A New Power Forecasting Method for Photovoltaic Plants under Hazy Conditions

Abstract

Power forecasting has become more and more important for the safe and economic operation of Photovoltaic (PV) plants. In practice, the power generation of PV plants may be affected by haze conditions. To improve the power forecasting accuracy of PV plants under hazy conditions, the Air Quality Index (AQI) is adopted for the power forecasting modeling in the paper. The relationship between the AQI and solar irradiance is analyzed firstly, which indicates that the AQI and solar irradiance have a significant negative correlation. It reveals the weakening effect of haze on solar irradiance. A clear boundary exists in the scatter diagram of the AQI and solar irradiance under different haze conditions which represents the maximum solar irradiance which the PV plant can receive at a moment. Then, a new method is proposed in which the AQI is adopted to correct the solar irradiance from Numerical Weather Prediction (NWP). The corrected NWP solar irradiance serves as an input variable for the Artificial Neural Network (ANN) model. Finally, the measured data is used to compare the different power forecasting methods. The results show that the proposed method can effectively improve the power forecasting accuracy of PV plants under hazy conditions.