Key Issues and Future Trends on Solar Power Forecasting

Abstract

Owing to the effect of cloud, solar irradiance is intermittent and causes uncertainty in power system operations. Thus, accurate solar power forecasting helps system operators reduce the risk and cost of generator scheduling. Compared to wind power forecasting, solar power forecasting has a higher challenge because the prediction of solar irradiance is difficult. However, the curve of solar irradiance includes a daily pattern. Thus, statistical methods or deep learning models capture the important characteristics of solar power generation, which improves forecasting accuracy. This study summarizes essential issues about solar power forecasting, including potential training models, data processing, efficient weather classification, and other corresponding technologies. In addition, possible trends for solar power forecasts are also addressed.