Prediction of building energy consumption based on PSO - RBF neural network

Abstract

At present, building energy conservation is a hot topic in urban construction and energy conservation research. Predicting the trend of energy consumption is very meaningful for a whole building energy management. Compared with the other feed-forward neural networks, RBF network learning faster and the ability of function approximation is stronger, but its performance still need to be improved. We use particle swarm optimization algorithm (PSO) to optimize RBF neural network and use the optimized RBF neural network to predict energy consumption in this article. Used the statistical data of the whole society's monthly electricity consumption published online as a sample, and simulated the forecasting method by MATLAB.