基于自适应模糊推理系统模型的径流中长期预报

Long-Term Prediction of Discharges Using Adaptive-Network-based Fuzzy Inference Systems Models

介绍了自适应模糊推理系统ANFIS的原理结构及学习算法。以漫湾和双牌两座水库实测月径流序列为研究对象,研究不同的输入及不同的模糊数对自适应模糊推理系统模型做中长期预报的影响,并通过与人工神经网络模型的预报结果进行比较,显示本模型是中长期水文预报方法中较为准确的方法之一。

Accurate time and site-specific forecasts of streamflows and reservoir inflow are required for effective hydropower reservoir management and scheduling.Intelligent computing tools,such as artificial neural network and fuzzy logic approaches,are proven efficient when they are applied individually to a variety of problems.Recently there is a growing interest in combining both these approaches,and as a result,a neuro-fuzzy computing technique, ANFIS,emerges.The principle and structure of the Adaptive-Network-based Fuzzy Inference System(ANFIS) is presented as well as a hybrid learning algorithm. Using the long-term observations of discharges of monthly river flow discharges in Manwan Reservoir and Shuangpai Reservoir,different antecedent input flows and types of membership functions associated with ANFIS model are tested.Comparing with the ANN model performance,it is illustrated that the ANFIS model is an effective algorithm to forecast the long-term discharges.