基于BP神经网络的旬降雨径流相关预报模型  

Ten-day Correlation Forecast Model of Rainfall and Runoff Based on BP Neural Network

作　　者：梁国华[1] 习树峰[1] 王本德[1] 

机构地区：[1]大连理工大学土木水利学院,辽宁大连116024

出　　处：《水力发电》2009年第8期10-12,39,共4页Water Power

基　　金：国家"十一五"科技支撑项目(2006BAB14B05);国家自然科学基金项目(50809011)

摘　　要：参照流域暴雨与径流相关的特点,以实际的旬初土壤蓄水量、旬降雨量为输入条件,以实际的旬径流量为输出条件,建立BP神经网络旬降雨径流相关模型。实例表明,模型结构简单,可操作性强,利用该模型进行的旬径流预报具有较好的模拟精度,并为利用旬降雨预报信息实施旬径流中期预报奠定了基础。The ten-day correlation forecast model of rainfall and runoff based on BP neural network was established by taking actual initial soil water content and ten-clay rainfall as its inputs and the actual ten-day runoff as its output, according to the watershed characteristics of rainfall and runoff correlation. The example results showed that this model is simple in structure and easy to operation, and its ten-day runoff forecast has good simulation precision. This model laid a foundation for the ten-day runoff forecast by using ten-day rainfall forecast information.

关 键 词：降雨径流预报模型 旬径流 旬初土壤蓄水量 旬降雨 BP神经网络 

分 类 号：TV121.1[水利工程—水文学及水资源]