Abstract: In probabilistic slope stability analysis, the selection of slip surface for calculation usually varies from critical deterministic surface to critical probabilistic surface. The reliability index for the former is not the minimum based on a deterministic surface, the latter is closer to the actual failure probability as a similarity to reliability of the slope system. In this paper, a technique put forward by Hassan and Wolff for the determination of slip surface with the minimum reliability index is presented. The critical probabilistic surface could be simply determined by several stability calculations according to specific combinations of soil parameters, which requires no special programs. The method’s validity is examined by two cases including simple homogeneous and heterogeneous soil slopes. Besides, the orthogonal design is applied to slope stability analysis for comparison. The results show that the critical probabilistic slip surface corresponding to the minimum reliability index conforms well with the surface corresponding to a minimum safety factor according to the combinations of soil parameters. Therefore, the critical probabilistic slip surface could be determined by this method in engineering applications.

Key words: slope stability, slope reliability analysis, critical deterministic slip surface, critical probabilistic slip surface, FLAC^3D, Rosenblueth method