BP神经网络算法预测路基边坡滑坡变形的应用研究

为了准确掌握降雨量与路基边坡滑坡变形之间的关系，在对金温货线铁路K135+900～K136+400滑坡区域降雨量及路基边坡滑坡变形进行长达3年监测的基础上，采用BP神经网络算法，以前期累计降雨量和累计变形进行样本训练，建立网络关系，并以实测数据加以验证，模型能够对降雨条件下的变形演化发展做出较为准确的预测。研究成果已作为重要的基础性数据被当地政府管理部门、铁路运营管理单位和滑坡治理设计单位所采用，用于分析预测滑坡变形趋势，指导滑坡应急管理和永久治理，以确保滑坡影响区内铁路运营安全和人民生命财产安全。

Application of BP Neural Network Algorithm Predicting Landslide Deformation of Subgrade Slope

In order to accurately grasp the relationship between rainfall and landslide deformationof subgrade slope, on the basis of monitoring rainfall and landslide deformation ofsubgrade slope in K135+900～K136+400 landslide area of Jinwen freight line railway for three years, BP neural network algorithm is adopted to conduct sample training with the accumulated rainfall and accumulated deformation in the early stage and establish the network relationship. The model is verified by the measured data and can accurately predict the deformation evolution under rainfall conditions. The research results have been adopted by the local government management departments, railway operation management unitsand landslide treatment design units as an important basic data, being used to predicte landslide deformation trend, guide landslide emergency management and permanent treatment and ensure the safety of railway operation and people's life and property in the landslide affected areas.