Prediction of Grid-connected Photovoltaic Power Generation Based on Multivariate Statistics
Ji Xu, Hong Zhou and Yanjun Fang
Photovoltaic grid-connected generation technology is an effective way to make full use of solar energy, but its power generation has great randomness; therefore it is necessary to predict its output power. Firstly, this paper briefly introduced the influencing factors of photovoltaic grid-connected power generation. Then three factors, solar radiation intensity, temperature and humidity, were chosen, and multivariate statistical regression model was established using multivariate statistical theory, through which the output power of power generation system was predicted. Then the model was used to predict a photovoltaic grid-connected power generation system. The results showed that the predicted data curve basically fit the measured data curve, and the error value was very small. The root-mean-square error (RMSE) value and mean absolute percentage error (MAPE) value of the model were also small, which showed that the model had good prediction accuracy. The multivariate statistical regression model in this study can accurately predict the output power of photovoltaic grid-connected power generation, which is worth further promotion and use.
Keywords: photovoltaic grid connection, multivariate statistics, power generation forecasting, output power, regression model