Short-Term Electric Load Forecasting Model based on PSO-BP

To improve the accuracy of load forecasting and improve the efficiency of power generation, a PSO-BP model is proposed. Particle Swarm Optimization (PSO) is used to optimize the parameters such as parameter weight and learning rate of the BP neural network, and then the PSO-BP model is used for load forecasting. The algorithm is verified on the load data set constructed in this paper, and the feasibility of the combined algorithm is verified. The prediction error is 3%. It can be seen from the experimental results that compared with BP and LSTM neural networks, the RMSE values of the PSO-BP algorithm are increased by 63.6 % and 54.90 % respectively, which can achieve higher prediction accuracy.