Power Load Forecasting Based on Improved Elman Neural Network

In order to reduce the economic loss of power plants, it is necessary to improve the forecasting accuracy of power generation. In this paper, the Elman Neural Network (ENN) is used to predict the short term load of the power grid. And the Simulated Annealing (SA) algorithm and Particle Swarm Optimization (PSO) algorithm are adopted to optimize the parameters. SA is introduced into the PSO with hybridization and Gaussian mutation. The SA can jump out of the local optimal solution, prevent the PSO algorithm from falling into the local minimum value, and improve the global optimization ability. In this paper, the fusion algorithm is deduced and proved. The experimental results show that the proposed algorithm has better predictive precision and generalization ability than the traditional ENN.