A method of short-term wind power forecast based on wavelet transform and neural network

With the increasing scale of grid connected wind forms, it is important to predict the wind power in order to ensure stability of the power system, and make a reasonable dispatching scheme, and improve the wind form competitiveness in generation market. A novel method was proposed and applied to forecast the short-term wind power in this paper. Wavelet transforms and neural networks were combined in this method. First, the history wind speed and history wind power was decomposed by multi-resolution analysis. Then, the general signals and detail signals of wind power were forecasted by neural networks separately, which introduced the general signals and detail signals of wind speed as the effect factors. Finally, the general wind power and detail wind power were reproduced to obtain the forecasting wind power. The validity and feasibility of the method were verified through the actual data from a wind farm in China.