Title On the Rheological Characters of Clay Part 1

Author(s) MURAYAMA, Sakuro; SHIBATA, Toru

Bulletins - Disaster Prevention Research Institute, Kyoto

Citation University (1958), 26: 1-43

Issue Date 1958-10-20

URL http://hdl.handle.net/2433/123681

Right

Type Departmental Bulletin Paper

Textversion publisher

Kyoto University
*34. € 1.

DISASTER PREVENTION RESEARCH INSTITUTE

BULLETIN No. 26 OCTOBER, 1958

ON THE RHEOLOGICAL CHARACTERS OF

CLAY
PART 1

BY

SAKURO MURAYAMA

AND
TORU SHIBATA

KYOTO UNIVERSITY, KYOTO, JAPAN

 1

DISASTER PREVENTION RESEARCH INSTITUTE

KYOTO UNIVERSITY
BULLETINS

Bulletin No. 26 October, 1958

On the Rheological Characters of Clay

Part 1
By
Sakurd MURAYAMA
and TOru SHIBATA
Contents
Page
Synopsis 2
Introduction 2
Chapter 1 The Flow Mechanism of Clay
1. Introduction 3
2. Structural Viscosity proposed by the Authors 4
3. Mechanical Model and Flow Mechanism of Clay 8
Chapter 2 Studies on the Flow Tests
4. Introduction 11
5. Some Data of the Clay Specimen and the Compression
Plastometer 11
6. Flow Characteristics 13
7. Strain-rate Characteristics 15
8. Stress—Strain Characteristics 19
9. Characteristics of Flow under Repeated Load and Flow
Recovery 21
10. Summary 27
Chapter 3 Long-term Strength
11. Aim and Brief Contents 28
12. A Method to Measure the Upper Yield Value of Clay by
Stress-controlled Compression Test 29
13. Effect of Water Content on the Upper Yield Value and
on the Compressive Strength 34
14. Long-term Strength 35
15. Effect of Flow on the Strength of Clay 39
16. Summary 41
Conclusion 42
References 43
 2

Synopsis

This is one part of the report* of a theoretical research on the rheological

properties of clay. In this paper, the authors indicate that the representation
of clay by the complex combination of simple mechanical model units causes
the complication on the mathematical treatment, but does not always contribute
to the analysis of essential nature of clay, and insist that the solution of
rheological behaviour of clay should be treated with the theory deduced from
the micrometric standpoint.
From this standpoint, applying the structural viscosity derived from the
statistical mechanics on viscosity of clay, a new fundamental formula concern-
ing the deformation or strength characters of clay—such as flow or long-term
strength is derived. This formula well agrees with the various kinds of
experiments—experiments on compression flow, recovery, flow by repetitional
load, yielding, failure, etc.. Furthermore, results of tests on flow, yielding,
failure or thermal effect can be explained micrometically by this formula.
Applying the formula and experiments, some important problems on soil
structure—such as strength—water content relation, relation between yielding
valure and failure strength, time lapse necessary for failure, variation in st-
rength by flow—are investigated and a new measuring , method of yielding
value of clay is proposed.

Introduction

This is a report on the rheological characters of clay theoretically re-

searched from the micrometric standpoint.
Formerly the apparent mechanical behaviour of clay was represented fre-

quently by the behaviour of the complex combination of simple mechanical

model units such as dashpots, springs or sliders. But the real character
of clay is too much complicated to be represented by such a simple model
analogy. Moreover, this method of representation dose not always contribute
to the analysis of the essential character- of clay behaviour in spite of its corn-

* This is an English report translated from one part of the Transactions of Japanese
Society of Civil Engineers No. 40 pnblisheil in Dec. 195.
 3

plicated mathematical treatment: Therefore' the authors believe that the solu-
tion of the theological behaviour of clay should be treated with the theory
deduced from the micrometric standpoint considering the structure of clay.
From this standpoint, in this paper, the theological characters of clay
are researched, and the contents of the research are as follows :
In chapter 1, applying the structural viscosity derived from the statistical
mechanics on viscosity of clay, a new fundamental formula concerning the
flow character of clay is derived.
Chapter 2 is a report on the experimental results and their considerations
of a series of compression flow tests of clay to refer and ascertain the formula
derived in Chapter 1. Any flow characters of clay in the experiments exact-
ly agree with the formula so far as the intensity of applied flow stress is less
than either the pre-consolidation pressure or the upper yield -value.
In Chapter 3, some important problems on soil strength—such as the
relations between clay strength and its water content, relation between failure
strength and the time lapse necessary for failure and variation in strength
by flow—are researched theoretically and experimentally and a new measuring
method of the yield value of clay is proposed.
In this report rheological terms are used following the International Rhe-
ological Nomenclature ; for instance, the time effect of the deformation of
clay is called "flow" but not "creep", because the latter term implies com-
plete recovery.

Chapter 1 The Flow Mechanism of Clay

1. Introduction

In this chapter, the fundamental relation of strain character of clay under

constant external stress, that is, the flow character of clay is deduced from
the micrometric standpoint applying following 3 assumptions.
(1) The viscosity of clay exerting on the flow of clay is assumed as
the structural viscosity which is derived by applying the statistical mechanics
on the frequency Of the mutual exchange of position between a clay particle
and the neighbouring hole or the point of irregularity in the arrangement of
particles (c.f. Fig. 1.1).
(2) The mechanical model of clay is assumed from the consideration on
the structure of clay as the model as shown in Fig. 1.3 which consists of the
 4

modified Voigt element with series coupling of spring El. The coefficient of
viscosity 7;2in the modified Voigt element is represented by the structural vis-
cosity deduced from (1). The modified Voigt element, therefore, is composed
of a spring element E2, a dashpot element r2 and a slider element ao connected
in parallel coupling each other whose elastic modulus, coefficient of viscosity
and restraining resistance are E2 (=constant), 7j2( = coefficient of structural
viscosity deduced from (1)) and ao (=constant) respectively.
(3) The fundamental relation of flow of clay obtained applying above
assumption is supposed to be valid so far as the external stress applied on
clay is less than either the stress of pre-consolidation or the stress which breaks
the structure of clay.
The structural viscosity of clay which was deduced in this paper may be
said to be a extention of the viscosity deduced by H.Eyring. Thus obtained
formula will be ascertained with experiments in Chapter 2, and the results
of this formula exactly agree with the experimental results of Chapter 2.

2. Structural Viscosity proposed by the Authors

Before the structural viscosity proposed by the authors is explained,

it would be convenient to explain about A.V.Tobolsky and H.Eyring's
theory,2) which is founded on the concept that the flow process of polymeric
material is produced by exchange of positions between particles and their
neighbouring holes. Such a set of a particle and its neighbouing hole which
exchange their positions is called a unit process of deformation.
The detailed mechanism considered by Eyring" in his rate process
theory is depicted in Fig. 1.1. Since the point of irregularity in the arran-
gement of the particles can be cosidered as the hole, the exact size and shape
of which need not be known in advance. The flow process is then considered
to be the jump of a particle which is denoted by (a) in Fig. 1.1 from its
potential energy well into a neighbouring hole by (b), that is, from one eq-
uilibrium position to the next.
In the absence of external stress, forward and backward jumps will be
equally frequent. Eyring obtained by statistical mechanics the following ex-
pression for this jump frequency I:

J.=—KTexpHE°-)

h
 5

cr
With external
Stress

400 (6)tress
.00••0 E.\
e
thout
+c/a sxstress
reersnsa

o
‘A,I -CICT
(6)
Fig. 1.1 Schematicsketch of the Fig. 1.2 Representationof potential
structurecomposedof particlesand energybarrierwithor withoutexter-
holes nal stress

Where Ic is Boltzmann's constant, T is the absolute temperature, h is Planck's

constant and Eo is the free energy of activation for the jump, that is the
potential energy barrier as shown in Fig. 1.1. In the presence of a shearing
stress a the potential barrier will be deformed so as to favour the forward
jumps as shown in Fig. 1.2., i.e., the potential barrier is biased in the
direction of motion by the amount —ao-and by the same amount with oppo-
site sign in the backward direction. Therefore the net frequency of jump or
the net rate of motion should be represented as

T {—(E0—ao.)T—
—h-exp (Eo+(La)}
.
Th exPlT

=((ao- ) —hTexp—E5)sinn
?Tr(2.1)
where a=2/2N, and 2 is the average distance projected in the direction of
stress between equilibrium positions and N is the number of unit process of
deformation in a unit area of cross section perpendicular to the direction of
stress. And also, if there are n unit of processes of deformation in series per
unit length in the direction of stress, the rate of shear D should be

D=C el)sinh(x-C1-26-.)
2,11T (2.2)
—
hL..=--

The equation for the apparent viscosity coefficient of flow would then
become

^
 6

V=5=
a 1Et, a
cTJ
sinh asT) (2.3)

The most significant result of the Eyring's theory, therefore, is to provide

a basis for equation (2.3) which has been employed so successfully to repre-
sent the structural viscosity and the viscosity-temperature dependence of ordi-
nary liquids at ordinary temperatures.
In order to represent the rheological behaviour of the material of more
complex structure, the more general relationship can be derived by assuming
the generalized model of the material which consists of r parallel groups of
different types of unit processes of deformation described as follows and by
applying the W. Kuhn's theory" which gives the stress a and strain e of the
material as the sum of the each ,elementary stress at and strain gi of the
group of unit processes of deformation of type i, thus : a= Xaf and = Xei.
In the group of unit processes of deformation of type 1, there are Ni units
of parallel process of deformation of type i in a unit area of cross section per-
pendicular to the direction of stress aa, and unit stress as distributed on N, units
(viz. the average force acting on each unit is a"), and also there are such
nr unit of processes of deformation of this type in series per unit length in
the direction of stress Cl.
Applying these assumptions to Eyring's theory which is represented by Eq.
(2.2), Tobolsky and Eyring led the following equation relating the rate of
strain dei/dt with the stress at.

der221nIKT—E01.Maas
dt = hexp7.—sink)=Ai'sinh (131'at)
whereRhAi' =-22'n'T——01)(2.4)
exP
KT
Bfi =
2NIc T

In the treatment above stated, the number of the unit process of defor-
mation is considered to be independent of applied stress, and does not refer
to the existence of restraining resistance against flow which acts on particles
at all.
On the solution of rheological behaviour of clay, in this paper, the
authors have taken the restraining resistance acting on clay particles in con-
sideration and further have assumed that N1 or ;If varies with a function of
total stress (a) applied on the clay skeleton. Consequently, the unit process
 7

of deformation is restrained to participate in deformation until the total stress

overcomes the restraining stress ao, and when 6' become greater than ao, it is
assumed that the number of unit processes of deformation which participate
in deformation increase in direct proportion to the value of (a —ao). The latter
assumption is based on the consideration that no special regularity in the
arrangement of clay particles is expected.
Hence, the above assumption can be expressed in the following equations.
where 0<cr<ao ni or Nt= 0
cro<a a(tt —co) (2 .5)
Ni= b(a—a0)
where a and 11are constants respectively.
On the basis of above assumptions, the frequency of forward jump j+
should be

j teT —fEot- 2b(a —ao) (2.6)

+= — exp T
and the frequency of backward jump j- should be

j -= hT—1"
1172Atql
)1 ab-(6- 60)1 (2.7)
exp T

where Eot is the free energy of activation of the unit process of deformation
of type i.
Therefore the net rate of jump should be
.2KT(—E01\ tat
5=1+—i-- exP ) s'12 bcT(a— co)} (2.8)
Correspondingto Eq. (2.4), we obtain the following equation between
the rate of strain dei/dt and the stress at.
dot 2 Ata(a —ao)cTEot .2tat
dtexpTt20.0))1
=Ai(a—ao)sinh(B“"(2.9) — ao

where
2,117.(—E
At = — h— -expot ), Bt=2btcT

Hence, for the unit process of deformation of type i, the apparent coe-
fficient of viscosity proposed by the authors is represented as
1
t B
Atsinh
(tat(2.10)
a — Co,
 8

3. Mechanical Model and Flow Mechanism of Clay

As the results of many preparatory tests on clay, it has been observed

that the clay behaves visco-elastically but has a lower yield value corresponding
to the restraining resistance, below which no flow deformation can be detected.
From these test results and following consideration, the authors adopt a spe-
cial model as the mechanical model of clay. This special model, as shown in
Fig. 1.3, consists of a series coupling of a spring
element E1 whose elastic modulus is El. and a
modified Voigt element composed of a spring element
E2, a dashpot element va and slider element ao con-
E2 '12 nected in parallel position each other, whose elastic
modulus, coefficient of viscosity and restraining
resistance are E2, v2 and ao respectively. The co-
Fig. 1.3 Mechanicalmo- efficient of viscosity v2 of the Voigt element is
del representing clay represented by the structural viscosity expressed in
skeleton E
q. (2.10).
On the construction of the above mechanical model, the following consi-
deration may be applied. Structure of clay is supposed as it is composed of
aggregates of matrices accumulatively composed of very fine fractions, so
that the elastic properties E1 must be ascribed to the elastic repulsion
between matrices as if elasticity of sand mass is caused by the elasticity bet-
ween sand grains in it. On the other hand, as the application of external
force causes the exchange of positions between particles and neighbouring holes
in the matrix, clay shows the structural viscosity 722 on the flow process.
Although the frequency or the number of activating unit process of de-
formation per unit time may change accordance with (a —ao), it may be su-
pposed that the total existing number of unit process of deformation in the
matrix is unchanged independent of the magnitude of external stress so far as
the external stress is less than the stress which breaks the structure of the
matrix, because the flow process of clay is only caused by the mutual exchange
of positions between particles and holes in the matrix so far as the external
stress is less than the pre-consolidation stress. Since the elastic property of
the matrix, furthermore, is supposed to be dependent of the elastic property
of the part of the matrix in which no unit process of deformation exists,
the elastic modulus E2 of the matrix is assumed to be constant in spite
 9

of the magnitude of external stress.

From Eq. (2.9) and (2.10), the strain rate de2/ dt and the apparent
coefficient of viscosity v2 for this mechanical model are given as follows.

dtde2
=A2(a
—ao)sinh(B2
0.20) (3.1)
722= (3
.2)
A2sinh(-8262)
a —ao

where E2 is strain of modified Voigt model and a2 is applied stress on the

dashpot v2. The relation between total strain e and time t, can be obtained
by solving the following simultaneous equations,

fe•=.21-Fez
a= (3.3)
(3.4)

a=e2,(aco)
c.2-r
132{A2(0.1-110)
di}F(7° (3.5)
Transformation of Eq. (3.5) leads to :

de2
dt=A2(aao)sinh
[B2{(a-6°)—e2E2}]
(a— co)
Integrating, we obtain
de2 = A2(a —cro)dt-F C
sinh
[132{(6
6°)€2E2})
(a —ao)
from which

—B(a—o)1132{(a—ao)—E2E2)]
__
A2(a
—ao)t±C (3.6)
2E-2log tank2(a—ao)
a—ao)
Thearbitrary
constant
Cis equal
to C=(132
.E2log
tanh(B2),since
we have e2= 0 at t= 0.
Substituting this constant into Eq. (3.6), we obtain.

logtanb
[B2{(aa°)—
2(a—ao)
ezE2}]=
A2B2E2t±
logtanh(
B22) (3.7)
From Eq. (3.7) we get the relation between e2 and t

e2— E(a—
ao)2(aao)tanh-'lexu(A2B2E2t)tanh(B2
(3.8)
2 B2E22'
If B2/2 is larger than unity and

2B2>2B2{ (a(—ao)— ezE2}>1

a —ao)
Or

1
 10

0<e2<(3 (a2 —a)o(2B2 —1) .9)

.B2Ez
approximate result is given as

logtanlog tanh(-2
[;(6,;20.—o,e2E2}] B2)
=lOgri —
L,(a
2exp[—B4(a—'5°).-82E21]+
—ao)L
2exp,-._213.2{
.
(a— co)—
(6 —co )
e2E2}
] ]
—log (1 —2 exp (—B2)+2 exp ( —2./3z) J

-=-:log[l
—2exp[ —132{ ((a—')—ezEzrj+
—co)—o)J 2ex,r1...(r—2B2{(aa—
aoa)
—e2E2}] l
- B,{ (o--cro) —ezE2P
4—2exp[ (a —ao) ,..1,
Accordingly,

—2
expr
L
,
—
B2{
(a — ao)
—ezE2}]=
(a —co)
_A2B2Rat (3.10)
then, strain of modified Voigt element ez will be
e2=R(a —ao)(a —ao)k
2+B2E2gAzi132E2t (3.11)
where Az' = A2/2.
Substituting Eq. (3.11) into Eq. (3.3), we obtain
e=++la(a —ao)(a —ao)(a —ao)
Et ogA2 '132E2+logt(3.12)
Ey BzEzBzEz
Eq. (3.12) is the "flow equation" representing the relation between
strain e and time t for the condition (3.9), and in this equation e is propor-
tional to log t.
If ez beyonds the limits of (3.9), viz.
e2>--(2.B2—
(a)—co 1) (3.13)

(o2.82E2 (a —co)

C.5
\rt.)•
T1Eq. (3.8)
ez=B
becomes
y(3.14)

VS 4 b at t—).°0
li
6. bid
SubstitutingEq. (3.14) into
brurEq. (3.3), we obtain
a _,_(a —ao) (3.15)
109t Et-.=El.'Ey
Fig.1.4Schematic
diagram
offlowstrainAccording to the theory pro-
E vs.timet posed by the authors,therefore,
 11

the curve representing the relation between the strain of clay and the loga-
rithm of the loading time should approach to a horizontal asymptote as shown
schematically in Fig. 1.4.

Chapter 2 Studies on the Flow Tests

4. Introduction

In this chapter, the authors report on the experimental results and their
considerations of a series of compression tests at constant uni-axial load, carried
out on Osaka Alluvial Clay, in order to refer and ascertain the new fundamental
formula derived in the preceding chapter.
As the experimenting apparatus for flow tests of clay, there are various
kinds of apparatus such as apparatus for simple shear, torsion shear,
unconfined compression, triaxial compression or other testing method and
some of them have been used by some researchers. But, in this case, for
simplicity and general applicability of testing process, the authors performed the
compression flow tests with triaxial cell applying each constant uni-axial load,
as R.Haefeli" performed.

5. Some Data of the Clay Specimen and the Compression

Plastometer

Undisturbed clay specimens used for the compression flow tests were
obtained from the Umeda Alluvial Clay Stratum in Osaka City by means of
the thin-walled sampler with stationary piston. The tube of this sampler has
the inner diameter of 73 mm, the thickness of the wall of about 1 mm and
the length of 760 mm. The soil profile of the Osaka Alluvial Layer is shown
in Fig. 2.1, in which clay stratum extending between a depth from 6 to 18 m
below ground level is called Umeda Alluvial Clay Stratum. As the clay in
this stratum has the sensitivity ratio of about 3 in Terzaghi's representation,
it belongs to the category of an ordinary clay. The results of the physical
tests of the clay are as follows : — specific gravity : 2.67, L.L. ; 8-3-63q,
P.L. ; 36 --25q, natural water content : 92-58%, void ratio : 2.52--1.84,
degree of saturation : 100%, and the grain size distribution is shown in Fig.
2.2. The maximum pre-consolidation pressures measured by oedometer tests
are as follows : — sample No. 4 ; 0.88 kg/cm2, No. 10 ; 1.60 kg/cm2. As
 12

_c otr,4Mechanic) Consistency&these pre-consolidation pressures

12, 7,-,0 a
a„E,,‘''
-o u3'E6- Analysis Natural W.Cont. are equal to their overburden
• 20 4o 66do% 40 E'oio Kic.% pressures respectively,the clay in
2 . ---. G A'..,„:;;;::::'';: this stratum is normally loaded
":/::•1:::: .:__---a
. § „: ::::::/.. :•'.:::
'... • --- clay.
In the caseof the compression
flow tests, a triaxial compression
6„ .0•-5, ,,,,/,. apparatus was used as compression
• 2
. 45 ',.” ,,' :• •4
• I,
plastometer. In these tests, a con, 8

5,,,-,,,,stant uni-axial load was applied on

Jo. 6 'Clay :Silt ::: • sl the clay specimenin a triaxial
7 ///
s11•4,;apparatus
:://,',::::ii!fr.. whoseschematicappear-
12-q'i•'':::::' ance is illustratedin Fig. 2.3,
•;0r7, •<is where the clay specimen is denoted
,,,
,,,,,:
14 12 by A, the pressure chamber B,
. 13 ,,, 0
14ci• ,,'

the loading lever (lever ratio ;

16 i 1/10) b
• G',:%/-"::i• y C and the dialgaugebyD.
18.17
-1—/,;;;;:i'l...•The
20. 0,.'''.:..:.'.:
lg,,',."by ..i. clay
ed shaven
samplespecimen
out
to from was
havethe obtained
undisturb-
a cylindrical
Fig. 2.1 Soil profileof the Osaka form with a height of h =8.0 cm
AlluvialClay
and a diameterof 3.5cm. On the
compression test, longitudinal displacementzlh (mm) of clay specimenwas
measured by the dialgauge. Then the strain a is given by zlh/h so far as

a. IllI
100 1
_
„..iiIIIIIMM
..,....

E .^411111111111
C
;-; 50

(r)
HI Ai .

Nor...-...r#411
—.
• 10'4,,

w
0 Fig. 2.3 Schematic
102 2 54 5 102
2 54 5 10 2 54 appearance of corn-
Fttrticle size mm pression plastoine-
Fig. 2.2 Grain size distribution curve ter
 13

displacement is assumed to be infinitesimal,

In these tests, compression loads were so restrained as the vertical stresses
a in the specimens to be less than the pre-consolidation pressures to avoid the
effect of consolidation. On a series of compression flow tests, to avoid the
effects of the stress-history given to the specimen by the compression test,
each "fresh specimen" obtained from a same clay sample which had never
received any compression test was used for each constant uni-axial load. To
protect from the evaporation at the surface of the clay specimen, the specimen
was coated by grease, and covered by thin gum membrane and the pressure
chamber was filled by water for the duration of the tests.

6. Flow Characteristics

Using clay samples above stated, compression flow tests were performed
in which stress is kept as constant as possible during the tests of 24 hrs. The
results of the tests are shown as strain—time diagrams in Fig. 2.4, 2.5 and
2.6, in which each abscissa shows elapsed time t in logarithmic scale and
each ordinate strain e respectively.
From the strain—time curves, the features worth noticing are as follows.
(1) The flow strain a increases proportionally to the logarithm of time
(log 06'7, if the .applied stress a is smaller than a certain critical value.

180 •
•

1405'-‘14701111111.er
2.0 160

..-•041'''''

".01.5(2Oppm,11.1
< 100

8°111111m•-1011111W^"”--1111111.111
04_,/111

60 I- I- I
0.5 40 0208

20 C' • . - • • • - •

"5::V:1.1101...1101.1111Mill
0 0 li•MilaWilliM10.001101111.111.111
10 1.0 10 16 icf
t min
Fig 2.4 Flow straina vs. time I curves(sampleNo.13, watercontent; 65%)

•
 14

2.0- 160
140 _—^11111111

0
CC)
15- E120
70

<
100
'0 428 TEI
,080 ••.
0.51f, . • • ' • MI
- 60 '
0250.9/ow
JIMMIMMOR'F"v-•—•— •

05- 40
• °161'8/'''11110.0".101001000.11.
0124K9/Mr
I
20—
0.051")=1"..mioNomomeinut
0
0 IO 10'
t Min
Fig. 2.5 Flow strain E vs. time t curves (sample No. Bi water content ; 75%)

- 100

10- E 8•
co0.265
70.000.01.0.• KW •
%.1 _c60

05- 400.16.6 • C"2?

KV'n'
1111.1111111.111/0111111..1
2008.5K9/601111.1.11111111
0- 010.
=mom 1.0 Id 10 t• t
miON11111
min
Fig. 2.6 Flow strain E vs. time t curves (sample No. 4, water content ; 92%)

Therefore; it may be confirmed that Eq. (3.12) representing the linear relation
between the flow strain and the logarithm of time is valid for the stress
range up to a certain critical value.
(2) If the applied stress is larger than a certain critical value, the curves
in these figures rise concave upwards and this tendency suggests the occurence
of failure in future after a certain duration.
It will be discussed in the following Article 7 that each critical value
coincides with the upper yield value.
(3) The slopes of these lines related in (1) increase with applied stresses,
and it will be showed that the tangent of the slope angle of each line is
directly proportional to the applied stress as shown in Fig. 2.7 whose ordinate

•
 15

and abscissa show the tangent of the S

ample
slope angle of the line and the applied —o—o—
No.t3
—o—o—
No.
81
stress of the curve illustrated in Fig. 2.4, 0
.15- No.4
2.5 and 2.6 respectively.

7. Strain-rate Characteristics
a 1 r • /
/Inthis
article,
theeffects
ofstress
a or time t on strain-rate de/dt are an- WON
alysed. 0,05
From Eq. (3.12) the strain-rate can
be represented as
de (a —cro) 1
dtB2E2 t (7. 1) 0 02 Cr race
04
Taking the logarithm of each term, Fig. 2.7 Relations between the
we getcoefficient b and applied stress a-

log(de
—log
611)=(a—ao)
Bz8,log t (7. 2)

On the other hand, Eq. (3.12) is transformed :

e=-6 +(a—ao)2.3(a
E • ±(7. —c)olop-A2'B2E2t 3)
2 B2E2
= ad-b login t (7 . 4)

where a=+-Ea(a —ao)+2.3(a—cro)

2 B2E2I -OgleA2 'B2E2

b= 2.3(a
B —ao)
zE2

From the above treatment on strain-rate equations (7.1) and (7.2), the
following characters of strain-rate can be presumed so far as the flow of clay
are ruled by the theory stated in Article 3.
(1) The relation between strain-rate and stress for various constant values
of the time is represented by straight line.
(2) The logarithmic relation between strain-rate at a constant stress and
time is represented by straight lines.
These characters of strain-rate are ascertained as follows from the experi-
mental results on -flow of clay shown in Fig. 2.4, 2.5 and 2.6.
 16

1.0- 80 Elapsed time

-o-o- 45° •

-0--0- 15'

OE07_I. GO
70 •

-o •
0.5- 40

025- 20

°- 0 a6 0.2 0.4
0.6 0.8
cr
Fig. 2.8 Viscositycharacteristic (sample No. 13)

The consideration for (1) :

In Fig. 2.8 the strain-rate de/dt obtained from Fig. 2.4 (in the case of
clay sample of No. 13) is plotted against the applied stress a for various
constant values of the time, taking the time as t=20 sec, 45 sec, 1.5 min and
15 min. This figure shows that the curves are linear within the stress range
up to a=0.45 kg/cm2, whereas they rise concave upwards at a>0.45 kg/cm2.
Moreover all linear parts of these curves concentrate at a point on the a-axis.
The abscissa of the point of intersection of these lines with the a-axis is
called the lower yield value ao, below which the strain-rate equals to zero and
no flow deformation takes place. The stresses corresponding to the points of
first inflection of curves or the upper end points of linear parts are equal
for every lines and this stress is called the upper yield value a., below which
no failure can occur even after long periods of time elapsed. Thus the
lower and upper yield values can be uniquely determined by strain-rate--stress
diagram as shown in Fig. 2.8 which corresponds to so-called relationships
or viscosity curves. As above stated, therefore, the part of the curves between
ao (=0.025 kg/cm2) and Cu (=0.45 kg/cm2) are straight lines and within this
range Eq. (7.1) is valid.
 17

From the same experimental viscosity curves as Fig. 2.8 obtained for
samples No. 4 and No. 8, we get the values of Co and Cu, which are shown
in Table 2.1. Within the stress ranges up to the upper yield values tabulated
in Table 2.1, the s---log t relations of these clay samples under a constant
uni-axial stress are linear as shown in Fig. 2.4, 2.5 and 2.6, and in these
ranges of stress Eq. (3.12) is valid.

Table 2.1
Sample No. •okg/cm2 I Oukg/cm2
13 0.025 0.45
8 0.020 0.33
4 0.020 0.19

The thick lines in Fig. 2.7 show the relations between the coefficient of
b in Eq. (7.4) and the applied stress a. The value of b, in this figure, was
obtained as the tangent of the slope angle of each straight line due to the
comperssion stress smaller than the upper yield value Cu in Fig. 2.4, 2.5 and
2.6. Since the experimental data lie in well agreement on thick straight lines
representing the relations for stresses smaller than the upper yield value,
it can be said that the value of B2E2 in Eq, (7.4) does not depend on stress
within the range up to the upper yield value. Therefore, the values of B2E2
can be computed from Eq. (7.4) and the results of calculations are shown in
Table 2.2.

Table 2.2
Sample No. B2E2 103kg/cm2 Water content % Temperature °C
13 0.688 65 22.4
8 1.000 75 2.7
4 0.712 92 2.7

The value of B2E2 which influences the strain-rate depends on a large

extent on the water content, temperature and other factors of the testing
procedure. If we consider the effect of temperature on the value of B2, it
may be presumed that B2 decreases with increasing absolute temperature,
because B2 is represented as B2=22/21w T in Eq. (2.9). The dotted lines in
Fig. 2.7 show the relations between the tangents of slope angle of tangent
 18

lines on each curve for the part of a>au at t=1 min in Fig. 2.4, 2.5 and
2.6 and applied stress a. These dotted lines do not hold linear relation as
thick lines do in Fig. 2.7.
The consideration for (2) :

to to'
=•

_ .
_ • 0,
_ so
6't
•• f._
.». allure
. - 10
,_. 1d;°.j09 t, ••
• . .. .
• I —
.-^ -) _.
.....•
3o 1..k, -,
., ,
,,,
,s0'4. .
.0•q5..• . • ••

1‘...•6:1.01..>_,.:.
_ •. • .
- C
.)°,
,,,,, • . •. I • • ••

Id-101.0 t M.n !Ol— 10

1.0
,•1
Fig. 2.9 Strian rate dEldt vs. time t curves (sample No. 13).

1c
5F
)1.014 • •

Id_%lg
101
i-•^°.0
0'
2 1•—o.. 0•-t6-.,

..I..1%
.-. -

Ai t-2 - o. \,
-98-,6 o. 4

Id- - 1
-

16' ' '''.'

la 10to t Mt10' 10
n
Fig. 2.10 Strain rate dEldt vs, time t curves (sampel No. 8)

_ _ •
 19

Fig. 2.9 and 2.10 show the logarithmic relations between the strain-rate
de/dt and time t obtained from the strain--time curves in Fig. 2.4 and 2.5
respectively. The straight lines in these figures represent Eq. (7.2) computed
by using the values of B2E2 tabulated in Table 2.2. Since the experimental
data lie in well agreement on straight lines representing Eq. (7.2) for the
stresses smaller than the upper yield values, it can be said that the relation
between the strain-rate and the time is quite satisfactory expressed by Eq.
(7.2). On the other hand, for stresses larger than the upper yield value the
curves rise concave upwards and the failures occur after a certain duration as
shown in Fig. 2.9. In this figure, it is found that curves for compression
stresses of 0.509 kg/cm2 and 0.676 kg/cm2 rise concave upwards and for the
curve of 0.676 kg/cm2 failure occurs after 96 min.

8. Stress Strain Characteristics

From Eq. (3.12) the strain e can be represented as

e=(ao-0)(1+1 +B2E.2
logA2'.B2E2+B2Ez
logt)+so (8.1)
where eo =E6°-

Since so nearly equals to zero, above equation becomes approximately

6—(6-6°)W-E
.±B2E21°g
Ao'
1
B2E2+B.0
logt)
1 11
(8.2)
This equation shows that e and (a —ao) lie in proportional relation if
elapsed time t is equal.
From Eq. (8.2), as time-function or stress—strain relation regarding to
time, 0(t) are given as follows.
1 1+E+ 1
(a —e ao)0=
= 0( (Et+E.,13..!-logA
E2go'B22B,E„,log
t) (8.3)
= a' +b' login t
11, 2.3 ,
where a,,-r-logio E2B2E2° A2'B2E2 (8. 4)
b' — 2.3
BoE2
Refering to Eq. (7.4), the coefficients of a' and b' in Eq. (8.4) are
represented as a'F-2(2/(a —60), b' =Ka— ao).
In Fig. 2.11(a) and 2.12(a), each stress in Fig. 2.4 and 2.5 is plotted against
the strains for various constant values of the time. In these figures values
 20

0.6
(a) Thl....
'
, 0.4
71 iii/rrA111111^^
b

Elapsed time
0.2'Io'
vi"' .-o—o-i'

.': '-0---e-
100
—o--0—24'
1,-
0 100 200 . 300
.o.41ics=n—

1.0 20 3.0 ' 4.0

E Id'
"
.k (b)
7E
_3.51-,
,.... i
PW Woi

g 3.0 I- I

2.5
,...iiiillall
16' 1.0 10' 10 10
t min
Fig. 2.11 (a) Relationsbetween applied stress , and strain E
at various instants I (samale No. 13)
(b)1Variation of 0(1) with time
,c)

1.1•111111jra
04Willlow"
o. . . .
.1
j • • Elapsed time
0,2

Q
zi?111a•100,
Zo o
o

to
o 10
o . 1.

o 24•

0 50 100 . 150
oh len.
0 0.5 1.0 L5 20
E le

7b)
.,.
:,.3.5-
g -
.3.0- I •
_
.
25_ .
10' LO le 10' 10
t min
Fig. 2.12 (a) Relationsbetweenapplied stress 0• and strain a
• at various instants t (sample No. 8)
(b) Variation.of 0(t) with time

a O 1
 21

of parameter t are 1=10 see, 1, 10, 100 min and 24 hr. All curves concentrate
at very near points to the lower yield value ao on a-axis. The lower part of
these curves up to the first inflection points or the upper yield value au are
approximately straight lines. Within this range, therefore, Eq. (8.2) is valid
and the ratio e/(a—ao) or 0(t) plotted against time must give the same
time-function 0(0. The values of 0(t) obtained from Fig. 2.11(a) and
2.12(a) are plotted in Fig. 2.11(b) and 2.12(b).
The experimental results of E.C.W.A. Geuze and T.K. Tan8' had indicated
that 0(t) can be uniquely determined from torsional tests at constant moments
and further that the time function in simple shear is identical with the time
function. 0(t) in torsion.
The thick lines in Fig. 2.11(b) and 2.12(b) are 0(t)-lines obtained by the
calculation. of Eq. (8.4) substituting a/ (a —ao) for a' and B2E2 in Table 2.2
for b', and they are represented by each straight line on the semi-logarithmic
coordinate system. By above stated calculation, these lines are given as
follows;
For Fig. 2.11(b) (for sample No. 13)
0(t) =2.65+0.33 logio t 10-2/kg/cm2
For Fig. 2.12(b) (for sample No. 8)
0(0=2.80+0.23 logio t 10-2/kg/cm2
As all experimentally obtained plotts of 0(t) perfectly lie on the theo-
retically obtained 0(t)-lines as shown in Fig. 2.11(b) and 2.12(b), it may be
said that the- stress—strain characteristics of clay can be deduced from the
theory established by the authors and containing elements of time-function
0(t) which was introduced as the experimental results by E.C.W.A. Geuze
and T.K. Tan.° are clarified.

9. Characteristics of Flow under Repeated Load and Flow

Recovery

In order to get an insight in the flow mechanism of clay, it is of

fundamental importance to study the behaviour of the flow strain due to
repeated load and flow recovery after the load is taken away. For this purpose,
repetitional loads of same magnitude and same loading duration are applied on
the clay specimen at same interval of removal of the load according to the
method stated in Article 5.
 22

Fig. 2.13(a) and (b) :show the schematic relations of the stress—time and
the strain--time respectively. The flows and flow recoveries due to the 1st,
2nd, and nth loading ar.d load-removing are called the 1st, 2nd, and
nth flow and flow recovery in this paper.
Fig. 2.14 and 2.15 show the ex.
(O. amples of flow and recovery curves

b plotted on the semi-logarithmic pa-

Isi 1st 2nd 2nd 3rd 3rd
leadingload lordingload loadin8load pers in which the repetitional stresses
removing removing reinotnno
smaller than. the upper yield values
0 t, t, tim were applied as noted in both figures .
e t
From the flow and recovery curves

to \46cv,w0. in these figures, the following features

are noticeable.
a_ (1) The tangent of the slope an-

„awrygle of the 2nd flow curve which is cor-

t, t, t, t,ts respondent to the coefficient b in. Eq.
time t (7 .4) decreases remarkably in corn-
Fig. 2.13 (a) Schematic relations of
stress vs. time due to repetitonalload- parison.with that of the 1st flow curve.
ing (2) The value of b gradually
(b) Strain vs. time curve due to re- decreases with increasing number of
petitional loading
repetition of loading.
(3) The strain during recovery decreases in direct proportion to the
logarithm of time and the tangent of the slope angle of each line representing
the strain and the logarithm of time is constant independent of the number
of repetitional removal of load.
Above features can be explained theoretically as follows. The flow equa-
tion for repetitional loading test should have different relation from Eq. (3.1),
because the flow due to the exchange of positions between a particle and its
neighbouring hole only takes place at the stress larger than (a —ao—erE2),
where sr denotes a residual strain at the beginning of the loading.
Then refering to Eq. (3.1), the strain rate del/ di is given as follows.
de2
d =Byar, oErp.)
t242(a —a—Esink((a
— aoErE2)
=A2(a
—ao—
erE2)
sin.h-1132(6—
(a —ao6°
——62E2)1(9. 1)
erE2)
and refering to Eq. (3.5)

_
 23

- 74 (0'
OPI- 8th . mall
1111111110.11w
- 70

0(a)
MP - e 011,001.1' III.....
E eaEallii".11111w--.4'1
11111.111111111
E
II.: -.: '208-'9- - . 3 cd•1•

- .

(l.) 4
0.7
. • -

06El .i 50
'

—
-

06-°
- 50
\5‘•

I:min,. _0::::
:_z_E42
04 .....wetIIIIIII

7th=..1_15'1.00
.
4+,1a11.1611 .2..........to-
10'34 < .

Fig. 2.14 Flowand

MU
22 rb)-30 2230.vo"Q
Kr
10
1.0
t

recovery curves due12lo"

min 10
t min
10 0.2-
-
,-

1.0
(b)

30
t Min

to repetitional loading (sample ; No. 11, Fi t mir,

water content ; 71% , 0-=0.319 kg/cm2, g 2.15 Flowandrecovery curves due
to repetitiondloading (sample ; No. 9,
0-.=0.38 kg/cm2, loading time ; 30 min,to content; 74% , a =0.218 kg/em2,
interval of removal of loading ; 10 min.) cr.=0.34kg/cm2. loading time ; 5 min,
(a) Flow strain vs. time curves interval of removal of loading; 5 min)
(b) Flow recovery strain vs. time curves (a)
(b) Flow
Fl strain vs. time curves
ow recovery strain vs. time curves

0.=4E2+
(a—
ao—
132
er.E0)
sink''{---Fao
. {A2(a —co1—etE2) de,,
(
dt I
(9.2)
Integrating Eq. (9.1), we obtain
dE2
.,/a—ao—ezE2)('
=A.(a
—ao—
ErE2)
idt+C
si'l(cB2( —coerE2)i

from which
—(a
—(to—
zrE2)log
.B2E22(0_6._
tanh{.132(a.—
ao—eoE2)1=._A2(a—
erE2)f.
ao—
ErE2)t+C
(9.3)
Thearbitrary
constant
Cisequal
to(a —
co—erE")B
2E22 B2)since
,logtank(B2\
we have a2= ar at t= 0.
Substituting this constant into Eq. (9.3), we obtain

1
•
 24

logtanh{-(
.130(a—0E.)
a—
a_60o_tr-E2")
= A2B2E2t+
logtarh(11) (9.4)
From Eq. (9.4) we get the relation between e2 and t.

e2=tank-Oexio(
(a —ao) 2(a— ao—SrE2)
E2 B2E22
—ABoEtan.h(1}-(9.
5)
If B2/2 >1 and

2/32>2132(aao—
( a2E2) > 1
a —ao —27E2)
or

er<aa< 2B '5°)(2B (9.6)

2E222B2
Eq. (9.5) is given approximately as

e2=(a E
—ao)(a— ao—ErE2)•og A21.82E2t (9. 7)
2 B2E2
where A2/= Az/2.
Substituting Eq. (9.7) into Eq. (3.3), we obtain
6 +(6.6°)— a°erE2) l
6= E 1 B2E2B2E2ogA2'B2E2+ (a —— E'E2) logt (9. 8)
E2
The tangent of the slope angle of the straight lines representing a
relations for repetitional loading test is expressed by

b =2.3(a B—ao—erE2) c (9. 9)

2E2
where c = 2.3(a —ao)/B2E2 and d = 2.3/B2 are constants.
As a, increases with the number
Q2 of repetition of loading, Eq. (9. 9)
indicate that the value of b decreases
o with increasing number of repetition
0.1 • of loading. This gives rize to the
•
suggestion that the strain-hardening ef-
fect and increase of apparent elastic
0 5 10 15 20
modulus by deformation of clay can
0 0102.0.3E
,lo' be explained froni the micrometric
Fig. 2.16 A relation between the standpoint .
coefficient
b and residual strain Er
obtainedfrom Fig. 2.15 (sample An experimental reality of Eq.
No. 9) (9 .9) can be showed by Fig. 2.16 which
ohtain.ed' from Fig. 2.15. and illustrates

)
 25

the linear relationship between b and er. Since this linear relation gives
b = c —d •sr
= 0 .00175 —0.50 . Or,
the calculated values of E2 and B2 are as follows : E2 = 56 .5 kg/cm2, B2 = 4.6
(In this case, a = 0.218 kg/cm2, ao = 0.02 kg/cm2).
Besides, if €2 is beyond the limits of (9.6), viz.
—ao)
213 (2B21)+ (9. 10)
2E2,2B2
Eq. (9.5) becomes

e2= E—ao) (9. 11)

2
at t—>0o
.
Substituting Eq. (9.11) into Eq. (a)
10- 800

(3.3), we obtain ,Collure

a— cro
e) 8-
E2(9. 12)
According to the theory proposed by
the authors, therefore, the strain —log/ 6- 500
curve should approach to a horizontal -
asymptote
represented
byEq.(9.12). 4- It
2nd 1111111111111111
Above mentioned results were obtain--
St

ed by experiments performedat the 2-

stresses
lowerthanthe upperyieldvalue. 1.0 50
In orderto givea further illustration ont min
rheological
behaviour of clay,repetitional (b)
flowcurvesobtainedby the experiment CO4 5--,0E
403
WWWWIMPOIMMINO
MEI
performed at a stresslargerthantheupper . 111111111111 - <

yieldvalue
areshown
inFig.2.17(a).zooMIM
2 Curvesshown in Fig. 2.17(a) are flow111111111111111.11
curves obtained by repetitional loading t 1.0
min
with a stress of a=0.436 kg/cm2, larger Fig. 2.17 Flow and recoverycurves
than the upper yield value of a.= 0.34 due to repetitional loading (sample
No. 9. 0---=0.436kg/cm2, au=-0.34 k
g/cm2, and it is found that all curves kg/cm2, loading time ; 5 min,
rise concave upwards on the semi-logarith- interval of removal of loading; 5
mic paper and failure occurs on the way min)
Flow strain vs. time curves
(a)
of 4th flowing. It deserves attention that, (1) Flow recovery strain vs. time
as above facts indicate, the effect of curves
 26

repetitional loading on clay may be harmful if the applied stress is larger

than the upper yield value.
In spite of non-linear character of the flow curve due to the repetitional
stress larger than au, each flow curve due to recovery after the removal of
such stress holds linear character, and the each slope of the flow recovery
curve is nearly equal as shown in Fig. 2.17(b).
The theoretical analysis of the flow recovery is as follows. Since the
restraining stress ao acts in the opposite direction to the recovering stress e2E2
after removal of the stress, we put ao= -ao and a=0 in Eq. (3.5). And ei
in Eq. (3.3) is equal to zero in this case, because the total recovering defor-
mation of the spring element E1 shown in. Fig. 1.3 takes place immediately
after removal of the load. Therefore, we get the following equation.
,
€E2- a°1-
)=0
E(702-(421ao
dtde (9.13)

Transformation. of Eq. (9.13), leads to ;

——
de =—
dt(70)
A260I}(9
;B2 (
.14
from which

B2E2logBytanh),,
EGa (2E2-60)= —A2aot-I-C (9.15)
Since we have the residual deformation sa at the beginning of the flow

recovery,
thearbitrary
constant
Cisdecided
toB°.
2°E
logtanh((eaE2-cro)}
by putting e=ea and t=0 in Eq. (9.15).
Substituting this constant into Eq. (9.15), we obtain

logtanhBo ao(5E5ao)1=A2132E2t+
log
tank
j2"(eaE2-0'0)1.(9.16)
From Eq. (9.16) we get the relatioD. between e and t.
ao2a0
e=tanh-
E2132E5 1exn(
-2ao - A2B2E2t)Co)}(9.17)
So far as

ea>e>2-Bo—(1 +2B2) (9.18)

.E2
The following approximate result is given as

ao ao Appr,ao
E=——
E 1.0p-12'Amp-LI,Vag/.(9. 19)
2 ByEyB2E2

a
 27

where A2 A2/2
If e is beyond the limits of (9. 18), viz.

e< 2B
2E2 (1 +2B2) (9. 20)
Eq. (9. 17) becomes
Co
(9. 21)
at CC
Eq. (9. 19) and Eq. (9. 21) indicate that the strain—logt relation of flow
recovery is linear and the slope of the line, viz. the value of cro/B2E2, is constant
independent of residual strain ea, and moreover the strain—logt curve should
approach to a horizontal asymptote represented by Eq. (9. 21).

10 Summary

In this chapter, various experiments and their considerations of compres-

sion flow are carried out on undisturbed clay, in order to refer and ascertain
the theory derived in Article 3. On the basis of the theory and experiments
on rheological properties of undisturbed clay reported in this paper, the follow-
ing conclusions are believed to be warranted so far as the applied stress a is
less than the pre-consolidation stress and the upper yield value au of the clay.
(1) The results of the compression flow tests on undisturbed clay well
agree with the theory stated in Article 3 in which the authors proposed the
existence of restraining resistance ao which prevents the displacement of particle
to neighbouring hole, and the new structural viscosity obtained by the assump-
tion that the number of the unit processes of deformation Ni or ni are directly
proportional to stress (a —ao) when the applied stress is less than the upper
yield value au and the pre-consolidation stress of the clay.
The details of (1) are summarized in the following articles.
(2) Taking flow strain as ordinate in ordinal scale and loading time as
abscissa in logarithmic scale, this relation is expressed by straight line, and
the tangent of slope angle of this straight line increases linearly with applied
stress (c.f. Fig. 2.4, 2. 5, 2.6).
(3) If the stress larger than the upper yield value is applied, the flow
strain increases acceleratively and failure takes place after a certain duration
(c. f. Fig. 2.4, 2.5, 2.6).
Only in this case, theory stated in (1) does not hold.
 28

(4) The linear relation between strain rate of flow and loading time
holds if they are plotted on the logarithmic paper, and the slope of the straight
line is constant independent of applied stress (c.f. Fig. 2. 9, 2. 10).
(5) There holds also linear relation between strain rate of flow and
applied stress for various constant values of loading time (c.f. Fig. 2. 8).
(6) It is presumed that the rate of strain increases with increasing values
of temperature at constant water content.
(7) The relation between the flow strain at a certain elapsed time and
the applied stress are linear and the authors analysed the character of the time-
function 0(t) (c.f. Fig. 2. 11, 2. 12).
(8) Taking flow strain as ordinate in ordinal scale and loading time as
abscissa in logarithmic scale for repetitional loading test, the relation is expres-
sed by straight line. The tangent of slope angle of this straight line decreases
direct proportionally to the residual strain Er which rests at the biginning of
the loading and increases with the number of repetition of loading (c.f. Fig.
2. 16).
(9) The strain during flow recovery decreases in direct proportion to
the logarithm of time and the tangent of the slope angle of each line repre-
senting the strain and the logarithm of time is constant independent of the
number of repetitional removal of load (c.f. Fig. 2. 14 (b), 2. 15 (b)).

Chapter 3 Long-term Strength

11. Aim and Brief Contents

The investigations on the long-term strength or the stress expressed at

flow failure are important for the problems on the critical bearing capacity of
foundation, the stability of slope or the analysis of land-slide. As the long-
term strength has close connection with the upper yield value mentioned in
the preceding chapter, a new measuring method of the yield value founded
on. the preceding theory is proposed and a new formula relating between
long-term strength and the time lapse necessary to flow failure is derived.
Researching experimentally on the relations between the upper yield value,
strengths by the stress- or strain-controlled tests and the water content with
undisturbed clay samples of various water contents, new important results
have been obtained. In addition, the effect of flow on the strength of clay
 29

is experimented and considered.

The clay samples and the testing apparatus used on the experiments re-
rated in this chapter are the same with those described in Article 5. Therefore
the sample- number denoted in this chapter can be refered to the soil profile
shown in Fig. 2. 1.

12. A Method to Measure the Upper Yield Value of Clay by

Stress-controlled Compression Test

If a stress exceeding the upper yield value is applied on clay, the clay
fails after increasing its flowing speed. Though the upper yield value car . be
determined graphically with the viscosity curve obtained by means of the com-

pression flow tests as stated in Article 7, such determining procedure is trouble-

some a"d requires much time and much fresh specimens made of the same
sample.
As an improved method to measure the upper yield value of clay, instead
of the above stated complex procedure, the authors propose a new method
using the so-called stress-controlled compression test which can be performed
simply and quickly with the common type of load-controlled compression
apparatus. In this case, load-controlled test is considered to be stress-controll-
ed test if the slight change in sectional area caused by lateral expansion is
neglected. On this test, the compressive stress is applied by equal stress
increment at uniform time interval stepwise as shown in Fig. 3. 1 and com-
pressive strain is measured at each end of the step of stress. The ratio of
stress increment creand time interval to is defined as the rate of stress-incre-
ment a, therefore a=a0c.

di 50;
b cl 40'c
3a`
A
a 2
O cr,

0 t, 2 t, 4 t, 3t, 0
t toe E
Fig. 3.1 Schematic relation of stress Fig. 3.2 Logarithmic relation of stress
vs. time in stress-controlled compre- vs. strain in stress-controlled compres-
ssion process to determine the upper sion process to determine the upper
yield value yield value
 30

A stress—strain relation obtained by a stress-controlled test with an a on

an undisturbed clay sample can be illustrated as a broken line OABCD as
shown in Fig. 3. 2 on the logarithmic paper. In Fig. 3. 2 points 0, A, B,
C, D correspond to the points o, a, b, c, d in Fig. 3.1 and first few
points e.g. 0, A, B lie on the same straight line, therefore point B becomes
a first inflection point.
The linearity of the line OB can be explained theoretically as follows. In
the case of stress-controlled test, we have to consider the effect of the residual
strain Er as stated in Article 9, because the process of stress-controlled test
is considered to be an.alogus with the repetitional loading test with successively
applied loads of cc, 2a,, 36e,...and no,. In Eq. (9. 7) representing the strain--
time relation of modified Voigt element for repetitional loading test, the lower
yield value co is considered negligibly small compared with the applied stress
a, therefore the strains of modified Voigt element e2.1, e2.2, 62.8,....and e2.n
can. be expressed as follows. Where the 1st suffix 2 of e2.c expresses the strain
of modified Voigt element at the end of each loading step and the 2nd suffix
i of e2.t the number of loading step.

e2.1=E
y(1 + /)=er. 1
2o-
e2 2=P•(l4-l)—ler.i= er.1(2 —1)=--sr.2

3a,
E2.3=- -(1+1)— /er.2= zr.1(3 —2/ +

ea.n= er. i{n+(n —1)( (n —2)(— 02+

+2( —/)7'-2+( —T)n-1)
Cie
=—E
2Cn—{(—l)+ (— 1)2+ (— 1)2+ +(— })
(12. 1)

where
1=A1002'112E24and
1a1
—l<l<0 because
0<E2(1+
<-6er•Er.a,
E er•3 and er•mdenote residual strains at the ends of the
y,"
loading steps of 1, 2, 3 n respectively.
Then the total strain at the end of the nth loading step en is given by
6.-= e3•n+F.2.2,6
no.na, +( —l).) (12 . 2)
E
i,+ E2—E,-2{( —1)±( 1}2+ (-1)3 -1-
 31

where ei.n is the strain of a spring element El .

Eq. (12. 2) represents the stress (a nac)—strair. relation for stress-controlled
compression test.
The results of the stress—strain -E 1.0
relations numerically calculated by Eq.
(12. 2) are shown. in Fig. 3. 3 for tj 5
4-
various values of 1 and these calcula-
3 41'
, ted curves are represented by approxi-40. cp
mate straight lines on the logarithmic2
paper. In these calculations the values
of constant terms in. Eq. (12. 2) are 1610-3.5102
zAA4 52
founded on the experiments of Fig.
. 15 (in the case of clay sample No. Fig.
l 3.3 Results of numericalcalcu-2
ationsof Eq. (12.2)
9) as follows. The first elastic modu-
lus E1 (=393kg/cm2) was obtained as the ratio of the stress a= 0.218kg/cm2
to the in.stantaneous strain ei =5.55 X10-4 Comparative calculation of E1
showed so excellent agreement as the value above obtained is equal to the
instantaneous modulus of recovering deformation (El.= 392 kg/cm2) computed
from the flow recovery curve shown in. Fig. 2. 15 (b). The second elastic
modulus E2 was equal to 56.5 kg/cm2 as stated in Article 9. The stress incre-
ment ac is 0.1 kg/cm2 The calculated values of Eq. (12. 2) against 1= —1/2,
—1/10, and —1/100 are shown in. Fig. 3. 3 and the slope angle of line
increases with increasing the value of 1. It is obvious that the upper limit
value of le is given when 1=0 and is equal to 45°.
As Eq. (9. 7) holds only for the stress below the upper yield value, Eq.
(12. 2) holds also within this stress region and the stress at the first inflection.
point from which Eq. (12. 2) does not hold should correspond to the upper
yield value. Therefore the upper yield values of clay can be measured as the
stress corresponding to the first inflection point of stress—strain curve on logari-
thmic paper obtained by the stress-controlled unconfined compression test,
which is performed by adding equal stress increment at uniform time interval.
Fig. 3. 4 gives an example of logarithmic relations between stress and
strain obtained with three clay samples applying the rate of stress increment
of a= 5 X 10-3 kg/cm2/min. From the experimental curves in Fig. 3. 4, stresses
corresponding to the points of first inflection of curves are obtained as shown
in center column. of Table 3. 1. Each stress tabulated in the center column
 32

Table 3.1

SamleNo.Stress fi corresponding to the point of Upper yield value

rst inflection of curve kg/cm2 kg/cm2

13 0.44 0.45
8 0.34 0.33
4 0.19 0.19

well agrees with the each value tabulated in the right column of Table 3. 1
respectively which is the upper yield value obtained by the graphical procedure

1mmilm^Imulienowi.0^Nommmoommen^o...^in
nmsmil^
Fig. 2. 8 and tabulated in
k' NEN^mimmmmoon•IMMEMIIII.reZIW Table2. L '''€1===IMMEMEMEMIN
11^INIMENEMEMErSiiMMIN
MINIIMEIIILINIIIIIPMEAccordingto
V5 the abovemen-
b 4 MEIlliill11111111 tioned theoretical consideration
10511116
2=11111and actual
proof
with
experimental
Pr-
111MAA obtain the upper yield value by
1010
id^Iliresults,
2
itbecomes
5 4 5
possible
to 2 5 4
4h 1(52MT
means of the unconfined stress-

1I012345IIIII 11I34IIIIcontrolled
1 compression test.
R52Besides undisturbed,clay, it
Fig. 3.4 Logarithmicrelationsbetweenthe was observed experimentally that
appliedstress crand strain e concrete° and some kinds of metal
have the same flow character as
1.0 . mentioned above : that is , stress--
..E
NEEMMUMMIIM•EN^MIIIIIIII k'
IIMMININIMENELOSIMstrain relations of such materials co5
4represented on logarithmic paper
be.Rill.ffalll
3
,,,eo.\.04-\--the
upperp
AJIrf6Pr-8werestraightlines continuing
to

I0'''
'''-44All
II,I*)' Ifactiisrayield
#1,

2 3 4 5
well.knownstres
102
that
\IIIIIIIMMIPM1111111111

2 3 4 the so-called stress--strain curve 10

4k1 Id'InM and the failure strength obtained by
IIIIIIIIIIII
101234 51
0234the ordinary compression test are
E t= influenced to a certain extent by
Fig. 3.5 Logarithmic relations between the
applied stressuand strainaat various either the rate of deformation or
rates of stress-increment (sample No. 8) the rate of stress-increment and
the failure strength increases with the rate of stress-increment. On the other
hand; as shown in Fig. 3. 5, the effect of the rate of stress-increment on the
upper yield value is negligible ; that is, the each stress corresponding to the

2
 33

point of first inflection of the curve obtained by the each stress-controlled test
of different rate of stress-increment (a =5 X 10-3, 1 X 10-2 or 2.5 X 10-2 kg/cm2
/min) on the same clay (sample number of clay : No. 8) is nearly equal to
0.33 kg/cm2
Next, a series of stress-controlled undrained triaxial compression tests at
a =5 X 10-3kg/cm2/min was performed with undisturbed saturated clay of 7090
water content (sample number of clay : No. 17) to fined the failure strength
and the upper yield value. In these tests, applied compressive stresses were res-
trained not exceed the pre-consolidation pressure of the clay. Fig. 3. 6 shows

0.5

Failure envelope Upper yield value envelope

p-7,—....„\ ------„, ,....__...s., „,..._____,,:'•..
\ 7,.._-_-z-_,....c.__
.1.
‘`
\ ' \ ‘
1I
A ,',
01 05 1.0 A t52.0as
cr K9/cm'
Fig. 3.6 Envelopesof both failure and upper yield value by stress-controlled
9-test (samle No 17)

the Mohr's circles and their envelopes for the five tests in which no consolida-
tions either before or during compression tests are permitted. Each circle for the
upper yield value in Fig. 3. 6 is drawn with the ambient pressure and the
stress at the point of first inflection of the stress--strain curve obtained by the
same procedure as that of the unconfined stress-controlled compression test.
Since the ambient pressure is carried by the pore water on the undrained
triaxial test and, accordingly, generates no additional strength in clay, both
the envelope of circles of ruptures and that of the upper yield values cru are
horizontal inspite of intensity of the ambient pressure. In Fig. 3. 6, deviator
stresses for rupture and upper yield value are measured as 0.55 kg/cm2 and
0.40 kg/cm2 respectively.
As the fact that the upper yield value does not be ;nfluenced by the
ambient pressure of triaxial stress-controlled compression test means that
rheological constants (A2, B2) of clay in Eq. (12. 2) also indifferent from the
ambient pressure, it may be concluded that the strain rate of clay does not
be influenced by the ambient pressure. This conclusion had been observed
experimentally by R.Haefeli5) with undrained triaxial compression test.
 34

13. Effect of Water Content on the Upper Yield Value and

on the Compressive Strength

In order to research the effect of water content on the upper yield value
of undisturbed saturated clay, a large number of stress-controlled unconfined
compression test introduced in preceding article has been performed on un-
disturbed saturated clay of same kind but with various water contents. In ad-
dition to these tests, using the same kind of undisturbed saturated clay with
various water contents their failure strengths were measured by two kinds of
unconfined compression tests ; —one is the strain-controlled test whose strain
speed is 1% of the original height of the specimen per minute and the other
is the stress-controlled test whose rate of stress increment a is 5 x 10-3 kg/cm2
per minute.
Results of these tests are
90 0 shown in Fig. 3. 7. From this
-0• fi
gure the followingmatters may
4-) 80
Li151;6
,,beconcluded.
GC'(1) If the unconfined corn-
I2>1 % pressive strengths measured by

70 the strain-controlled test and the

L.4) •t0
ov
,1 cllstress-cnotrolled test and the upper

a—-56
1 tkyield
values
ofundisturbed
saturat-
GO
1.,1111Led
clay
are
cowater
plotted
against
their contents on a semi-logarith-50
mic paper as shown in Fig. 3. 7,
it is clear that the linear relations
40
0-2 3 4 5 b.) 2.0hold for each kinds of strength
ica/cm2 and the upper yield value within
Fig. 3.7 Relations between water content a range of permissible error.
and strengths (upper yield value, corn- Among these relations, rela-
pressive strengths by stress-controlled tions between water content w and
and strain-controlled test)
compressive strength by the strain-
controlled test an have been reported to be represented by following equation
as a result of a theoretical consideration").

am=
Anexp
(-2.3-1'1
.) (13.
1)

0 B
 35

where A., Bn; constants.

If the same form of equation is applied for the compressive strength by
the stress-controlled test as and the upper yield value au, relations between
water content w and as or au as shown in. Fig. 3. 7 can be represented as
follows ;

as= Asexp
Bs) (13. 2)
au=
-Auexp(-
2.3-w-
B„
where As, Bs, Au, Du; constants.
(2) As the left line and center one in Fig. 3. 7 are parallel each other,
the ratio of the upper yield value and the failure strength obtained by the stress-
controlled compression test is constant at the same water content and this ratio
is nearly equal to 0.71, independent of water content of clay.
From this result, following relations hold among the constants of Eq.
(13. 2).
Bs =Bu (13. 3)
then
Cu.Au
== a 71
as As (13 . 4)

(3) The ratio of the upper yield value and the failure strength obtained
by the strain-controlled test (au/a.) is not constant but increases with decreas-
ing of water content. This ratio varies from 0.36 to 0.57 according as water
content decreases from 90 to 50% for the clay tested here.
Results obtained in this article are important and very useful for the practi-
cal investigation such as the calculation of stability of slope or bearing capacity
of foundation which support almost constant load. As the upper yield value of
clay is smaller than its strength measured by the strain-controlled or stress-con-
trolled tests, the calculated critical strength of soil structure or foundation which
is based on the failure strength of clay obtained by the ordinary strain-controlled
testing method lies in dangerous side as compared with true critical strength
(or upper yield value) below which the load is supported permanently.

14. Long-term Strength

As already stated, if any constant stress exceeding the upper yield value is
applied on clay, the clay fails after it flows. Since the authors defined such
 36

stress as exceeds the upper yield value the long-term, strength, the long-term
strength does not mean a definite value,

• - E
:cr-`18002.7"4q0
„10 - .,E ,o'0 .676K9/cm 0567Wa41.0
0.64611/o?
(&) -C

5- .4ePe'11111.1.1"1111111
, „ MPAPEZT111111111.1=
x Failure

020050 100 150250 1700 1750 1800 1850

t min
Fig. 3.8 Characteristic curves of the flow failure under various stresses

Fig. 3.8 illustrates the flow curves and points where failures took place
obtained by the unconfined compression tests under various constant stresses a
exceeding the upper yield value on undisturbed clay specimens made from
sample No. 13. The general feature of such a flow curve consists of three
main stages of deformation ; that is, a stage of flow at a decelerating rate which
appears a while after the application of compression, next to this stage, a
stage at an approximately constant rate and finally a stage at an accelerating
rate leading to failure. As this feature is quite different from that of the flow
at the stress below the upper yield value, the mechanism of flow at the stress
exceeding the upper yield value must differ from that stated in Article 3.
Moreover, in Fig. 3. 8 it is observed that every flow strain value at failure
is almost same irrespective of the intensity of applied stress", but the
higher the intensity of the long-term strength, the shorter the elapsed time
until the flow failure.
About the relation between the long-term strength and the elapsed time
to failure the authors solved as follows, applying the theory deduced from
the statistical mechanics on the structure of clay.
In the Articles 6 and 9, it was clarified that, when the stress exceeding
the upper yield value was applied on clay, the rate of flow strain showed the
marked increase and acceleration as compared with the rate of flow strain at
the stress lower than the upper yield value. This fact is considered to be
caused by the successive breaking of bond between clay particles, which
 37

is caused by the application of external stress exceeding the upper yield value
and leads to flow failure after a certain duration. Therefore, if the number of
such bonds per unit cross sectional area perpendicular to the direction of the
applied stress of clay is Nb, the failure takes place when the number of re-
maining bonds become zero (when Nb=0).
Since the ratio of number of activating bonds per unit time (dNb/dt) and
total bonds (Nb) is equal to the probability or the frequency of activation of
one bond per unit time, if repair of broken bonds are prospective, the rate
of breaking of such bonds at constant
1
bdtdNb
Nstress as follows ; (14. 1)
=2KhTe(--E/CTI0)sm.h(2NbK
7)2a2
a is written
where K, h, 2 and E0 are the same notations as denoted in Article 2, 02 is
the stress applied on the dashpot shownin Fig. 1. 3, and
0'2=(cr—
60—E2E2
At the state near failure, elastic property of clay become unconspicuous
with decreasingthe value of Nb and the lower yield value Co is negligibly
small comparedwith the applied stress a, hence 02 is,
c2-4(a — a = const. (14. 2)
Therefore Eq. (14. 1) becomes

N1 dNb=2Te h( Aa
b dt hxP \ICT.2Nbk T)(14. 3)
When the intensity of applied stress a is high or the repair of the bonds
does not expecting,the equation is
1 dNb.KT
=N (—E0\ Aa (14. 4)
b dth exP\ T /exP^2 Noc T)
By the substitutionu=o-/BNb, we obtain the following:

uo uexp( —u)duAti (14. 5)

where

A=KTh ex(—E',)B=2 KT
A a
, 110=BNbb

and tr is the time lapse necessary to the flow failure, Nbb the initial number
of bonds per unit area of clay.
The upper limit of the integral in Eq. (14. 5) is infinite because when
 38

the clay fails the number of remaining bonds is zero. ••

Th 1
e integral--ti-exp(— u)du is the well-known logarithmic integral ex-

pressed as —Ei(— u), and the table of EE(u) is published and applicable for
the calculation. But for sufficiently large value of u,

—Et(— u)= •u.-1exp(— u) (14. 6)

Therefore Eq. (14. 5) becomes approximately

BN
abo
exp(ff-W--)=
Atr (14.
7)
Taking the logarithm of each term, we get
log t 1 a a
r logRAr
bo BNbo

= h FEoAa2cr
ET log2
—box
iv T 2 Nbo T (14. 8).
When the intensity of stress is so high as the 3rd term log(2o-/2Nbotc7')
of the right hand side in Eq. (14. 8) becomes negligibly small compared with
the 4th term (20/2Nrgx T), Eq. (14. 8) can be written approximately by
l h :o
og tr = logx TE-k--T 2N
bo T (14. 9)
or
h oE
logiotr =logiox 7.+2.3frc(14.
T 4.6NbotrAa 10)

aq

0.8

v 0.7
(7.' A".65

0.6

0.5

0.4
• w3ci2i
0.3

0.2
1
.0 10 10' 10'
tf min
Fig. 3.9 Relationsbetween the long-term strength and the time to failure (sample
No. 13, water content ; 65%, sample No. water content ; 92%)
 39

The results of some tests for the flow failure are given
in Fig . 3. 9
which shows the relations of a —logtr obtained for the clay samples of No .
13 (water content =65W3) and No . 2-5 (water content=92 IV). Since the
experimental data lie in well agreement on straight lines on a semi-logarith-
mic scale, it can be said that the each relation between the long-term strength
of clay and the time to failure is quite satisfactorily expressed by Eq . (14. 10),
and the tangent of the slope angle of each line is equal to the values of
2/4.6Nbox T.
If logiotr at a=0 is written as logiotr.i, the actual value of logiotf .i is
obtained by the abscissa of the point where the axis of a =0 and extrapolating
the line in Fig. 3. 9 experimentally obtained intersects . As the equation of
logiotr•t is given by
log E0
ioif.= logtoxhT+2
.3KT (14. 11)
the value of E0 can be computed from Eq. (14. 11). Curves of a—logtr
shown in Fig. 3. 9 are given for clay samples of No. 13 and No. 2-5.
Clay No. 13 ; a = 0.845 —0.0£34logioEr
Clay No. 2-5 ; a =0.437 —0.053logiotf
therefore, by putting a =0 we obtain loglotf.i as follows
Clay No. 13 ; logiotf.i =10.07
Clay No. 2-5 ; logiotf.t =8.24
On the other hand, as logiotf.t is represented by Eq. (14. 11) and temperature
during the experiment was 10°C (or absolute temperature T =273.2+10 =
283.2°K), the values of activating free energy E0 can be computed by substi-
tuting the values in Table 3. 2 into Eq. (14. 11). The values of E0 thus
calculated are shown in the 5th column of Table 3. 2 and have order of 10-12
erg respectively.
Table 3.2
Sapmle No. Water content % tpi min min E0 erg
13 65 1.18 x1010 2.82 x10-10 2.21 x10-12
2 ^•5 92 1.74 x 108 2.82 x 10-16 2.05 x 10-12
Note : Planck's const h=6.626 x 10-27erg. sec
Boltzman's const lc=1.3808 x 10-18erg. deg-1

15. Effect of Flow on the Strength of Clay

Some results of this effect above tilted were reported by other researchers.
 40

A. Casagrande and D. Wilson12' said that this effect is often considerable and
it may be either favourable or unfavourable effect on the strength of clay
according to the type of soil. R. Haefeli5) gave the results of two weeks' flow
by the triaxial compression test which showed a certain increase of about 20
—30W, in the compressive strength according with the applied shearing stress
of intensity of 0.55 kg/cm2.
No attention was paid in their reports, however, to the relation of intensi-
ty of the applied flow stress and the upper yield value. It may be suggested
that the strength of clay will increase after the flow at the stress lower than
the upper yield value but will decrease at the stress higher than it, because
it is clarified that, from the results of flow tests by repetitional loading as
stated in Article 9, the rate of strain by the stress lower than the upper yield
value decreases with the number of repetition, on the other hand the rate of
strain by the stress higher than it has another behaviour as shown in Fig.
2. 17.
On the basis of the above suggestion, the authors investigated the effect
of the flow on the compressive strength of clay when the flow is caused by
applying the constant stress lower than the upper yield value. As a series of
experiments, constant compressive stress of 0.24 kg/cm2 lower than Cu= 0.36
kg/cm2 was applied on the clay samples of No. 11 and 12 with the several dura-
tions of the loading from 1 to 5 days, and the common unconfined compres-
sion tests (whose rate of strain is 1% of the height of specimen per minute)
were performed after removal of the load. Results of these tests (q.) are shown
in the 2nd column in Table 3. 3. Strengths shown in the 4th' column in
Table 3. 3 are the estimated compressive strengths of the undisturbed clay
which receive no flowing stress yet. These "estimated compressive strengths
(que)" are the strengths of the undisturbed clay of the same sample correspond=
Table 3.3

Duration ofWater Unconfined!EstimatedcompressiveStrength

contentato ratio
compressivestrength of
application
load(day)flowq
ofthetest
strength after for
(q
.(to%undisturbed clay)
. kg/cm2) (que kg/cm2) CquIque)

1 0.924 71.0 0.860 1.075

2 1.008 67.6 0.930 1.085
3 1.176 60.0 1.070 1.098
4 0.976 69.4 0.880 1.108
5 1.092 64,2 0.990 1.107
 41

ing to the water contents at the above mentioned tests for q. by applying
the water content--strength line of strain-controlled test at the same strain rate
shown in Fig. 3. 7.
The 5th column in Table 1.12
3. 3 and Fig. 3.10 show the re-0 .....(M1111111111.111
lationof ratio of strength beforea1.1:
and after flow (4,./que) and dura-
tion time of the flow . As shownc 3, 1.04

in Fig. 3. 10 the strength of clay

after application of the constant1.000 1 2 3 4 5
stress lower than the upper yieldTime day
value increases with the duration Fig. 3.10 A relation of strength ratio of
before and after flow and duration time
time of flow. of the flow

16. Summary

In this article the results of experimental and theoretical studies in this

chapter are summarized.
(1) The upper yield values of clay can be measured as the stress corres-
ponding to the first inflection point of stress—,strain curve on logarithmic
paper obtained by the stress-controlled compression test, which is performed
by adding equal stress-increment at uniform time interval.
(2) Though the failure strength is effected by the rate of stress-increment
of above mentioned stress-controlled compression test, the upper yield value
is almost indifferent from the rate of stress-increment.
(3) The envelope of the Mohr's circles of the upper yield value obtained
by undrained triaxial test are horizontal as well as the envelope of failure
strength obtained by the same test.
(4) The ratio of the upper yield value and the failure strength obtained
by the stress-controlled test is constant and almost equal to 0.71 irrespective
of water content of clay.
(5) Eq. (14.9) representing the relation between the long-term strength
and the time to failure well agrees with the experimental results.
(6) The strength of clay increases gradually with duration of the flow
when a constant stress smaller than the upper yield value is applied on the
clay.
 42

Conclusion

In this paper, the rheological characters of clay are researched, and the
summaries of the research are as follows (besides, the detailed summary have
been already reported in Articles 10 and 16).
In chapter 1, the fundamental character of clay under a constant external
stress, that is, the flow character of clay is derived from the micrometric
standpoint applying following assumptions. (1) The viscosity of clay exerting
on the flow of clay is assumed as the structural viscosity which is derived by
applying the statistical mechanics on the frequency of the mutual exchange of
position between a clay particle and the neighbouring hole or .the point of
irregularity in the arrangement of particles. (2) The mechanical model of
clay is assumed from the consideration on the structure of clay as the model
as shown in Fig. 1.3. (3) The fundamental relation of flow of clay obtained
applying above assumption is supposed to be valid so far as the external stress
applied on clay is less than either the stress of pre-consolidation or the upper
yield value.
Chapter 2 is a report on the experimental results and their considerations
of a series of compression flow tests of clay to refer and ascertain the for-
mula derived in Chapter 1. Any flow characters of clay in the experiments—
such characters as strain'-time, strain-rate, stress—strain expressed by time-
function, flow by repetitional loading, flow recovery —exactly agree with the
formula so far as the intensity of applied flow stress is less- than either the
pre-consolidation pressure or the upper yield value.
In Chapter 3, some important problems on soil strength—such as the
effect of water content on the failure strength as well as the upper yield value,
the relation between the failure strength and the time lapse necessary for
failure and the variation in strength by flow—are researched theoretically and
experimentally. Then, a new measuring method of the upper yield value of
clay is proposed ; that is, the upper yield value of clay can be measured as
the stress corresponding to the first inflection point of stress—strain curve on
logarithmic paper obtained by the stress-controlled compression test performed
by adding equal stress increment at uniform time interval.
Furthermore, the values of rheological constants (B2, E1, E2) and activat-
ing free energy of clay (E0) are computed and it is ascertained that these
calculated values are of the right order of magnitude.
 43

As the rheological characters of clay are important not only for the pro-
blems on the soil mechanics but also for the applications on the practical
works, this paper would give some contributions on the soil mechanics and
the soil engineering.
Although this paper, subtitled as "Part 1", is only a partial report on
the rheological characters of clay under the statical stress less than the pre-
consolidation pressure, the authors have arleady researched on the characters
of clay on the consolidation and their secondary time effect under the stress
larger than the pre-consolidation pressure and on the dynamic behaviours under
the vibrating stress, and these above stated researches have reported on some
Japanese papers. These results shall be translated in English and published
in near future in the Bulletin of this Institute as the supplementary paper.

References :

1) J.M. Burgers and G.W. Scott Blair : Report on the principles of rheological nomen-
clature, Proc. 1st Int. Congr. Rheol., 1949.
2) A.V. Tobolsky and H. Eyring : Mechanical Properties of Polymeric Materials, J.
Chem. Phys.. Vol. 11, 1943. pp. 125-134.
3) S. Glasstone, K.J. Laidler and H. Eyring : The Theory of Rate Process, New York,
1941, p. 477.
4) W. Kuhn : Beziehungen zwischen Viscositat and elastischen Eigenschaften amorpher
Stoffe, Z. Phys. Chem., B 42, 1939, pp. 1-38.
.5) R. Haefeli : Creep Problems in Soils, Snow and Ice, Proc. 3rd Int. Conf. Soil
Mech. and Found. Eng., Vol. 1, 1953, pp. 238-251.
6) E.C.W.A. Geuze : Compression. an Important Facter in the Shearing Test, Proc.
2nd Int. Conf. Soil Mech. and Found. Eng., Vol. 3, 1948, pp. 139-142.
7) S.S. Vialov and A.M. Skibitsky : Rheological Process in Frozen Soil and Dense
Clays, Proc. 4th Int. Conf. Soil Mech. and Found. Eng., Vol. 1, 1957, pp. 120-124.
8) E.C.W.A. Geuze and T.K. Tan . The Mechanical Behaviour of clays, Proc. 2nd
Int. Cong. Rheol., pp. 247-259.
9) A. Yoshimoto : An experimental Study on the Deformation of Concrete, J. S. C. E.,
Vol. 40, No. 9. 1955, pp. 22-27.
10) S. Murayama, K. Akai and T. Shibata : The Effect of the Moisture Content on the
Strength of an Alluvial Clay, Disaster Prevention Research Inst. Kyoto Univ. Bulletin
No. 12, Dec., 1955.
11) M. Goldstein and G.T. Stepanian : The Long-term Strength of Clays and Depth
Creep of Slopes, Proc. 4th Int. Conf. Soil Mech. and Found. Eng., Vol. 2. 1957.
pp. 311-314.
12) A. Casagrande and D. Wilson : Effect of Rate of Loading on the Strength of Clays
and Shales at Constant Water Content, Harvard Soil Mech. Series, No. 39, Cam-
bridge, Mass., 1950.
 Publications of the Disaster Prevention Research
Institute

The Disaster Prevention Research Institute publishes reports of the

research results in the form of bulletins. Publications not out of print may
be obtained free of charge upon request to the Director, Disaster Prevention
Research Institute, Kyoto University, Kyoto, Japan.

Bulletins :
No.
N 1 On the Propagation of Flood Waves by Shoitiro Hayami, 1951.
o. 2 On the Effect of Sand Storm in Controlling the Mouth of the Kiku River
by Tojiro Ishihara and Yuichi Iwagaki, 1952.
No. 3 Observation of Tidal Strain of the Earth (Part I) by Kenzo Sassa, Izuo Ozawa
and Soji Yoshikawa. And Observation of Tidal Strain of the Earth by the
Extensometer (Part II) by Izuo Ozawa, 1952.
No. 4 Earthquake Damages and Elastic Properties of the Ground by Ryo Tanabashi
and Hatsuo Ishizaki, 1953.
No. 5 Some Studies on Beach Erosions by Shoitiro Hayami, Tojiro Ishihara and
Yuichi Iwagaki, 1953.
No. 6 Study on Some Phenomena Foretelling the Occurrence of Destructive Earth-
N quakes by Eiichi Nishimura, 1953.
o. 7 Vibration Problems of Skyscraper. Destructive Element of Seismic Waves for
Structures by Ryo Tanabashi, Takuzi Kobori and Kiyoshi Kaneta, 1954.
No. 8 Studies on the Failure and the Settlement of Foundations by Sakur6 Murayama,
1954.
No. 9 Experimental Studies on Meteorological Tsunamis Traveling up the Rivers and
Canals in Osaka City by Shoitiro Hayami, Katsumasa Yano, Shohei Adachi and
Hideaki Kunishi, 1955.
No.10 Fundamental Studies on the Runoff Analysis by Characteristics by Yuichi Iwa-
N gaki, 1955.
o.11 Fundamental Considerations on the Earthquake Resistant Properties of the Earth
Dam by Motohiro Hatanaka, 1955.
No.12 The Effect of the Moisture Content on the Strength of an Alluvial Clay by
Sakur6 Murayama, KOichi Akai and TOru Shibata, 1955.
No.13 On Phenomena Forerunning Earthquakes by Kenzo Sassa and Eiichi Nishimura,
1956.
No.14 A Theoretical Study on Differential Settlements of Structures by Yoshitsura
Yokoo and Kunio Yamagata, 1956.
No.15 Study on Elastic Strain of the Ground in Earth Tides by Izuo Ozawa, 1957.
No.16 Consideration on the Mechanism of Structural Cracking of Reinforced Concrete
Buildings Due to Concrete Shrinkage by Yoshitsura Yokoo and S. Tsunoda. 1957.
No.17 On the Stress Analysis and the Stability Computation of Earth Embankments
by Koichi Akai, 1957.
No.18 On the Numerical Solutions of Harmonic, Biharmonic and Similar Equations by
the Difference Method Not through Successive Approximations by Hatsuo Ishizaki, 1957.
No.19 On the Application of the Unit Hydrograph Method to Runoff Analysis for
Rivers in Japan by Tojiro Ishihara and Akiharu Kanamaru, 1958.
No.20 Analysis of Statically Indeterminate Structures in the Ultimate State by Rvo
Tanabashi, 1958.
No.21 The Propagation of Waves near Explosion and Fracture of Rock (I) by Soji
Yoshikawa, 1958.
No.22 On the Second Volcanic Micro-Tremor at the Volcano Aso by Michiyasu Shima, 1958.
No.23 On the Observation of the Crustal Deformation and Meteorological Effect en It
at Ide Observatory and on the Crustal Deformation Due to Full Water and Accumu-
lating Sand in the Sabo-Dam by Michio Takada, 1958.
No.24 On the Character of Seepage Water and Their Effect on the Stability of Earth
Embankments by KOichi Akai, 1958.
No.25 On the Thermoelasticity in the Semi-infinite Elastic Soid by Michiyasu Shima
No.26 On the Rheological Characters of Clay (Part 1) by Sakur6 Murayama and TOru
Shibata, 1958.

Bulletin No. 26 Published October, 1958

IfefIl 33 10 Fl 15 El
rrgfri 33 V 10 Fl 20 Ei 11

* tr 5r. Fili
F13hIO Uitt E r‘T
74,11,111PS
A
FP IIIUiN W 4t Ell -R