Research on Short-Term Load Forecasting Approach for Smart Grid

To solve the problems including improving the precision of power load forecasting, speeding up the convergence speed, improving the possibility of falling into the local minimum and the single optimization process, a short-term forecasting model for the power supply load for smart grid system is established in this article. The particle swarm optimization(PSO) is used to initialize the reverse neural network, and the improved genetic algorithm(GA) is also introduced. In the crossing process, the optimal value of parent generation is integrated with the next generation population to optimize the network weight and improve the model performance. Then, by using the advantages of high nonlinearity of neural network and PCA reduction dimension principle, the prediction results are obtained through Matlab simulation. The results show that the approach proposed in this paper is reasonable and effective for short-term load forecasting.