

CPT

the

test

obtained as 0.17.

Combining

the

This qb/q c value

is

approximately 0.13 for a relative density of

D R = 90%.

Calibration

8.2.4

series of

chamber

30

plate load tests

calibration

chamber load

tests

were carried out by Ghionna

(1994) in sand samples with two different relative densities, dense (D R

et al.

= 90%) and

202

medium dense (D R = 50%).


shafts.

Calibration

chamber

Figure 8.1

small.

These

size effects

0.

relative densities are in reasonable

= 5%, and

D R 90% at s/B

scatter than those for

DR

earlier to

chamber

calibration

be

tests

resistance q c used to prepare Figure 8.1

D R = 90% at s/B

for

10 - 0.14 range for

50% show more

The cone

for the

al.

in Figure 8.1, the values of qb/q c are in the 0.09

= 50% and 0.07 - 0.10 range


and

on the values of q b were shown

the penetration resistance analysis of Salgado et

As can be seen

D R 50%

simulate the loading of the base of drilled

shows the values of qb/qc

corresponding to s/B = 5 and 10%.

was obtained from

tests

= 90%,

(1997a,

b).

- 0.14 range

in the 0.1

= 10%. Although

for

DR

- 0.19 range for

the results for

DR

the average values of qb/q c for both

agreement with the proposed

results given in

Table

7.4.

8.3

8.3.1

Displacement Piles

Purdue University load

test

The Purdue University load


University

campus

in

West

was performed on

test

Lafayette, Indiana (Goble et

Wabash

the edge of a large terrace along the


loess covering stratified sand

was driven using a

and gravel

DELMAG

D-12

sometimes referred to as small displacement


section does not cause as

much

1972).

al.

This

site is

Purdue

located on

river with a variable depth of weathered

layers.

diesel

the western side of the

15-m long

hammer and

steel H-pile

load tested.

(10HBP57)
H-piles are

piles because their relatively small cross-

disturbance and densification of the surrounding soil

during the installation process as concrete or pipe piles would.

The base

resistance q b

and

relative settlement s/B

were calculated based on the

equivalent circular base area transformed from the half perimeter area of the H-pile.

obtained q b values for s/B

5 and

Because the standard penetration

test

10% were

2.43

MPa

The

and 3.37 MPa, respectively.

was used instead of cone penetration

test,

the

cone

203

U.JU

0.25

0.20
a-

0.15

;.

0.10

0.05

n nn

20

40

60

80

100

Dr(%)
(a)

100

(b)

Figure 8.1

Values of qt/q c
(b) s/B

= 10%.

in calibration

chamber

plate tests for (a) s/B

= 5% and

204

was estimated from

resistance q c

SPT-CPT

SPT blow

was about

MPa.

9.0

0.27 and 0.37 respectively.

The values of
These

the values of qt/q c for displacement piles

8.3.2

NGI

in the

shown

homogeneous sandy

middle of the

soil layer consists

in

Table

Drammen

The

soil.

From

D in Table 8.1.

of 28 cm, while pile


top to 20

cm

Piles

C was

at the pile base.

with a diameter of 28

diameter of 20

MPa

Although

cm

cm

and

Piles

down

Drammen, Norway.

to

The four

al.

Four instrumented

1973).

piles are referred to

had the same length (8 m);

B and

pile

by the

from the top

cm

had the same length (16 m);

to half the pile length,

The cone

letters

at the pile

pile

B had

5%

a pile

and 16

were

for each pile are given in Table 8.1.


,

within the 0.29 - 0.32 range for s/B =

A,

and then tapered with a

resistances at depths 8

The values of qb/q c

piles

A had a diameter

from 28

shows an exceptionally high value of qb/q c most of

Considering that the

The

30 m, underlain by a clay

D was made by connecting piles A and C, resulting in

at the pile base.

and 5.0 MPa.


pile

located in a small island,

site is

a tapered pile with diameter varying

diameter of 28 cm, while pile

3.1

5 and

the geological history of this area, the subsoil condition

with two lengths, 8 and 16 m, were tested.

and

7.6.

river near the city of

believed to be normally consolidated (Gregersen et

qb/q c corresponding to s/B

at

appear to be near the upper range of

of uniform, loose upper sand layer

layer and finally bedrock.

B,

The estimated q c value

were performed on precast-concrete piles driven into a very

tests

loose deposit of quite

is

results

base level based on the

at the pile

load tests

The NGI load

Holmen,

count

Campanella (1983).

correlation of Robertson and

the pile base level

10% were

the

the qb/q c values fall

and the 0.36 - 0.39 range for s/B = 10%.

soil is in a loose state, these

values proposed for displacement piles in Table 7.6.

measured values agree well with the

205

8.4

Summary

In order to verify the normalized base resistance value qb/q c > several case histories
for both non-displacement

and displacement

piles, the

observed results of field pile load

chamber

plate load tests

piles

tests, finite

element analysis, and calibration


in

Chapter

7.

good agreement with the proposed values

7.

For displacement

piles, 5 pile

qb/q c for displacement piles


piles,

For non-displacement

were compared with the qt/q c values presented

Overall, the observed values of qt/q c were in

given in Chapter

were examined.

the comparison

load tests were investigated.

were established based on the

Although the values of

results of

non-displacement

between the measured and the proposed qt/q c values showed

reasonably good matches.

206

CHAPTER 9 PILE DESIGN USING CPT RESULTS

9.1

Introduction

In this chapter, the design of piles using


shaft resistances are addressed.

be used

proper

Since the standard penetration

SPT-CPT

when only SPT

A
CPT

correlation

makes

Both base and

results is discussed.

between the SPT blow count

practice, the correlation

discussed.

CPT

it

test is still

widely used

and the cone resistance q c

possible for

is

in

also

CPT-based methods

to

results are available.

computer program developed for the estimation of

results is introduced.

In this

pile load capacity

based on

program, base and shaft resistances can be estimated

using different methods.

9.2

9.2.1

Determination of Base and Shaft Resistances

Base resistance

As

CPT

discussed in Chapter

results (DeRuiter

2, there are several

methods available for

pile design using

and Beringen 1979, Schmertmann 1978, Aoki and Velloso 1975,

Bustamante and Gianeselli 1982).

All of these

methods define the base resistance

terms of the cone penetration resistance q c and correlation parameters.

methods were developed under

different conditions, the selection of the

in

Since these

method should be

207

made with

consideration of the differences and recommendations of the methods.

The

important differences between the methods include (Bandini and Salgado 1998):

(1)

the criterion adopted to define pile load capacity;

(2)

the type of equipment used to obtain q c

(3)

the selection and relative importance of q c values above and

(4)

soil

below the

pile base;

types and conditions under which the methods were developed.

Regarding the criterion adopted to define the

pile load capacity,

design methods do not clearly define the pile load capacity.

CPT-based

Although for small-diameter

piles the differences in pile capacity are not large, as the diameter increases the

which load capacity

is

defined becomes

critical.

pile

way

in

relative settlement-based criterion for

pile load capacity definition is favored, as discussed previously.

Following Franke (1991), piles must typically undergo relative settlements s/B
greater than
structures or

10%

to reach a limit state,

damage

including either a loss of functionality of

to the superstructure and/or foundations.

of the pile diameter, in some design situation,

may

Settlements less than

10%

cause the foundations or the supported

Hence, in the evaluation of the value of the relative

structures to reach a limit state.

settlement associated with the limit states design of piles, the type, functionality, location,

and importance of structures should also be taken into account.


Figure 9.1 shows a simple example for estimating pile base resistance using

CPT

m and a diameter of 50 cm.

The

results.
soil is a

The

pile is a drilled shaft

having a length of 10

normally consolidated, medium dense sand.

varying linearly with depth from

For these given

pile

different methods.

and

MPa

The cone

at the surface to

soil conditions, the

10

resistance q c

MPa

was assumed

at the pile

base level.

base resistances q b were calculated using

208

10.9

8.2

Drilled Shafts

6.4

10

11.8 q c

(MPa)

B = 50 cm
Depth (m)

qb =

8.2

+ 10.9

Dutch method:

Aoki and Velloso's method:

q.

9.55

MPa

= -^ =
3.25

LCPC

Proposed method:

Figure

9.

method:

= 3.07

q b = k c x q ca = 0.4 x 10 = 4.0
qb =0.16x10.9 = 1.74

MPa

3.25

MPa

MPa

Estimation of pile base resistance using different methods.

209

For the method proposed


to the
is

medium dense

in this study, the value of qt/q c

was equal

Although the profile of cone resistance

condition.

corresponding

to 0.16

in this

example

extremely simplified, the values of q b obtained using different methods show a quite

The highest value of q b equal

wide range.

method while the lowest one equal


in this study.

It

to 1.74

to 9.55

MPa

MPa

was obtained from

the

Dutch

was obtained from the proposed method

should be noticed that the lowest base resistance obtained from the

proposed method was based on the relative settlement s/B equal to 10%.

9.2.2

Shaft resistance

The

maximum

shaft resistance of piles in

base resistance

is

most cases

is

reached (Franke 1993).

be estimated with greater simplicity,

at least

fully mobilized well before the


It

follows that shaft resistance can

from a conceptual point of view, than base

resistance.

Based on the review of the methods


and Velloso 1975) and the

LCPC method

in

Chapter

2,

Aoki and Velloso's method (Aoki

(Bustamante and Gianeselli 1982) appear to be

effective approaches to estimate the shaft resistance using

a simple

example of estimation of

two clayey

layers

layers.

Pile condition

example of base resistance estimation discussed


representative cone resistance q C for each sub-layer
layer.

The

Figure 9.2 shows

shaft resistance using these methods.

and three sandy

CPT results.

in

the

is

the

Soils consist of

same

previous

was obtained from

as that for the

section.

The

the center of the

calculated values of shaft resistance are given in Figure 9.2.

In Figure 9.2, the shaft resistances of layer 2, 4,

and 5 for the

LCPC method

were

taken as the limit values given in Table 2.5 because the calculated shaft resistances were
greater than the limit values.

The estimated value of

the shaft resistance

Velloso's method was significantly lower than that from the

LCPC

from Aoki and

method.

210

Drilled Shafts

1.9 2.5

5.5

6.0

9.1

q c (MPa)

B = 50 cm

LCPC method

Aoki-Velloso method

qs

(kPa)

qc
(kPa)

ks

6.5

4.1

1900

60

31.6

6.0

6.5

23.1

2500

40

35.0

5500

1.4

6.5

11.8

5500

100

55.0

6000

6.0

6.5

55.4

6000

60

35.0

9100

1.4

6.5

19.6

9100

100

80.0

Layer
(kPa)

(%)

1900

1.4

2500

Figure 9.2

F2

114.0

qsi

(kPa)

236.6

Estimation of pile shaft resistance using different methods.

211

This indicates that Aoki and Velloso's method produces very conservative results in the
estimation of the shaft resistance compared to the

was studied by

several authors (Briaud et

found

to

9.2.3

Factor of safety

al.

LCPC

method.

The LCPC method

1989, Milovic and Milovic 1993), and

was

be satisfactory.

The

selection of factors of safety in pile design

factors of safety

may be

is

an important.

In general, large

required in the following cases (Canadian Geotechnical society

1992):

(1)

friction piles in clay;

(2)

sites

where only a limited number of

tests

where

are performed and

soil

conditions are variable;


(3)

piles in loose sands

When

and

the pile load capacity

silts

is

for

which the capacity may decrease with time.

determined from

pile load tests, factors of safety in the 2 to 3 range

Canadian Geotechnical society 1992).

field tests including in-situ tests

have been proposed (Broms

et al.

and

1988,

Table 9.1 shows the values of factor of safety

proposed by Canadian Geotechnical society (1992) as a function of the type of


based on the ultimate pile load capacity corresponding

field test,

to an ultimate limit state.

In the

table, fp is referred to as the resistance modification factor representing the ratio of

allowable pile load capacity to ultimate pile load capacity.

Some

authors have proposed the use of the partial factors of safety.

earlier, the shaft resistance, in

reaches a limit

state.

most cases,

Based on

is

As mentioned

fully mobilized before the base resistance

this observation, the use of separate factors

the base and the shaft resistance

is

sometimes suggested to

different rates of mobilization of shaft

and base

resistance.

artificially

of safety for

account for the

212

Table 9.1

Resistance modification factor fp and factor of safety for different field tests
(after

Canadian Geotechnical Society 1992).

Type of filed

Cone

test

penetration test

Standard penetration

test

Static pile load test

fP

Factor of safety

0.5

0.3

3.3

0.5

0.6

1.7

0.5

(routine test)
Static pile load test

(high technical level test)

Dynamic

analysis using

measured data of
and acceleration

strain

213

According to Franke (1993), the allowable design

Q^=Q^ +

Qd=

where

= unfactored

FS b

FS g

Q d = allowable design pile


Q

pile load capacity is given by:

Q>

FS

(9.1)
S

load capacity

pile load capacity

FS g =

global factor of safety

FS b =

factor of safety for the base resistance

FS =

factor of safety for the shaft resistance

Qb = base

resistance

from a

limit state

shaft resistance.

Since

Q =Q +Q
u

(9.2)

the following relationship can be obtained:

FS h -FS-(l + ^-)
FS.

(9.3)

FS+FS

If

FSg = 2 and FS S =

Table 9.2

it

are assumed,

is

Qb

obtained as in Table 9.2.

From

(9.3)

and

can be seen that for a constant global safety factor no unique value of factor

of safety for the base resistance


is

FS b

-Q-

allowed to vary.

is

obtained

when

partial safety factors are

used and Qt/Q s

214

Table 9.2

Partial factor of safety for the base resistance.

Qs/Qb

FS b

0.4

0.8

0.6

1.0

2.5

3.33

10

CO

Since the goal in design


safety factors

partial

0.2

is

is

to achieve a target level of safety (say

yet variable, as fixed values of

>

<0

FS g =

2), the

use of

FS b and FS would produce


S

depending on the relative contribution of base and shaft

different levels of global safety

resistance to pile capacity.

Use of SPT Blow Counts

9.3

The method

for estimating pile load capacity proposed in this study

cone penetration resistance q c


is still

widely used

the

CPT

pile design

approach

in

many

SPT blow

between the

may be

useful

CPT-based Method

in

Although the

CPT

is

CPT-SPT

count

more

considerable scatter
that the

mean

q c/N

ratio

correlation

was found

data.

This

was proposed by Robertson and Campanella (1983).

in the values of the

be used for sandy

ratio

soils.

of the cone resistance q c to the

mean

q c /N ratio for

grain size
silty clay,

(D 5 o).

Since

they suggested

Figure 9.3 shows the q c/N ratio with respect to

grain size.

As can be seen
increases.
silty

proper correlation

uncertainties, but is believed to be useful in design.

not a constant, varying with the

is

SPT

and the cone penetration resistance q c can be established,

According to Robertson and Campanella (1983), the

SPT blow

If the

methods can also make use of the SPT blow count

subjected to

based on the

a superior test for pile design,

geotechnical engineering projects.

count

is

in the figure, the value of

For sands, the values of the q c/N

sand were

in the 3

-4

range.

ratio

q c /N increases as the mean grain size

were

in the

4 - 6 range while those for

215

bars

N,blowj/(ool

(Ibor'IOOkPo)

CL*EY

SILTS

8 SiLTT CLAY

SANDY
ft

SILT

SAND

SILTY SAND

SILT

10

4
4

9
.

/I
7

<I
cr

y
*4

12

4
12

i *4.

%^h* 45

12

L^^J

u"

f "
001

01

MEAN GRAIN

Figure 9.3

CPT-SPT
(after

0.1

SIZE

D 5Q

correlation with the

mm

mean

grain size

Robertson and Campanella 1983).

<6

216

For further investigation of the value of the q c/N


performed

at the

same

Evansville, Indiana.

The

site.

The

tests

site

was located

mostly consists of sandy

penetrate in the

few

first

starting test depth

Two
taken

at

feet,

Both

fill

down

at the

material

it

CPT

a depth of from 14.0

maximum and minimum


(D 50 ) was about 0.3

and

SPT were

66 over Garvin

Street,

SPT

The ground was very hard

20 m.

difficult to

same elevation

boring logs indicate that the

CPT

anchor down the


as

all

to

The

rig.

the surrounding terrain.

No

was observed.

m down

soil

and

to about

sand samples were obtained using the

were performed with these

size

soil

making

appeared to be

appreciable presence of

State route

CPT

were performed under the North East end of the bridge, on

the grassy area next to the bridge wall.


test site

US

at

both

ratio,

to 14.7

CPT

Some

m.

sampler.

soil

Both samples were

of the standard laboratory tests

samples, including grain size distribution analysis, the

void ratio

mm, and

tests,

the

and the specific gravity

maximum and minimum

The mean

test.

grain

void ratios were 0.826 and

0.454, respectively.

Figure 9.4 shows the

SPT blow

The corresponding values of q c/N

counts

and the cone resistance q c with depth.

for this site are given in Table 9.3.

values in Figure 9.3 were based on the energy ratio of


corrected for the energy ratio of

Table

9.3.

As can be

60%, were used

however, were observed

in the 3

was about

a depth of 13.47

8 range.

the lowest and highest values of q c /N


in reasonable

at

can be used for

This result

CPT pile

is

was approximately

one more indication

design using the

N 60

which was

1.15 at a depth of 4.33

Most

m.

The averaged q c/N

agreement with the value for sand given

Campanella (1983).

the values of

SPT N

for the calculation of the q c /N ratio in

seen, the lowest value of q c /N

and the highest one was about 14.96

55%,

Since the

SPT blow

5.63.
in

the q c /N ratios,

ratio without including

This value appears to be

Figure 9.3 by Robertson and

that the correlation of Figure 9.3

count

N for Indiana soils.

217

SPT blow

q c (MPa)
5

10

15

20

25

30

35

10

10

20

30

count, N
40

50

10

Q.

Q.

Q
15

15

20

20

25

25

Figure 9.4

Cone

resistance q c and

SPT blow

count

N with depth.

60

218

Table 9.3

Correlation between

CPT

and SPT.

Depth (m)

N 60

0.21

36

33.12

236.9

7.12

0.98

17

15.64

112.0

7.16

1.12

40

36.80

120.0

3.26

2.80

11

10.12

26.7

2.63

4.33

15

13.80

16.0

1.15

5.85

10

9.20

49.8

5.40

7.38

18

16.56

85.0

5.10

8.90

37

34.04

138.5

4.10

10.42

40

36.80

55.6

1.51

11.95

13

11.96

88.7

7.4

13.47

7.36

110.0

14.96

14.99

33

30.36

146.6

4.8

16.52

15

13.80

120.0

8.7

18.04

27

24.84

156.7

6.3

N 6o

= SPT

N value corrected for the energy ratio of 60%

bar = 100 kPa

q c (bar)

q c/N 60

219

9.4

Program

CONEPILE

For the effective use of the CPT-based


practice,

method proposed

in this study in

was programmed together with other methods discussed herein

it

"CONEPILE".
program.

pile design

Calculation of both base and shaft resistances

The methods

available in the

program

CONEPILE

is

in

program

addressed in the

are:

for the calculation of the base resistance,


(1)

Aoki and Velloso's method;

(2)

LCPC

(3)

proposed method;

method;

for calculation of the shaft resistance,

Program

(1)

Aoki and Velloso's method;

(2)

LCPC

CONEPILE

method.

can be used to calculate base and shaft resistances for both non-

displacement and displacement

piles.

Users can select a preferred method to use for a

given soil condition and pile type.

The program CONEPILE


interface

program

calculation.

The

for pre-

consists

of two different programs:

and post-processing, and the

user-friendly interface

FORTRAN

a user-friendly

code used

in actual

program was developed using Visual-C for easy

input and output data processing.

For Aoki and Velloso's method and the

LCPC

method, the correlation parameters

given in Tables 2.1, 2.2, 2.5, and 2.6 were included in the program
of table.

as a

form

For the calculation of base resistance using the proposed method, regression

equations matching with the normalized load-settlement curves


used.

CONEPILE

The

regressions were
2

coefficient of correlation (r

made with

shown

in

Figure 7.9 were

fourth order polynomials, and the square of the

were about 0.999.

Several intermediate steps were also

220

used

in the

These include the estimation of the

program.

for the coefficient

Ko of

lateral earth

resistance ratio for displacement piles.


stress at the pile base level

pressure

at rest,

relative density, the correction

and the calculation of the base

Since the effect of the pile length or the confining

appeared to be small (as can be seen

in

Figure 7.10),

it

was not

considered for calculating the base resistance.

The estimation of
Salgado

et al.

was made using

relative density

the following relationship by

(unpublished paper):

^=C (^)

~ C:

<Vexp(C D
3

C3 =

/?

(9.4)

where q c = cone

resistance; Ci, C2,

and

= 100 kPa =

MPa

effective horizontal stress;

The

0.1

tsf;

oh =

correlation parameters C\, C2,

relative density

C3

state

correction for

program CONEPILE).

the effect of

Ko was based on

Ko was small

70%.

the relationship

CONEPILE

are given in

relative density.

Using

(9.4), the

relative density

shown

in

70%,

Figure 7.1
it

1.

Because

was only considered

Figure 9.5 shows the procedure for the estimation of

the base resistance using the proposed method.

program

The estimated

for relative densities greater than

for relative densities less than

DR

reference stress

and cone resistance q c (which need

then used for calculating the normalized base resistance q\Jq c

The

and

Pa =

are given in Table 9.4.

can be obtained for a given stress

to be assigned as input data for the


is

and

correlation parameters;

Appendix.

More

detailed information for the

221

Table 9.4

Correlation parameters for estimation of relative density


Friction angle at the critical state (0 C )

30

31

32

33

34

35

36

c,

40.0

44.7

49.7

55.3

61.6

68.6

76.4

c2

0.524

0.519

0.514

0.508

0.501

0.496

0.490

c3

0.0195

0.0196

0.0197

0.0197

0.0198

0.0199

0.02

222

Input parameters
a' v Kq,
,

qc

(|)c,

s/B

Estimation of
Relative Density

Calculation of

(D R )

qt>/q c

with

Estimated Relative Density

Modification for
a given

Ko value

Determination of

Base Resistance q b

Figure 9.5

Estimation of the base resistance for a given

soil condition.

223

9.5

Summary

In this chapter, the design of piles using

methods available

for pile design

CPT results was

discussed.

Since the

CPT

were developed under different conditions, the selection

of the method should take into account the differences and recommendations of the

methods.

From

the example for calculating the base resistance,

from the proposed method was based on

was observed

method while

highest value of q b was obtained from the Dutch

obtained from the method proposed in this study.

it

that the

the lowest one

The lowest base

was

resistance obtained

a relative settlement s/B equal to

10%.

For the

shaft resistance,

Aoki and Velloso's method produced lower values compared with the

LCPC

method.

was found

to

global factor of safety (and not partial base and shaft factors of safety)

be more correct for use

in pile design.

For more general use of CPT-based

SPT blow

count

N and the

pile design

methods, a correlation between the

cone resistance q c can be used.

for an Evansville sandy soil site suggests the correlation

Robertson and Campanella (1983)

is

SPT and CPT

between q c and

likely applicable to Indiana

The program CONEPTI JB developed


introduced.

Field

test

data

proposed by

sandy deposits.

for the estimation of the pile load capacity

was

This program has a user-friendly interface, which controls an underlying

FORTRAN-based DLL.

224

CHAPTER

10.1

10

SUMMARY, CONCLUSIONS AND RECOMMENDATIONS

Summary

Pile foundations are often used for supporting large,

heavy superstructures

where competitively priced shallow foundations would lead

Based on

the

method of

installation, piles are classified as either

consists of both base resistance and side resistance.


fully

side resistance

resistance

is

mobilized well before the

is

maximum

The

side resistance of piles

base resistance

methods using

in-situ test results

(CPT).

test

is

into the ground.

the

study, in order to take advantage of the

CPT

pile load test.

counts

Such a
cone

Static

better related to the pile loading process.

performed quasi-statically and resembles a scaled-down

The

test is

In the present

for pile design, load-settlement curves of

loaded piles bearing in sand were developed in terms of normalized base

resistance (qt/qc) versus relative settlement (s/B).

the load-settlement curves

Salgado

SPT blow

not well related to the quasi-static pile loading process.

penetration, on the other hand,

axially

in

have been mainly based on the

from dynamically driving a standard sampler one foot


is

is

As

reached.

is

piles

mobilized early in the loading process, the determination of pile base

standard penetration test (SPT) and the cone penetration

process

displacement or non-

a key element of pile design.

Pile design

result

to excessive settlements.

The bearing capacity of both displacement and non-displacement

displacement.

most cases

at sites

et al.

was determined from

(1997a) using the program

Cone

resistance q c used to normalize

the penetration resistance analysis of

CONPOINT. The

proposed as a modern design approach,

in

limit state concept has

been

which the adequate technical quality of

225

foundations and superstructures

is

Although the limit

considered.

for pile design has been used mostly with respect to either s/B

normalized load-settlement curves obtained

in this study

= 5%

design concept

or s/B

= 10%,

the

allow determination of pile base

- 20%

resistance at any relative settlement level within the

state

resistance for both non-displacement and displacement piles

range.

The normalized base

were addressed.

order to obtain the pile base load-settlement relationship, a 3-D non-linear

In

elastic-plastic constitutive

model was used

linear elastic-plastic constitutive

in finite

model takes advantage of the

variables that can be uniquely determined for a given soil type


linear stress-strain

model represents changes of

and the bulk modulus K) according to the

For the description of

reached.

dimensional stress

state, the

failure

elastic

stress

The 3-D non-

element analyses.
intrinsic

and

and condition.

state soil

This non-

parameters (the shear modulus

level before a failure condition is

and post-failure

soil

response for the three-

Drucker-Prager plastic model with non-associate flow rule

was used.
Calibration

chamber

have been used to investigate both the load-settlement

tests

response of the base of non-displacement piles and cone penetration resistance under a
variety of conditions.
size

and boundary

Because calibration chambers have

effects arises.

If size effects are

finite sizes, the possibility

well understood, calibration chambers

can be used to experimentally assess pile base resistance under controlled conditions.
series of calibration

chamber

tests

were modeled and analyzed using the

approach with the 3-D non-linear elastic-plastic stress-strain model.


results

plate load tests.

The

predicted load-settlement curves

with measured load-settlement curves.

Calibration

chamber

finite

The

were compared with the measured values of plate resistance

chamber

of

in

element

analytical

calibration

showed good agreement


size effects

were also

investigated for different relative densities and boundary conditions using the finite

element analysis.

For effective use of the


in this study

CPT

pile design

methods

and some of existing methods reviewed

user-friendly interface procedure.

in practice, the

in study

method proposed

were programmed with the

This program can be used in practice to estimate pile

226

load capacity for a variety of pile and soil conditions without significant difficulties for
input and output procedure.

10.2

(1)

Conclusions

Based on findings of the present

study, the following conclusions are drawn.

The values of

and g for the 3-D non-linear

strain

the parameters f

model vary according

elastic-plastic stress-

As

to the relative density level.

the relative density

increases, the value of f decreases while the value of g increases.


that the ratio of the elastic

modulus

at failure to its initial

value

is

This indicates

higher for denser

than for looser sand, and the rate of degradation of elastic modulus

higher for

is

looser than for denser sand;

(2)

Plate unit loads in calibration

higher (for BC2,

BC3) than

chamber

tests tend to

pile base unit loads.

pile base level also influences calibration

chamber

size

effects

under

be lower (for BC1, BC4) and

The confining

chamber

size effect.

BC1 were more pronounced

corresponding to shorter piles, while size effects under


at

(3)

at

stress level at the

The

calibration

low confinement,

BC2 were more

pronounced

high confinement, corresponding to longer piles;

The value of

the normalized base resistance qb/q c

not a constant, varying as a

is

function of the relative density, the confining stress, and the coefficient of lateral
earth pressure at rest;

(4)

The values of
within the 0.07

the normalized base resistance qb/q c for non-displacement piles


-

0.13 range for s/B

5%

and the 0.10

0.20 range for s/B

fall

= 10%;

227

(5)

The

effect of relative density

on the normalized base resistance qb/q c was most

was

significant, while that of the confining stress at the pile base level

higher relative densities, the value of qb/q c was smaller (qb/q c

90%)

(6)

than

at

lower

relative densities (qb/q c

The values of the normalized base

= 0.19

0.2 for

= 0.12

(7)

= 5% and

the 0.22

of lateral earth pressure

at rest

0.13 for

DR =

resistance qb/q c for displacement piles are higher

0.35 range for s/B

The value of the normalized base

At

D R = 30%);

than those for non-displacement piles, showing typically the 0.15

s/B

small.

0.25 range for

= 10%;

resistance qb/q c tends to decrease as the coefficient

(Ko) increases.

This trend

is

more pronounced

at

lower relative densities, and negligible for very dense sand;

(8)

The values of

the normalized base resistance qb/q c for silty sands are in the 0.12

0.17 range, depending on the relative density and the confining stress at the pile

base level.

The confining

stress is another

important factor that influences the

For lower

relative density, the value of qb/q c

value of qb/q c for silty sands.

decreases as the pile length increases while that for higher relative density increases;

(9)

Field test results of both

SPT and CPT performed

at the

same

site

show reasonable

agreement with the correlation proposed by Robertson and Campanella (1983).


This indicates that

CPT

pile design

method can be used with

the

SPT blow

count

for practical purpose, if the proper value of q c /N can be obtained for a given soil

condition.

228

10.3

(1)

Recommendations

Since pile design methods using in-situ


conditions, the selection of the

test results

were developed under different

method should be made with consideration of

the

differences between the methods, such as:

the criterion adopted to define pile load capacity;

the type of equipment used to obtain q c ;

the selection and relative importance of q c values above and below the pile
base;

(2)

soil

types and conditions under which the methods were developed;

Use of global
is

aimed

in

factor of safety

is

more

suited to pile design if a target level of safety

design irrespective of the relative contribution of base and shaft

resistances to total pile capacity;

(3)

The

relative density is the

piles in sands.

density be
since

(4)

it is

therefore,

made through

recommended

that the estimation of the relative

a reasonable correlation with in-situ tests such as

very difficult to obtain undisturbed granular

Piles typically
state

It is,

most important factor influencing the base resistance of

soil

CPT,

samples;

undergo relative settlements s/B greater than 10%

to reach a limit

corresponding to either a loss of functionality of structures or damage to

superstructures and/or foundations.

However, the evaluation of the

relative

settlement associated with the limit states design of piles should be done with

consideration of the type, functionality, location, and importance of structures.

229

LIST

AASHTO

(1994).

OF REFERENCES

AASHTO LRFD Bridge Design

H., Sacilotto, A.

C, De

Specification.

N. R., Teixeira, C.

Albiero,

J.

(1995).

"Successive Load Tests on Bored Piles", Proceedings of

Mantilla,

J.

Z.,

and Carvalho, D.

Pan-American

Conference of Soil Mechanics and Foundation Engineering, Guadalajara, Vol.

2, pp.

991

- 1002.
Aoki, N. and Velloso, D. A. (1975). "An Approximate Method to Estimate the Bearing
th

Pan-American Conference
Coi
of
Foundation Engineering, Bueno Aires, Vol. 1, 367 376.
Capacity of Piles", Proceedings of 5

Soil

Mechanics and

Baldi, G., Bellotti, R., Ghiona, V.N., Jamiolkowski, M., and LoPresti, D.C.F. (1989).

"Modulus of Sands from CPT and DMT", Proceedings of 12th International Conference
on Soil Mechanics and Foundation Engineering, Balkema, Rotterdam, Vol. 1, pp 165170.

Baligh,

M.M.

(1975). "Theory of

Deep

Static

Cone

Penetration Resistance", Report No.

R75-56, Department of Civil and Environmental Engineering, Massachusetts

Institute of

Technology, Cambridge, Mass.


Bandini, P. (1999). "Static Response and Liquefaction of Silty Sands", Master Thesis,

Purdue University, West Lafayette, EN.


Bandini, P. and Salgado, R. (1998). "Methods of Pile Design Based on
results",

Proceedings of

Robertson and P. Mayne

BCP

st

ed.),

CPT

and SPT

International Conference on Site Characterization


Balkema, Rotterdam, pp. 967 - 976.

Committee. (1971). "Field Tests on Piles

in

(P.

Sand", Soils and Foundation, 11(2), 29

-49.

Been, K.,

Jefferies,

M.G., and Hachey,

Geotechnique, 41(3), 365-381.

J.

(1991).

"The

Critical

State

of Sands",

230

Been, K., Crooks, J.H.A., Becker, D.E., and Jefferies, M.G. (1986). 'The Cone
Penetration Test in Sands: Part I; State Parameter Interpretation", Geotechnique, 36(2),

239 - 249.
Jamiolkowski, M., Lo

Bellotti, R.,

Presti, D.C.F.,

and O'Neill, D.A. (1996). "Anisotropy

of Small Strain Stiffness in Ticino Sand", Geotechnique, 46(1), 115-131.

Berezantev, K. and Golubkov (1961). "Load Bearing Capacity and Deformation of Piles

Foundations", Proceedings of 5

Foundation Engineering, Vol.


Biarez,

Burel, M., and

J.,

1,

th

Des Fondations", Proceedings of


1,

Conference on Soil Mechanics and

pp. 11.

Wack, B.

Foundation Engineering, Vol.

International

(1961). "Contribition
th

A L'etude De La Force Portante

International Conference on Soil

Mechanics and

603 - 609.

Bolton,

M. D.

(1979).

Bolton,

M.D.

(1986).

"The Strength and Dilatancy of Sands", Geotechnique,

Guide

to Soil

Mechanics, Macmillan

Inc.,

Greenwich, Conn.
36(1), 65-

78.
th

'The development of codes of practice for design", Proc. of 12


International Conference on Soil Mechanics and Foundation Engineering, Vol.3,
Balkema, Rotterdam, pp. 2073 - 2076.
Bolton,

M.D.

Borja, R.

I.,

(1989).

Lee, S. R. and Seed, R. B. (1989). "Numerical Simulation of Excavation in

Elastoplastic Soils", International Journal for Numerical

and Analytical Methods

in

Geomechanics, 13(3), 231 - 249.


Bowles,

J.

E. (1988). Foundation Analysis

and Design, McGraw-Hill Book Company,

Singapore.

Bozozuk, M. (1978). "Bridge Foundation Move",


Board, pp. 17-21.
Briaud,

J.

L.

and Tucker, L. (1984), "Piles

Stresses", Journal of Geotechnical Engineering,

TRR

in Sand:

ASCE,

678, Transportation Research

Method Including Residual


1), 1666 - 1680.

10(1

and Gibbens, R. (1994), "Test and Prediction Results for Five Large spread
Footings on Sand", Proceedings of a Prediction Symposium Sponsored by the Federal
Briaud,

J.

L.

Highway Administration
Texas

at the

Occasion of the Settlement '94

ASCE

Conference

A&M University, ASCE, Geotechnical Special Publication No. 41, pp. 92

128.

at

231

"Load Settlement Curve Method for Spread


Footings on Sand", Proceedings of Settlement '94, Vertical and Horizontal Deformations
of Foundations and Embankments, ASCE, Vol. 2, pp. 1774 - 1804.
Briaud,

L. and Jeanjean, P. (1994).

J.

G.B

Briaud, J.L, Moore, B.H, and Mitchell,

and

Dam

(1989). "Analysis of Pile

Load

Tests at

Lock

26", Proceedings of the Congress, Foundation Engineering: Current Principles

and Practices, ASCE, Vol.


Brinch Hansen,

J.

1,

pp.

925 - 942.

(1963). "Discussion, Hyperbolic Stress-Strain response, Cohesive

Soil", Journal of Soil

Mechanics and Foundation Engineering Division, ASCE, 89(SM4),

241-242.
Broms, B.B., Chang, M.F., and Goh, A.T.C. (1988). "Bored Piles in Residual Soil and
st
Weathered rocks in Singapore", Proceedings of the 1
International Geotechnical
Seminar on Deep Foundations on Bored and Auger Piles (Van Impe, ed.), Balkema,
Rotterdam, pp. 17 - 34.
Burland,

J.

B.

(1973).

"Shaft Friction of Piles in Clay

Approach", Ground Engineering,

6(3),

Simple Fundamental

30 -42.

and Wroth, O.P. (1974). "Allowable and differential settlement of


damage and soil-structure interaction", Proc. of Conference on
Settlement of Structures, Cambridge University, pp. 61 1 - 654.

Burland, J.B.

structures, including

Bustamante, M. and Gianeselli (1982). " Pile Bearing Capacity Prediction by Means of

CPT", Proceedings of 2
Testing, Amsterdam, Vol. 2, 493 - 500.
Static

Penetration

nd

European Symposium on Penetration

H.D. and Hoy, H.E. (1977). "Users Manual for the Texas Quick-Load Method for
Foundation Load Testing", Federal Highway Administration, Office of Development,
Butler,

Washignton, pp. 59.


Campanella, R.G., Robertson, P.K., Davies, M.P., and Sy, A. (1989). "Use of In Situ
Tests in Pile Design", Proceedings of International Conference on Soil Mechanics and
Foundation Engineering, pp. 199 - 203.

Canadian Geotechnical Society (1992). Canadian Foundation Engineering Manual, The


Canadian Geotechnical Society, Canada.
and Baladi, G. Y. (1985). Soil Plasticity: Tfieory and Implementation,
Elsevier Science Publishing Company Inc. New York.

Chen,

W.

Chen,

W.

F.

F.

and Han, D.

New York Inc.

Virginia.

J.

(1988). Plasticity for Structural Engineers, Springer- Verlag

232

Chin, F.V. (1970). "Estimation of the Ultimate Load of Piles Not Carried to Failure",

Proceedings of 2

nd

Southeast Asian conference on Soil Engineering, pp. 81

90.

Coduto, D.P. (1994). Foundation Design: Principles and Practices, Prentice Hall,

Englewood
Coyle, N.

Cliffs,

M. and

New Jersey.
"New Design Correlations for Piles
Engineering Division, ASCE, 107(GT7), 965 - 986.

Castello, R. R. (1981).

Journal of the Geotechnical

in

Sand",

Davisson, M.T. (1972). "High Capacity Piles", Proceedings, Lecture Series, Innovation

Foundation Construction,

ASCE,

DeBeer, E.E. (1948). "Donness Concamat La Resistance

An

En Ais De

22 - 40.

De

Penetration

in

Illinois Section, pp. 52.

En Profondeur", Geotechnique,

1,

Oisaillement Decluities Des

Beer, E.E. (1967). "Proefondervindelijke Bijdrage tot de studie van zand onder

funderingem op
Beigie

staal", Tijdshrift

Nos 6-67 and

1-, 4-, 5-,

der Openbar Werken van het grensdraag vermogen van

6-68.

De

Beer, E.E. (1984). "Different Behavior of Bored and Driven Piles", Proceedings of VI
Conference on Soil Mechanics and Foundation Engineering, Budapest (G. Petrasovits
ed.), pp.

De

307-318.

Beer, E.E. (1988). "Different Behavior of Bored and Driven Piles", Proceedings of

Deep Foundation on Bored and Auger

Piles

(Van Impe

ed.),

Ghent, pp. 47

82.

De

Mello, V.F.B. and Aoki, N. (1993). "Updating Realism on Large-Diameter Bored


nd
Piles", Proceedings of 2
International Geotechnical Seminar on Deep Foundations on

Bored and Auger


DeRuiter,

J.

Piles

(Van Impe,

and Beringen,

F.

ed.),

L.

Balkema, Rotterdam, pp. 35 - 42.

(1979). " Pile Foundations for Large North Sea

Structures", Marine Geotechnology, 3(3),

Desai, C.S. and Christian,

Mcgraw-Hill

Inc.,

New

J.

267 - 314.

T. (1977). Numerical

Methods

in

Geotechnical Engineering,

York.

Desai, C. S. and Siriwaden, H.

J.

(1984). Constitutive

Laws for Engineering Materials

with Emphasis on Geologic Materials, Pretice Hall, Englewood, Cliffs.

Duncan,

J.

M. and Chang,

C. Y. (1970). "Nonlinear Analysis of Stress-Strain in Soils",

Journal of Soil Mechanics and Foundation Engineering Division,

- 1653.

ASCE, 96(SM5), 1629

233

Duncan, J.M. and Tan, C.K. (1991). "Engineering manual for estimating tolerable
movements of bridges", Manuals for the Design of Bridge Foundations, NCHRP Report
No. 343, Transportation Research Board, National Research Council, Washington, D.C.,
pp.

219-228.

Durgunoglu, H.T. and Mitchell, J.K. (1975). "Static Penetration Resistance of Soils I:
Analysis", Proceedings of ASCE Special Conference on In Situ Measurement of Soil
Properties,

ASCE, New York,

Vol.

1,

151 - 171.

Eurocode 1 (1993). "Basis of design and actions on


Committee for Standardization, TC 250.

structures", Sixth Draft,

European

Eurocode 7 (1993). "Geotechnical Design", Fourth and Final Draft, European Committee
for Standardization, TC 250/SC7.
Fahey, M., and Carter, J.P. (1993).

Sand Using a Non-Linear

"A

Finite

Elastic Plastic

Element Study of the Pressuremeter Test

in

Model", Canadian Geotechnical Journal, 30(2):

348-361.
Fahey, M., Robertson, P.K., and Soliman, A. A. (1994). "Towards a Rational Method of
Predicting Settlements of Spread Footing on Sand", Vertical and Horizontal Deformation

of Foundation and Embankments,


1,

ASCE,

Geotechnical Special Publication No. 40, Vol.

pp. 599-611.

Fellenius, B. (1975). "Test Loading of Piles

of the Geotechnical Engineering Division,


Fellenius, B.H. (1980).

Engineering, Sep. pp.

and

ASCE,

New

Proof Testing Procedure", Journal

101 (GT9), 855

"The Analysis of Results from Routine

Pile

869.

Load

Tests",

Ground

19-31.

Fioravante, V., Ghionna, V.

N,

Jamiolkowski, M. and Pedroni, S. (1995). "Load Carring

Capacity of Large-Diameter Bored Piles

in

Sand and Gravel", Proceedings of

Asian

Regional Conference in Soil Mechanics and Foundation Engineering, pp. 1-13.


Franke, E. (1989). "Co-Report to Discussion, Session B, on Large Diameter Piles", 12

ICSMFE, Rio

th

de Janeiro, Brazil.

Franke, E. (1990).

"THE EUROCODE - SAFETY APPROACH AS UPRIGHT TO

FOUNDATIONS",

Proc. of 9

Foundation Engineering,
Franke, E. (1991).

th

International

th

Danube-European Conference on Soil Mechanics and

AKADEMIAI KIADO, Budapest, pp.

"EUROCODE

173 - 182.

Safety Approach as Applied to single Piles", Proc. of

DFI Conference, Balkema, Rotterdam,

pp. 13

18.

234

Franke, E. (1993).

"Design of Bored

Horizontal Loading", Proceedings of 2

Including Negative Skin Friction

Piles,

nd

International Geotechnical Seminar on

Foundations on Bored and Auger Piles (Van Impe,

ed.),

and

Deep

Balkema, Rotterdam, pp. 43 -

57.

S. (1993). "Base Capacity


nd
International
from In-Situ Tests", Proceedings of 2
Geotechnical Seminar on Deep Foundations on Bored and Auger Piles (Van Impe ed.),
Balkema, Rotterdam, pp. 67 -74.

Ghionna, V. N., Jamiolkowsi, M., Lancellotta, R. and Pedroni,


of

Bored

Piles

Sands

in

and Salgado, R. (1994). 'The Tip


Displacement of Drilled Shafts in Sands", Proceedings of Settlement '94 (Yeung and
Felio ed.), Vol. 2, Geotechnical Engineering Division, ASCE, June, pp. 1039 - 1057.
Ghionna, V. N., Jamiolkowsi, M., Pedroni,

Gibbens, R. and Briaud,

J.

S.

(1994). "Data and Prediction Request for the Spread Footings

Symposium Sponsored by the Federal


Occasion of the Settlement '94 ASCE Conference at

Prediction Events", Proceedings of a Prediction

Highway Administration
Texas

at the

A&M University, ASCE, Geotechnical Special Publication No. 41, pp.

11

85.

Giuseppe, B. (1991). "Modellazione Numerica Delle Prove Pressiometriche In Camera Di


Calibrazione Per

La Sabbia Del

Ticino", Graduate School Thesis, Politecnico Di Torino,

Italy.

Goble, G. G., Kovacs, A. M., and Rausche, F. (1972). "Field Demonstration: Response of
Instrumented Piles to Driving and Load Testing", Proceedings of Specialty Conference on

Performance of Earth and Earth-Supported Strfuctures, Vol.


Gregersen, O.

S.,

Aas,

G, and Dibiagio,

Sand", Proceedings of the 8


Engineering, Vol. 2,

USSR,

th

ASCE,

pp. 3 -38.

"Load Tests on Friction Piles in Loose


Conference on Soil Mechanics and Foundation

E. (1973).

International
ati

pp. 109

3,

117.

Grover, R.A. (1978). "Movements of Bridge Abutments and Settlements of Approach


slabs in Ohio", Transportation Research Board 678, Tolerable Movements of Bridge
Foundation, Sand Drains, K-Test, Slopes, and Culverts, Transportation Research Board,
Washington, D.C., pp. 12 - 17.

Hardin, B. O. and Black,

W.

L. (1966).

"Sand

Stiffness

under Various Triaxial Stresses",


ASCE, 92(SM2), 27 -

Journal of Soil Mechanics and Foundation Engineering Division,


42.

Hardin, B.O. and Drnevich, V. P. (1972). "Shear

Modulus and Damping

in Soils;

Design

Equations and Curves", Journal of Soil Mechanics and Foundations Division,

ASCE,

98(7),

667 - 692.

235

Wave

Hardin, B.O., and Richart, F.E. (1963). "Elastic

Velocities in Granular Sands",

Journal of Soil Mechanics and Foundation Engineering Division,

ASCE, 89(SM1),

33-

65.

Harris, D. E.

and Mayne,

P.

W.

(1994). "Axial

Compression Behavior of Two Drilled

Shafts in Piedmont Residual Soils", Proceedings of International Conference Design and

Construction of

Deep Foundations,

Vol.2, Orlando, pp. 352

367.

Harrop-Williams, K. (1989). "Risk Correction Factors for Bearing Capacity Analysis",


Proceedings of the Congress on Foundation Engineering (H. Kulhawy

ASCE,

Hirany, A. and Kulhawy, F.H. (1989). "Interpretation of


1:

ed.),

Vol.

2,

pp. 848-856.

Load Tests on

Drilled Shafts Part

Axial Compression", Proceedings of the Congress on Foundation Engineering (H.

Kulhawy

ed.),

Vol. 2,

ASCE,

pp.

132

- 1

149.

"Comparison of Predicted and


Sponsored by the Geotechnical
National Convention, pp. 413 - 433.

Horvitz, G.E., Stettler, D.R., and Crowser, J.C. (1981).

Observed

Pile Capacity", Proceedings of a Session

Engineering Division

at the

ASCE

Hu, G. (1965). "Bearing Capacity of Foundations with Overburden Shear", Sols-Soils,


No. 13, 11-18.
Jamiolkowski, M. and Lancellotta, R. (1988). "Relevance of In-Situ Test Results for
Evaluation of Allowable Base Resistance of Bored Piles in Sands", Proceedings of
International Geotechnical

Impe

ed.),

Seminar on Deep Foundations on Bored and Auger

Piles

(Van

Ghent, pp. 107 - 120.

Janbu, N. (1963). "Soil Compressibility as Determined by Oedometer and Triaxial Tests",

Proceedings of European Conference on Soil Mechanics and Foundation Engineering,

Wiesbaden, Germany, Vol.1, pp. 19-25.


Kondner, R. L. (1963). "Hyperbolic Stress-Strain Response: Cohesive Soil", Journal of
Soil Mechanics and Foundation Engineering Division, ASCE, 189(SM1), 1 15 - 143.

Lambe, W.T. and Whitman, R.V. (1986).

Soil Mechanics, SI Version,

John Wiley and

Sons, Korea.

Lee, C. Y., Allman,

M. A. and

Cemented Calcareous
Congress, Vol.

Lee,

J.

1,

Sands",

Poulos, H. G. (1989). "Static Behavior of Piles in

Proceedings

ASCE, New York,

pp.

of

the

1989

Foundation

Engineering

485 - 499.

H. and Salgado, R. (1999a). "Analysis of Calibration Chamber Plate Load Tests",

Canadian Geotechnical Journal, December.

236

H. and Salgado, R. (1999b). "Determination of Pile Base Resistance


Journal of Geotechnical and Geoenvironmental Engineering, ASCE, August.

Lee,

J.

Lee,

J.

characteristics of Geomaterials, Torino,

J.

Sand

Sands",

H. and Salgado, R. (1999c). "Pre-failure Loading Response of Foundations on


nd
International Symposium on Pre-Failure Deformation
Proceedings of 2

Sand",

Lee,

in

H., Salgado, R.,

Lightweight

as

Engineering,

Mayne,

ASCE,

W. and

P.

Shaft Foundation

Bemal,

A.,

C.W

and Lovell,
Journal

Backfill",

125(2), 132

Accepted for publication.


(1999). "Shredded Tires and Rubber-

and

Geotechnical

of

Geoenvironmental

141.

Harris, D. E. (1993). "Axial Load-Displacement Behavior of Drilled

in

Piedmont Residium", Technical Report, No. 41-30-2175,

FHWA.

Mayne, P.W. and Kulhawy, F.H. (1991). "Calibration Chamber Database and Boundary
Effects Correction for CPT Data", Proceedings of the First International Symposium on
Calibration

Chamber Testing,

Elsivier,

New

York, pp. 257 - 264.

Meyerhof, G. G. (1956). "Penetration Tests and Bearing Capacity of Cohesionless Soils",


Journal of Geotechnical Engineering, ASCE, 82(1), 1- 19.
Meyerhof, G. G. (1976). "Bearing Capacity and Settlement of Pile Foundations, Journal
of Geotechnical Engineering,

ASCE,

102(3), 197

Meyerhof,

"Scale

Effect

G.

G.

(1983).

Geotechnical Engineering,
Milovic,

ASCE,

D.M. and Milovic, S.D.


Piles,

Pile

Capacity",

of

Journal

806.

(1993). "Predicted and Observed Behavior of Piles",

st

1
International Geotechnical Seminar on Deep Foundations on Bored
Balkema, Rotterdam, pp. 381 - 384.

Proceedings of the

and Auger

228.

Ultimate

of

109(6), 797

Moulton, L.K. (1986). "Tolerable movement

highway bridges", Final Report


NO. FHWA-TS-85-228, Federal Highway Administration, Washington, D.C.
criteria for

Moulton, L.K., GangaRao, H.V.S., and Halvorsen, G.T. (1985). "Tolerable movement
criteria for highway bridges". Report No. FHWA/RD-85/107, Federal
Highway
Administration, Washington, D.C.

Naylor, D.J., Pande, G.N., Simpson, B., and Tabb, R.

(1981).

Finite Elements

in

Geotechnical Engineering, Pineridge Press, Swansea, Wales, U.K.


Neely,

Sand",

W.

(1989). "Bearing Pressure

Proceedings

Practices,

ASCE,

of Congress

Vol. 2, 979

990.

- SPT

Correlation for

Foundation

Engineering:

Expanded Base
Current

Piles in

Principles

and

237

Neely,

W.

(1990). "Bearing Capacity of Expanded-Base Piles in Sand", Journal of

ASCE,

Geotechnical Engineering,
Neely,

W.

(1991).

"Bearing

87.

Capacity of Auger-Cast Piles

ASCE,

Geotechnical Engineering,

116(1), 73

117(2), 331

in

Sand",

Journal

of

345.

Ovesen, N.K. and Orr, T. (1991). "Limit State Design - The European Perspective",
Geotechnical Engineering Congress, Geotechnical special publication No. 27, Vol. 2,

ASCE,

pp. 1341

Parkin,

A.K.

- 1352.

Proceedings of the
Elsivier,

New

"Chamber Testing of Piles in Calcareous Sand and Silt",


First International Symposium on Calibration Chamber Testing,

(1991).

York, pp. 289-302.

Polshin, D.E. and Tokar, R.A. (1957).


structures", Proc. of

th

"Maximum

allowable non-uniform settlement of

International Conference
Con
on Soil

Engineering, London, Vol.

1,

pp.

Mechanics and Foundation

402 - 406.

Poulos, H. G. (1989). " Developments in the Analysis of Static and Cyclic Lateral

Response of Piles", Proceedings of rv International Conference on Numerical Methods


Geomechanics, Vol. 3, Rotterdam, Balkema, pp. 1 1 17 - 1 135.
Price, G.

and Wardle

I.

F.

(1982).

"A Comparison Between Cone

in

Penetration Test

Results and the Performance of Small diameter Instrumented Piles in Stiff Clay",

Proceedings of 2

nd

European Symposium on Penetration Testing, Amsterdam, Vol.

2,

775

- 780.
Purzin, A.

M. and

Burland,

J.

B. (1996).

"A Logarithmic

Stress-Strain Function for

Rocks

and Soils", Geotechnique, 46(1), 157 - 164.


Reese, L. C. and O'neill,

M. W.

and

(1988). Drilled Shafts; Construction Procedures

Design Methods, Report No. FHWA-FfJ-88-42, U.S. Dept. of Transportation, Washington


D.C.
Reese. L. C. and O'Neill,

Common

Soil

M. W.

(1989).

"New Design Method

for Drilled Shaft

from

and Rock Test", Proceedings of Congress Foundation Engineering: Current


ASCE, Vol. 2, 1026 - 1039.

Principles and Practices,

Robertson, P. K. and Campanella, R. G. (1983).

Geotechnical Engineering, 109(11),

ASCE, 1449 -

"SPT-CPT

correlation", Journal of

1459.

Robertson, P. K. and Campanella, R. G. (1989). Guidelines for Geotechnical Design

Using the Cone Penetrometer Test and

Hogentogler

& Co., Columbia, MD.

CPT

with Pore Pressure Measurement, 4

ed.

238

Robertson, P. K., Campanella, R. G., Gillespie, D., and Rice, A. (1985). "Seismic

CPT

to

Measure In-Situ Shear Wave Velocity" Proceedings of Measurement and Use of Shear
Wave Velocity for Evaluating Dynamic Soil Properties, ASCE, 34 - 48.
Salgado, R. (1993). "Analysis of Penetration Resistance in Sand", Ph.D. Thesis, Univ. of
California, Berkeley, California.

Salgado, R. (1995). "Design of Piles in Sands Based on

CPT

Results", Proceedings of

Pan-American Conference of Soil Mechanics and Foundation Engineering, Vol.


Guadalajara, pp. 1261

3,

- 1274.

M.

Salgado, R., Mitchell, J.K., and Jamiolkowski,


Penetration

Resistance

Engineering,

ASCE,

in

of

Journal

Sand",

(1997a). "Cavity Expansion and

and Geoenvironmental

Geotechnical

123(4), 344-354.

Salgado, R., Boulanger, R.

CPT

Liquefaction

W. and

Resistance

Mitchell,

J.

K. (1997b). "Lateral Stress Effects on


Journal

Correlations",

of

Geotechnical

and

Geoenvironmental Engineering, ASCE, 123(8), 726 - 735.


Salgado, R., Dmevich, V.P.,

Ashmawy,

A., Grant, W.P., and Vallenas, P.

(1997c).

"Interpretation of Large-Strain Seismic Cross-Hole Tests", Journal of Geotechnical and

Geoenvironmental Engineering,

ASCE,

123(4), 382-388.

Salgado, R., Mitchell, J.K., and Jamiolkowski,


Penetration Resistance

Measured

Geoenvironmental Engineering,

in Calibration

ASCE,

M.

(1998a).

"Chamber Size

Effects on

Chambers", Journal of Geotechnical and

124(9), 878

888.

Salgado, R., Jamiolkowski, M., and Mitchell, J.K. (1998b). "Penetration Resistance in

Sand: Analysis and Applications to Liquefaction Potential Assessment and Estimation of


Pile

Base Resistance", Rivista Italiana

di Geotecnica, Vol,

XXXII, No.4, 5 -17.

Salgado, R., Bandini, P. and Karim, A. (1999). "Stiffness and Strength of Silty Sand",
Journal

of Geotechnical

and Geoenvironmental Engineering, ASCE,

accepted

for

publication.

Schmertmann,

J.

H. (1970). "Static Cone to Compute Static Settlement Over Sand",

Journal of Soil Mechanics and Foundation Engineering Division,

ASCE, 96(SM3), 1011

- 1042.
Schmertmann,

J.

H. (1978). "Guidelines for Cone Penetration Test, Performance and

Design", U.S. Department of Transportation,

FHWA-TS-78-209.

239

Schmertmann,

J.

H., Hartman,

and Brown,

J. P.,

P. R. (1978).

"Improved Strain Influence


ASCE, 104(GT8),

Factor Diagrams", Journal of the Geotechnical Engineering Division,

1131-1135.
Schnaid,

and Houlsby, G.T. (1991). "An Assessment of Chamber Size Effects

F.,

in the

Calibration of In Situ Tests in Sand", Geotechnique, 41(3), 437-445.

Seed, B., Tokimatsu,

Procedures

in

Shibuya,

S.,

111(12),

Tatsuoka,

F.,

M,

Harder,

and Chung, R. (1985). "Influence of

Resistance Evaluations",

Liquefaction

Soil

ASCE,

Engineering,

K.,

1425-

Journal

SPT

of Geotechnical

1445.

Teachavorasinskun,

S.,

Kong,

X.J.,

Abe,

F.,

Kim,

Y.,

and Park,

C. (1992). "Elastic Deformation Properties of Geomaterials", Soils and Foundations,


32(3), 26-46.

Simonini, P. (1996). "Analysis of Behavior of Sand Surrounding Pile Tips", Journal of

Geotechnical Engineering, 122(11),

ASCE, 897 - 905.

Skempton, A. W. (1986). "Standard Penetration Test Procedures and the Effects in Sands
of Overburden Pressure, Relative Density, Particle Size, Ageing and Overconsolidation",
Geotechnique, 36(3), 425 - 447.

Skempton, A.W. and MacDonald, D.H. (1956). "Allowable settlement of buildings",


Proc. of Institute of Civil Engineering, Part m. Vol. 5, pp. 727 - 768.

Move", by Bozozuk,
Transportation Research Board 678, Tolerable Movements of Bridge Foundation, Sand
Drains, K-Test, Slopes, and Culverts, Transportation Research Board, Washington, D.C.,
pp. 17-21.
Stermac,

A.

G.

(1978).

"Discussion

of Bridge

Foundation

E., and Warden, P. (1994). "Footing Load Tests on Sand",


Proceedings of Settlement '94, Vertical and Horizontal Deformations of Foundations and

Tand, K., Funegard,

Embankments, ASCE, Vol.

1,

pp. 164

178.

F., Siddiquee, M. S. A., Park, C, Sakamoto, M. and Abe, F. (1993).


"Modelling Stress-Strain Relations in Sand", Soils and Foundations, 33(2), 60-81.

Tatsuoka,

Teachavorasinskun,

Monotonic

and

S.,

Shibuya,

Cyclic

S.,

Torsional

and Tatsuoka, F. (1991). "Stiffness of Sands in


Simple Shear", Proceedings of Geotechnical

Engineering Congress, Geotechnical special publication No. 27, Vol.1,


878.

ASCE,

pp. 863

240

Teixeira, C.Z. and Albiero,

"A Evolucao da Reacao de Ponta de

H. (1994).

J.

Escavadas Submetidas a Sucessivas Provas de Carga", Proceedings of


Conference on Soil Mechanics and Foundation Engineering, Vol.1, Foz
-9.

Do

Estacas

Brazilian

Iguacu, pp. 3

Terzaghi, K. (1943). Theoretical Soil Mechanics, John Wiely and Sons Inc.,

New

York,

N.Y.
Terzaghi, K. and Peck, R. B. (1948). Soil Mechanics in Engineering Practice, John Wiely

and Sons

Inc.,

New

York, N.Y.

Terzaghi, K. and Peck, R. B. (1967). Soil Mechanics in Engineering Practice, 2

John Wiely and Sons


Tomlinson, M.

J.

New

Inc.,

(1971).

Edition,

York, N.Y.

"Some

Conference on Behavior of

Piles,

Trochanis, A. M., Bielak,

J.,

Effects of Pile Driving on Skin Friction", Proceedings of

ICE, London, 107 -

and Christiano,

14.

P. (1991).

Study of Piles" Journal of Geotechnical Engineering,

"Three-Dimensional Nonlinear

ASCE,

17(3),

429

447.

Metodo Degli Elementi Finiti Delle Prove


E Pressiometriche Utilizzando La Legge Constitutiva Di Lade", Graduate

Vecchia, G. (1991). "Modellazione


Triassiali

nd

School Thesis, Politecnico

di

Torino,

Con

II

Italy.

Vesic, A.S. (1972). "Expansion of Cavities in Infinite Soil Mass", Journal of Soil

Mechanics and Foundation Engineering Division, ASCE, 98(3), 265 - 290.


Vesic, A. S. (1977). "Design of Pile Foundations",

NCHRP

Synthesis of Practice No. 42,

Transportation Research Board, Washington, D.C., pp. 68.


Vesic, A. S. (1963). "Bearing Capacity of

Rec. No. 39,

Deep Foundations

in

Sand", Hwy. Res. Board

112-153.

Viggiani, G.

and Atkinson, J.H. (1995). "Interpretation of Bender Element Tests",

Geotechnique, 45(1), 149-154.


Vijayvergiya, V. N. and Focht,

Clay",

OTC paper

1718,

th

J.

A. (1972).

"A New Way

to Predict Capacity of Piles in

Offshore Technology Conference, Houston, TX.

"Estudo da Compressao Unidirecional do Sedimento Moderno (Solo

Vilar (1979).

Cidade de Sao Carlos", M.S. Thesis, School of Engineering of the


University of Sao Paulo (USP), Sao Carlos, Sao Paulo.
Superficial) da

241

Wahls, H.E. (1981). "Tolerable settlement of buildings". Journal of Geotechnical


ASCE, 107 (GT11), pp. 1489 - 1504.

Engineering Division,

Wahls, H.E. (1990). "Design and Construction of Bridge Approaches", National


Cooperative Highway Research Program Synthesis of Highway Practice 159,
Transportation Research Board, National Research Council, Washington, D.C. pp. 45.
Wahls, H. E. (1994). "Tolerable Deformations", Proceedings of Settlement '94, Vertical
and Horizontal Deformations of Foundations and Embankments, Vol. 2, Geotechnical
Special Publication No. 40, ASCE, pp. 1611 - 1628.

Walkinshaw,

J.L. (1978).

TRR 678, Transportation

"Survey of bridge movements

in the

western United States",

Research Board, pp. 6-12.

Xanthakos, P. (1995). "Bridge Substructure and Foundation Design, Prentice Hall,

New

Jersey.

Yokel, F.Y. (1990). "Proposed Design Criteria for Shallow Bridge Foundations", Dept. of

Commerce, National

Inst,

of Standard and Technology, Gaithersburg,

Yu, H.S. and Houlsby, G.T. (1991). "Finite Cavity Expansion


Analysis", Geotechnique, 41, 173

MD,

pp. 50.

in Dilatant Soil:

Loading

183.

Yu, H.S. and Mitchell, G.T. (1998). "Analysis of Cone Penetration Resistance: Review of
Methods", Journal of Geotechnical and Geoenvironmental Engineering, ASCE, 124(2),
140 - 149.
Yu,

P.

and Richart, F.E. (1984). "Stress Ratio Effects on Shear Modulus of Dry Sands",

Journal of

Geotechnical Engineering,

ASCE,

110(3), 331-345.

242

APPENDIX: THE

A.l

PROGRAM CONEPILE

Introduction

The program CONEPILE


base and shaft resistances.

base resistance:

is

for calculating the pile load capacity including both the

The methods adopted

in the

program

CONEPILE

are:

Aoki and Velloso's method;

LCPC

method;

proposed method (Lee-Salgado method);

shaft resistance:

Aoki and Velloso's method;

LCPC

The

details of

method.

each method can be found

in

Chapter

The program CONEPILE can

2.

calculate the base and shaft resistances for both non-displacement and displacement piles.

Users can select the method desired to use for a given

The program CONEPILE


interface

program

calculation.

The

for pre-

consists

soil

condition and pile type.

of two different programs:

and post-processing, and the

FORTRAN

user-friendly

code used

program was developed using Visual-C for easy

user-friendly interface

input and output data processing.

A.2

A. 2.1

Guidelines for Running

Starting the

(1) Click the icon

CONEPILE

program

named "CONEPILE"

in

Windows.

(2) Click the "start" button (Figure A-l).

(3) Click the "exit" button if the

in actual

program needs

to be terminated (Figure A-l).

(4) Click the "continue" button after reading the introduction (Figure A-2).

243

^^s^i!?m*hWMmi\mKtJ M~'i-'^\
*

[4

r>'-

J Debug
AIR

li

ill

Wl Iffl

I' |

|iiF"'""'-

TT-rr

lo* S*o

sl xl^iBl

-1

xlsffl a

->l

|>|s|S

ConJenti of T>ebu9'

Name

Seel Twe

> ImroOlsobi
IcpahaltlOlgoti

Modfied

58KB

Intermedwte Fie

10/26/38

27KS

Iniwmediate Fie

10/26/38

857 PM
857 PM

Lcpcs*wft2Dlgobi

^ lcocshettDlo,o6i
C

.
"

LesatmeDirjoci

Ql lpbaseDlg.abt
^ LpmetacDig, at*

6
n

IF
IP
fCONEPlLL

LcJs2D^ob
LpmdexiDlg.o&i

OutovavOl^obi
Ouiffv()Dlo.oCn

VERSION

"

i|

1.0

OubpavDtoobi
outttoDiaotj,

E1

^ OutbtOlo.obr
SCONEPILE
*J protect *

*? pfQWClO&

637KB

imetmed.ateFie

plOfeCtltt

103KB

imemedtaie Ft c

63KB

imemediale Fie

40K3

imetmediale Fie

7KB

imeimediale FJe

11GKB

Intermediate Fe

SlcJfch.oti

01

40

21

>db

vc40pdb

If?

& 56 PM
PM
10/23/38 1255 PM
1 C/2S/38 & 57 PM
1 0/26/98 8 56 PM
0/2S/38 8 57 PM
0/26/38 & 57 PM

F*e

2.849*B

p<aecipdb

wowctDlgott

01

immediate

piwecipch

0/26/38

10/23/38 1255

z.

233KB

3BSll| ^JEntorgDabug

3yHicTMoHWotd Arerefa

Figure A-l.

Ek E

I*

>>

f CONEPIIT

||

Start of the

(0O HSPM

program CONEPILE.

a*

JO ebug

EI

jd

*|ib|kfi|

-*|

XIeS

*a|SU8

Aire Contort* of "Dabugf

Siw|T*

Nana
;
.

-1

iModfed

IrXtoDtflOb

T> LepexhaftlDlgobj

INTRODUCTION

P< Lcpcstw!i2Dlg-ob
;

LcpctfwftDlaotn
1

Ca Lat>aseOlg.oti

O LotwieOlfjobi
O LoeidwDlftobi

T h proojam can

be ad for defending pia beanie capacity bora CPT

tesuk*.

lAraJabb method:
AflfttWetae method: shaft and bate tendance

^UjrxJewlObaab,
tano>, tiy and dayey

^ U>ndea2Dlaobi

sot

O LpndexsOirjObi
"*

Lsa-Salgada tnslhed base nortancs

OuiavsvDloobi

O OJavipOlgobi

undyioi

^OiibavOlsotx

ncfMBptaceisonl and dnptarianert ofcj

? OudobDlooDi

3
Q

LCPC method:

sheR and baas isattance

OuiMvOigobi
sand;. s*v and dayey

ti

OutlilpDIg-obi,

BCONEPILE
uf]

projects

pioiect obj
|!

01 DfOiecLpch

-J

piaec Pdb

5)

ptoieci

Sic^Kob,

^O
1

vciOdb

0/26/38 8 57

69KB

Iniemiecfcate Fas

40KB

Iraenrwtoe Fie

10/26/98

7KB

Intennedete Fie

0726/38

(rtetmedete

0/26/98 8.57

21

11&KB

voULpdb

oupci[ii*

^BSuyl|

ik

r> pioiectOlg.obj

!X~>

File

PM

856 PM
857 PM

PM

23SKB

.^JE<r*yT>g- Debug

3?MooollWii-acw*i
|

Figure A-2.

|l-CONEnLE

Introduction of the program.

^j^* 12I8AM

244

A.2.2

Selection of the

method

for calculating the base resistance

(1)

Click a button corresponding to the method to be used (Figure A-3).

(2)

Aoki-Velloso method (Figure A-4):

Required input parameters:

pile

index number;

pile

diameter (B);

cone resistance

(3)

at the pile

base level (qc );

Click "pile index" button to find a pile index number for a given pile type;

Click "continue" after entering

required parameters.

all

Lee-Salgado method (Figure A-5):

Required input parameters:

pile

index number;

pile

diameter (B);

representative cone resistance between

below

and 2B

pile base level (q c );

effective vertical stress at the pile base level (a' v );

coefficient of lateral earth pressure at rest (Ko);


relative settlement (s/B);
critical state friction

base

(4)

Click "continue" after entering

LCPC method (Figure

is

all

located

(<)) c

angle for the material where pile

);

required parameters.

A-6):

Required input parameters:

index numbers for pile and


pile

soil types;

diameter (B);

cone resistance

at the pile

base level (q c );

Click "index box" button to find index numbers for given pile and

Click "continue" after entering

all

required parameters.

soil types;

245

ammmmsmzm^
aJFt Corarts

ci

T>!bt5

Kae

3]

s|i5i_

...

msMsmmm

lntioOlg.obi

? UpcshaftlDlgobi

OLcpcshart2Dlsobi

lcpcshaft01g.ob|
?" LesobajeDbobi

CALCULATION OF BASE RESISTANCE

Olj*eDloobi
E? LpnctexOtBObi
? LprdewlDlgobi

Pteaee

Lpr*ie2Dlgob,
"* LDrtcwOlpoh

Mtod the me*od j*u mh to un fa the etfeaHbn of ban

tanctt

JovovOlgobi
O Oiivte0lg.obi
^ OiibavOigobi
O
O

OutlpipDlgobi

Outbv0igobi

AoktVcloatT^

[|

LCPC

lae-5alpado

OutdpOlgobi

HCONERLE
*)pioiect*.

*? pqeci.o6i

O pioieclpch
I? p>oicc(.pdb

f
O pioieclDlgobi
pfoiect.ies

Ga SidAfxob,

db

O vc40.pdb

liT
1

vc)

63KB

Iniermediale Fie

10/25/38857 PM

40KB

Intermediate Fie

10/26/98

ln(etmio(e

10/26/98

217KB
116KB

Imetmedale Fie

File

856 PM
857 PM

10/26/98 8 57

PM

239KB

obJK&)jdccted

3BSlt| ^|>piorng- Debug

^MboaoftWttd-Appcndb.

Selection of the

Figure A-3.

T CONEPILE

method

.^.-OO

1221AM

for calculating the base resistance

EMI]

REnilbnKi'-Oelan':-

Be * frm

look

U*

-J Debug

zi

xNel

xirffl

.1>1=l

AIFt:Conrjt>obLtf

Hart,
ni InlioO^OC-

Tjt

Mcdfed

AJi.LMJJ.U.I..UI,!.l.l

-1

-_J:.J.-_^_

II

t? LcpcshafMDIgobi

LcDChart2Dlgobi

Cl LcpcshaltDlg.obi

K
d

Ack>V3D method ha &a= reatfoncc

<

D LabeOlg.ob,
O LpbateOlgobi

>

Bdton >cu enter the input data, find tho appropriate inde* mebai
pie tjK id wtkh jou
tocabiale the ban wjstanec.

mh

C? LOMteaOkobl

for

the

<e

LoMJewlOlgobi
re

Lprriew2Dlg

t
D-re

PwnoexriuroDec

8elndr

jj

OutavipOlgobi

[? Outr&ovOJgobi

OudpteDlgobi

Pfc danet-r (en*

Com mmiancs at pis base level IkPaJ

|"~

OutliavDigobt

or
c

TodoH^pbanokfctenilmbcM.

obi

^ LpndewOlg-Dbi
D OiiavavDlg.061

D OotlsloOlg

obi

HCONEPILE

frrtw*

*\ pcorect.*
r? pmjfirt obj
prorectoch

C? p(Oiect.odb
01 pra.ect.ies

Imermedaie Fie

10/26/38

Wermaoate Fie

0/26/98

IntenwdMte Fie

0/26/38 & 57

116KB

Irtwmeiiate Fie

0/26/38

StoAhtobi

vc40.D<fc

lot* ctli] mbjcpm

23!

aiSIa| JjEstong -Debug

Figure A-4.

857 PM
856 PM

69KB

40KB
217K8

proiectDIg obi

C> v C .idb

PM

857 PM

SB
|

yHjoojcHWonl-Ocoon-.l

-jCDPIaye. -Kg] 00:25

tcOHEPILE

SS&SS

123SAM

Aoki-Velloso method for calculating the base resistance.

'

246

tdl

B*

In*

"3 a]

^CletuB

^l>|s( s_

j|na|ej_^ [x|iff|

AlFeContenttofTJebuo'

3
iwcOlftobi
Lcpciheft1Dlg,obi

Laxihaft2DtG,obi
< Lee-Saioado method fa bats raeetanc* >

LcccshaftOlo,oti
..-'.

jbasrf

.i...^-.

...

-Fie aide* rwrter


(0 (or nGrxfepiaceewt pin.

1 ft*

dBpiaceraent pies)

LcrttoolDlfl ooi
:

:'r

Fie dametM font

:-.."

.
i

:r

Cone tesriance

tDndexiip,obi

OuavavDlg.061

- Vafical

OiisvbOigob

Cooftaent of

OutHcDigoOi
OuiiavOlcobi
-Cl

pie bene level (kPa$

Ask at pie ba level (kRa)


pnaue at ie*t Rot

lateral earth

[faN.Ceand.Ko-0410 05]

OuHpovO-gobi

.'.

at

afiecove

Retetwe

Critical

mUbm4 (nuifcer

ii

decinafc

Iricbonanoje at pie baactovrifrdeoj ewt

i<*

CONERLE

proectoeh
DfOtecLDOb

frora

cored 'ei
piorectOlg. ooi

StoAfxobi

vc40.db

""

ri*

,,
:.

_- TT
,

:_-^T

40KB

InlefmedaleFte

857 PM
1C/26/38B56PM

217KB

IrtermedateFie

10/26/98 &57PM

116K9

Intermediate fie

10/26/36657 PM

J|Stt| ^JEfftema- Debug

10/26798

^Mna^Wcd-Appv_| .ffiDBaiwIOSj 01.22

\^s- S?

V 1246AM

Lee-Salgado method for calculating the base resistance.

Figure A-5.

dt>aw-Ioob H<*

fib
'^l

Eranaaic

IntemecWe F*e

239K1

lobpoctfri selected

ni

S9KB

~H gO

Debug

*I^M

rt
l

X|gl

a ftl>lst5"

AlFt Cottttrf'Mug'

SwJT*

Han
C InDoOiflobi

O
'

l>

.1

JMadfcd

LcpcshaH1Dlo,obi

" LcpcsheiT2Dlftobi

O LcpcjhahDlo,obi
Q

'

"*>

D
IC

al

HJUU^'..,".l'.|.".H,i.yWHiliiiiiiiiiiiiMiiMi^ilMxt

O
O

LptwaOigobi

Mm

*e

LpoTdexDig.otH

LordoalDloobi
|

Todot^iacpleMeciektherdeibox
.

LondemDlgobt

GS OiiovavOlgooi

re

[RC method lot bsss wsBcicb >

numberfcr&ie
appropriate
Baron yaii enter (he input date, find
pis and soi type* ha Wich joj nidi to caJcatats (he base wwtance.

? Lpnde20lg.ooi

<

LwabaseOlsiobi

f? O^avipD^obi

index njoto to pie and sol t^cxr

fndexBo*

OmlDevOlaobi

-PfedametefaBi

01 OutlplpDlo,ooi

O OutypOlsobi

Cone wwetanco at pie be** level {kPat

g CONERLE
)pioiecL*,

O
.

Continue

pioiecLobi

protect

pdb

protect, res

pwctDle,

or*

StoAhtoti

l^
1

*? proiectpchi

P
?

63KB

Intemedaie Fie

1Q/2S/38

857 PM

40KB

Intefroedate Fie

0/26/38

217KB

IniermecWeFJe

0/25/98 8 57

856 PM
PM

IrtermeoiaieFJle

10/26/38857PM

vciO jdb
vciO.pdb

116KB

obrdEsJs
|

Figure A-6.

ayMcie.JWd-Acpji.| >jCDPtoyg-HBHB23

|-rcotlPILE

9Va*

LCPC method for calculating the base resistance.

247

Selection of the

A.2.3

method

for calculating the shaft resistance

(1)

Click a button corresponding to the method to be used (Figure A-7).

(2)

Aoki-Velloso method (Figure A-8):

Required input parameters:

pile index

number;

pile diameter (B);

number of layers

for shaft resistance calculation;

thickeness, soil index number, and cone resistance q c

of each sub-layer;

Click "pile index" and "soil index" buttons to find pile and

given pile and

LCPC method (Figure

(3)

index numbers for

soil types;

Click "continue" after entering

soil

all

required parameters;

A-9):

Required input parameters:

pile diameter (B);

number of

layers for shaft resistance calculation;

thickness,

LCPC

index number, and cone resistance

q c of each sub-layer;

"LCPC

index" button to find index numbers for given pile and

Click

Click "continue" after entering

A.2.4

all

soil types;

required parameters;

Output (Figure A- 10)

(1)

Click "show results" button to get the calculated base and shaft resistance.

(2)

Click "end" button to terminate the program.

248

fe

loob Uet>

yiew

J Debug

Al

Ft

~3

1 xifctel

xtet Hi^lsSr

Cortonfc of Debug'

N.,

Sc-slTjce

Modified

LcpcihafilDlgobi

Lcohatt20lflobi

'

^ LcpcshaftDljobi

''

O Labe01sL0b!
LpbaieOlgobi
^ UxndodDlftobi

"

CALCULATION OF SHAFT FCStSTAHCE

OLwxtedlDlgobi

R Lpnde2Dtg

Pleme tded

obi

*ie method you

wsh to tae tor the calciiebon of shed

wwlwu.

" UwTdewOlQobi
OutavavOlgobi

^OjavlpDtgob)

O OuHoavOlflob.
OiilpbDlflobi

"T

!?"

(X

OiibavDlgobi

HCONERLE
H

kCPC

IVr*SSF3

OutbloDtg.obi

project, ft.

owed obi

piotecLpch
pdb
_ ^

cxoiect

project, res

9 proiectDlg at*

^StdAKobi

Li

vciOidb

S?

vcittodb

69KB

Irtermeciaie Fie

40KB

Irtferrnectate

Re

217KB

Intwmedwte F3e

116KB

Intermediate Fie

0/2S/98 B57

10/26/98

PM

&5E PM

57PM
107207386- 57 PM
10/26/38 8

zl

2333
Jj

Eistarrq

Debus

3a>PteimHL.| -rcoNEPBj

STMicrowftWcni-

^s

SW

130AM

Figure A-7. Selection of the method for calculating the shaft resistance.

^^^^fei^afeifeaissifefea -

<

5 IntJcOkjoci

Belare

*? LcpdhaHlDlgobi

pis

O Lcocihari2D 15001
^ LcpcshattDlg,0O]
O LesebeseO

gg

AokrVdbso method fa shaft rentance >

you ertat tvj npU date. &id the appropnaM ridex nmber to the
aid Hri i^se* la -*hch you wjh to c*ia/ate th jhart wnsanca

-Ptendon^her

Fjfatod

[j

Soitniex

10,0b,

D LpbaseO^ot,
^ LpndeDlg.obi

-PSedHneterfcrri

f~

Naoilaacofo-jhoHn

cAjiattsnPOmoxt

OLondewlDlflobi

O Lpndew20lo.obi

LaperKo.

Layer thcJcnest [mt

SoJrdexPta.

PiWa)

T" Lpnde*sDlgot>

O OL*avavDlo,obi
P> OutovipDl^obi
*? OutpavDlg.ot>
OuibtoDioobi

E? 0ullHlv0IO.Ob|

^ OuibJoDlg.obi
HCONERU
_*) protect,

ft,

^
pioiecLpch
protect obi

01

ptoiect.pab

O ctoiectres

PoiectDla.ofci

StdAhtobt

If?
1

Contrae

*0-Pdb

etweti:!

o^KttC

^BSw[

TTOT

0726799 8 57

PM

Z33K8

gjExploio- Debug

Figure A-8.

InterroecieieFile

^yMipowftWord-.

-jCPPhi-|tt3)ll-|
I

l-CONEPHE

^J-i

*>

134AM

Aoki-Velloso method for calculating the shaft resistance.

249

f*pfewng.-;Dbug
Efe

a*
-l.l.l.'lJJIIJIJJJiWimiMP.MIimUM.UJ

|_i Debug
l Fc

:J

* <* let

"
7

CMe^oToebug

3[^

<
IntroOtflobi

LTPC method far. ihrft rarisance >

Before jou enter the incut this, fhd the apcHpoate nfex number tor the
P6c and joJ tjipry f whcti >x*j wei to cataiore ihe tfwft icjiaance.

S LcpcshoH1DI*obi
O l_CDCthaH2Dlg.obi
n LcpcahattDlg.otn

Ptedamaretlcml

LpbweOlgobi

Mo. of layer*

tor

Aaft

femarwcaksiabonpOtMxJ

9* LpfidwDlgobi

O LpB^tealDlg-obt
LorKJeo20lo.ob(

UverHa

UTCndnNo,

Uy thickness (m}

QcftPal

jCPC Index

LorxteeOlBobi

O Q^avavOlgobi
D OUavlpDlgobt
$ OufcavOlsobi

D OutbDiB-obi
5> CMbavOlg
* OiibbOlg

obi

obi

gCONERLE

projedobj
O

DrneClpCh

f? projecLpdb
'T1 protect

praectOlg obi

Cam*

^StdAfxobt

L^
1

vc40Jdb

D vc40.p*

116KB

^BSHrt| .aJExptang- Debug

^fMiootdiWord- ]

ffCP

Playgr

[0610-j

,^W

TCONEPtLE

LCPC method for calculating the

Figure A-9.

n txptoww

1lVX^5S57PM

lo.ermrd.are Fie

'239KB

obrcdtj tetecwj

i*4AM

shaft resistance.

v0s6uo

E
AlFtfContertioCDetug*

"-1|

Haw
KMtodEi) UtednrheCosfUWion

? IfWoOltjob

O LcrxihartlDlg.ol
&* Lctxshatt2Digo(

n
O

LcpcthoftOlg.ob

LesisswOlg.ob|
Shaft

mistncc

AofcWefetornefcod

> LoboaeOtgobi

O
O

Lj^noexDlftobi

LpndewlDlgob

P* Lowfcxs2Dlg.ob

ShowRosJrif

O LorxiewOlgobi
O OutavvDloc*i
?* OuiovlpDigobi

D OuitpovO^obi
P OuHDODlaobi

Sm retiaaxs.

'

fcNl

?* OuiiavOl&ob

O OutlsbOlg

obi

IkM)

HC0NERL
^J pmjed.lt

O prO|ecLob(
^ ptotectpch
r*

oroiec!

pdb

~^~3

protect, res

01 proredDlg

obi

O StdArxoh
O
n
iii
1

vc40db
vc4Q.pdb

ohecl|l**>cMd

lr*eimedw*e

He

lu/A/3Bb/HM

233KB

iJ8SUrt[ .jJEataiig Debug

3TMioi>iBWcri

jCDPtwtlBlILl TCONEFftE

Figure A- 10.

Output

^is--.0O 150AM

results.

You might also like