A Novel Power Load Forecasting Method Considering Output Rate of Wind and Photoelectric Power

Accurate power load forecasting is highly significant for efficient and reliable operation control of modern power grid. In this paper, we propose a novel power load prediction method by considering the output rate of wind and photoelectric power, which aims to improve the energy management level and promote grid development. First, an optimized support vector regression (OSVR) is proposed to build the prediction space, where a combination kernel is deployed to effectively fuse multiple feature information including time, wind power output rate, and photoelectric power output rate. The output rate of wind and photoelectric power can be measured and obtained by the related measurement equipment. To enhance the prediction performance of OSVR, then, an adaptive genetic algorithm (AGA) is used to adjust its parameters dynamically. Finally, the proposed AGA-OSVR model is tested on the real power load dataset of a province in China. The experimental results prove that the proposed model can obtain a higher prediction accuracy.