Matlab仿真平台下大坝位移BP神经网络模型研究

BP Neural Network Model to Monitor Dam Deformation in Matlab Simulation Platform

基于人工神经网络的非线性映射能力,应用Matlab7.1网络仿真平台,结合辽宁省白石水库多年大坝位移实测数据,建立了3种不同改进BP算法的多层前馈神经网络模型。并通过LM算法、BR算法、GDX算法的BP网络模型的拟合、预报结果,对3种模型的应用效果进行了比较分析,得出了LM算法的BP网络更适合用于建立坝顶位移监控模型的结论,以实现对大坝位移实时、有效的监控。

On the basis of the nonlinear reflection ability of artificial neural network,we established three multi-layer feedforward neural network models in Matlab 7.1 simulation platform to monitor the Baishi reservoir deformation in Liaoning Province.The three models adopt different modified BP algorithms,i.e.LM algorithm,BR algorithm,and GDX algorithm.According to the fitting and prediction results,we compared the application results of the three models and concluded that the BP network based on LM algorithm was more suitable for building dam＇s displacement monitoring model to realize real-time and effective monitoring.