基于 BP 神经网络的面板堆石坝参数反演分析

摘要

Dams with years of loads can see changes in their physical and mechanical parameters,which pose as a direct threat to its safety and stability. In order to track these parameters,this paper makes use of nonlinear mapping feature of BP neural network,and,combined with the field data,puts forward a method of inversion analysis for the dam. Its feasibility is proved by the match of inversion analysis results and the experimental data.%运行多年的大坝在各种荷载的作用下，大坝及其基础的物理力学参数发生一定的变化，对大坝的安全与稳定有着直接的影响。为了准确地获取这些参数，可以利用 BP 神经网络非线性映射的特性，结合大坝实测资料，提出了大坝反演分析的方法，并利用反演分析的成果对大坝进行分析，计算结果与实测资料基本吻合。