基于LMBP算法的崇阳溪流域降雨径流预报模型研究  

Research on rainfall runoff forecast model of Chongyang River basin based on LMBP arithmetic

作　　者：金保明[1] 王伟 杜伦阅 李斐斐 JIN Baoming;WANG Wei;DU Lunyue;LI Feifei(College of Civil Engineering,Fuzhou University,Fuzhou,Fujian 350108,China)

机构地区：[1]福州大学土木工程学院

出　　处：《福州大学学报（自然科学版）》2019年第6期842-847,共6页Journal of Fuzhou University(Natural Science Edition)

基　　金：福建省自然科学基金资助项目(2016J01734)

摘　　要：采用泰森多边形法对流域进行划分,分别确定崇阳溪上游流域6个雨量站控制子流域的面积权重.选择1997至2014年的14场流域降雨径流过程为训练样本,以上游6个雨量站的时段雨量和武夷山水文站前期流量为输入,武夷山水文站相应流量为输出,采用3层网络,其中隐含层节点数采用试算法确定,建立崇阳溪上游流域LMBP神经网络降雨径流预报模型.利用余下的7场降雨径流过程对模型进行检验,结果表明,模型运算速度快、时效性好,预报精度符合要求,可以用于流域的降雨径流预报.Thiessen polygon method was used to divide the basin,and the area weights of the control sub-basin of six rainfall stations were determined respectively in the upper reaches of Chongyang River.Fourteen rainfall runoff processes from 1997 to 2014 in the upper Chongyang River were selected as training samples.The period rainfall of six precipitation stations in the upper reaches and the previous flow of Wuyishan station were input,and the corresponding flow of Wuyishan station was taken as output.Three-layer network was adopted,in which the number of hidden layer nodes was determined by trial and error method,and LMBP neural network rainfall runoff forecast model of Chongyang River upstream basin was established.The remaining seven rainfall runoff processes were used to test the forecast model.The results showed that the LMBP neural network model had the advantage of fast operation speed and good timeliness,and the prediction accuracy met the requirement of the specification,and it could be used to forecast rainfall runoff in river basins.

关 键 词：降雨径流 预报模型 LMBP神经网络 崇阳溪流域 

分 类 号：TV124[水利工程—水文学及水资源] P338[天文地球—水文科学]