Short-Term Photovoltaic Output Forecasting Based on Multivariable Phase Space Reconstruction and Support Vector Regression

摘要

A comprehensive forecasting method of the short-term photovoltaic output is presented in this paper to forecast the short-term output of two types: the days with weak weather fluctuates (sunny days) and the days with strong weather fluctuations. Firstly, the support vector regression (SVR) is used to forecast and track the reconstructed irradiated intensity and temperature, the forecasted data of the radiation intensity and temperature are used as the input of the forecasting model. At last, the high nonlinear fitting of the photovoltaic output is realized by the SVR model also. Through the analysis of the actual data of photovoltaic power plants, the rationality and effectiveness of this method are verified in this paper. Compared with the traditional forecasting method using SVR only, the useful information contained in the historical data of photovoltaic output can be further explored and the accuracy of the PV output forecasting is also improved.