崇阳溪流域PRBP神经网络洪水预报模型研究

打开文本图片集

SI Qi'， JIN Baoming1*，LU Wangming²， CHEN Zhaoqingl （1.ColegeofCiviEineinguzouUvesityuzo8，Ca;2.aningFlooddroughtisastereveiote Nanping ， China）

Abstract：ThePoak-Ribiereconjugategradientbackpropagationalgorithm（PRBP）ofnumericaloptimizationtechnologywasused， and21rainstormandfloodprocessesfrom1997to2O2 intheupperreachesofChongyangRiverbasinwerestudied.Therainfal volumeofsixrainfalltationsintheupperreachesofChongyangRiverbasinandthepreviousdischargeof Wuyishan Hydrological Stationwereregardedasinput，anditscorespondingdischargewasregardedasoutput;thenumberofhiddenlayerunitswas determinedbytrialcalculation，andthenPRBPneuralnetworkfloodforecastingmodelofChongyangxiRiver Basinwasestablished. Theremainingeightfloodswereusedtotestandvalidatethemodel.TheresultsshowthatcomparedwiththatoftheconventionalBP neuralnetworkodel，tonergencespdofeodelissterndthalculationspeedisviouslyiproed;thteistic coefficient of the model is greater than O.87，and the relative error of peak flowof six floods is within 10% . The forecasting accuracy meets the requirements，which can provide a basis forthe flood control department to forecast the flood.

Keywords：PR conjugate gradient method; BP neural network; flood forecasting; Chongyang River Basin

山区流域洪水往往具有破坏性强、预见期短、预报难度大的特点。（剩余8880字）