Short-Term Power Load Forecasting Method Based on Empirical Mode Decomposition-Glowworm Swarm Optimization-Back Propagation

The short-term prediction of power load is of great significance for the planning and development of the power industry. With the reform and development of the power market, the short-term prediction of power load is extremely important for industrial manufacturing enterprises to effectively reduce energy costs. However, the actual load series data exhibit multiple complex properties, such as nonlinearity, non-stationarity, and time variation. In this paper, we propose a three-level hybrid ensemble short-term load forecasting method consisting of Empirical Mode Decomposition (EMD), Glowworm Swarm Algorithm (GSO)and BP Neural Network (BP). EMD is used to decompose load data to obtain good electricity consumption characteristics; GSO is used to obtain the optimal weights and thresholds required for BP prediction. The hybrid ensemble method (EMD-GSO-BP) was evaluated using the data of Chaoyang County, Liaoning Province in 2017. The EMD-GSO-BP method was compared with two other mainstream coupling methods (BP, GSO-BP). Statistical analysis shows that the proposed method shows better prediction accuracy on the three standard scales of MAPE, MAD and RMSE, which reflects the advanced nature of the method.