摘要
大坝在运营过程中由于受多种外界环境影响,表现出一种非线性、非稳定性变形特征。为了从大坝变形监测数据中有效提取变形规律,提高大坝变形的预测精度,本文在RBF(Radial Basis Function)神经网络模型的基础上,结合局部均值分解(LMD,Local Mean Decomposition)技术在对非线性、非平稳性信号分解中的优势,提出了一种LMD-RBF神经网络预测模型。该组合预测模型实现变形预测的步骤为:首先对变形监测数据进行LMD分解得到若干个PF(Product Function)分量以及余量;其次,使用RBF神经网络模型对各分量与余量进行预测得到各分量与余量预测值;最后,重构各分量与余量预测值得到最终预测结果。将本文提出的LMD-RBF神经网络模型应用于大坝变形预测中,结果表明,相比于BP(Back Propagation)神经网络模型与RBF神经网络模型,本文提出的组合预测模型的预测精度最高且稳定性最好,可为大坝等工程的变形预测提供思路与参考。
Due to the influence of various external environments,the dam shows a kind of nonlinear and unstable deformation charac-teristics.In order to effectively extract the deformation law from the dam deformation monitoring data and improve the prediction accu-racy of dam deformation,an LMD-RBF neural network prediction model is proposed based on the radial basis function(RBF)neural network model and the advantages of local mean decomposition(LMD)technology in the decomposition of nonlinear and non-stationa-ry signals.The steps of realizing deformation prediction by the combined prediction model are as follows:firstly,the deformation mo-nitoring data are decomposed by LMD to obtain several PF(product function)components and margins;Secondly,RBF neural net-work model is used to predict each component and residual,and the predicted values of each component and residual are obtained;Fi-nally,the predicted values of each component and residual are reconstructed to obtain the final prediction results.The LMD-RBF neural network model proposed in this paper is applied to dam deformation prediction.The results show that compared with BP(back propagation)neural network model and RBF neural network model,the combined prediction model proposed in this paper has the highest prediction accuracy and the best stability,which can provide ideas and references for deformation prediction of large dams and other projects.
作者
周志广
刘伟昌
ZHOU Zhiguang;LIU Weichang(Zhejiang Institute of Surveying and Mapping Science and Technology,Hangzhou 311100,China)
出处
《测绘与空间地理信息》
2023年第9期153-156,共4页
Geomatics & Spatial Information Technology
关键词
局部均值分解原理
RBF神经网络模型
变形预测
精度分析
local mean decomposition principle
RBF neural network model
deformation prediction
accuracy analysis
