神经网络在高坝基岩多点位移计监测分析中的应用

针对高坝基岩多点位移计监测的实际情况，采用神经网络预报模型，对基岩变位中一孔多点监测、多孔联测和裂隙开合度等工程监测分析实际问题进行研究。通过对反向传播（BP）模型输入层因子的比较分析，并以实测资料加以训练，建立了多种基岩变形的预测模型，预测效果好，有利于多点位移计监测资料的综合分析及对高坝基岩状态的监控。

Application of Neural Network to Multi-point Extensometer Monitoring of High Dam Bedrock

By use of the artificial neural network model, some practical engineering problems in multi-point monitoring in one hole, multi-point monitoring in different holes and the crack opening monitoring are studied based on the survey data of bedrock deformation of a high dam measured by the multi-point extensometer. A predicting model is established for the deformation of various kinds of bedrock based on the comparative analysis of the input factors of the BP model, and it is trained with the survey data. The established model performs very well, favourable for the comprehensive analysis of survey data and monitoring of the bedrock state of high dams.