Measurement of soil strength in simple shear tests: Discussion

Abstract

The use of the conventional simple shear apparatus for measuring the shear strength parameters of soils has some merits over many of the existing laboratory testing apparatuses. However, one of the drawbacks of the conventional simple shear apparatus is its inability to measure all the stress components, thus necessitating the need for various assumptions to construct the Mohr circle of stress (Wroth 1987) and to evaluate the shear strength in terms of principal stresses. In the above paper, it was assumed that the ratio of the horizontal stress (a,!,) to the vertical stress (a;) at the critical state increases with overconsolidation ratio (OCR)~ and that for heavily overconsolidated specimens, failure occurs well in the passive state, i.e., a,!, > a;. Based on these assumptions, it was deduced in the paper that, given that the critical state friction angle +,', = arcsin[(oi a;)/ a{ + a;)],, is independent of the consolidation history, the simple shear critical state stress ratio angle p,', = arctan(~;/o:),, increases with OCR. In the conventional simple shear apparatus, such assumptions and any subsequent deduction cannot be adequately evaluated. As part of a study into the behaviour of reconstituted kaolin (liquid limit (LL) = 84.2%, plasticity index (PI) = 43.6%, and G, = 2.65) in simple shear from very small strains to failure and the effect of consolidation history, we used an automated hollow cylinder torsional simple shear apparatus, which in addition to several other capabilities can consolidate specimens along various stress paths and also measure all stress components accurately during all stages of testing. In these undrained simple shear tests, all the incremental strains in the axial, circumferential and radial directions were maintained to be zero. For details of the apparatus, test procedure, and test results see Ampadu (1991) and Ampadu and Tatsuoka (1992). Some relevant extracts from this study are presented to discuss some of our findings with respect to the above-mentioned assumption. Table 1 summarizes the consolidation and undrained simple shear failure conditions of some of our results of tests on hollow cylinder specimens of initial dimensions as follows: 16 cm height, 10 cm outer diameter, and 6 cm inner diameter. The relationship between the effective axial stress a; and the effective circumferential stress a; for these tests is shown in Fig. la. In the hollow cylinder, the stresses a; and a; are equivalent to the vertical stress a; and the horizontal stress a,!,, respectively, referred to in the paper. The specimens were first KO-normally reconsolidated (test A201-1) and, for overconsolidated tests, later KO-rebounded along KO = OCR'.~ from about the same preconsolidation pressure a;,. The KO-rebound stress path is based on Mayne and Kulhawy (1982) using +' = 17. The variation of the stress ratio K = a;/u,' as shearing progresses is shown in Fig. 2. This figure and Fig. l a illustrate our observation that, despite their different initial values, the stress ratios K = a;/a; change rapidly, moving towards a common value close to unity but always on the active side, i.e., a; l), it rapidly crosses over from this state to the active stress state (K < 1) and remains in this state until failure. The implications of this behaviour in terms of the direction a of the major principal stress from the vertical are shown in Fig. 3. (In the hollow cylinder apparatus, a can be determined, since all the stress components are measured.) The figure shows that a changes rapidly from 0 for 'paper by J.H. Atkinson, W.H.W. Lau, and J. J.M. Powell. normally and lightly overconsolidated specimens and from 1991. Canadian Geotechnical Journal, 28: 255-262. 90 for heavily overconsolidated specimens towards a com20verconsolidation ratio OCR is given as R in the DaDer under mon value slightly less than 45. There is also the tendency * discussion. for a slight increase of a with OCR at large strain levels.