Study on Monthly Runoff Forecasting Model Based on the Wavelet Transform

Wavelet transform is applied to the monthly runoff forecast of a hydropower station in southwest China, and the forecast model is established. The model selects the 40-year monthly runoff data from 1960 to 1999 for analysis and the 21-year monthly runoff data from 2000 to 2021 for check. Firstly, the monthly runoff time series is decomposed by using the MATLLAT Pyramid Algorithm, and several high-frequency signals and one low-frequency signal at different scales are obtained. Then, the obtained signal sequences are extended and predicted by the periodic mean superposition method, and finally, the prediction results are obtained by the MALLAT reconstruction algorithm. For the results predicted by wavelet theory, the mixed linear regression model is used to analyze and correct the results, so as to obtain a more reliable monthly runoff forecasting model.