基于BP神经网络的隧道位移反分析方法研究

The Research on the Inverse Analysis Method of Tunnel Displacement Based on the BP Neural Network

通过设置合理的神经网络结构参数,采用基于数值模拟与正交试验方法建立的训练样本,最终由数学计算得到BP神经网络反演模型。通过神经网络反演模型位移反演分析,将分析结果与地质勘察成果相结合,对围岩性质等目标参数进行优化调整,对于隧道支护参数设计有实际意义,有助于进一步加强隧道稳定性。

By setting the reasonable parameters of neural network structure, this paper established the training sample based on the numerical simulation and orthogonal experiment method, and finally obtained the BP neural network inversion model by the mathematics calculation. Through the inversion analysis of displacement by neural network inversion model, it also combined the analysis results with geological survey results, optimized and adjusted the target parameters such as the nature of the surrounding rock properties, which had practical significance for the design of tunnel support parameters and was helpful to further strengthen the stability of tunnel.