计算机技术与发展
計算機技術與髮展
계산궤기술여발전
COMPUTER TECHNOLOGY AND DEVELOPMENT
2015年
4期
200-203
,共4页
BP神经网络%粒子群算法%洪水流量%洪水预报
BP神經網絡%粒子群算法%洪水流量%洪水預報
BP신경망락%입자군산법%홍수류량%홍수예보
BP neural network%particle swarm optimization%flood flow%flood forecasting
受四川盆地地形与北部秦岭山脉的影响，达州市河流众多，洪灾频频发生，每次洪灾都给达州市政府和人民带来巨大的经济损失和惨重的人员伤亡。文中以四川省达州市州河流域为研究对象，采用达县水文站月平均流量作为洪水属性，提出了一种基于PSO算法优化的神经网络，并建立了洪水预报模型。通过实验仿真对比，其预报精度高于传统的BP神经网络，具有预报结果合理、相对误差小、收敛速度快、预报精度高等优点，从而能更有效地帮助防汛部门预报洪水，降低洪水带来的风险，还能够为达州市的防汛工作提供一定的参考意见。
受四川盆地地形與北部秦嶺山脈的影響，達州市河流衆多，洪災頻頻髮生，每次洪災都給達州市政府和人民帶來巨大的經濟損失和慘重的人員傷亡。文中以四川省達州市州河流域為研究對象，採用達縣水文站月平均流量作為洪水屬性，提齣瞭一種基于PSO算法優化的神經網絡，併建立瞭洪水預報模型。通過實驗倣真對比，其預報精度高于傳統的BP神經網絡，具有預報結果閤理、相對誤差小、收斂速度快、預報精度高等優點，從而能更有效地幫助防汛部門預報洪水，降低洪水帶來的風險，還能夠為達州市的防汛工作提供一定的參攷意見。
수사천분지지형여북부진령산맥적영향，체주시하류음다，홍재빈빈발생，매차홍재도급체주시정부화인민대래거대적경제손실화참중적인원상망。문중이사천성체주시주하류역위연구대상，채용체현수문참월평균류량작위홍수속성，제출료일충기우PSO산법우화적신경망락，병건립료홍수예보모형。통과실험방진대비，기예보정도고우전통적BP신경망락，구유예보결과합리、상대오차소、수렴속도쾌、예보정도고등우점，종이능경유효지방조방신부문예보홍수，강저홍수대래적풍험，환능구위체주시적방신공작제공일정적삼고의견。
Influenced by the Sichuan basin and the north Qinling mountains, there are many rivers that flooding frequently occurs in Dazhou. Every time,floods bring huge economic loss and heavy casualties to the government and people. Taking Zhou River in Dazhou City,Sichuan Province,as the study objects,use the monthly average flow as flood properties collected by Daxian hydrological stations, a neural network based on PSO algorithm is proposed,and a flood forecasting model is established. By the simulation experiments,the forecast accuracy is higher than traditional BP neural network,the prediction result is reasonable with small relative error,fast convergence rate and high prediction accuracy. It will help flood control departments to predict flood flow effectively and reduce the risks of flooding, and also can provide some reference opinions to flood control work in Dazhou.
