Knowledge-based clustered partitioning method for reliability analysis of slope stability

A new global optimization reliability method, knowledge-based clustered partitioning (KCP) method, is proposed. The proposed method includes five steps, namely, partitioning, random sampling, calculation of the polar radius, backtracking, and calculations of reliability index and design points. A flowchart for the proposed method is presented. Moreover, a C-language based computer program is developed to carry out the reliability computations. Two examples of reliability analysis for rock slope stability with plane failure are presented to demonstrate the validity and capability of the proposed method. The results indicate that the proposed method can obtain the reliability index and the design points simultaneously. Furthermore, the global optimization solutions can be obtained. The proposed method can ensure sufficient accuracy for reliability computations; and its efficiency is significantly higher than the traditional Monte Carlo simulations, which can be considered as a potential method for reliability analysis of slope stability, especially for slope stability involving implicit and nonlinear performance function. The proposed KCP method with equal-step-length can search the angles systematically, which results in the accurate design points. It is recommended that angle below ten degree should be adopted to ensure sufficient accuracy and reduce the computational effort as low as possible.