Missouri University of Science and Technology

Scholars' Mine
International Conferences on Recent Advances 1991 - Second International Conference on
in Geotechnical Earthquake Engineering and Recent Advances in Geotechnical Earthquake
Soil Dynamics Engineering & Soil Dynamics

12 Mar 1991, 10:30 am - 12:00 pm

The Effect of Frequency of Cyclic Loading on Earth Structures and

Foundation Soils
Daniel E. Normandeau
Woodard & Curran Inc., Consulting Engineers, Portland, ME

Thomas F. Zimmie
Rensselaer Polytechnic Institute, Troy, NY

Follow this and additional works at: https://scholarsmine.mst.edu/icrageesd

Part of the Geotechnical Engineering Commons

Recommended Citation
Normandeau, Daniel E. and Zimmie, Thomas F., "The Effect of Frequency of Cyclic Loading on Earth
Structures and Foundation Soils" (1991). International Conferences on Recent Advances in Geotechnical
Earthquake Engineering and Soil Dynamics. 39.
https://scholarsmine.mst.edu/icrageesd/02icrageesd/session01/39

This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License.

This Article - Conference proceedings is brought to you for free and open access by Scholars' Mine. It has been
accepted for inclusion in International Conferences on Recent Advances in Geotechnical Earthquake Engineering
and Soil Dynamics by an authorized administrator of Scholars' Mine. This work is protected by U. S. Copyright Law.
Unauthorized use including reproduction for redistribution requires the permission of the copyright holder. For more
information, please contact scholarsmine@mst.edu.
A Proceedings: second International Conference on Recent Advances in Geotechn
W March 11-15,199 1, St. Louis, Missouri, Paper No.1.55
ical Earthquake Engineering and Soil Dynamics,

The Effect of Frequency of Cyclic Loading on Earth Structures and

Foundation Soils
Daniel E. Normandeau Thomas F. Zimmie
Engineer, Woodard & Curran Inc., Consulting Engineers, Civil Engineering Dept., Rensselaer Polytechnic Institute, Troy,
Portland, Maine 04102 USA New York 12180 USA

SYNOPSIS
Earth dams, earth embankments, slopes, and certain foundation soils are subject to
shear stress due to earthquak es, blasting operations, and in some cases machine cyclic loading with an initial locked-in static
vibrations. In order to investigate the effect of the
frequency of the applied cyclic loading, thirty-on e Norwegian Geotechnical Institute
Frequencies for the applied cyclic loading of 0.2 Hz, 0.05 Hz, and 0.025 Hz were used Direct Simple Shear (NGI-DSS ) tests were run.
Silt (SFSS), from the Lower San Fernando Dam which failed in 1971, was the material throughout this investigation. San Fernando Sandy
tested in this study.
This investigation determined to what extent the frequency of the applied loading affects
initial locked-in static shear stress. The results show that the accumulated deformati accumulated deformation of SFSS with an
applied loading. on is inversely proportional to the frequency of the

INTROD UCTION
In the past few decades, much interest has been
developed in the area of earthquak e engineering. Much used in this investigation. However, it can be considered as
research has been performed in this area, with the study of an upper bound of the frequency effect, and is simply used as
dams perhaps leading the list. With the failure of the Lower a basis for comparison.
San Fernando Dam in 1971, hydraulic fill dams have been the
subject of many studies (Seed et al., 1973). The Norwegian Geotechnical Institute Direct Simple
Shear (NGI-DSS ) device was used in this investigation.
The research performed in this investigation pertains Combinations of initial locked-in static shear stresses were
to earth dams and embankm ents, slopes and machine used with the application of different cyclic loads. Constant
foundation soils. All are subject to cyclic loading conditions volume (undraine d) and drained tests were performed.
from earthquak es, blasting operations and machine Within these various test series, varying frequencies of cyclic
vibrations. Initial locked-in static shear stresses are present loading were investigated.
in all of these cases, and play a significant role in their
behavior under cyclic loading (Lien, 1983; Goulois et al., Soil from the Lower San Fernando Dam, San Fernando
1985; Zimmie and Lien, 1986). Sandy Silt (SFSS), was used in this investigation. This
hydraulic fill dam experienced a major flow slide as a
The objective of this study was to determine the effect consequence of the San Fernando Earthquak e in 1971. Since
of the frequency of cyclic loading on the accumu~a~ed that time, it has been the focus of extensive earthquake
deformation of soils. Newmark used the analogy of a sliding engineering research. The test results from this investigation
block on an inclined surface to analyze deformations in slopes add to an extensive data base on SFSS.
due to cyclic loading (Newmark, 1965). Goodman and S~ed
(1966) verified that the ~mount. of accumula ted .deformatiOn
in sand, due to the cychc loading, was proportiOnal to t~e
inverse square of the frequency, as Newmark's analysis
predicted. Strictly speaking, Newmark's analysis applies. to
yield conditions, and is not applicable to the type of loading

153
 represent the initial locked-in static shear stress, whereas in
BACKGROUND this report (as in Zimmie and Lien, 1986) the term r 5 is used.
Locked-in Static Shear Stresses
The presence of an initial locked-in static shear stress Effect of Frequency
can be found in both sloping and level ground conditions.
This may be more obvious for a slope than for flat ground, As can be seen from the previous discussion, the effect
simply due to the effect of gravity. Figure 1 shows several of the frequency of the applied cyclic load on the accumulated
soil elements and their associated stresses. Soil element A deformation of soil, has not been the subject of extensive
(Fig. 1a) has no locked-in static shear stress, because it is study, but it has been cited as a necessary area of research
located in a position where the static vertical loads impinging (Malek et al., 1987).
on it are symmetrical. Soil element B, on the other hand,
does not have symmetrical static vertical loads acting on it. Newmark (1965) used basic physical principles to
It has an initial locked-in static shear stress acting on it, and suggest that frequency has an inverse--square effect on
because it lies on a potential failure surface, the value of this accumulated deformation when an initial locked-in static
stress is relatively high (Seed et al., 1968). Figure 1b shows shear stress is involved. Figure 2 explains Newmark's
soil element C, which could be considered under level ground analogy of a sliding block to the cyclic loading of a soil
conditions, but because it has non--symmetrical vertical loads element. The case of level ground conditions (symmetrical
acting on it and because it is located on a potential failure loading) is shown in Figure 2a. The initial locked-in static
surface, it experiences an initial locked-in static shear stress. shear stress on the block is zero. A cyclic shear stress is
applied to the block causing it to cycle around its center-line,
Figure 1c shows a better example of the initial Y. The displacement (x) of the block is determined from
locked-in static shear stress in level ground conditions. This Equation 1.
is due to soil element E being located away from the line of
symmetry of the vertical load of the structure. This figure x=l/2at 2 ( 1)
shows a foundation which is subject to cyclic loading from
machine vibrations. where a= acceleration and t = time.
As can be seen from the previous discussion, locked-in Because of the level ground condition, the block cycles
static shear stresses can be found in several different an equal distance on each side of the Y axis. Hence, the net
circumstances, and are of considerable concern because of displacement of the block is zero. It does not matter how
their relevance to the safety of earth dams and embankments. long (t=time) the cycle is, the net displacement (x) is always
Evidence of the concern with initial locked-in static shear zero. Therefore, the frequency of the cyclic loading has no
stresses and their applicability to current issues can be found effect on the level ground condition. (The block or soil
in the recent work done in this area (Moriwaki et al., 1982;; element in laboratory testing will eventually fail in one
Zimmie et al., 1985; Zimrnie and Lien, 1986; Malek et al., direction, but this is usually caused by the limits imposed by
1987; Andersen et al., 1988; Andersen and Lauritzsen, 1988). the test apparatus. In theory, the direction (either left or
Although the recent work has been quite thorough with right) should be random.
respect to initial locked-in static shear stress, the effect of
frequency has not been studied. For the case with an initial locked-in static shear
stress ( r 5 ), the block also cycles around its center-line, Y.
Most, if not all of the work involving initial locked-in
static shear stress, has been done on clay samples. Moriwaki However, because of r 5 , the block's center-line moves in the
et al. (1982), Zimmie et al. (1982), Zimmie and Lien (1986) direction which T 5 is acting (Fig. 2b). The block slides down
and Goulois et al. (1985), investigated the undrained cyclic
simple shear response of clay subjected to combined cyclic the incline a certain distance during the first half of the
and initial locked-in static shear stresses. These loading cycle (assuming the first half of the loading cycle is in
investigations led to the development of strain and pore the direction of r 5 ), but never returns to its original position
pressure contour diagrams, which are useful for practical when the loading cycle is complete. For this case, the length
design problems. Andersen and his colleagues have performed of time (t) of the loading cycle in the direction of r 5 is
extensive research in this area as well. He has developed a
procedure for predicting the results of variable cyclic stress directly related to the displacement of the block. Hence, the
level tests from tests performed with a constant cyclic stress, frequency of the cyclic loading does have an effect on the
as well as a procedure to calculate the bearing capacity for accumulated deformation for the case with an initial
foundations subjected to a combination of static and cyclic locked-in static shear stress ( r 5 ).
loads under undrained conditions (Malek et al., 1987;
Andersen and Lauritzsen, 1988). Andersen et al. (1988) has Newmark used Equation 1 to propose that for the case
compiled an extensive database on cyclic soil testing which is with an initial locked-in static shear stress ( r 5 ), the amount
important in the design of 9,ravity structures. Malek et al.
(1987) developed an 'Apparent Overconsolidation" of deformation accumulated from cyclic loading is equal to
hypothesis, which provides good predictions of the number of the inverse--square of the frequency (see Equation 2).
cycles to failure, as well as the development of shear strains,
during undrained cyclic simple shearing.
These previous studies are similar in regards to their (2)
testing procedure and the testing notation used in their
analysis. However, Andersen et al. (1988), Malek et al. where x = displacement and
(1987), and Goulois et al. (1985) use the term rave to f = frequency

154
 t

'
'' '
I
'

8~---
;:/..

(a) Earth Dam

I ,;. I
D E
line of Syn~~~etry

1 FILL ""-._
', ~ /¥

SOFT CLAY
' ... _
'
-
--
I c
-l~t.:-----­

T
___ . . . "'
/

-
-
r ~,
' sc
-tJI lt!- ah
DENSE SAND
(b) Earth Embankment
Element D
T
Element E
(c) Machine Foundation

Figure 1 Soil Elements (a) Earth Dam, (b) Earth

Embankment (Lien, 1983), (c) Machine
Foundation.
Iy
+I-
Therefore, as can be seen form Equation 2, if frequency f is I
1
ten times smaller (i.e. slower) than frequency f , then
2
displacement x 1 will be one hundred times larger than
displacement x 2. level surface

Newmark's inverse-squar e relationship makes

intuitive sense, however, little research has been done in this
area. Goodman and Seed (1966) conducted drained shaking (a) Symmetrical Loading
table tests and verified the validity of Newmark's
sliding-block analogy for predicting accumulated deformation
in earth slopes caused by earthquakes. Through this
verification process they found that displacement varied with
the inverse-squar e of the frequency. Little else has been done
to determine the effect of frequency.
As already noted, Newmark's analysis applies to yield
conditions (i.e. the blocks yield or slide, or in the case of soil
yield strength of the soil is applied on the first and
subsequent cycles). To explain the gradual degradation of inclined surface
soil using cycli.c loads less than the yield strength of the soil,
probably reqmres the use of a more complex constitutive
model, of which there are many. This was beyond the scope
of this project. In addition, most of the constitutive models (b) Non-;;ymmetrica.l Loading
available are strain rate independent, and would predict no
frequency effect. Therefore, Newmark's model was used as a Figure 2 Newmark's Sliding Block Analogy as it
basis for comparison (upper bound limit), and as a simple Relates to Cyclic Loading of
analogy to illustrate that there should be a frequency effect a) Level Ground Condition
for the test boundary conditions used in this study. The (Symmetrical Loading),
results of this study neither validate nor invalidate b) Condition with an Initial Locked-in
Newmark's Method. Static Shear Stress (Non-symme trical
Loading).

155
 TABLE 1

Summary of Lab Analysis of San Fernando

Sandy Silt (G.E.I., 1988)

Mineralogical Analysis

Mineral Type Volume(%)

Quartz 67
Feldspar 15
Clays 10
Opaques 5
Other 3

Index Properties *
Liquid Limit = 24
Plastic Limit = 20
Plasticity Index =4
Specific Gravity = 2.69

* The Atterberg Limit Test was performed on the minus No. 40 fraction of the SFSS.

100
..
US SU.NDAIIID allVl OPI:NIN8 •

'
...
IEHlS

·~ 1 '•
~
~
I
U.S STAIIIOAIID at(V[ -...Eita
910 141S lO 5040 ~ 70DOM0200
HYDROIUTER

00
-- - - '\ '0
80 - \ zo
\ f- --
70
1--
40
-
f- -
40

1- - - -I- - - 70
- -I- -
zo -r-- 80
..............
,_ --·-
10 90
-r--
0 00
D O> 01 OD5 001 0~ 0001
"""
COML(I liLT 011' CLAY

IIOTI: 0.61. of the total •-wale

••• Tat~tned on the llo4 4
•l•we. The - · • - -
particle ala• ••• l/1-tn.

Figure 3 Sieve Analysis of the San Fernando

Sandy Silt (G.E.l., 1988).

156
 necessary to determine both the constant volume (undrained)
EQUIPMENT DESCRIPTION and the drained shear strength of the SFSS.
The Norwegian Geotechnical Institute Direct Simple Both the static and the cyclic tests were run with an
Shear Device (NGI-DSS), with certain modifications was
used for this investigation. Originally designed by NGI and initial vertical stress of 1 kg/em 2• All samples were prepared
now manufactured by Geonor, the NGI-DSS is similar to the in an identical fashion, however, slight differences in the void
device described by Bjerrum and Landva (1966). ratio of the samples did occur (see Table 2). The testing
began after the sample had been consolidated and saturated.
The soil sample in the NGI-DSS device is surrounded A strain rate of approximately 14% per hour was used for the
by a wire reinforced latex membrane. This membrane static monotonic loading of the SFSS.
enables the sample to be uniformly strained, under conditions
of simple shear and plane strain. The NGI-DSS closely
simulates the aspects of cyclic loading caused by earthquakes. TEST RESULTS AND ANALYSIS

A detailed discussion of the NGI-DSS device is

omitted in this paper. However, several previous reports Data Definitions
done at Rensselaer Polytechnic Institute (RPI) discuss the
NGI device in great length. (Floess, 1979; Dyvik, 1981; In order to facilitate the understanding of the data
Pierce, 1982; Lien, 1983; Vucetic, 1986). reduction and the deformation behavior of the SFSS, typical
data obtained for the development of strains as a function of
time is illustrated in Figure 4. as shown in Figure 4a, the
LABORATORY TESTING PROGRAM strain is symmetrical. It increases with the number of cycles
for the case of zero initial locked-in static shear stress, i.e.
Description of Soil Tested symmetrical loading conditions (Test Series I, r 5 fr = 0). the
cyclic shear strain ( 'Yc) changes sign for each cycle and it is
The soil used for this investigation is from the Lower
San Fernando Dam in San Fernando, California and was exactly equal to the total shear strain ( 'Yt) developed in the
obtained through the U.S. Arm~ Corps of Engineers. The soil sample.
San Fernando Sandy Silt (SFSS) came in a bag and was a
batch mix of several individual samples taken from the site.
The SFSS is classified as an ML soil (inorganic silts and very
fine grained sands). All laboratory classification tests were
conducted by Geotechnical Engineers, Inc. of Winchester,
Massachusetts, U.S.A. (G.E.I., 1988), and were performed on
the soil fraction passing the No. 4 sieve. The results are
shown in Table 1 along with a sieve analysis of the SFSS,
shown in Figure 3.
Soil Preparation
In an effort to make the results of this investigation
comparable to previous studies performed with SFSS, the
sample was frepared from air dried SFSS at an initial water
content (we equal to 4% (Vasquez, 1988; G.E.I., 1988).
(a)
Because previous studies were primarily interested in
liquefaction, the loosest consistent void ratio for which the
SFSS could be prepared was determined and used in this
study.
Tests Performed
A total of thirty-one tests were performed for this
investigation. Twenty-five constant volume, load controlled,
direct simple shear tests were performed; twenty-two cyclic z
and three monotonic. Sixteen of the cyclic tests had an ~ 1'4~\::J-----l-...-------- T!~E
initial locked-in static shear stress. Six drained tests were
run; five cyclic and one monotonic. All five cyclic drained
~
(b)
tests had an initial locked-in static shear stress.
Three frequencies (0.2 Hz, 0.05 Hz, and 0.025 Hz)
were used for the cyclic loading tests with an initial locked-in
shear stress. Due to the difficulties associated with
maintaining constant volume conditions from faster
frequencies, 0.2 Hz was the fastest frequency used in this
investigation. Table 2 summarizes the stress conditions, the
frequency of applied loading, and the results of each test Figure 4 Definition of Strain Data Obtained
performed. From
a) Symmetrical Cyclic Tests,
Four static monotonic shear tests were performed in
this investigation; three constant volume (undrained) and one
drained. These tests were necessary for several aspects of the
~b) Non-Symmetrical Cyclic Tests
Lien, 1983)

study. Most importantly, the monotonic shear tests were

157
 TABLE 2

Summary of Test Conditions and Results for Cyclic Tests

Test Series I-IV, Undrained

Initial
Void N
Test Test r 5 /r Tc/T Ratio Frequency (to failure, F
Series Number (%) (%) e f no failure, NF)

1 0 34 0.655 0.2 1400 NF

2 0 54 0.669 0.2 143 F
3 0 54 0.658 0.2 148 F
4 0 54 0.632 0.2 760 NF
5 0 69 0.640 0.2 8F
6 0 69 0.625 0.2 SF

II 1 36 34 0.672 0.2 650 NF

2 36 34 0.635 0.025 200 NF
3 36 54 0.658 0.2 37 F
4 36 54 0.589 0.05 !OF
5 36 54 0.650 0.025 12 F
6 36 69 0.667 0.2 3F
7 36 69 0.639 0.05 2F

III 1 55 34 0.650 0.2 500 NF

2 55 34 0.672 0.025 200 NF
3 55 54 0.655 0.2 10 F
4 55 54 0.659 0.025 3F
5 55 69 0.643 0.2 2F

IV 1 72 34 0.644 0.2 69 F
2 72 34 0.658 0.025 41 F
3 72 54 0.593 0.2 4 F
4 72 69 0.627 0.2 2F

Test Series V-VI, Drained

Initial
Void N
Test Test r 5 fr rcfr Ratio Frequency (to failure, F
Series Number (%) (%) e f no failure, NF) 'n
v 1 83 59 0.640 0.2 200 NF 0 62%
2 83 59 0.631 0.025 50 NF 2 187c
VI 1 94 92 0.647 0.2 200 NF 10.09'/'c
2 94 92 0.635 0.025 40 NF 9 26%
3 94 92 0.663 0.025 40 NF 1244 7c

N = Number of cycles
1'n = Residual strain value for the first cycle

T = 0.21kg/cm 2 (20.61kPa)
uvi = 1kgjcm 2 for all tests (98.16kPa)

158
 The development of strain becomes more complicated
~r-------------------------------------~
for non-symmetrical loading, due to the initial locked-in
static shear stress. Figure 4b shows that for the
non-symmetrical tests, the plot of total shear strain versus
time exhibits cyclic characteristics, but increases in the
direction of the initial locked-in static shear stress. The total
shear strain ( 'Yt) is comprised of a residual shear strain ( 'Yr)
and a cyclic shear strain (I'c); ( 'Yt) = ('Yr) + ('Yc). The
residual shear strain is a measure of the portion of the total
strain that accumulates in the direction of the initial
locked-in static shear stress (Andersen et al., 1988; Malek et
al., 1987; and Goulois et al., 1985, refer to this as the average
shear strain, 'Yave)·

When applied to field conditions, 'Yr represents the

drifting portion of 'Yv and causes the permanent downslope 8
deformation of the soil element. f'c, in the non-symmetrical or-----:-----~~--~------~----~
8 12 16

tests, is the portion caused by the cyclic loading, and cycles SHEAR STRAIN, 1 t'l.l
around 'Yr· The relationship, 'Yt = 'Yr + 'Yc• is also applicable
to the symmetrical tests (with 'Yr = 0). Therefore, Figure 6 Stress-Strain Curve for the Drained
comparisons of shear strains between symmetrical and Monotonic Test
non-symmetrical loading cases can be made.
The strain accumulated from the drained creep
initiated by the application of the initial locked-in static
shear stress is not part of the strain values ( 'Yc, 'Yr or 'Yt)·
Therefore, for all of the tests with an initial locked-in static
shear stress, the initial strain is zero. It is also important to average value of_ the pea.!c normalized shear strength ( r)
note that the shear strength of the soil sample may change equals 0.21. Th1s value IS taken as the normalized shear
slightly with the application of the initial locked-in static stre~gth o_f the SFSS and _will be referred to throughout the
shear stress. However, the value of r used for the stress cyclic testmg as r, the static shear strength of the soil.
ratios; 'Ys/ r and ref r , is the value obtained from running the
static monotonic shear tests without an initial locked-in Stress-Strain Behavior (Drained)
static shear stress.
Figure 6 shows the stress-strain behavior for a drained
Static Tests static monotonic shear test of SFSS. Because the sample
gains strength throughout the entire test, the determination
Stress-Strain and Pore Pressure Behavior (Undrained) of the normalized drained shear strength of the SFSS is
arbitrary, and was chosen to be equal to the normalized
Figure 5 exhibits the stress strain behavior of three undrained shear strength. The normalized shear strength
undrained static monotonic tests. As can be seen, the ( r = 0.21) occurred at approximately 9% strain (see Fig. 6).
stress-strain curves of the three tests are very consistent. This ensured that sizable strains would develop in the drained
They all reach a peak shear strength at_ about 1% strain and cyclic loading tests, so that comparisons of accumulated
drop off slightly throughout the remamder of the test.The deformation for varying frequencies could be made.

Cyclic Tests

~~-----------------------------------------------, Background
Twenty-seven load controlled cyclic NGI-DSS tests
were performed; twenty-two constant volume (undrained),
and five drained (see Table 2). Twenty-one of these tests
had initial locked-in static shear stresses applied to the
sample prior to the cyclic loading. The cyclic tests are
52,3 differentiated by the ratio of their respective normalized
51
locked-in static shear stress to the normalized undrained
shear strength of the SFSS, (r 5 fr). Sixteen constant volume
cyclic shear tests had initial locked-in static shear stress
ratios ( r 5 / r) of 0%, 36%, 55%, and 72%. The five drained
cyclic shear tests had initial locked-in static shear stress
ratios (r5 /r) of 83% and 94%. Various values of cyclic shear
stress ratios (rcfr, determined in the same manner as r5 /r)
~~------~~-------T--------~---
0
SHEAR STRAIN,
•
t'l.l
-----~------~~
12
<
18 20 were used. rc/r equal to 34%, 54% and 69% were used for
the constant volume tests, where as rcfr equal to 59% and
Figure 5 Stress-Strain Curves for the Constant 92% were used for the drained tests.
Volume Monotonic Tests.

159
 Since the primary goal of this research was to
determine the effect that the frequency of cyclic loadi.ng had
on the accumulated deformation of the SFSS, different 2,-------------------------------~
frequencies of cyclic loadin_& were applied to the samples
tested. Three frequencies (0.2 Hz, 0.05 Hz, and 0.025 Hz)
were used in this study. The variation of frequency along 69%
.2Hz
with r 5 /r and rcfr can be seen in Table 2.
Strains and Pore Pressure Buildup (Undrained)
The results of Test Series II, with r 5 / T = 36% and 51%
.2Hz
varying frequencies of cyclic loading, are shown in Figures 7,
8, and 9. Both the residual and cyclic strains accumulate
quickly, and the pore pressure builds up faster than would be
expected for a symmetrical loading test, due to the initial
locked-in static shear stress. The normalized pore pressures 31% 31%
(Fig. 9) increase to values around 0.8 for the tests reaching
tailure and remain low for those which do not. Zimmie and I Ill
NlH!ER Of' CYCLES, N
Lien (1983) have also observed that for non--symmetrical
undrained cyclic loading the normalized pore pressures are Figure 7
between 0.6 and 0.8 at failure. Development of Cyclic Strain in the
This is in contrast to Constant Volume
symmetrical loading tests, where pore pressures equal about Tests With
1.0 at failure. rsfr = 36% (Curves Labeled by rcfr
and Frequency of Loading Respectively)
Figure 8 shows the effect of frequency on the residual
strain for Test Series II. The number of cycles to failure
decreases with slower frequencies. Test #4, with a frequency ~r-----~----~5~1%~.------------------------~
four times slower than Test #3, fails 3. 7 times faster. In 51%
Test Series II (refer to Table 2), Test #4 with a faster .2Hz
1%
frequency than Test #5, fails in less cycles than does Test #5 02SHz
having a lower void ratio than the other test samples in Test
Series II.

Strain Development (Drained)

Five drained NGI-DSS tests, with two different initial
locked-in static shear stresses and varying frequencies of
clclic loading, were run in Test Series V and VI (see Table
2 . Due to the contractive nature of the SFSS, there is no
c ear definition of failure for the drained testing in this
31% 31%
investigation. The sample densified and gained strength .2Hz
throughout the entire drained static testing. Thus, the I Ill I Dill
drained cyclic tests must be analyzed differently than the NU'1BER Of' CYCLES, N
constant volume (undrained) cyclic tests.
Figure 8 Development of Residual Strain in the
Unlike the constant volume (undrained) cyclic tests, in Constant Volume Tests With
which the cyclic strain ( 'Yc) increases until the sample fails, in rsfr = 36% (Curves Labeled by rcfr
the drained tests a maximum value of cyclic strain is and Frequency of Loading Respectively)
produced by the first cycle, but decreases rapidly with
additional loading cycles (see Figure 10). The residual strain
( 'Yr) increases moderately throughout the entire test (see
Figure 11). Because pore pressures are not built up and
residual strain is not accumulated to the point of failure, a
comparison between the number of cycles at failure (N r) and
the frequency of applied cyclic loading (f) cannot be made.
For these tests, the residual strains accumulated by
the first cycle of loading ( 'Yn) are compared to the frequencies
of the applied loading. The first cycle is used for comparison 31%
because it is the only point in the test at which the void ratio .2Hz
of the soil specimen is known. Because the sample's void
ratio changes by varying amounts depending on how the
sample is tested (i.e. r 5 /r, rcfr, and f), it would not be
useful to make any comparisons between the frequency and
the strains accumulated by any cycle other than the first.
This procedure has been followed previously (Baziar et all, 10 100
Nut1BER Of CYCLES, N
I Dill

1990).
Figure 9 Development of Normalized Pore
Table 2 lists the residual strain produced by the first Pressure in the Constant Volume Tests
loading cycle in the five drained tests (Test Series V and VI), With r5 /r = 36% (Curves Labeled by
and it can be seen in Figure 11 as well. In Test Series V, the rcfr and Frequency of Loading
residual strain in Test #2 was 3.5 times larger than in
Test #1. Respectively)

160
 ~r--------------------------------------------, ~r-------------------------------.

;~~
~G
tlo 62%

~~ ~ ~=---

1___________:::::::~~~==~:·:0:25;H;z~====~~_j
62% 62%

0
.2Hz
1 10 100 10 100
NUI1B!:R Of CYCLES, N NUt1BER Of CYCLES, N

Figure 10 Development of Cyclic Strain in the Figure 11 Development of Residual Strain in the
Drained Non-Symmetrical Tests With Drained Non-Symmetrical Tests With
T sl T = 83% (Curves Labeled by T cl T
rs/T = 83% (Curves Labeled by rc/T
and Frequency of Loading Respectively) and Frequency of Loading Respectively)

SUMMARY AND CONCLUSIONS

The constant volume (undrained) behavior of the
TABLE 3 SFSS, under both monotonic and cyclic loading, is
determined to be quite similar to that observed by others.
Summary of Freauency Effect
The development of strains and pore pressures in the constant
volume (undrained) tests is similar in nature to other te~ts
performed using the NGI-DSS (Malek et al., 1987; Goulms,
Effect of
Frequency 1982; Andersen, 1975, 1976; Andersen et al., 1980; Lien, 1983;
Difference Effect Expected Zimmie and Lien, 1986; Dyvik, 1981).
Frequency on Nf (Undrained) !rom Newmark's
Tests Difference 1,, (Drained) Analys1s Of the tests performed, six good comparisons can be
made (four constant volume (undrained) and two drained) in
Constant Volume which the effect of frequency can be seen (see Table 3). In all
!Undrained) Tests tests, the frequency at" which the test was run did affect the
Test Series II amount of deformation accumulated in the sample. All tests,
#3 & 4 4x 3.7x 16x with two exceptions, show that the slower the frequency, the
Test Series II greater the accumulated deformation. The lower void ratio of
#6 & 7 4x 1.5x 16x these two exceptions is most likely the reason for this (see
Test Series lil
Table 2).
#3 & 4 8x 3.3x 64x
The results of this investigation show that there is a
Test Series IV
#I & 2 8x 1.7x 64x definite inverse effect of the frequency of the cyclic loading on
the accumulated deformation of SFSS with an initial
Drained Tests
locked-in static shear stress. Although the inverse effect of
frequency found in this investigation is small, it is not due to
Test Series V the scatter of the test results. The results of this study are
#1 & 2 8x 3.5x 64x
consistent and reproducible. The effect of frequency
Test Series VI measured in this investigation is substantially less than the
#1 & 3 8x !.2x 64x
inverse-square effect obtained from Newmark's sliding block
analogy. This is as expected, since Newmark's method can be
considered as an upper bound of the frequency effect.
Both the constant volume (undrained) and the drained
tests exhibit an effect of frequency. This effect appears to be
similar between the constant volume (undrained)tests and the
drained tests. However, due to the difference in the methods
by which the effect of frequency is determined for the
constant volume (undrained) tests (Nf) and the drained tests
( 'Yr 1), it is difficult to make any direct comparisons. Baziar
(1988) also witnessed an effect of frequency for undrained
In Test Series VI, Test #3 accumulated a residual non-symmetrical loading triaxial tests on Ottawa F-125
strain in the first cycle which was 1.2 times larger than in sand. (It may be important to note that the specimens tested
Test #1. Test #2, in Test Series VI, is not used for by Baziar et al. were dilative, whereas those in this
comparison due to its lower actual accumulated residual investigation where contractive).
strain in the first cycle than in Test #1. The void ratio in
Test #2, being much less than that in Test #3, is most likely The results of this investigation should be considered
the reason for the lower accumulation of residual strain in the when making any comparisons of soil behavior in the
first cycle of Test#2. laboratory to that of soil in the field. Cyclic testing in the
laboratory may be done using slower frequencies than which

161
are found in the field. This is the usual case for earthquake to the U.S. Army Corps of Engineers, Waterways
loading, especially for tests on fine grained soils (silts and Experiment Station, February.
clays), where it is difficult to conduct proper tests at typical
earthquake frequencies (about 1 Hz). Goodman, R.E., and Seed, H. Bolton (1966),
"Earthquake--Induced Displacements in Sand
For example, one could conclude that due to a faster Embankments," Journal of the Soil Mechanics and
frequency found in the field, the accumulated deformation Foundations Division, ASCE, Vol. 92, No. SM2,
will be considerably less than that found in the laboratory, March, pp. 125-146.
and that for this reason the situation is safe. However, with
respect to the results found in this investigation, accumulated Goulois, Alain M., Whitman, Robert V., Hoeg, Kaare (1985),
deformations occurring in the field may not be much different "Effects of Sustained Shear Stresses on the Cyclic
from those observed in the laboratory. Degradation of Clay," Special Technical Testing
Publication 883, ASTM, Philadelphia, P A,
pp. 336-351.
ACKNOWLEDGEMENT
Lien, Chin-Yue (1983), "Response of Pelita Clay Subjected
This work was supported by the National Center for to Combined Cyclic and Initial Static Shear Stresses,"
Earthquake Engineering Research, Buffalo, New York, under M.S. Thesis, Rensselaer Polytechnic Institute, Troy,
grant number 86-2021B and 87-1308 under auspices of the N.Y.
National Science Foundation, master contract number
ECE 86-D7591. Funds were also provided directly by the Malek, Aziz M., Azzouz, Amr S., Baligh, Mohsen M., and
National Science Foundation. Germaine, John T. (1978), "Undrained Cyclic Simple
Shear Behavior of Clay with Application to Pile
Foundations Supp~rting Tension Leg Platforms,"
REFERENCES Massachusetts Institute of Technology Sea Grant
Publication (MITSG) 87-20, 381 pgs.
Andersen, K.H. (1975), "Repeated Loading on Clay.
Summary and Interpretation of Test Results," NGI Moriwaki, Y., Idriss, I.M., and Doyle, E.H. (1982),
Research Report #74037-9, October. "Earthquake--Induced Deformations of Soft Clay
Slopes," Journal of the Geotechnical Engineering
Andersen, K.H. (1976), "Behavior of Clay Subjected to Division, ASCE, Vol. 108, No. GT11, November,
Underdr~ned Cyclic Loading," Proceedings, pp. 1476-1493.
InternatiOnal Conference on the Behavior of Off-Shore
Structures, Norwegian Institute of Technology Newmark, N.M. (1965),, "Effects of Earthquakes on Dams and
Trondheim, Vol. 1, pp. 392-402. Also NGi Embankments,' Geotechnique, Vol. 5, No. 2,
Publication No. 114, pp. 33-44. pp. 139-160.

AnderseWn, K.(H., gPo~~, J.H., Brown, S.F., and Rosenbrand, Pierce, William G. (1982), "The Behavior of Monterey
.F. 1980 , Cyclic and Static Laboratory Tests on #0 Sand Under Cyclic Strain Controlled Simple Shear
Drammen lay," Journal of Geotechnical Engineering Loading," M.S. Thesis, Rensselaer Polytechnic
Division, ASCE, Vol. 106, No. GT5, May, Institute, Troy, New York.
pp. 499-529.
Seed, H.B., Lee, K.L., and Idriss, I.M. (1968), "An Analysis
Andersen, K.H., Kleven, A., Heien, D. (1988) "Cyclic Soil of the Sheffield Dam Failure," Report No. TE--68-2
Data for Design of Gravity Structures'," Journal of University of California, Berkeley, CA. '
Geotechnical Engineering, ASCE, Vol. 114, No. 5,
May, pp. 517-555. Seed, H.B., Lee, K.L., Idriss, I.M., and Makdisi, F.l. (1973),
"Analysis of the Slides in the San Fernando Dams
Andersen, K.H. and Lauritzsen, R. (1988), "Bearing Capacity During the Earthquake of Feb. 9, 1971, "Report No.
for Foundations with Cyclic Loads," Journal of EERC 73-2, Earthquake Engineering Research
Geotechnical Engineering, ASCE, Vol. 114, No. 5, Center, College of Engineering, University of
May, pp. 540-555. California, Berkeley, June.

Baziar, M.H., S,uccarieh,. M., Elgamal, A. and Dobry, R. Vasquez-Herrera, A. (1988), "The Behavior of Undrained
(1990), 'EvaluatiOn of Permanent Deformation Contractive Sand and Its Effect on Seismic
Through Laboratory Strain-Controlled Cyclic Liquefaction Flow Failures of Earth Structures "
Testing," Draft Report for NCEER, Buffalo, New Ph.D. Thesis, Rensselaer Polytechnic Institute, Troy,
York, Rensselaer Polytechnic Institute, Troy, New New York.
York.
Vucetic,, Mlad~n (1986), "Pore Pressure Buildup and
Bjerrum, L. and Landva, A. ( 1966), "Direct Simple Shear LiquefactiOn at Level Sandy Sites During
Tests on Norwegian Quick Clay," Geotechnique, Earthquakes," Ph.D. Thesis, Rensselaer Polytechnic
Vol. 16, No. 1, pp. 1-20. Institute, Troy, New York.
Dyvik, R. (1981), "Strain and Pore Pressure Behavior of Zimrnie, T.F., ~ask~ratos, P., and Lien, C.Y. (1982), Static
Fine Grained Soils Subjected to Cyclic Shear and Cyclic Simple Shear Tests on Pelita Samples for
Loading," Ph.D. Thesis, Department of Civil the Alicura Dam site (Argentina). Reports Civil
Engineering, Rensselaer Polytechnic Institute, Troy, Engineering Department, Rensselaer Polytechnic
New York. Institute, Troy, New York, January and March.
Floess, C.H.L. (1979), "Direct Simple Shear Behavior of Fine
Grained Soils Subjected to Repeated Loadings," Ph.D. Zimrnie, T,.F. and Lien,. C.Y. (1986), "Response of Clay
Thesis, Rensselaer Polytechnic Institute, Troy, New SubJected to Combmed Cyclic and Initial Static Shear
York. Stress," Proceedings, 3rd Canadian Conference on
G.E.I., Geotechnical Engineers, Inc. (1988), "Re-evaluation Marine Geotechnical Engineering, Vol. 2 Memorial
of the Lower San Fernando Dam," Report submitted University, St. John's, Newfoundland, pp. 655--673.

162