A Sample Preparation Method Enhancing Shear Strength for Remolded Unsaturated Soil
Remolded soil is usually substituted for undisturbed soil in shear tests to measure the strength of natural soil. However, the traditional remolding process of unsaturated soil may lead to a wide difference on the strength of soil and therefore fail to represent the strength properties of natural soil. Based on the traditional remolding methods, this paper introduces an experimental method which could enhance the shear strength of remolded soil significantly by readjusting the water distribution and restoring the matric suction between soil layers. Compared to the traditional compacted method, the improved method could effectively raise the cohesion of unsaturated soil, which enables the remolded soil fully to retain its strength and hence to be applied accurately to predict the strength of natural soil.
KeywordsShear strength Unsaturated soil Remolded soil Matric suction Cohesion
The shear strength of soil could be measured through laboratory tests or in-situ tests, while considering the large cost and the inconvenience for the research under different conditions, laboratory tests are more likely to be adopted. For sand and gravel, remolded soil is usually substituted for undisturbed soil since its strength could hold the value basically after the remolding process. The clay remolded by the traditional compacted method, by contrast, with finer granularity and complex structures, presents a vast strength difference from natural soil under non-saturation state (mostly lower than natural soil, according to several previous studies), which forces researchers to use undisturbed or saturated specimen for tests. However, the undisturbed soil is often hard to obtain and the divergence of the sample is also not a negligible issue. The stress-strain characteristics of the saturated soil are apparently unable to determine the properties of real natural unsaturated soil as well. As a result of the absence of reliable values, it is imperative to search for an experimental method which enables remolded unsaturated soil to fully retain its strength and to be applied to predicting the strength of natural soil.
The strength properties of soil are largely dependent on the particle structure, water content and mineral composition. Despite the disturbance of the structure caused by the remolding process, the shear tests’ results of remolded sand are still qualified to be the reference for the strength of natural sand or saturated clay respectively, which proved that the macro-structural changes are not the key reason to the wide difference between remolded soil and natural soil. The relation between water content and shear strength has been investigated by many previous studies. Fredlund et al. [1, 4] had performed a lot of measurements on natural clay and a power law was obtained, that is, the shear strength of clay could be consequently reduced with increasing water content.
But for clay remolded by compacted method, as the water content rises, the shear strength would initially increase and soon dropped after reaching its maximum. Lin and Sun et al. [2, 3] found that the inadequate water-soil interaction was the main factor led to the variation, or to put it another way, the matric suction of natural soil failed to be restored through the remolding process. Expanding upon this idea, the authors aimed at emphasizing the effect of the matric suction on soil strength and put forward an improved remolding method which could recover the suction and thus enhance the shear strength significantly during laboratory tests.
In the present paper, several triaxial-shearing tests are performed on silty clay originated from YiChang, HuBei Province. The experimental results of undisturbed soil, traditional remolded specimens as well as the improved specimens, all in unsaturated conditions, are compared to assess the feasibility of the new sample preparation method.
2 Materials and Methods
To analyse the effects of different remolding methods on soil strength, a series of shear tests were designed to compare the properties of undisturbed soil, compacted specimens and specimens prepared by the improved method.
All soil samples were collected from Three Gorges area, which were identified as silty clay. The sampling depth is between 6 to 8 m. The relative density is 2.76 g/cm3, the maximum natural saturation is around 66%, the optimum water content is measured as 13.7%.
The undisturbed soil was cut by a thin-walled split mould with the diameter of 39.1 mm and height of 80 mm. A total of 15 specimens were prepared and then divided in five groups. All specimens were dried by drying ovens to the required saturation of 25%, 35%, 45%, 55%, respectively.
The preparation of compacted specimens is strictly following The Standard for Soil Test Method . Disturbed soil was dried, ground and sieved, then mixed with water to make moist paste with saturation of 25%, 35%, 45%, 55%, 65%, 80%, respectively. After the distribution of moisture became even, the soil was compacted in four layers to obtain 6 groups of specimens in line with the dry density of undisturbed soil.
Similar to the previous preparation of traditional remolded soil, the disturbed soil was also dried and ground. To obtain the best compact effect, the dry soil was mixed into wet soil by 50% saturation degree and compacted into specified dimensions. The samples were initially saturated and then dried gradually into designated saturation. After the moisture run through the soil pores fully, 18 columns with varying saturation (35%, 45%, 55%, 65%, 80%) were sheared at a constant rate of 0.5 mm/min under the confining pressure ranged from 100 kPa to 300 kPa, the other 6 columns were also set as alternatives.
3 Results and Discussion
3.1 The Comparison of Shear Indexes
For remolded specimens, it can be easily observed that the cohesion of compacted soil decreases not before the saturation degree exceed 45%. This is caused by the traditional remoulding method that directly compact soil paste (dried soil and certain amount of water) in several layers. Shear failure can be easily caused by the weak suction between soil layers and particles when the water content is very low, especially at the interfaces. As the water content goes up, the suction force becomes stronger which contribute to an increasing cohesion and shear strength of the soil, but the strength in this period is still relatively low. After the saturation rises above a certain threshold (45% in this experiment), the bound water membrane becomes thicker, the cohesion of soil could be impaired by the decreasing effective stress, accounting for the subsequent down trend of the curve. However, the cohesion of the sample used new remolding method decreases linearly as the saturation of soil increases, which is consistent with the corresponding relationship of undisturbed soil. Such differences could be explained as follows.
Compared with the traditional sample making, the improved method achieved the best compacting results by making the sample at the optimum moisture content, which insures closer connections between soil particles. Through the process of saturation and slow drying, samples successfully keep its matric suction to the full extent by simulating the way that natural soil loses moisture, which means, a more evenly distribution of water in the columns. Therefore, the cohesion of the samples used improved method shows similar variation with undisturbed soil. The improved remolding method is proved to be able to enhance the shear strength of unsaturated clay effectively from the above analyses.
3.2 The Comparison of Shear Characters
Figure 4 presented the stress-strain curves for 25%, 35%, 55% saturation degree with two different remolding methods. It can be observed that the peak values of improved specimens are apparently higher than that of traditional specimens. Moreover, as the saturation increases, the stress-strain curve gradually transits from strain softening state to strain hardening state, which is similar to the shear properties of natural structural clay [6, 7]. But for traditional specimens, the stress-strain relation presents properties similar to light sand, that is, stress hardening. On the one hand, the differences in shear properties of the two kinds of specimens demonstrate that matric suction contributes in shear strength differently under different saturations, and also, proved that the shear strength of unsaturated soil could be enhanced by retaining its matric suction. On the other hand, although the peak value of the stress rises with greater matric suction, the stress would then drop rapidly after the maximum value with a relatively low residual strength, which could cause sudden failure especially for cohesive slope. Hence, considering the probability of such brittle failure, it is more secure for engineering projects to evaluate its safety by the improved method developed in this article when conducting soil tests.
4 Summary and Conclusions
From the perspective of recovering the matric suction of soil, this article presents an improved method to remold unsaturated clay through saturation and dehydration process. By comparing a series of shear tests conducted on different specimens, it is found that the cohesion changes of improved specimens under different saturation degrees accord with that of undisturbed soil, moreover, the strength of the two are very close. While the cohesion of specimens prepared by compacted method will increase firstly and then decrease as the saturation grows, the shear strength is also much lower than that of undisturbed soil. In addition, using the improved method could better simulate the shear properties of natural soil and take the sudden failure of dry clay into full consideration when evaluating the stability of realistic slopes. However, the stress-strain relation of compacted specimens is unrepresentative.
- 2.Sun, D., Gao, Y.: Water retention behavior of soils with different preparations. Chin. J. Geotech. Eng. 37, 91–97 (2015)Google Scholar
- 3.Lin, H., Li, G., Yu, Y.: Influence of matric suction on shear strength behavior of unsaturated soils. Rock Soil Mech. 28, 1931–1936 (2007)Google Scholar
- 5.GBT 50123-1999: The Standard for Soil Test Method. MOC (Ministry of Construction) (1999)Google Scholar
- 6.Delage, P., Yu, J.C.: Elasto-plastic behavior of a compacted silt. Int. J. Rock Mech. Mining Sci. Geomech. Abs. 32, 317–324 (1995)Google Scholar