LM-BP神经网络在大坝变形预测中的应用

Application of LM-BP neural network in predicting dam deformation

为了对大坝进行切实有效的监控,需要建立一个良好的大坝预测模型。针对传统BP(Back-Propagation)神经网络存在的收敛速度慢和泛化能力弱等缺陷,利用LM-BP(Levenberg Marquardt Back Propagation)算法对大坝变形进行预测,并根据丹江口大坝1996和1997两年的变形观测数据,对大坝挠度预测结果进行分析。结果表明,所建立的LM-BP神经网络的预测精度和收敛速度明显提高。

It is significant to establish an effective and practical dam safety monitoring model.The shortcomings of the traditional BP neural network lie in the slowness in the convergence rate and the weakness in the generalization ability.Based on the above,LM-BP neural network is adopted for predicting the dam deformation.With the measured data of Danjiangkou dam deformations in the year of 1996 and 1997 as examples,the deflection of dam is predicted using LM-BP.The results show that the proposed method can obviously enhance the forecasting precision and convergence rate.