A Three-Dimensional Nonlinear Back-Analysis Method for Rock Mass Mechanical Parameters in Large Underground Caverns

摘要

This paper presents a three-dimensional nonlinear finite element-based back-analysis method for the rock mass mechanical parameters in underground caverns. The proposed approach incorporates the theory of back-propagating (BP) neural networks, relative incremental displacement monitoring results, and the orthogonal design method for rock mass parameter sets. Because the relative incremental displacement computation results are clearly influenced by the mechanical parameters, the back-analysis of mechanical parameters can be ascertained using the numerical simulation and BP neural network methods. The paper further applies the proposed method to ascertain the rock mass mechanical parameters in the underground caverns of the Xiangjiaba Hydro Power Station. The computational results show that the back-analysis method can be applied in to calculate the rock mass mechanical parameters effectively.