基于BP神经网络的岩体力学参数反演及边坡整体稳定分析  

Inversion Analysis of Rock Mass Mechanical Parameters and Slope Stability Analysis Based on BP Neural Network

作　　者：柏俊磊[1] BAI Junlei(Powerchina Northwest Engineering Corporation Limited,Xi'an 710065,China)

机构地区：[1]中国电建集团西北勘测设计研究院有限公司,西安710065

出　　处：《西北水电》2020年第6期61-67,共7页Northwest Hydropower

摘　　要：岩质边坡应力应变分析,需要确定岩体变形参数与强度参数,而岩土体力学参数的确定是工程界难题之一。文章根据南方某水电站实际运行过程中边坡岩体力学性质的动态变化,利用坡体上变形监测资料对边坡岩体力学参数进行反演,选取三维有限元模型关键节点的监测数据与计算数据进行对比,结果表明测点两者大小接近、变化趋势一致,即利用BP神经网络反演得到的参数值具有一定可靠性,可为边坡的稳定性分析以及位移预测提供参考。Deformation parameters and strength parameters of the rock mass are needed in the analysis of the stress and strain to the rock slope,and to determine the rock and soil mechanical parameters is one of the engineering problems.Based on dynamic mechanical properties of slope rock mass during the actual operation of a hydropower station in south China,the mechanical parameters of rock mass on the slope is back analyzed with the monitored displacement data,then the monitoring data selected from key nodes of the three-dimensional finite element model are compared with the calculated data.The results show that the two values of the measuring points are close and the variation tendency is the same,that is,the parameter values obtained by inversion based on BP neural network is reliable and can provide reference for slope stability analysis and displacement prediction.

关 键 词：BP神经网络 监测数据 参数反演 稳定性 

分 类 号：TU457[建筑科学—岩土工程]