Very short term Solar Irradiance Prediction for a microgrid system in Taiwan based on Hybrid of Support Vector Regression and Grey Theory

Abstract

In recent years, energy crisis becomes a global issue. The use of renewable energy in electricity generation has increased significantly. Natural resource such as solar energy is available in large amount, but it is unpredictable. Solar Irradiance Prediction (SIP) is very important to estimate Photovoltaic (PV) power generation. The generated power of PV will affect power dispatch, scheduling, or even the stability of a microgrid system. This paper proposes to use the Hybrid of Support Vector Regression (SVR) and Grey Theory Models for very short term (VST) SIP (in the range of minutes). The proposed model has been validated with the actual measured data and will be implemented in the Energy Management System (EMS) of a microgrid system in Taiwan. Comparing to the models based on either the Grey theory or the SVR, the proposed method yields higher accuracy.