You are on page 1of 4

Indian Geotechnical Conference 2010, GEOtrendz

December 1618, 2010


IGS Mumbai Chapter & IIT Bombay

Behaviour of Soil Under Cyclic Loading

Raghunandan, M.E. Juneja, A.


Research Scholar Assistant Professor
e-mail: raghunme@iitb.ac.in e-mail: ajuneja@iitb.ac.in

Department of Civil Engineering, IIT Bombay, Mumbai

ABSTRACT
During seismic loading, natural soil is subjected to cyclic shear stresses at different amplitudes and frequencies
that will induce transient and permanent deformations. Literatures demonstrate the significant influence of dynamic
soils properties on the response and stability of soil layers. The objective of this paper is to address some of these
issues. Strain controlled cyclic triaxial tests were conducted at +/-0.4% cyclic shear strains and 1Hz frequency on
large 100mm diameter and 200mm high reconstituted triaxial samples prepared using soil samples collected from
a pilot area in Mumbai, and a clean sand sample. The response of soil samples to applied shear strains are
presented as decrease in cyclic stress and development of pore pressure with number of loading cycles. Further,
response of the soil is compared with that of clean sand. Results showed that the soil samples are less susceptible
to liquefaction compared to clean sand, with G and D values around 9 104kN/m2 and 15% respectively at 0.4%
shear strains, .

1. INTRODUCTION or the sample preparation techniques used in case of


Adequate information on dynamic soil properties, including remoulded and reconstituted samples (Raghunandan and
dynamic shear modulus, damping ratio, pore pressure Juneja 2011). Many researchers (Lee and Seed 1967, Seed
response and cyclic strength, are more essentially considered and Idriss 1970, Mulilis et al 1977, Ladd 1977, Okur and
in ground response and soilstructure interaction problems. Ansal 2007, Lee and Sheu 2007) have studied the effect of
Previous works demonstrates many in-situ and laboratory this issue on the measured dynamic soil properties. Few
tests to determine these properties, from which, considering literatures also attempt to compare the dynamic soils
the economy involved in field tests such as cross bore hole properties obtained from laboratory and field tests (eg.
survey, geophysical techniques, laboratory cyclic tests are Kokusho 2004).
more commonly used. Cyclic triaxial, bender elements, cyclic This paper examines the behaviour of saturated soil
simple shear, resonance column are some of the commonly samples under cyclic loading and further compared with that
used laboratory test equipments to evaluate the dynamic of a sand sample (Juneja and Raghunandan 2011). The tests
properties of soil, under low strain to high strain amplitudes conducted in this study are first described. Sample
(Lee and Seed 1967, Seed and Idriss 1970). preparation technique and the procedure used in the cyclic
Soil response to cyclic or monotonic loading as obtained triaxial tests are then briefly described. The data from
in laboratory mainly depend upon the quality of the soil laboratory testing are used to study the variation of cyclic
sample used and the errors associated with the testing stress, pore pressure built up, and liquefaction resistance of
equipment. Jardine et al. (1984) explained the important locally available soil samples compared with that of the sand
sources of large unaccountable errors in conventional triaxial sample.
experiments (1) difficulty in trimming a sample so that the 2. EXPERIMENTAL PROGRAM
end faces are perpendicular to the vertical axis of symmetry,
All tests were conducted in ADsoil laboratory at IIT
(2) play in the connection between the load cell and the sample
Bombay. Triaxial samples of 100mm diameter and 200mm
top cap, and (3) bedding down at the ends of the sample, due
long were prepared using locally available disturbed soil
to local surface irregularities or voids. Whereas the quality
samples. Samples from four different sites were collected
and repeatability of the sand samples is highly dependent on
from a pilot area located in north eastern region of Mumbai,
the methodology used for procuring undisturbed soil samples
between the Vihar and Powai lakes. The bore log data of
196 M.E. Raghunandan and A. Juneja

Table 1: Properties of the Soil Samples Collected from the Pilot Area

Sample Specific Gravity, GS Gravel % Sand % Silt % Clay % Liquid Limit Plastic Limit
Sample-1 2.67 0 13 40 47 48 23
Sample-2 2.66 15 60 17 8 44 25
Sample-3 2.68 7 53 30 10 48 26
Sample-4 2.67 21 46 15 18 38 18

the site showed the presence of rocky strata within a shallow back pressure increments. Saturation was continued till the
depth of 0.1- to 4.0m, and the top layers composed of loose Skemptons (1954) B-factor was achieved close to 1. At
fills of sandy and gravelly soils with varied percent of clay. the end of saturation, the samples were consolidated at
Samples were collected from the top layer (1.0m depth). required effective stress (c) and then sheared under cyclic
Table 1 shows the properties of all the soil samples as loading with drainage valve closed. Shear strains of
obtained from various soil test reports. Also, it is noted +/ 0.4% were maintained in the tests. Figure 2a-c shows
that all the samples collected were above the ground water the input loading in terms of cyclic axial strains and
table, with 10 to 12% water content and dry density around recorded data of cyclic deviator stress and pore water
18kN/m2. pressure with number of loading cycle from a cyclic triaxial
Also used for comparison is a river bed sand obtained test conducted on sand sample.
from Gujarat region. Figure 1 shows the particle size 0.3
distribution curves of all the samples. Results that the sand (a)
sample used is uniformly graded with very less fines, having Cyclic axial strain (%) 0.2

mean grain size, D50 = 0.3mm, uniformity coefficient, 0.1


CU = 2.12, and coefficient of curvatures, CC = 1.47. Specific 0
gravity of the sand is equal to 2.63. The maximum and
minimum void ratios of the sand sample were 0.795 and -0.1
0.492, respectively. -0.2
100 -0.3
0 5 10 15 20 25 30
Number of cycles
80
200
Percentage finer (%)

Deviator stress (kN/m2)

(c)
150
60
100

40 Sample-1 50
Sample-2 0
Sample-3
20 -50
Sample-4
Sand sample -100
0 0 5 10 15 20 25 30
0.01 0.1 1 10 100 Number of cycles
Sieve size (mm)
Fig. 1: Grain Size Distribution Plots of Soil Samples Used in 1
Pore pressure ratio, ru

the Present Experimental Study (b)


0.8
Triaxial samples were prepared using dry tamping
technique. Details of sample preparation technique are 0.6
discussed elsewhere (Raghunandan and Juneja 2010). In
0.4 RD = 60%; c = 310kN/m2;
essence, oven dried soil samples were tamped in 3 layers
with a 50mm diameter tamper of 240g weight. The triaxial 0.2
b = 110kN/m2; p'0 = 200kN/m2;
samples were prepared directly on the cell base to minimize Frequency = 1Hz
the disturbance involved during placing the samples. Dry 0
unit weight was maintained close to the field measurement. 0 5 10 15 20 25 30
Samples were then flooded with carbon-di-oxide gas Number of cycles
Fig. 2: Applied Loading and Response of Soil Samples at
followed by de-aired water under an effective stress of about Different Loading Cycles (a) Axial Strain (b) Deviator Stress
15- to 20kN/m2. Further, saturation was achieved using (c) Pore Water Pressure
Behaviour of Soil Under Cyclic Loading 197

3. RESULTS AND DISCUSSIONS 0.4

Five soil samples were tested using a strain controlled Sample-1


cyclic triaxial tests, conducted on a servo controlled, Sample-2
pneumatic actuator to study the cyclic behaviour of 0.3 Sample-3

Cyclic stress ratio


collected soil samples compared with that of clean sand. Sample-4
The deviator stress and pore pressure build-up was Sand
recorded. 0.2
Figure 3 shows the variation of pore pressure ratio (ru)
with number of loading cycles (N) for the soil samples.
In this paper, pore pressure ratio (ru) is defined as the ratio 0.1
of excess pore water pressure to the effective confining
pressure. As r u reaches unity, the excess pore water
pressure reaches the total stress, soil deposit suddenly 0
moves from solid to a liquefied state, this phenomenon in 0 100 200 300 400 500
general is described as soil liquefaction. Observations from Number of cycles, N
Fig. 4: Variation of Cyclic Deviator Stress with Number of
Figure 3 show that the soil samples showed peak ru equal
Loading Cycles
to 0.74 even after 500 loading cycles, whilst in case of
sand sample, ru reached unity at around 32 cycles. 2.5
(a) = 0.15%
1
2 = 0.30%
Cyclic stress ratio, CSR
= 0.45%
0.8 1.5 = 0.60%
Pore pressure ratio, ru

= 0.75%
1 = 0.90%
0.6
= 1.05%
ru at N=500 0.5
0.4 Sample-1 0.74
Sample-2 0.62 0
0 0.2 0.4 0.6 0.8 1 1.2
Sample-3 0.62 Normalised number of cycles, N/NL
0.2
Sample-4 0.64
1.2
Sand NL = 10 (b)
0
1
Pore pressure ratio, ru

0 100 200 300 400 500


Number of cycles, N = 0.15%
Fig. 3: Pore Water Pressure Response with Number 0.8 = 0.30%
of Loading Cycles
= 0.45%
Figure 4 shows the variation of cyclic deviator stress 0.6 = 0.60%
with number of loading cycles. Figure explains the decrease
in strength of soil samples with cyclic loading, and = 0.75%
compared for different samples tested in this study. The 0.4 = 0.90%
deviator stress is expressed as cyclic stress ratio (CSR). = 1.05%
CSR is the ratio of cyclic deviator stress to the initial mean 0.2
effective principle stress. As expected the soil samples 0 0.2 0.4 0.6 0.8 1 1.2
collected from the site showed CSR value of about 0.15 Normalised number of cycles, N/NL
even after 500 loading cycles, whilst in case of sand sample, Fig. 5: Effect of Shear Strains on Cyclic Behaviour of
CSR reached to zero at 32 cycles indicating liquefaction. Sand Samples; (a) CSR; (b) Pore Pressure Ratio
Hence, as evident from previous literatures, the soil samples Figure 5 shows the effect of applied shear strains on
collected from the study area is less susceptible to the cyclic behaviour of sand samples. Results showed a
liquefaction. huge variation in the number of cycles required for
198 M.E. Raghunandan and A. Juneja

liquefaction (NL) from around 8 to 45 cycles. Hence, the CSR reached to zero at 32 cycles indicating liquefaction.
number of cycles is normalised with the number of cycles G and D values for all the samples were observed around 9
required for liquefaction, N/NL. Figure 5a shows CSR value
4
kN/m2 and 15% respectively at 0.4% shear strains, .
1 0

of 0.5 to 2 in the first loading and further decrease with


ACKNOWLEDGEMENT
increase in number of loading cycles. This is due to the
accumulation of excess pore water pressure as shearing is The first author gratefully acknowledges the support of IIT
continued as shown in figure 5b. Further, CSR reaches to Bombay in providing research scholarship to pursue his
zero and ru reaches unity as N/NL reaches unity, indicating PhD program at the institute. The authors are also
liquefaction. profoundly grateful to Estate office of IIT Bombay for its
help to obtain the borehole data.
80
Sample: L5-1 REFERENCES
d = 13.3kN/m3 E Jardine, R.J., Symes, M.J. and Burland, J.B. (1984). The
'c = 150kN/m2 measurement of soil stiffness in the triaxial apparatus.
Deviator stress (kN/m2)

40
(a)cyc = 0.25% Geotechnique. 34(3), 323-340.
AT Juneja, A., and Raghunandan, M.E. (2011). Effect on cyclic
Cycle No. 3
0 response and liquefaction resistance due to de-saturation
of sand. Geotechnical Special Publication, ASCE
AL (GEO-FRONTIERS-2011) (Accepted).
-40 Kokusho, T. (2004). Nonlinear site response and strain
dependent soil properties. Current science 87(10), 1363-
1369.
-80 Ladd R.S. (1977). Specimen preparation and cyclic stability
-0.4 -0.2 0 0.2 0.4 of sands. Journal of geotechnical and geoenvironmental
Axial strain (%) engineering ASCE. 103(6), 535-547.
Fig. 6: Loading-Unloading (Hysteresis) Loop of Third Lee, C.J. and Sheu, S.F. (2007). The stiffness degradation
Cycle on Sample-4
and damping ratio evolution of Taipei silty clay under
Recent works on dynamic soil properties explains the cyclic straining. Soil dynamics and earthquake
importance of dynamic material properties to evaluate engineering. 27, 730740.
response and stability of surface soil layers, basic dynamic
Lee, K.L. and Seed, H.B. (1967). Drained strength
soil characteristics needed for dynamic analyses including
characteristics of sands. Journal of Soil Mechanics and
shear modulus and damping ratio (Kokusho 2004, Lee and
Foundation Division ASCE, 93(SM6), 117-141.
Sheu 2007, Okur and Ansal 2007). In this study, shear
modulus (G) and damping ratio (D) are calculated using Mulilis, J.P., Seed, H.B., Chan, C.K., Mitchell, J.K., and
the hysteresis loop as suggested in ASTM D 3999-91. The Arulanandan, K. (1977). Effects of sample preparation
third loading cycle was considered as the representative on sand liquefaction. Journal of geotechnical
cycle and hence used to calculate G and D. Figure 6 shows engineering division ASCE. 103(2), 91107.
a typical hysteresis loop obtained from test conducted on Okur, D.V., and Ansal, A. (2007). Stiffness degradation of
sample-4. G and D values for all the samples varied within natural fine grained soils during cyclic loading. Soil
a range of 9 104kN/m2 and 15% respectively at 0.4% dynamics and earthquake engineering. 27, 843854.
shear strains, . Raghunandan, M.E., and Juneja, A. (2011). Effect of sample
preparation on particle packing. International Journal
4. CONCLUSIONS
of Geomechanics and Geoengineering (tentatively
Much useful data obtained from laboratory strain controlled accepted, under review).
cyclic triaxial tests on soil samples obtained from the study Seed, H.B., and Idriss, I. (1970). Soil moduli and damping
area is presented, and further compared with a clean sand factors for dynamic response analysis, Report No. EERC
sample. Results showed that the soil samples are less 70-10, Fundamental of Soil Mechanics, Elsvier.
susceptible to liquefaction with, CSR value of about 0.15
Skempton, A.W. (1954). The pore-pressure coefficients A
even after 500 loading cycles, whilst in case of sand sample,
and B. Geotechnique. 4(4), 143-147.

You might also like