Downloaded from http://qjegh.lyellcollection.org/ at UQ Library on June 11, 2014
Quarterly Journal of Engineering Geology and Hydrogeology
Prediction of swelling characteristics of remoulded and compacted expansive soils using free swell index A.S. Rao, B.R. Phanikumar and R.S. Sharma Quarterly Journal of Engineering Geology and Hydrogeology 2004, v.37; p217-226. doi: 10.1144/1470-9236/03-052
Email alerting service
click here to receive free e-mail alerts when new articles cite this article
Permission request
click here to seek permission to re-use all or part of this article
Subscribe
click here to subscribe to Quarterly Journal of Engineering Geology and Hydrogeology or the Lyell Collection
Notes
© The Geological Society of London 2014
Downloaded from http://qjegh.lyellcollection.org/ at UQ Library on June 11, 2014
Prediction of swelling characteristics of remoulded and compacted expansive soils using free swell index A.S. Rao1, B.R. Phanikumar2 & R.S. Sharma2 1
Department of Civil Engineering, JNTU College of Engineering, Kakinada 533 003, India Department of Civil, Construction and Environmental Engineering, Iowa State University, Ames, IA 50011-3232, USA (e-mail:
[email protected])
2
Abstract easonal changes in moisture content result in volume change in expansive soils, which may damage structures founded on them. Evaluation of swelling characteristics of expansive soils, namely, swell potential and swelling pressure, is important for the design of foundations. Many relationships have been suggested for prediction of swell potential and swelling pressure from various index properties such as liquid limit, plasticity index, shrinkage index, activity, clay content, etc., and placement conditions such as initial dry unit weight, initial water content and initial surcharge pressure. Free swell index (FSI) indicates the potential expansiveness of a soil. FSI, being determined on the soil fraction <425 µm sieve like the other index properties of clays, is also an index property of an expansive soil. Hence, it can be used as a parameter in the relationships for swell potential and swelling pressure. This paper proposes relationships for predicting swell potential and swelling pressure of remoulded and compacted expansive soils using FSI and placement conditions. The relationships are based on experimental data for soil samples from 10 different sources.
S
Keywords: expansive soil, swell potential, swelling pressure, free swell index, placement conditions
Changes in moisture content result in volume changes in expansive soils. Expansive soils swell during the wet season by absorbing water and shrink during the dry season as a result of loss of water by evaporation and transpiration. Furthermore, upon wetting during a monsoon, expansive soils can exhibit swell or reduction in volume depending upon the stress and suction history of the soil (e.g. Sharma 1998; Sharma & Wheeler 2000; Gallipoli et al. 2003; Wheeler et al. 2003). The alternating swelling and shrinkage may damage civil engineering structures such as walls, light structures, pavements and canals founded on such soils. Hence, successful design and construction of foundations on expansive soils requires a good understanding of the swelling characteristics. Expansive clays are identified directly by the measurement of swelling characteristics, or indirectly from index properties and clay mineralogy. Quarterly Journal of Engineering Geology and Hydrogeology, 37, 217–226
Relationship of swelling characteristics with index properties and placement conditions In previous research Seed et al. (1962), Ranganatham & Satyanarayana (1965) and Bandyopadhyay (1981) proposed relationships for swell potential of remoulded natural expansive soils using index properties such as plasticity index (numerical difference between liquid and plastic limits), clay content, activity and shrinkage index (difference between liquid and shrinkage limits), which have only an indirect bearing on the degree of swelling. Activity (Skempton 1953) is defined as the ratio of plasticity index to the percentage of the soil fraction <2 µm. The definition of activity was later modified (Seed et al. 1962) as activity =
plasticity index % clay fraction<2 µm 5
.
(1)
Chen (1975) proposed a relationship for swell potential of undisturbed natural expansive soils in terms of plasticity index. Swell potential and swelling pressure depend not only on index properties and clay content but also on placement conditions such as initial dry unit weight, initial water content and initial surcharge pressure. The higher the initial dry unit weight, the greater will be the swell potential and swelling pressure. Both swell potential and swelling pressure decrease with increasing initial water content. Increase in surcharge pressure obviously reduces the amount of swell potential. However, various opinions prevail about the effect of surcharge pressure on swelling pressure. According to Satyanarayana (1966), swelling pressure is dependent on initial surcharge, but Chen (1973) observed that it is independent of initial surcharge. Having a direct influence on the values of swell potential and swelling pressure, the placement conditions are important parameters for predicting swelling characteristics. Swell potential (S) of a soil is defined as the ratio of increase in thickness (H) to the original thickness (H) of the soil sample compacted at optimum moisture content in a consolidation ring and soaked under a token surcharge of 6.9 kPa (Seed et al. 1962). It is expressed as S = (H/H) 100.
(2)
1470-9236/04 $15.00 2004 Geological Society of London
Downloaded from http://qjegh.lyellcollection.org/ at UQ Library on June 11, 2014
218
RAO ET AL.
Table 1. Existing relationships for swelling characteristics. Quantity Swell potential, S (%) Swell potential, S (%)
Source
Parameters used
Equation
Vijayvergiya & Ghazzaly (1973) Nayak & Christensen (1974)
Liquid limit (%), wL Dry unit weight (lb ft3), d Plasticity index (%), IP Initial water content (%), wi Clay content Liquid limit (%), wL Dry unit weight (kg m3), d Initial water content (%), wi Liquid limit (%), wL
log S = 1/19.5(d + 0.65wL 130.5)
Swelling pressure (kg cm2), ps
Komornik & David (1969)
Swelling pressure (tons ft2), ps
Vijayvergiya & Ghazzaly (1973)
Swelling pressure (lb inch2), ps
Nayak & Christensen (1974)
Dry unit weight (lb ft3), d Plasticity index (%), IP
S = 2.3 102(IP)1.45C/wi + 6.4 log ps = 2.1 + 0.021wL + 0.00067d 0.027wi log ps = 1/19.5(d 0.65wL 139.5) ps = 3.6 102(IP)1.12C2/wi2 + 3.8
Initial water content (%), wi Clay content
Swelling pressure (ps), as determined by the free swell method, is defined as the pressure that is required to recompress a completely swollen soil sample to its original unloaded volume or volume under a small token surcharge of about 10 kPa (Jennings 1963; ASTM 1996, D4546A). It is obtained from e–log p curve as the pressure corresponding to the initial void ratio. Some researchers (Komornik & David 1969; Vijayvergiya & Ghazzaly 1973; Nayak & Christensen 1974), on the basis of experimental data, proposed relationships for swell potential and swelling pressure involving both placement conditions and index properties. Some of these relationships are given in Table 1.
and compacted expansive soils in terms of FSI and placement conditions.
Experimental investigation To propose relationships for swelling characteristics with FSI and placement conditions, an experimental investigation was carried out on 10 remoulded expansive soil samples collected from 10 districts of the state of Andhra Pradesh, India. All the soils were black cotton soils. Table 2 shows the index properties of the soils tested.
Quantities determined and variables studied Importance of free swell index (FSI) Free swell index (FSI) can be considered as an index property of an expansive soil and it reflects the potential for expansion of the soil. The FSI test is carried out on oven-dried soil passing a 425 µm sieve. It is defined as the ratio of the difference in volumes of soil in water and kerosene to the volume of soil in kerosene (Holtz & Gibbs 1956), expressed as
FSI =
[(final volume of soil in water final volume of soil in kerosene) 100] final volume of soil in kerosene
Swell potential and swelling pressure were determined by the swell-consolidation method, which is a free inundation method, at the following placement conditions: (i) initial water content, wi (%): 0, 5, 10, 15 and 20; (ii) initial dry unit weight, di (kN m3): 10, 12, 14, 16 and 18; (iii) initial surcharge, qi (kPa): 5, 50, 100, 150 and 200. In total 1250 swell-consolidation tests were performed.
.
(3)
Soils having FSI >200, FSI between 100 and 200, FSI between 50 and 100 and FSI <50 are respectively characterized as having a degree of expansion described as ‘very high’, ‘high’, ‘medium’ and ‘low’ (Mohan, 1977). It is considered here that swelling characteristics could be predicted with the FSI as a parameter instead of other index properties. This paper proposes relationships for swell potential and swelling pressure of remoulded
Sample preparation and test procedure The clay sample was air-dried and pulverized to pass through a 4.75 mm sieve. It was then oven-dried at a constant temperature of 105( C to reduce the water content to 0% and to obtain the maximum value of swell potential. The sample was allowed to cool to room temperature. The required weight of the oven-dried sample was mixed with the required amount of distilled
Downloaded from http://qjegh.lyellcollection.org/ at UQ Library on June 11, 2014
SWELLING OF EXPANSIVE SOILS
219
Table 2. Index properties of the soils investigated. Soil source
S1
S2
S3
S4
S5
S6
S7
S8
S9
Particle density (ASTM D854-02) Gravel (%) (ASTM, 98 D422-63) Sand (%) (ASTM, 98 D422-63) Silt (%) (ASTM, 98 D422-63) Clay (%) (ASTM, 98 D422-63) Liquid limit (ASTM D4318-00) Plastic limit (ASTM D4318-00) Plasticity index (ASTM D4318-00) Shrinkage limit (ASTM D4318-00) USCS Classification (ASTM D2487-00) Free swell index (ASTM D5890-02) Soil fraction <425 µm (%) (ASTM, 98 D422-63)
2.75
2.70
2.73
2.75
2.68
2.76
2.71
2.80
2.70
2.72
0
0
0
0
0
0
0
0
0
0
2
7
14
1
14
11
8
0
34
16
13
42
44
39.5
35
43
36
12
20
48
85
51
42
59.5
51
46
56
88
46
36
102
93
79
74.5
61.4
69
78.5
20
18
26
22
19.4
22
82
75
53
52.5
42
47
10
9
11
11.63
12
9.2
S10
121.5
60.5
148
25
23.5
24.5
17
53.5
98
36
131
8
11
12
16
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
200
170
190
173
120
110
165
161
140
254
49
49
50
50
48
49
49
49
44
43
Soil sources: S1, Waddilanka; S2, Kesawaram; S3, Amalapuram; S4, Bhimavaram; S5, Tanuku; S6, Munnaluru; S7, Guntur; S8, Ongole; S9, Wyra; S10, Warangal.
water (based on the weight of the oven-dried soil) and allowed to equilibrate for 6 h. After equlibration, the sample was statically compacted in four layers, each of 5 mm thickness, in a consolidation ring of thickness 20 mm and diameter 60 mm, to the desired placement water content and dry unit weight. A thin layer of silicon grease was applied to the walls of oedometer to reduce friction between the sample and the oedometer wall. A filter paper and a porous stone were placed at each end of the sample. This unit was placed in the oedometer and the loading pad positioned centrally on the top porous stone. The required initial surcharge pressure was applied on the specimen after setting the dial gauge reading (initial reading) to zero. Swell potential and swelling pressure were determined by the free inundation method, in which the sample is completely inundated with water and allowed to swell freely under the applied surcharge. Dial gauge readings were taken up to equilibrium swell (no further change in the dial gauge reading). The swelling process took 3–4 days depending upon the soil. The increase in the thickness (H) of the sample was noted after equilibrium swell. No material was lost over the sides of the oedometer ring. The sample was consolidated under increased applied pressures until the dial gauge reading was less than the initial reading. Swell potential (Seed et al. 1962) was determined as S = [(H/H) 100], and swelling pressure from the e–log p curve as the pressure corresponding to the initial void ratio (Jennings 1963).
Results and discussion Effect of placement conditions (initial surcharge, initial dry unit weight and initial water content) on swell potential Figure 1 is a typical example of e–log p curves for the soil (source S1) compacted at an initial water content (wi) of 10% and at a dry unit weight (di) of 16 kN m3 and subjected to various initial surcharge pressures (qi). As the initial surcharge increased, swell potential decreased as indicated by the decreasing void ratio. However, the swelling pressure was not influenced by the initial surcharge. This is in accordance with the observations of Chen (1988). The swelling pressure was about the same for all the samples (Fig. 1) irrespective of the initial surcharge, because swelling pressure is the pressure required to bring back a fully swollen soil sample to its initial void ratio (as in the free inundation method). Therefore, as long as the initial surcharge is less than the swelling pressure, there is bound to be some swelling and the total amount of pressure required to bring back the swollen sample to the initial volume (or void ratio) remains unchanged. Figures 2–11 are typical summary plots for all the soils showing the variation of swell potential with initial dry unit weight for various initial water contents. The plots are shown only for an initial surcharge pressure of 5 kPa, as a similar trend was observed for the other
Downloaded from http://qjegh.lyellcollection.org/ at UQ Library on June 11, 2014
220
RAO ET AL.
Fig. 3. Effect of initial dry unit weight on swell potential.
Fig. 1. Effect of surcharge on swell potential and swelling pressure.
Fig. 4. Effect of initial dry unit weight on swell potential.
Fig. 2. Influence of initial dry unit weight on swell potential.
surcharge pressures. For a given water content, swell potential increased with increasing dry unit weight. As dry unit weight increases, the closer packing increases the number of clay particles and their interactions, resulting in greater potential swelling. The plots also show that swell potential decreased with increasing water content for a given initial dry unit weight. With an increase in the initial water content, the initial degree of saturation also increases for a given dry unit weight (or void ratio). As the degree of saturation approaches 100%, the amount of uptake of water by the soil will be less, and hence a lower swell potential results. Figures 12–21 summarize the effect of the initial water content on swelling pressure for various dry unit weights for all the soils. Swelling pressure decreased with increasing water content. As the amount of swelling decreased with increasing water content, the pressure required to nullify the amount of swelling would also
Fig. 5. Effect of dry unit weight on swell potential.
decrease. Figure 22 shows the variation of swell potential and swelling pressure, determined at the placement conditions of di = 16 kN m3, wi = 10% and qi = 5 kPa, for all the soils with their respective FSI. Both swell potential and swelling pressure increased with FSI, indicating that FSI is an important parameter for expansive soils.
Downloaded from http://qjegh.lyellcollection.org/ at UQ Library on June 11, 2014
SWELLING OF EXPANSIVE SOILS
221
Fig. 6. Influence of initial dry unit weight on swell potential.
Fig. 9. Influence of initial dry unit weight on swell potential.
Fig. 7. Variation of swell potential with initial dry unit weight.
Fig. 10. Effect of initial dry unit weight on swell potential.
ment conditions and FSI by performing multiple linear regression analysis on the entire experimental dataset, with the following results: S% = a1di b1wi c1qi + d1(FSI) K1
(4)
logps = a2di b2wi + d2(FSI) K2.
(5)
and
Fig. 8. Effect of initial dry unit weight on swell potential.
Relationships developed from the experimental data Relationships for swell potential (S%) and swelling pressure (ps) have been developed in terms of the place-
The constants or the regression coefficients for the above equations are a1 = 4.24,, b1 = 0.47,, c1 = 0.14,, d1 = 0.06 and K1 = 55; and a2 = 0.30,, b2 = 0.02, d2 = 0.005 and K2 = 3. Because the initial surcharge pressure (qi) had no effect on the swelling pressure, as shown in Figure 1, it was not included in the analysis. The standard errors for the regression coefficients a1, b1, c1 and d1 are 9.69, 3.22, 1.76 and 1.15, respectively. The standard deviation is 4.69. Similarly, the standard errors for the regression coefficients a2, b2 and d2 are 4.82, 1.24 and 0.62 respectively. The standard deviation is 8.77. This means that most of the values of the
Downloaded from http://qjegh.lyellcollection.org/ at UQ Library on June 11, 2014
222
RAO ET AL.
Fig. 13. Influence of initial water content on swelling pressure.
Fig. 11. Variation of swell potential with initial dry unit weight.
Fig. 14. Effect of initial water content on swelling pressure.
Fig. 12. Effect of initial water content on swelling pressure.
predicted swell potential and swelling pressure do not show much deviation from the mean of the measured values. Hence, the linear model proposed for the prediction of swell potential and swelling pressure from FSI and placement conditions is a successful model. The statistical data obtained for swell potential and swelling pressure are summarized in Tables 3 and 4.
Validation of the relationships developed Figure 23 shows the measured values of swell potential at water contents w of 5% and 10% for all the soils tested in comparison with those predicted from the equation proposed by the authors, as shown in series 1 and 3, and by Nayak & Christensen (1974; Table 1), as shown in series 2 and 4. Each series in the figure shows the
Fig. 15. Effect of initial water content on swelling pressure.
measured and predicted values for all 10 soils. The measured swell potential tallied closely with the swell potential predicted from the equation proposed in this study (equation (3)). For other water contents also the measured and predicted values tallied (not shown for want of space). However, the equation proposed by Nayak & Christensen (1974) shows significant difference between measured and predicted values of swell potential (series 2 and 4). The predicted values are much
Downloaded from http://qjegh.lyellcollection.org/ at UQ Library on June 11, 2014
SWELLING OF EXPANSIVE SOILS
223
Fig. 18. Influence of initial water content on swelling pressure. Fig. 16. Variation of swelling pressure with initial water content.
Fig. 19. Effect of initial water content on swelling pressure.
Fig. 17. Effect of initial water content on swelling pressure.
higher than the measured ones. The values predicted from the equation of Vijayvergiya & Ghazzaly (1973) are too high to be plotted alongside those shown in Figure 23. Figure 24 shows the measured values of swelling pressure at w of 10% for all the soils tested in comparison with those predicted from equation (3) of the present study (series 1) and from the equations of Nayak & Christensen (1974) and Komornik & David (1969) (series 2 and 3). The values predicted from the equation of Vijayvergiya & Ghazzaly (1973) are also too high to be plotted alongside those shown in Figure 24. The swelling pressure predicted from equation (4) in the present study showed a good relationship with that measured. The values predicted from the equation of Nayak & Christensen were much higher than the
measured values for all the soils. The swelling pressure predicted from the equation of Komornik & David (1969) gave a good relationship only for two soils, S2 and S6. Komornik & David (1969) did not consider the initial surcharge pressure, whereas Nayak & Christensen (1974) considered only initial water content and Vijayvergiya & Ghazzaly (1973) only dry unit weight among the placement conditions. Apart from this inadequacy of not considering all the placement conditions together, these expressions included plasticity index, clay content and liquid limit, which do not directly reflect the swelling nature of expansive soils. Hence, the existing equations did not give good relationship between the measured and predicted values of swelling characteristics. The equation proposed here, however, considers the placement conditions together and also includes FSI, which directly reflects the swelling characteristics. Hence, the predicted values are close to the measured ones. Figure 25 compares the measured and predicted values of swelling pressure for soils from other sources using the proposed equation. Measured values of swell
Downloaded from http://qjegh.lyellcollection.org/ at UQ Library on June 11, 2014
224
RAO ET AL. Table 3. Regression coefficients for swell potential. Regression coefficient for swell potential a1 = 4.24 b1 = 0.47 c1 = 0.14 d1 = 0.06
Standard error
9.69 3.22 1.76 1.15
Standard deviation
4.69
Table 4. Regression coefficients for swelling pressure. Regression coefficient for swelling pressure
Fig. 20. Effect of initial water content on swelling pressure.
a2 = 0.30 b2 = 0.02 d2 = 0.005
Standard error
Standard deviation
4.82 1.24 0.62
8.77
sources, the water content and dry unit weight at which swelling pressure was measured, the FSI values of the soils, and the measured and predicted values of swelling pressure. From the close agreement between the measured and predicted values of swelling pressure of soils from other sources also, it can be said that the proposed equations could be used for the prediction of swell potential and swelling pressure for a wide range of remoulded and compacted expansive soils all over the world.
Conclusions
Fig. 21. Effect of initial water content on swelling pressure.
Fig. 22. Variation of swell potential and swelling pressure with FSI.
potential from other sources could not be obtained by predicting with the new equation. The figure shows that the new equation predicts the swelling pressure values for the other soils very well. Table 5 summarizes the data
Experimental data obtained from 10 different expansive soil samples were used to develop relationships for swell potential and swelling pressure in terms of placement conditions and free swell index (FSI). The following conclusions are reached. (1) Swell potential decreased with increasing surcharge. However, the swelling pressure is not affected by the initial surcharge, consistent with previous research (Chen 1975). Both swell potential and swelling pressure increase with increase in the initial dry unit weight, and decrease with increase in the initial water content. (2) It is necessary to include in the relationships for swell potential and swelling pressure all the placement conditions and the free swell index (FSI), as these parameters directly affect the values of swell potential and swelling pressure. Other index properties, such as plasticity, are not directly related to the swelling nature and should be excluded. (3) Based on the analysis of the experimental data from the 10 soil samples, a general equation for swell potential (S) may be given in terms of the initial water content wi, the initial dry unit weight di, the initial surcharge qi and FSI, in the form S
(%) = a1di b1wi c1qi d1(FSI) K1.
Downloaded from http://qjegh.lyellcollection.org/ at UQ Library on June 11, 2014
SWELLING OF EXPANSIVE SOILS
Fig. 23. Swell potential values for w of 5% and 10% measured and predicted for all 10 soils tested.
Fig. 24. Swelling pressure values for w of 10% measured and predicted for all 10 soils tested.
225
Downloaded from http://qjegh.lyellcollection.org/ at UQ Library on June 11, 2014
226
RAO ET AL.
Table 5. Comparison of measured and predicted swelling pressure values for soils from other places (after Komornik & David 1969). Source
Hadera Ekron Ramat Hasharon Kiryat Gat Hedera Nesher MMD Kfar Jeruham
Water content (%)
Dry density (kN m3)
FSI (%)
Measured swelling pressure (kPa)
Predicted swelling pressure (kPa)
16 16 15.3 2.7 17.6 17 16.7 16.9
17 20 17.10 14.6 17.65 20 23.7 22.2
100 100 100 130 106 100 100 50
191 75 151 84 248 143 120 60
190 79 197 95 300 158 110 75
Fig. 25. Swelling pressure of soils from other sources predicted by the proposed equation.
(4) Similarly, a general equation for swelling pressure (ps) may be proposed in terms of the initial water content wi, the initial dry unit weight di and FSI, as logps = a2di b2wi + d2(FSI) K2. (5) The equations proposed for swell potential and swelling pressure gave a good relationship between the predicted and measured values for the soils tested and those for soils from other sources, whereas the existing equations did not. Hence, it is proposed that FSI should also be included in the relationships for swelling characteristics along with the three important placement conditions, namely, initial water content, initial dry unit weight and initial surcharge pressure, so that the relationships could be used for predicting swelling characteristics of a wide range of remoulded, oven-dried and compacted expansive soils all over the world.
References B, S.S. 1981. Prediction of swelling potential for natural soils. ASCE Journal of Geotechnical Engineering Division, 107 No. GT 1, 658–661. C, F.H. 1988. Foundations on Expansive Soils. Elsevier, Amsterdam. G, D., G, A., S, R. & V, J. 2003. An elasto-plastic model for unsaturated soil incorporating the
effects of suction and degree of saturation on mechanical behaviour. Géotechnique, 53 (1), 123–135. H, W.G. & G, H.J. 1956. Engineering properties of expansive clays. Transactions of the American Society of Civil Engineers, 121, 641–677. J, J.E. 1963. Discussion on ‘The heaving of buildings and associated economic consequences with particular reference to Orange Free State gold fields’. Civil Engineers, South Africa, 5, 122. K, J. & D, A. 1969. Prediction of swelling potential for compacted clays. Journal of the Soil Mechanics and Foundation Engineering Division, American Society of Civil Engineers, 95 (1), 209–225. M, D. 1977. Engineering of expansive soils. Inaugural Address, Proceedings of the 1st National Symposium on Expansive Soils. HBTI, Kanpur, India.. N, N.V. & C, R.W. 1974. Swelling characteristics of compacted expansive soils. Clays and Clay Minerals, 19 (4), 251–261. R, B.V. & S, B. 1965. A rational method of predicting swelling potential for compacted expansive clays. In: Proceedings, 6th International Conference on Soil Mechanics and Foundation Engineering, Canada, 1, 92–96. S, B. 1966. Swelling pressure and related mechanical properties of black cotton soils. PhD thesis, Indian Institute of Science, Bangalore. S, H.B., W, R.J. & L, R. 1962. Prediction of swelling potential for compacted clays. Journal of the Soil Mechanics & Foundation Engineering Division, American Society of Civil Engineers, Part I, Proceedings 3169, 88(SM3), 53–87. S, R. S. 1998. Mechanical behaviour of unsaturated highly expansive clays. DPhil thesis, University of Oxford. S, R.S. & W, S.J. 2000. Behaviour of an unsaturated highly expansive clay during wetting/drying cycles. In: R, H., T, D.G. & L, E.C. (eds) Proceedings of the International Conference on Unsaturated Soils (UNSAT-ASIA), Singapore. Balkema, Rotterdam, 721–726. S, A.W. 1953. The colloidal activity of clays. In: Proceedings, 3rd International Conference on Soil Mechanics and Foundation Engineering, Zurich, 1, 57–61. V, V.N. & G, O.I. 1973. Prediction of swelling potential for natural clays. In: Proceedings of the 3rd International Conference on Expansive Soils, Haifa, Israel, 1, 227–236. W, S.J., S, R.S. & B, M.S.R. 2003. Coupling of hydraulic hysteresis and stress–strain behaviour in unsaturated soils. Géotechnique, 53 (1), 41–54.
Received 10 October 2003; accepted 18 June 2004.
