The use or direct shear testing to provide data on jointed rock mass behavior to applied loads has been limited. Unlike other methods, such as triaxial and UCS testing, the location of the shear plane can be controlled relative to the joint orientations. allowing a closer examination of the effects of anisotropy as well as other parameters. However, little direct shear testing has been carried out on jointed, soft rock masses. This may be due in part to the difficulty of obtaining suitable samples for testing and the more complex stress conditions arising from such testing. A study currently in progress at Monash University is using direct shear testing to provide data for the development or an analytically based criterion. This criterion is based on the hypothesis that the critical surface along which shear failure occurs can be considered as a macro-joint between the upper and lower shear boxes. This paper presents some preliminary results of direct shear testing (relatively) large-scale samples of soft, jointed rock masses made from a synthetic soft rock. The samples have been sheared under constant normal stiffness (CNS) conditions, and the strength and dilation of the jointed rock mass sample recorded. The measured strength and dilation are compared with the anticipated strength of an intact rock, allowing the influence of the jointing to be assessed.
Current practices for estimating the response of jointed soft rock masses are predominantly empirically based, although more numerical modelling is being undertaken. The development of analytical criteria to predict the response ofjointed rock masses to loads appears to have stalled in recent years. The approach of the empirical criteria is to treat the rock mass as a "black box". i.e. the interaction between the intact rock and the joints within a jointed rock mass is overlooked in favour of an outcome based result. Currently available numerical techniques, such as discrete and finite element methods, do not easily or accurately model a jointed rock mass that is failing both as a result of block rotation and of crack initiation and propagation. Previous laboratory studies have mostly concentrated on triaxial or uniaxial compressive tests on plaster based samples to obtain data on jointed rock mass response to loading. These tests may be expedient, but they oversimplify the complex failure mechanisms withina rock mass, as well as using a material that may not sufficiently satisfy similitude criteria. This paper presents preliminary results of direct shear tests on relatively large jointed rock mass samples made from a synthetic soft rock known as "Johnstone". These preliminary results are compared to the behavior or a very rough rock joint.
Naturally occurring rock masses possess an inherent variability in their properties. When used in laboratory testing, this variability can result in scatter that can mask the effects of changes to parameters. This can be overcome by using laboratory prepared materials that have consistent properties.
