i

OF TECHNolo

JUN 19 1963

LIBRA R*

THE RELATIONSHIP OF THE EFFECTIVE STRESS-STRAIN

BEHAVIOR OF A SATURATED CLAY TO THE PATE OF STRAIN

by

Archie Merrill Richardson, Jr.

B. S., Allegheny College 1949

B.S. in C.E., University of Pittsburgh 1952
M.S. in C.E., University of Pittsburgh 1956

SUBMITTED IN PARTIAL FULFILLMENT OF THE

REQUIRMENTS FOR THE DEGREE OF
DOCTOR OF SCIENCE

at the

MASSACHUSETTS INSTITUTE OF TECHNOLOGY

1963

Signature of Author
Department of Civil EngineeringV, March 2d, 1963

Certified by
Thesis Supervisor

Accepted by
Chairman, Depap en binmittee,/1J
Dduate Students

V/
 Abstract of

THE RELATIONSHIP OF THE EFFECTIVE STRESS-STRAIN

BEHAVIOR OF A SATURATED CLAY TO THE RATE OF STRAIN

by

Archie Merrill Richardson, Jr.

Submitted to the Department of Civil Engineering on Febru-

ary 1, 1963, in partial fulfillment of the requirements for the degree
of Doctor of Science.

Results of a laboratory investigation of the influence

of a rate of strain on the effective stress-strain behavior of a
remolded, saturated clay are presented and discussed. Triaxial
compression tests were performed on 1A-" dia. x 3" long specimens
at rates of strain ranging from 1% in -A minute to 1% in 1440
minutes. Two extreme preshear stress histories are represented,
i.e. normally consolidated and heavily overconsolidated. Pore
pressures were measured in all tests at the sample midheight
employing a porous probe, electric pressure transducer system.
In addition, base pore pressures were measured in selected tests,
also with an electric pressure transducer.
The results indicate an increase in strength in the
fast tests of about 12o over the slow tests in the normally
consolidated and about 20%o in the overconsolidated tests. The
pore water pressure at failure in both normally consolidated
and overconsolidated samples was lower in the fast tests. In
the overconsolidated samples, evidence of a pore water migration
was obtained while pore water migration appears to have had
little influence on the normally consolidated samples.
In tests with increases in strain rate from slow to
fast at predetermined strain levels, the evidence of a dilat-
ancy component of the behavior of normally consolidated samples
was demonstrated.

Thesis Supervisor: Robert V. Whitman

Title: Associate Professor of Civil Engineering
 ACKNOWEDGMENTS

The author wishes to express his sincere appreciation

of the advice and encouragement of Professor Robert V. Whitman
who supervised this thesis. The research described herein
could not have been completed without this help and without
discussions with Mr. Richard H. Harkness, Mr. Kent A. Healy
and the help of Mr. Carl M. Stahle.

1
 CONTENTS

Page No.

Preface

Previous Reports

List of Symbols iii

Chapter 1 - Introduction and Background

A. Introduction

B. Background 3
Chapter II - Description of Test Program and Test Procedure 9
A. Introduction 9
B. Description of Soil 10
C. Preparation of Samples 11
D. Preparation of Individual Samples of Testing 13
E. Consolidation in the Triaxial Cell 17

F. Testing 18

Chapter III - Normally Consolidated Specimens with Uniform 21

Strain-Rate

A. Introduction 21

B. Presentation of Results - Normally Consolidated 21

Samples - Uniform Strain-Rate

C. Behavior of Normally Consolidated Samples at 27

Small Strains

D. Behavior of Normally Consolidated Samples at Large 33

Strains

Chapter IV - Normally Consolidated Specimens with Stepped 42

Strain Rate

A. Introduction 42
 CONTENTS (Continued)

Page No.

B. Testing Procedure 43

C. An Investigation of Stepped Strain Rate on Piston 45

Friction

D. Results of a Negative Step (Step Increase) in 46

Strain Rate

E. Results of Positive Steps (Step Increase) in Strain 49

Rate Imposed at 41" Strain in Normally Consolidated
Samples

F. Results of Tests in Which Equal Positive Steps in 53

Strain Rate were Imposed at Varying Strain Levels -
Normally Consolidated Samples

G. Relaxation Tests - Normally Consolidated Samples 56

Chapter V - Behavior of the Normally Consolidated Samples 60

A. Introduction 60

B. Summary - Effect of Strain-Rate on the Deviator 60

Stress - Normally Consolidated Samples

C. Summary - Effect of Strain Rate in the Measured Pore- 62

water Pressure - Normally Consolidated Samples

D. Effect of Strain Rate on the Obliquity Ratio - 63

Normally Consolidated Samples

E. Discussion - Effect of Strain Rate on the Behavior 65

of Normally Consolidated Samples at Small Strains

F. Discussion - Effect of Strain Rate on the Behavior 66

of Normally Consolidated Samples at Large Strains

Chapter VI - Overconsolidated Specimens with Uniform Strain 70

Rate

A. Introduction 70

B. Laboratory Test Program - Overconsolidated Samples - 70

Uniform Strain-Rate
 CONTENTS (Contined)

Page No.

C. Presentation of Results - Overconsolidated 72

Samples - Uniform Strain Rate

D. Effect of Rate of Strain on the Stress Versus 75

Strain Behavior of Overconsolidated Samples

E. Effect of Strain Rate on Maximum Deviator Stress, 76

Strain at Failure and Mode of Failure, Over-
consolidated Samples - Uniform Strain Rate

F. Relationship Between Rate of Strain and Pore 78

Water Pressures, Overconsolidated Samples,
Uniform Strain Rate

G. Pore Water Pressure Gradients and Internal Volume 80

Change Phenomena, Overconsolidated Samples

H. Effect of Rate of Strain on Obliquity (0/3) 82

Overconsolidated Samples

I. Summary and Conclusions: Overconsolidated Samples - 83

Uniform Strain Rate

Chapter VII - Overconsolidated Specimens with Stepped Strain 85

Rate

A. Introduction 85

B. Sample Preparation, Consolidation and Testing 85

C. Presentation of Data 85

D. Test OC5-FS with Step Decrease in Strain-Rate 86

E. Effect of a Step Decrease in Strain Rate Upon 86

Deviator Stress - Overconsolidated Samples

F. Effect of a Step Decrease in Strain Rate Upon the 87

Observed Pore Water Pressure, Overconsolidated
Samples

G. Effect of a Step Decrease in Strain Rate Upon the 88

Deduced Principal Effective Stress Ratio - Over-
consolidated Samples
 CONTENTS (Continued)

Page No.

H. Mode of Failure of Sample with a Step Decrease

in Strain Rate 89

I. Tests 0010-SF and OC11-SF with Step Increases in 89

Strain Rate

J. Effect of a Step Increase in Strain Rate Upon 90

Deviator Stress - Overconsolidated Samples

K. Effect of a Step Increase in Strain Rate Upon 90

Developed Pore Water Pressure - Overconsolidated
Samples

L. Effect of a Step Increase in Strain Rate Upon 91

the Effective Principal Stress Ratio (/ ) -
Overconsolidated Samples

M. Mode of Failure - Step Increase in Strain Rate - 92

Overconsolidated Samples

N. Summary - Effect of Steps in Strain Rate on Observed 92

Behavior - Overconsolidated Samples

Chapter VIII - Behavior of Overconsolidated Samples 94

A. Introduction 94

B. Stress Distribution in Triaxial Samples 94

C. Effect of Stress Distribution on Pore Water 95

Pressure Distribution

D. Evidences of the Existence of Pore Water Pressure 97

Gradients - Overconsolidated Samples

E. Migration of Pore Water as a Function of Strain Rate 97

F. Evidence of Internal Volume Change - Overconsolidated 98

G. Relationship Between Failure Plane Formation and 98

Pore Water Pressure

H. Effect of Pore Water Migration and Internal Volume 99

Changes on Observed Behavior of the Overconsolidated
Samples
 CONTENTS (Continued)

Page No.

I. Conclusions 102

Chapt er Ix 103

A. Introduction 103

B. Tests with Uniform Strain Rate 103

C. Effect of Strain Rate on the Hvorslev True Friction 1o4

and True Cohesion

D. Effect of a Step Increase in Strain Rate on the 106

Hvorslev Shear Strength Parameters

E. Conclusions 107

Chapter X 109

A. Objectives of Research 109

B. Summary of the Research Program 109

C. Significance of Research, 110

D. Pore Pressure Measuring System 111

E. Strain Rate Effect in the Normally Consolidated 111

Samples

F. Strain Rate Effect in the Overconsolidated Samples 112

G. Conclusions 113

Bibliography 114

Appendix A - Pore Pressure Measurement: Devices, Techniques

and Performance

Appendix B - Plots of Laboratory Data

 LIST OF FIGURES

1. Grain Size Distribution of Vicksburg Fat Clay

2. Plasticity Characteristics of Vicksburg Fat Clay
3. Consolidation Characteristics of Vicksburg Fat Clay
4. Maximum Deviator Stress Versus Strain Rate, Normally Consolidated
Samples
5. Summary Plot - Fast Tests - Normally Consolidated Samples
6. Summary Plot - Slow Tests - Normally Consolidated Samples
7. Average Deviator Stress and Midheight Pore Water Pressure
Versus Strain Behavior, Normally Consolidated Samples
8. Obliquity Ratio Versus Strain Behavior, Normally Consolidated Samples
9. Relationship Between Pore Pressure and Obliquity Ratio, Normally
Consolidated Samples
10. Stress Vector Paths, Normally Consolidated Samples
11. Maximum Obliquity Ratio Versus Strain Rate, Normally Consolidated
Samples
12. Pore Water Pressure Versus Strain Rate, Normally Consolidated Samples
13. Pore Pressure Parameter A Versus Strain Rate, Normally Consolidated
Samples
14. Pore Pressure Gradient Versus Strain Rate, Normally Consolidated Samples
15. Summary of Water Content Redistribution Data, Normally Consolidated
Samples
16. Summary, Behavior at Small Strains, Normally Consolidated Samples
17. Summary, Behavior at Small Strains (Continued)
18. Plot of Midheight Pore Pressure Versus Deviator Stress; Normally
Consolidated Samples
19. Plot of A Versus Strain, Normally Consolidated Samples
20. Summary Plot; Tests with Step Increase in Strain Rate at 4% Strain
21. Relationship of Post Step cr/C and (0- ) to Strain Rate
Following Step Increase in Strdin Rate, Nrmally Consolidated
Samples

22. Summary Plot; Tests with Step Increases in Strain Rate at Varying
Strain Levels, Normally Consolidated Samples
 LIST OF FIGURES (Continued)

23. Deviator Stress Behavior Following Step Increase in Strain Rate

from t = 500 min. to t+i = 3/4 minute; Normally Consolidated
Sample

24. Relaxation Behavior, Normally Consolidated Samples

25. 0- ) tj = 1 min. - (CF-C) t = 500 min]. vs. Strain

26. Percent Increase in (j-c) from tj = 500 min. to ti = 1 min.

Versus Strain

27. Pore Pressure Difference, Slow and Fast Tests Versus Strain

28. Idealized Behavior Models, Normally Consolidated Samples

29. Summary Plot - Fast Tests Overconsolidated Samples

30. Summary Plot - All Tests - Overconsolidated Samples

31. Maximum Deviator Stress Versus Strain Rate, Overconsolidated Samples

32. Strain at (cT- )m Versus Strain Rate, Overconsolidated Samples

33. Midplane Pore Pressure Versus Strain Rate - Overconsolidated Samples

34. Midheight Pore Pressure Versus Deviator Stress, Overconsolidated Samples

35. Pore Pressure Parameter A, Versus Strain , Overconsolidated Samples

36. A at (Gy-g )mx Versus Strain Rate, Overconsolidated Samples

37. Pore Pressure Gradient Versus Strain Rate - Overconsolidated Samples

38. Post Shear Distribution of Water Content, Overconsolidated Samples
39. Obliquity Versus Strain Rate - Summary - Overconsolidated Samples
40. Obliquity Versus Strain Rate - Overconsolidated Samples
41. Failure Envelopes
42. Behavior of Overconsolidated Samples in Terms of Stress Vectors
43. Summary Plot - Stress Vectors - Fast Tests - Overconsolidated Samples
44. Summary Plot - Stress Vectors - Slow Tests - Overconsolidated Samples
45. Summary Plot, All Stress Vectors - Overconsolidated Samples
 LIST OF FIGURES (Continued)

46. Plot for Determination of Hvorslev Parameters at Failure

47. Analysis of Step Increase in Strain Rate in Terms of Hvorslev
Parameters
 PREFACE

This document is the sixteenth in a series of reports issued

by M.I.T. under its present contract with the U. S. Army Engineers
Waterways Experiment Station. A list of the earlier reports follows
this preface. The work has been carried out in the Soil Laboratories
of the Civil Engineering Department. Dr. T. W. Lembe heads the soils
group. This work was performed by Archie M. Richardson, Jr., formerly
Instructor in the Civil Engineering Department and now Assistant
Professor at Swarthmore College, under the immediate supervision of
Dr. Robert V. Whitman, Associate Professor of Civil Engineering.
A documentation of the stress-strain behavior of the saturated
clay utilized in this research effort over a very wide range of strain
rates appears in the tenth report in this series. Report No. 10 also
includes a preliminary report based on a rough draft of the present
document.

i
 PREVIOUS REPORTS

1. "Scope of Test Program and Equipment Specifications," November 1957

2. "Test Equipment for High Speed Triaxial Tests," January 1959

3. "First Interim Report on Dynamic Soil Tests," October 1959

4. "One-Dimensional Compression and Wave Velocity Tests," August 1960

5. "Pore Pressure Measurements During Transient Laodings," November 1960

6. "Effects of Rate-of-Strain on Stress-Strain Behavior of Saturated

Soils," April 1961

7. "Adaptation and Use of the Boynton Device for Rapid One-Dimensional

Compression Tests," June 1961

8. "Laboratory Measurement of Dilatational Wave Propagation Velocity,"

June 1961

9. "Shearing Resistance of Sands During Loading," May 1962

10. "Strength of Saturated Fat Clay," June 1962

11. "Triaxial Tests Upon Saturated Fine Silty Sand," September 1962

12. "Static Tests Upon Thin Domes Buried in Sand," December 1962

13. "The Dependence of Dilatation in Sand on Rate of Shear Strain,"

February 1963

14. "Propagation Velocity of Ultrasonic Waves through Sand," March 1963

15. "Undrained Strength of Saturated Silty Clay," March 1963

11
 LIST OF SYMBOLS

= slope of failure envelope on plot of ( - 3)/2 vs. ( 1+ 3)/2

v = coefficient of compressibility
A = Skempton's pore pressure parameter
B = Skempton's pore pressure parameter
Cc = consolidating compressibility
C = swelling compressibility
c = coefficient of consolidation
V
c = Hvorslev's truc cohesion
d = cohesion intercept in terms of effective stresses
CU = consolidated undrained
Ec = modulus of elasticity - compressing
E = modulus of elasticity - swelling
e = void ratio
( ) = failure value of quantity in parenthesis
e = strain
= strain rate
A = pore pressure coefficient

P c = Poisson's ratio - compressing

P s = Poisson's ratio - swelling
n v = coefficient of volume compressibility
N.C. = normally consolidated
O.C. = overconsolidated
O.C.R. = overconsolidated ratio
p = Hvorslev's equivalent consolidation pressure
a = stress
= effective stress
t = time to 1% strain (usually minutes)
t+1 = time to 1% additional strain (usually minutes)
= angle of shearing resistance in terms of effective stresses
= Hvorslev's true friction angle

iii
 Chapter 1

INTRODUCTION AND BACKGROUND

A. Introduction
The following work presents and analyzes the results of an
extensive series of triaxial compression tests performed on remolded
samples of a saturated fat clay. These tests were performed to determine
the nature of the effect of strain rate ( ) upon the shear resistance
of this clay. Forty one consolidated undrained (CTU) tests with pore
pressure measurements were performed upon samples subjected to two extreme
preshear stress histories. Twenty nine tests were upon samples normally
consolidated (N.C.) from a deaired slurry. In these samples the preshear
effective stress was the largest stress the sample had carried since
preparation of the slurry. Twelve tests were upon samples having an over-
consolidation ratio of approximately 16 (O.C.R. = 16). In these samples
the preshear effective stress was on the order of 1/16 of the maximum
effective stress carried by the sample since preparation of the slurry.
These tests were performed at constant strain rates varying
from 1% strain in 45 seconds (t1 = 3 ), to 10% strain in 24 hours, (t =
1
1140). In all tests, pore water pressure was measured at the mid-height
of the test sample employing a porous stone probe and a very rigid pore
pressure measuring system of novel design. This system relied upon an
extremely low compliance electric pressure transducer for measurement
of induced pore water pressure. In addition, selected tests were
performed at representative strain rates with the pore water pressure
being measured at both the mid-height of the sample and at the base.
These tests were performed in order to determine the existence and magni-
tude of any pore pressure gradient between the mid-height of the sample
and the base. Water content determinations at six levels in each sample
were carried out and an attempt made to correlate the observed gradients
with the observed water content variation.
Using the information gathered in the manner outlined above, the
results are analyzed in terms of effective stress concepts (where effective
stress equals total stress minus pore water pressure; i.e., (a = tu).
The manner in which each of the effective stress components is affected
by the strain rate (e) is shown. Possible macroscopic mechanisms con-
tributing to the observed effect of strain rate upon shear resistance
are suggested. Results of additional tests in which the strain rate ( )
is varied during the course of each test are presented to support and
supplement the qualitative relationships deduced from this analysis.
The work is organized for presentation in ten chapters as
follows: The remainder of this first chapter gives the background ma-
terial necessary for a realization of the motivations behind this work
and for the attainment of perspective as to the standing of this work
historically. Chapter II describes the test program and test procedure.
In Chapter III are presented the results of tests upon normally con-
solidated samples with uniform strain rates while Chapter IV presents
test results obtained in tests on normally consolidated samples where
the strain rate was altered during the course of the test. In Chapter V
the behavior of the normally consolidated specimens is summarized. This
summary draws together and describes all the features of behavior which
must be satisfied by any mechanism proposed to explain this behavior.
An attempt is then made to describe this observed behavior in terms of
a macroscopic and microscopic structural behavior mechanism.
The results of tests upon overconsolidated specimens tested with
uniform strain rates are presented in Chapter VI and the results of tests
upon overconsolidated samples in which the rate of strain was changed
drastically during the course of the tests are presented in Chapter VII.
Again the significant features of these tests in terms of observed'
behavior that must be satisfied by any proposed mechanistic theory are
drawn together in Chapter VIII.
Chapter IX presents the results of all of the tests in terms
of the Hvorslev "true friction' and "true cohesion" components of shear
strength. (IVORSLEV, 1937). In so far as possible, the observed

-2-
relationship between shear strength and strain rate is explained in terms
of change in these "true" strength-effective stress parameters. Finally,
Chapter X snmiarizes the entire work and points up areas of weakness as
avenues for further research.
In addition, there are two appendices to the work. Appendix A
describes in detail the pore pressure measuring system employed in this
research and presents data indicating the performance of this system.
In addition it describes the author's experience gained in the develop-
ment of this system. Appendix B presents and documents the results from
each individual test.

B* Background
The understanding of the magnitude, nature and mechanism of the
effect of rate of strain on the stress vs. strain behavior in soils is of
far more than academic interest. Any attempt to utilize the results of
laboratory tests in an analysis of field stability requires that the
effects of the differences in rates of strain between the laboratory
and the field be either slight or predictable. In general, when used
in situations where these parameters are appropriate, the usual effec-
tive stress parameters and c or the undrained shear strength as
determined by customary laboratory tests or by field vane tests, where
such laboratory or field tests are carried out within the range of
convenient strain rates, give results that have been shown to be within
the limits of engineering acceptability. However, certain noteworthy
exceptions do occur.
At present, no analysis of stability can be made with any
certainty in stiff, fissured clays. The clay shales of the prairie
regions of the U. S. and Canada (for example the Bearpaw Shale) and the
London Clay are two well known examples. The best that can be said
at this time is that (the intercept on the y axis of the plot of
shearing resistance on the failure plane at failure vs. the effective stress
on the failure plane at failure) seems to decrease with time. Thus, one of

-3-
the present day problem areas in soil mechanics is almost certainly a
strain rate or at least a time problem. The other notable situation in
which the use of laboratory parameters of shear strength behavior gives
dubious results is found with some of the highly sensitive clays where
* in the field is much less than laboratory tests indicate. This may
or may not be partly or entirely a strain rate effect. A discussion of
these situations is found in (BISHOP AND BJERRUM, 1960).
Present day concern for protective construction to resist blast
effects and associated impact loading problems in foundation engineering
gives the search for an understanding of the nature of strain rate effects
much of its current urgency. In addition, the analysis of the stability
of building foundations and slopes against stress imposed by earthquakes
is currently receiving much attention. Highway and airfield pavement
design and trafficability studies in general are additional areas in
which the need for strain rate effect information is felt keenly. How-
ever, of much more importance is that present day parameters of stress
vs. strain behavior in soils represent descriptions of observed laboratory
behavior and are probably not fundamental. To begin to really understand
the nature of the resistance of soils to applied stresses and the reason
for the resulting strain patterns, regardless of the rate of stress or
strain application, should be an underlying motive of soil mechanics
research and it is hoped that the present work contributes, no matter
how slightly, and thus lives up to this philosophy.
Previous research has been performed in several laboratories
by numerous investigators. Most of this work has been concerned with
observation of the increase of shear strength and the alteration of.
the shape of the shearing resistance versus strain curve with increased
strain rates and a good deal of the available data concerns unsaturated
(compacted) soils. (See for instance M.I.T. 1959). Tests on saturated
sands with rapid load applications in which attempts were made to
measure pore water pressure were begun at the Massachusetts Institute
of Technology in the early part of the 1950's (M.I.T. 1954). Tests of

-41-
this nature are being performed on a much refined basis at M.I.T. today
and a demonstration of the nature of the strain rate effect in sands has
been made. (IITmAN & HEALY, 1961, HEALY 1962).
Significant work with saturated clay shales and soft rock was
undertaken at Harvard in the latter 1940's and early 1950's (CASAGRANDE
and SHANNON, 1948) and (CASAGRANDE and WILSON, 1951). All of this early
work can be found very nicely summarized in (IUTMAN, 195Ta) and
(wITMAN 1957b).
Most of this research, however, has been in terms of total stress
application with pore pressure measurements being,attempted only in a few
isolated instances and then only in saturated sands. Therefore, little
light has been shed on the fundamental mechanisms contributing to the
observed and documented relationship of shearing resistance to rate of
strain. However, two notable studies of the relationship of effective
stress parameters of shear strength behavior to rate of strain have
appeared quite recently and are summarized below:

1. (cRAwFORD, 1959)
This research consisted of triaxial compression tests with base
pore water pressure measurements performed upon samples of undisturbed
Leda clay cut from a single block. The Leda clay is a widely distributed
clay found along the Ottawa and St. Lawrence Rivers in the U. S. and
Canada. It is usually highly sensitive and exhibits compression and
shear strength properties indicative of a slight to moderate degree of
overconsolidation.
The soil tested was rather plastic (liquid limit = 65%,
plasticity index = 40%) and had a natural water content slightly above
the liquid limit. Tests were performed with preshear consolidation
pressures ranging from 2.0 kg./cm.2 to 6.O kg./cm. where the adjudged
2
preconsolidation load was 2.75 kg./cm. , .The rates of strain considered
ranged from an'actual time to failure of 7.8 minutes (t1 = 6 minutes) to
an actual time to failure of 1320 minutes (t1 = 1300 minutes).

-5-
 The results of this research can be summarized as follows:
i. The deviator stress at failure increases with increasing strain
rate. The absolute magnitude of this strength increase seems independent
of the consolidation pressure, and therefore represents a greater percent-
age increase at lower consolidation pressures.

ii. The measured pore water pressure at failure (uf) was higher in the
tests with the lower strain rates, (i.e. as the strain rate increased the
pore water pressure at failure decreased) in the range of strain rates
considered. This decrease was slight in the slightly overconsolidated
samples (co- cons = 2.0 kg./cm.2) and of considerable magnitude in the
0 -cons = 6.0 kg./cm. .
samples with

iii. Looked at in terms of the pore water pressure parameter A, where

A = u (SKEMPTON, 1954), in all cases A was less in
3
the tests with rapid strain rates (i.e., as
de
d- increased, Af decreased).

iv. The magnitude of this decrease in A was sufficient to more than

offset the noticed decrease in < (where A and 4 were measured at
maximum ( - a3) ). The value of > was found to be smaller at the
higher strain rates than at the lower strain rates, the range being
0 10
from ( = 23 in the tests at the slow strain rate to <{ =1{ in
the fast test. This was accompanied by a slight increase in the cohesion
intercept (c) from the slow tests to the fast tests.

v. The strain rate effect in the slightly overconsolidated samples

was attributed to a loss of the effects of preconsolidation with decreas-
ing rates of strain.

vi. Evidences of gradients in the pore water pressure within the sample
at the faster strain rates was noted in the post shear water content
distribution. These data seem to indicate higher pore water pressures
in the center of the sample than at the base.

-6-
 vii. In creep type tests in which a constant deviator stress is applied
and allowed to remain, the pore water pressure was seen to be a linear
function of strain.
2. (BJERRUM, SIMOKS and TOBIAA 1960)
This research consisted of 36 consolidated undrained tests, with
pore water pressure measured in all but the most rapid tests, and 21
consolidated drained tests. The variation of the rate of strain in the
undrained tests was from 100%per hour (45 seconds to 1%) to 0.0036%
per hour (300 hours to 1%). The range of strain rates in the drained
tests was from 1.3%per hour to .033% per hour.
The clay tested was an undisturbed, normally consolidated marine
clay from Fornebu, Oslo, Norway, both less plastic and less sensitive
2
than the Leda clay. Preshear consolidation pressures of 1.0 kg./cm.
2.0 kg./cm.2 and 4.0 kg./cm.2 were used for both the undrained and the
drained tests.
Following is a summary of the principal results obtained from
this research:

i. In the undrained tests, the deviator stress at failure in-

creased as the time to failure decreased. The absolute increase
increased with increasing preshear consolidation pressure and thus
differed slightly from the Leda clay.

ii. In the undrained tests, the excess pore water pressure at

failure increased as the time to failure increased. It should be
mentioned here that in undrained tests of as long duration as the slow
tests described in this work, membrane leakage can significantly in-
fluence the measured pore water pressure.

iii. In these tests, the pore water pressure parameter A increased

as the time to failure increased, but not sufficiently to account for
the loss of strength.

iv. The additional loss of strength was accounted for by an observed

decrease in the angle of shearing resistance in terms of effective
stresses, , of from 309 to 25.50 as the time to failure increased.
This occurred with no change in the cohesion intercept c.
 v. The axial strain necessary to produce failure increased as the
time to failure decreased in the undrained tests and to a lesser degree
in the drained tests.

vi. The maximum deviator stress in the drained tests was constant
regardless of the strain rate. It was postulated that the decrease in
effective stress parameters was offset by a decrease in water content
due to secondary consolidation type phenomena.

In comparing and contrasting the research just described, the

following summary statements may be made. As the work at the Norwegian
Geotechnical Institute was confined to normally consolidated samples,
this sunmary pertains only to such samples. In both investigations the
maximum deviator stress increased as the strain rate was increased (i.e.,
a higher strength was attained in the more rapid tests). In both
investigations the pore water pressure at maximum deviator stress and the
pore pressure parameter A increased in tests with slower strain rates.
In the Canadian work the strain necessary to achieve maximum
deviator stress was more or less independent of the strain rate. In
contrast, the Norwegian investigation showed larger strains at failure
in the faster tests. More importantly, the Norwegian work showed larger
values in the faster tests, while the Canadian work produced exactly
opposite results (i.e., j larger in the slower tests). The Canadian
work shows ' (the '~C axis intercept in a plot of shear stress on the
failure plane at failure versus the effective normal stress on the
failure plane at failure) increasing with increasing strain rate, while
the Norwegian work suggests that E remains unchanged.
In addition, two very recent publications of the Soil Engineering
Division at the Mssachusetts Institute of Technology present research
that paved the way for the work to be described in the following chapters.
In (M.I.T. 1961) the results of rapid tests on a saturated loess and of
a very preliminary series of tests on the same clay used in the research
to be described, are presented, The assumption is made that the Hvorslev

-8-
parameters of shear strength behavior $eand c are independent of
time to failure and the pore water pressures necessary at failure
to produce the observed strain rate effect are deduced. In light
of the present work this assumption is somewhat unjustified, but
nevertheless the analyses shed light on the probable patterns of
pore water pressure behavior.
(NASIM 1961) presents the results of an extensive series
of triaxial compression tests with no pore water pressure measure-
ments. These tests were performed on samples of the same clay
described in the following work and prepared in a similar manner.
The rates of strain employed overlapped the present work and ex-
tended to ouite rapid rates. Thus the results of that research
fully document the nature and magnitude of the effect of rate of
strain on the total stress vs. strain behavior of this soil.
 Chapter II

DESCRIPTION OF TEST PROGRAM AND TEST PROCEDRE

A. Introduction
This chapter describes the test program undertaken in the
course of this work and the procedure employed in its performance. This
test program consisted of a total of 43 triaxial compression tests on a
single very fat clay with a permeability on the order of 10 9 cm./sec.
Prior to shearing, the samples were subjected to two stress histories.
Ten samples were overconsolidated to an overconsolidation ratio of
approximately 8 atmospheres gauge and subsequently allowing rebound to
approximately 1/2 atmosphere. The exact pressures used tend to differ
from test to test due to the calibration difference of the several gauges
used. These pressures were corrected by comparison with one accurately
calibrated gauge. The remnder of the samples were normally consolidated
to approximately 4 atmospheres gauge.
In general the tests were performed at rates of strain which can
be grouped in two categories, fast and slow. The majority of the slow
tests were performed at strain rates of approximately 1% in ten hours.
(t1 = 600) This strain rate is about half that which would be considered
"normal" for triaxial testing with pore pressure measurements on this clay.
The fast tests were performed at rates of strain on the order of 1% in
one minute, (t 1 = 1) which is, of course, considerably more rapid than
the strain rate which would be considered "normal". In addition, scattered
tests performed at strain rates faster, slower, and intermediate were
performed in an attempt to gain a complete picture.
Most of the tests were performed at a constant rate of applica-
tion of deformation to the sample-proving ring system. This resulted, for
all practical purposes, in a uniform rate of strain since the proving ring
chosen was of sufficient stiffness to contribute little to the deformation.
In addition, in several tests, the strain rate was changed abruptly during
the course of the test. In all tests, pore water pressure was measured
at the mid-height of the sample by a rather unique pore water pressure

-9-
measuring system. In selected tests, pore water pressure was measured
at the base of the test sample in addition to the mid-height.
The principal results of all of the tests, the preshear
history of each sample and other pertinent data are presented in
Appendix B.

B. Description of Soil
The soil used for these tests was considered to be representa-
tive of a very fat clay. The soil itself was obtained from the Water-
ways Experiment Station of the U. S. Army Corps of Engineers, Vicksburg,
Mississippi. Several shipments of this soil have been sent to the Soil
Engineering Division at M.I.T. in the past several years. The pre-
liminary series of 11 tests were performed upon assorted samples of
this soil remaining from previous research efforts, primarily (M.I.T. 1959)
and (NASIM, 1961). The soil used in the main series of tests was set
aside from a shipment of the soil received in the summer of 1960, being
"Barrel D' described in (M.I.T., 1959).
The soil used in the main series was shipped to M.I.T. in a
55 gallon drum. It had been air dried and coarsely pulverized. The
sample was obtained from a typical backswamp deposit of the lower
Mississippi valley. The geological history of such a deposit is
described in (KOLB and SHOCKLEY, 1957). The soil is a fat clay,
impermeable and quite plastic. A plot of grain size distribution
of the soil used in the main series of tests is given in Figure 1.
The plasticity characteristics of the soil used for the main series of
tests are plotted in a Casagrande type plasticity chart in Figure 2.
In addition, on this chart, are plotted the plasticity characteristics
of several other shipments of this soil and references are given to
the sources. Table II-1 contains a complete mineralogical analysis of
a sample of the minus T4 micron fraction of the soil from a previous
shipment, but there is no reason to suspect that this fraction of the
soil differs in mineralogical composition from shipment to shipment.
In addition, engineering properties appropriate to the soil utilized

-10-
in this research appear in this table.

C. Preparation of Samples
The preliminary series of tests, P-IS through P-9F, were
performed upon samples remaining from the research described in
M.I.T. (1959) and NASIM (1961). As little other than preliminary
feelings and experience in setting up tests was gained from these
tests, little concern is expressed regarding the exact source of
each sample.
As the successful use of the device employed to measure the
pore water pressure at the mid-height of the sample demands a completely
saturated soil sample, and as some variation from absolute saturation
was encountered with the preliminary samples, a revised method for
preparation of samples consolidated from a slurry was deemed necessary.
A suiall batch, Batch A, was prepared from a quantity of the soil in the
drum to be used for this research.

Table IVI-1

DESCRIPTION OF THE CIAY

A. Mineralogic Composition ,of Minus 74iA Fraction BY Weight

1. Illite 25% + 3
2. Montmorillonite 25% + 3
3. Quartz 26% + 3
4. Feldspar 20%+ 10
5. Fe2 0 1.9% + 0.1
6. Orgaric Matter 1.1% ~ 0.1

B9 Engineering Properties
1. # normally consolidated + 230 + 20
2. Permeability K > 10 cm/sec.
3. Consolidation characteristics 2
Cc = 0.52 + 0.04, c, = 2 x 10- Cm/sec

-11-
 (a) Preparation of Batch A
This soil was air dried and reduced to a fine powder by
mechanical grinding. A quantity of deaired, demineralized water was set
boiling in an enameled pail. The powdered soil was sprinkled in a very
fine layer on the surface of this water, and allowed to saturate itself
from the bottom and sink into the pail. The mixture was stirred with
minimum disturbance with a small paddle, and the sequence of events
repeated. A very small flame was maintained on the bottom of the pail
as this operation proceeded, to maintain the.mixture at boiling. When
a slurry of the proper consistency had been obtained, it was transferred
to a vacuum dessicator in the following manner;
A 1-inch layer of boiling, deaired water was placed in the
dessicator.. A scoop of the slurry was carefully obtained from the pail
and deposited beneath this layer of water. Extreme care was taken to
introduce no air bubbles, any small surface air bubbles being removed
with a spatula. The slurry in the pail was gently and continually
stirred. When the slurry had been entirely transferred to the vacuum
dessicator, the lid was applied and the chamber over the surface of the
slurry evacuated and kept evacuated for 48 hours. It was felt that in
this manner any small air bubble in the body of the slurry would go
into solution as the slurry cooled, but that the slurry would still
not be in equilibrium with dissolved air due to the maintenance of a
vacuum over its surface.
A 1-inch layer of deaired water was then placed in the bottom
of a standard 4 inch diameter consolidometer which had been modified
by the addition of a lucite sleeve to extend its height to 8 inches.
The stone at the base of this consolidometer had previously been
deaired by boiling and the bottom drainage system had been saturated.
The slurry was then transferred to this consolidometer in exactly the
same manner as it had been placed in the vacuum dessicator, i.e., by
careful deposition beneath the surface of the layer of deaired water.
When the consolidometet had been filled, the actual consolidation

-12-
load was applied in the usual manner on a Fairbanks-Morse Scale, in
increments starting with 1/16 of a ton per sq. ft., and doubling
until 1 ton per sq. ft. was attained. Each increment was permitted
to remain three days, except the final one, which was maintained for
one month. The cake of soil thus obtained was ejected from the
consolidometer and stored in Mobil AB Transformer oil, an extremely
inert oil, until wanted for testing. Two samples only were obtained
from this Batch A.

(b) Preparation of Batch B and Batch C

As the general method of batch preparation was deemed
satisfactory, a quantity of the soil sufficient for the main series
of tests was taken from the shipment, air dried and finely ground.
An identical procedure for the preparation of a slurry was followed.
This slurry was placed in a large 9J in. I.D. x 7 in. high Teflon-
lined consolidometer (described fully in M.I.T. 1961). The consolida-
tion pressure was applied to the top piston through transformer oil
and an air-oil accumulator. The load was applied in increments as with
Batch A. At the completion of consolidation, the large "cake" was
ejected and stored in transformer oil until needed. Two batches,
Batch B and Batch C were prepared in this manner. Approximately
20 samples could be cut from each large "cake". Although every
precaution was taken to insure that the two batches, B and C, be
identical, the water content from trimmings of Batch B samples
averaged approximately 41%, while Batch C samples averaged about 40%.
However, no immediately discernible difference in the water contents
of the samples from Batches B and C was evident following testing.

D. Preparation of Individual Samples for Testing

Individual samples were cut from the large cakes of soil
stored in the transformer oil as they were needed for testing. Samples
were trimmed on the Norwegian Geotechnical Institute trimming frame.

-13-
This frame produces samples having a height of 8.0 cm. (3.15 in.) and
a cross-sectional area of 10.0 sq. cm. (1.53 sq. in.).
Samples P-lF, P-2P, P-58 And P-6S of the preliminary series
were set up as follows:
The base of the triaxial cell to be used was thoroughly
deaired. It was then placed in a basin of deaired water. The level of
the water in the basin was at least 2 inches above the top of the stone
on the pedestal of the cell. One latex membrane was fastened to the
base of the cell with a rubber "O" ring and the trimmed sample set
upon the stone on the pedestal. The membrane was rolled up the sample,
bringing a ring of deaired water up along with it; rolled up on the
lucite top cap and secured with "O" rings. The top cap used in these
tests was provided with a deaired stone and saturated drainage line
leading out of the cell, so that double drainage was provided.
The membrane was grasped with the tips of a pair of longnosed
pliers at mid-height and snipped in front of the pliers with a sharp pair
of scissors making a small, circular hole in the membrane, and the de-
aired pore water pressure probe pushed horizontally into the sample
through this hole. This operation was conducted beneath the surface
of the water. The water level in the basin was then lowered beneath
the level of the needle, and the opening cleansed with alcohol and
sealed with rubber cement. Approximately 6 hours were required to
permit drying of the rubber cement. The drainage lines were opened
all this time, and allowed to remain open, as was the line leading
from the probe. In earlier experimenting it was observed that cavitation
of the water occurred in the tube leading to the probe if it were to be
closed off.
The chamber of the cell was then replaced and filled with
deaired water. The top half inch of the chamber was supplied with a
layer of oil, the piston inserted, and the cell attached to the source
of chamber pressure. Following application of the chamber pressure the
tubing leading from the pore water pressure probe was blocked off.

-14-
 The chamber of the cell was then replaced and filled with
deaired water. The top half inch of the chamber was supplied with a
layer of oil, the piston inserted, and the cell attached to the source
of chamber pressure. Following application of the chamber pressure the
tubing leading from the pore water pressure probe was blocked off. The
progress of consolidation was checked from time to time by measuring the
mid-height pore water pressure.
Two objections to this method rapidly became apparent:
1. Since no filter strips were used, an excessive time was
required for consolidation of the samples. This time was in excess of
45 days.
2. As the pore water pressure probe was inserted, the sample
was seen to develop a split. This split was not horizontal, but was
usually quite steep, oriented at more than 60 0 to the horizontal. In
attempt to avoid this splitting, a piece of stainless steel tubing
slightly smaller in diameter than the probe was sharpened at one end,
and a hole formed in the sample beneath the surface of the water before
the probe was inserted. In all cases a nice core was removed with this
sampling tube, but the sample split very soon after the core was with-
drawn, and long before the probe was inserted.
In additional experimenting, it was found that a hole could
be formed in the sample with the sample out of the water and with the
stainless steel borer completely dry. If one drop of water were then
to be placed in the hole thus formed, the sample split quite rapidly.
This splitting represents a tension failure of the sample due to forces
associated with release of the pore water tension adjacent to the hole
and consequent central swelling of the sample.
One additional objection to this method of setting up the
samples is the fact that only one membrane can be provided for, and
the problem of membrane leakage becomes overwhelming.

-15-
 I

To avoid the problems associated with the previously described

procedure, the following system was devised and employed in setting up
all of the main series of tests with only very minor exceptions as noted.
The base of the triaxial cell was modified by drilling a 0.03"
diameter hole vertically through the very edge of the pedestal as de-
scribed in Appendix A. The small diameter Teflon tubing from the pore
water pressure probe was passed out of the triaxial chamber through this
hole. The space between the outside of the tubing and the walls of the
hole was sealed with an Epoxy resin. This base, stone, tubing and pore
pressure probe were deaired by boiling, flushing boiling water through
all tubing and drainage connections and allowing the base, in boiling
water, to cool beneath a vacuum in a vacuum dessicator. With the cell
base beneath water, two latex membranes were affixed to the base with
"O" rings.
A hole slightly smaller in diameter and slightly shorter than
the pore water pressure probe was formed at mid-height of the soil sample.
This operation was performed using the stainless steel boring tool
described previously, fixed in the chuck of a drill press to insure
vertical alignment. The trimmed sample in the NGI trimming cradle was
placed on the bench of the drill press, the hole located and formed by
vertical motion. No rotation of the boring tool was permitted. The
sample was then carefully and squarely sliced in half horizontally
through the axis of this hole with a fine wire saw.
The bottom half of the sample was placed on the stone of the
pedestal of the triaxial cell base, submerged as before in a basin of
deaired water to a depth sufficient to cover by one-half inch the top
of the bottom half of specimen. The porous stone probe was then placed
carefully in the trough representing the bottom half of the bored hole
and the top half of the sample fitted in place with the utmost care.
Gentle pressure axially was then applied with the specimen cap to-
insure seating of the two surfaces.

-16-
 A band of latex membrane 3/8 inch in width was placed over the
cut, thus forming a dam insulating the region of the probe from the four
filter strips which were then applied to the sides of the sample. One
latex membrane was then rolled up and the water level lowered to the
level of the pedestal of the cell. It is to be noted that this arrange-
ment allows the Teflon tubing to the pore water pressure probe to spiral
around the sample under the rubber membrane. A thick layer of silicone
grease was then applied to the sample and the second membrane rolled up
and both membranes sealed to the top cap with two "0" Rings. The cell
chamber was fastened in place and filled as before.
The exceptions to this procedure are as follows:
1. Samples P-8s, P3F, P4F, P95 had only one membrane.
2. In samples Oc6s, OC1F, 002F, OCTS and OC8S, outer membrane
was also rolled up under water. This was seen to be
equivalent to having only one membrane as a layer of water
was trapped between the two membranes and remained there
during testing.

E. Consolidation in the Triaxial Cell

Following setting up the samples, the triaxial cells were
connected to a source of pressure and consolidated. In the normally
consolidated samples, this source of pressure consisted of Bishop type
self-compensating mercury pots, (BISHOP & HENKEL, 195T). Since the
available mercury pressure system was limited to 4 kg/cm2, it was
necessary in the case of the overconsolidated samples to apply the
pressure through water to a water-nitrogen gas accumulator. The
pressure in the nitrogen gas was controlled with a regulator. This
later source of pressure was somewhat less than satisfactory. Two
samples had to be discarded due to the escape of nitrogen gas bubbles
from solution following diffusion through either the membrane or tube
leading from the porous probe.

-17-
 In the normally consolidated samples, a preliminary pressure of
30 lb./in.2 was maintained for approximately one hour before the entire
2
consolidating pressure of approximately 60 lb./in. was applied. Drainage
from the tube leading to the probe was closed off immediately following
application of the first pressure to the chamber. Consolidation was
permitted to proceed until no more than 0.1 cc. of water per day was
being expelled from the sample. This occurred in seven days.
The same procedure was followed with the overconsolidated
2
samples, up through the application of the 60 lb./in. pressure. This
pressure was maintained for three days and then increased to approximately
120 lb./in. . This latter pressure was held for about a week, or until
less than 0.1 cc./day was being expelled by the sample. The tubing
leading to the probe was then opened and placed in a container of
deaired water. This was to prevent cavitation of the water in the
2
tubing. The chamber pressure was gradually reduced to about 7.0 lb./in. ,
allowing several hours in most cases for this reduction. This rebound
pressure was maintained for about a week or until the criteria estab-
lished above was met.

F. Testing
Following consolidation, the samples were sheared. Connection
was made to the pore water pressure measuring transducer, outside the
cell, as described in Appendix A, and electrical connections made to the
recording equipment. The triaxial cells were placed in a Wykeham-
Farrance load frame and a proving ring seated atop the piston. Two
rather stiff proving rings were calibrated and set aside for this
research. They were of a stiffness to strike a happy medium between
accuracy and nonuniformity of strain rate due to compression of the
proving ring. These rings were calibrated one against the other.
Following attachment of the pore water pressure measuring
devices as described in Appendix A, check was made of the response
time of the system. This was accomplished in the case of normally

-18-
consolidated samples by increasing the chamber pressure 10.0 lb./in.2
recording the time to 90%, 95%, and approximately 100% response, and
then reducing the chamber pressure to its original value. In all
samples in the main series, essentially 100% response was obtained
in 60 seconds or less. This is an excellent response for a pore
pressure system when used in a clay with as low a permeability as
this. It represents performance superior to that recorded by
(TAYLOR, 1954) in which 100% response was attained in 10 seconds in
Boston Blue Clay.
In the overconsolidated samples, the same procedure was
followed, but more than 10.0 lb./in.2 was applied and this pressure
was allowed to remain as a back pressure. If the response was less
than 100% in 60 seconds, another 10 lb./in.2 increment was added,
and this repeated until the desired response was attained. However,
it was a rare occurrence indeed when the response to the first
increment was unsatisfactory.
The question of whether the 60 second response represents,
in part, a compression of the Teflon tubing by the chamber pressure
is answered by the results of a test, appearing in Appendix A, in
which increments of deviator stress were applied and the resulting
response in pore water pressure measured. Here the response to a
deviator stress application is seen to occur fully in about the
same time as a response to a chamber stress application. The Teflon
tubing is quite rigid, being in effect a thick walled cylinder, and
the effect of compression of the tubing is seen to be quite small.
Compression at a constant rate was carried out until a
distinct failure plane was observed, or until reasonable equilibrium
of all measurements had occurred. The normally consolidated samples
were usually carried to 12% strain. Following testing, the samples
were dismantled and water content samples taken from six levels in
the sample, with the very top and the very bottom being discarded.

-19-
Approximately 5 minutes were required for this dismantling and water
content determination.
The usual question of the effects of piston friction crop
up at this time. Extreme precautions were taken to have a properly
fitting piston at all times. The pistons were cleaned thoroughly
after each use, lubricated and checked to insure that they fell under
their own weight when fitted in their bushing. Occasionally they were
polished with "Noxon" as the program progressed. No notice of any
reasons for concern appeared either as the tests progressed or as the
data were plotted. Cells both of Norwegian Geotechnical Institute
and of Clockhouse Ltd. manufacture were used, with no preference
noted. However, it should be kept in mind that piston friction is a
factor in all triaxial testing with external load measurement.

-20-
 Chapter III

NORMALLY CONSOLIDATED SPECIMENS WITH UNIFORM STRAIN-RATE

A. Introduction
The following chapter presents and discusses the results of a
series of triaxial compression tests upon normally consolidated samples
of the Vicksburg clay. These tests were performed at uniform strain
rates representing times to 1% strain varying from 45 seconds to
24 hours (t, = $ to t, = 1440). In some of these tests, the strain
rate was "stepped" (i.e., the test was begun at one strain rate, and
the strain rate was abruptly shifted at some predetermined strain to
one either faster or slower). The data regarding the behavior of
these samples subsequent to this "step" in strain rate are presented
in Chapter IV. However, data from the earlier portions of these tests,
prior to the step in strain rate, are included with the data from the
tests run to completion at a constant strain rate, and are included in
the summaries appearing in this chapter.

B, Presentation of Results - Normally Consolidated Samples - Uniform

Strain-Rate
This section presents in summary fashion, the results of all
of the tests performed upon the normally consolidated samples. Complete
plots of stress versus strain, pore water pressure versus strain,
obliquity versus strain and stress vector plots are presented in
Appendix B.
The results of nine of the eleven preliminary tests are
included in Appendix B in Figure B-1. The results of tests Pl-F,
P2-F, P5-S, P6-s, and P7-S are seen to be quite erratic. This is
felt to be due primarily to two factors: splitting of the sample
at the time of insertion of the pore water pressure probe and incomplete

-21-
consolidation of the sample. The results of tests P8-S, P3-F and
P4-F are felt to be reasonably reliable, and quite interesting in
themselves as will be discussed later. However, some doubt can be
cast on even these tests in the preliminary series as in all but
P3-F there is evidence of irregularities in the test. Test p8-s
had to be stopped in the middle to correct trouble with the measuring
system, and P4-F shows evidences of membrane leakage. In test P9-S
an attempt was made to use a needle or probe type measuring system
at a level 1/8 inch above the base in addition to mid-height. There
was excessive lag in both measuring systems and the needle at mid-
height appeared to be non-functioning. In summary, the results of
these preliminary tests are generally not reliable. However, tests
P3-F, P4-F and P8-S do point out interesting trends.
The void ratio (or water content) data following triaxial
consolidation are presented in Figure 3. This data was obtained from
three research efforts: that described in this thesis, M.I.T. (1959)
and NASIM (1961). The majority of the data gives parallel
virgin and rebound lines. The difference in placing of these lines
is felt to be due to differences in water content of the slurry from
which they were consolidated, and, in addition, the extreme dif-
ference recorded by the data from M.I.T. (1959) may include the
effect of non-saturation. If the slightly flatter normal consolida-
tion curve for Batches B and C is real, it no doubt represents the
effect of heating the slurry. Scatter was noted in the data from
all sources, and the range of scatter in the data from Batches B and
C is indicated. In the performance of the research described in this
work, several gages were employed for the measurement of consolidation
pressure. When the gages were subsequently calibrated, slight devia-
tions from the anticipated consolidation pressure were seen to have
occurred. However, the scatter of water content data within the
small range of this pressure deviation is completely random, as is

-22-
the scatter in the maximum deviator stress for any nominal consolidation
pressure. However, marked differences in undrained shear strength at a
given consolidation pressure occurred from one research effort to another,
I.e., those appearing in this work, in M.I.T. (1959) and in NASIM (1961).
These marked deviations can be directly attributed to water content
variations for a given consolidation pressure from one research effort
to another.
Figure 4 presents the maximum deviator stresses obtained from
all tests on the normally consolidated samples ploted against the logarithm
of time to 1% strain.E(- - u3 ) max. vs. log tl3/. As tests EC9-SF,
NC10-SF, NCll-SF and C18-SF involved a "stepped" strain rate, and as
this "step" occurred in most cases prior to the attainment of maximum
deviator stress, the values of maximum deviator stress from these tests
were obtained by an extrapolation of the stress vs. strain curves from
these tests. In addition, data from NASIM (1961) are also shown on this
plot and the slope of the deviator stress versus logarithm of time to 1%
strain is seen to be identical with that of data from the current work.
The results of all of the rapid tests on normally consolidated
samples are plotted in Figure 5. The deviation of test NC2-F from the
average stress versus strain curve drawn in Figure 5 is felt to be due
to the fact that this sample was allowed to consolidate considerably
longer than the others and the effects of aging may be evident. Test
NC18-SFLC was performed with a modified triaxial cell having a load
measuring cell in the base. There is some question about the calibra-
tion of the load cell, and the drainage system permitting consolidation
of the sample was known to leak. Tests NC3-F, NC4-F and E5-FS will be
seen to give almost identical data and have been relied upon heavily
when drawing the average stress versus strain plot. Below the stress
versus strain plot in Figure 5 is a pore water pressure versus strain
plot. The same comments applying to the stress strain curve apply here
with tests EC3-F, NC4-F and NC5-FS yielding quite consistent data.

-23-
 The results of all the slow normally consolidated tests that

had been consolidated to near 60 lb./in. are plotted in Figure 6. This

plot represents the results of eleven tests. Unfortunately, the curves
at large strains must rely heavily on the results of NCT-S. The curves
as given (Fignres, 5 and 6), however, nox doubt represent good typical
curves for the 'apid arid slow strain rate, with NC2-F representing the
only major deviation. (see Figure 5)
These average curves from Figures 5 and 6 have been reproduced
in Figure 7 with the curves of stress and pore water pressure versus
strain obtained from Test NO6-s superimposed. This test is assumed to
be representative of a test performed at an intermediate strain rate.
The average time to 1% strain as a measure of strain rate is marked on
each curve.
The information presented in these average or typical curves
was used to deduce the plot shown in Figure 8. This figure is a plot of
the principal stress ratio ( f )versus strain, with the times shown
on the curve being the time to 1% strain as a measure of strain rate.
This plot, in conjunction with Figure 7, presents the two components
of resistance to shear stress and shows how they develop with strain.
The comparison of the development of these two components is shown
in Figure 9, which is a plot of versus strain and thus
)-3
shows graphically the manner in wch pore water pressure is developed
in relationship to obliquity of the principal stresses as the sample
is strained. In such a plot, a horizontal line indicates that both
the pore water pressure (u) and the obliquity of principal effective
stresses ( /-3 ) are developing at the same rate. This plot can
be significant in that, if Au& is taken as a measure of structural
collapse with "/c0 being a measure of the mobilization of
structural resistance to displacement, then (c, ~31 - I 3
is an indication of the comparative rates of structural breakdown vs.
development of structural resistance.

-24-
 If the sign of f() is (-), structure is being
mobilized at a faster rate than it is being broken down. If d (E)
is 0, structure is being mobilized and destroyed at the same rate, while
if g is (+) structure is being broken down faster than bonds
can be formed. The implications of this plot will be discussed in a
later paragraph in this chapter.
Another familiar and convenient manner of looking at the
development of resistance to shear displacement and pore water pressure
3r7- -
development is presented in Figure 10. This is a plot of

as a measure of resistance to shearing and the ratio of

as a measure of effective stress normal to the failure plane development.
Such a plot is usually termed a "stress-vector" plot. If a line be
drawn on this plot starting at the chamber pressure on the x axis and
extending up and to the right at an angle of 450, such a line represents
the path of total stresses during the test. The distance from a point
on the effective stress path horizontally to an intersection on this
450 line represents the pore water pressure. Hence such a plot shows
the interrelationship of the resistance to shear, the effective stress
normal to the potential failure plane and the pore water pressure as a
test progresses.
In addition, plots of maximum obliquity of principal effective
stresses ( /os ), pore water pressure (u), pore pressure parameter
A (= _) and probable differences between mid-height and base pore
water pressure versus time to 1% strain have been developed and are pre-
sented in Figures, 11, 12,' 13 and 14. (Figure 11 is /& vs. t 1 ; Figure 12,
u vs. tj; Figure 13, A vs. t and Figure 14 is ubase - uneedle vs* t )
The other data gathered frm the tests on normally consolidated
samples, neglecting temporarily those tests in which the rate of strain
was changed in the midst of the test, are the water content data at
failure and the data on pore water pressure gradients. Every sample
in the main series of tests was sliced into six slices by horizontal

-25-
 1

cuts. A very thin slice at the bottom and at the top, if top
drainage was employed, was discarded to eliminate the possibility
of the water content in these areas having increased due to picking
up water from the stone as the sample was dismantled. These data are
listed in Appendix B on the stress versus strain curve from each
individual test. The plots in Figure 15 represent an attempt at a
statistical analysis of these data. In tests NCl-F, NC2-F, and NC3-F,
which were all rapid tests, the water content was determined as rapidly
as possible (2 to 5 minutes) following the end of shearing. Therefore,
the results of these determinations probably indicate the distribution
of water contents before shearing. The data from these determinations
are plotted in (c) of Figure 15. All of the data were plotted in
Figure 15(a) and envelopes enclosing the data were drawn. Inner
envelopes enclosing 85% of the data were drawn and the average of
this 85% area is plotted in 15(b), Subtracting the preshear distribu-
tion from this average leaves 15(d), i.e. (b) - (c) = (d)j, which
is an indication of the possible migration of water within the samples,
either during shear or following shear and while the sample was being
dismantled.
In Figure 14 are plotted all of the data from tests in which
the pore water pressure was measured at both the base and the mid-
height of the sample. In general, two points are plotted from each
such test, one being the pore pressure difference at maximum deviator
stress, and the other being the maximum pore water difference occurring
during the course of the test. The solid line drawn in this figure
represents the best estimate of the gradients in samples of this soil
that have to be contended with as they are related to the strain rate.
C. Behavior of Normally Consolidated Samples at Sal Strains
The section presents and sumnarizes the behavior of the normally
consolidated samples during the first 1% strain. These data were obtained
from the stress versus strain and pore water pressure versus strain plots
presented in Appendix B. The data from the individual plots were averaged
numerically and sumnary plots of (r- - o379 ) versus strain and pore water
pressure versus strain are presented in Figure 16. The plots presented
represent the numerical average of 4 fast tests (t 1 1 min.) and 9 slow
tests (t 1 = 500 min.). The numbers of the individual tests are listed
in Figure 16. In addition to the average behavior, the range of the
data is indicated in all cases.
From these averages of stress versus strain and pore water
pressure versus strain data, the plots in Figure 17 were obtained. In
this figure are plotted, on the basis of numerical averagesc
versus strain, the pore pressure parameter A ( ) versus strain
and the quantity [" , + ' . as a function of strain, all
representing behavior to one percent strain.
All of the factors acting to introduce inaccuracies into the
observed magnitudes of stress versus strain and pore pressure versus
strain parameters are magnified in importance vhen behavior at small
strains is considered. During the course of this research, every pre-
caution was exercised to insure that the data obtained were reliable
and accurate, including the data describing the behavior at small strains.
If it is assumed that the observed and recorded data plotted in the curves
of individual tests in Appendix B are true and accurate records of actual
measured quantities continuous from 0 to 1% strain, then any inaccuracy
lies in interpreting just what the measured quantity represents. This
assumption is, on the whole, justified. The pore water pressure in most
instances was obtained from a continuous plot of pore water pressure
transducer output versus either strain or time. The rate of strain of
the more rapid tests was such that a sufficient number of readings of

-27-
the proving ring dial gage could be and were taken during the first
one percent of strain.
In the case of deviator stress versus strain data, any
inaccuracy would seem to arise from inaccuracy in the relationship of
stress to strain rather in any inaccuracy in recorded value. By this is
meant the effect of so-called seating errors which cause the recorded
change in length to be not representative of a uniform length change
over the entire sample length. In other words, the early strains might
be occurring in localized zones adjacent to the base or cap of the sample.
In situations where this problem was obviously present, for example NC6-S,
the stress versus strain curve was corrected for this effect. Corrected
areas were used in the computation of deviator stress, but this correction
was quite small at less than one percent strain and any effect of in-
accurate strains upon this correction would be negligible.
In addition to these seating errors, several other possibilities
of sources of inaccuracies in interpretation of pore water versus strain
behavior should be pointed out. Three major sources of error plague any
investigator attempting to measure developed pore water pressures and
the severity of these is considerably magnified at small strains. These
sources of error are listed below and comments on the nature of the
possible error at small strains are presented:
(1) Errors due to compliance in the measuring system. These
errors are, in the main, more severe in the faster tests and doub~y
influence the data at small strains. Response of the pore pressure
system was checked before each test, as described in Appendix A, and
kept to a standard of full response in one minute or less. Therefore,
the gross effect of this should be more or less constant for all tests.
However, a minute bubble of air, small enough to be ineffective in
introducing error into the pore water pressure observations of the
slower tests at all strains and the fast tests at larger strains, could
influence the observations made in the most rapid tests at small strains.

-28-
This effect would tend to lower the observed pore water pressure at
small strains in the fast tests.
(2) Minute leaks in the external system. Gross leaks would be
picked up in the response check before commencing the test. Such leaks
would spoil any test. Minute leaks, undetected in the response check,
would affect the observed pore water pressure in slow tests at all
strains, but would not affect the data from fast tests. The effect
would cause a lower pore water pressure to be recorded in the slow
tests.
(3) Factors exist which can cause the measured water pressure to
not be indicative of pore water pressure in the sample at all, but to
be simply a water pressure generated by the presence of the measuring
device. For example, a flexible stone on the base pedestal will
obviously produce a pressure in the water in the stone or beneath it
which would approach the deviator stress in magnitude. If the test
was slow enough or the soil being tested was permeable enough this
effect would be equalized by minute volume changes within the sample
and would cause no gross inaccuracy.
In the case of the fast tests at small strains, there is the
possibility of a minute void in the soil structure adjacent to the
porous probe and containing water. This void filled with water could
be compressed by the soil structure around it as stress was applied,
and in fast tests in soils of low permeability this compression would
produce pressure in the water in this void in excess of the pore water
pressure in the bulk of the sample. This would cause the measured pore
water pressure in the fast tests to be higher than the true value.
Considering these possibilities of error, the data plotted
in Figure 16 are remarkable consistent. This is especially true of
the recorded pore water pressures. In fact there is more scatter in
the data from the slower tests, while the maximum scatter would be
expected in the fast tests.

-29-
 A careful study of the data plots presented in Figure 16
leads to the following conclusions. There can be little question that
at any particular strain level in excess -of 0.1%, the magnitude of the
deviator stress is higher in the fast tests. The percentage increase
in deviator stress in the fast tests as compared to the slow tests is
much greater at low strains as compared to large strains. In fact,
the actual magnitude of the excess of deviator stress in the fast
tests as compared to the slow tests seems to be a maximum at approxi-
mately one percent strain. Little can be said about the modulus of
the stress vs. strain curve at either strain rate at less than 0.2%
strain. The secant modulus to any strain level is greater in the fast
tests. However, the tangent modulus in the fast tests, while starting
out greater than that exhibited by the slower tests, decreases as
straining progresses at a much more rapid rate than the tangent modulus
to the stress vs. strain curve of the average slow tests, and approaches
the slow test tangent modulus in value at one percent strain.
The pore water pressures recorded in the slow tests are felt
to be quite accurate in spite of the previously mentioned sources of error
possible at small strains. It is also felt that, if any inaccuracy is
affecting the observed pore water pressures in the fast tests, the effect at
low strains is to make the observed values slightly lower than the true values.
Considering the confidence of the author in the observed pore water pressures,
the plot of pore water pressure at small strains in Figure 16 seems to
indicate that there is little, if any, effect of rate of strain on the
relationship between pore water pressure and strain at small strains.
If any, the trend is toward slightly higher pore water pressures in
the more rapid tests.
At small strains, little, if any, destruction of structure is
occurring. The resistance to shear displacement is due to the mobiliza-
tion of bonds or bonding forces between particles and to particle inter-
ference. This interference can be either body interference or force field
interaction. It would appear that larger displacements would be necessary

-30-
to mobilize resistance due to particle interference than would be
necessary to mobilize resistance due to bonding. Both of these
components of shear resistance probably demonstrate a viscous type
behavior. However, it is felt that the more viscous component is the
resistance to displacement afforded by particle interference. In
addition, it is felt that the resistance to stresses, or the stress
versus strain behavior reflecting mobilization and stressing of bonds
between particles, is the more elastic of the two components until
straining sufficient to begin breaking bonds has occurred.
SKMPTON (1948) at the Second International Conference on
Soil Mechanics and Foundation Engineering presented a theory relating
pore pressure to applied stress changes assuming elastic behavior.
The relationship u = -- + 2 >F-.7 where N= C was developed by
1 + 2
C
by Skempton. Cc = 3(1-2 c is the compressibility (volume
E
decrease per unit all-sided effective pressure increase) and
CS = 3(1-2--- s ) the expansibility (volume increase per unit all-

sided effective pressure decrease) of the soil structure. E and.A

5 5
are the modulus of elasticity and Poisson's ratio operating in an
unloading situation, and Ec and . c are the same quantities where
compression is occurring. This equation can be written u = - 3 + -
1 + 2 X 3
Thus the pore pressure parameter A = 1 + 2 )s If the resistance
of the soil structure to an increase in deviator stress is purely
elastic and isotropic, with Cc = Cs = 1 and A = 1/3. A value of
A in excess of 1/3 would occur due to an increase in Cc or a decrease
in C
5E

Since = Ec x s , and increase in Ec without a

s - c
corresponding increase in E5 would cause a decrease in A, while an equal

-31-
mm 2

change in Ec and Es would not change A, all else remaining constant.

In Figure 17 are plotted the deduced values of "A" versus
strain. The plot of A values in the fast tests seems to indicate the
possibility of A - 1/3 at very small strains, while the plot of A
values from the slow tests is rather inconclusive regarding strains
less than .2%, seeming to head for A values -: 1/3 at small strains.
A values higher than 1/3 can be interpreted as indicating deviations
from purely elastic behavior. Such deviations are seen to play more
of a role in the resistance to deviator stress increases in the slower
tests.
In Figure 17 are also presented plots of obliquity )
versus strain and ( ) ) versus stralin.
The first quantity (di7/ 0 3 ) is usually taken as an indication of
mobilization of structural resistance to shear displacement, while
the second value, Bjerrum's parameter A (BJERRUM: 1961) is taken
as an indicator of the relationship of structural breakdown. In the
plot of o versus strain, it is seen that structural resistance
to shear displacement is mobilized much more rapidly in the faster
tests.
As has been pointed out in a previous section, a plot of
(--~ ) -j- (~/ -i) versus strain is an indication of the
relative magnitude of structural mobilization and breakdown. A
horizontal slope of such a plot indicates a constant rate of structural
mobilization and breakdown. A positive slope indicates an increasing
rate of structural breakdown, or a decreasing rate of structural
mobilization, while a negative slope indicates structure being
mobilized at an increasing rate relative to structural breakdown.
The study of stress versus strain behavior and pore pressure
development at small strains and its relationship to rate of strain is
seen to be a promising area of future research. On the basis of the
data presented the following conclusions can be drawn:

-32-
1. The resistance to shear displacement at small strains exhibits a
viscous type behavior pattern.

2. Either:
(a) The resistance to compression also exhibits a viscous
type behavior of such magnitude to cause the relationship between
pore water pressure and strain to be unchanged with change in
strain rate, or
(b) The development of pore water pressure at small strains
is a function of strain and independent of both strain rate and the
total stress level.

D. Behavior of Normally Consolidated Samples at Large Strains

This section presents and discusses the results of the tests
on normally consolidated samples emphasizing the behavior patterns at
strains in excess of 1%. Figures 4, 5, 6, and 7 summarize and average
the data obtained from 17 tests on normally consolidated samples having
strain rates varying from 1% in one minute (t = 1 min.) to 1% in one
day (t = 1440 min.).
In Figure 4 maximum deviator stress is plotted versus the
logarithm of time to 1% strain. Data is, of course, plentiful at the
two extremes of strain rate and hence the best estimate of intermediate
behavior is based on a linear logarithmic interpolation. A study of this
plot and the upper plot of Figure 7 (r -<r vs. strain) lead to the
conclusion that the maximum deviator stress increases with increasing
strain rate. Also, this maximum deviator stress occurs at smaller
strains in the fast tests.
Figure 4 indicates that, in the range of strain rates
considered, an increase in deviator stress on the order of 10% over
the slow test value occurs in the fast tests. The nature of this
increase in deviator stress is shown very clearly in Figure 10. If
a line be drawn at a slope of +454, intersecting the x axis at the

-33-
I

point of origin of the particular stress vector being studied, the

horizontal distance from this line, which, if no back pressure, is the
total stress vector, and the effective stress vector under considera-
tion represents the excess porewater pressure at that point of the
test. A line drawn from any point on the stress vector through the
origin is a measure of the obliquity of stresses at that point, or
is a measure of f (angle of shearing resistance in terms of effective
stresses). If & be the slope of a line through the origin on Figure 10,
then Tan & = Sin, . Two effects of changing the rate of strain are
then seen to contribute to the increase of deviator stress at failure in
the faster tests. These are: (1) a decrease in maximum obliquity as the
rate of strain is increased; this would tend to produce an opposite to the
observed effect (i.e., a decrease in maximum deviator stress) were it not
for (2) the accompanying lower porewater pressure at failure in the faster
tests. Therefore it may be said that the increase in strength accompany-
ing an increase in rate of strain within the limits of this test program
is due mainly to a decrease in porewater pressure at failure'and is in
spite of a decrease in . The magnitude of this decrease in /
is illustrated in Figure 11 (where Sin ), and the increase
s+
in porewater pressure at the lower rates of strain is shown in Figure 12.
The relationship between the A factors at maximum deviator stress
(SKEMPTON, 1954), where A = - is shown in Figure 13. From this
figure, the increase in Afailure with decrease in strain rate can plainly
be seen.
The relationship between deviator stress and porewater pressure
as the test progresses is shown clearly in Figure 18. After the initial
1% strain, the porewater pressure increases in almost direct proportion
to the increase in deviator stress in the slow tests. In the fast tests,
in contrast, after 1% strain (and also prior to 1%), each succeeding
increment of deviator stress produces an incremental increase in porewater
pressure much greater than the preceding (i.e. .- increases).

-34-
a

The A factor, however, while higher at any strain level in the slower
tests, continues to increase at more or less the same rate in both the
fast and slow tests as straining progresses. These trends are clearly
shown in Figure 19.
It has been demonstrated that in the samples tested in this
program the effect of a decrease in the time to failure is accompanied
by an increase in the deviator stress at failure. In addition, it has
been demonstrated that this increase in strength is due to a decrease
in the porewater pressure at failure and occurs in spite of a decrease
in j as the time to failure is decreased.
It seems important to decide upon the mechanism by which the
porewater pressure at strains in excess of 1% attains its higher level
in the slower tests. Three possibilities immediately suggest themselves:
(1) membrane leakage, (2) migration of porewater (i.e., pore pressure
gradients), and (3) a difference in the structural behavior at rapid
strain rates resulting in a lower porewater pressure. These possi-
bilities are discussed in the ensuing paragraphs. As the following
discussions with their accompanying data point out, the difference in
porewater pressure is attributable to a real difference in structural
behavior in the fast versus the slow tests.
(1) Membrane leakage: - Little reason exists to suspect that
membrane leakage represents much of a problem in the tests in which
two membranes and a layer of silicone grease were employed. Several
evidences available tend to minimize this possibility. The porewater
pressure versus strain curve presented for the slow tests in Figure 6
shows little change in porewater pressure from 8 to 12% strain. In
time, this represents nearly one and one-half days. In addition,
Tests NC 15SF and NC 16SF were brought to a failure condition and
allowed to remain at constant strain for approximately four days.
During this time, after the initial change in porevater pressure
as straining was stopped (which was slight), the only change in

-35-
porewater pressure of any significance was a drop-off attributed to
a temperature change. (The straining was stopped in the midst of a
heat wave and during the night the ambient temperature dropped 209F.
as the heat wave was broken). One other bit of evidence is available.
Test Nc-8S (see Appendix B) was performed using Cylrex as a cell fluid.
This is an inert heavy oil manufactured by the Sunocy-acuum Oil
Company and does not attack rubber. Here again, the ambient tempera-
ture changes interfere with direct use of the test results, but if
the drops in porewater pressure due to drops in ambient temperature
are added onto the end porewater pressure, and a correction made as a
ratio of the differences in chamber pressure between this test
(66.5 lb./in.2) and the average test (60 lb./in.2), the equivalent
porewater pressure turns out to be 40 lb./in. , or higher than that
developed in the average slow test, This falls in line as test NC-8S
was considerably slower than average. Therefore it seems reasonable
to discard the possibility that any gross portion of the higher
porewater pressure in the slower tests is due to membrane leakage.
(2) A second possibility is lowering of the mid-height porewater
pressure due to migration of porewater from regions of higher pressure
nearer the base. As is the case with the increase in porewater
pressure that could occur due to membrane leakage, this possibility
does not represent a fundamental process of soil behavior. The
possibility of the existence of a porewater pressure gradient
represents the little understood distribution of stresses in a tri-
axial test specimen. BISHOP, BLIGHT AND DONALD (1960), in a study
of the effect of rate of strain on base porewater pressure measure-
ments, have described this mechanism in the following manner:
The unequal distribution of stresses in a triaxial specimen
occurs due to end restraint at the base and at the cap. End restraint
has the effect in these regions of increasing the major principal
stress ( ;', ) slightly and considerably reducing the deviator

-36-
stress (:7- ) The basic pore pressure equation La 8 [aLo-3 + A (&i3-.Aq-3]
(SKENPMN, 1954) can be rearranged in the form 6 5 [ <7 - (/-A) 6e- .
Both of the previously discussed effect of end restraint will tend to
increase the porewater pressure at the base and cap over that at the mid-
height of the sample unless A >- 1. If the test is run slowly enough,
the porewater pressure will be equalized throughout the length of the
sample. If A sii 1, the effect of this equalization will be to increase
the mid-height porewater pressure over that which would be recorded in
a rapid test.
Another way of looking at this same phenomenon is to liken
the region of the test sample close to the stone to a consolidation
test specimen with lateral restraint. In such a specimen of saturated soil,
the porewater pressure set up is initially equal to the applied load. In
the case of the triaxial test specimen, it would seem reasonable to expect
that the porewater pressure in the region immediately adjacent to the
restraining stone would equal the deviator stress at the instant of
initial application. Thus, unless A was greater than 1, there would
be induced against the stone a porewater pressure greater than at the
center of the sample. The thickness of this region would depend on the
completeness of restraint and upon the stress versus strain characteristics
of the soil.
In several of the tests on normally consolidated samples, pore-
water pressure was measured at the base in addition to being measured
at the mid-height of the sample. The results of all these measurements
are summarized in Figure 14. A curve of probably typical gradient is
sketched through the points on this plot. The maximum measured
gradient was 5.0 lb./in.2 (i.e., the porewater pressure at the base of
the sample was 5 lb./in.2 higher than at mid-height) but this measure-
ment seems to be extreme. A more typical value would be 2.0 lb./in,
higher at the base than at the midplane in the faster tests. This
higher porewater pressure presumably exists in a thin zone of the

-37-
sample adjacent to the stone and rapidly grades to the mid-
height porewater pressure, so it would seem inconceivable that
this gradient, if allowed to be dispelled by migration of porewater,
could raise the mid-height porewater pressure any significant amount.
Since, at maximum deviator stress, the porewater pressure in the
2
typical slow test is more than 5 lb./in. higher than that in the
typical fast test, there would seem to be little reason to expect
that porewater migration would account for any significant amount of
this increase.
Additional evidence of the insignificant effect of porewater
pressure gradients on the increase of deviator stress concurrent with
a decrease in the time to 1% strain is given by the water content data
as presented in Figure 15. These data scatter considerably and
it is rather hard to draw any conclusions from them. Looking at (b)
of Figure 15, one can note a suggestion of a possible migration of
porewater from slices 2 and 5 to slices 3 and 4, overlooking slices
1 and 6. However, this suggestion disappears when one studies (d)
of Figure 15. The best that can be said is that there is no gross
evidence of any migration of porewater.
(3) The third possibility is that the lower porewater pressure
in the more rapid tests is a real phenomenon resulting from the
manner in which the soil structure itself responds to different rates
of stress application. The author feels that one of the major contri-
butions of this research is the demonstration of the effect of strain
rate on the manner in which the soil structure resists shear displace-
ment as manifested in the pattern of porewater pressure development.
At small strains the porewater pressure, if considered
per se, appears to be independent of strain rate. It was suggested
that this may be a fortuitous occurrence indicating an earlier and
greater deviation from purely elastic behavior in the slower tests.

-38-
044900pp

In contrast, at larger strains and presumably after a degree of

structural breakdown has occurred, the porewater pressure at any
strain level is higher on the slow tests.
The phenomenon of the lower porewater pressure at any
strain level past 1% in the faster tests, and the higher A value in
the slow tests at any strain level, can be attributed to a resistance
to compression at any strain level that is time dependent and exhibits
viscous behavior. If only the tests carried to completion at a constant
strain rate are considered, the observed behavior follows with this
explanation. However, a series of tests in which the strain rate was
stepped at preselected strain levels (i.e., at certain predetermined
values of strain, the strain rate was instantaneously increased or
decreased) was run. The detailed results of these tests will be
presented in the next chapter. However, the behavior of the porewater
pressure in the tests which progressed to a strain of 2% or greater at
ti = 500 min. before a step in strain rate to t= 1 min. show a decrease
in porewater pressure following this switch which seems to the author to
be attributable only to some dilatent component in structural behavior.
That this dilatent component exists is shown in Figure 20.
In this figure are plotted the results of four tests in which the strain
rate was stepped from 1% in 500 min. to 1% in 3/4 min. In all tests
where this step in strain rate occurred at strains sufficiently large
to represent a reasonable degree of structural breakdown, the porewater
pressure following the step in strain rate leveled off at a value lower
than that existing before the step. This occurred even in the presence
of a considerably higher deviator stress. The complete discussion of
these tests appears in Chapter IV.
While the increase in deviator stress that has been observed
with decreased time to failure in the tests on the normally consolidated
samples has been attributed to the lower porewater pressure in the more
rapid tests, this increase had to occur in spite of a decrease in the

-39-
angle of drained shearing resistance f. This decrease is evidenced
in the decrease in the principal effective stress ratio at failure as
shown in Figure 11. This decrease in was also noted by
(CRAWFORD, 1959) but the opposite trend was recorded by (BJERRUM, 1960).
However, this decrease appears to be a real phenomenon occurring in the
fat clay used in the present work.
Several mechanism have been postulated which offer explanations
for this decrease in i . One of the simplest explanations of this
observation lies in attributing it to a prestress effect as illustrated
in Figure 28(d). The higher porewater pressure in the slower tests,
indicated by the horizontal projection of the distance x-y, results
in a lower effective stress normal to the failure plane at the time of
failure. The postulation of i Hvorslev line independent of strain rate
is not necessary.
The author feels that in this case the relationship between
effective normal stress and shear resistance reflects the relative
rates of bond or resistant force link mobilization and destruction.
The behavior of 3 as straining progresses is illustrated in
Figure 9. In this plot, the author interprets a positive eC(Z &3 )
to indicate that resistive force links are being formed at more
rapid rates than the rate of dstruction of these force links. A
horizontal slope (j It Z ) indicates bond destruction and bond

mobilization are occurring at the same rate while a negative slope,

indicates a more rapid rate of bond destruction.
Thus, in the rapid tests, bonds are initially mobilized
much faster than is the case in the slow tests. However, an equal
rate of mobilization and destruction is reached at approximately
3J% strain. Following the attainment of this level, resistance
force destruction occurs at a rate in excess of the rate of bond
formation, However, in the slow tests, bond formation occurs at a
rate in excess of bond destruction until quite large strains are reached.

-40-
 If the resistance to shear displacement were largely due to
particle (or micelle) interference, the behavior in terms of obliquity
could be expected to be more strictly viscous and thus a higher obliquity
at any strain level in the faster tests could be expected.

In Figure 9 is plotted the parameter 3-/ vs. strain.

If aL&is a measure of structural breakdown and r is a measure
of structural mobilization, the shape of this plot can indicate much
regarding the manner in which the soil structure is behaving as it
mobilizes resistance to shear displacement. Two features of such a
curve shed light on the behavior of the soil structure, (1) the actual
value of a point on the curve at any level and (2) the slop of the curve
at any point.
The value of this parameter at any strain level is 'anindica-
tion of how much structure has been destroyed in relationship to the
amount of structural bonds being mobilized up to the strain level under
consideration. For example, referring to Figure 9, at 1i2% strain more
shear displacement resisting bonds have been mobilized per bond destroyed
in the fast tests. At 8% strain the relationship between statistical
number of bonds destroyed and statistical number of bonds mobilized is
about the same in both tests. However, the shape of the two curves tell
much about the nature of the bond breaking versus bond formation process.
In the fast tests, following 1% strain, the relationship between bond
formation and bond destruction is more or less constant until a strain
of about 7% has been reached. This is indicated by the horizontal
slope of the curve. The positive slope of the curve for the fast
tests after 7% strain indicates that the rate of bond destruction
relative to bond formation is beginning to increase. In contrast,
from small to large strains, the curve for the slow tests shows a
negative slope, indicating bond formation at an increasing rate
relative to bond destruction.
 I

Chapter IV

NORMALLY CONSOLIDATED SPECIMENS

WITH STEPPED STRAIN RATE

A. Introduction
This chapter presents and discusses the results of a series
of fourteen tests on normally consolidated samples of the fat clay.
These samples were prepared and set up in exactly the same manner as
described in the previous chapter. The tests in this subseries differ
from those of the main series only in the fact that at some pre-
determined strain level during the progress of each test, the strain
rate was drastically changed. This series of tests was originally
embarked upon with the limited objective of demonstrating the so-
called "dilatency" effect, described previously, and its relationship
to strain rate. It was postulated that some portions of the soil
structure being strained exhibit a tendency toward collapse and thus
a volume decrease (positive porewater pressure development) while other
portions exhibit a tendency toward volume increase (dilatency) due to
structural viscosity and thus tend to reduce the developed porewater
pressure (negative porewater pressure development). Statistically, in
this and most (if not all) normally consolidated soils, the dilatency
effect plays a relatively minor role. However, it was felt that a
rapid change in strain rate would upset the balance in these two
behavior patterns and, if a lowering of porewater pressure with an
accompanying increase in deviator stress would ensue, the existence
of a dilatency component would be demonstrated. The data resulting
from the tests at a constant strain rate can be explained (at greater
than 1%strain) without postulating any dilatency effect, only
attributing viscous behavior patterns to both shear displacement
resistance and resistance to compression. Since it was suspected by the
author that a dilatency component did exist, this series of tests
was planned. In the original planning for this supplemental series
of tests, the sole object was to vary the rate of strain during the
course of any test and to observe the effect of this change on the

-42-
developed porewater pressure.
However, while the preliminary few tests in which the strain
rate was stepped can be interpreted as clearly demonstrating the presence
and importance of this structural viscosity effect on porewater pressure
development in the normally consolidated samples, other effects of equal
or greater importance were noted in the effect of this change in strain
rate on the deviator stress . To follow up and pin down the cause of
this rather significant phenomena, the scope of this supplementary test
program was extended.
The effect of a step in strain rate on the deviator stress
has been noted by others in the Massachusetts Institute of Technology
Soil Engineering Laboratory and at other establishments. However, the
rate changes were nowhere near so drastic in these cases and the change
in porewater pressure had usually been overlooked. The effectiveness
of the strain rate change technique in the current research is due to
the ability of the novel pore pressure measuring device to rapidly
(almost instantaneously in this case) detect changes in porewater
pressure occurring as a result of the change in strain rate, and to
detect these changes in a portion of the test specimen which is
relatively unaffected by anomolies in stress distribution. Consequent-
ly, it is felt that the results of these tests stand by themselves as
an important contribution to knowledge resulting from this research
effort.

B. Testing Procedure
This subseries of tests involved fourteen normally consolidated
samples. These tests are NC4FS, NC5FS, NC9-SF through NC17-SF, tests
NC18SFID, NC19SFIC and NC20SFIC. Some of these represent tests from
the main series in which the strain rate was stepped following attain-
ment of peak deviator stress.

-43-
 Now

Most of these tests were performed using the same load

frame, triaxial cells and proving rings as were employed in the main
series of tests. However, as the tests in this subseries were begun,
considerable skepticism was voiced regarding the validity of the
observed phenomenon of increased deviator stress with increase in
strain rate. The reason for doubting the observed behavior patterns
stmed frm a feeling that some frictional effect in the bushing of
the piston and bushing assembly of the triaxial cell could be making
a major contribution to the effect of step in strain rate upon the
observed deviator stress. A careful investigation of this possibility
was undertaken and is described in the next section.
In an attempt to satisfy the skepticism, several tests were
set up in the cells employed in a companion research effort (NASIM 1961).
The cells had been modified to provide a method of measurement of the
load on the soil sample by means of a load cell inside the triaxial
chamber. This load cell consisted of a lucite platten, a stainless
steel piston riding in a Thompson Ball Bushing and activating an
electric force transducer similar in design to those employed in
measuring porewater pressure by means of the porous probe. The
electronic equipment necessary for reading out the information gave
considerable trouble, and there is doubt as to the calibration of the
load cell and thus to the information regarding the peak values of the
deviator stress. In addition, the drainage and porewater pressure
measuring systems were never perfected in that leakage was known to
occur. Hence, it is rather difficult to make any comparison of the
gross results of these tests with those performed in the standard
cells. However, there is no question that the results of changes in
strain rates in tests performed in these cells gave quite similar
results to tests in standard cells. A total of three tests were
attempted in these cells, NC-18SFIIC, NC-19SFIC and NC-20SFIC. Of
these, NC-19SFIC was the most satisfactory. The deviator stress
increase following the change in strain rate is the same percentage
as the increase obtained using standard cells.

-44-
C. An Investigation of Stepped Strain Rate on Piston Friction
At the conclusion of each test in this series, the piston of
the triaxial chamber was withdrawn to a position such that the lower end
of the piston was well above the cap of the sample. The load frame was
then started causing the piston to descend into the cell at the slow
rate. The proving ring reading was noted. In all cases this reading
in excess of that caused by chamber pressure was less than 1 lb. and
in most cases considerably less than 1 lb. There was no difference
between the Clockhouse Cells and the Norwegian Geotechnical Institute
Cells in this regard. The speed of descent of the piston was then
stepped to the fast rate. In no case was any change in proving ring
reading recorded.
A Clockhouse Triaxial Cell was set up in the standard manner
with the standard oil in the top and a chamber pressure of 60 lb./in.2
applied. A horizontal thrust of 5 lb. was applied to the piston 3j
inches above the top of the bushing. Even with this load, less than
1/2 lb. of piston friction was noted. This was a constant value,
independent of the rate of strain or of any sudden changes. In fact,
the load frame was cranked by hand at a quite rapid rate and no change
in the amount of piston friction was noted.
The same setup was repeated using a Norwegian Geotechnical
Institute Cell. Here the piston friction with the eccentric 5 lb. force
was 2j lb., and was again independent of strain rate and of any sudden
changes in rate. These observations coupled with the results of the
tests in triaxial chambers equipped for internal load measurement lead
to the conclusion that while piston friction may slightly influence
the gross test results, especially with the Norwegian cells, the effect
of rate of strain and sudden change in rate of strain is negligible.
There exists the possibility, with the Norwegian cells, of a 2 or 3%
error in maximum recorded deviator stress on the gross test results.

-45-
D. Results of a Negative Step (Step Decrease) in Strain Rate
Two tests on normally consolidated samples were performed
which involved a step decrease in strain rate at some strain level.
The results of these two tests, NC4Fs and NC5FS, are presented in
Figures B-4 and B-5 of Appendix B. In test NC4FS, the test was
carried to a strain of 7% at a strain rate of 1% in 45 sec. (t = 3mi.).
At 7% strain, the rate of strain was decreased to 1% in 10 min.
(t, = 10 min.), 10 minutes being the time required for l% additional
strain. The test was stopped at 7.2%strain.
In test E5FS, the initial strain rate was again 1% in 45
seconds (t = 'min.). At 42% strain the strain rate was decreased to
1% in 60 min. (t1 = 60 min.). Straining was continued at this slower
strain rate until a strain level of 8j% had been attained.
Test E-4FS contributes relatively little to this discussion
as the test was stopped before the effects of the step in strain rate
had reached an equilibrium state. Comparing this test with the average
results of tests on normally consolidated samples leads to the following
observations. Prior to the step decrease in strain rate, this sample
was behaving in a manner almost identical to the average behavior for
t = 1 min. as shown in Figure 7 and Figure 8. At 7% strain in the
average tests, the deviator stress (Z~ - C9 ), porewater pressure
( u ) and obliquity ( /I ) were 43 lb./in. 2, 32 lb./in.2 and
2.70 respectively in test NC-4FS. If the single test in the "average"
series which represents an intermediate strain rate is taken as being
comparable with the portion of NC4-FS after the step decrease in strain
rate, the following comparison can be made. At 7.2%in the test
representing t 1 = 20 min., Figures 7 and 8, the deviator stress, pore-
water pressure and ratio are 39.5 lb./in.2, 35 lb./in.2 and
2.60 respectively. In test NC-4FS following the step decrease in strain
rate, these values are 38 lb./in. , 33 lb./in. and 2.30 respectively.
However, as will be demonstrated when the results of test NC-5FS are

-46-
discussed, these values do not represent equilibrium structural
behavior following the change in strain rate.
In test NC-5FS, however, straining was continued for a
sufficient period following the step decrease in strain rate to
establish an equilibrium yielding condition. In this test, the strain
rate was decreased from ti = 3/4 minute to t = 60 minutes. This step
decrease in strain rate was imposed at approximately 4j% strain in this
test. The deviator stress prior to the step decrease was 42.0 lb./in.2
the porewater pressure 30 lb./in.2 and the obliquity ratio 2.48
(43, 312 and 2.52 respectively in average test). Immediately following
the step decrease, these quantities were 34 lb./in.2, 31l4b./in. and
2.35 (deviator stress, pore pressure and obliquity ratio). However,
further straining at the new and slower strain rate caused a rise in
these same quantities (i.e., deviator stress, porewater pressure, and
obliquity ratio) to 37.6 lb./in.2, 35.3 lb./in.2 and 2.62 at 8% strain.
(As compared with 39 lb./in.2 35 lb./in.2 and 2.60 in the t = 20 min.
test in Figure 7.)
In summary, the following description of behavior of this fat
clay following a step decrease in strain rate represents in the author's
opinion, the observed behavior, Following an instantaneous or "step"
decrease in strain rate from t1 = 3/4 minute to t = 60 minutes:
1. The deviator stress drops to a level below that commensurate
with the new strain rate and rises gradually to such a deviator stress
as straining at the new rate is continued. (The deviator stress at the
new or slow strain rate is lower than that attained before the strain
rate step.)
2. The porewater pressure jumps slightly immediately follow-
ing the step decrease in strain rate and continues to rise, attaining
an equilibrium level approximating the pore pressure in a test at the
slow strain rate at a constant rate of strain. (Higher than the pore
pressure in a fast test.)
 3. The obliquity ratio ( /&a ) drops following the
step decrease in strain rate, but rises as straining continues at
the new (slow) strain rate to a level approximating the obliquity
ratio in a slow test at a constant strain.

-48-
 E. Results of Positive Steps (Step Increases) in Strain Rate
Imposed at 4L Strain in Normally Consolidated Samples

In three tests on normally consolidated samples of the

saturated fat clay, step increases of varying magnitudes were imposed
at 4j% strain. In these tests the samples were all brought to 4}% strain
at the same rate, t 1 = 500 minutes. The stress versus strain, pore
pressure versus strain, and obliquity ratio versus strain data for these
tests, NC-9SF, NC-10SF and NC-llSF are presented in plots in Appendix B.
The step increase in strain rate is given as the time required for 1%
additional strain (t+1 ). In test NC-9SF, t+ = 3 minutes; NC-10SF, t
1 +1
1
50 minutes, and in EX-llSF, t = minute. Table IV-1 briefly summarizes
this series of tests.

Table IV-1

TESTS WITH POSITIVE STEP IN STRAIN RATE AT 4}% STRAIN

et 4}
EC E> 4} Deviator 2Stress Pore Pre sure Obliquity Ratio
Test tin.) n (lb./in.2 )__ (lb./in.
1 tBefore After % In- Before After %,In- Before After % In-
Step Step crease Step Step crease Step Step crease

NC9SF 500 3 37.5 45. 20% 32.3 32.8 lj% 2.38 2.74 15%
E10SF 500 50 37.5 40.8 9% 31.T 32.7 3% 2.27 2.44 7%
EllSF 500 36.5 46.6 27% 35-1 36.0 2j% 2.53- 2.95 l6b%
-

Summary plots of the results of these three tests are

presented in Figure 20. In this figure are presented plots of
deviator stress, porewater pressure and obliquity ratio versus
strain. In addition, a plot showing the behavior of the pore
pressure parameter "A" following the step increase in strain rate
is presented. This figure was derived by arriving at curves
representing the average or typical behavior of these quantities
up to 44% strain. (Actually the typical curves for t 1 = 500 minutes
in Figures 7 and 8 were used. The behavior of the three tests in this
series deviated from this typical behavior only to a very minor degree.)
The actual observed behaviors following the step increase in strain rate
were applied to these "typical" curves. For this reason, the actual
magnitudes of values following the step in strain rate may differ slightly
from those noted in Table IV-1; however, the percentage changes will be
more or less the same.
The deviator stress, pore pressure and obliquity ratio behavior
of these samples can be described as follows. All tests were bronght to
4k% strain with ti = 500 minutes with all measured quantities behaving
quite comparable to the typical behavior portrayed in Figure 7.- Behavior
typical of a test at a uniform strain rate of tj = I minutes would be
expected to differ only slightly from that shown as typical for t =

1 minute as shown in Figure 7. Ideas of behavior typical of t1 = 50 min-

utes and t1 = 3 minutes in steady strain rate tests are not available,
and the conclusiveness of the following description suffers in this
regard. Following the step in strain rate, the deviator stresses rise
to values about 10% higher than would have been attained in constant
strain rate tests at the new strain rate and at the same level of
strain. A considerable poxtion of this increase in deviator stress
was obtained almost instantaneously in terms of strain, but additional
strain on the order of 4% in tests NC-9SF and NC-loSF and j% in tests
NC-llSF were needed to fully realize the maximum deviator stress. As

-50-
additional strain at the new strain rate was imposed, the deviator
stress gradually fell to values that appear to be compatible with
values that would be attained at comparable strain levels in tests
having a constant strain rate equal to the post-step rate in these
tests.
The porewater pressure following the step increase in strain
rate in Test NC-13SF continued to rise in approximately the same manner
as before the step until peak deviator stress had been reached. After
the attainment of peak deviator stress, the pore pressure began falling
off and, as straining at the new rate was continued, seemed to attain
a value and a behavior pattern characteristic of a steady rate test at
the new strain rate. In Test NC-10SF, the behavior was similar with
the exceptions that a large initial jump (in comparison to NC-llSF) in
porewater pressure, a subsequent leveling off until a strain-pore
pressure level compatible with a steady rate test had been reached,
followed by a behavior comparable to a steady rate test. Test NC-98F
exhibited porewater pressure behavior intermediate to NC-10SF and NC-llSF.
Prior to the step in strain rate, the porewater pressure had been behaving
typically, rising gradually. Following the step, the porewater pressure
remained constant as strain was continued until a strain-pore pressure
situation compatible with a steady rate test at the new strain rate had
been reached.
Looked at in terms of the porepressure - deviator stress
relationship, all tests exhibited a similar behavior pattern consistent
with the strain rate following the step. The pore pressure parameter
"A" is plotted versus strain following the step in strain rate in
Figure 20. After the increase in strain rate, the A values begin
dropping rapidly with additional strain, and gradually achieve a
lower level. The amount of additional strain necessary to achieve
this lower level increases as the new (post-step) strain rate increases.
Following this drop, the A values rise to values that seem to be

-51-
comparable to those in steady rate tests at strain rates comparable
with the post-step strain rate.
The measured porewater pressures following the change in
strain rate would be influenced by any minor lag in the measuring
system to only a very slight extent as the actual magnitude of the
change to be measured is relatively small. In addition, the system
is already operating under a positive pressure in excess of two
atmospheres which, if looked at as a back-pressure in terms of
post-step pore pressures, would tend to minimize several of the
factors contributing to any slight lag in the measuring system.
Hence it is felt that the porewater pressures measured following
the step in strain rate are of the utmost reliability and accuracy.
Also plotted in Figure 20 is a summary of the obliquity
ratio behavior exhibited in these three tests. The obliquity ratio
is sensitive to slight irregularities in either pore pressure or
deviator stress and thus the behavior shown is somewhat erratic.
However, the gross pattern is compatible with the following descrip-
tion which must follow from the description of pore water pressure
and deviator stress behavior following the step increase in strain
rate. Following the step, the obliquity ratio jumps sharply, coming
to a rather sharp peak in the tests with the most rapid post-step
strain rate (NC-9SF and NC-11SF). Following the attainment of the
peak, the obliquity ratios drop off and appear to reach levels
compatible with those that would be achieved in steady rate tests
at the post-step strain rate.
In Figure 21, the relationship between deviator stress and
obliquity ratio attained following the step increase in strain rate
are plotted against the logarithm of the post-step strain rate.
In addition, the relationship between deviator stress at 4if1strain
and strain rate, and obliquity ratio at 44$% strain and strain rate
for typical constant strain rate tests, are shown by way of comparison.

-52-
F. Results of Tests in Which Equal Positive Steps in Strain Rate Were
Imposed at Varying Strain Levels - Normally Consolidated Samples
Described in this section are a series of four tests in which
a step increase in strain rate of an equal magnitude was imposed (Tests
NC-13SF, NC-14SF, NC-153F, and NC-16SF). In each case, the initial
strain took place at a strain rate giving a time to 1% strain of 500
minutes (t, = 500 minutes). At certain preassigned strain levels, a
step increase to a strain rate giving a time to 1% additional strain
of 3/4 minutes was imposed. This strain rate step was imposed at
progressively larger strains in each sample. In Test NC-15SF the strain
rate step was imposed at 1.05% strain; in Test NC-13SF at 1.75%, NC-16SF
at 4.40%, and NC-14sF at 7.25%.
The complete stress vs. strain and porewater pressure versus
strain data from these tests will be found plotted in Appendix B. A
sunianry of the behavior of these samples is presented in Figure 22.
Here again the curves of deviator stress and pore water pressure at
t 500 minutes (i.e., pre-step curves) are the typical slow rate
curves from Figure 6. However, Tests NC-15SF and NC-16SF exhibit
stress versus strain behavior almost coincident with this typical
behavior. Test NC-13SF differs in both deviator stress and pore
pressure only to a minor degree, being slightly high in both quanti-
ties at any strain. Test NC-14SF differs to a slightly greater extent
with both deviator stress and porewater pressure being lower than
typical at any strain level. In all cases, however, the numerical
magnitude of the post-step increases were added to the typical curves.
All of these samples were trimmed from prisms of soil cut
from Batch C. The typical curves for the fast strain rate presented
in Figure 5 rely quite heavily on Batch B samples with no Batch C
samples being represented. Hence, exact conclusions regarding the
fast rate behavior for these samples in a steady strain rate situation
cannot be drawn. However, for comparison, the typical fast curves from
Figure 5 are shown in Figure 22.

-53-
 In terms of deviator stress, the effect of a step in strain
rate from t1 a 500 min. to t+1 = 3/4 minute can be described as follows:
In all cases a considerable increase occurred. The actual magnitude
of this increase differed only slightly, in all four cases being 13.1 lb./in,
when the step was imposed at 1.05% strain, 12 lb./in.2 at 1.75%, 10 lb./in.2
at 4.40%, and 8.5 lb./in.2 when the step was imposed at 7.25% strain. In
each case, approximately ,%additional strain was required to completely
develop the full increase in deviator stress due to the step in strain rate.
This is rather hard to define in Test NC-15SF. However, a strain of I
following the step in strain rate seems to have been required to bring
the deviator stress versus strain behavior to a pattern compatible with
the new strain rate. Table IV-2 following summarizes the immediate
deviator stress increases following the step increase in strain rate.

Table IV-2

Deviator Stress Behavior Following Step Increase in

Strain Rate - Normally Consolidated Samples

Strain ( y7o~)
("~-43) ( -3) % %
Level at Test t t.1 beforg afteE typicl iscrease in excess
Step (%) No. (mn.) (min.)(lb/in ) (lb/in ) (lb/in ) over of typical
slow rate fast rate

1.05% E15SF 500 3/4 29.4 42.5 41.5 45%

1.75% E13SF 500 3/4 33.5 45-5 42.5 36% 7 %
4.40% NC16SF 500 3/4 38.6 48.2 43.2 25% 11b
7.25% NCl4SF 500 3/4 40.0 48.5 42.7 21% l3b%

-54-
In addition, both the percentage increase over the pre-step deviator
stress and the increase over the typical fast rate deviator stress are
plotted against the strain at which the step in strain rate occurred
in Figure 23.
From a study of Table IV-2, Figure 22 and Figure 23, the
following conclusions can be drawn regarding the effect of the strain
level at which a step increase in deviator stress is imposed and the
subsequent deviator stress behavior.
1. Both the actual magnitude of the increase in deviator
stress following the step in strain rate and the percentage increase
is greater the lover the strain level at which the step is imposed.
However -
2. The actual magnitude and the percent increase in the
deviator stress over that typical for the new strain rate are less
at low strains, becoming greater as the strain level at which the
step in strain rate is imposed increases.
3.The exact location of a curve representing deviator stress
versus strain behavior for a steady rate of strain yielding a time to
1% strain of 3/4 minute is not known. In the opinion of the author,
however, the deviator stress versus strain behavior following the step
in strain rate tails off to approach a behavior pattern typical of the
new strain rate as straining continues at the new rate.
The pore pressure behavior following the change in strain rate
is somewhat similar to the behavior described above in that it ultimately
approaches the pore pressure versus strain behavior typical of the new
strain rate. Immediately following the step in strain rate, the pore
pressure rises above that which it would have attained if no step had
occurred. As straining continues, however, the porewater pressure
drops to the value it would have assumed if no rate change had
occurred and approaches the level typical of the new strain rate. The
change in pore pressure seems to be more fundamentally described by
the behavior of the pore pressure parameter "A" following the step

-55-
in strain rate. The value of "A" in each case (with the possible
exception of test NC-14SF) falls quite rapidly to a value typical of
the new strain rate and thereafter exhibits values at any strain a3most
identical to those typical of the post-step strain rate.
The obliquity ratio versus strain behavior following the change
in strain rate is of the same pattern as the deviator stress versus strain
behavior. The actual magnitude of the increase is not greatly different
when the shift takes place at differing strain levels. However, the per-
centage increase decreases as the strain level at which the step in strain
rate was imposed becomes greater. The increase in obliquity ratio in
excess of that typical of the new strain rate becomes progressively much
greater as the step strain level increases. As post-step strain at the
new strain level continues, the obliquity ratio assumes a value compatible
with that expected in a steady rate test at the new strain rate.

G. Relaxation Tests - Normally Consolidated Samples

In compression testing, two types of studies can be made involv-
ing behavior during a period of stopping in the course of a test. The
more common is the study of "creep" behavior following the attainment
of a predetermined deviator stress. In this type of a test, following
the attainment of a particular deviator stress, the deviator stress is
held constant and the resulting deformation, or strain, measured as a
function of time. Such a test has come to be referred to as creep test.
In a relaxation test, on the other hand, following the achievement of a
predetermined strain level in a compression test, the strain is held
constant by varying (reducing) the deviator stress. The magnitude of
the deviator stress needed to maintain this constant strain is observed
as a function of time.
This section describes and presents the results of two tests
on normally consolidated samples of the fat clay (Tests NC-15SF and
NC-16SF) involving a period of relaxation. Both of these tests were
of samples previously involved in the series of tests with a positive

-56-
step in strain rate described in the preceding section. The complete
stress versus strain curves of these tests, found in Appendix B, include
the relaxation and subsequent continuation of loading data.
In both of these tests, a period of relaxation for a time 'of
approximategy10,000 minutes, at a strain level reached in identical manners
(i.e., t1 = 3/4 minute), were imposed upon the samples. The strain level
was held within + 1 inch by carefully observing the strain
10,*000
dial and correcting the load manually to maintain this constant strain.
The actual strain level differed in the two samples, being 7 S in
3
NC-15SF and 9 1% in NC-16SF, but both strains are felt to represent approxi-
ately the same degree of structural breakdown. Following the period
of relaxation, the samples were reloaded, NC-15SF at t+1 = 500 minutes
and NC-16SF at t+ = 3/4 minutes. The results of this continuation
of loading are shown in the figures in Appendix B.
In Figure 24 are presented the observations of behavior in
terms of deviator stress and porewater pressure. These tests were
commenced in the midst of a period of hot weather (avg. temp. = 900F).
During this period, there was one day of cooler weather (avg. temp. = 700F).
This change in ambient temperature complicates the observed data to a
considerable degree as is seen in Figure 24. Between points A and B in
the plots in Figure 24 for both tests, a drop in ambient temperature of
200F. occurred. This was accompanied by a significant drop in porewater
pressure in both tests, (drop of 8.5 lb./in.2 in NC-15SF and 9.0 lb./in. 2
in NC-16SF) and a drop in the deviator stress necessary to maintain the2
2 in Nc-15sF and 6.2 lb./in.
constant strain level (drop of 2.4 lb./in.
in NC-16SF). The greater drop in deviator stress observed in NC-16SF
may reflect the greater strain at which relaxation occurred.
One of the first items of importance shown by these tests (see
Figure 24) is the apparent lack of any increase in porewater pressure as
a result of membrane.leakage. This observation serves to bear out the

-57-
 I

conclusion reached earlier that membrane leakage was not a significant

factor in these tests.
Following a period on the order of 10,000 minutes at these
strain levels, (9500 minutes in NC-15SF and 7250 minutes in IC-16SF)
straining was resumed. In NC-15SF the rate of strain following
resumption of testing gave a time to 1% additional strain of 500 minutes
(t+1 = 500 minutes) while in Test NC-16sF straining was resumed at the
prerelaxation rate which gave 1% additional strain in 3/4 minute
(t+1 = 3/4 minute). In both of these tests, deviator stresses were
obtained far in excess of values appropriate to the new (post-relaxation)
strain rates. In NC-15SF, the peak post-relaxation deviator stress of
43.7 lb./in. is 10j% higher than typical of a constant strain rate
test having a ti - 500 minutes at 82% strain. In NC-16SF, the peak
post-relaxation deviator stress of 54.2 lb./in.2 is 30% greater than
would typically occur at 10% strain in a constant strain rate test
having a t1 = 3/4 minutes. In Test NC-16SF, the post-relaxation
deviator stress had every appearance of decaying to a magnitude com-
mensurate with the new strain rate. In Test NC-15SF on the other hand,
little if any decay in deviator stress occurred as post-relaxation strain-
ing was continued.
Little analysis of the pore pressure data from these two tests
is possible due both to the anomolous behavior introduced by the tempera-
ture change and to the lack of extensive straining following resumption
to testing. However, the data available point to the strain dependency
of the porewater pressure. In both cases, following resumption of
testing and the initial increase in porewater pressure, both curves
of porewater pressure appear to be seeking levels appropriate to the
respective rates of straining.
In terms of obliquity ratio ( /- ), following the
resumption of straining, test NC-15SF did not attain a 2, ratio

of a magnitude typical of the post-relaxation strain rate, but the

-58-
obliquity ratio was rising at the cessation of testing. In Nc-16sF
as straining was resumed at a strain rate yielding a t+1 = 3/4 sec.,
the obliquity ratio rose to a value far in excess of that typical
for the new strain rate, but the obliquity ratio was dropping off
rapidly at cessation of testing.

-59-
 Chapter V

BEHAVIOR OF TIHE NORMALLY CONSOLIDATED SAMPLES

A. Introduction
All of the important behavior characteristics related to
strain rate effect on normally consolidated samples of saturated fat
clay are summarized and discussed in this chapter. Any model or
mechanistic theory of shear strength behavior as related to strain rate
must satisfy these observed features of behavior.

B. Simary - Effect of Strain-Rate on the Deviator Stress - Normally

Consolidated Samples
The effect of the rate of strain on the deviator stress in
tests on normally consolidated samples can be summarized as follows:
1. The maximum deviator stress attained in any test increases
as the time to 1% strain decreases (Figure 4). This increase is on the
order of 10%as the strain rate is increased from t 1 = 500 minutes to
t= 1 minute.
2. At any strain, the deviator stress is greater in a test
with t = 1 minute than in a test with t 1 = 500 minutes.
3. The magnitude of strain required to attain maximum deviator
stress decreases as the rate of strain is increased (i.e., time to 1%
strain decreased). At t 1 - 500 minutes, 8*% strain was required to reach
maximum deviator stress while with t = 1 minute, 5)1 strain was required.
4. The maximum difference in deviator stress between a fast
and a slow test (both numerical and percentage-wise) occurs at less than
1%strain (0.6%) and is on the order of 55% of the slow test deviator
stress (Figures 25 and 26).
5. The increase in deviator stress is descriptively seen to
occur as a result of both a steepening of the prepeak stress vs. strain
curve and an elevation of the peak of such a curve and thus exhibits
elements of both viscous and brittle behavior patterns.

-60-
 6. As found by NASIM (1961) the greatest effect of strain
rate upon maxim%= deviator stress occurs at rates of strain more rapid
than the fast rate employed in this research.
7. The increase in maximum deviator stress within the range
of strain rates considered in this research can be mainly attributed to
the lower porewater pressure at failure in the faster tests.
8. If the strain rate is increased from t1 = 500 minutes to a
more rapid strain rate following the attainment of approximately 4% strain,
the deviator stress immediately attains a value far in excess of that
normal for the new strain rate, but decays to a value compatible with
that expected in a steady rate test at the new strain rate. (Figure 20).
9. The magnitude of this excess over normal deviator stress
for the new strain rate increases as the new strain rate is increased.
(Figure 21).
10. If a step increase in strain rate from t, = 500 minutes
to t+1 = 3/4 minutes is imposed at any strain level, a substantial
increase in deviator stress occurs. (Figure 22). The actual magnitude
of this increase in deviator stress decreases as the strain level at
which the step in strain rate is imposed increases. (Figure 23).
11. An increase in deviator stress over what would be normal
for the appropriate strain level in a steady rate test occurs as the
strain rate is stepped from t, = 500 minutes to t+1 = 3/4 minutes
regardless of the strain level at which such a step is imposed. How-
ever, the actual magnitude of the post-step deviator stress over that
expected at a comparable strain level in a strain rate test increases
as the level at which the strain rate is stepped increases. (Figure 23).
12. Following a period of relaxation at a constant strain
level, the deviator stress after resumption Of straining rises to a level
far in excess of that normal for the post-relaxation strain rate in
constant rate of strain tests.
C. Sinnny - Effect of Strain Rate in the Measured Porewater Pressure -
Normally Consolidated Samples

The effect of strain rate on the developed porewater pressure

in tests on normally consolidated samples can be summarised as follows:

1. The maximum porewater pressure developed in a test is lower

at the more rapid strain rate.

2. The porewater pressure at maximum deviator stress was lower

on the faster tests. (Figure 12).
3.Thus the "A" factor at failure decreases as the strain rate
is increased. (Figure 13). This would happen if the pore pressure did
not change as the strain rate increased.
4. Pore pressure gradients had only a minor effect on this
data as did migration of porewater (i.e., internal volume change)
(Figures 14 and 15). Membrane leakage was not a factor.
5. At strains, less than 1% the porewater pressure is independent
of rate of strain within the strain rates considered in this research.
Therefore, the pore pressure parameter "A" is greater in the slower tests
at small strains. (Figure 17).
6. At any strain greater than 1%, the pore pressure in the
faster tests is less than the pore pressure in the slower tests. The pore-
water pressure in the slower tests (t1 - 500 minutes) rises continuously
until straining ceases. In the faster tests, the porewater pressure
reaches almost a "plateau" rising only 0.5 lb./in.2 from 6% to 12% strain.
7. The porewater pressure parameter "A" in the fast tests at
any strain is a value approximately 0.2 lower than that in the slower
tests. This difference in "A" between fast and slow tests remains
constant as straining progresses. (Figures 17 and 19).
8. The porewater pressure following step increases in strain
rate at approximately 41% strain from t 1 = 500 minutes to t+ 1 = 50 minutes,
t = 3+1minutes and t +1 = g minute result in the behavior shown in

-62-
Figure 20 (i.e., after an initial change, the post-step porewater
pressure seeks a level commensurate with the post-step strain rate).
9. In terms of "A" factors, the post-step pore pressure
behavior represents an initial drop in A. The magnitude of this drop
depends upon the value of t+1 , increasing as the post-step strain
rate becomes greater. Following this initial drop, the "A" factor
rises to a value approaching that appropriate to the new strain rate.
10. Following steps in strain rate from t, = 500 minutes to

t+= minutes at succeeding greater strain levels, the pore pressure

increases immediately and subsequently drops to a level approximating
that appropriate to the new strain rate. The exact post-step behavior
is as shown in Figure 22.
11. In terms of "A" factor, the post-step porewater pressure
behavior results in a considerable drop in "A" and a subsequent rise
or leveling off to a value appropriate with the new strain rate.

D. Effect of Strain Rate on the Obliquity Ratio - Normally Consolidated

Samples

The quantity , herein termed "obliquity ratio" is

a measure of the development of shearing resistance on the prospective
failure planes with respect to the effective stress acting normal to
these planes. If it is assumed that the observed sample strain is a
shear yielding on the prospective failure planes, then the obliquity
ratio is a fair indicator of the manner in which resistance to shear is
being mobilized. The obliquity ratio is related to the angle of shearing
resistance by the relationship Sin =

This section summarizes the effect of strain rate on the

observed behavior of the developed obliquity ratio. This value, of
course, is completely dependent on the previously discussed externally
measured quantities. However, it is felt by this author that this,

-63-
rather than deviator stress, is the fundamental quantity. The
interaction of the obliquity ratio as a measure of the developed
resistance to shear displacement and the developed porewater pressure
give the sample its ability to resist the observed deviator stress.
The following observations summarize the effect of strain
rate on the obliquity ratio. The discussion will again minimize the
results of the one test representing the intermediate rate.
1. At failure (maximum deviator stress) the obliquity ratio
is lower in the test with the more rapid strain rate. The obliquity
ratio typical of t1 = 1 minute at (,a- -3)m is 2.55 (represent-
ing f of 250 55') and that typical of t = 500 minutes at
-3a is 2.83 (representing a = 280 35').

2. At any strain less than 3.8%, the obliquity ratio is

lower in a typical slow test (t1 - 500 minutes). At 3.8% strain the
obliquity ratio is the same in both the typical t1 = 500 minute and
t = 1 minute test. However, the scanty data at intermediate strain
rates suggests that this "crossover" point is not unique to all strain
rates.
3. At any strain rate greater than 3.8% the obliquity ratio
is higher in a typical slow test. (t1 = 500 minutes).
4. In a typical fast test, the obliquity ratio peaks at
about the same strain at which the deviator stress peaks. Following
this peaking, the obliquity ratio decreases with further strain. In
contrast, no distinct peak in obliquity ratio occurs in the slow
test. The obliquity ratio levels off at quite large strains ( > 10%).
5. If, at approximately 4%, the strain rate in a slow test
(i.e., t 1 = 500 minutes) is stepped, the obliquity ratio shows an
immediate jump to a level higher than typical of the post-step strain
rate, followed by a tailing off toward a level compatible with the
new strain rate.

-64-
 6. The same behavior can be expected if the strain rate is
stepped from the slow (i.e., t1 = 500 minutes) to the fast (t 1 = 1 minute)
rate at any strain. The obliquity ratio shows a jump followed by a
subsequent increase or decrease as straining continues to a level com-
patible with the post-step strain rate. However, in Figure 20, the strain
rate steps at less than 3% strain produced obliquity ratios higher than
would have occurred in a typical steady rate test having the post-step
strain rate. This can be seen in the rather wide variation between
these curves (i.e., obliquity vs. strain for steps in strain rate at
less than 3% strain) and the curve given for t1 = 500 minutes, if a
"crossover" in obliquity were expected at 31 - 4% strain for steady
rate t 1 = 1 minute and t1 = 500 minute curves.

E. Discussion - Effect of Strain Rate on the Behavior of Normally

Consolidated Samples at Small Strains
In this section is presented a discussion of the observed
behavior of the normally consolidated samples at strains less than 1%.
The key to the explanation of this behavior proposed by this author
lies in the assumption that "bond" destruction or destruction of the
interparticular forces resisting a shear deformative at these small
strains is minor.
At small strains, the pore pressure is seen to be a function
of the magnitude of strain and appears to be independent of the deviator
stress. An admittedly crude model of this behavior is presented in
Figure 28. As a shear displacement of particle A relative to particle B
is effected, rotation of particle C must occur if the bonds at ( and (2)
are not broken. Thus, particle A must move closer to particle B induc-
ing a pressure in any fluid filling the intervening space. If the bonds
at ( and are not broken, it would seem that this induced pore-
water pressure would be independent of the shearing stress (C) neces-
sary to effect the displacement, and be a function only of the tendency
toward change in spacing of particles A and B, (h 1 -h 2 ), which is in turn
a function of the horizontal or shear displacement. This is the
observed behavior in terms of porewater pressure at small (i.e, (1%)
strains.
If the "hinging" action of the bonds Q and ® is assumed
to incorporate elements of "viscous" behavior, the observed behavior
in terms of deviator stress can be explained. By "viscous" behavior
is meant here behavior at the hinge such that the moment necessary
to effect a certain angular rotation at the hinge is a function of
the rate at which this rotation is caused to occur. Thus a greater
shearing force is necessary to effect the same shearing displacement
in a short time than in a longer time.
In summary, the observed behavior of the normally consolidated
samples at small strains can be explained if it is assumed that:
1. The majority of the bonds between particles remains
unbroken.
2. Some sort of a "structural viscosity" in the relative
shearing displacement as modeled by the hinges in Figure 28 exists.
3. The pore pressure is a function of the tendency toward
change in relative particle spacing (h 1 -h 2 ) as shown in Figure 28, at
constant volume.
4. Porewater migration or interval volume change plays a
relatively minor role.

F. Discussion - Effect of Strain Rate on the Behavior of Normally

Consolidated Samples at Large Strains
This section discusses and presents the author's explanation
of the observed behavior of the normally consolidated samples at strains
in excess of 4%. Here the key to the author's explanation of the observed
behavior lies in the assumption that bonds between adjacent particles
have been broken (they may be constantly breaking and reforming), and
that this type of interparticular behavior described in the previous
section as applying at small strains plays a relatively minor role.

-66-
 In Figure 28 (b) is presented a highly idealized model
representing the author's conception of the behavior at large strains
in the normally consolidated samples. The rotational moment carrying
capacity of' the bonds (l) and y) has been destroyed by the strain up
to this time. Further displacement of particle A relative to particle B
occurs by translation at the points of interparticular proximity 1) and C
It would seem that the resistance to displacement at these points would be
very much a function of the normal stress effective at these points of
contact. Since this normal effective stress is a function of the induced
porewater pressure and since the induced porewater pressure in the tests
in which the strain rate was rapid was lower than in slower tests, it
follows that the shear stress necessary to produce an additional shear
displacement would become greater as the strain rate is increased, as
has been observed. Therefore, the greater strength in the faster tests
can be attributed to the lower porevater pressure.
This increased strength occurs in spite of an observed decrease
in obliquity ratio, which is a measure of the "coefficient of friction"
active at points 1 and (D2 in Figure 28 (b), as the strain rate is
increased. This phenomenon can be explained by postulating that the
resistance of points and ®
against translation and in comparison
to the normal stress acting to maintain contact at these points is a
function of the distance of separation at these points (which would be
very small). However, as the rate at which the induced displacement
occurs increases, the spacing at the points of proximity could increase
due to an action somewhat akin to planing of a boat. Hence, the obliquity
ratio would be expected to be lower in the more rapid tests.
This observation of decrease of obliquity ratio with increase
in strain rate can also be attributed to the same phenomena causing the
tpre-stress effect" which has been postulated as a portion of the
explanation of an observed higher o or angle of shearing resistance
in undrained triaxial compression tests as compared to drained tests.
(CASAGRANDE and WIISON, 1953). This effect can best be explained

-67-
- a

by accepting as fundamental the Hvorslev Parameters of shear strength

behavior and by additionally assuming that such parameters are inde-
pendent of the rate of strain. The validity of the second assumption
will be discussed in Chapter IX. The first assumption leads to the
adoption of the line A-B in Figure 28 (d) as representing the relation-
ship between the shear resistance on the failure plane at failure
(t ) and the normal effective stress on the failure plan at failure
( ) at a particular water content. The maximum pressure in terms of
effective stresses acting on the potential failure plane occurred at the
time of consolidation or possibly early in the test. (Actual average
stress paths are shown in Figure 10). Point in Figure 28 (d)
represents failure in a slow test while point ) is the corresponding
point from a fast test. Since the porewater pressure is higher in the
slow test, point ® is nearer the Z( axis than is point 0 A line
drawn from the origin through either of these points would represent
the standard failure envelope in terms of effective stresses. (OC and OD).
It can be seen that since the effective stress on the failure plane in a
fast test is higher due to a lower porewater pressure, the obliquity
ratio (measured by in Figure 20 (d)) will be lower.
The fundamental relationship in the effect of strain rate on
the observed behavior of the normally consolidated samples at strains
in excess of 4I% is seen to be the relationship between the rate of
strain and the porewater pressure at any strain. To explain the higher
porewater pressure in the slower tests, a dilatency effect will be
introduced using an idealized model similar to those of Figure 28 (a)
and 28 (b). This is shown in Figure 28 (c). It can be postulated that
at any instant, statistically, some of the displacement must be occurring
in situations as shown in Figure 28 (c). If this displacement is occurring
slowly, translation at the points of particular proximity can occur. How-
ever, if the resistance to such translation is of a viscous nature, a more
rapid displacement of particles would lead toward a tendency of rotation

-68-
a

of the diagonal particle. If the resulting tendency toward increase in

the particle spacing, h, is prevented by the porewater, a lower pore-
water pressure will be induced. Consequently, the more rapid the dis-
placement of the particles, the lower the expected porewater pressure.

-69-
 Chapter VI

OVERCONSOLIDATED SPECIMENS WITH UNIFORM STRAIN-RATE

A. Introduction
This chapter presents and discusses the results of a series of
triaxial compression tests upon overconsolidated samples of the Vicksburg
clay. These tests were performed at uniform rates of strain representing
times to 1% strain varying from 45 seconds to 500 minutes. In some of
these tests the strain-rate was "stepped" as was the case with the tests
on normally consolidated samples described in Chapter IV. It was the
2
original intent to prepare these samples by consolidation to 120 lb/in.
and rebound to 71 lb./in. . Thus an overconsolidation ratio of 16 would
be attained. However slight inaccuracies in gages (subsequently cali-

Kj brated)led to OCRs varying from 15b to 19. (Test OC 9-S had OCR - 13.
However, this test is not utilized in any analysis due to other dif-
ficulties as will be described). The OCR of each test is listed in
Table VI-l. The data regarding the behavior of these samples subsequent
to the change in rate of strain is described in Chapter VII. However,
data from the earlier portions of these tests, prior to the "step" in
strain rate are included with the data from the tests run to completion
at a constant strain rate and are included in the summaries appearing
in this chapter.

B. Laboratory Test Program - Overconsolidated Samples - Uniform Strain-Rate

This section presents, in summary fashion, a listing of the
tests performed upon the overconsolidated samples. Complete plots of
stress versus strain, pore water pressure versus strain obliquity
versus strain and stress vector plots from individual tests are
presented in Appendix B. Also included in Appendix B are brief
descriptions of each test, including any information judged to be
necessary to completely document the tests.

-70-
 A total of 11 tests were performed on the overconsolidated
samples. The setup of these tests is described in Chapter II as is the
testing procedure. Tests OCl-F and OC6-s were on samples from Batch A
as described in Chapter II, Section C. Tests OC2-F, 0C3-F, 004-F, OC5-F,
OCT-S, oc8-s, and 009-S were performed on samples from Batch B while
OC10-SF, and OC11-SF were performed on samples from Batch C. There is
little reason to suspect any gross deviation in the properties of the
three batches of soil. However, the small differences which did exist
may have considerable effect on the observed results. The nature and
cause of the strain rate effect in the overconsolidated samples are
rather elusive and hinge on small differences rather than on the more
pronounced differences noted in the normally consolidated samples.
Considerable difficulty was experienced in the performance
of Test OC9-S. This test was meant to be an extremely slow test, and
was performed using Cylerex as the cell fluid. The results are
obscured by seating difficulty in the piston to top cap seat and in
the pedestal to internal load cell seat. In addition, the electricity
was cut off to the load frame sporadically due to a faulty connection
in the wiring. The results of this test are discussed, where pertinent,
in light of conclusions drawn from the other tests and are not used to
influence the conclusions.
All tests with the exception of Test 008-S were performed
upon samples encased in 2 rubber membranes separated by a thin layer
of silicone grease. Test OC8-S, however, was performed upon a sample
having 2 membranes, but without the benefit of a silicone grease
layer. Since the sample was assembled under water, there was a film
of water trapped between the two membranes and the effect regarding
membrane leakage was to nullify the advantage to be gained from two
membranes.

-Tl-
C. Presentation of Results - Overconsolidated Samples - Uniform
Strain Rate
The following several sections present the results of the
tests on the overconsolidated samples tested at a uniform strain rate.
These results are presented by describing the effect of strain rate on
observed stress versus strain behavior, pore water pressure versus strain
behavior and conventional parameters of stress versus strain behavior
deduced from these observations. (i.e., obliquity ratio and various
pore water pressure parameters.) In these sections it is not the
intention to pinpoint causes for the observed effect of rate of
strain on these quantities, but only to describe the observed effect.
There is considerably more scatter in the data obtained from
the tests on the overconsolidated samples than was the case with the
normally consolidated samples. This scatter is attributed to dif-
ferences in individual samples. As will be pointed out in a later
section, migration of pore water to planes or zones of failure result-
ing in dissipation of, in this case, negative (with respect to back
pressure) pore water pressure and thus tendencies toward internal
volume changes in the samples were of paramount importance. The
effect of strain rate on this migration of pore water probably leads
to the majority of the observed effects of rate of strain on the stress
versus strain behavior of these samples. Thus, as will be pointed out
in following sections, the overall picture of the effect of rate of
strain on stress versus strain behavior in the overconsolidated samples
is obscured by the effect of rate of strain on the internal migration
of pore water and the effects of this internal migration on observed
behavior.
The cause of pore pressure gradients within individual test
specimens is found in the stress distribution occuring as a result of
the nature of the test iteself. In analysing the test results, a
uniform distribution of stresses is assumed. This assumption would
be valid if the sample pedestal and top cap of the triaxial test

-72-
 I

apparatus were completly frictionless. However, this is far from the

case. The resulting phenomena have been described in terms of a re-
arranged pore pressure equation by Bishop et al (1960) and this has
been presented in Chapter III, Section D.
The following description of this effect, while crude, is
presented to point out the importance of internal pore pressure gradients
in overconsolidated samples and as such will be repeated and elaborated
upon in a later section. In addition, it is presented to point out that
considerable scatter in data regarding the effect of rate of strain
observed in these tests is probably due to differences in individual
test specimens.
In a triaxial sample of saturated soil undergoing testing,
the portions of the sample immediately adjacent to the stone on the
pedestal and immediately adjacent to the specimen top cap are restrained
from lateral expansion due to friction at the contact surfaces. In the
no-lateral-strain situation, the pore water pressure induced due to an
application of vertical load can be expected to be essentially equal to
the vertical stress. Thus, in the thin zone adjacent to the ends of the
sample, the pore pressure response to an instantly applied axial load can
be expected to be essentially equal to the deviator stress a- - Q-3
applied to the sample as a whole. In the central portion of the sample
a uniform stress distribution in a horizontal plane is a reasonable
assumption and here the pore pressure response to an instantaneous load
is A &(c- - cr~3 . If A = 1, the induced pore water pressures at
the sample mid-height and at the base are similar. In normally
consolidated samples of clay soils, A approaches 1 and internal pore
pressure gradients and resulting internal volume changes are of a
minor effect. However, in overconsolidated samples with Ae4O. it can
be seen that a considerable pore pressure gradient within a test
specimen is possible.

-T3-
 Thus, in overconsolidated samples, if a test is performed
rapidly, the midplane pore pressure is the value to use in determining the
effective stress controlling the shear resistance, whereas the base pore
pressure would not. In this situation, little volume change is possible.
If a test is performed slowly enough to allow pore pressure equalization,
the measured pore pressue is inappropriate to the assigned Overconsolidation
Ratio and equalization of pore pressures has led to internal volume changes.
At intermediate strain rates intermediate degrees of pore pressure equali-
zation and internal volume change have occurred. The assignment of appro-
priate strain rates to the extremes (i.e., no migration of pore water to
complete equalization of pore water pressure) hinges on the time-compression
parameters, c (compression) and cv (rebound) applicable to the test speci-
men. These parameters are in turn a function of volume change behavior
and permeability. It is felt that only minor differences in volume change
behavior (as described by a V or My) occurred either between batches or
between individual test specimens in this test series. However, the
permeability of the samples is influenced so greatly by small differences
in grain size composition and particle structural arrangement that
considerable differences from batch to batch and sample to sample are
possible. Therefore, the observed data scatter is felt to be attributable
to differences in permeability among test samples.
To sum up, it has been attempted in this section, to prepare the
reader for the presentation of results which follows by pointing out that
the observations are influenced more by the effect of rate of strain on
internal pore water migration than by the effect of rate of strain on
fundamental parameters of stress versus strain behavior, and that the
importance of differences in permeability on the time rate of this
pore water migration has lead to considerable data scatter.

_74-.
D. Effect of Rate of Strain on the Stress Versus Strain Behavior of
Overconsolidated Samples

This section presents, in a general form, the observed results

from the series of tests on the overconsolidated samples performed at
uniform strain rates. Complete stress versus strain and pore water
pressure versus strain curves for each test appear in Appendix B.
All of the pertinent observed results are presented in
Table VI-1 following:

Table VI-1

RESULTS OF TESTS ON OVERCONSOLIDATED SAMPIES

, ax. .-A
To 1010 . T (a - a-) (a /3)
W/No) max, Ma ax. .ay.- Max- F- lulie
Ps-r ST.i * 6b./i l .a)x

OCl-F 3 28.5 23 -7.0 +3.8 -11.0 7.5 3.3 8.5 1.9 29.3 A 19
0C2-F 1.7 32.6 25 -4.1 +1.6 -9.6 4.48 3.45 7.2 1.9 30.0 B 17
Q03-F -7 32.6 23 -5.8 +0.1 -lo.6 4.15 3.45 8.4 2.2 28.6 B 16
0C4-F 1.5 33.5 24.7 -8.0 +0.1 -16.5 4.10 3.25 11.1 1.75 29.8 B 152

OC5-FS -75 32.5* 21.5 -6.4 +2.5 -8.0 6.4 3-5* 8.5* 1.3 29.4 B 16
oc6-s 18 27.3 17.0 -5.8 +2.6 -7.0 5.7 3.2 7.5 2.3 29,1 A 18
007-S 30 27 17.2 -3.2 +3.7 -4.8 4.7 3.34 5-2 1.4 30.3 B 17
oc8.-s 250 24.3 19 -1.2 +1.7 -2.2 4.65 3.95 5.0 2.1 29.9 B 16
__o9s 300 -- ____ -_--- __-_?_
-1. -- 28. B 13

oClO-SF 500 28.2* 20.5 -5* +0-6 -7.5 4.30 3.5* 6.6* 2.2 29,4 C 17
OCll-SF 500 26.5* 19.2 -4* +1.1 -7.0 4.28 3-5* 6.0* 2.0 29.7 C

*NOTE: Data extrapolated from tests with stepped strain rate.

-75-
I

In Figure 29 are plotted all data from tests on the over-

consolidated samples perfoned at strain rates having times to 1%
strain short enough to be grouped as "fast" tests. Average stress versus
strain and excess pore water pressure versus strain are drawn using these
plotted points. Test OCl-F was the only rapid test run on a sample cut
from Batch A soil and hence is excluded from consideration in this plot.
In Figure 30, all data from tests considered to be "slow"
tests are plotted. Data from tests OC10-SF and OCll-SF were almost
coincident and were averaged. Both of these tests were tests in which a
step increase in strain rate was imposed at about 4% strain. Consequently
the curves beyond 4% strain are extrapolations of observed data.
From Figure 30 the following conclusions can be drawn:
1. The stress versus strain curve from fast tests rises more
steeply and attains a higher deviator stress level at any strain than do
stress versus strain curves from any slow test.
2. The pore pressure versus strain curve from fast test does
not rise initially to quite as high positive pore water pressure levels as
do curves from slow tests. Following attainment of maximum positive pore
water pressure, the pore pressure in the fast tests falls off more rapidly
with strain, decreases to lower negative pore water pressures and con-
tinues decreasing to larger strains than does the pore water pressure
in the slow tests.
3- Fast tests attain maximum deviator stress at larger strains
than do slow tests.

E. Effect of Strain Rate on Maximum Deviator Stress, Strain at Failure

and Mode of Failur. Overconsolidated Samples - Uniform Strain Rate

This section summarizes the effect of strain rate upon the

maximum deviator stress obtained, the strain level at the attainment
of maximum deviator stress and the mode of failure observed in the
tests on overconsolidated samples at uniform strain rates. These

-76-
m

observations are not to be construed as conclusions regarding any

fundamental soil properties but are, as was pointed out in Section C
of this chapeter, observations of the behavior of laboratory samples
of a particular size.
In Figure 31 are plotted the maximum deviator stresses
obtained in all of the tests on the overconsolidated samples. The
points representing tests 0010-SF and OC11-SF are plots of values
obtained by extrapolation of the stress versus strain curves follow-
ing the step in strain rate. The solid line is the author's impression
of the probable relationship between maximum deviator stress and rate of
strain. The exact location of the intersection of the two tangents is
unknown and probably hinges on local permeability deviations from sample
to sample. In addition, the results of a series of tests on the same
soil but at a slightly lower OCR are presented. (NASIM 1961). In the
range of strain rates of Nasim's tests, the relationship between
deviator stress and strain rate roughly parallel the author's.
In Figure 32 are plotted strains at maximum deviator stress
versus time to 1%strain as a measure of strain rate. There is seen
to be considerable data scatter. However, the solid line represents
what appears to be the general trend of the data and is the author's
impression of the probable relationship between rate of strain and
strain at failure (i.e., maximum deviator stress). It will be seen
from this figure that as the rate of strain decreases (time to 1%
strain increases), the strain necessary to attain maximum deviator
stress becomes less.
In Table VI-2 are listed the modes of failure of each test
and, where failure planes occurred, the strain level at which these
were first noted. All three batches of soil are represented in this
data and the formation of failure planes is probably a phenomenon
very sensitive to local variations in samples. Thus no quantitive
conclusions can be drawn from this data. However, it seems obvious

-77-
that there is less tendency toward formation of failure planes in the
more rapid tests. In fact, no failure planes were noted at all in the
fastest tests (OCs-F, 0c4-F & 0C5-FS) while quite distinct failure
planes were formed at relatively low strains in the slowest tests.

F. Relationship Between Rate of Strain and Pore Water Pressures,

Overconsolidated Samples, Uniform Strain Rate

Presented in this section are all data obtained regarding the

pore water pressures recorded in this series of tests. In all of the
tests on overconsolidated samples, the pore water pressure was measured
at the sample mid-height employing the probe described in Appendix A.

Table VI-2

MODES OF FAIIURE-OVERCONSOLIDATED SAMPIES

Test Time to 1% Mode Strain at which

No. strain of! failure plane
(minutes) failure ()
001-F 3 Bulging Failure Plane 17%
002-F 1.7 Bulging Failure Plane 13%
003-F .75 Bulging --
004-F 1.5 Bulging none at 32%
OC5-FS .75 Bulging
006-s 18 Failure Plane 10%
OC7-S 30 Failure Plane 9-2
008-s 250 Failure Plane 7 3/4%
009-S 3000 Failure Plane ?
0010-SF 500 Failure Plane 6%
0011-SF 500 Failure Plan 10%

-78-
In addition, the pore water pressure was measured at the base of the
sample in tests OCl-F, 002-F, OC3-F, oc4-F, OC5-FS, oc6-s, OC7-S and OC8-s.
In Figure 30 are plotted summary curves of pore water pressure
versus strain representing all tests. The average "fast" curve represents
the four fast tests as plotted in Figure 29. A study of Figure 30 leads
to the following conclusions:
1. The maximum positive pore water pressure is attained at about
the same strain regardless of strain rate.
2. A higher positive pore water pressure is attained at these
low strains in the slower tests. However, the actual magnitude of the
2
difference is only on the order of 1 lb./in.
3. The pore water pressure goes negative with respect to back
pressure at lower strains in the faster tests. (i.e., more strain is
necessary to develop a negative pore water pressure in the slower tests).
4. The pore water pressure in the faster tests attains a much
more negative value with respect to back pressure than do the pore water
pressures observed in the slow tests.
5. At a certain strain, the pore water pressure in the slower
tests stops decreasing and tends to remain at a constant level. However,
the observed pore water pressure in the fast tests continues to become
more negative to quite large strains.
The relationship between midplane pore water pressure and strain
rate is shown in Figure 33. Points representing pore water pressure at
maximum deviator stress and points representing the maximum pore water
pressure obtained in the tests are plotted. Since the maximum negative
pore water pressure attained in any test is a function of the strain at
which shear plane formation occurs and is a function of the strain level
at which testing was discontinued, the considerable scatter in data
representing this quantity is to be expected.
The data representing pore water pressures at maximum deviator
stress, as plotted in Figure 33 are remarkably consistent. The solid line
in that figure represents the author's impression of the relationship

-79-
between pore water pressure at maximum deviator stress and strain rate
for the overconsolidated samples. From this figure, the conclusion can
be drawn that a more negative (lower) pore water pressure is attained
as the rate of strain increases.
The relationship between mid-height pore water pressure and
deviator stress is shown in Figure 34. The values represent typical or
average values with tests 0010-SF and OC11-SF averaged for the slow
values. (See Figure 30). Using the same "average" curves, the pore
pressure parameter A, (A )is plotted as a function of strain
in Figure 35. The discontinuity in the curve representing the slow tests
and occurring at about 4% strain is felt to be the start of the shifting.
of the strain to a zone of local yielding and thus represent the start of
failure plane formation. It will be observed in this figure that no such
discontinuity occurs in the curve representing the fast tests, and the A
factor is gradually decreasing at a decreasing rate throughout the progress
of the test.
In Figure 36, the relationship between the pore water pressure
parameter A and strain rate as given by the midplane pore pressure is
shown. It can be seen in this figure that the A factor tends to decrease
slightly (become more negative) as the rate of strain increases.

G. Pore Water Pressure Gradients and Internal Volume Change Phenomena,

Overconsolidated Samples

In this section are presented all the data regarding the

differences in pore water pressure at various heights (base versus mid-
height-needle pore water pressures) and the distribution of moisture
content throughout individual test specimens obtained as a result of
this research.
Pore water pressure was measured at both the sample base and
the sample mid-height in tests 001-F, OC2-F, 0C3-F, oc4-F, OC5-FS,
oc6-S, OC7-S and 008-S, and these data are plotted with each individual

-80-
test in Appendix B. Typically, in a fast test, the mid-height pore water
pressure rose more rapidly to its maximum positive value (lower than the
base maximum positive value). After attaining its maximum positive
value, the mid-height pore water pressure fell rapidly to a negative
value (with respect to back pressure). The base pore pressure rose more
slowly to a higher positive value than the midplane value, fell less
rapidly, going negative at a considerably greater strain than was the
case at midplane. At any strain level following an initial crossover
at to 3/4% strain, the midplane pore pressure is lower than the base
pore pressure. Tests OC3-F and 004-F illustrate this typical behavior
quite well.
The slow tests, typified in this case by OCT-S and OC8-S, show
midplane pore pressure higher than the base pressure at strains less than
lj$ although a portion of this difference existed initially in both tests
and is probably anomalous. In test OC8-S, which was slower than OCT-S,
the midplane pore water pressure became slightly lower than the base pore
water pressure, but the two became coincident at 8% strain and remained
so to the completion of the test.
In Figure 37 are plotted all the data available in this test
series regarding pore water pressure gradients, or a difference in base
and mid-height pore water pressure. It can be seen that there is a
good deal greater tendency for the development of gradients in the over-
consolidated samples than was the case with the normally consolidated
samples. This plot also demonstrates the relationship between base
and mid-height pore pressures and rate of strain.
In all tests, with the exception of 003-F, post shear water
contents were determined from six levels in each sample as shown in
Figure 38. The preshear, post consolidation distribution of water
content is, of course, unknown. However, Nasim (1960) in tests on the
same clay found an essentially uniform distribution of post consolida-
tion water contents in his overconsolidated samples. This assumption

-81-
seems justified when the data from tests 001-F and OC2-F are con-
sidered. In these tests, sheared rapidly and quickly dismantled,
an essentially uniform water content through the sample was observed.
The water content data from the slow tests all fall within
the shaded area shown in Figure 38. In addition, the water content
distribution from test 0c4-F in which considerable time elapsed following
testing before the water contents were determined, fell within the shaded
area. On the basis of the assumption of essentially uniform preshear
water content, there is seen to be considerable evidence of migration of
water toward the middle portion of the sample from the ends with a result-
ing increase in water content in the middle portion and decrease at the
specimen and regions.

H. Effect of Rate of Strain on Obliquity Ratio (r'/&3 ) Overconsolidated

Samples

This section presents and discusses the effect of strain rate

on obliquity ratio as deduced from measured deviator stress and pore water
pressure. In Figure 39, the / ratios as deduced from Figure 30
are plotted against strain. The value of maximum obliquity which occurs
between 1% and 2% strain is felt to be a good average as shown on Figure 39.
However, as the results of individual tests are studied, this value will
be seen, in the faster tests to exhibit considerable scatter. This
maximum ' is quite sensitive to the maximum positive pore water
pressure, and can change radically within the limits of the accuracy of
pore pressure measurements in this region. It can be seen clearly from
this figure that, while the faster tests, at any strain level, exhibit
a higher obliquity, the ratio at maximum deviator stress is
only slightly higher in the faster tests due to the attaining of this
condition at a larger strain. This trend is illustrated in Figure 40
where, in addition to the points from the average curves, data from
each test is also plotted.

-82-
 In this figure are plotted both maximum obliquity ratio
versus strain rate and obliquity ratio at maximum deviator stress
versus strain rate. Straight lines representing possible relationships
between obliquity ratios and logarithm of strain rate (time to 1% strain)
have been drawn. Considerable scatter of data casts doubt on the rela-
tionship between maximum obliquity ratio and strain rate. However,
obliquity ratios at maximum deviator stress seem to form a consistent
pattern indicating a slight increase in obliquity ratio at failure as
the strain rate increases. This trend is opposite to that exhibited by
the normally consolidated samples.

I. Summary and Conclusions: Overconsolidated Samples - Uniform Strain Rate

The following conclusions drawn from observations of the behavior

of the overconsolidated samples seem justified on the basis of the data
available:
1. At any strain level, a faster test exhibits a higher
deviator stress than a slow test.
2. Within the range of strain rates considered in this
research an increase in maximum deviator stress in the more rapid tests
of 20% over the slower tests occurred.
3. At any strain level, a lower pore water pressure was
recorded in the more rapid tests. The pore water pressure at failure
in the faster tests was much lower than that in the slower tests, partly
due to the fact that the faster tests "held together" to much greater
strains.
4. At any strain level, a higher obliquity ratio in terms of
effective stresses ( o§ci) existed in the faster tests.
5. Thus the higher deviator stress attained in the faster
tests can be attributed to a more negative pore water pressure and a
slightly higher obliquity ratio at failure.

-83-
 6. The above effects are observations based on the behavior
of a particular size laboratory specimen. Fundamental behavior patterns
were obscured by pore pressure gradients between the base of the sample
and midplane. equalization of which led to internal void ratio changes
in the slower tests.
7. Associated with the above phenomenon, the faster tests
tended to fail by bulging, while quite distinct shear planes were
observed in the slower tests.
 Chapter VII

OVERCONSOLIDATED SPECIMENS WITH STEPPED STRAIN-RATE

A. Introduction
Presented in this chapter are the results of a series of
three tests on overconsolidated samples, having overconsolidation ratios
of between 16 and 18, which were subjected to a step in strain rate,
either increase or decrease. Tests OC5-FS, 0010-SF and OC11-SF are
included in this series.

B. Sample Preparation, Consolidation and Testing

The specimens tested in this series of tests were prepared,
set up for testing and consolidated in the triaxial chamber in exactly
the same manner as the specimens described in the previous chapter.
In test 0C5-FS, the sample was brought to 5.6% strain at the most
rapid strain rate employed in this research, t1 = 3/4 minutes. At
this strain level, the strain rate was decreased to t 1 = 17 minutes.
This test was prepared from a Batch B sample.
Tests 0010-SF and OCll-SF were prepared from Batch C samples.
In these tests, the samples were brought to approximately 4% strain
(4.3% in OC10-SF and 3.9% in 0011-SF) at a strain rate of t = 500
minutes. At this strain level, the strain rate was instantaneously
increased to a strain rate of t 1 = 3/4 minutes in both tests. In
test OC10-SF, shortly following the step in strain rate, the- clutch
on the load frame began to slip, so the actual strain rate is unknown.

C. Presentation of Data

The data from these three tests appears in Appendix B, in

Figures B-43, B-44, B-51, B-52 and B-53, and the reader is referred to these
figures during the ensuing discussion. Plots of deviator stress,
pore water pressure and ratio of principal effective stresses

-85-
versus strain are presented. In test 0C5-FS plots of both midplane
and base pore pressures versus strain are presented.
In addition, the water content distribution following testing
was determined in all three samples. This information is also presented
on Figures B-43, B-51 and B-52 in Appendix B.

D. Test 0C5-FS with Step Decrease in Strain-Rate

In test OC5-FS, the strain rate was instantaneously decreased
from t = 3/4 minutes to ti = 17 minutes at 5.6% strain. In this test,
pore water pressure was continuously measured at both the sample base
and the sample mid-height (via the porous probe),
From the start of the test, up to the strain at which the strain
rate was decreased, the sample behaved quite similarly to the "average
fast test" presented in Figure 30. The base pore water pressure at any
strain level was considerably greater than the mid-height pore water
pressure. The mid-height pore pressure went "negative" with regard to
initial back pressure at 2.8%strain, while the base pore water pressure
remained positive until after the decrease in strain rate. Prior to the
change in strain-rate, the deviator stress versus strain behavior was
also quite typical of the rapid rate of strain (t1 = 3/4 minutes).

E. Effect of a Step Decrease in Strain Rate Upon Deviator Stress - Over-

Consolidated Samples

Following the step decrease in strain-rate from t = 3/4 minute

to t 1 = 17 minutes, the deviator stress, which had nearly attained its
peak at the rapid strain rate, fell abruptly from 32 lb./in. to
26.5 lb./in.2, or 17%. Following this drop, the deviator stress rose
steadily to a new peak of 29 lb./in.2 at 9% strain. This is a value
about 11% below the expected peak at the fast rate (t1 = 3/4 minutes)
and occurs at a strain level similar to the expected peaking strain
at the fast rate.

-86-
F. Effect of a Step Decrease in Strain Rate Upon the Observed Pore
Water Pressure, Overconsolidated Samples

Prior to the step decrease in strain rate, the pore water

pressure, as measured at the sample mid-height by the porous probe
described in Appendix A, behaved in a pattern typical of the fast
strain rate. No change in the pattern of pore pressure behavior as
recorded at the sample mid-height was noted following the step decrease
in strain rate, The pore water pressure as measured at mid-height did
not rise to a level typical of a test at t 1 = 17 minutes performed at a
uniform strain rate.
The base pore pressure measurement did change following the
step decrease in strain rate. Prior to the decrease in strain rate,
the base pore pressure was greater than the midplane pore pressure
with the magnitude of this difference increasing. However, following
the step decrease in strain rate, the base pore water pressure dropped
drastically rather quickly and continued falling as straining progressed.
The behavior of the base pore pressure as post-step straining continued
was such as to be gradually diminishing the magnitude of the difference
between it and the mid-height pore pressure. The two measurements ran
parallel, with the base showing 2} lb./in.2 higher from 11% strain until
cessation of testing at 12* strain.
This behavior of base pore water pressure illustrates nicely
the phenomenon described in Section C of Chapter VI. Due to lateral
restraint in regions of the sample adjacent to the base and to the end
cap, a tendency toward a higher pore water pressure exists. If the
equalization of this pore water pressure due to internal flow of pore
water and resulting volume changes were prevented (e.g. by extremely
rapid testing) the measured base pore water pressure could be expected
to approach the magnitude of the applied deviator stress. However,
testing at finite strain rates allows some internal flow of water to
occur causing the measured pore water pressure to be always less

-87-
m

than the deviator stress. At the mid-height of the sample, however,

lateral strains can occur, stress distribution is uniform across the
sample and the pore water pressure will be.A = if the test is
performed rapidly. At slow strain rates, the base and mid-height pore
water pressures will be equal and be greater algebraically than
A(-; -: in overconsolidated samples (or in any situation where A< l).
In test OC5-FS, at the fast strain rate a higher pore pressure
is observed at the base of the sample than at the sample mid-height.
Reduction of the strain rate serves to reduce the magnitude (algebraic)
of the base pore water pressure by allowing time for a greater degree of
internal flow of pore water to occur.

G. Effect of a Step Decrease in Strain Rate Upon the Deduced Principal

Effective Stress Ratio - Overconsolidated Samples

This section presents the obliquity ratio ( , . ) versus

strain behavior deduced from observed deviator stress and midplane pore
water pressure versus strain behavior. It should be emphasized that the
effects of strain rate and steps in strain rate are largely a function
of the degree of internal migration of pore water and thus the true
strain rate effect is largely obscured.
Figure 44 of Appendix B presents deduced obliquity ratio
versus strain behavior from test OC5-FS, In the early portion of the
test the behavior was similar to the expected in a test performed at the
fast (t1 = 3/4 minutes) rate. The exact "typical" behavior of obliquity
at small strains is a rather elusive factor as was previously pointed
out. Following the step decrease in strain rate, the obliquity fell
to a value below that typical of the new strain rate (t 1 = 17 minutes)
at the same strain level. However, the obliquity ratio rose quickly
to a value typical of uniform strain at the new strain rate and
subsequently behaved similarily to test performed at a uniform
strain rate of t 1 = 17 minutes.
H. Mode of Failure of Sample with a Step Decrease in Strain Rate

The mode of failure of sample OC5-FS is quite interesting.

It has been observed in the main tests on overconsolidated samples at
uniform strain rates that failure at slow strain rates occurred at
relatively small strains by formation of a distinct failure plane.
However, failure in tests at fast strain rates was by bulging with no
failure plans being noticed generally at quite large strain levels.
Failure in test OC5-FS was by bulging with no failure plane
being noted at cessation of testing at a strain of 12"* even though the
test was carried on at a slow rate of strain for a sufficient magnitude
of strain to allow failure plane development.

I. Tests OC10-SF and OCll-SF with Step Increases in Strain Rate

This and the following sections of this chapter describe
and present the results of two tests, performed on overconsolidated
samples, in which the strain rate was "stepped" up or increased at a
strain level of approximately 4%. In test OC10-FS the step increase in
strain rate was imposed at 4.3% strain and in test 0C11-FS at 3.9%strain.
These samples were set up and consolidated in exactly the same
manner as the rest of the overconsolidated samples. Testing was insti-
tuted at a slow strain rate of ti = 500 minutes. At approximately 4$
strain the strain rate was abruptly "stepped" to a new rate of
t1 = 3/4 minutes. Following this step in strain rate, the test was
carried to completion at the new strain rate.
Some difficulty was experienced in the performance of test
OC10-SF. Following the step in strain rate, the clutch of the load
frame was observed to slip. Thus the exact post step strain rate is
unknown. For this reason, most of the results presented in the follow-
ing sections are the results of test OCll-SF. Complete stress versus
strain, pore water pressure versus strain and obliquity versus strain
from both tests appear in Figures 51 and 52 in Appendix B.

-89-
U

J. Effect of a Step Increase in Strain Rate Upon Deviator Stress -

Overconsolidated Samples

This section discusses the effect of a step increase in strain

rate upon deviator stress in the overconsolidated samples. Prior to the
step increase in strain rate, the deviator stress versus strain behavior in
both tests 0C10-SF, and OCll-SF was identical and typical of the stress
versus strain behavior of the pre-step strain rate of t1 = 500 minutes.
Following the step increase in strain rate, the deviator stress jumped
to a value in agreement with the deviator stress that would have been
attained in a test at a uniform strain rate equal to the post-step
strain rate at the same strain level. Following this, the deviator
stress in test 0C11-SF fell slightly and remained below the deviator
stress at any strain past the strain level of the step increase. The
same behavior with a much more pronounced dip in deviator stress was
noted in test OC10-SF, but this excess dip in deviator stess is probably
attributable to the load frameclutch slippage as discussed in the
previous section.

K. Effect of a Step Increase in Strain Rate Upon Developed Pore

Water Pressure - Overconsolidated Samples

One of the more interesting phenomena observed in this series

of testing was the effect of a step increase in strain rate upon the
observed pore water pressure versus strain pattern. Since there is
some question regarding the results of test OC10-SF due to suspected
clutch slippage, only test OCll-SF will be discussed here. The
reader is referred to Figure 52 in Appendix B.
Prior to the step increase in strain rate, the developed pore
water pressure versus strain pattern was typical of that expected in a
test at a uniform strain rate of t1 = 500 minutes. Following the step
increase in strain rate, there was no discernable change in the pattern
of pore pressure development. The post step pattern of pore pressure
development was one of continuing decrease in pore water pressure until

-90-
cessation of testing at 10 3/4% strain.
Also presented in the bottom half of Figure 52 of Appendix B
is a plot of the pore water pressure parameter A versus strain before
and after the step in strain rate. The apparent behavior pattern is a
follows: Prior to the step in strain rate, A is decreasing with
becoming less negative and following a pattern typical of the slow rate
of strain (t1 = 500 minutes). Following the step in strain rate to
t+1 = 3/4 minutes, the A factor assumes a constant value and holds this
value until reaching a strain level appropriate to this value of A at
the faster strain rate. Following the attainment of this strain level,
the behavior pattern of A related to strain becomes one assumed to be
typical of the rapid rate of strain.
The implications of this observed pore pressure development
will be discussed in the following chapter.

L. Effect of a Step Increase in Strain Rate Upon the Effective Principal

Stress Ratio ( . )- Overconsolidated Samples
3

In this section is presented and discussed the effect of a step

increase in strain rate upon the observed obliquity ratio (&~/jj) versus
strain pattern as deduced from measured deviator stress and pore water
pressure. It should again be emphasized that the following is an observed
behavior pattern of a certain dimensioned test sample and cannot be
construed to represent fundamental soil behavior as will be described
in the following chapter. Here again the emphasis is upon test OCll-SF
due to the suspected clutch slippage of the test frame during test
OC1O-SF.
Prior to the step increase in strain rate, the obliquity ratio
versus strain pattern is seen to be typical of that expected in a uniform
strain rate test of t = 500 minutes (the reader is referred to Figure 52
in Appendix B). Incidentally, the curves of obliquity ratio versus strain
from tests OC10-SF and 0C11-SF are almost coincident. Following the step

-91-
increase in strain rate, the obliquity ratio jumps rapidly to a quite
high level, higher in fact than any value of pre-step obliquity ratio
attained, and much higher than could be considered typical of the new
strain rate. As strain is continued at the new rate, t1 = 3/4 minutes,
the obliquity ratio decreases quite rapidly at a gradually decreasing
rate and approaches a behavior pattern typical of the fast rate.

M. Mode of Failure - Step Increase in Strain Rate - Overconsolidated Samples

Both samples OC10-SF and 0011-SF showed distinct failure planes
at cessation of testing. In test OC10-SF, a quite distinct failure plane
was noted at 6% strain while in test 0011-SF no failure plane was noted
until 9"% strain. Failure plane formation is typical of the slow strain
rate. In test OC1-SF, the step increase in strain rate appears to have
delayed the fornation of a shear plane over the strain level typical of
the slow strain rate.

N. Summary - Effect of Steps in Strain Rate on Observed Behavior -

Overconsolidated Samples

This section summarizes the effect of step increases or decreases

in strain rate upon the behavior of overconsolidated samples:
1. Deviator Stress
(a) A step decrease in strain rate at 5% decreases the
deviator stress to a level below that appropriate
for the post-step strain rate. Following the step,
the deviator stress rises to a level and follows a
pattern typical of the new strain rate.
(b) A step increase in strain rate at 4% strain increases
the deviator stress to, but not in excess of, a value
appropriate to that strain level at the faster rate.
Following the step, the deviator stress follows a
pattern typical of but slightly lower than that of
the post-step strain rate.
~m

2. pore Water Pressure

(a) A step decrease in strain rate at 5% strain appears to
have no effect on the pattern of pore pressure develop-
ment as measurei with the porous probe at midplane.
However, the base pore pressure drops drastically
following the step decrease in strain rate.
(b) A step increase in strain rate at 4% strain appears to
have no effect on the pattern of pore pressure develop-
ment as measured at the sample mid-height.
3. Principal Stress Ratios
(a) Following a step decrease in strain rate at 5% strain,
-3 drops to a value lower than that typical
of the post-step strain rate at the same strain level.
However, the magnitude of C/ rises as strain at
at the slower rate continues, and follows a pattern
typical of the slow strain rate.
(b) Following a step increase in strain rate at 4% strain,
rises to a value far in excess of that typical
of the post-step strain rate, in fact higher than
previously recorded in the test. Following this drastic
jump,"/r 3 rapidly decays as straining continues, and
behaves in a pattern typical of the new, fast strain
rate.

-93-
 Chapter VIII

BEHAVIOR OF OVERCONSOLIDATED SAMPLES

A. Introduction
This chapter sets forth explanations for the observed
behavior of the overconsolidated samples as related to strain rate.
The observed behavior patterns have been described in Chapter VI for
tests at uniform strain rates and in Chapter VII for tests with steps
in strain rate. The final sections of those two chapters present
summaries of the observed behavior patterns.

However, it is not felt by this writer that the behavior

patterns exhibited in the overconsolidated samples are representative
of true soil behavior. Rather, it is felt that the true soil behavior
patterns as influenced by strain rate are obscured by an internal volume
change or pore water migration phenomenon. The effect of this migration
and evidences of its occurence are presented in the following sections.

B. Stress Distribution in Triaxial Samples

In the theory of triaxial testing, a uniform distribution of
stresses throughout the sample is assumed. This uniform distribution of
stress would be attained if the specimen base and top cap were friction-
less. However, this is not the case. In fact, at the specimen base and
often at the cap of the specimen, a porous stone is employed that, at
the end of testing, is usually found bonded to the specimen. Consequently,
a uniform stress distribution throughout a triaxial sample is not attained.
Theoretical solutions for the stress distribution in triaxial samples have
been presented by D'APPOIONIA and NEWMARK (1951) and BAIA (1960).
A zone of uniform stresses exists for a distance above and
below a plane passed through a triaxial sample at mid-height, if the
proper length to diameter ratio is maintained. The longitudinal
dimension (thickness) of this zone probably varies throughout testing
and is probably a function of the ratio at any instant. A zone
of restraint, usually assumed to be conical, exists adjacent to the
specimen cap and base and extending into the sample. In fact, these
two zones do not exist per se but represent boundaries of a continuous
stress distribution at any instant.
The stresses on a plane adjacent to the base or top cap vary
radially also. At the periphery of the sample, the minor principal
stress is radial and equal to the chamber pressure and the major
principal stress is vertical and related to the piston load. Thus
in a differential element at the periphery of the triaxial sample
stress conditions like the midplane of the sample exist.
On an element at the axis of the sample and adjacent to the
base, no shear stress between the element and the base exists due to
radial symmetry. Hence the principal stresses are horizontal and
vertical. Since the element is restrained from outward motion, the
major principal stress and the minor principal stress approach each
other in magnitude and are possibly of the same order of magnitude
as the major principal stress at the sample mid-height. With these
boundaries, a radial distribution of stresses occurs. Thus,9 through-
out the triaxial sample, both a radial and longitudinal distribution
of stresses is attained.

C. Effect of Stress Distribution on Pore Water Pressure Distribution

Assuming an increment of piston (axial) load is applied to
a triaxial specimen, the pore water pressure at any point is given by
the basic pore pressure equation t4 = B A 4% +A (6 v; -A 0-3). At the mid-
height of the sample, and throughout the zone of uniform stresses, the
pore water pressure at the first instant of incremental load applica-
tion is thus A a (ag since
-,o-), c- remains constant.

-95-
 However, instantaneously at the planes of the sample base
and top, the pore water pressure varies radially. At the periphery of
the specimen the pore water pressure is again AA (7 - 0~). However,
if the axial stresses are as postulated in the previous section, the
pore water pressure, instantaneously, at the axis on the top and base
planes is nearly B& (o-, - ).
At the base, since a relatively very permeable stone is present,
the radial distribution of pore water pressure is immediately equalized
to some average value. If an initially parabolic distribution of pore
water pressure varying from A 4 ( -) at the periphery to B A (a7-oy)
at the axis, an instantaneous average pore water pressure adjacent to
the base is ( A+ LB)[& or approximately (.5A + .5B)L(0, -o-.
In a saturated sample, with B = 1, this becomes (.5A + 5)
Thus, if A = 1, the instantaneous average pore water pressure at the
plane of the sample base (and top if a porous stone is present) should be
approximately equal to the pore pressure in the uniform stress zone near
the sample mid-height.
However, if A<1, the pore water pressure at the sample base
will be greater than the pore water at the sample mid-height. For example,
if A = A, w at the midheight = i the deviator stress increment, while q
adjacent to the base = .75 times the deviator stress increment. In the
case of overconsolidated soils with A Z 0, the immediate pore water
pressure at the specimen base due to the addition of an increment of
deviator stress can be much greater than the pore pressure at the
specimen mid-height. This discussion should emphasize the fact that
the pore pressure gradient in normally consolidated samples with A near 1
is a minor consideration, while this pore pressure gradient assumes
paramount importance in the overconsolidated samples.
D. Evidences of the Existence of Pore Water Pressure Gradients -
Overconsolidated Samples
Direct measurements of pore water pressure at both sample mid-
height and sample base were made in Tests 001-F, 00C2-F, OC3-F, 00c4-F,
005-FS, oc6-s, OC7-S and oc8-S. The reader is referred to plots of
base and needle pore water pressures as a function of strain for each
of these tests appearing in Appendix B. These tests provide direct
measurement of the pore water pressure gradients as a function of strain
rate.
Figure 37 presents the data from the above tests plotted against
strain rate and thus shows the relationship between pore pressure gradients
and rate of strain. Test 0C8-S, performed at a strain rate of t1 = 250
minutes, appears to have been slow enough to allow complete equalization
of pore water pressure to occur.

E. Migration of Pore Water as a Function of Strain Rate

If a triaxial compression test is performed at a slow rate of
strain, equalization of pore water pressures throughout the sample can
occur. This results frm a movement of pore water from high to low
pressure areas with resulting internal volume changes. This migration
occurs as the test progresses in slow tests, and can occur after cessation
of testing in rapid tests. The result of this migration of pore water
in overconsolidated samples would be expected to be a decrease in water
content near the sample ends and an increase in water content in the
middle region of the sample.
The extent of this internal volume change depends, of course,
on the rate of testing and, as residual pore water pressure remains
following release of axial load, on the length of time elapsing from
the end of testing until water content determinations are undertaken.
F. Evidence of Internal Volume Change - Overconsolidated Samples
During the course of this test series, considerable evidence
of migration of pore water leaing to internal volume change was obtained.
The pore pressure gradient data, of course, shows the tendency for
internal volume change. In addition, as was the case with the nmally
consolidated samples, each test specimen was sliced into six horizontal
segments following each test. 'This was done as rapidly as possible, the
elapsed time being on the order of 5 minutes.
One missing bit of evidence is the post-consolidation, pre-
shear water content distribution. NASIM (1961), on the same soil with
an OCR of 16 found essentially a uniform distribution of water content
at the completion of consolidation. Test 002-F, which was one of the
fastest tests, was disassembled particularily rapidly. In this case a
uniform distribution of water content was obtained.
All of the post-shear water content data are plotted in Figure
38. In Figure 38 (a), the data from all of the fast tests are presented
and in Figure 38 (b), all of the data from the slow tests are presented.
If a uniform pre-shear water content distribution is assumed, there is
seen to be sae evidence of migration or internal volume change in the
fast tests and everhIelming evidence of internal volume change or
migration of pore water in the slow tests.

G. Relationship Between Failure Plan Formation and Pore Water Pressure

The picture of internal volume change or pore water migration
is further complicated by the process of failure plane formation. In
this series of tests it was noted that all of the overconsolidated
samples with the exception of 005.FS failed by shear plane formation
In addition, evidence from the M.I.T. laboratory and other soil mechanics
laboratories shows that moat heavily overconsolidated samples tested at
conventional strain rates fail by shear plane formation.
This process also affects the measured pore water pressures.
The picture of pore pressure developuent in overconsolidated samples
probably is as follows:

-98-
 As the sample is loaded and straining progresses, pore
pressures are developed in the sample. At low strains, less than 2%,
these are positive throughout the sample and little tendency for pore
water migration occurs. As straining progresses, the pore water pressure
at the midplane zone tends to be lower than at the base, for reasons
already presented, and in the slow test, a migration of pore water
starts internal volume change occurring. In fast tests, time is not
available for this to occur.

H. Effect of Pore Water Migration and Internal Volume Changes on Observed

Behavior of the Overconsolidated Samples
One of the major observations of the test series on the over-
consolidated samples is that moisture redistribution occurs in the slow
tests on the overconsolidated samples. This section deals with the effect
of this moisture redistribution on the observed behavior of overconsoli-
dated samples. To do this, recourse will be made to stress vector plots
as shown in Figure 41.
In Figure 41, circle 1 represents the failure condition in a
normally consolidated sample at a certain water content. A line drawn
through the top point of this circle and the origin represents a usual
normally consolidated failure line, Circle 2 is the failure circle
from a test brought to the same water content as the test of circle 1,
but by a different stress history (i.e., overconsolidated). A line
drawn through the top point of the two circles represents a failure
line for the water content of these two tests as shown by HVORSLEV (1937).
A family of lines parallel to this "Hvorslev" line can be drawn for other
water contents.
A "stress vector" representing the progression of 3
versus as the overconsolidated sample is strained
is shown. This stress vector rises, in a pattern dictated by the
manner of mobilization of obliquity of effective stresses on the
potential failure plane and the pattern of development of pore water

-99-
pressure, to the failure line, The stress vector rides along this
"Hvorslev" failure line as the further development of negative pore
water pressure increases the effective stress on the failure plane.
At same point no more increase in effective stress is possible due
either to cavitation of the pore water, migration of pore water to a
failure zone and/or failure plane formation. At this point the stress
vector breaks away from a "Hvorslev" failure line and tends to progress
horizontally until a "failure" occurs.
The relationship of rate of strain to this process is
diagrammed in Figure 42. In this Figure, the assumption is made that
a "Hvorslev" line is non-time dependent. "Hvorslev" lines for several
water contents are shown. Stress vectors from slow and fast tests are
shown. In region A, the exact path of the stress vector is probably
time dependent. In this investigation the data obtained is sketchy as
quite small strains are involved, and thus no conclusions can be drawn
regarding any "structural viscosity" effect on either principal effective
stress ratio or pore pressure development.
The stress vector from a slow test will rise through region A
to a "Hvorslev" line appropriate to the water content in the portion of
the sample in which yielding is occurring. Depending on the degree of
slowness, this may or may not be the initial water content of the sample
as a whole. The sample reaches this line somewhere between 1.5% and 2%
strain. The pore pressure has not, as yet, gone negative.
Upon reaching a "Hvorslev" line appropriate to the water
content at that instant in the zone that is yielding (i.e., the middle
portion of the sample), the stress vector tends to progress along
this "Hvorslev" line until a flow of water internally in the sample
increases the water content. At this point, point 1 on Figure 42,
the stress vector breaks away from the failure line and tends more
toward a horizontal path as the moisture content in the yielding
zone becomes increasingly greater. The stress vector, reaches a peak

-100-
at a situation where the rate of migration of water to the yielding .
zone is the same as the rate of development of negative pore pressure,
and as this rate of migration becomes great enough with respect to
the development of negative pore water pressure, the stress vector
path assumes a downward trend. As failure plane formation retards
any further development of pore water pressure, migration of water
to the failure zone decreases the effective stresses and the stress
vector turns back toward the origin.
In the case of a fast test, the stress vector path may reach
a higher "Hvorslev" line, the one appropriate to the initial uniform
sample water content if the test is rapid enough, as the time necessary
for migration is not available. This failure line will be followed by
the stress vector for a distance until either failure of the soil
structure in the yielding zone has occurred, or time has elapsed
allowing internal volume change to occur. Hence the higher strength
in the faster tests seems probably to be mainly due to less internal
volume change. This results in: 1. The forcing of yielding to
occur at a lower water content (i.e., reaching a higher "Hvorslev"
line) and, 2. the forcing of yielding to occur longer at this lower
water content with the development of more negative pore water pressure
and the resultant increase of effective stresses (i.e., progression
further up a "Hvorslev" line).
That this does occur is demonstrated by the actual stress
vector paths observed in this series of tests. In Figure 43, the
stress vector paths from all of the fast tests have been reproduced
while those from the slow tests appear in Figure 44. These stress
vectors have been summarized in Figure 45. The stress vector paths
from all fast tests fall within the shaded zone. The slow tests
present a close enough grouping of stress vector paths to allow
a single typical stress vector path to be drawn. That the processes
described in this section do in fact occur is felt to be clearly
shown by Figure 45 .

-101-
I. Conclusions
The behavior of the overconsolidated samples has been
discussed in this chapter. Theoretical analyses have been presented
to show that pore pressure gradients could exist in overconsolidated
triaxial compression samples and thus migration of pore water and
consequent internal volume changes during testing could occur.
Experimental evidence of differences in pore water pressures
between the base of overconsolidated triaxial samples and the sample mid-
heights have been presented and the relationship of these pore pressure
differences to strain rate has been shown. Thus the presence of pore
pressure gradients has been demonstrated experimentally. Results of
moisture content determinations have been presented to show that
internal volume changes have occurred, and that this moisture content
redistribution is related to the strain rate.
The effect of this internal volume change tendency upon the
relationship between stress versus strain behavior and rate of strain
has been shown by use of stress vector path plots on "Hvorslev" yield
criteria plots. Thus it has been shown that behavior similar to that
experimentally observed in the overconsolidated samples could be
attributed to internal moisture content redistribution or internal
volume change.
On the basis of this evidence, the conclusion is drawn that
the observed behavior of overconsolidated samples can, to a large
degree, be attributed to internal migration of pore water producing
internal moisture content redistribution. This is a result of pore
pressure gradients which exist in triaxial samples, and to a great
degree in overconsolidated samples, as a consequence of non-uniform
stress distribution (end restraint) and failure plane formation
phenomena.
The effects of rate of strain upon fundamental parameters
of shear strength behavior in the overconsolidated samples are thus
assumed to be obscured to a great extent by effects relative to sample
geometry and the triaxial test itself.

-102-
 Chapter IX

ANALYSIS IN TERMS OF HVORSLEV PARAMETERS

A. Introduction
This chapter presents an analysis of the observed behavior in
uniform strain rate tests and in two tests with a step-increase in strain
rate, in terms of shear strength parameters advanced by HVORSLEV (1937).
These parameters of shear strength behavior were introduced in Section H
of the preceding chapter. In brief, HVORSLEV proposed two parameters of
shear strength, one termed "true angle of internal friction" ( e and
the other termed "true cohesion"(Cc)These represent the angle of shearing
resistance and cohesion intercept of samples tested at varying stress
histories, but identical failure water contents.
The analyses performed in this chapter use a method of
determination of the true friction angle and true cohesion presented
by BISHOP and HENGL (1957).

B. Tests with Uniform Strain Rate

In this section the failure (i.e., maximum deviator stress)
shear strength behavior as related to strain rate is analyzed in terms
of "Hvorslev" parameters for those tests performed at a uniform rate
of strain. Data used in the determination of the appropriate
"Hvorslev" parameters are listed in Table IX-l.
These data are normalized using an equivalent pressure (f)
as suggested by Hvorslev. This equivalent pressure in this case is
the triaxial chamber pressure required to produce the water content
of the test specimen in a normally consolidated sample. The
equivalent consolidation pressure (-f, ) was determined from the
virgin consolidation branch of the Batch B and Batch C consolidation
plots of Figure 3. The equation of this average Batch B and C virgin
consolidation curve is w (%) = 41.8 18.05 (log-f - log 10). Thus

using loge = 41.8 - V .

- as calculated
e 10 18.05

-103-
_______________ _________________ - .- ~----------- ~

Table IX-l

DATA FOR DETERMINATION OF HVORSIEV PARAMETERS AT FAILURE . (1 )ax

0 -
( ) c-

Test 3 2 1 3"max wavg P

e 3 3 1 3 t
P
No. (lb/in2) (lb/in2 ) I (lb/in2 ) (lb/in2) (lb/in2 ) e 2 Pe (minutes) Class

NClF 117.6 82.2 23.4 104.5 62.0 55.6 .532 .394 2

2 58.5 47.5 27.6 61.0 29.4 29.1 .476 .390 1
3 59.0 42,5 29.6 47.4 31.0 28.0 .590 .449 3/4
4FS 59.0 44.5 28.9 51.8 31.8 27.2 .525 .428 3/4
5FS 58.5 42.8 28.8 52.5 29.0 29.5 .563 .408 3/4
6s 60.5 39.5 28.8 52.5 34.4 26.1 .500 .376 20
7S 6o.o 40.0 29.1 50.5 38.2 21.8 .432 .396 350
8S 66.5- 47.4 26.4 71.4 39.0 27.5 .385 .332 1500
93F* 59.0 39.3 27.6 61.2 34.6 24.4 .399 .321 500
lOSF* 61.0 38.5 27.8 59.6 33.0 28.0 .470 .323 500
llsF* 59.8 37.0 28.3 60.6 35.5 24.3 .400 .305 500
143F* 58.5 39.5 28.0 58.2 35.2 23.3 .400 .339 500
16sF* 6o.5 39.5 28.3 60.6 36.2 24.3 .401 .326 500
17SF* 61.0 39.8 27.5 62.0 38.0 23.0 .372 .321 1000
OClF 6.0 28.5 29.3 49.1 -7.0 13.0 .265 .290 .3
2F T.0 32.6 30.0 45.1 -4.1 11.1 .246 .362 1.7
3F 7.2 32.6 28.6 53.8 -5.8 13.0 .241 .302 .75
4F T.5 33.5 29.8 46.2 -8.0 15.5 .336 .364 1.5
5FS* T.5 32.5 29.3 48.0 -6.4 13.9 .290 .339 .75
6s 6.4 27.3 29.1 50.5 -5.8 12.2 .242 .270 18
7S T.0 27.0 30.3 43.4 -3.2 10.2 .235 .311 30
8s T.0 24.3 29.9 45.5 -1.2 8.2 .180 .267 250
10S* 6,6 28.2 29.4 48.6 -5.0 11.6 .238 .290 500
11* 6.T 26.5 29.7 46.7 -4.0 10.7 .229 .283 500

Notes: * = Tests extrapolated

S= Slow
F= Fast
Intermediate
 The resulting normalized data are plotted in Figure 46.
From the resulting plots, the values of 0. and Ce were determined
and are listed on this Figure. In the upper portion of Figure 46 are
plotted (ca -- 3) /2A vs. 1/ for the tests of this research
that could be deemed fast tests. All of these tests had a strain rate
in excess of that producing 1% strain in 2 minutes (i.e., t1 Z 2 mm.).
A moderate degree of scatter exists. A best fit line was drawn through
the plotted data points.
In addition, in the upper half of Figure 46 are plotted data
from three tests (one normally consolidated and 2 overconsolidated)
representative of an intermediate strain rate. The t1 from these
tests varied from 18 to 30 minutes.
In the bottom half of Figure 46 are plotted data from the
slow tests. These data represent all tests having a rate of strain
slower than that necessary to. produce 1% strain in 250,minutes
(i.e., t1 > 250 min.). Many of these data points represent extrapo-
lations of tests involving step increases in strain rate and are thus
noted. Here again a best fit line was drawn through the test data.

C. Effect of Strain Rate on the Hvorslev True Friction and True

Cohesion
Using the data plotted in Figure 46, the magnitudes of the
Hvorslev shear strength parameters from the fast and slow tests were
determined. The magnitudes of f e and ce were also determined from
the intermediate strain rate test data. However, it is felt that the
limited data available invalidates these values. Tabulated in
Table IX-2 following are the results of this determination.

-104-
 Table IX-2

EWECT OF STRAIN RATE

ON 'THE HVORSLEV SHEAR STRENGT PARAMETERS

Strain Rate e _ __

t 2 minutes 110 45' .218

t 250 minutes 110 45' *180

In can therefore be concluded that the observed increase in

strength in the fast tests occurs as a result of a lower pore water
pressure in the faster tests and an increase in the "true cohesion"
with increase in strain rate with the "true friction" being independent
of strain rate, when the shear strength behavior is discussed in terms
of Hvorslev parameters.
It should be emphasized that Oe and ce were determined in
the conventional manner using the average water content for the
sample as a whole. Since it was concluded that migration of pore
water played such an important role in the behavior of the over-
concolidated samples, these values are not representative of
fundamental shear strength behavior. If the true water content
at the failure zone were known, this value could have been used to
determine e e. In Figure 46, the effect of this on the fast test
data would be negligible and in the slow tests, the effect would be
to shift the points plotted from the overconsolidated tests up and
to the right. Since these are the data points at the lower left of
the plot, the effect of this shift would be to increase the y inter-
cept while decreasing the slope of the best fit line. Since ce/Pe
is a function of both the y intercept and the slope of the line

-105-
on such a plot, the magnitude of this value would be relatively
unchanged by this shift. However, , would be lowered to a
noticeable degree.
Hence it can be concluded that, in terms of fundamental
shear strength behavior, the effect of an increase in strain rate is
to increase both the "true cohesion" and the "true angle of internal
friction."

D. Effect of a Step Increase in Strain Rate on the Hvorslev Shear

Strength Parameters
An attempt to determine the effect of a step increase in
strain rate upon the observed Hvorslev shear strength parameters is
presented in this section. This determination is based on data from
only two tests, NC16SF and 0011SF and suffers considerably from the
lack of sufficient data. In addition, the step increase in strain
rate in these tests occurred at strain levels differing by about 1%.
The data used in this analysis are listed in Table IX-3 and are
plotted in Figure 47.
Here again the magnitude of pe was determined from the
average water content of the entire sample as determined at the
conclusion of testing. The results of the analysis made on this
basis are presented in Figure 47. From these limited data it appears
that the effect of a step increase in strain rate is to increase both

e and ce/pe. However, additional straining tends to reduce both

strain rate
and ce to values more commensurate with the new
as evidenced by the 10% strain data.
 Table IX-3

DATA FOR COMPUTING HVORSLEV PARAMETERS UE

TO CHANGE IN STRAIN RATE

at 3.9% Strain
Before Change 1 -3
Test No. 3 1 - 3 p, U 3 pe e

NC16sF 60.5 37.8 60.6 33.8 26.7 .441 .312

OC11SF 6.7 25.1 46.7 -1.5 8.2 .175 .269
at 4.9%
After Change

NC16SF 60.5 48,2 6o.6 33.8 26.7 .441 .398

0011SF 6.7 31.3 46.7 -2.6 9.3 .198 .335
at 10%
Nci16sF 60.5 44.3 60.6 31.8 28.7 .473 .366
0011SF 6.7 31.5 46.7 -7.0 13.7 .293 .337
Typical at 4.9% Fast

NC16SF 59.7 43.4 60.6 31.5 28.2 .465 .358

0011SF 7.3 32.0 46.7 -3.5 10,8 .231 -342

E. Conclusions
No firm conclusions can be drawn from the foregoing analysis
due both to scatter of data and the influence of migration of pore water
in the slow overconsolidated tests. It appears that the values of O,
and c are related to the rate of strain. If determined in the

-107-
conventional manner, e at failure is relatively unaffected by the
strain rate in tests with uniform rate while c is higher in the
e
faster tests. However, if the water content existing at failure in
the actual zone of failure were known and used in the analysis, it
would appear that both 0 and c would be a function of strain
rate and would be greater in tests at a more rapid strain rate.
This observation is borne out by the tests with a step
increase in strain rate. However, the lack of sufficient data severely
hampered this analysis.
 Chapter X

SUMMARY AND CONCLUSIONS

A. Objectives of Research
The research described in the preceding chapters was instituted
for the purpose of determining the cause, in terms of fundamental contri-
butors to stress versus strain and shear strength behavior, of the strain
rate effect in a saturated fat clay. The effects of rate of strain upon
the shear strength and observed stress versus strain behavior of the
particular fat clay employed have been shown by NASIM (1961). However,
that investigation, while covering a larger range of strain rate than
that covered by the program described herein, suffered in that only
total stresses were considered (i.e., no pore water pressures were
measured) and thus no deductions could be made regarding the causes of
the observed behavior.
Consequently the program of research described in this paper
was undertaken with the determination of the cause, in terms of effective
stresses, of the effect of rate of strain on the stress versus strain
behavior of a fat clay as its principal objective.

B. SumMary of the Research Program

The research program described in the proceding chapters
consisted of a total of 31 triaxial compression tests on a fat clay
having a liquid limit of 60% and a plasticity index of approximately
32%. The predominant clay mineral constituents of the clay are illite
and montmorillonite.
These tests were performed at rates of strain varying from that
necessary to produce 1% strain of a 3.15 inch high sample in 45 seconds
(t 1 = 3/4 minutes) to that rate of strain producing 1% strain in
24 hours (t1 = 1440 minutes). In 17 of the tests, a step in strain
rate was imposed at some predetermined strain level,

-109-
 Twenty of the tests were upon normally consolidated samples,
19 of which were consolidated to a triaxial cell pressure of 60 lbs./in .
In 14 of these tests on normally consolidated samples, a step in strain
rate was imposed. Eleven tests were upon samples having an overconsolida-
tion ratio of approximately 16. Three of these overconsolidated samples
involved steps in strain rate at predetermined strain levels.
In all of the tests, the pore water pressure was measured in the
zone of the sample in which yielding was occurring. This was accomplished
by employing a small porous probe inserted into the sample at sample mid-
height. This probe was connected by a length of very small, exceptionally
noncompliant teflon tubing to a quite rigid electric pressure transducer
and the output of this transducer was monitored by an X-Y plotter.
In addition, in several tests, the pore water pressure was also
measured at the base of the test specimen and thus some measure of the
distribution of pore water pressure throughout the test specimen was
obtained.
The post shear distribution of water contents throughout each
test specimen was obtained as rapidly as possible at the completion of
each test. This was accomplished by slicing each test specimen into
six horizontal slices and determining the water content of each slice.

C. Significahce of Research
The following three sections summarize what is thought, by the
author, to represent the significant results of this research. Three
features of this research stand out: (1) A pore pressure measuring
system capable of measuring the pore water pressure at the sample mid-
height in tests having a strain rate in excess of that necessary to
produce 1% strain in 1 minute was developed. Techniques necessary for
the successful utilization of this system were perfected. (2) In the
range of strain rates considered, the cause, in terms of effective stress
behavior, of the observed strain rate effect in the normally consolidated

-110-
sanples was determined. (3) The results of the test program on the
overconsolidated samples showed the presence of and demonstrated the
importance of internal migration of pore water leading to internal
volume changes. This process was shown to contribute substantially
to the observed effect of strain rate on the stress versus strain
behavior.

D. Pore Pressure Measuring System

The pore water pressure measuring system and the techniques
for its utilization are described fully in Appendix A. Briefly, the
system as finally perfected, consisted of a porous stone probe 1/8 inch
in diameter and 3/4 inch long, inserted into the specimen at iid-height.
This probe connected to a length of fine teflon tubing .01 inches ID.
by .03 inches O.D. The tubing spiraled around the test specimen beneath
the rubber membranes but over the filter strips and passed out of the
triaxial chamber through the pedestal of the triaxial cell. This tubing
connected to a minute volume chamber into which was threaded a very rigid
electric pressure transducer.
The response of this system in a triaxial sample submitted to
an increase in all around pressure (chiamber pressure) was 100% in less
than 1 minute which, considering the low permeability of the clay, is
considered exceptional. The development of this rapid response system
capable of measuring pore pressures at the middle zone of the samples
was the key to the successful completion of this research.

E. Strain Rate Effect in the Normally Consolidated Samples

It was demonstrated in this research that the deviator stress
achieved at any strain level in a triaxial compression test on normally
consolidated samples of the saturated fat clay was higher in tests with
more rapid strain rates. At failure (i.e., maximum deviator stress),
this higher deviator stress in the fast tests was shown to be primarily
due to a lower failure pore water pressure in these tests. In fact,
the angle of shearing resistance in terms of effective stresses,
was shown actually to be slightly lower in the fast tests.
At strains less than 1%, the higher deviator stress in the
fast tests was shown, on the other hand, to be attributable to a
higher principal effective stress ratio (0/ ) with the pore
pressure at these low strain levels being independent of strain rate.
The migration of pore water due to pore water pressure
gradients existing during testing of normally consolidated samples
and resulting in internal volume changes was shown to be a minor factor
in contributing to the strain rate effect observed in normally consoli-
dated samples.
The existence of a "dilatency" component of pore water pressure
development in the normally consolidated samples was shown to exist and
to be time dependent by tests with step increases in strain rate.

F. Strain Rate Effect in the Overconsolidated Samples

As a result of this research on the fat clay, it was shown
that the major portion of the observed strain rate effect on the
observed stress versus strain and pore water pressure versus strain
behavior is a function of internal volume change or moisture content
redistribution phenomena. This internal volume change is due to
migration of pore water to equalize pore pressure gradients set up by
the irregular stress distribution in triaxial samples.
The reason for a tendency toward a higher base than mid-
height pore water pressure in overconsolidated samples was demonstrated.
The dissipation of this pore pressure gradient is a time dependent
phenomenon and thus contributes heavily to the observed strain rate
effect and would lead to the dependency of the observed strain rate
effect upon sample geometry.
The author feels that the results of the test program on
overconsolidated samples of the fat clay demonstrates that in searches
for fundamental explanations of stress versus strain behavior, tri-

-112-
axial test data on heavily overconsolidated samples is to be severely
suspected. It is up to any future investigator to demonstrate that
internal volume change phenomena is not influencing his test results.

G. Zonalusions
The cause, in terms of effective stress behavior, of the strain
rate effect on triaxial compression samples of this fat clay has been
determined. The existence of similar phenomena in other clay soils needs
to be demonstrated. The tools necessary for this demonstration are
available.
Further work is necessary particularly on overconsolidated
samples of this and other clays. Testing techniques eliminating -and
restraint in these samples could be developed. The. fundametal behavior
and strain rate effect of these samples has yet to be demonstrated.
The pore water pressure measuring system developed in the
course of this research is a valuable research tool and should lend it-
self to further investigations of fundamental stress versus strain be-w
havior of clay soils.

-11l3-
1--WAU.000000.............

BIBLIOGRAPHY

BAIA, A. 1960: "Stress Conditions in Triaxial Compression," Journal

of the Soil Mechanics and Foundations Division of the ASCE,
Vol. 86, December 1960, pp 5T-84

BISHOP, A. W., ALPAN, I., BLI(GT, G. E. and DONALD, I. B., 1960:

"Proceedings ASCE Research Conference on the Shear Strength
of Cohesive Soils," pp 503-532 and 1027-1042

BISHOP, A. W., and HENKEL, D. J., 195T: "The Measurement of Soil

Properties in the Triaxial Test," Edward Arnold, London

BJERRUM, L. 1961: Lecture Series in Spring of 1961 at M.I.T.

BJERRUM, L. 1954: "Theoretical Investigations on the Shear Strength

of Soils," Publication No. 5 of the Norwegian Geotechnical
Institute, Oslo, 1954

BJERRUM, L., SIMONS, N., and TORBIAA, I. 1960: "The Effect of Time on
the Shear Strength of a Soft Marine Clay," Publication No. 33
of the Norwegian Geotechnical Institute, Oslo

CASAGRANDE, A., and WILSON, S., 1951: "Effect of Rate of Loading on

the Shear Strength of Clays and Shales at Constant Water
Content," Harvard Soil Mechanics Series No. 39, Cambridge,
Massachusetts

CASAGRANDE, A., and SHANNON, W. L. 1948: "Stress Deformations and

Strength Characteristics of Soils Under Dynamic Loads,"
Second International Conference on Soil Mechanics and
Foundation Engineering, Rotterdam, Vol. 5

CRAWFORD, C., 1959: "Influence of Rate of Strain on Effective Stresses

in a Sensitive Clay," Annual Meeting of ASTM, Atlantic City,
June 1959

D'APPOLONIA, E., and NEWMARK, N. M., 1951: "A Method for Solution of
the Restrained Cylinder Under Axial Compression," Proceedings
of the First U. S. National Conference of Applied Mechanics,
ASME, 1951, pp 217-226

HEALY, KENT A., 1962: "Dependence of D lation on Rate of Shear Strain,"

M.I.T. Doctors Thesis, Department of Civil Engineering

HVORSLEV, M. J., 1937: Thesis at Vienna Abstracted in "The Shearing

Resistance of Remolded Cohesive Soils," Proceedings of the Soils
and Foundations Conference of the U. S., Engineers Department,
Boston, Massachusetts, 1938

-114-
KOLB, C. R., and SHOCKLEY, W. G., 195T: "Engineering Geology of the
Mississippi Valley," Transactions of the ASCE, Vol. 124, p 633

M.I.T., 1954: "The Behavior of Soils Under Transient Loadings,"

Report No. 3 to the Office of the Chief of Engineers, U. S. Army

M.I.T., 1959: "Dynamics No. 3, First Interim Report on Soil Tests,"

Report to U. S. Army Engineers Waterways Experiment Station

M.I.T., 1961: "Dynamics No. 6, Effect of Rate of Strain on Stress Vs.

Strain Behavior of Saturated Soils, "Report of U. S. Army
Engineers Waterways Experiment Station

NASIM, N. M., 1961: "Laboratory Investigation of the Effect of Rate

of Strain on the Shear Strength of a Saturated Clay," M. S. Thesis,
Department of Civil Engineering, M.I.T.

SKE4PTON, A. W., 1948: "A Study of the Immediate Triaxial Test on

Cohesive Soils," Second International Conference on Soil Mechanics
and Foundation Engineering, Vol. I. Rotterdam, 1948

SKf4PTON, A. W., 1954: "The Pore Pressure Coefficients A and B,"

Geotechnique Vol. No. 4, pp 143-147

TAYLOR, D. W., 1954: "Review of Research on Shearing Strength of Clay

at M.I.T., 1948-1953," Report to U. S. Army Waterways Experiment
Station

WHITMAN, R. V., 195Ta: "Testing Soils with Transient Loads," Conference

on Soils for Engineering Purposes, ASTM, Mexico City, p 243

WHITMAN, R. V., 195Th: "The Behavior of Soils Ursder Transient Loadings,"

Proceedings of the Fourth International Conference on Soil Mechanics
and Foundation Engineering, London, 1957 Vol. I, p 207

WHITMAN, R. V., and HEALY, K. A., 1961: "Shearing Resistance of Sands

During Rapid Loadings," Publication No. 114 of the Soil Engineer-
ing Division M.I.T.

-115-
m
M.I.T SAND I
SILT C
100 - '-

390
>-80
70
w60
z
F50
140
r20

-
1-0
10 1.0 0.1 0.01 0001 0.0001
DIAMETER IN M.M.
FIGURE I GRAIN SIZE DISTRIBUTION
VICKSBURG FAT CLAY

60[-

x 50

40
® \

-
30

20
10
L
I I I I I I
01 I
'
C 10 20 30 40 50 60 70 80 90 100
LIQUID LIMIT (%)

FIGURE 2 PLASTICITY CHARACTERISTICS

VICKSBURG FAT CLAY
~m a

45

I.
0

'- 35
z
w
1--
z
0
u

w
< 3-

1.0 2.0 5.0 10 20 50 100

CONSOLIDATION PRESSURE - lb/in 2
(Log Scale)

FIGURE 3 SUMMARY PLOT - CONSOLIDATION CHARACTERISTICS

OF VICKSBURG FAT CLAY
 50,

0 NC4-F
-NC5-FS
NE33 F NC8S
NC20S ---'
0
NUCS (Avg. NC17SF
c40 ,NCl18SF
*NCS '
NC9S
NCIOSF NC14SF
ONCIISF
0

E
0o
Test w (%) 03(lb/in2 ) Legend
NC2F 27.5 58.5 o Batch D-l
NC3F 29.5 59.0 2
NC6S 28.8 60.5
--= Test NC8S corrected to (T3 = 60 lb/ in
30i- NC7S 29.1 60.0
. o= Nasim data from ref. Fig.8
corrected to W= 28 %
NC4F 28.9 59.0
NC5SF 28.8 58.5 A =Batch D-2
NCI7SF 27.5 61.0 = From average curves
NC8S 26.4 66.5
NC9SF 27.6 59.0
NCIOSF 27.8 61.0
NC18SF 28.1 60.0
NCI ISF 28.3 59.8
NC 20SF *See note in legend
27.4 61.0

20
0 .1 1.0 10 100 1000
TIME TO I % - minutes

FIGURE 4 MAXIMUM DEVIATOR STRESS vs STRAI.N RATE

NORMALLY CONSOLIDATED SAMPLES
N
c

.0

u)

STRAIN-%
4

0 - 31.6

3
C
0
.0o

w2 0
U)

a-
o I
0 0
a.
.
o113
0 I 2 3 4 5 7 t3 9 10 11 12
STRAIN

FIGURE 5 SUMMARY PLOT - FAST TESTS - NORMALLY

CONSOLIDATED SAMPLES
 0

(f)
w
cr
0

w
0

STRAIN- %

4 0

301
cu

c0

w
L20
U)
U)
w
w
X 10
0

STRAIN- %

FIGURE 6 SUMMARY PLOT - SLOW TESTS NORMALLY

CONSOLIDATED SAMPLES
c

U)
Lu

Ld
Q

30
c'J
C

0
I- 20
(I)
C',
0
1~

0
L.
0
10

0o
STRAIN - %

FIGURE 7 AVERAGE DEVIATOR STRESS AND MIDHEIGHT

PORE PRESSURE vs STRAIN BEHAVIOR
NORMALLY CONSOLIDATED SAMPLES
 3,

lb 2.0

1.

STRAIN - %
FIGURE 8 OBLIQUITY RATIO vs STRAIN
NORMALLY CONSOLIDATED SAM PLES
0.7

0.6

0.5

0.4

0.3

II 0.2
(4.)
0.1

0L
0 12
STRAIN - %

FIGURE 9 RELATIONSHIP BETWEEN PORE PRESSURE 8 OBLIQUITY

RATIO - NORMALLY CONSOLIDATED SAMPLES
 2320

20

200 i10 .5%

lb +

Sorri (M.I.T.1959)
1b to\

- lb/in2
2

FIGURE 10 STRESS VECTOR PATHS - NORMALLY CONSOLIDATED SAMPLES

 NC7S, NCIOSF, NC13SF,
NC19SF, NCI4SF,
NC15SF and NC16SF
i
3.0 f

NC2F NCI8SF NC17SF

NC4F 0 0 ONC9SF
i i ONCIISF

Obliquity from NC5FS

E *
hand cranking
in NC12SF

2.0

o = Points from individual tests

= Points from "Average curves

to'0 .1
__ ____ __ _L_

1.0
__ -1 _ _ _ __ _ _ _

10
_ __ _ __ _ __ _

100
_____I____I__ I

1000
TIME TO 1%Y min.

FIGURE 11 MAXI MUM OBLIQUITY RATIO vs STRAIN RATE

NORMALLY CONSOLIDATED SAMPLES

- -,- . w
w

w
U,
(I,

0
wL

TIME TO I % STRAIN - min.

FIGURE 12 PORE WATER PRESSURE vs STRAIN RATE
NORMALLY CONSOLIDATED SAMPLES

b
~ I

0.61
0.1 1.0 10.0 100 1000
TIME TO 1% STRAIN - min.

FIGURE 13 PORE PRESSURE PARAMETER A vs STRAIN RATE

NORMALLY CONSOLIDATED SAMPLES
 i 1

0 = Preliminary tests - Gradient at

w +5.0k I P2-F max. deviator stress
Cn a = Preliminary tests- Max. gradient
+4.0
o Main tests - Gradient at (Oi0-3)max
NC5FS NC6 - S
w +-3.01- gf Main tests - Max. grad'ient
a. NC6-SO
0 P4F NC2-F PtF
""
+2.0
w Probable typical gradient
0.
w + 1.0 I SNC2F
0
NC7-S NC17-SF
2
w
0
Also maximum
Mi P2-F nrndiant ,
W 1.01-

0 ____________________________________________________

1.0 10 100 1000

TIME TO 1% STRAIN - min.

FIGURE 14 PORE PRESSURE GRADIENT vs STRAIN RATE

NORMALLY CONSOLIDATED SAMPLES
 p
I
/
I
I
I
I
I
3
I
I
4 I

(a) (b)
5

6
Dicrde
0.4 0.2 0 0.2 0.4 0.2 0 Q2
DECREASE (-) INCREASE (+) (-) (+)
POSTSHEAR DEVIATION POSTSHEAR DEVIATION
FROM AVG. W(%) FROM AVG. W(%). AVG.
OF 85%, FIG.15a

= (2-3)
Avg.. Tests NCI-F
NC2-F
NC3-F (c) (d)

0.2 O 0.2 0.2 0 0.2

(-) (+)W (-) (+)
POSSIBLE PRESHEAR POSSIBLE MIGRATION
DEVIATION FROM AVG. DUE TO GRADIENTS
W (%) DURING SHEAR

FIGURE 15 SUMMARY OF WATER CONTENT REDISTRIBUTION DATA

NORMALLY CONSOLIDATED SAMPLES
 1

40 Average fast
0-

cm
c
-- o
30
Average slow
41
W Fast Tests Slow Tests
I- NC2F NC7S
20 NC3F NC9S
0 NC4F NCIOSF
NC5FS NCII SF
NC 12SF
W NC 13SF
0 NC 14SF
10
NC 15SF
NC 16SF

Note: Shaded areas show

0 range of data
u .2
e .4 .b .8 I.u
STRAIN- %
(0)- 0~3 ) vs STRAIN

o = Avg. fast
0 = Avg. slow

W low

-Slow range

U) Shaded area = fast range

co
WI
a-
W

D .2 .4 .6 .8 1.1
STRAIN - %
PORE PRESSURE vs STRAIN

FIGURE 16 SUMMARY - BEHAVIOR AT SMALL STRAINS

NORMALLY CONSOLIDATED SAMPLES
 1.0

Avg. slow tests

< Avg. fast tests ;>

*-+-A" value if isotropic

and elastic

0' I I - - I -I
0C 0.2 0.4 0.6 0.8 1.0
STRAIN- % STRAIN- %

Note: Tests averages are

listed in Figure 16

.500 ~44c4~L4Avg.
slow tests
slow tests
WE)
b 1 I
fE)
*1*lb
1t5 .400
Avg. fast tests

.300L
0
\4~TT1
0.2 0.4 0.6
STRAIN- %
0.8 1.0

FIGURE 17 SUMMARY - BEHAVIOR AT SMALL STRAINS (cont'd)

N

a-

w
a:
0

a-

DEVIATOR STRESS - Ib/in 2

FIGURE 18 PLOT OF MIDHEIGHT PORE PRESSURE vs DEVIATOR
STRESS - NORMALLY CONSOLIDATED SAMPLES

+1.0

-I-

STRAIN- %

FIGURE 19 PLOT A vs STRAIN - NORMALLY CONSOLIDATED

SAMPLES
 30

:D
wl

lb
, 2.0
16~

I
STRAIN - %

FIGURE 20 SUMMARY PLOT - TESTS WITH STEP INCREASE

IN STRAIN RATE AT 4% STRAIN
 3.0 'I-
owing step n

strOin rote at 44
Steady rote tests
1b

2.0

cQJ
c

w
I-
Ur)

w
a

RATE OF STRAIN FOLLOWING CHANGE

(TIME TO 1% ADDITIONAL STRAIN)
t+l (minutes)

FIGURE 21 3 ) TO STRAIN
RELATIONSHIP OF POST STEP O1 /O3 AND (0-1 - O~ RATE
FOLLOWING STEP INCREASE IN STRAIN RATE - NORMALLY
CONSOLIDATED SAMPLES
c

U)
U)
wd
m~

3.0 I I N

b' 20
16

l.c I _________ I __________ I

0 I 2 3 4 5 6 7 8 9 10 11 12
STRAIN

FIGURE 22 SUMMARY PLOT - TESTS WITH STEP INCREASE

IN STRAIN RATE AT VARY ING STRAIN LEVELS
NORMALLY CONSOLIDATED SAMPLES
cr

z -
w

WW
cr cr
L).

I 2 3 4 5 6
STRAIN AT WHICH STEP IMPOSED - %

FIGURE 23 DEVIATOR STRESS BEHAVIOR

FOLLOWING STEP INCREASE IN
STRAIN RATE FROM t, = 500min.
TO t., = 3/4 min. - NORMALLY
CONSOLIDATED SAMPLES
 50

Somple held ot 7.33 % Strain

(A= 0. 2175 & .0001)
4C

Pore Pressure A Ambient temperature

cQJ \ fell 20* F, A to B - most
E30 ---- _____--- __
--- -- X recovered B-C

A C- -
S2C
N BGi
- - (73
w
a-

IC
NC 15- SF

C
0 2 3 4 5 6 7 8 9 1U
TIME - 1000 min.

FIGURE 24a RELAXATION BEHAVIOR - NORMALLY

CONSOLIDATED SAMPLES
 50

*-Initial deviator Held at 9.75% strain

II __

40 _stress = 44.3 (A= 0.2895 :.000)

10 min pore pressure 33.6 1 ,

- 4- -.....-... A Ambient temperature fell 200 F
C
Initial pore from A to B - Most recovered B- C
30 pressure = 31.8 N . I
4

L]-
B C
Pore water pressure

u20
0- Deviator stress

10
NC16- SF

0
0 1 2 3 4 5 6 7 8 9 10 II
TIME - 1000 minutes

FIGURE 24b RELAXATION BEHAVIOR - NORMALLY

CONSOLIDATED SAMPLES
 to

w
0 5
2
LI
w
w
U-
u-

0
0 I 2 3 Il 12
STRAIN - %

FIGURE 25 [(oi - 3 ) f, = I min. - (0j -0^3 ) t = 500 min] vs STRAIN

60

u-
E

O lI I I i I I I I I I I I I
0 I 2 3 4 5 6 7 8 9 10 I 12
STRAIN - %

FIGURE 26 % INCREASE IN (0) -0~ 3 ) FROM t =500 min. TO

t, = I min. vs STRAIN
 101 I0 II II II I I
I I
C- Pore pressure excess - slow tests
81F- I I ---
tA
tr KLti = 500 min) - I ti = I min)

Z)
WWU 4
o-w
CrLL
2
b-

OC I 2 3 4 5 6 7 8 9 10 HI 12
STRAIN - %

20
0
D 15
(I)
(I)

a3- z 10
W
Xr % excess pore pressure, slow tests
W
over fast tests
5
0

0 '7
I 2 3 4 5 6 7 8 9 10 11 12
STRAIN- %

FIGURE 27 PORE PRESSURE DIFFERENCE - SLOW AND

FAST TESTS vs STRAIN
 Particle A
r 2

Particle B

(a) IDEALIZED SMALL STRAIN BEHAVIOR MODEL

Translation at bond (D

Translation of bond (2

(b) IDEALIZED LARGE STRAIN BEHAVIOR MODEL

Translation

interparticular
h,
I

Slow displacement

7-+ A~r*

hi >h

Rapid displacement

(c) IDEALIZATION OF "DILATANT" BEHAVIOR COMPONENT

7-f
C D
B
Hvorslev Failure Envelope
A
slow
fast
C Chae p
Chamber pressure

(d) EFFECT OF "PRESTRESS" ON OBLIQUITY RATIO

FIGURE 28 IDEALIZED BEHAVIOR MODELS NORMALLY

CONSOLIDATED SAMPLES
 I I
I I I I I I I
0 Time to 1%
LLi+I Symbol Test Avg. W( /)
a: 0 OC2- F 30.0 100 sec
U) 0 OC3-F 28.7 45 sec
N OC4- F 29.8 90 sec
wL .c + OC5-F 29.4 45 sec
a: A+
0 C* 411.9 4 [280_
0L AO
0 Needle pore pressure. 29.5% 70 sec
W)
Un
wL
U
x 0-
Lii -I
0 I 2 3 4 5 6 7 E 9 10 I 12 13 14 15 16
STRAIN- %

30
C
.0

Ja

20

65

0 I

STRAIN - %

FIGURE 29 SUMMARY PLOT - FAST TESTS - OVERCON SOL IDATED SAMPLES

D
V)
CLi
U)

CL

0.-
x
(.1)
U)
wL
U
w-

STRAIN -%
40 I

Failure plane (Slowest of

70 sec Avg. fast fast tests)

30
.0 OCIO 8 OCIl-SF (Avg. slow)
..-''-- b0C7-S
O Failure plane noted OC7-S

U) 1, N 0 C8..S
eO
20
Note: Time on curve = Time to 1%
./

(n
0 I f%

Note: Tests OCIO-SF and OCII-SF extrapolated

beyond 4% strain

0 | 1 1
0 I 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
STRAIN- %
F1IGURE 30 SUMMARY PLOT - ALL TESTS - OVERCONSOLI DATED SAMPLES
 40

0 30
C OCI- F E
OC6-SOCIO- SF C
oc 6 OClil- SF
U) Nasim
U) Corrected to OOC7-S
ui
xW =29. 5%OC8S
O
U)008S

> 20

o Batch A samples
o Batch B samples
+ Avg. Batch B fast
A Batch C samples

0.1 1.0 10 100 1000

TIME TO 1%-min.

FIGURE 31 MAXIMUM DEVIATOR STRESS vs STRAIN RATE OVERCONSOUIDATED SAMPLES

 0

0
E
9-
b8
8-

z
6-

C.I 1000
TIME TO 1% STRAIN - min.

FIGURE 32 STRAIN AT (0) -O~03)mlx vs STRAIN RATE - OVERCONSOLIDATED SAM PLES

 TIME TO 1% -min.
0.1 1.0 10.0 100 1000
0I
OC8S SF I.Aocss
|CII-
C- 0C5-F Q02 F 10O
OC -F 0 Pore Pressure at OC7S
0 CIO-SF
OC3-FF00-F--
OC3-F loMax. (0 1 -J13 ) OC7S
OCI-F
0C7\
0- OC4-F
OC5.-F 0 OC6 S
-i
T10.0
--
002-F
%OF O
01- O Avg. Fast
OC3-F A Batch A, At (0j - 0-3)max.
0C4-F -Ar Batch A, Max, Pore Pressure
C0
o Batch B, At (( - T3)max.
a. -0- Batch B, Max. Pore. Pressure
-20.0 0 Batch C, At (ci -O-3)max.
-- Batch C, Max. Pore Pressure
0
a. I

FIGURE 33 MIDPLANE PORE PRESSURE vs STRAIN RATE - OVERCONSOLIDATED SAMPLES

 + 40

c
0
O

w tO 20 40
DEVIATOR STRESS Ib/in 2
-
Slow
tests
w
-10 Fast tests

W Note. Deduced from average curves

0
.

-20

FIGURE 34 MIDHEIGHT PORE PRESSURE vs

DEVIATOR STRESS
OVERCONSOLIDATED SAMPLES

+ 0.2
Slow tests

0
-0.1
6- 1 2 4 5 6 7 8 9 10 11 12
b STRA IN - %
-0.1Fast
tests
-0.2

FIGURE 35 PORE PRESSURE PARAMETER A vs STRAIN

OVERCONSOLIDATED SAMPLES

L
 TIME TO I% - minutes
x 0.1
0 1.0 10 100 1000
E 0Or-

4-
0

FIGURE 36 A AT (0j -O~3)max vs STRAIN RATE - OVERCONSOLIDATED SAMPLES

z
wdr

wLJcr
cr0
0 a.

00 0

Mz

TIME TO 1% - minutes

FIGURE 37 PORE PRESSURE GRADIENT vs STRAIN RATE - OVERCONSOLI DATED SAMPLES

-. mE

OCI-F (t,= 3min.)

OC2-F (t =3

OC3-F (t=1 min. -F (t I- min.)

-0.6 0 +0.2 +0.4 +0.6

NEGATIVE POSITIVE
DEVIATION FROM AVERAGE
WATER CONTENT - %

(a) FAST TESTS

o = OC5-FS
0 = OC6- S
A = OC7-S
= OC8-S
O 6)0,+ = OC9-S
4- OCO-S
* OCi-S
-0, e, *

"0+)-

-1.0 -0.8 -0.6 -0.4 -0.2 0 +0.2 +0.4 +0.6 +02 +1.0
NEGATIVE POSITIVE
DEVIATION FROM AVERAGE
WATER CONTENT - %

(b) SLOW TESTS

FIGURE 38 POST SHEAR DISTRIBUTION OF WATER CONTENT

OVERCONSOLIDATED SAMPLES
 5.0 IIiI

4.0
(0-J3max - fast

_3)mx
-I.. I.(9-
3.0 - (TI-93)max''..-..0 7
lb
lb

2.0

I i I I I I

1.0

0
0 I 2 3 4 5 6 7 B 9 10 11 12 13 14 V5 lb I
STRAIN- %

FIGURE 39 OBLIQUITY vs STRAIN RATE - SUMMARY - OVERCONSOLIDATED SAMPLES

5.0
A
OC5F =6.40 OCIF= 7.48 OC6S = 5.64

COC7- S

0C9-S
O
OBoth
- plots

0 0C3-F 0OC4-F

4.0
O
CC8-S

OC3-F a OC12-SF

O OCl-F OC6-S DC-

OC4-F

3.0

o =Points from Avg. curves

= Batch A
0 = Batch B
= Batch C

2.0
CD.A 1.0 10 100 1000
TIME TO 1% - min.

FIGURE 40 OBLIQUITY vs STRAIN RATE - OVERCONSOLIDATED SAMPLES

I (m

C1 + U3
2
FIGRE 41 FAILURE ENVELOPES

r')
YN

ii
Oi+( 3
2

FIGURE 42 BEHAVIOR OF OVERCONSOLI DATED SAMPLES

IN TERMS OF STRESS VECTORS
 20 -
....... OC4-F t =
1 5 min.
OC5-FS t, = min. Rote change, OC5- FS

OC5-FS--.
C

dol
/.-

0020 30 40
__ _ F_ -*- lb/in2
2

FIGURE 43 SUMMARY PLOT - STRESS VECTORS -FAST TESTS -OVERCONSOLIDATED SAMPLES

. . . . .. .. . . . . ||

20

0
In
'N

0L 0
O0
- lb/in2
2

FIGURE 44 SUMMARY PLOT - STRESS VECTORS - SLOW TESTS - OVERCONSOLIDATED SAMPLES

 20

"c

10

S)- Ib/in 2

FIGURE 45 SUMMARY PLOT - ALL STRESS VECTORS - OVERCONSOLIDATED SAMPLES

 W
CN

0.4
I

0.3

I
S~

(D
0.2
N

0.\1

0}
3 Pe

FIGURE 46 PLOT FOR DETERMINATION OF HVORSLEV

PARA METERS AT FAILURE
 to kfl- - -

0.5

0.

0.3
a.

b .2

0.1

01 1 I 1 1 1
0 0.1 0.2_ 0.3 0.4 0.5
0-3 / Pe

FIGURE 47 ANALYSIS OF STEP INCREASE IN STRAIN RATE IN TERMS OF

HVORSLEV PARAMETERS
 Appendix A

PORE PESSURE MASUREMENT: DEVICES,

TECHNIQUES, AND PERFOMANCE
 CONTENTS

Page No.
A. Introduction A-
B. Electric Pressure Transducer and Associated A-2
Instrumentation

C. Porous Probe and Connections for Midplane Measurements A-3

D. Technique for Installation of Midplane System A-4

E. Base Measuring Systems A-8

F. Response of Systems A-9

G. Conclusions A-10
 LIST OF FIGURES

-A-1 Final Method of Installation - Midheight Pore Pressure Measuring

System

A-2 Top View of Triaxial Cell Pedestal Prior to Placement of Specimen

A-3 Final Arrangement - Probe Measuring Chamber

A-4 Detail of Porous Probe

A-5 Base Pressure Device

A-6 Instrumentation Layout

A-7 Response Curves NC6S, OC3F

A-8 Response to Deviator Stress Increment

 Appendix A

PORE PRESSURE MFASUR0!r: DEVICES,

TW!CHNIQUES, AND PERF1RMACE

A* Introduction
The research described in the preceding chapters became
possible only through the development of rapidly responding pore
water pressure measuring devices and techniques for the utilization
of these devices. This Appendix describes these devices in detail.
In addition, the techniques employed in the successful utilization
of these devices are described, and data supporting response claims
are presented.
The basic element of the devices employed in measuring pore
pressures in this research is an electric pressure transducer. This
pressure transducer ccmbines outstanding sensitivity with extreme
rigidity and thus it was possible to develop a system having an
exceptionally low compliance (i.e., very small volume change required
to accurately measure pressures).
This pressure transducer was incorporated into two measuring
systems, one allowing measurement of pore pressure in the triaxial
compression specimen at the conventional specimen base location but
with a rapidly responding electronic system. The other system
provided measurement of the pore water pressure developed at the
elevation of the sample mid-height by employing a small porous probe.
Both of these systems required the development of techniques
for their successful application and the success of both these systems
as applied to triaxial compression samples was the main experimental
factor in this research.

A-1
B. Electric Pressure Transducer and Associated Instrumentation
The electric pressure transducer employed in this research was
developed and is commercially available from the Dynisco Division of
American Brake Shoe Company, 42 Carleton Street, Cambridge 42,
Massachusetts, as Model PT-25. This "Dynisco" gage consistis of a four-
active-anm unbonded strain gage bridge which senses the deflection of
a small, rigid diaphragm. It is accurate to better than 0.25 percent
of its full scale output and exhibits a maximum non-linearity of less
than 0.50% of its full scale output. The particular gages employed
in this research had fu3l scale ranges of 0 to 150 lb./in.2 . These
gages have a compliance on the order of 3 x 10~6 cubic centimeters
per pound per square inch pressure change which i's less than the
change in volume of the water in a typical triaxial compression
specimen. (The cubic compressibility of water is 3.4 x 106 lb.
per square inch. An 80 cc sample at a water content of 33% contains
20 cc of water which will compress 68 x 10 cc/lb./in.2 or 20 times
the gage compliance.)
These gages were activated with an extremely accurate, 6 volt
D.C. power supply of the type used to replace standard reference cells
in accurate recording equipment (EvenVolt Power Supply, Instrulab Inc.,
1205 Lamar Street, Dayton 4, Ohio). The power supply is insensitive to
changes in ambient temperature, a factor influencing earlier tests
employing a storage battery as a power supply.
The output of these pressure transducers is extremely small,
being of the order of from 0 to 10 millivolts D.C., at the range of
pressure changes represented by the tests performed during this
research. This output was displayed against time on a Moseley X-Y
plotter on which 10 millivolts represented about 10 inches, and thus
the required precision of measurement was possible.
The final layout of the instrumentation is pictured
schematically in Figure A-6. This layout featured a sampling switch

A-2
allowing both base and midplane pore pressure transducer output to
be displayed on the same channel of the X-Y plotter. This was
accomplished by switching input channels approximately 15 times per
one percent strain of the specimen. Prior to the development of this
apparatus, two X-Y plotters were employed when both midplane and base
measurements were required.

C. Porous Probe and Connections for Midplane Measurements

The porous probe used for midplan pore water pressure
measurements is shown in Figure A-4. Stones were cut from larger pieces
of Norton grade 2120 porous stones with a hacksaw and hand-shaped with
a file. Both the file and the hacksaw were destroyed in the process.
The shaped porous probe was fitted to a small brass cap and fastened
securely with small patches of epoxy cement.
The brass cap was drilled as shown to allow slipping the end
of a length of teflon tubing into place. No cement was used at this
connection and this connection was made prior to each installation of
the probe.
The teflon tubing employed was purchased from the Supernant
Manufacturing Company of Newton, Massachusetts, and is marketed as
"spaghetti" insulation for wiring. The type used was insulation for
#31 gage wire and had an I.D. of 0.01 inches and 0.D. of 0.03 inches.
Thus this shape can be classed as a thick walled tube. This tubing
was found to combine the properties of sufficient flexibility with
extremely low volume change due to pressure change.
This tubing led from the triaxial cell through a small hole
in the specimen pedestal. In the Norwegian cells, one of the base
drainage holes was utilized, but in the English cells, a small hole
was drilled to allow this tubing to pass from the triaxial chamber.
In both cases the holes were quite close to the edge of the specimen
pedestal. This exit was securely sealed with a plug of epoxy cement
around the tube.

A-3
 The external free end of the tube was connected to the
measuring chamber holding the pressure transducer by a small (27 gage)
hypodermic needle inserted into the tubing as shown in Figure A-3.
Since insertion of the needle into unprotected tubing tended to split
the tubing, the tubing was first fitted with a split sleeve as shown
in the detail in Figure A-3.
Figure A-2 shows the final modification of the specimen
pedestal of the triaxial cells. The method of construction was as
follows: (1) The porous base stone was cut out in the region of the
tubing outlet hole as shown. A layer of grease was applied to this
cut surface to prevent future adherence of the epoxy cement. The cut
stone was clamped to the pedestal. (2) The tubing was roughened with
emery paper in the region of passage through the pedestal and the
tubing was passed through the pedestal. (3) Using a piece of oiled
paper as a circumferential form, the cutout in the stone was filled
with epoxy cement. Care was taken to insure bond with the specimen
pedestal. After the epoxy had set, the oiled paper form and the
porous stone could be removed, leaving the epoxy addition securely
bonded to the specimen pedestal and thus sealing the tubing as it
passes out of the triaxial cell. Care was taken to keep each stone
with its respective triaxial cell.
The complete base measuring assembly then consisted of
(1) the porous probe and its brass cap, (2) the length of teflon tubing
fitted into the hole in the brass cap, spiraling around the specimen
and passing out of the cell via the base pedestal, and (3) the trans-
ducer chamber connected to the teflon tubing by a hypodermic needle.

D. Technique for Installation of Midplane System

This section describes the procedure for setting up a tri-
axial sample using the midplane pore pressure measuring system.- This
setup was acccmplished in the following manner:

A-4

L
 1. The porous probe, the base stone, transducer chamber and
washer, two hypodermic needles and a hypodermic syringe, were placed in
boiling water in a large vacuum dessicator. The dessicator was fitted
with a. lid, evacuated and kept evacuated for 24 hours. The water, of
course, boiled vigorously until cool.
2. An additional quantity of deaired water was prepared and
transferred to a large plastic tray about 6 inches deep. The base of
the triaxial cell was placed in this tray and air bubbles flushed from
the base drainage hole.
3. The hypodermic syringe was fitted with a needle below
water in the dessicator and filled with deaired water. This syringe
was used to flush thoroughly the length of teflon tubing attached to
the triaxial cell base. This flushing was done beneath the surface
of the water in the plastic tray.
4. The porous probe and base stone were transferred to the
plastic tray containing the triaxial cell base in a beaker of deaired
water. At no time were they removed from deaired water.
5. The two membranes, in a rolled condition, were attached
by "O" rings to the pedestal and the now saturated porous base stone
fitted into place.
6. The teflon tubing was pushed into the hole in the brass
end cap of the porous probe.
T. The specimen was trimmed. A horizontal hole slightly
smaller in diameter and length than the porous probe was formed in
the specimen at mid-height. In some soils this may be sufficient,
but in the Vicksburg clay specimens splitting of the specimen in
tension occurred upon imersion in water. The hole could be formed
in the sample dry, but if a drop of water were placed in the hole,
the sample would split.

A-5
 8. To avoid this splitting, the sample was carefully and
squarely cut into two pieces with a fine wire saw. This cut split the
hole. Now the sample existed as two cylinders, the top of one and the
bottom of the other having I of the hole.
9. The bottom half of the specimen was placed on the pedestal
of the triaxial cell. The water in the tray was deep enough to cover
this half entirely. All adhering air bubbles were removed by flushing
with water using the hypodermic syringe.
10. The teflon tubing was spiralled around the triaxial
specimen and the probe fitted into the half-hole in the bottom portion
of the specimen.
ll. The top half of the specimen was fitted into place taking
care to avoid any entrapment of air bubbles.
12. A rubber dam cut from a membrane and I inch wide, was
placed over the specimen and over the cut as shown in Figure A-1.
13. The filter strips were positioned, the top cap affixed
on the specimen and the first membrane rolled up. The water level was
then lowered to the level of the pedestal. Care was taken to avoid
having either the loose end of the teflon tubing or the base drainage
connections come above the water surface.
14. A layer of silicone grease was applied to the inner
membrane, the outer membrane was rolled up, and both membranes fastened
securely with "0" rings to the top cap.
15. The triaxial chamber was secured in place and fitted with
deaired water. A burette to measure drainage from the sample was
attached to the base drainage system and a chamber pressure of about
one atmosphere was applied. The loose end of the teflon tubing was
now clamped off with a pinch clamp and the triaxial cell removed from
the water. The chamber pressure was raised to the desired level and
consolidation proceeded. Note: If the teflon tubing is sealed prior
to application of a chamber pressure, or left sealed as an over-

A-6
consolidated sample rebounds, cavitation of water in the tube will
occur introducing air bubbles into the system. On the other hand,
if the tubing is not sealed during consolidation, it may clog as some
soil particles may wash out through the stone.
The the sample is to undergo rebound, the free end of the teflon
tubing must be immersed in deaired water and opened prior to reduction
of the chamber pressure.
After consolidation and immediately prior to testing, the
transducer chamber was attached. This was accomplished in the following
manner:
1. The transducer, washer, chamber and hypodermic needle were
assembled beneath the surface of deaired water in the dessicator.
2. The triaxial cell was again placed in a plastic tray of
deaired water.
3. The pinch clamp was removed from the teflon tubing and
the free end clamped in the split sleeve with a small bit protruding.
This protruding bit, amaged by the clamp, was cut off flush with the
sleeve using a new razor blade.
4. The transducer chamber assembly was removed from the
dessicator, keeping the tip of the needle immersed in a small beaker
of deaired water, The transducer was connected to the readout devices
and activated. After a warm-up period, the zero pressure reading was
noted.
5. The hypodermic needle was inserted into the teflon
tubing. The pressure developed by this operation commonly amounted to
1 to 3 lb./in.2 and decayed in a matter of minutes.
6. The base drainage was either sealed or a base measuring
device attached, as will be described in a following section.
7. A chamber pressure of 10 lb./in.2 was applied and the
response of the system (or systems) was noted. If this response was
100% in less than 60 seconds, as was usually the case, the 10 lb./in.2

A-T
pressure was removed and testing proceeded. If the response was
inadequate, this 10 lb./in. increase in chamber pressure was
permitted to remain as a back pressure for 24 hours and the response
to another 10 lb./in. increment noted. In a few instances a final
back pressure of 30 lb./in.2 resulted before the response criterion
was met.

. Base Measuring Systems

The base measuring systems consisted of chambers holding the
"Dynisco" transducer and fitted with adapters for connection either to
the Clockhouse cells or the NGI cells. In addition, a bleeder valve
arrangement permitted relief of any pressure developed during con-
nection. These devices are illustrated in Figure A-5.
The block, washer, and transducer were saturated in the same
manner as the midplane system (i.e., boiling in the evacuated dessicator).
They were assembled under water and transferred under water to the
plastic tray containing the triaxial chamber. Again electrical con-
nection was made before installation and the zero reading noted.
The bleeder valve arrangement consisted of a small ball
forced to seat in a conical seat by a thumbscrew. The threads were
quite loose and thus drainage past the threads could take place. The
transducer-block assembly was connected to the base drainage system
with this bleeder valve open. Tightening was accomplished slowly so
drainage past the bleeder valve would prevent the development of any
pressure. The bleeder valve was then closed tightly. The pressure
developed during this operation was monitored and amounted to on the
order of 2 lb./in.2 which decayed in a minute or two.
The response of this system was checked as was the midplane
system, Little trouble developed in the utilization of this system.

A-8
I

F. Response of Systems
Following consolidation of each specimen and prior to the
performance of the actual triaxial test, the response of the measuring.
system (or systems) was determined. This was accomplished by applying,
as rapidly as possible, an increment of chamber pressure
(usually 10 lb./in.2) and allowing this increment of chamber pressure
to remain until complete response was attained. Figure A-7 presents
response curves of the needle and base measuring systems from tests
NC6S and 003F which are typical of those of every test.
In no instance was the response less than 100%. In OC2F
and OC7S the initial response time was slightly greater than 60 seconds,
so an additional back pressure was applied and allowed to remain for a
period of time. The response time following this back pressuring was
less than 60 seconds.
In addition, the response of the systems to an increment of
deviator stress was checked by assembling a dummy sample, consolidating
it in the usual manner and connecting the pore pressure measuring ap-
paratus following consolidation and adding dead weight loads to the
piston of the triaxial cell.
Figure A-8 shows the results of such a test. In the upper
figure the results of the application of an increment of chamber
pressure (LaO) are shown. In the lower figure is shown the response
to an addition of an increment of deviator stress fA(1 - )
The response of the base system to an increment of chamber pressure
appeared to be more rapid than to an increment of deviator stress.
However, the response of the needle system to an increment of deviator
stress was about twice as fast as the response to an increment of
chamber pressure.
As a result of this experiment it can be concluded that the
pre-test criterion of full response to a chamber pressure increment in

A-9
60 seconds or less seems quite reasonable as an insurance of rapid
system response.

G. Conclusions
As a result of this research, it has been shown that measure-
ment of pore pressures with electronic pressure transducers is not only
feasible, but is superior in ease of measurement, rapidity of measure-
ment, and reliability of measurement to the conventional manual rule
balance system. This can be said with confidence since the particular
clay used in this research represents an extreme to the range of clays
to be tested.
In addition, an electronic measuring system utilizing a
sensitive millivoltmeter, such as the Daystrom-Weston Model No. 1477
instead of the relatively expensive X-Y plotter as a readout device
would be economically competitive with the currently marketed manual
systems.

A-10
 Top Loading Cap

.9

K
I I "O" Ring Seal
I
2 rubber membranes with
separating layer of silicone
- Il-I grease
I Filter Strips (4 used)

~L1~ Rubber "Dam" beneath

filter strips
Porous Probe

Teflon Tubing
.03" OD x .01" ID
II spiralled around specimen

i~1' ~
II
Built up with epoxy
111(See Fig. A-2)
This hole is as shown
Epoxy Seal
in N.G.I. cells but at
center of pedestal This hole exists In NGI cell
in Clockhouse cells and was drilled in Clock-
house cells

I
i To measuring
chamber

Standard base drainage connector V-Base of Clockhouse or

(either Clockhouse or N.G.J.) N.G.I triaxiol cell
connected to burette during
consolidation and either to base
measuring device or sealed
during testing

FIGURE A-I FINAL METHOD OF INSTALLATION - MIDHEIGHT

PORE PRESSURE MEASURING SYSTEM
 Porous stone
Base drainage
hole, NGI cell Built up with epoxy
C.o -Teflon tube

Base drainage hole

Clockhouse cells

FIGURE A-2 TOP VIEW OF TRIAXIAL CELL PEDESTAL

PRIOR TO PLACEMENT OF SPECIMEN

fit To recorder.

Lucite sleeve
Teflon tubing
To triaxial--
cell

/ \
-- Dynisco sealing
washer
Std. 17 gage hypodermic needle
Sleeve Detail

FIGURE A-3 FINAL ARRANGEMENT - CONNECTING

CHAMBER - TEFLON TUBING TO
PRESSURE TRANSDUCER
 dia. x " long
Porous stone
Norton grade 2120

brass cap

-Teflon tubing
exactly fits
into cap hole (no glue)

FIGURE A- 4 DETAIL OF POROUS PROBE

I

*Also has adapter to fit

Clockhouse cells

FIGURE A-5 BASE PRESSURE DEVICE

I
 Case grounds
hielded cable to X-Y plotter

-DPDT Swich
(Allows "sampling"
of channels)

Chassis

"Zero" adjustments
Notes:
B = Bose transducer
N = Midplane transducer
4 Wire shielded cable from transducers to chassis
P S. = Even volt 150 ma., 6V. , D C power supply
Plotter and transducer cases grounded to chassis

FIGURE A-6 FINAL INSTRUMENTATION LAYOUT

 10
w
Needle pore pressure
(I)
w

5
o OC 3-F
A603 = 12.3 lb/in 2
No back pressure

0 10 20 30 40 50 60
TIME - seconds

Nore
10 0c p
B e pressure

Needle pore pressure

0
z)5
w NC6-S
a: A(3 =10 lb/in2
a.
cr No back pressure

0 10 20 30 40 50 60
TIME - seconds

FIGURE A-7 TYPICAL RESPONSE - PORE PRESSURE SYSTEMS

 8
Bose_
N4
C
6
Needle
4
AG3 6.5 lb/in2
I
2

00 I I
10
__ I__II
20 30 40
II
50 60
ELAPSED TIME - seconds
RESPONSE TO CHAMBER
PRESSURE INCREMENT
2.C

Also 30 min.
CQ
A-- Needle (O -03) before = 0
,o I.0 0-- Bose
2
(91 0-)ofter = 3.9 lb/in
-

LLo 0
(0-j-A 3) = 3.91b/in 2 2
Lec = 1.6 lb /in
Aequil. = 0.41
0
0) 10 20 30
ELAPSED TIME - seconds
RESPONSE TO DEVIATOR
STRESS INCREMENT

FIGURE A-8 RESPONSE TO DEVIATOR STRESS INCREMENT

 Appendix B

PLOTS OF IABORATORY DATA

 LIST OF FIGURES

B-1 Results of Preliminary Tests

I B-2 NCl-F - Stress, Pore Pressure and Obliquity Ratio vs. Strain
B-3 NCl-F - Stress Vector Plot
B-4 NC2-F - Stress, Pore Pressure and Obliquity Ratio vs. Strain
B-5 NC3-F Stress, Pore Pressure and Obliquity Ratio vs. Strain
I B-6 NC4-FS - Stress, Pore Pressure and Obliquity Ratio vs. Strain
B-7 NC4-FS - Stress Vector Plot
B-8 NC5-FS - Stress, Pore Pressure and Obliquity Ratio vs. Strain
B-9 NC6-S - Stress, Pore Pressure and Obliquity Ratio vs. Strain
B-10 NC6-S - Stress Vector Plot
B-l NCT-S - Stress, Pore Pressure and Obliquity Ratio vs. Strain
B-12 NC7-S - Stress Vector Plot
B-13 NC8-S - Stress, Pore Pressure and Obliquity Ratio vs. Strain
B-14 NC8-S - Stress Vector Plot
B-15 NC9-SF Stress, Pore Pressure and Obliquity Ratio vs. Strain
-

B-16 NClO-SF - Stress, Pore Pressure and Obliquity Ratio vs. Strain
B-17 NCll-SF - Stress, Pore Pressure and Obliquity Ratio vs. Strain
B-18 NCl2-SF - Stress, Pore Pressure and Obliquity Ratio vs. Strain
B-19 NCl3-SF - Stress, Pore Pressure and Obliquity Ratio vs. Strain
B-20 NCl3-SF - Stress Vector Plot
B-21 NClk-SF - Stress, Pore Pressure and Obliquity Ratio vs. Strain
B-22 NCl4-SF - Stress Vector Plot
B-23 NC15-SF - Stress, Pore Pressure and Obliquity Ratio vs. Strain
B-24 NCl5-SF - Stress Vector Plot
B-25 NCl6-SF - Stress, Pore Pressure and Obliquity Ratio vs. Strain
B-26 NCl6-SF Stress Vector Plot
B-27 NC17-SF Stress and Pore Pressure vs. Strain
B-28 NCl7-SF Obliquity Ratio vs. Strain
B-29 NClT-SF Stress Vector Plot

I B-30 NC18-SFLC - Stress, Pore Pressure and Obliquity Ratio vs. Strain

L
B-31 NCl9-SFLC - Stress and Pore Pressure vs. Strain
B-32 NC20-SFLC - Stress and Pore Pressure vs. Strain
B-33 001-F - Stress and Pore Pressure vs. Strain; Stress Vector Plot
B-34 ocl-F & OC6-S - Obliquity Ratio vs. Strain
B-35 0C2-F - Stress and Pore Pressure vs. Strain
B-36 002-F - Obliquity Ratio vs. Strain
B-37 002-F - Stress Vector Plot
B-38 003-F - Stress and Pore Pressure vs. Strain
B-39 0C3-F - Obliquity Ratio vs. Strain
B-40 004-F - Stress and Pore Pressure vs. Strain
B-41 004-F - Obliquity Ratio vs. Strain
B-42 0C3-F & 004-F - Stress Vector Plot
B-43 005-FS - Stress and Pore Pressure vs. Strain; Stress Vector Plot
B-44 005-FS - Obliquity Ratio vs. Strain
B-45 0C6-S - Stress and Pore Pressure vs. Strain; Stress Vector Plot
B-46 0C7-S - Stress and Pore Pressure vs. Strain; Stress Vector Plot
B-47 OC7-S - Obliquity Ratio vs. Strain
B-48 008-S - Stress and Pore Pressure vs. Strain; Stress Vector Plot
B-49 008-S - Obliquity Ratio vs. Strain
B-50 OC9-S - Axial Load vs. Piston Movement
B-51 OC10-SF - Stress, Pore Pressure and Obliquity Ratio vs. Strain
B-52 OC1-SF - Stress, Pore Pressure and Obliquity Ratio vs. Strain
B-53 0010-SF & 0011-SF - Stress Vector Plot
 PI-F P2- F P3- F ) =23
N 2
P4 - F
A
40 20
#6=16 -' 0
$'B 17 0
C- Lp (0i -c3 )
30 30

0 W%
z 20 20 32.9
I- 32.3
32.4
10 03 59.5 101 Time to 1/ 543 sec 32.4
Time to 1% : 60 sec 43 sec Old Batch 32.5
Batch 3 Filter Strips

2 3 4 5 1 2 3 4 5 6 1 8 9
STRAIN - % STRAIN - % STRAIN- % STRAIN - %

Rote x 5
P5-S P6-S + P8-S 4)= 251 P9-S
41 40 40
Needle PP
Needle PP
N
C
--- (0T- T3 ) 1 03 .. --
N. PP
.0 -- "-Base PP
30 30 -Needle 30
Base PP agse PP

20 W = 31.5 20 20 Needle PP,---

I.-
C3 = 57.5 W = 31.5
CD Time to 1% =15min 59 Old Batch
(03=
10 (Leak patched) 10 Time to 1% 15min 10 W = 30.8 10
No Filter Strips No Filter Strips 03= 56. 5 (73 = 59 lb/in
2

Old Batch
f) I
Old Batch
I I I I I JNo Filter Strips Time to 1% =22min. Time to 1% = 13min.
02 C
2 3 4 5 0 1 2 3 4 5 6 0 1 2 0 2 3 4 5 6 7 8
STRAIN - % STRAIN- % TIME TO 1% STRAIN- % STRAIN- %
= 75 min.
2
0-3 56 lb/in
W= 31.5%

FIGURE B-1 RESULTS OF PRELIMINARY SERIES

 0
STRESS & PORE PRESSURE
0~' Vs
STRAIN (0i- 3)
0

Needle Pore Pressure

c

4z
0
to / ~3 .5
wC

-
f
9L I
9 TEST NO. NC I -F
/4 Ib/Is 2
6cons.- 117.6 lb/In2 &OrOunD
TIME TO I % STRAINm
I 2 mIA.

II i : I a I I I I , I a

\ 4- 5 6 7 \C) \ \Z-
STRAIN (%I
0
Co

OBLIQUITY vs STRAIN

lb
-No
Ib7

PL
~0 I 2. 3 4- 5 Go 7 6 9 o7
STRAIN (%
 0
TEST NO. NCI-F
STRESS VECTOR PLOT
Needle Pore Pressure
1i

N.

0
70 90 i00 110 lzo

2
 2

STRESS S PORE PRESSURE

STRAIN
Base Pore Pressure

Needle Pore Pressufe.

w. 8~

/
TEST NO NC2-F
0 Ocons.- 58. lb/In2 Orebound Ib./In.2
TIME TO I% STRAIN* I min. -
4A.TCN F

S I Z. 316 - % 9 0o t V?-
STRAIN N

OBLIQUITY vs STRAIN

Needle Pore Pressure

1b~
7/
I
I, I . I ,I I I . I II I . I I I I
Ol I ' 5 4 SiI
% STRAIN
 0
in1
STRESS a PORE PRESSURE
vs
STRAIN 3

Needle Pore Pressure

-- .. ,-- ~- ID- - -- n---

< TEST NO. N C3-F

2 ebound. b/in
_cons&e 59.0 ib/1i
0 / TINE TO 1% STRAIN= 5/4 min.

w LU 0)

0 /

O Z ' A- 5 6 9 \

0 STRAIN M%

OBLIQUITY vs STRAIN

lb n -

C) ( . $ 4. 5 6 ~ 9 1o fl 't
% STRAIN! -
 I- 3

- STRE SS a PORE PRESSURE

STRAIN

Needle Pore Pressure

0n

C
.f [o ( /.D

es

Q
If TEST NO NC4-FS
p Ocons.- 59.0 tb/In
2
Orebound e Ib/In.2
I
rt
TIME TO 1% STRAIN a
3/4 mIN.
BA;TC" 3
I
I I I I I i I
C'I J. I I - I

c3 4 5 r 1 tI I

0l STRAIN (N

OBLIQUITY vs STRAIN

lb~

0 I I I III L L 1 _____________________

~o - 3 4 5 G - B 9 1o \ VL
% STRAIN R-G
 ZAI
C.

bU
NC4-FS
I C

STRESS VECTOR PLOT

Needle Pore Pressure 0

I- i I -A-I - I
so I 0
40
(T+ )
12 (ib/in ),.
 2.
G3 Q)
4 jx\

- -
A ,

IIr
C S S~~ TEST NO. NC5-FS
6c'ons.- 58.5 lb/In2 debound- Ib/in2
TIME TO I'% STRAIN= 3/4 m In.
w
0c / EBATC "
STRESS a PORE PRESSURE
CL vs
wQA)
STRAIN
i~iI:1
L~i~~J
r~±c~~I
1~is-l
L~1i

3 4- 5 to - 7 , 9 10 11 I
STRAIN (%N
9

- ~ -& ~ 0

LY
/
H /

lb

:7 OBLIQUITY vs STRAIN
Needle Pore Pressure

~0

0 I Z 3 4. 5 10?
% STRAIN
 STRESS S PORE PRESSURE
vs
STRAIN

N.

w
?Ol
cc
03 Ise '
0
at
IL

TEST NO. NC6-S

2
O'cons.* 60.5 lb/In Crebound lb./In.2
0
TIME TO 1% STRAIN - 20 min.

( 2.1 3 4 5 6 7 8
STRAIN (%
c)

lb
-fto
Ib7

/ OBLIQUITY vs STRAIN

-I Needle Pore Pressure

* II

01
C1
I I a I , , , I - I i I & I I i I i I
it 3 5 6 7 ~ 9 ~O II 12.
STRAIN (94)
B-9
 0

-,

NC6-S

STRESS VECTOR PLOT

Needle Pore Pressure

+ )
2 (Ib/ap

B-tO
 STRESS & PORE PRESSURE
vs
STR AIN

(0 - (T3)

c
/
/i ~ A~
4z

W
O

TEST NC. NC 7-S

acons.- 60.0 lb/i n2 dieboundl Ib/I 2 a.3
TIME TO I% STR AIN= 350 min'.

O0 a 3 A- 5 c -7 8 9 ?-
STRAIN (%

,D

OBLIQUITY vs STRAIN
/
/ Needle Pore Pressure
/
9 p
~0

lb0 'p

b O

S Z.. 3 - 5 I 9 \o \~ Iz~
% STRAIN
 c!( l-'9
STRESS VECTOR PLOT
Needle Pore Ptessure

G~o
40 50
-i--I (lb/i. 2 )
2
 STRESS a PORE PRESSURE
vs -crM
STRAIN ,-00 er

LT~IZOF ~

TEST NO. NC8-S

2
Mcons.. 66.5 lb/In2 &rebound- Ib/hi
TIME TO 1 % STRAIN a 1440 ml .
C BA:ToC e,

w (o/0,/

G -r5

.0

w9
0

1I , 1 I I I I
C) 3 4- 5 ro 7 s
STRAIN

'%0

lbo / OBLIQUITY vs STRAIN

/
-I/

0 I I. I
I
, I , F I I , I I I I I I
, .
0 I 2. 3 4 5 (
GZ- - S 9 to
% STRAIN
r
( D
'Le~v

-1 l

iN

A-
9
NC8-S
'N
STRESS VECTOR PLOT

140 5 ~
+ ()

B-14.
 )
p ~C
Need e pore Pressure

TEST NO. NC9-SF

- cons.-a 59.0 lb/In2 &ebound, lb/in2

./ TINE TO 1% STRAIN= 500 min.

wA / BATC \A 8
'"0 7
w /
C*/
U, STRESS a PORE PRESSURE

o.. STRAIN ?

0 - -
2. '5 4. 5 6 ~ 9 \o \ ?
STRAIN (%
0

P8I NTMD

A T

b OBLIQUITY vs STRAIN
lb7

- 3 4- 5 ( 7 "c
% STRAIW
 STRESS a PORE PRESSURE 12 m in.
YIN
STRAIN
(0--(r)
0
'4

Needle Pore Pressure

0
'I-

N. /

C -I,'
('I
I

0 TEST NO NC 10-SF
2
0-cons. * 61.01 b/In 0-rebound Ib./n.2
TIME TO %) STRAIN - 500 min.

I 5 o -T 9 to %1
STRAIN (%)
0

lb
Ib-0
.... ed /
/
OBLIQUITY vs STRAIN

Q I I a I . I ,I I . .. . I . I , -I , I
C) z. 3 4. 5 6 re~ 9 io0 .
% STRAIN
 t 1= 1/2min.

flTypical Fast
I - ftft. rn_ I&- -. - &
f, :zbbi:>0 rn IV,
-~i~.-.-%r~~ LD -
K'-TypicalFast u

TEST NO NCIl-SF
(Tcons.- 59.8 lb/In2 C-rebound Ib/In.2
C
TIME TO 1% STRAIN a 500 min.

STRESS & PORE PRESSURE

vs
STRAIN

2.4

2. 3 4- 5 -I B 9 to \\ '.
STRAIN (%)

~j+

lb! 0
OBLIQUITY vs STRAIN
/
U'
/
0
~o
I
I I I
I
S
I I
I
3
I I
I
4-
I
5
II I II
G
I - II
1
A I
8
I I
9
A I
I
to
*
a I I II -I
I i

STRAIN (0/c,
5-Pi
 cr-cy 4r min,
~

Q *0

o
0
(3 4)/
'I

w /
4"/t- w
TEST NO. NC12-SF
C /* 'cons.* 60.0 tb/In2 &eboun
'y /
l I TIME TO 1 % STRAINm 500 m -
BATCH C
?- 9
STRESS S PORE PRESSURE
O3 ve Z -c-
STRAIN

0O I I tj

I '.I T 9
STRAIN (%4
 tj /4
( I

4J

-~ ~E.LA 1 'e. .F~Ci7~

10
~

po(~ PG~

'1,/i TEST NO.NC 13-SF

/ cons.- 60.4 tb/In
2
Erebond a lb/hi
2

9 I
1 TINE TO 1 % STRAINs 500 mIn.
I BATCH C
[(*/&)

9 STRESS a PORE PRESSURE

vs

STRAIN

I 3 4 ~T to \\
5TRA4 (a/0)
ll
WA. O

'I
I
nj

)
/
- /
/ OBLIQUITY vs STRAIN
i/I
/
-I
-7
I,
(
-O
r I I
'.
.I
A
.I
5
,Ii
6
I .I
.
~7
I I I
1 9It 17-
STRAIN)
B-19

L
 0

NC13-SF

STRESS VECTOR PLOT

0
,%j N)
"/
b. ,'- I

;II N Ii,

N'

7"
I I, w t
40) 50 Go

a-zo
 tI 3 / 4 m

TEST NO. NC 14-SF

Ocons.*- 58.5 lb/In
2 Orobound Ib/In.2
/1
If TIME TO I% STRAIN - 500 "Is.
:30
BATCI HC

STRESS S PORE PRESSURE iN-)

vs
STRAIN
2-~1
a.19(

4
710 it
cA5
STRAI
0 -

OBLIQUITY vs STRAIN

.7
-
o 3. 4 5 ti SRI 'T & 1
% STRAIN g
 NC14-SF
Now STRESS VECTOR PLOT

d0f
ibJ

rIN
0

I
I
___

I- _--V
/4'
40
+
2 (bI2

B-?ZZ

L
 9,

(- (y 3' -y
3

O C
E
'3- t 1500 min

sr e Suree
- -- - -*
Jr e 4 .

STRESS & PORE PRESSURE

vs
c. STRAIN

la 0
0. 4f)
'I IWt71
C0
F~7~7i TE ST NO. NC15-SF
F~7~7I Mcon.- 59 .5 lb/In2 Ubo*nd Ib/In2
TIME TO I % STRAIN a 500 mn.
L~i~
t8T C" C

0
I 3 A-
STRAIN N

wt

6Y~
/
OBLIQUITY vs STRAIN

am%

0
3 4. 5 .= a 9 ic> Ii .
% STRAIN
B-23
 0

NC 15- SF
e~P~
El
.f
0 STRESS VECTOR PLOT
W.

11 ,,

I N

0 H-
i f)

a -I - 40.0
I< 11
60.0

(Ob /14)
2
 2
b/

II
--- SL G-
C
cc,
c
TE ST NO. NC 16- SF
.0 ( Ocons. - 60 .5 lb/In2 Orebound - lb./In. 2
TI ME TO I% STRAIN- 500 min
0
6~A,'C CN

0 STRESS S POR E PRESSURE

ST
STRAI N

4 I I I 8 |
3 4- 5 6 ~( 8 9 \7-
% STRAIN
0

3/4 man.

OBLIQUITY vs STRAIN

lb p

(~)

0 ' 3 4. 5 G ' ' 9 -

STRAIN (%
 0-

NC 16-SF
STRESS VECTOR PLOT

le~p
a" ~

"b'o

(o

04

I
0
40

(lb/Is2 )
a
 0LO

Needle Pore Pressure

C
~0
'4

Lii
oa TEST NO. NC17-SF
2 Orebound- Ib./In. 2
U(,)
-ycons.- 61.0 lb/In _W).
TIME TO I% STRAIN - 500 mIN.
l-A~
STRESS a PORE PRESSURE

I STRAIN

II I I . I I I I I I . I . I
O' t. 3 A- 5 -T 9 " it.
I-- -T; P.Ao iK! %'

B-zT
 NC 17- SF

S_ OBLIQUITY vsS TRAIN

Needle Pore Pressure

0
0 '~

b O . 5

. ~~ I * *

5 4- s 6 T 9 t IT-
6o
STP.AAIN c%>
r
0

NC 17- SF -'

STRESS VECTOR PLOT

Needie Pore Pressure

40 0
cy, +(O;b

B-Z9
 z

c I /

o W(0/0)
TEST NO. N 8S Lc
zcons.- 6 0.0 Wbin 2 Seon. Ib/In2 Z1-

a 0 TIME TO 1 % STRAIN= 3 min.

BAT CH 6 KZo

o ?.. $ 4- 1 r. -1 a
ST Atn ~c A/)
0

op

0-7 e 9 to \\ \'?-
ST GAd (/o) 3. 30
 STRESS & PORE PRESSURE
vs
STRAIN
tj 3/ r . ( (- T)

k~ ~ ~ ~ * 50 % %. .,Needle Pore Pressure

TEST NO. NC19-SFLC

Ocons.- 59.2 lb/In2 Orebound e b./In.
TIME TO 1% STRAIN - 500 min. g

BA TCH C e
LOAD CELL
O'TE -- LEAVA~Ca. taJO bAMPi"
0
4- 5 6 ~7 9 10 \3\ t'
5T RA,%
V' C(%)
 STRESS & PORE PRESSURE
vs
STRAIN
0-r
oo 0

2 0
I-'
Cd

N
CD
J IHL47T 4.035
'-I
I
[U
I
I
I LOAD CS- Tae5T
Li' I
0~ 9! -T 211

0 L
6 I ?.. C 4- 5 E -7 i 9
TQA~~4 C"/o

TEST NO. NC2 O-SFL C

Ocols.u- 61.0 lb/In 2 r ebound Ib./In. 2
TIME TO I% STRAIN = 3/4 fai

BATCH C
 TEST NO. OCt-F
con.-116.0 lb/in2 rebound- 6.0. lb/in2
TINE TO 1% STRAIN' = mn.
B/\TCH A r1ua PLA.je (7

0*n197 1b. /in2

C8

STRESS & PORE PRESSURE EI.3

vs
'I)
STRAIN

0~~
Base Pore

o.. Preresu

I
C 3 5 C6 -7 K~ ~I 1'L
S IL '6 9\
15 \9C to Z~ Z'~ 2.4-
ST PAt VJ (*0/0)

0
e'I

STRESS VECTOR PLOT

Needle Pore Pressure

'4

I'

0 0 1=14~
Z.0 30
(:7' 1- W-4
B-33
 a
0
p
I
I
a
- /
I OCt-F a OC6-S
a -~ ~4h/4~
0 - I a
/ OBLIQUITY vs STRAIN

a
'I

a
-
I

lb!
-IU

'I
/ Is mkwr .Q

0
-4 ~J
Ib

."N

U ~

I
1/

4. 5 6 8 9 K> H rs 4- is n
STRAIN (%)
 Base Pore
Pressure

- Needle Pore Presr

(0-a-) 4-
-0 - -a esr

dr .. u.0/'

TEST NO. OC2-F 300

O Mcons. 120.0 lb/in2 d'iboundu 7.0 lb/In2
TIME TO 1 % STRAIN I,7 mIn. 30.0
BNTCH 8 30.0
STRESS & PORE PRESSURE
0
STRAIN

0 1 Z 5 4- 5 G 1 8 9 0 t t ' 1 's 14- Is -1 8

 a
ui

0C2 FTAt
/ OBLIOUITY vs STRAIN

14
(0
lill
'Ui 131
z
CU -
M
to

0u

0 I-i- I I I I I I
0
S~AU~4 (%~)

Lo
0'
 0
OC2-F
N
STRESS VECTOR PLOT

Needle Pore Pressure

Oft
-o

br 0

/1 I
-~ II5I I
0 I I 30 arv
10 20
tJJ
I
-.4 2 (ib/lA
 4 mad

TEST NO. OC3-F

ficons IIT.5 lb/In2 drebound. 7.2 lb/in2
TIME TO 1% STRAINa 3/4 mIn

(KT-a-

CC4

STRESS & PORE PRESSURE

vs
STRAIN
a.4

Needre P Pre
s; eelePore Pres s 4
sure e --

0 3 4 5 6a 10 1\ 1z. k I- ' 6 1-7 \s

TqA,( *0
 OC3-F
OBLIQUITY vs STRAIN
0
Needle Pore Pressure

01
&*I
0
vi.

0l- 6 '7 IC9 ) 1-3 14. ;

0 I -6 4- 9-'
db- RAI/ ~
U)
 TEST NO. OC4-F
Mcon.- 116.5 lb/In2 d&ebound- 7.5 lb/ C-- =40 1
TIME TO 1 % STRAINs 1.5 mIn.
BATCH a

CI

0 Base Pore Pressure

-j
o-..Needje Pore p~
-e Pressure
dl
d
'~eJ
'p
di Z.9. 53-
ci I9 .- ..bt
.11
&

O1
\ e. 3 4-6 T 6 9 \C \\ \.. is \4. \) \ 0 I
STW.A\ .(*/
 Li

OC4-F
OBLIQUITY vs STRAIN
Needle Pore Pressure

ib-
0
N

3 4-. -

STP?.A, t
 0

lip
0C3- F
STRESS VECTOR PLOT
0 Needle Pore Pressure

O' I
10 Z0 30

3a)F

0C4-F

STRESS VECTOR PLOT-

-j Needle Pore Pressure

2D 40
 TEST NO. OC5-FS
mcons.8 120.0 lb/In2 '*bound= 7. lb/Wi 2

TIME TO 1 % STRAIN* 3/4 INI

0~~~a Ov
k- -Npa c-

u4. 5P t m

w \Base Pore Pressure C

soPo

re ss P re ss u.

STRESS S PORE PRESSURE

vs - - &-
STRAIN

0 9 .
STRAIN (%

STRESS VECTOR PLOT

N*ede Pore Pressure *~

br

0130
ot . o so
(0+5')
2
j (ib/lp2)
2

0-43
 0
0

OC5- FS
'.9
OBLIQU4TY vs STRAIN
Needle Pore Pressure

-1-

"C'
...
.....
.............

on

I QjI

OL
1 7- 5 (o 9 1
5TQAkW (b/o)

B-44-
 TEST NO. OC6-S
cons.- 116.0 lb/In2 &*boun- 6.4 lb/In2
TIME TO I%STRAIN= Is min.
B A-7CV A
Ii 0-

0 aZs94-
Lii

ct
90 -- "-lse e Pressure

STRESS a PORE PRESSURE

Needte Pore Pressure
ST RAIN

O
0 I 3 4- 5 6 7 8 9 1ct i -
,5Aud (*/)

0
ca
STRESS VECTOR PLOT
0" Needle Pore Pressure

/AI

01 I 1' I0 ao I 3 I
c C 10o - s 2.o 3o
 0
ct) ~~ (a- C'3
o-?Lk

--- * 5:
tp 0
(~J
TEST NO. 0 C 7- S
.>Co/o)
111 0-cons.- 120.0 lb/In 2 O'rebound - T.0 lb./n. 2
ci TIME TO I% STRAIN - 30 mir'.
0
STRESS & PORE PRESSURE

STRAIN

40 II 111 1 I , I I I
) 2 3 4- 5 6 7 B 9 10 1?
'5-rQAi K (%*')

8
0se pore Pressur

Needle Pore Pressure

STRESS VECTOR PLOT

Needle Pore Pressure
0

,~-0

crC~

0.

C' I1 I I a II

O0 \1 0O 30
CLB/ 2 )

?,- 4 =
 0
0d

lb!
ib
-0b
40

Q-

STRAIN (M4

B-47
K
TEST NO. 0C8-S
Mcons.- 116.5 lb/In2 Oveboundu 7.0 lb/In2
0 TIME TO 1% STRAIN 250 min.
'-I
(4
2 (0 --a)3)
0
-J
LIP
:2
Needlie p
'p Oe prs
Surew
a!
-
STRESS S PORE PRESSURE [ - -S

STRAIN
(
4- 5 6 ~7 810 It k?..

9
STRESS VECTOR PLOT

(4

btni<
t~AeJ

D IC 30

z.
B-45
 0
uii

AM
0
4

0 ~Q~- A&Q~VW~4t
co
to

I 0C8-S
OBLIQUITY vs STRAI N
40

-o

O I I , I I 1 |1 1 1 1 1 1I1 1
c Z 3 A- 5 6 ~7 9 \ '

B-49
 WO- \ot.)
\AJ\4 're4TwAlbC h
TEST N6. OC 9- S
6'cont= lig.5 tb/In 2 &ebo m 0. 0 Ib/ln2
SKTCb4 S~ - OOAF-\
OF Loki>C~

CA\J-0 0~ w~~J
LM.0 0FQ MS 5902JVNDJOPF
=rmztg
- tATkv-JCT VA.LUQF= PL~r-

oc-0 1.4
c5
LB PL

t A5

r-ts
IIZAt 1

to~I
0L1
01i

MA.-D .FQ-Ame CiIFF

7
I 1 I I
~j1 C LU a a a

I
040
I.
0.10
I .

DOwQwkp-lD P\157MN-l mo-rCVj

 TEST NO. OC1O- SF
(cons.- 116.5 lb/In2 dbound 6.6 lb/1n2
TIME TO 1 % STRAIN. 500 m Ip.
XCg HC t, =-3/4 min.
9 -- TN~Ome'CAL.
S Lc\s% P.AtTS

.4 s2 Needle
Pore Pr
essur,
i -
*f

'-2OC 29'.6l

STRESS S PORE PRESSURE

vs
STRAIN

2. 3 .A 5 G 7 8) 9 \O ii VL
5-TPA\Qi C%/7)

S
OBLIQUITY vs STRAIN
Needle Pore Pressure

MAN HAV AFFQC:ED) ST4

2Ama s*/o -l*/ me10
0

. 3 A- 5 6 ' 8 9 -7
 -- VE -(T3t
--- kt P'(T 3/4 m

_________r.5ThAT~ ED -ow)RAT
-- --- am
,gt.
b2.

/ -Neede Pore pr
0 - / ress~r

STRESS & PORE PRESSURE

vs
o1 / 1 STRAIN
BkTCH C

494= ?.r).-7

I I I I I , I . I
0 )
z 3 4- 5 6 "T Eb 9 \ \\ '2.

TEST NO. OC 11-SF

&cOns.- 116.5 lb/In2 db~ound 6-7 lb/hI 2
TIME TO 1 % STRAIN= 500 n.
.5I0 min

0 -
FAST CuPvE,%ZFic

4 b OBLIQUITY vs STRAIN
6'~~ Needle Pore Pressure
I

OL
'O I 5 4- 5 ~7 1& 2 G9 \o i
STeA" */o0
 0
0i]

OCIO-SF
STRESS VECTOR PLOT
Needle Pore Pressure

OI

30

t = 3/4mut

OC l- SF
STRESS VECTOR PLOT
Needle Pore Pressure

t0
0-

'a