Short-term Load Forecasting Based on Wavelet-Particle Swarm

In view of the power load with the randomicity and the complexity, the short-term power load forecasting based on optimal wavelet-particle swarm is introduced in this paper. First, the power load series is decomposed several frequency ranges by wavelet packet. Select the optimal wavelet tree to reconstruct the coefficients of the wavelet packet and form the number of power load components. Then, forecast the reconstructed series with the particle swarm optimization neural network, respectively, introduce hourly temperature factor for the low frequency components and promote the prediction precision by the newest temperature information. In addition, taken a city´s power system into test and simulation to test the advantage of this method, and proved that it has more advantage and better efficiency.