Reservoir Monthly Runoff Forecast Model Based on Wavelet-ANFIS Analysis
Huifang Guo · Water Resources and Power · 2008
According to the character of runoff changing,a wavelet-ANFIS model is presented for reservoir monthly runoff forecast.Monthly runoff series data are decomposed into individual components at the different scales by Mallat,and then low frequency signals and high frequency signals can be gotten according to different scales.Based on different signals,ANFIS model is applied to forecast respectively.Then the forecast results of every model are composed to form the forecast of original runoff.This model is used in Zhao Ping-tai reservoir monthly runoff forecast.The result shows that the runoff changing trend can be forecasted and the precision is improved comparing with the single ANFIS forecast model,but still needs improving.The reason engendering this aftermath is analyzed,and suggestion for improving model is made.