Mao 20031

Mao, X. & Fahey, M. (2003). Géotechnique 53, No.
8, 715–727
Behaviour of calcareous soils in undrained cyclic simple shear

X . M AO a n d M . FA H E Y
The behaviour of three calcareous sediments—a muddy Nous étudions le comportement de trois sédiments calcaires
silt, a silt and a sand from offshore areas on the North- une vase, un silt et un sable - au large de la plaque nord-
West Shelf of Australia—is studied through undrained ouest de l’Australie, grâce à une série toute simple d’essais
simple shear tests. In monotonic tests, the muddy silt de cisaillement non drainés. Dans les essais monotones, la
shows contractive behaviour and the silt and sand show vase montre un comportement de contraction tandis que le
dilative behaviour. In cyclic tests, a cyclic mobility failure silt et le sable montrent un comportement de dilatation.
mode is observed for the three soils. Responses in non- Dans les essais cycliques, nous observons un mode de
symmetrical cyclic tests were significantly different from rupture de mobilité cyclique pour les trois sols. Les ré-
those in symmetrical cyclic tests. A ‘cyclic phase trans- ponses obtenues dans les tests cycliques non symétriques
formation state’, which separates undrained cyclic stress sont assez différentes de celles obtenues dans les essais
paths into contractive and dilative phases, was identified: cycliques symétriques. Nous avons identifié un ‘état de
this was found to be different from the phase transforma- transformation de phase cyclique’, qui sépare les chemins
tion line obtained from monotonic tests, and it is unique de contrainte cyclique non drainée en phases de contraction
for a particular soil. The three soils show strain hard- et de dilatation : cette transformation s’est révélée différ-
ening behaviour in plots of stress ratio against shear ente de la ligne de transformation de phase obtenue lors des
strain, which can be characterised by a backbone curve tests monotones et elle est unique pour un sol donné. Les
and hysteresis loops. trois sols montrent un comportement de durcissement par
les efforts dans des tracés de rapport d’effort relativement
KEYWORDS: calcareous soils; sedimentation; shear strength; à la déformation de cisaillement, qui peut être caractérisée
simple shear strain localisation par une courbe principale et des boucles d’hystérésis.
INTRODUCTION terised with respect to cyclic stress paths and strain hard-
A variety of calcareous sediments have been encountered in ening behaviour in plots of stress ratio against shear strain
offshore areas on the North West Shelf of Australia, where plots.
oil and gas resources are exploited (Fahey, 1998). The use
of analytical methods for offshore foundation design requires
extensive knowledge of the properties of calcareous sedi- MATERIALS
ments. Most previous studies on calcareous soils have Gorgon muddy silt
focused on the compression and monotonic behaviour. Stud- The first soil, designated Gorgon muddy silt in this paper,
ies of their cyclic behaviour, which is the main concern in was recovered from 18–100 m depth beneath the seabed at
the design of offshore foundations, have been limited, and Gorgon site on the North West Shelf of Australia. The soil
concentrate mainly on the behaviour of calcareous sands is composed of over 95% calcium carbonate. Over 60% of
(Kaggwa, 1988; Airey & Fahey, 1991; Joer et al., 1995; this is in the form of aragonite, with the remainder being
Hyodo et al., 1998). Knowledge of the behaviour of fine- calcite. A typical grading curve of the soil shown in Fig. 1
grained calcareous sediments in undisturbed condition with reveals that approximately 60% of particles in this soil are
high void ratios is limited. Little attention has yet been paid clay-sized (, 2 ì), though it should be noted that none of
to the comparison of the behaviour of different-sized calcar- the particles is clay mineral. The specific gravity, Gs , lies
eous sediments. This paper describes the behaviour of three between that of calcite (2·7) and aragonite (2·9). The liquid
uncemented calcareous sediments covering a wide range of limits lie between 40 and 50, with plastic limits of about 30,
particle size.
Three calcareous sediments—Gorgon muddy silt, Gorgon
seabed silt and Legendre sand, recovered from the North 100
West Shelf of Australia—were studied through undrained Gorgon muddy silt
simple shear tests. A range of different types of test was 80
carried out to study the effects of cyclic and mean shear
Percentage passing: %
stresses on soil responses: undrained monotonic and symme-

trical cyclic simple shear tests on the three soils, and non- 60
symmetrical cyclic simple shear tests on the Gorgon muddy
silt. Similar cyclic responses are observed, even though
40
different monotonic responses are evident among the three
soils. The cyclic responses of the three soils are charac- Legendre
20 Gorgon sand
seabed silt
Manuscript received 21 February 2002, revised manuscript accepted 0

17 January 2003 0.0001 0.001 0.01 0.1 1 10
Discussion on this paper closes 1 April 2004, for further details see Particle size: mm
p. ii.
Centre for Offshore Foundation Systems, Department of Civil and Fig. 1. Grading curves of Gorgon muddy silt, Gorgon seabed
Resource Engineering, University of Western Australia. silt and Legendre sand
715
Downloaded by [ UNIV OF MASSACHUSSETS] on [13/09/16]. Copyright © ICE Publishing, all rights reserved.
716 MAO AND FAHEY
giving plasticity indices, Ip , of 10 to 20. According to the was first dried in an oven and then sieved through a 2·4 mm
classification system proposed by Fookes (1988), the soil sieve. Particles less than 0·6 mm comprise a very small
would be designated calcareous muddy silt. component, and were removed from the material in order to
The soil is uncemented, with a very high in-situ void ratio avoid particle segregation during sample preparation. The
(typically between 1·6 and 1·8 and up to 2·0 in some cases), grading curve of this sand (Fig. 1) indicates that this is
which is attributed to its highly structured fabric. Mao & uniform coarse sand.
Fahey (1999) show scanning electron micrographs of this In preparing samples, the sand was soaked in water and
material that clearly indicate this structure. The undisturbed the mixture was then de-aired. A membrane was attached to
samples were used for triaxial and simple shear tests asso- the bottom cap using two O-rings. A two-segment sample
ciated with a preliminary study for a platform project. The former, with diameter of 82 mm and height of 32 mm, was
leftover material, including material used in these tests then positioned. The saturated sand was then spooned gently
(some of this material had been oven-dried), was collected into the water layer by layer until the height was just above
with the intention of investigating the behaviour of the soil the top of the mould. The sand above the top of the mould
by use of reconstituted samples. However, it was found that was then trimmed and the top cap was then placed. Then
the reconstituted samples possessed much lower void ratios the membrane was pulled on the top cap and secured with
(between 1·23 and 1·31), and their undrained shearing be- two O-rings. The sample, together with the two-segment
haviour was quite different from that of the undisturbed sample former, was then placed into the simple shear
samples. Therefore a method using flocculation for prepara- apparatus. After the top and bottom caps were fixed, the
tion of reconstituted samples was developed (Mao & Fahey, two-segment sample former was removed, leaving a sample
1999). It was found that the compression and shearing ready for testing.
behaviour of the samples reconstituted using this flocculation
method is very similar to that of undisturbed samples (Mao
& Fahey, 1999). It is the behaviour of these reconstituted SIMPLE SHEAR TESTING
samples that is described in this paper as being representa- Apparatuses
tive of that of undisturbed Gorgon muddy silt. Two simple shear apparatuses designated old and new in
The samples were prepared by one-dimensional consolida- this paper were used. The principles of the two apparatuses
tion under a vertical effective stress of 100 kPa in a plastic are similar except for the loading-control systems and sizes
tube with diameter of 50 mm. After consolidation, each of samples. Two drawbacks of the old simple shear appara-
sample was extruded from the tube, and an 18 mm high tus are:
sample was prepared for simple shear testing.
(a) a pneumatically controlled loading system, which
results in difficulty in controlling the horizontal
displacement during cyclic loading
Gorgon seabed silt
(b) fixed sample size; only 50 mm diameter samples can be
The second soil, designated Gorgon seabed silt in this
tested.
paper, was recovered from the surface of the seabed at the
same site as the first soil. The motivation for recovery of In order to overcome the drawbacks of the old simple shear
this seabed material was to conduct centrifuge model tests, apparatus, the new simple shear apparatus was manufactured
in which a large volume of soil was required. The original in house at UWA.
seabed material was first dried in an oven and then sieved In both old and new apparatuses, the sample is enclosed
through a 150 ìm sieve. in an unreinforced latex membrane and contained in a
Preliminary simple shear tests on the seabed material pressurised cell, very similar to a triaxial apparatus. The
(after sieving) showed strong dilative behaviour in undrained vertical and cell pressures are controlled independently. A
shear tests. The purpose of the centrifuge model tests was to feedback system allows total vertical stress to be kept
simulate the behaviour of foundations on the deep material constant during the shearing phase while maintaining a
(Gorgon muddy silt), which had large void ratios and constant sample height. This is achieved by locking the
presented strain softening behaviour in undrained shear. Two vertical loading ram, and using the feedback system to vary
methods were used in order to diminish the dilative behav- the cell pressure to keep the total vertical stress constant. As
iour as follows: (1) a commercially produced carbonate the height and volume are both constant (for undrained
powder named Omya Carb was blended with the seabed tests), the average cross-sectional area is likely to remain
material as a filler, and (2) silicon oil was used as pore fluid constant also.
instead of water in order to reduce cv , the coefficient of
consolidation. Fig. 1 shows that the mixture of the seabed
material and the carbonate powder consists of 73% silt and Testing procedures
27% fine sand. The tests on Gorgon muddy silt and Gorgon seabed silt
The samples were prepared from a slurry, consisting of were conducted in the old simple shear apparatus, whereas
85% dried seabed material after sieving and 15% carbonate the tests on Legendre sand were conducted in the new one.
powder mixed with 70% silicon oil by weight of total dry After the sample had been set up in the apparatus and the
soil mass. After being thoroughly mixed and de-aired, the cell closed, the saturation phase was started. Using the old
soil–oil slurry was poured into a 72 mm diameter tube, and simple shear apparatus, a back-pressure of 200 kPa was
consolidated under a vertical effective stress, 9vc , of 100 kPa applied with a cell pressure of 210 kPa. For the tests in the
for 72 h. The consolidated sample was then extruded from new simple shear apparatus, 400 kPa back-pressure was
the tube and trimmed to 50 mm in diameter and 18 mm in applied with a cell pressure of 410 kPa. A B-value test was
height using a thin-walled ring cutter. conducted at the end of saturation, and generally a B-value
of 95% or greater was achieved in all tests.
After saturation, samples were consolidated. In the old
Legendre sand apparatus, the target cell pressure was applied instanta-
The third soil, designated Legendre sand in this paper, neously. The additional deviator stress required for anisotro-
was recovered from the seabed surface at the Legendre field pic consolidation was ramped up to the target value in 1 h
on the North West Shelf of Australia. The seabed material and maintained for at least 6 h. In the new apparatus, the
CALCAREOUS SOILS IN UNDRAINED CYCLIC SIMPLE SHEAR 717
consolidation stresses were applied by ramping the deviator triaxial test results. It is assumed in this case that there is
stress slowly to the required target value while adjusting the complementary shear stress on vertical planes of the cylind-
cell pressure to ensure that required consolidation stress ratio rical sample. This will be discussed later.
was maintained. The ramping lasted for 0·5 h and consolida-
tion was maintained for 2–3 h. A loading rate of 0·1 mm/
min was applied for monotonic shearing, and a frequency of MONOTONIC TEST RESULTS
0·1 Hz was maintained in cyclic shearing. Stress path (–9v ) plots and stress–strain (–ª) curves
from typical undrained monotonic simple shear tests on
Gorgon muddy silt, Gorgon seabed silt and Legendre sand
Interpretation of test results are shown in Figs 2, 3 and 4 respectively. The samples of
In simple shear testing, a cylindrical sample is enclosed in Gorgon muddy silt and Gorgon seabed silt were consolidated
an unreinforced latex membrane, and a loading shear force under a vertical effective stress, 9vc , of 150 kPa and a
is applied to the top and bottom planes of the sample. No horizontal effective stress, 9hc , of 60 kPa. This is equivalent
complementary shear stresses exist on the vertical boundary to the stress state of a soil sample at 20 m beneath the
planes of the sample. Therefore the distribution of horizontal seabed with K0 of 0·4, which was determined from one-
shear stresses on both the top and bottom planes is not dimensional compression tests in triaxial apparatus for both
uniform. As a result of the horizontal shear force applied to Gorgon muddy silt and seabed silt. Legendre sand—a seabed
the top and bottom planes of the sample, the vertical stress sand—was consolidated under two stress levels (9vc ¼ 75
distribution on these planes is also non-uniform owing to the and 300 kPa) and two consolidation stress ratios (9hc /9vc ¼
need to preserve moment equilibrium. In the interpretation 1 and 0·4), with the intention of seeing the effect of
of test results, horizontal shear stress, , is calculated as the consolidation condition as well as void ratio. As the same
average shear stress in the horizontal direction, whereas sample preparation method is used in each test, all tests
shear strain, ª, is the ratio of sample horizontal displacement started with an initial void ratio, e0 ¼ 1·16. Void ratios of
to sample height. In both monotonic and cyclic tests, no the four tests during undrained shearing varied owing to
strain localisation on testing samples was observed. volumetric strains during the consolidation stage, and are
Later in this paper there is a need to plot simple shear shown in Fig. 4.
stress paths in the q–p9 plane in order to compare them with In the stress path plot for Gorgon muddy silt (Fig. 2(a)),
150 80
Consolidation
σ′vc ⫽ 150 kPa

60
Shear stress, τ: kPa
100 σ′hc ⫽ 60 kPa
CSL 40
50
20
0 0
0 50 100 150 0 10 20 30
Vertical effective stress, σ′v : kPa Shear strain, γ: %
(a) (b)
Fig. 2. Undrained monotonic simple shear test on Gorgon muddy silt: (a) stress paths; (b) shear stress–strain curve
160 160
Consolidation
σ′vc ⫽ 150 kPa
120 120
σ ′hc ⫽ 60 kPa
80 80
CSL PT line
40 40 PT state
0 0
0 40 80 120 160 0 10 20 30
Vertical effective stress, σ′v: kPa Shear strain, γ: %
(a) (b)
Fig. 3. Undrained monotonic simple shear test on Gorgon seabed silt: (a) stress paths; (b) shear stress–strain curve
718 MAO AND FAHEY
250 CSL 250
PT line
200 200 e ⫽ 1.02

σ ′vc ⫽ 300 kPa
150 150 e ⫽ 1.05
σ ′vc ⫽ 75 kPa σ′hc ⫽ 300 kPa
σ ′hc ⫽ 30 kPa e ⫽ 1.02
100 e ⫽ 1.15 100
σ′vc ⫽ 300 kPa PT state
σ′hc ⫽ 120 kPa e ⫽ 1.15
50 σ ′vc ⫽ 75 kPa
e ⫽ 1.05 50
σ′hc ⫽ 75 kPa e ⫽ 1.10
e ⫽ 1.10
0
0
0 100 200 300 400 0 100 200 300
Vertical effective stress, σ′v: kPa Shear strain, γ: %
(a)
(b)
Fig. 4. Four undrained monotonic simple shear tests on Legendre sand: (a) stress paths; (b) shear stress–strain curve. The
void ratio indicated in each test is the value at the end of consolidation
contractive behavioury is observed, with 9v decreasing during after which a dilative response starts, and the whole process
shearing until the critical-state line (CSL) is reached. In a–b–c–d–e–a forms a butterfly-shaped loop. As the stress
contrast, in the equivalent plots for Gorgon seabed silt and path moves towards the critical-state line, this butterfly shape
Legendre sand (Figs 3(a) and 4(a)), dilative behaviour is becomes more and more obvious, indicating that a stronger
evident after the phase transformation (PT) state line dilative response is occurring as cycling progresses. The
(Ishihara et al., 1975) is reached. Even though conducted elbow points b and d, which separate the cyclic stress path
under different confining stresses and void ratios, the four into contractive and dilative phases, are defined as a cyclic
monotonic tests on Legendre sand show similar stress paths. phase transformation (PTcyc ) state in this paper. Features of
Some slight strain hardening behaviour is observed in the the PTcyc state will be discussed later.
stress–strain (–ª) curve for Gorgon muddy silt (Fig. 2(b)). The cyclic –ª curves for the three soils (plot (b) in Figs
However, Gorgon seabed silt and Legendre sand show much 5, 6 and 7) show a cyclic degradation response: that is, the
stronger strain hardening after the phase transformation state stiffness of each successive cycle decreases as a consequence
is reached, as shown in Figs 3(b) and 4(b). For Gorgon of cyclic loading. The shape of the –ª curves changes as
seabed silt, the PT state was reached at ª ¼ 5·4%. For shear strain accumulates, and an S-shaped –ª curve is
Legendre sand, shear strains corresponding to the PT state developed. For the three soils, cyclic strain developed in a
are in the range 2·2–4·4% for the four tests. symmetric pattern as shown in plot (c) in Figs 5, 6 and 7. In
these plots, each vertical line represents a cycle because no
fractions of cycles were recorded in the testing. Shear strain
SYMMETRICAL CYCLIC TEST RESULTS accumulated slowly in the beginning; then the rate of shear
Undrained symmetrical cyclic simple shear tests were strain increment accelerated. For Gorgon seabed silt and
performed on the three soils at the same consolidation Legendre sand, after critical state was reached the shear
stresses as those used in the monotonic tests. A series of strain increased dramatically, whereas for Gorgon muddy silt
symmetrical cyclic tests was carried out on each soil, with a critical state does not indicate a dramatic change in shear
cyclic shear stress ratio (cyc /9vc ) in the range 0·1–0·33. strain accumulation. A cyclic mobility failure mode occurred
Figures 5–7 show the results of one typical test for each for the three soils. For Gorgon seabed silt and Legendre
of the three soils. In plot (a) of each of these three figures, sand, the point at which the CSL is reached corresponds
the cyclic effective stress path in the –9v plane for each to double shear strain amplitudes of 4·5% and 3·2%
soil is shown plotted with the result from the corresponding respectively.
monotonic test. In monotonic shearing, Gorgon muddy silt
shows contractive behaviour, and Gorgon seabed silt and
Legendre sand show dilative behaviour, but in cyclic loading NON-SYMMETRICAL CYCLIC TEST RESULTS
the three soils show similar cyclic stress paths, with 9v Cyclic loading in simple shear tests can be carried out
continuing to decrease as a consequence of cyclic loading. either symmetrically, with the horizontal shear stress being
However, in each individual cycle, the three cyclic stress cycled around zero, or non-symmetrically, with the horizon-
paths show a dilative response after the initial contractive tal shear stress being cycled around a non-zero value, which
response. For example, in the 150th cycle of the test shown is denoted as mean shear stress, m . In order to investigate
in Fig. 6(a), 9v decreases (contractive) before point b and the effect of varying m , non-symmetrical cyclic simple
increases (dilative) after this point until point c is reached. shear tests were carried out on Gorgon muddy silt consoli-
In the unloading process (or loading in the opposite direc- dated anisotropically at 9vc of 150 kPa and 9hc of 60 kPa.
tion), a contractive response occurs until point d is reached, Four types of non-symmetrical cyclic test were performed:
(a) tests with mean to cyclic shear stress ratio, m /cyc ¼
0·4
(b) tests with m /cyc ¼ 1
† As these are undrained tests, there is zero volumetric change. In (c) tests with m /cyc ¼ 2·5.
this paper, the terms ‘contractive’ and ‘dilative’ behaviour are used
to signify a tendency for such behaviour, manifest as a generation Typical results from the three groups, which encompass test
of positive or negative pore pressure changes respectively. conditions of m , cyc , m ¼ cyc and m . cyc , are shown
150 120
130
100
CSL
110
80
90

70 60 150 N ⫽ 130
Monotonic test Monotonic test
190
50 CSL 40
30 50 40 c
20
10
b a
0
⫺10 d
⫺30 30 20 N ⫽ 11 ⫺20 e
⫺50 ⫺40
0 50 100 150 0 50 100 150
Vertical effective stress, σ ′v: kPa Vertical effective stress, σ′v: kPa
(a) (a)
40
40
N ⫽ 20 40
30 N ⫽ 150
30 50 170 190
30
20 160 180
20
10 Shear stress, τ: kPa
10
0
⫺10 0
⫺20 ⫺10
⫺30 ⫺20
⫺40
⫺24 ⫺12 0 12 24 ⫺30
⫺30 ⫺10 10 30
Shear strain, γ: %
Shear strain, γ: %
(b)
(b)
25
20 30
CSL is reached at N ⫽ 28
15
20
10
Shear strain, γ: %
5 CSL is reached at N ⫽ 150

Shear strain, γ: %
10
0
⫺5 0
⫺10
⫺15 ⫺10
⫺20
⫺20
⫺25
0 20 40 60
Cycle number, N ⫺30
(c) 0 50 100 150 200
Cycle number, N
Fig. 5. Symmetrical cyclic simple shear test on Gorgon muddy (c)
silt with ó9vc 150 kPa, ó9hc 60 kPa and ôcyc 25 kPa: (a)
stress paths; (b) shear stress–strain; (c) shear strain against Fig. 6. Symmetrical cyclic simple shear test on Gorgon seabed
cycle number silt with ó9vc 150 kPa, ó9hc 60 kPa and ôcyc 21 kPa: (a)
stress paths; (b) shear stress–strain; (c) shear strain against
cycle number
in Figs 8, 9 and 10 respectively. In the stress path plots for

the three tests (plot (a) in each case), 9v decreases as a reveals the effect of strain rate, which is 0·1 mm/min in the
consequence of cyclic loading until the stress path crossed monotonic test and is approximately 10–20 times this value
the CSL (for the muddy silt) or PT state line (for the silt in the cyclic test (with frequency of 0·1 Hz).
and sand). The discrepancy between the monotonic stress Different stress–strain behaviour is observed for the
path and cyclic stress path of the first cycle in the three tests three tests (plot (b) in each case). High rates of stiffness
720 MAO AND FAHEY
280 80
230 Monotonic test

60 CSL
180 CSL

Monotonic test 40
130
void ratio e ⫽ 1.10
80 20
30
0
⫺20
Cyclic test ⫺20
⫺70
void ratio e ⫽ 1.09
⫺120
0 100 200 300 400 ⫺40
0 50 100 150
Vertical effective stress, σ′v: kPa
Vertical effective stress, σ′v: kPa
(a)
(a)
80 50
60 40
40 30

20 20
0 10
⫺20 0
⫺40
⫺10
⫺60
⫺20
⫺80
⫺20 ⫺10 0 10 20 ⫺30
⫺5 5 15 25 35
Shear strain, γ: %
Shear strain, γ: %
(b)
(b)
15 35
10 CSL is reached at N ⫽ 15 30
γcyc
25
5
Shear strain, γ: %
Shear strain, γ: %
20
0 γt
15
⫺5
10 γm
⫺10
5
⫺15 0
⫺20 ⫺5
0 10 20 30 40 0 2 4 6 8
Cycle number, N Cycle number, N
(c) (c)
Fig. 7. Symmetrical cyclic simple shear test on Legendre sand Fig. 8. Non-symmetrical simple shear test on Gorgon muddy silt
with ó9vc 300 kPa, ó9hc 120 kPa and ôcyc 60 kPa: (a) stress with ó9vc 150 kPa, ó9hc 60 kPa, ôm 12 kPa and ôcyc
paths; (b) shear stress–strain; (c) shear strain against cycle 30 kPa: (a) stress paths; (b) shear stress–strain; (c) shear strain
number against cycle number
degradation and strain softening are evident in the test with Figures 8(c), 9(c) and 10(c) show ª–N plots for the three
shear stress crossing zero (m /cyc ¼ 0·4), as shown in Fig. tests respectively (where N is the cycle number). The ª–N
8(b). However, no stiffness degradation or strain softening responses of the non-symmetrical tests are significantly dif-
are observed in the test with shear stress greater than zero ferent from those in the symmetrical tests, but even among
(m /cyc ¼ 2·5), as shown in Fig. 10(b). It can be seen that, the non-symmetrical tests there are significant differences
as m /cyc ratios increase, the tendency towards stiffness also. The total shear strain, ªt , induced in each cycle can be
degradation and strain softening decreases. divided into mean shear strain, ªm , which is represented by
60 60
CSL
50 50
Monotonic test Monotonic test

40 40
CSL
30 30
20 20
10 10
0 0
0 50 100 150 0 50 100 150
Vertical effective stress, σ′v: kPa Vertical effective stress, σ′v: kPa
(a) (a)
50
50
40
20 40 60 80 100
40
30 Shear stress, τ: kPa
30
20
20
10
30 50 70 90
0 10
0 10 20 30
Shear strain, γ: %
(b) 0
0 5 10 15 20 25
Shear strain, γ: %
35
(b)
30
25
35
Shear strain, γ: %
20
30
15
25
Shear strain, γ: %
10
20
5
15
0 10
0 5 10 15 20
Cycle number, N
(c) 5
Fig. 9. Non-symmetrical simple shear test on Gorgon muddy silt 0

with ó9vc 150 kPa, ó9hc 60 kPa, ôm 22 kPa and ôcyc 0 20 40 60 80 100
22 kPa: (a) stress paths; (b) shear stress–strain; (c) shear strain Cycle number, N
against cycle number (c)
Fig. 10. Non-symmetrical simple shear test on Gorgon muddy

silt with ó9vc 150 kPa, ó9hc 60 kPa, ôm 30 kPa and ôcyc
the average value of shear strain in a cycle, and cyclic shear 12 kPa: (a) stress paths; (b) shear stress–strain; (c) shear strain
against cycle number
strain, ªcyc , which is the amplitude of shear strain within a
cycle, as defined in Fig. 8(c). In the test with m , cyc
(Fig. 8(c)), both ªm and ªcyc accumulate progressively as a
consequence of cyclic loading. In the test with m ¼ cyc a slow rate. In the test with m . cyc (Fig. 10(c)), ªm
(Fig. 9(c)), ªm developed quickly whereas ªcyc increased at increased significantly whereas very little ªcyc was observed.
722 MAO AND FAHEY
In non-symmetrical cyclic tests failure is identified by ªt the PTcyc line in the positive shearing direction is 0·18,
reaching a defined value. which locates below the critical-state line, and the pair of
PTcyc lines in the two shearing directions are symmetric in
the –9v plane.
CYCLIC PHASE TRANSFORMATION STATE These same PTcyc lines are superimposed on the –9v
In undrained monotonic stress paths, a well-defined PT plots of other three tests on the Gorgon muddy silt (Figs
line is observed for Gorgon seabed silt and Legendre sand. 11(b), (c) and (d))—though only one PTcyc line was drawn
In undrained cyclic stress paths, a cyclic phase transforma- in Fig. 11(d), as the effective stress path for this test exists
tion PTcyc state, which separates dilative response from only in one direction. It can be observed that the pair of
contractive response in individual cycles, has been found for PTcyc lines determined from Fig. 11(a) pass through all the
the three soils, even though no PT state is observed in the elbows in each test, even though these tests involved differ-
undrained monotonic stress path for Gorgon muddy silt. In ent cyc , different m and different consolidation stress ratios
the literature, the PT state has been widely recognised in (9hc /9vc ).
undrained monotonic shearing. In order to clarify the PTcyc A similar procedure is applied to the results from the
state and compare it with the PT state, stress paths from Gorgon seabed silt in Fig. 12. Figs 12(a) and (b) show –9v
different types of cyclic test for the three soils are shown in plots for two cyclic simple shear tests on this soil. A pair of
Figs 11–13. PTcyc lines is obtained by linking elbows on the plot for the
Figure 11 presents the effective stress paths (–9v ) from first test (Fig. 12(a)), and then these are superimposed on
two symmetrical and two non-symmetrical tests on Gorgon the stress path for the second test (Fig. 12(b)). This
muddy silt. Three of these tests were conducted on anisotro- reinforces the impression gained from Fig. 11 that the PTcyc
pically consolidated samples and one on an isotropically lines are not affected by the testing conditions. The /9v
consolidated sample. A pair of PTcyc state lines is deter- ratio corresponding to the PTcyc line in the positive shearing
mined by linking elbows of the stress path loops of the test direction is 0·22, which locates below the PT line obtained
in Fig. 11(a). The shear stress ratio (/9v ) corresponding to from the monotonic test, as shown in Fig. 12(a).
60 60
CSL Monotonic test

40 40
20 20
0 PTcyc lines 0
⫺20 ⫺20
⫺40 ⫺40
0 50 100 150 0 50 100 150
(a) (b)
60
60
40 50
40
20
30
0
20
⫺20
10
⫺40 0
0 50 100 150 0 50 100 150
(c) (d)
Fig. 11. PTcyc state lines in stress paths of four undrained cyclic simple shear tests on Gorgon muddy silt: (a)
symmetrical test with ó9vc 150 kPa, ó9hc 60 kPa and ôcyc 38 kPa; (b) symmetrical test with ó9vc 150 kPa,
ó9hc 150 kPa and ôcyc 30 kPa; (c) non-symmetrical test with ó9vc 150 kPa, ó9hv 60 kPa, ôm 12 kPa and
ôcyc 30 kPa; (d) non-symmetrical test with ó9vc 150 kPa, ó9hc 60 kPa, ôm 22 kPa and ôcyc 22 kPa
100 40
Monotonic test
80
20
60

PT line
40
0
20
PTcyc line
0
⫺20
⫺20
⫺40 ⫺40
0 50 100 150 0 50 100 150
(a) (b)
100
100
PTcyc determined from
80 triaxial test shown in Fig. 12(c)
50
Deviator stress, q: kPa
Deviator stress, q: kPa
60
0
40
⫺50
20
⫺100 0
0 50 100 150 0 50 100
Mean effective stress, p′: kPa Mean effective stress, p′: kPa
(c) (d)
Fig. 12. PTcyc state lines in stress paths of undrained cyclic triaxial and simple shear tests on Gorgon seabed silt: (a)
cyclic simple shear test with ó9vc 150 kPa, ó9hc 60 kPa and ôcyc 15 kPa; (b) cyclic simple shear test with ó9vc
150 kPa, ó9hc 60 kPa and ôcyc 21 kPa; (c) isotropically consolidated undrained cyclic triaxial test with p9c 150 kPa
and qcyc 75 kPa; (d) stress path plotted in q–p9 plane for the simple shear test shown in Fig. 12(a)
This is extended further to consider the PTcyc state in through the elbows of the stress path of the simple shear test
cyclic triaxial tests. In order to do this, an undrained cyclic (indicated by the solid dots on the curve), indicating that the
triaxial test was conducted on isotropically consolidated same PTcyc state applies in both triaxial and simple shear
Gorgon seabed silt with consolidation mean effective stress, conditions.
p9c ¼ 150 kPa. The stress path (q–p9) for this test is shown Finally, Fig. 13 examines the PTcyc line for Legendre sand
in Fig. 12(c), and this also shows PTcyc elbow points. A pair in simple shear. The first three plots (Figs 13(a), (b) and (c))
of PTcyc lines is drawn through the origin, linking the elbows present the –9v plots of three tests with different consolida-
on this stress path. The PTcyc line in the positive q quadrant tion stresses and consolidation stress ratios. A pair of PTcyc
is then superimposed in Fig. 12(d), in which the test data in state lines is obtained from the test in Fig. 13(a), which has
Fig. 12(a) are replotted in the q–p9 plane.y It can be seen a gradient of 0·41 and is below the PT line obtained from
that the PTcyc line obtained from the triaxial test passes the monotonic test also shown in this figure. These PTcyc
state lines are superimposed in Figs 13(b) and (c). It can be
y
Plotting simple shear results on the q–p’ plane requires the seen that the pair of PTcyc lines pass through all the elbows
assumption that there is a complementary shear stress on the in these figures.
vertical planes of the simple shear sample. The principal normal Figure 13(d) shows the –9v plot from a slow undrained
effective stresses (91 , 93 and 92 ) can be obtained from the equation cyclic simple shear test, in which unloading–reloading loops
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2 were carried out at a constant loading rate of 0·1 mm/min,
v9 þ h9 v9 h9 which is equal to that in monotonic tests and much slower
19,3 ¼ þ 2 , 29 ¼ h9 ,
2 2 than the rate used in cyclic tests. The pair of PTcyc lines
where 9v and 9h are vertical and horizontal effective stresses obtained from Fig. 13(a) is superimposed in this figure. It
respectively, and is horizontal shear stress. q and p9 are calculated can be seen that the two PTcyc lines pass through the elbows
from q ¼ 91 – 93 , p9 ¼ (91 + 92 + 93 )/3. The assumption may be in the –9v plot, indicating that change of loading rate does
not totally accurate. not affect the PTcyc state.
724 MAO AND FAHEY
280 40
Monotonic test Void ratio, e ⫽ 1.14

200 Void ratio, e ⫽ 1.04
20

120 PT line
40
PTcyc lines
⫺20
⫺40
Void ratio, e ⫽ 1.08
⫺120 ⫺40
0 100 200 300 400 0 20 40 60 80
(a) (b)
120
80
Void ratio, e ⫽ 1.09 Consolidation:

80
σ ′vc ⫽ 75 kPa
σ ′hc ⫽ 30 kPa

40 40 Void ratio, e ⫽ 1.13
⫺40 0
⫺80
⫺120 ⫺40
0 100 200 300 0 50 100
(c) (d)
Fig. 13. PTcyc state lines in stress paths of four undrained cyclic simple shear tests on Legendre sand: (a) cyclic simple
shear test with ó9vc 300 kPa, ó9hc 300 kPa and ôcyc 100 kPa; (b) cyclic simple shear test with ó9vc 75 kPa, ó9hc
75 kPa and ôcyc 25 kPa; (c) cyclic simple shear test with ó9vc 300 kPa, ó9hc 120 kPa and ôcyc 80 kPa; (d) slow cyclic
simple shear test with constant loading rate of 0·1 mm/min
From all of this, it can be concluded that the PTcyc state (which involves non-normalised shear stress, , increasing
is different from, and locates below, the PT state obtained in with increasing ª).
monotonic tests. The PTcyc state appears to be unique for a Figure 14 shows cyclic /9v –ª curves from symmetrical
soil regardless of shear stress levels, consolidation condi- cyclic simple shear tests on the three soils, with the equiva-
tions, loading rates or shearing modes (triaxial or simple lent monotonic /9v –ª curve superimposed in each case. In
shear). The PTcyc state is a function of soil intrinsic proper- each plot, the cyclic data show the type of strain hardening
ties such as particle size and angularity. The mobilised just described, though the loop stiffness (the average gradient
friction angles corresponding to the PTcyc state lines in the of the loop) reduces with each cycle. Because of this loop
–9v plane for Gorgon muddy silt, Gorgon seabed silt and shape, it is possible to superimpose a backbone curve on the
Legendre sand are 10·18, 12·38 and 22·18 respectively. loops, passing through the apices of the loops, in a manner
analogous to the backbone curve used by Hardin & Drne-
vich (1972) to describe the cyclic behaviour of sands when
plotted as –ª. These backbone curves are included in each
STRAIN HARDENING BEHAVIOUR of the three plots in Fig. 14.
The shear stress–strain (–ª) curves in Figs 5–10 show Figure 15 shows three non-symmetrical cyclic simple
quite different response: for symmetrical tests a gradual shear tests with different m (that is, m , cyc , m ¼ cyc
development of S-shaped –ª curves is observed, whereas and m . cyc ) on Gorgon muddy silt. Again, it can be seen
for non-symmetrical cyclic tests very different –ª curves that each /9v –ª curve shows hysteresis, and a backbone
are found as the m /cyc ratio increases. When the shear curve can be superimposed. It must be pointed out that,
stress, , is normalised by the current value of 9v at all owing to shortcomings of the pneumatic load control system
stages of each cycle, the resulting plots of /9v against ª are in the old simple shear apparatus in which the simple shear
of a somewhat similar shape, and show a response that could tests on Gorgon muddy silt and Gorgon seabed silt were
be described as strain hardening, in that the shear stress conducted, the apices of cycle loops were not always prop-
ratio, /9v , increases with increasing strain. This of course is erly recorded (see Figs 14(a), 14(b), 15(a) and 15(b)). How-
a different definition of hardening from that usually used ever, in the /9v –ª curve for Legendre sand, on which tests
σ′vc ⫽ 150 kPa, σ′hc ⫽ 60 kPa

1.0 σ ′vc ⫽ 150 kPa, σ′hc ⫽ 60 kPa 1.2
τm ⫽ 12 kPa, τcyc ⫽ 30 kPa
τcyc ⫽ 38 kPa
0.6
Shear stress ratio, τ/σ′v

0.7
0.2
0.2
Monotonic test
⫺0.2
Monotonic test
⫺0.3
⫺0.6
Backbone curve
⫺1.0 ⫺0.8
⫺15 ⫺10 ⫺5 0 5 10 15 ⫺5 0 5 10 15 20 25 30
Shear strain, γ: % Shear strain, γ: %
(a) (a)
1.6
1.0 σ′vc ⫽ 150 kPa, σ′hc ⫽ 60 kPa
σ ′vc ⫽ 150 kPa, σ′hc ⫽ 60 kPa τm ⫽ 22 kPa, τcyc ⫽ 22 kPa
1.2 τcyc ⫽ 30 kPa
0.8
0.8

0.4 0.6
0
Monotonic test 0.4
⫺0.4
⫺0.8 0.2 Monotonic

Backbone curve test
⫺1.2
⫺15 ⫺10 ⫺5 0 5 10 15 0
Shear strain, γ: % 0 5 10 15 20 25 30
Shear strain, γ: %
(b)
(b)
1.8
1.0
σ ′vc ⫽ 75 kPa, σ ′hc ⫽ 75 kPa
τcyc ⫽ 25 kPa σ′vc ⫽ 150 kPa, σ′hc ⫽ 60 kPa
1.4
τm ⫽ 30 kPa, τcyc ⫽ 12 kPa
1.0 0.8
Shear stress ratio, τ/σ ′v
0.6
0.6
0.2
⫺0.2 Monotonic 0.4 Monotonic

test test
⫺0.6
0.2
⫺1.0 Backbone curve
⫺1.4 0
⫺10 ⫺5 0 5 10 0 5 10 15 20 25 30
Shear strain, γ: % Shear strain, γ: %
(c) (c)
Fig. 14. Stress ratio–strain curves of three symmetrical tests on: Fig. 15. Stress ratio–strain curves of three non-symmetrical
(a) Gorgon muddy silt with ó9vc 150 kPa, ó9hc 60 kPa and ôcyc tests on Gorgon muddy silt: (a) ôm < ôcyc ; (b) ôm ôcyc ; (c) ôm >
38 kPa; (b) Gorgon seabed silt with ó9vc 150 kPa, ó9hc ôcyc
60 kPa and ôcyc 30 kPa; (c) Legendre sand with ó9vc 75 kPa,
ó9hc 75 kPa and ôcyc 25 kPa
were carried out in the new simple shear apparatus, a much bilinear, the multi-linear, the hyperbolic, Ramberg–Osgood
clearer, well-defined backbone curve is obtained (see Fig. and Davidenkov. Among them, the hyperbolic model has
14(c)). been widely used (Kondner, 1963; Hardin & Drnevich,
Idriss et al. (1978) stated that the backbone curve could 1972; Finn et al., 1977; Pyke, 1979; Pradhan, 1989), and is
be expressed in a number of ways, which included the expressed as follows:
726 MAO AND FAHEY
G0 ª 2. However, for unloading the backbone curve is not reached
¼ (1)
v9 G0 ª before reversal loading occurs.
1þ
(= v9 )f
CONCLUSIONS
where /9v is the stress ratio, ª is the shear strain, G0 is the Undrained behaviour of three calcareous sediments—Gor-
initial gradient of the /9v –ª curve, and (/9v )f is the gon muddy silt, Gorgon seabed silt and Legendre sand—has
asymptotic value of the stress ratio. The backbone curves been studied through a series of monotonic and cyclic
shown in Fig. 14 were drawn using this equation, which simple shear tests. Under monotonic loading, Gorgon muddy
shows that the hyperbolic model represents the backbone silt shows contractive behaviour whereas Gorgon seabed silt
curve for each soil very well. and Legendre sand show dilative responses. However, similar
Figures 14 and 15 also show for each case the results of responses are found in undrained cyclic tests among the
an equivalent monotonic test. For each test, the monotonic three soils.
/9v –ª curve is below the backbone curve of the cyclic In symmetrical tests (m ¼ 0), a cyclic mobility failure
results, because the loading rate used in monotonic tests was mode was observed for the three soils. Responses in non-
much slower than that in cyclic tests, in which a constant symmetrical tests (m . 0) are significantly different from
frequency of 0·1 Hz was maintained. those in symmetrical cyclic tests. For the non-symmetrical
Figure 16 plots the /9v –ª curve of a slow cyclic simple test with m , cyc , both ªm and ªcyc accumulated progres-
shear test, in which a loading rate of 0·1 mm/min rather than sively. For the test with m ¼ cyc , ªm developed quickly
a frequency of 0·1 Hz was applied. The result of a corre- while the ªcyc increased at a slow rate. In the test with m .
sponding monotonic test is also plotted in Fig. 16. In this cyc , ªm increased significantly, with very little ªcyc being
case the /9v –ª curve of the monotonic test represents the induced.
backbone curve of hysteretic loops very well. Therefore it A cyclic phase transformation (PTcyc ) state, analogous to
appears that the backbone curve corresponds to the mono- the phase transformation (PT) state observed in monotonic
tonic curve if the shearing rate is similar. tests, has been discussed. The PTcyc state, which separates
From these results, it appears that the hysteresis loops in the stress path (–9v ) in individual cycles into contractive
each symmetrical cyclic test can be described by Masing and dilative phases, is different from the PT state obtained
rules, (Masing, 1926; quoted by Pyke, 1979) which are in undrained monotonic tests, and is found to be unique for
stated as a particular soil in spite of shear stress levels, consolidation
conditions, loading conditions and shearing modes.
(a) the initial shear modulus on each loading reversal
For the three soils, strain hardening behaviour is found in
assumes a value equal to the initial tangent modulus for
/9v –ª plots, which can be characterised by a backbone
the backbone curve
curve and hysteresis loops. Hysteresis loops in /9v –ª
(b) the shape of the unloading or reloading curves is the
curves in symmetrical cyclic tests follow the two Masing
same as that of the backbone curve except that the
rules. In non-symmetrical cyclic tests, the size and shape of
scale is enlarged by a factor of 2.
the hysteresis loops change depending on m /cyc ratios, and
Although the size and shape of the hysteresis loops of the shear modulus on each loading reversal is equal to the
non-symmetrical cyclic tests change, depending on m /cyc initial tangent modulus for the backbone curve. Unloading
ratios, the shear modulus on each loading reversal is equal and reloading curves have the same shape as that of the
to the initial tangent modulus for the backbone curve. Also, backbone curve with the scale enlarged by a factor of 2.
unloading and reloading curves have the same shape as that However, for unloading the backbone curve is not reached
of the backbone curve with the scale enlarged by a factor of before reversal loading occurs.
1.6
Slow cyclic simple shear test

1.2
with σ ′vc ⫽ 75 kPa, σ′hc ⫽ 30 kPa and
constant loading rate ⫽ 0.1 mm/min
0.8
0.4
0.0
0.4
0.8
Monotonic test with
σ′vc ⫽ 75 kPa, σ′hc ⫽ 30 kPa and
⫺1.2 constant loading rate ⫽ 0.1 mm/min
⫺1.6
⫺10 ⫺5 0 5 10
Shear strain, γ: %
Fig. 16. Results of monotonic test and slow cyclic simple shear test with same loading rate
for Legendre sand
ACKNOWLEDGEMENTS Hardin, B. O. & Drnevich, V. P. (1972). Shear modulus and
The work described in this paper forms part of the damping in soils: design equation and curves. J. Soil Mech.
activities of the Centre for Offshore Foundation Systems, Found. Div., ASCE 98, No. SM7, 667–692.
established and supported under the Australia Research Hyodo, M., Hyde, A. F. L. & Aramaki, N. (1998). Liquefaction of
Council’s Research Centres Programme, with additional sup- crushable soils. Géotechnique 48, No. 4, 527–543.
Idriss, I. M., Dobry, R. & Singh, R. D. (1978). Nonlinear behaviour
port provided by the Minerals and Energy Research Institute of soft clays during cyclic loading. J. Geotech. Engng Div.,
of Western Australia, the West Australia Petroleum Com- ASCE 104, No. GT12, 1427–1447.
pany, and Woodside Offshore Petroleum. The first author Ishihara, K. Tatsuoka, F. & Yasuda, S. (1975). Undrained deforma-
was supported by a University Postgraduate Award, and an tion and liquefaction of sand under cyclic stresses. Soils Found.
ad hoc scholarship from the Department of Civil and 15, No. 1, 29–44.
Resource Engineering at UWA. All of this support is grate- Joer, H. A., Fahey, M., Bhattarai, B. & Randolph, M. F. (1995).
fully acknowledged. Assessing liquefaction potential of calcareous soil. Proc. 11th
Eur. Conf. Soil Mech. Found. Engng, Danish Geotechnical
Society, Denmark, 3, 107–112.
Kaggwa, W. S. (1988). Cyclic behaviour of carbonate sediments.
PhD thesis, School of Civil and Mining Engineering, University
of Sydney.
REFERENCES Kondner, R. L. (1963). Hyperbolic stress–strain response: cohesive
Airey, D. W. & Fahey, M. (1991). Cyclic response of calcareous soils. J. Soil Mech. Found. Div., ASCE 89, No. SM1, 115–143.
soil from the North West Shelf of Australia. Géotechnique 41, Mao, X. & Fahey, M. (1999). A method of reconstituting an
No. 1, 101–121. aragonite soil using a synthetic flocculant. Géotechnique 49, No.
Fahey, M. (1998). Some unusual calcareous soils. Proc. 14th Int. 1, 15–32.
Conf. Soil Mech. Found. Engng, Hamburg 4, 2197–2198. Masing, G. (1926). Eigenspannugen und verfestigung beim Mes-
Finn, W. D., Lee, K. W. & Martin, G. R. (1977). An effective stress sing. Proc. 2nd Int. Cong. Appl. Mech., 332–335.
model for liquefaction. J. Geotech. Engng, ASCE 103, No. GT6, Pradhan, T. B. S. (1989). The behaviour of sand subjected to
517–533. monotonic and cyclic loadings. PhD Thesis, Graduate School of
Fookes, P. G. (1988). The geology of carbonate soils and rocks and Civil Engineering, Kyoto University, Japan.
their engineering characterisation and description. In Engineer- Pyke, R. (1979). Nonlinear soil models for irregular cyclic loadings.
ing for calcareous sediments (eds R. J. Jewell and M. Khorshid), Journal of the Geotechnical Engineering Division, ASCE, 105,
pp. 787–806. Rotterdam: Balkema. GT6, 715–726.

Mao 20031

Uploaded by

Copyright:

Available Formats

Mao 20031

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Mao 20031

Uploaded by

Copyright:

Available Formats

Mao, X. & Fahey, M. (2003). Géotechnique 53, No.

Behaviour of calcareous soils in undrained cyclic simple shear

stresses on soil responses: undrained monotonic and symme-

Manuscript received 21 February 2002, revised manuscript accepted 0

σ′vc ⫽ 150 kPa

Shear stress, τ: kPa

100 σ′hc ⫽ 60 kPa

Shear stress, τ: kPa

Shear stress, τ: kPa

5 CSL is reached at N ⫽ 150

in Figs 8, 9 and 10 respectively. In the stress path plots for

230 Monotonic test

Shear stress, τ: kPa

Shear stress, τ: kPa

Shear stress, τ: kPa

Fig. 9. Non-symmetrical simple shear test on Gorgon muddy silt 0

Fig. 10. Non-symmetrical simple shear test on Gorgon muddy

CSL Monotonic test

Shear stress, τ: kPa

Shear stress, τ: kPa

Monotonic test Void ratio, e ⫽ 1.14

Shear stress, τ: kPa

Void ratio, e ⫽ 1.09 Consolidation:

Shear stress, τ: kPa

σ′vc ⫽ 150 kPa, σ′hc ⫽ 60 kPa

Shear stress ratio, τ/σ′v

Shear stress ratio, τ/σ′v

⫺0.8 0.2 Monotonic

Shear stress ratio, τ/σ′v

⫺0.2 Monotonic 0.4 Monotonic

Slow cyclic simple shear test

You might also like