Short-term Power Prediction of the Photovoltaic System Based on QPSO-SVM

Lei Yang (State Grid Hubei Electric Power Research Institute)
Zhou Shiping (State Grid Hubei Electric Power Company)
Xia Yongjun (State Grid Hubei Electric Power Research Institute)
Shu Xin (State Grid Hubei Electric Power Research Institute)

Indonesian Journal of Electrical Engineering and Computer Science
Vol 12, No 8: August 2014

Short-term power prediction of the photovoltaic system is one of the effective means to reduce the adverse effects of photovoltaic power on the grid. Since the efficiency of the traditional support vector machine(SVM) prediction method is low, this paper proposes the SVM based on the parameter optimization method of quantum particle swarm optimization(QPSO), and then apply into the power short-term prediction of the photovoltaic system. After comparing and analyzing the prediction results of SVM based on three optimization methods, we find that the QPSO-SVM method has better precision and stability, which provides reference to forecast generation power of the photovoltaic system.