基于Elman神经网络的面板堆石坝沉降预测模型  

Settlement Prediction Model of CFRD Based on Elman Neural Network

作　　者：吕飞[1,2] 沈振中[1,2] 

机构地区：[1]河海大学水文水资源与水利工程科学国家重点实验室,江苏南京210098 [2]河海大学水利水电学院,江苏南京210098

出　　处：《水电能源科学》2011年第12期56-59,共4页Water Resources and Power

基　　金：河海大学水文水资源与水利工程科学国家重点实验室专项基金资助项目(2009586012)

摘　　要：分析了某面板堆石坝运行初期坝体的监测资料,选择水压分量与时效分量为影响因子构建逐步回归分析模型,应用反馈神经网络理论建立Elman神经网络模型,并与逐步回归模型预测精度做了对比分析。结果表明,Elman神经网络模型预测精度高、可靠,有助于分析大坝的安全性态。This paper analyzes the monitoring data of a CFRD in initial operation. The stepwise regression analysis model is established by choosing hydraulic and aging components as influence factors. An Elman neural network model is built for settlement prediction. Compared with stepwise regression analysis model, the results show that Elman neural network model has high precision and reliability for predicting settlement of CFRD. Thus, it has contributed to analysis of dam safe state.

关 键 词：面板堆石坝 沉降 反馈神经网络 ELMAN神经网络 预测模型 

分 类 号：TV641.4[水利工程—水利水电工程] TV698.1