【摘要】 本文从潮流泥沙模型出发 ,以同时刻的U、V、H和前一时刻的S、U、V、H为输入 ,本时刻的S为输出 ,建立了杭州湾含沙量模拟的模糊神经网络模型。经初步实验表明含沙量的预报较为困难 ,本模式只能作实时的预报 ,其预报尺度有限 ,只能有效预报 1～ 3步 ,这一方面说明杭州湾含沙量变化的复杂性和不确定性 ,另一方面也表明试图对杭州湾含沙量过程作长期预报是不现实的。更多还原

【Abstract】 This paper presents a new network named AFNN (Adaptive Fuzzy Neural Network) for simulation of continuous sediment concentration on the basis of tidal flow\|sediment model.In this AFNN,S is considered as output,the value of the instant (U\,V\,H) and at previous instant (S\,U\,V\,H) are considered as input,the fuzzy relationships between sediment and various factors are discussed by the field data training and learning with fuzzy clustering and neural network in Hangzhou Bay.The results of primary application shows the network is only for real\|time forecast or short\|term sediment concentration forecast,which indicates sediment forecast is difficult,the forecast scale is very finite and the long\|term sediment forecast is unpractical.更多还原