大坝变形监测的BP网络模型与预报研究

建立有效实用的大坝安全监测模型，对于馆控大坝运行意义重大。针对目前国内外常用统计模型、确定性模型等的不足，提出将基于误差逆传播算法的BP神经网络模型用于大坝变形监测数据的拟合分析及其预测预报研究，最后以福建水口混凝土重力坝变形监测为例，对坝顶垂直位移实测值建立了BP网络模型，并将模型用于坝顶垂直位移预报，结果表明，BP网络模型的拟合和预报精度明显优于相应的统计模型。

Dam Deformation Monitoring Model and Forecast Based on BP Algo rithmof Artificial Neural Networks

With an aim at the shortage of statistic models and determinative models used in dam safety monitoring both home and abroad, the BP neural networks model based on error back propagation algorithm is suggested to use in research on the fitting analysis and prediction and monitoring of dam deformation. With the deformation monitoring of Shuikou concrete gravity dam in Fujian province as an example, BP networks model is established for the real measurement of vertical displacement of the dam crest, which can also be used in forecasting the vertical displacement of dam crest. The outcome shows that the accuracy of BP networks model is obviously superior to that of statistic model.