Size-dependent Behaviour of Weak Intact Rocks

Characterising the size-dependent behaviour of rock has been a significant challenge in rock engineering particularly during the design of structures on or within a rock mass. Generally, the mechanical characterisation of rock starts at the laboratory scale where the intact rock is tested and then the resulting data is extrapolated to the field conditions using a suitable size effect model. Despite extensive research on the size effect of intact rock, very few have included the weak rocks with low uniaxial compressive strength (UCS). Thus, in this study, the size-dependent behaviour of two different weak rocks namely, Gambier limestone and artificial rock with uniaxial compressive strength of less than 10 MPa were investigated experimentally and analytically. The diameter of the cylindrical Gambier limestone samples varied from 26 to 285 mm while the diameter of artificial rock samples ranged between 26 and 139 mm. For Gambier limestone, the uniaxial compressive, Brazilian and point load experiments were carried out while for artificial rock, only the uniaxial compressive tests were conducted. In both rock types, the ascending and then descending size effect trend was a pronounced behaviour for UCS and Young’s modulus data while the size effect behaviour of Poisson’s ratio was inconclusive. Also, the tensile strength and point load index data obtained from Gambier limestone revealed only descending size effect trends. The unified size effect law and its improved version were fitted to the UCS and Young’s modulus data leading to a very good agreement between the data and the model predictions. It was confirmed that an improved version of unified size effect law can predict a more realistic strength value for a sample with an infinite size. Finally, the applicability of two descending size effect models to the tensile strength and point load data was assessed and concluded that the multifractal scaling law is a suitable model for the point load data while the size effect law can better predict the tensile strength data.