The load forecasting using the PSO-BP neural network and wavelet transform

An integrated BP neural network and particle swarm optimization (PSO) for load forecasting method is presented in this paper. From the signal analysis point of view, load can also be considered as a linear combination of different frequencies. The proposed approach decomposes the historical load into an approximate part and several detail parts through the wavelet transform. Then based on the maximum and minimum loads of the approximate part, the similar coefficients are given. The PSO-BP neural network, trained by low frequencies and the corresponding temperature records, is used to forecast the maximum and minimum load of the forecasting day. The short term load forecasting is forecasted by summing the predicted approximate part and the weighted detail parts. We knew that the PSO has a good capability for searching for optimal value. There are have few parameters to adjust than genetic algorithm (GA). The result demonstrated the accuracy of the proposed load forecasting scheme.