Professional Documents
Culture Documents
Compressibility and Strains
Compressibility and Strains
net/publication/248018778
Article in International Journal for Numerical and Analytical Methods in Geomechanics · July 2000
DOI: 10.1002/1096-9853(200007)24:83.0.CO;2-V
CITATIONS READS
39 455
4 authors, including:
Chandra Desai
The University of Arizona
311 PUBLICATIONS 7,786 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
Analysis and modelling of the pre-failure behaviour of structured clay View project
All content following this page was uploaded by John P. Carter on 23 September 2018.
SUMMARY
The aim of this note is to quantify the in#uence of soil structure on the compression behaviour of natural
soils using the disturbed state concept (DSC). The behaviour of the fully adjusted state is chosen to be that of
the corresponding soil in a reconstituted condition so that the disturbance function is a direct measure of the
e!ects of soil structure. A new DSC compression model is proposed. This model is able to describe the
compression behaviour of structured soils under loading, swelling and reloading. Special versions of the
proposed model are also described for situations (a) where the compression behaviour of the corresponding
reconstituted soils is linear in the e}ln p space and (b) where the compression is one-dimensional. The ability
of the proposed model and its various versions to describe the compression behaviour of structured soils has
been veri"ed. Copyright 2000 John Wiley & sons, Ltd.
KEY WORDS: soil structure; disturbed state concept; compressibility; constitutive equations
1. INTRODUCTION
It is rational as well as advantageous to study the behaviour of natural soils by using knowledge
of the corresponding reconstituted soils as a frame of reference [1,2]. The disturbed state concept
theory (DSC) proposed by Desai [3}6], describes the response of a material by relating it to the
responses of the material at selected reference states. It is believed that the DSC framework may
therefore provide a useful method for describing the di!erence in structured soil behaviour from
that of reconstituted soil, and it may also have the potential to describe the destructuring of
a natural soil with increasing loading. In this technical note, a study of the in#uence of soil
structure on soil behaviour by means of the disturbed state concept is presented. A DSC
compression model is formulated based on a new disturbance function. Simulations of the
proposed DSC model for both structured soil and reconstituted soil are made and compared to
experimental data.
*Correspondence to: M. D. Liu, Department of Civil Engineering, University of Sydney, NSW 2006, Australia
Contract grant sponsor: Australian Research Council
2.3. Fully adjusted state and relative intact state for structured soil
The disturbed state concept allows #exibility in choosing ways to de"ne the two reference states
according to the practical problem of interest and the knowledge available, including its response
in laboratory and "eld tests. A special selection of the reference states is introduced here for
quantifying the in#uence of soil structure on soil behaviour.
The fully adjusted state is chosen to be the corresponding reconstituted state. For example, the
fully adjusted state for the virgin compression of a structured soil is based on the following two
assumptions, viz. (1) the material is reconstituted and has the same mineralogy as the structured
soil, and (2) the soil is in a state of virgin compression and the stress state is the same as that
applied to the structured soil. The relative intact state is chosen to be the &zero state', i.e. the state
with no response to stress (a perfectly rigid material). Therefore, e ,0 and general equation (1)
GH
can be simpli"ed to
Copyright 2000 John Wiley & Sons, Ltd. Int. J. Numer. Anal. Meth. Geomech., 2000; 24:723}735
ANALYSIS OF STRUCTURED SOILS 725
results in the di!erence in observed soil responses. Within the DSC theory the di!erence in soil
structure between a natural soil and its corresponding reconstituted state is therefore de"ned as
disturbance. Consequently, the disturbance function thus de"ned is a direct measure of the e!ects
of soil structure.
It may be seen from Equation (2) that under the two reference states thus de"ned all the
components of the observed strain for a structured soil are related to those of the strain for the
corresponding reconstituted soil. In the study of the disturbance function for structured soil, it is
found that the disturbance function can be investigated relatively easily if an incremental form of
the soil response is used. Therefore, the following manipulation is made. By setting
*De
DM e"De# e (3)
*e GH
GH
it is possible to express the increment of the observed strain as
de "D e de (4)
GH GH
This is a particular form of the more general equation (1), resulting from the assumption that the
relative intact state corresponds to a rigid material. As will be demonstrated, it is found to be very
convenient for the description of structured soils.
The in#uence of soil structure generally results in both isotropic and anisotropic variation of
mechanical properties. It is therefore reasonable to propose that the in#uence of soil structure on
the volumetric deformation and the distortional deformation may be described separately.
Consequently, from Equation (4) we may infer
de "D e de (5)
de "D e de (6)
where DM e and DM e are the separate disturbance functions for the volumetric deformation and the
distortional deformation. The volumetric deformation e and the distortional deformation e are
expressed as
e "e #e #e (7)
and
(2
e" ([(e !e )#(e !e )#(e !e )#6(e #e #e )] (8)
3
In this article, only the compression behaviour of structured soils is of concern. Therefore,
constitutive equation (5) will be examined in detail.
follows that a rational selection of the zero-strain condition for the fully adjusted state should also
ensure that the total FA strain consequently de"ned is an indicator of soil structure, and varies
uniquely with the structure. It would seem reasonable to suggest that there is no destructuring for
pure elastic deformation.
Let p represent the mean e!ective stress at which virgin yielding occurs. If it is assumed that
the compression behaviour of structured clays before virgin yielding, i.e. for p(p , is elastic,
then the strain associated with loading with p(p should not have any in#uence on the value of
the disturbance function. Consequently, it is suggested that the initial zero strain state for e is
selected at the initial virgin yielding point, i.e. p"p . For loading before yielding, i.e. p(p ,
there is no change in soil structure, and so for this condition e"0 is de"ned. Therefore
0 for p)p
e" NY (9)
de for p'p
NY
It should be noted that the selected initial strain state is valid only under the assumption that soil
behaves purely elastically for loading with p(p .
p before initial yielding
p " (10)
p after the occurrence of initial yielding
The current yield stress for a natural soil is, therefore, dependent on the initial soil structure as
well as the stress history. Hence, the compression behaviour of natural clays is assumed to be
elastic for p(p and elastoplastic for p*p .
It is also assumed that soil structure has negligible in#uence on the elastic deformation of the
material. Hence, the elastic deformation for structured soils may be expressed as a function of
intrinsic soil properties (i.e. those independent of soil structure), represented here by the general
parameter m*, the current mean e!ective stress p, and dp, i.e.
de "f (m*, p) dp for p(p (11)
Copyright 2000 John Wiley & Sons, Ltd. Int. J. Numer. Anal. Meth. Geomech., 2000; 24:723}735
ANALYSIS OF STRUCTURED SOILS 727
The key step in building a DSC model lies in "nding a disturbance function. Based on a study of
a large body of experimental data on the compression behaviour of structured clays and other
soils (e.g. [1,8,10]), the following new disturbance function for the compression behaviour of
naturally structured soils is proposed:
p
D e "1#b (12)
p
f (m*, p) dp for p(p
de " p (13)
1#b de for p*p
p
where de is the strain increment for the corresponding reconstituted soil under the same stress
change.
If the compression behaviour of the corresponding reconstituted soil can be assumed to be
linear in the e}ln p space, as suggested in the Critical State Soil Mechanics (e.g. [11]), Equation
(13) can be written as
i* dp
for p(p
1#e p
de " (14)
j* p dp
1#b for p*p
1#e p p
where j* and i* are intrinsic soil properties.
f (m*, p ) dp for p (p
de "
p (15)
1#b de for p *p
p
where p is the vertical e!ective stress, p is the initial vertical yield stress associated with initial
soil structure, p is the current vertical yield stress, and b is the disturbance index corresponding
Copyright 2000 John Wiley & Sons, Ltd. Int. J. Numer. Anal. Meth. Geomech., 2000; 24:723}735
728 M. D. LIU E¹ A¸.
to the one-dimensional compression conditions. The de"nition of the vertical yield stress p is
similar to p , which varies with stress history.
The advantage of this substitution is that the behaviour of a structured soil during one-
dimensional compression can be modelled in terms of the vertical e!ective stress only, thus the
di$culty in determining the values of the horizontal e!ective stresses is avoided.
Because the behaviour for many reconstituted soils can be assumed as linear in the e}ln p
space for a broad range of stresses, Equation (15) can be further simpli"ed as
i* dp
for p (p
1#e p
de " (16)
j* p dp
1#b for p *p
1#e p p
where j* and i* are intrinsic soil properties for one-dimensional compression.
4. EXPERIMENTAL EVALUATION
4.1. Background
The behaviour of seven di!erent soils is considered and the results of a total of ten tests on
structured soil are reported here. The sources of the data and details of the compression tests are
listed in Tables I and II. Because the proposed model describes the variation of the volumetric
strain with stress, comparisons between the simulations and experimental data are presented in
the form of the volumetric strain and stress relationships, which have been calculated from the
original data. The values of the initial yield stress, p , have been identi"ed directly from the
original compression curves plotted in the e}ln p co-ordinates. The method used to determine p
is shown schematically in the inset on Figure 2.
Copyright 2000 John Wiley & Sons, Ltd. Int. J. Numer. Anal. Meth. Geomech., 2000; 24:723}735
ANALYSIS OF STRUCTURED SOILS 729
Figure 2. One-dimensional compression tests on an arti"cially bonded soil (test data after Maccarini [12]).
The value of parameter b was determined by "nding the best match between the simulations
using Equation (13) or (14) and the test data for virgin compression. Parameters j* and i* were
determined from data on reconstituted soil according to their de"nitions. The methods
for determining parameters for Equations (15) and (16) are similar to those for Equations (13)
and (14).
The behaviour for both the structured soil and the reconstituted soil has been simulated for
cases where the behaviour of the reconstituted material is linear in the e}ln p or e}ln p space. For
other type of soils, only the behaviour of the structured soil was simulated, and the strain at the
fully adjusted state, i.e. e, was obtained from the experimental data on the behaviour of the
corresponding reconstituted soil. The stress units adopted here are kPa. The model simulations
for the behaviour of structured soils are represented by solid lines, and those for reconstituted
soils by the broken lines. Experimental data for structured soils are represented by solid circles or
Copyright 2000 John Wiley & Sons, Ltd. Int. J. Numer. Anal. Meth. Geomech., 2000; 24:723}735
730
7 Leda clay Quignley and 185 4.1 0.02 0.16 e}ln p linear
Figure 8 Thomson [16]
8a, 8b Guangshen Wang and 35.5 0.83 * * Di!. degrees of disturbance,
Figure 9 clay Wei [17] 30 0.45 e}ln p non-linear
Figure 3. One-dimensional compression tests on Mexico City clay (test data after Terzaghi [13]).
Figure 4. One-dimensional compression tests on Mexico City clay (test data after Terzaghi [13]).
squares, and those for reconstituted soils by open circles. For some simulations, the
selected initial yield point may also be detected on the plots from the sharp change in the
theoretical curve.
4.2. Evaluation
For the "rst group of tests, the mean e!ective stress is known. Therefore, Equations (13) and (14)
have been used in the simulations. The values of soil parameters identi"ed from the test data are
listed in Table I. Experimental data and theoretical curves are shown in Figures 2}4. Test 1 is
a one-dimensional compression test on an arti"cially and weakly bonded mixture of sand and
kaolin. The behaviour of the reconstituted soil is essentially linear. Tests 2 and 3 are two
one-dimensional compression tests on Mexico City clay, which has an extremely high initial voids
ratio, as high as 14. The mineralogy of the two samples is not the same, and the two samples are
considered as two di!erent clays. The behaviour of the reconstituted soils is non-linear. For the
second group of tests, the mean e!ective stress is unknown, therefore, Equations (15) and (16)
were employed in the simulations. The values of soil parameters identi"ed are listed in Table II.
Five types of soil are included and the behaviour of the corresponding reconstituted soils in this
group includes both linear behaviour and non-linear behaviour. Comparisons of the test data and
the theoretical curves are shown in Figures 5}9.
Copyright 2000 John Wiley & Sons, Ltd. Int. J. Numer. Anal. Meth. Geomech., 2000; 24:723}735
732 M. D. LIU E¹ A¸.
Figure 5. One-dimensional compression tests on Bothkennar clay (test data after Smith et al. [10]).
Figure 6. One-dimensional compression tests on Nagasaka clay (test data after Murakami [14]).
Figure 7. One-dimensional compression tests on Winnipeg clay (test data after Graham et al. [15])
The development of soil structure has been clearly demonstrated in the test on Nagasaka clay
(Figure 6). A reconstituted sample was "rst consolidated to a vertical e!ective stress of 158 kPa.
This compression pressure was then held constant for seven days. Subsequently, the sample was
subjected to further compression under increasing vertical e!ective stress. The results during the
Copyright 2000 John Wiley & Sons, Ltd. Int. J. Numer. Anal. Meth. Geomech., 2000; 24:723}735
ANALYSIS OF STRUCTURED SOILS 733
Figure 8. One-dimensional compression tests on Leda clay (test data after Quigley and Thompson [16]).
Figure 9. One-dimensional compression tests on Guang-shen clay (test data after Wang and Wei [17]).
subsequent compression loading are typical of those for a soil that has developed a structure. In
fact, the subsequent compression response indicates that the soil behaves as if it had been
previously loaded to a stress state with p "225 kPa, not the actual value of p "158 kPa.
Hence, during the 7 day period of constant e!ective stress, the reconstituted soil apparently
developed a structure producing a subsequent yield point which is larger than that produced by
its stress history alone. The experimental data and simulation of the behaviour of the structured
Nagasaka clay during one-dimensional compression from 158 to 410 kPa are shown in Figure 6.
Two samples of natural Bothkennar clay at the same location were tested (Figure 5). One was
obtained at a depth of 6.2 m below the surface soil by the Sherbrooke sampling method, and the
other at a depth of 5.3 m by the Laval method. It is generally accepted that the degree of
disturbance for the two methods is di!erent [10]. The values of b are found to be very di!erent
for the two samples. Two samples of similar Guang-shen clay were tested (Figure 9). The sample
for Test 8a had undergone less disturbance than that for Test 8b. It appears that b is sensitive to
the original sampling disturbance, as might be expected.
Overall, it is seen from the comparisons between the simulations and the test data in Figures
2}9 that the proposed DSC model describes very well the compression behaviour of these
Copyright 2000 John Wiley & Sons, Ltd. Int. J. Numer. Anal. Meth. Geomech., 2000; 24:723}735
734 M. D. LIU E¹ A¸.
structured soils under virgin yielding, swelling and recompression. However, it should be pointed
out that further work on the disturbance index, b or b , particularly its variation with sample
disturbance, is needed to enable the proposed model to serve as a predictive tool for engineering
calculations.
5. CONCLUSION
A DSC compression model for structured soils is proposed, and this is suitable to describe the
compression behaviour of structured soils under virgin yielding, swelling and reloading. Simula-
tions have been made using the proposed DSC model over a range of mean e!ective stresses from
10 to 2000 kPa. For one soil (Mexico City clay) the magnitude of volumetric strain reached was as
high as 120 per cent. It was found that the proposed DSC model can describe successfully the
compression behaviour of both naturally structured and arti"cially structured soils, independent
of whether the behaviour of the corresponding reconstituted soil is linear or non-linear in the
e}ln p space. The proposed models describe the compression behaviour of structured soils under
general loading conditions. The authors plan to formulate a complete three-dimensional model in
the future, by quantifying the shearing response of structured soil using the DSC.
ACKNOWLEGEMENTS
This work is supported by a grant from The Australian Research Council. The authors also wish to thank
the anonymous reviewers for making several useful suggestions, whose incorporation in this paper has
resulted in its signi"cant improvement.
REFERENCES
1. Burland JB. On the compressibility and shear strength of natural soils GeH otechnique 1990; 40(3): 329}378.
2. Liu MD, Carter JP. On the volumetric deformation of reconstituted soils. Research Report No. R765, Sydney
University, also accepted for publication in International Journal for Numerical and Analytical Method in
Geomechanics, 1998.
3. Desai CS. A consistent "nite element technique for work-softening behaviour. In Proceedings of the International
Conference on Computation Methods in Nonlinear Mechanics, Austin, Oden et al. (eds), 1974.
4. Desai CS. Constitutive modelling using the disturbed state as microstructure self-adjustment concept. In Continuum
Models for Materials with Microstructure, Muhlhaus (ed.) 1995; 239}296.
5. Desai CS. Mechanics of Materials and Interfaces: the Disturbed State Concept, 1991, under publication.
6. Desai CS, Toth J. Disturbed state constitutive modelling based on stress}strain and non-destructive behaviour.
International Journal of Solids and Structures, 1996; 33(11); 1619}1650.
7. Mitchell JK. Practical problems from surprising soil behaviour. Journal of Geotechnical Engineering, ASCE 1986;
112(3):257}289.
8. Leroueil S, Vaughan PR. The general and congruent e!ects of structure in natural soils and weak rocks. GeH otechnique
1990; 40(3):467}488.
9. Liu MD, Carter JP. Virgin compression of structured soils. GeH otechnique, 1999; 49(1):43}57.
10. Smith PR, Jardine RJ, Hight DW. On the yielding of Bothkennar clay. GeH otechnique 1992; 42(2):257}274.
11. Wood DM. Soil Behaviour and Critical State Soil Mechanics. Cambridge University Press: Cambridge, 1990.
12. Maccarini M. ¸aboratory investigations on arti,cial soil Ph.D. ¹hesis, University of London, 1987.
13. Terzaghi K. Fifty years of subsoil exploration. Proceedings of the 3rd International Conference on Soil Mechanics and
Foundation Engineering, 1953; 227}274.
14. Murakami Y. Excess pore-water pressure and preconsolidation e!ect development in normally consolidated clays of
some age. Soil and Foundation, 1979; 19(4):17}29.
Copyright 2000 John Wiley & Sons, Ltd. Int. J. Numer. Anal. Meth. Geomech., 2000; 24:723}735
ANALYSIS OF STRUCTURED SOILS 735
15. Graham J, Li CC. Comparison of natural and remoulded plastic clay. Journal of Geotechnical Engineering, ASCE
1985; 111(7):865}881.
16. Quigley RM, Thompson CD. The fabric of anisotropic consolidated sensitive marine clay. Canadian Geotechnical
Journal Vol. 3:61}73.
17. Wang N, Wei R. Evaluation of sample quality of soft clay. Proceedings of the Second International Conference on Soft
Soil Engineering Nanjing, 1996; 120}125.
Copyright 2000 John Wiley & Sons, Ltd. Int. J. Numer. Anal. Meth. Geomech., 2000; 24:723}735