Shear behaviour of rock discontinuities and soil-rock interfaces

Basic aspects of the shear behaviour of rock-rock and rock-soil interfaces are investigated by physical modelling. Shear characteristics of rock discontinuities are studied by conducting direct shear tests on identical models of various natural rough surfaces. A strong synthetic rock is developed and used to cast several sets of identical samples by using a rubber moulding material. All samples are subjected to direct shear testing in the same relative direction of shearing and under various normal stresses, over the full range of dilatant behaviour. The effect of scale on peak shear strength is studied by testing models of natural rock surfaces in full and divided into sub-samples of various sizes. Identical samples are reproduced by using two different casting materials of different mechanical properties and are tested under various normal stresses. The test results indicate that independent of rock material, roughness, scale and normal stress level, the peak shear strength can be interpreted as comprising two components, one purely :frictional and one geometrical. The controlling parameter of peak shear strength is the dilation angle, which for a particular joint type, reduces logarithmically with normal stress over four orders of magnitude and becomes zero at a normal stress which in most cases is about one order of magnitude lower than the unconfined compressive strength. Beyond this normal stress, the shear behaviour of the sample becomes purely frictional. A simple, theoretical but readily applicable criterion is proposed, for a realistic mechanism of shearing and based on the suggestion that peak shear strength comprises one frictional and one geometrical component. Changes in peak friction angle due to normal stress, sample size or roughness are interpreted in terms of change in the geometrical component only. The magnitude of the frictional component is determined from the shear strength of the rock wall material, whereas the magnitude of the geometrical component is predicted from consideration of surface roughness and normal contact theory. The results of a compilation of data concerning the friction angle of various types of fresh rock are provided. The advantage of the new criterion is its simplicity, the use of physically meaningful parameters and the ability to explain the observed behaviour of natural rock discontinuities. Verification of the new criterion is established by testing against published data. The behaviour of soil-rock interfaces is studied by direct shear tests on interfaces between granular materials and modelled rock surfaces of various roughness. Two ii different granular materials of various sizes and a diversity of surfaces with different roughness, are used to model the soil-rock interface. It is found that for a particular type of rock and soil material, the shear behaviour of the interface depends upon the rock surface roughness, the grain size and shape of the soil particles and the normal stress. Some aspects of double soil-rock interfaces are presented. Recommendations are made concerning the suitability of casting materials in reproducing natural rock surfaces, the expected friction angle for various types of fresh rock and the determination of in-situ shear strength both for rock-rock and rock-soil interfaces.