UNDRAINED BEHAVIOR OF C L A Y UNDER C Y C L I C

SHEAR STRESSES
By Atilla M. Ansal 1 and Ayfer Erken 2

ABSTRACT: The cyclic behavior of normally consolidated clays is investigated by

Downloaded from ascelibrary.org by Ryerson University on 10/06/16. Copyright ASCE. For personal use only; all rights reserved.

conducting a series of cyclic simple shear tests on one-dimensionally and isotrop-

ically consolidated kaolinite samples. The response of two types of specimens are
evaluated and compared based on tests performed with different cyclic shear stress
amplitudes and frequencies. The cyclic stress-strain-pore pressure and cyclic shear
strength properties of the clay samples are studied and a simple procedure is sug-
gested to define cyclic yield strength. The results also indicate that there is a thresh-
old cyclic shear stress level below which no excess pore pressure will develop.
An attempt is made to develop an empirical procedure to estimate the response of
a soil element subjected to cyclic shear stresses for a given number of cycles.

INTRODUCTION

During earthquakes soil layers experience cyclic shear stresses with dif-
ferent amplitudes and frequencies which will lead to cyclic deformations.
These deformations are going to affect structures located on these layers and
may cause damage. In addition, the change in the stress-strain and strength
properties of soil layers during cyclic loading may have a significant influ-
ence on the stability of earthdams, embankments, retaining structures, and
natural slopes. As a result, geotechnical engineers have shown expanding
interest in the evaluation and analysis of cyclic behavior of soils.
The behavior of clays subjected to cyclic loading has been studied by a
large number of researchers up to the present. Some of the early investi-
gations reported in the literature were performed by Seed and Chan (1966)
and Thiers and Seed (1968, 1969). In those studies stress-strain and strength
properties of clay samples were evaluated based on cyclic triaxial and cyclic
simple shear tests. Another group of studies reported in the literature were
performed by Sangrey (1968), Sangrey et al. (1969), France and Sangrey
(1977), Sangrey et al. (1978a, 1978b), and Sangrey and France (1980). In
those investigations the cyclic stress-strain-pore pressure behavior of clays
was studied based on cyclic triaxial tests performed at relatively slow rates.
The cyclic behavior of clays was also studied by other researchers, such as
Brown et al. (1975), Wood (1980), Koutsoftas and Fisher (1980), Matsui et
al. (1980), Yasuhara et al. (1982), Dyvik et al. (1983), Goulis et al. (1985),
Ishihara (1980, 1985), and Ishihara and Kasuda (1984).
The study reported in this paper was also concerned with the behavior of
normally consolidated saturated clays under cyclic shear stresses. The sam-
ples were prepared in the laboratory using kaolinite clay. The first group of
samples was obtained by consolidating kaolinite slurry in large-diameter
'Assoc. Prof, of Geotech. Engrg., Civ. Engrg. Fac, Istanbul Tech. Univ., Aya-
zaga, Istanbul, Turkey.
2
Res. Assoc, Geotech. Engrg. Dept., Civ. Engrg. Fac, Istanbul Tech. Univ.,
Ayazaga, Istanbul, Turkey.
Note. Discussion open until December 1, 1989. To extend the closing date one
month, a written request must be filed with the ASCE Manager of lournals. The
manuscript for this paper was submitted for review and possible publication on Oc-
tober 16, 1987. This paper is part of the Journal of Geotechnical Engineering, Vol.
115, No. 7, July, 1989. ©ASCE, ISSN 0733-9410/89/0007-0968/$1.00 + $.15 per
page. Paper No. 23657.
968

J. Geotech. Engrg., 1989, 115(7): 968-983

 odometers under axial stresses. The second group of samples was prepared
by completely remolding one-dimensionally consolidated clay blocks and by
reconsolidating them under isotropic stresses.
All of the tests were conducted under undrained conditions using a cyclic
simple shear testing system. The purpose was to investigate the effects of
cyclic shear stress amplitude, frequency, number of cycles, and the change
in the soil structure due to remolding on cyclic behavior of clays. An effort
Downloaded from ascelibrary.org by Ryerson University on 10/06/16. Copyright ASCE. For personal use only; all rights reserved.

was made to develop an empirical procedure to evaluate the cyclic response

of normally consolidated clay samples.

TESTING PROCEDURE

The cyclic simple shear testing system used in this investigation consists
of a pneumatic stress controlled system capable of generating cyclic shear
stresses at frequencies between 0.0001 Hz and 5 Hz. The test sample has a
diameter of 70 mm and a height of 30 mm and may be placed in a reinforced
or a flexible membrane with lateral confinement provided by a stack of cir-
cular thin polished metal plates similar to ones used in the Swedish direct
simple shear testing system. Axial pressure is applied to the sample by the
bottom platen, which is fixed in the horizontal direction but is able to move
axially.
The sample is placed in a cell so that confining and back pressure can be
applied. This aspect of the test system assures full saturation in the sample
prior to testing and leads to more reliable pore pressure measurements during
testing. Castor oil was used as the cell fluid to minimize air diffusion through
membranes. The horizontal shear stresses are applied at the top cap con-
nected to the horizontally movable shaft going through the cell. The applied
shear forces are monitored by a load cell located in the cell to eliminate the
effects of friction of the shaft. The pore pressure is measured by a pressure
transducer connected to the bottom platen. Axial and horizontal deformations
are measured by sensitive displacement transducers located outside of the
cell.
The samples used in this study were prepared in the laboratory under iden-
tical conditions using finely ground kaolinite clay with liquid limit wL = 65,
plastic limit wp = 25, and specific gravity Gs = 2.65. The first group of
samples were prepared as a slurry at a water content slightly higher than the
liquid limit, and then consolidated in large diameter odometers to axial pres-
sures of 100 kPa. The test samples were trimmed from these blocks and
were reconsolidated under isotropic stress of 100 kPa in the test chamber.
The second group of samples were obtained by completely remolding one-
dimensionally consolidated clay blocks and by reconsolidating the test sam-
ples from these blocks under isotropic stress of 100 kPa. There were basi-
cally two reasons for adopting the second method of sample preparation.
One of these reasons was to investigate the effect of excessive remolding
and the other was to prepare samples with isotropic properties.

CYCLIC BEHAVIOR OF ONE-DIMENSIONALLY CONSOLIDATED SAMPLES

In the first part of this investigation, the cyclic stress-strain-pore pressure

behavior of normally consolidated clay samples is studied with respect to
different shear stress amplitudes for sets of tests conducted at the same cyclic
969

J. Geotech. Engrg., 1989, 115(7): 968-983

Downloaded from ascelibrary.org by Ryerson University on 10/06/16. Copyright ASCE. For personal use only; all rights reserved.

FIG. 1. Shear Strain and Pore Water Pressure Behavior under Cyclic Shear
Stresses:/, = 0.1 Hz

frequency. The first group of tests were performed on samples obtained from
one-dimensionally consolidated kaolinite blocks. In this group, sets of cyclic
tests are carried out at frequencies of 1.0, 0.5, and 0.1 Hz. The response
patterns observed in two of these sets conducted at frequencies of 1.0 and
0.1 Hz are shown in Figs. 1 and 2. On these plots, the cyclic shear stress
amplitudes are given as a ratio with respect to the consolidated undrained
shear strength determined by stress-controlled, quasi-static tests carried out
in the same system. Logarithmic scales are used to show the variation of
shear strain amplitude and pore pressure versus number of cycles. In this
way, it is possible to demonstrate the differences in the behavior for small
numbers of cycles more distinctly. As can be observed from these data, it
appears possible to consider a critical shear stress ratio which can be defined
as "the critical level of repeated stress." This definition was first given by
Larew and Leonards (1962) as the maximum level of repeated stress that
will not lead to failure. Later Sangrey (1968), Sangrey et al. (1969), France
and Sangrey (1977), and Sangrey et al. (1978) demonstrated the validity of
this concept.
As shown on Fig. 1, if the clay sample is subjected to cyclic shear stresses
with stress ratios T/T 7 =1.10 and 0.60, pore pressure will accumulate rapidly
during the first 10 cycles and the sample will undergo large cyclic shear
deformations. However, if the applied stress ratio is smaller than the critical
970

J. Geotech. Engrg., 1989, 115(7): 968-983

Downloaded from ascelibrary.org by Ryerson University on 10/06/16. Copyright ASCE. For personal use only; all rights reserved.

FIG. 2. Shear Strain and Pore Water Pressure Behavior under Cyclic Shear
Stresses:/, = 1.0 Hz

cyclic stress ratio, as in other tests, the accumulated pore pressure will be
limited and the sample will experience relatively small shear deformations.
It is evident that pore pressure and cyclic shear strain amplitude are inter-
related (Koutsoftas 1978; Matsui et al. 1980). However, the determining
factor controlling the behavior in these cases needs to be the cyclic shear
strain amplitude. As pointed out by Lo (1969) and Wilson and Greenwood
(1974), the development of pore pressures in normally consolidated clays is
directly proportional to strains. The application of high cyclic shear stress
ratios would lead to strain softening (Ogawa et al. 1977; Ishihara 1985) and
to large shear strain amplitudes, and, as a result, to higher pore pressures.
By conducting a series of cyclic triaxial tests with different cyclic stress
amplitudes, Sangrey et al. (1969, 1978) have defined a "cyclic limit state
line" as the equilibrium line in the effective stress space for different over-
consolidation ratios and reported that residual pore pressures after cyclic loading
were linearly related to cyclic stress amplitudes for stress levels below the
critical value. For higher cyclic stress levels, large strains would occur and
the accumulation of pore pressures would cause the effective stress path to
migrate towards the failure envelope. It was also shown that a linear rela-
tionship can be established between the critical stress ratio and soil com-
pressibility. For one-dimensionally consolidated kaolinite samples tested at
0.1 Hz, the critical stress ratio may be estimated approximately as T/T/ =
0.50. These experimental observations, in general, are similar to the findings

971

J. Geotech. Engrg., 1989, 115(7): 968-983

 of other researchers, as reported in the literature (Sangrey et al. 1980; Kout-
softas 1978; Matsui et al. 1980).
As estimated from the sets of cyclic tests given, the critical cyclic shear
stress ratio is approximately equal to 50% of the static shear strength. How-
ever, if the creep aspect due to sustained shear stresses is considered, then
it may even be necessary to use a lower value for the critical cyclic stress
ratio. As shown on Fig. 1, the tests with stress ratios 0.42 and 0.36 have
Downloaded from ascelibrary.org by Ryerson University on 10/06/16. Copyright ASCE. For personal use only; all rights reserved.

yielded relatively high residual pore pressures, which may accelerate creep
phenomena and may cause large deformations under sustained shear stresses
if pore pressures are not allowed to dissipate. This aspect of the problem
can be treated in two ways, first, in terms of shear strength, and second, in
terms of stress-strain properties after cyclic loading. There are differences
in the results reported in the literature in relation to the decrease of the static
shear strength after cyclic loading. Castro and Christian (1976), Andersen
et al. (1980), and Koutsoftas (1978) have presented data including that the
loss in shear strength is not so drastic even for large cyclic strain levels.
However, Thiers and Seed (1969), Taylor and Bacchus (1969), and Lee and
Focht (1974), presented data showing a significant decreases in the shear
strength if the cyclic strain/static yield strain ratio is large. Sangrey and
France (1980) observed that the static shear strength after cyclic loading may
decrease as much as 50% as the cyclic stress amplitude approaches the crit-
ical level of repeated loading.
There appears to be a general agreement for the second aspect of the prob-
lem in relation to stress-strain behavior after cyclic loading. Almost all of
the data in the literature indicate a significant decrease in initial elastic mod-
ulus after cyclic loading. This aspect of the problem is more important when
considering the behavior of clays under sustained shear stresses, since this
strain softening may enhance creep rate and may lead to excessive defor-
mations (Arulandan et al. 1971; Holzer and Hoeg 1973). In actual in situ
conditions, there may be situations where the pore water pressures may not
dissipate rapidly. In addition, due to strain softening because of cyclic load-
ing, intolerable large deformations and eventually failure may occur. There-
fore, a pseudo-static analysis, based on static shear strength, may in some
cases yield erroneous information concerning the safety of the soil structure
investigated if the pore pressure buildup and creep aspects are neglected.
The other important parameter controlling the cyclic stress-strain charac-
teristics of clays is the number of cycles. This parameter plays a crucial role,
especially in analyzing the behavior of soil layers under earthquake loads.
As shown in Figs. 1 and 2, even if the applied cyclic stress is larger than
the critical level, the accumulated pore water pressure and cyclic shear strain
amplitudes may still be negligible if the number of cycles is small. There-
fore, to show the effect of number of cycles, the results obtained from the
test set conducted at 0.1 Hz are plotted as stress-strain and pore pressure-
strain relationships in Fig. 3.
One approach generally adopted in analyzing the response of soil layers
under earthquake loads uses the concept of equivalent number of cycles de-
termined to represent the effects of random cyclic stresses generated by an
earthquake. If such an approach is adopted, cyclic test results plotted as
given in Fig. 3 can be used to estimate the pore pressure buildup and cyclic
strain amplitude that would develop in a soil element subjected to specified
cyclic shear stresses for a given number of cycles.
972

J. Geotech. Engrg., 1989, 115(7): 968-983

Downloaded from ascelibrary.org by Ryerson University on 10/06/16. Copyright ASCE. For personal use only; all rights reserved.

SHEAR STRAIN (%) PORE PRESSURE RATIO

FIG. 3. Shear Stress-Shear Strain-Pore Pressure Behavior under Cyclic Shear

Stresses

As can be observed from the response patterns given on Fig. 3, cyclic

shear stress ratio versus shear strain amplitude relationships yield hyperbolic
stress-strain curves. Under these conditions, it appears possible to define a
yield point for each curve as the intersection of the tangents drawn to the
initial and final portions of the curve, as shown on Fig. 4. Since each stress-
strain curve is obtained for a specified number of cycles, it appears realistic
to consider the cyclic shear stress amplitude corresponding to this yield point
as the cyclic yield strength for this number of cycles. The data obtained
indicate that if this critical cyclic shear stress amplitude is exceeded, large
deformations may take place. And, as in most engineering structures, large
deformations may be considered as a failure criterion.

CYCLIC SHEAR STRAIN AMPLITUDE A%)

CYCLIC SHEAR STRAIN AMPLITUDE ,{X)

FIG. 4. Estimation of Cyclic Shear Strength from Cyclic Shear Stress-Shear Strain
Behavior

973

J. Geotech. Engrg., 1989, 115(7): 968-983

 I

— - ^^Z^t-!
Downloaded from ascelibrary.org by Ryerson University on 10/06/16. Copyright ASCE. For personal use only; all rights reserved.

• — i
:_ 2% hear strain

yield strength

l 2 l . . I 1 1 1
1 1 i
1 — I — I 1
' 1 2 3 5 10 20 40 70 100 150 250

NUMBER OF CYCLES

FIG. 5. Cyclic Yield Strength versus Number of Cycles

The results obtained from the cyclic shear tests shown on Fig. 4 are eval-
uated as previously outlined and the variation of cyclic yield strength with
respect to number of cycles is determined as shown on Fig. 5. As can be
observed from this figure, there seems to be a linear relationship between
the number of cycles and cyclic yield strength ratio when a semilogarifhmic
interpretation is adopted. This introduces an additional capability to evaluate
the cyclic yield strength ratio for different number of cycles and to derive
a simple mathematical expression such as

( - j = a - b log N (1)

which can easily be implemented in any numerical procedure developed to

analyze the behavior of soil layers. Here a and b are material constants rep-
resenting the properties of the clay tested. They can be determined very
rapidly by a linear regression analysis.
In addition, it appears also possible to use this approach for a specified
cyclic shear strain amplitude which may be defined as the upper allowable
limit for a specific design purpose (selected arbitrarily as y = 2% in this
study). As shown on Fig. 5, it is again possible to obtain a linear relationship
in terms of the number of cycles. This aspect can also be used to interpret
the strength degradation due to the number of cycles.
It is also interesting to observe that the pore water pressure versus cyclic
stress ratio, as given on Fig. 3, shows a linear variation. This feature of the
response pattern indicates that it may be possible to derive a simple linear
regression expression to estimate the slope of the cyclic stress ratio-pore
water pressure line based on the number of cycles. This aspect of the cyclic
behavior is illustrated in a more detailed manner on Fig. 6, where the results
obtained for the sets of tests shown on Fig. 3 are replotted with a larger
scale and with more data points. As can be seen, the pore water pressure
increases linearly with the cyclic shear stress ratio for almost all numbers of
cycles. It is interesting, however, to note that all of the straight lines drawn
to represent cyclic stress ratio-pore water pressure relationships for different
numbers of cycles intersect the cyclic stress ratio axis almost at the same
point. This indicates that, for normally consolidated clays, there is a critical
stress ratio level below which no pore water pressure will develop. This

974

J. Geotech. Engrg., 1989, 115(7): 968-983

Downloaded from ascelibrary.org by Ryerson University on 10/06/16. Copyright ASCE. For personal use only; all rights reserved.

0.2 0.4

PORE WATER PRESSURE RATIO

FIG. 6. Cyclic Stress Ratio-Pore Pressure Relationship for Different Number of

Cycles

value of the cyclic stress ratio may be defined as the threshold cyclic stress
ratio. This type of pore pressure accumulation is slightly different than the
behavior reported by Sangrey et al. (1969), where no threshold stress level
was observed for residual pore pressures, even though a similar linear re-
lationship was observed between the number of cycles and cyclic stress. One
possible reason for such a difference is most likely changes in mean con-
fining stress during cyclic triaxial loading. However, in the tests conducted
by Matsui et al. (1980), where the mean normal stress was kept constant
during cyclic loading, a similar threshold cyclic stress level, defined as "a
lower boundary value of cyclic shear stress level," was observed. A similar
behavior was also observed for saturated sands subjected to cyclic strains by
Dobry et al. (1982), when they evaluated the results in terms of strain versus
pore pressure. In this investigation, even though all of the samples tested
are identical (which means they all will have an identical shear modulus),
it was not possible to define a threshold cyclic shear strain amplitude, since
the tests are stress-controlled. However, it is believed that with strain control
tests it would be possible to demonstrate the existence of threshold strain
amplitude for clays, too.
It also appears possible to define the variation of the slopes of the pore
pressure lines with respect to the number of cycles, as given in Fig. 7. A
semilogarithmic regression
m = k + log N . (2)
can be determined where m = the slope of the pore pressure line AM/A ( T /
Ty); TV = the number of cycles; and k and p = material constants obtained
975

J. Geotech. Engrg., 1989, 115(7): 968-983

 - — - — - • -•• - - ^r-^T-

H
Downloaded from ascelibrary.org by Ryerson University on 10/06/16. Copyright ASCE. For personal use only; all rights reserved.

20 50 200 500

NUMBER OF CYCLES

FIG. 7. Slope of Pore Water Pressure Lines versus Number of Cycles

from the regression analysis. Once m is evaluated, and knowing the thresh-
old cyclic stress ratio, (S.R.),, it would be possible to estimate the magnitude
of the pore pressure buildup as

-(S.R.), \m (3)

CYCLIC BEHAVIOR OF REMOLDED CLAY SAMPLES

Sets of cyclic simple shear tests at frequencies of 0.1, 0.5, and 1.0 Hz
were also performed on completely remolded kaolinite samples. Because of
remolding, these samples most likely had isotropic stress-strain properties.
The results obtained for sets of tests conducted at 0.1 and 1.0 Hz are shown
on Figs. 8 and 9. As can be observed from these figures, the overall response
patterns are very similar to those obtained for one-dimensionally consoli-
dated samples.
The behavior of samples prepared by the two methods and tested at the
same cyclic stress ratios and same frequency are compared on Fig. 10. The
observed difference in the cyclic strain response of the two samples indicates
that isotropic or remolded clay samples are more resistant to cyclic shear
stresses. However, in general, the pore water pressure responses appear to
be similar where in two out of three tests the accumulated pore pressures
are slightly larger for remolded samples. One possible explanation for these
differences may be to consider the differences in the soil fabric. The re-
molded samples most likely have a more flocculated particle structure in
comparison to one-dimensionally consolidated ones. It is generally reported
that clay particles are oriented much more in the direction perpendicular to
the major stress axes for one-dimensionally consolidated clays, and as a re-
sult these samples are less resistant to shear stresses in the horizontal direc-
tion. Since most of the particle movements in the simple shear test are in
the sliding mode, the volume change tendency will be relatively small. How-
ever, in isotropic samples the particles will rotate and slide at the same time;
therefore, the sample will show a higher resistance to shear stresses—but in
the meantime, the tendency to change volume will also be greater. As a

976

J. Geotech. Engrg., 1989, 115(7): 968-983

Downloaded from ascelibrary.org by Ryerson University on 10/06/16. Copyright ASCE. For personal use only; all rights reserved.

1 2 5 10 20 50 100 200 500 1000

NUMBER OF CYCLES

FIG. 8. Shear Strain and Pore Water Pressure Behavior under Cyclic Shear
Stresses for Remolded Samples:/, = 0.1 Hz

result, it is reasonable that the isotropic clay samples would develop higher
pore pressures.
It is interesting to note that if these observations are used to evaluate the
effects of remolding, or, in the other words, sample disturbance on the cyclic
behavior of one-dimensionally consolidated, transversely isotropic clays. The
response patterns obtained as shown in Fig. 10 are contradictory to the con-
ventional understanding that remolding would decrease the shear strength.
However, it is important to realize that time effects were neglected in this
study. Besides, excessive remolding may not represent sample disturbance,
since the soil structure is altered completely. This aspect of observed be-
havior should be investigated in more detail before reaching any conclusions.
At present, efforts are being made to evaluate the response of natural soils
with different degrees of remolding.

EFFECT OF FREQUENCY

In the second part of this investigation, the rate-dependent response of

normally consolidated clays was studied based on a series of tests conducted
at constant cyclic shear stress ratios with different frequencies from 1 Hz to
0.01 Hz. All of the tests were performed under the same conditions on clay
977

J. Geotech. Engrg., 1989, 115(7): 968-983

Downloaded from ascelibrary.org by Ryerson University on 10/06/16. Copyright ASCE. For personal use only; all rights reserved.

0.1 I I I I I i I I I I
1 2 5 10 20 50 100 200 500 1000

NUMBER of cycles

FIG. 9. Shear Strain and Pore Water Pressure Behavior under Cyclic Shear
Stresses for Remoided Samples:/, = 1.0 Hz

1 2 5 10 20 50 100 200 500 3000

NUMBER OF CVCLES

FIG. 10. Comparison of Shear Strain and Pore Water Pressure Behavior of One'
Dimensionaiiy Consolidated and Remolded Samples

978

J. Geotech. Engrg., 1989, 115(7): 968-983

 20.0

Z 10.0

at
t-
« 4.0
ac
«t
w 2.0
Downloaded from ascelibrary.org by Ryerson University on 10/06/16. Copyright ASCE. For personal use only; all rights reserved.

1.0
1 2 5 10 20 50 100 200

1.0

Ul
fie
D 0.2
in
ui
£ 0.1
w
a:
O
u. 0.04
1 2 5 10 20 50 200

NUMBER OF CYC I E S

FIG. 11. Effect of Cyclic Frequency on Shear Strain and Pore Water Pressure
Behavior of One-Dimensionally Consolidated Clay Samples

samples prepared in the same manner as in the first part of this study.
The variation of shear strain amplitude and accumulation of pore water
pressure observed in the set of tests with a cyclic stress ratio of 0.6 are
shown on Fig. 11. The rate of cyclic loading, in the range of frequencies
considered, appears to have a significant influence on the response patterns.
Similar cyclic behaviors were also observed in other tests with different cyclic
shear stress ratios. However, the measured pore pressures at higher fre-
quencies may not reflect the actual values, especially during the initial cycles,
since sufficient time is not allowed for equalization (Sangrey et al. 1978).
But as the number of cycles increases, the rate of pore pressure buildup
decreases significantly, and it is believed that at this stage the pore pressures
measured are sufficiently equalized. It was reported by Matsui et al. (1980)
and Ogawa et al. (1977) that the pore pressures monitored after cyclic load-
ing do not show a significant increase in cyclic tests conducted at similar
frequencies.
The responses obtained under different frequencies indicate that the de-
crease in the rate of loading leads to an increase in the accumulated pore
pressures with respect to the number of cycles, as was also observed by
Brewer (1972), Ozaydm and Erguvanh (1980), and Matsui et al. (1980).
The effect of rate of loading is more dominant during the initial cycles. As
the number of cycles increases the rate effect diminishes progressively. As
is visible from Fig. 11, both the accumulated value of pore pressures and
the magnitude of the shear strain amplitude for all of the three tests con-
ducted approach each other asymptotically. This indicates that for a large
number of cycles, the frequency of the cyclic loading has no significant
effect on dynamic behavior of normally consolidated clays. From a practical

979

J. Geotech. Engrg., 1989, 115(7): 968-983

Downloaded from ascelibrary.org by Ryerson University on 10/06/16. Copyright ASCE. For personal use only; all rights reserved.

SHEAR STRAIN (ft) PORE PRESSURE RATIO

FIG. 12. Effect of Cyclic Frequency on Cyclic Stress Ratio-Shear Strain-Pore Water
Pressure Behavior

point of view, it appears that it is possible to neglect the influence of fre-

quency in problems such as offshore platforms (where the number of cycles
with respect to wave action will be large). However, in considering the
earthquake motions the number of cycles will be relatively limited; therefore,
the frequency content of the input earthquake motions should be taken into
account. These response patterns are different than the response patterns re-
ported by Matsui et al. (1980) and by Procter and Khaffaf (1985), where
they observed no significant variation in the effect of frequency with respect
to the number of cycles, while Yasuhara et al. (1982) reported that there is
no significant effect of frequency on cyclic undrained strength.
In light of the observed response patterns, it may be more suitable to
display the influence of frequency in terms of shear stress-shear strain-pore
pressure relationship, as in Fig. 12. The frequency of the cyclic shear stresses
appears to play a dominant role for a small number of cycles, in this case
10, and the relative influence of frequency increases with the magnitude of
cyclic shear stress amplitude. In addition, there seems to be a critical shear
stress amplitude below which the effect of frequency disappears. The vari-
ation of shear strain amplitude shows an increase with the decrease in the
frequency. This is partly due to the increase of the duration of shear stresses
and partly due to a decrease in effective confining stresses. As can be seen
on Fig. 12, the threshold cyclic shear stress amplitude appears to be valid
for the frequency range considered in this investigation. As shown previ-
ously (in Fig. 6), pore pressure accumulation can also be modeled as linear
relationships in terms of shear stress amplitude. Under these conditions, it
appears logical to use these types of plots to estimate pore pressure accu-
mulation and the magnitude of shear strain amplitude, if the equivalent shear
stress amplitude, number of equivalent cycles, and dominant frequency con-
tent of the earthquake input are defined for a clay layer.

980

J. Geotech. Engrg., 1989, 115(7): 968-983

 CONCLUSIONS

The cyclic behavior of a normally consolidated clay has been studied by

conducting several series of undrained cyclic simple shear tests with different
shear stress amplitudes and cyclic loading frequencies. The first group of
samples used in the testing program were prepared in large-size odometers
assuming that clay samples will possess anisotropic properties, as in natural
Downloaded from ascelibrary.org by Ryerson University on 10/06/16. Copyright ASCE. For personal use only; all rights reserved.

deposits. The second group of samples was prepared by completely re-

molding the clay blocks obtained from large-size odometers. The experi-
mental results indicate a reduction in the shear strength under cyclic shear
stresses. The magnitude of this reduction may be as large as 6 5 % , if possible
creep phenomenon under sustained shear loads are taken into account. It was
also possible to define a cyclic yield stress ratio which appears to vary lin-
early on a semilogarithmic scale with the number of cycles. The pore pres-
sure behavior has shown that there is a threshold cyclic shear stress below
which no pore pressure accumulation will occur.
Simple regression expressions were developed to estimate the cyclic yield
strength and magnitude of pore pressure buildup based on the number of
cycles and shear stress amplitude. The effect of the rate of cyclic loading
was studied with respect to different shear stress amplitudes and the number
of cycles. It was observed that the frequency effect diminishes with an in-
creasing number of cycles and with decreasing shear stress amplitude. The
influence of cyclic frequency appears to be significant if relatively small
numbers of cycles are used in evaluating the cyclic behavior, as in the case
of earthquakes.
Similar behavior patterns were also observed for remolded samples; how-
ever, these were most likely due to particle structure. Remolded samples
appear to be more resistant to cyclic shear stresses; cyclic shear strain am-
plitudes developed in these tests are smaller in comparison to cyclic shear
strains measured in one-dimensionally consolidated samples. The pore pres-
sure accumulations observed for both types of samples are similar, but the
remolded samples did show a tendency to yield slightly higher pore pres-
sures.

APPENDIX. REFERENCES

Andersen, K. H. et al. (1980). "Cyclic and static laboratory tests on Drammen clay."
J. Geotech. Engrg., ASCE, 106(5), 499-529.
Amlandan, K., Shen, C. K., and Young, R. B. (1971). "Undrained creep of a coastal
organic silty clay." Geotechnique, 21(4), London, U.K., 359-375.
Brewer, J. H. (1972). "The response of cyclic stress in a normally consolidated
saturated clay." Thesis presented to the Civil Engineering Dept., North Carolina
State Univ., at Raleigh, North Carolina, in partial fulfillment of the requirements
for the degree of Doctor of Philosophy.
Brown, S. F., Lashine, A. K. F., and Hyde, A. F. L. (1975). "Repeated load triaxial
testing of a silty clay." Geotechnique, 25(1), London, U.K., 95-114.
Castro, G., and Christian, J. T. (1976). "Shear strength of soils and cyclic loading."
J. Geotech. Engrg., ASCE, 102(9), 887-894.
Dobry, R. et al. (1982). "Prediction of pore water pressure buildup and liquefaction
of sand during earthquakes by the cyclic strain method." NBS Bldg. Sci. Series,
138.
Dyvik, R., Zimmie, T. F., and Schimelfenyg, P. (1983). "Cyclic simple shear be-
havior of fine grained soils." Norwegian Geotech. Inst., Publ., 149, 1-6.
France, J. W., and Sangrey, D. A. (1977). "Effects of drainage in repeated loading

981

J. Geotech. Engrg., 1989, 115(7): 968-983

 of clays." J. Geotech. Engrg., ASCE, 103(7), 769-785.
Goulis, A. M., Whitman, R. V., and Hoeg, K. (1985). "Effect of sustained shear
stresses on cyclic degradation of clay." ASTM STP 883, Strength Testing of Marine
Sediments: Lab. and In-Situ Strength Measurements, 336-351.
Holzer, T. L., and Hoeg, K. (1973). "Effect of seismic loading on undrained clay
creep." J. Soil Mech. and Found., ASCE, 99(1), 153-158.
Ishihara, K. (1980). "Strength of cohesive soils under transient and cyclic loading
conditions." Proc, 7th World Conf. on Earthquake Engrg., State-of-the-Art Vol.,
Downloaded from ascelibrary.org by Ryerson University on 10/06/16. Copyright ASCE. For personal use only; all rights reserved.

159-169.
Ishihara, K. (1985). "Stability of natural deposits during earthquakes." Theme Led.,
Proc, 11th Int. Conf. on Soil Mech. and Found. Engrg., San Francisco, Calif.,
2, 321-376.
Ishihara, K., and Kasuda, K. (1984). "Dynamic strength of a cohesive soil." Proc.,
6th Conf. on Soil Mech. and Found. Engrg., Budapest, Hungary, 91-98.
Ishihara, K., Nagao, A., and Mano, R. (1983). "Residual strain and strength of clay
under seismic loading." Proc, 4th Canadian Conf. on Earthquake Engrg., Van-
couver, Canada, 602-613.
Koutsoftas, D. C. (1978). "Effect of cyclic loads on undrained strength of two ma-
rine clays." J. Geotech. Engrg., ASCE, 104(5), 609-620.
Koutsoftas, D. C , and Fischer, J. A. (1980). "Dynamic properties of two marine
clays." J. Geotech. Engrg., ASCE, 106(6), 645-657.
Larew, H. G., and Leonards, G. A. (1962). "A repeated load strength criterion."
Proc, Highway Res. Board, 41, 529.
Lee, K. L., and Focht, J. A., Jr. (1976). "Strength of clay subjected to cyclic load-
ing." Marine Geotech., 1(3), 165-185.
Lo, K. Y. (1969). "Pore pressure-strain relationship for normally consolidated un-
disturbed clays." Can. Geotech. J., 6(4), 383-412.
Matsui, T., Ohara, H., and Ito, T. (1980). "Cyclic stress-strain history and shear
characteristics of clays." J. Geotech. Engrg., ASCE, 106(10), 1101-1120.
Ogawa, S., Shibayama, T., and Yamaguchi, H. (1977). "Dynamic strength of sat-
urated cohesive soil." Proc, 10th Int. Conf. on Soil Mech. and Found. Engrg.,
„ 2, 317-320.
Ozaydin, K., and Erguvanh, A. (1980). "The generation of pore pressures in clayey
soils during earthquakes." Proc, 7th World Conf. on Earthquake Engrg., 3, 326-
330.
Proctor, D. C , and Khaffaf, J. (1984). "Cyclic triaxial tests on remolded clays." J.
Geotech. Engrg., ASCE, 110(10), 1431-1445.
Sangrey, D. A. (1968). "Behavior of soils subjected to repeated loading." Thesis
presented to the Civ. Engrg. Dept., Cornell Univ., at Ithaca, New York, in partial
fulfillment of the requirements for the degree of Doctor of Philosophy.
Sangrey, D. A., et al. (1978). "Cyclic loading of sands, silts and clays." Proc,
Specialty Conf. on Earthquake Engrg. and Soil Dynamics, ASCE, Pasadena, Calif.,
836-851.
Sangrey, D. A., and France, J. W. (1980). "Peak strength of clay soils after repeated
loading history." Proc, Int. Symp. on Soils under Cyclic and Transient Loading,
Rotterdam, The Netherlands, 1, 421-430.
Sangrey, D. A., Henkel, D. J., and Esrig, M. I. (1969). "The effective stress re-
sponse of a saturated clay soil to repeated loading." Can. Geotech. J., 6(3), 241-
252.
Sangrey, D. A., Polard, W. S., and Egan, J. A. (1978). "Errors associated with rate
of undrained cyclic testing of clay soils." ASTM STP 654, Dynamic Geotech. Test-
ing, 280-294.
Seed, H. B., and Chan, C. K. (1966). "Clay strength under earthquake loading
conditions." J. Soil Mech. and Found., ASCE, 92(2), 53-78.
Taylor, P. W., and Bacchus, D. R. (1969). "Dynamic cyclic strain tests on a clay."
Proc, 7th Int. Conf. of Soil Mech. and Found. Engrg., Mexico City, Mexico, 1,
401-409.
Thiers, G. R., and Seed, H. B, (1969). "Strength and stress-strain characteristics of
clays subjected to seismic loading conditions." ASTM STP 450, Vibration Effects

982

J. Geotech. Engrg., 1989, 115(7): 968-983

 of Earthquakes on Soils and Foundations, 3-56.
Thiers, G. R., and Seed, H. B. (1968). "Cyclic stress-strain characteristics of clays."
J. Soil Mech. and Found., ASCE, 94(2), 555-569.
Wilson, N. E., and Greenwood, J. R. (1974). "Pore pressure and strains after re-
peated loading of saturated clay." Can. Geotech. J., 11(2), 269-277.
Wood, D. M. (1982). "Laboratory investigation of the behavior of soils under cyclic
loading: A review." Soil Mech.—Transient and Cyclic Loads, G. N. Pande and
O. C. Zienkiewicz, eds., John Wiley & Sons, New York, N.Y., 513-582.
Downloaded from ascelibrary.org by Ryerson University on 10/06/16. Copyright ASCE. For personal use only; all rights reserved.

Yasuhara, K., Yamanouchi, T., and Hirao, K. (1982). "Cyclic strength and defor-
mation of normally consolidated clay." Soils and Foundations, Tokyo, Japan, 22(3),
77-91.

983

J. Geotech. Engrg., 1989, 115(7): 968-983