FLAC和神经网络在隧道位移反分析中的应用
来源期刊：有色金属科学与工程2011年第6期
论文作者：彭剑文 赵奎 曹宗权 马乾天 刘明荣
文章页码：79 - 82
关键词：反分析；位移；正交实验；神经网络；侧压系数
Key words：back-analysis; displacement; orthogonal test; neural network; lateral pressure coefficient
摘 要：应用三维有限差分程序FLAC3D和BP神经网络对隧道位移进行分析，使用正交实验和FLAC3D正演结果作为样本，用神经网络建立围岩位移与反演参数的映射关系反演得出了围岩的弹性模量和初始地应力测压系数，并使用FLAC3D正算验证反演参数的精度结果表明可搜索得出反演参数的最优解，实现在隧道围岩中的位移反分析可将反演结果用于隧道的设计，反演精度满足工程要求。
Abstract:
This paper studies the application of neural network and FLAC3D to the back-analysis of displacements of tunnel displacement, using the learning and testing samples based on orthogonal test design and FLAC3D numerical simulation. The potential mapping between parameters and surrounding rock displacement was established using neural network. The modulus of elasticity and lateral pressure coefficient of surrounding rock were obtained by verifing the precision of inversion parameter. The results show that it can solve the problem by searching parameters of back-analysis, which can be derived to achieve the displacement back analysis in tunnel displacement. The inversion results can be feedback for the design of tunnel. The results of back-analysis are accord with the accuracy of engineering.
