Statistical analysis of bimslope stability using physical and numerical models

Abstract

Bimrocks are complex geological mixtures of strong blocks and weak matrix occurring various geological environments such as melanges, fault rocks and weathered rocks. Slopes underlain by these and other bimrocks (bimslopes) occur in weathered and near-surface conditions. In this study, artificial bimslopes were generated using physical and numerical approaches. The physical bimslopes were constructed as a matrix of blended sand-petrolatum surrounding ellipsoidal, cementitious blocks. The principle geomechanical properties were determined by appropriate laboratory tests. The physical models were fabricated with Volumetric Block Proportions (VBP) of 20%, 30%, 40% and 50%. Ten different arrays of blocks were fabricated for each VBP, yielding a total of 40 different physical models. To investigate the slope stability of physical bimslopes at laboratory scale, the physical models were tested using a tilt table apparatus, which has been constructed by the authors of the current paper, and the threshold face angle of the bimslopes at failure were measured. The numerical models, developed using the results of the physical model bimslopes, were analyzed using the finite difference method. To analyze the results of physical and numerical modeling, statistical indices, such as standard error of the mean, were utilized. The results show that by increasing the VBP, both the stability of bimslopes and the uncertainty increases in the stability analysis findings. An equation between variations in the range of safety factor and VBP was derived with a significant coefficient of determination: VarSF = 0.008VBP [(R2) of 0.95]. By using this equation, it is possible to evaluate the range of safety factors based on VBP and conservatively estimate the safety factor of a bimslope. Finally, using a validated numerical model, a sensitivity analysis was carried out to investigate the effects of strength contrasts between the materials on the stability of the bimslopes.