基于人工神经网络的多泥沙洪水预报  

Application of the Artificial Neural Network Method to Forecast High Sediment Percentage Flood

作　　者：苑希民[1] 刘树坤[1] 陈浩[1] 

机构地区：[1]中国水利水电科学研究院灾害与环境研究中心,北京100044

出　　处：《水科学进展》1999年第4期393-398,共6页Advances in Water Science

摘　　要：阐述了人工神经网络方法对多泥沙洪水演进的辨识机理， 建立神经网络预报模型， 采用最小二乘快速收敛法， 通过模型联想实现水沙演进预报。研究结果与实测结果吻合良好， 体现了人工神经网络的应用价值和发展前景。In the foundation of traditional forecasting flood methods,the neural network technology is adopted to forecast flood transfer history with high sediment percentage.Using the neural network method,the intelligent models are set up,and the recognition mechanism of flood with high sediment percentage is explained.The least mean square method is adopted to train intelligent models and through the neural network thinking flood water level,the discharge and sediment percentage are forecasted.There are very important value and far reaching significance to study the theory and technology of forecasting the flood with high sediment percentage.

关 键 词：人工神经网络 智能辨识 多泥沙洪水 洪水预报 

分 类 号：P338.6[天文地球—水文科学] TP18[水利工程—水文学及水资源]