3D slope reliability analysis based on the intelligent response surface methodology
LaiFu Song,Xiang Yu,Bin Xu,Rui Pang,ZeYu Zhang
DOI: https://doi.org/10.1007/s10064-020-01940-6
IF: 4.2
2020-07-26
Bulletin of Engineering Geology and the Environment
Abstract:At present, there are two problems in 3D slope reliability analysis: (1) the slope stability analysis method; (2) how to effectively extend the reliability analysis method from 2D to 3D slopes. This paper combines the advantages of the finite element method and the limit equilibrium method, develops a 3D slope stability analysis finite element sliding surface stress method program, and introduces the radial basis function network (RBFN) intelligent response surface method, which has the characteristics of strong adaptability, high fault tolerance, greater flexibility, and strong nonlinearity. With the response surface function built for the general framework for an intelligent response surface methodology for the reliability analysis of a 3D slope system, the reliability analysis of a slope is extended from 2D to 3D. The process involves generating samples, creating an intelligent response surface, and calculating the failure probability of the 3D slope system. Through the reliability analysis of typical examples and a comparison with Monte Carlo simulation (MCS) method results, the accuracy, feasibility, and superiority of the proposed intelligent response surface methodology for application to the reliability analysis of a 3D slope system were verified. A comparison with the analysis results of the 3D slope shows that the 2D slope stability calculation is too conservative and seriously overestimates the probability of slope instability. Furthermore, 3D analysis can consider the true stress state of a slope and is able to specify a definite range for a sliding body, thereby providing a basis and reference for determining reinforcement regions.
engineering, environmental,geosciences, multidisciplinary, geological