EMD-ARI模型在大坝渗流预报中的应用研究  

Application of EMD-ARI Model in Seepage Forecasting of Dam

作　　者：刘彩花[1] 景浩[1] 成一雄 牛月[1] 

机构地区：[1]太原理工大学水利科学与工程学院,山西太原030024

出　　处：《水力发电》2014年第7期49-52,共4页Water Power

基　　金：山西省国际合作项目(2013081034);2013年太原理工大研究生教改课题

摘　　要：利用EMD秉留原数据属性的特点,基于MATLAB平台对汾河水库左坝岸渗流数据进行分解,初步提取可能周期项与趋势项,对剩余值序列基于SAS平台进行时间序列分析,建立ARI模型,各参数重构建立EMD-ARI渗流耦合模型。将此模型对大坝渗流量的预报结果与ARI单一模型的预报结果进行比较,结果表明,EMD-ARI耦合模型的预报精度明显优于ARI单一模型,可为大坝安全运行状态监控及模型构建提供参考。The EMD-ARI seepage coupling model for the dam of Fenhe Reservoir is built by reconstructing various parameters, in which, the seepage data of the left dam shore of Fenhe Reservoir are decomposed by using the Empirical Mode Decomposition （EMD） adhered to keep the characteristics of the original data and based on the MATLAB platform to preliminarily extract possible cycle and trend terms, and the Auto Regressive Integrated （ARI） model is established by using time series analysis on residual sequence based on SAS platform. The comparison of the forecasting results between EMD-ARI model and ARI model shows that the EMD-ARI coupling model is superior to the ARI single model in accuracy. The EMD-ARI coupling model provides a reference for dam safety monitoring and model building.

关 键 词：时间序列分析 EMD分解 渗流预报 汾河水库 

分 类 号：TV698.1[水利工程—水利水电工程]