Application of surface solar radiation estimation model based on CCD/IRS data in ultra-short-term photovoltaic power prediction

In order to study the application of the surface solar radiation estimation model in ultra-short-term photovoltaic power prediction, and reduce the interference caused by cloudy weather to the prediction, the surface solar radiation estimation model is optimized based on charge-coupled device/infrared spectroscopy (CCD/IRS) data. Then, the effective model for ultra-short-term photovoltaic power prediction is constructed, and its reliability is verified through experiments. This paper first introduces the principle of CCD/IRS data, second, the surface solar radiation estimation model is improved with the consideration of the influence of cloud cover. Two different weather types, cloudy and cloudless, are selected to verify the estimation model. The results of the two experiments respectively show: (1) compared with traditional algorithms, the proposed surface solar radiation estimation model has more advantages in estimating surface solar radiation under cloudy weather and complex terrain conditions, and can be applied to the photovoltaic power prediction; (2) the constructed ultra-short-term photovoltaic power estimation model based on the surface solar radiation estimation model can be used to obtain accurate and effective prediction results in cloudless weather. The proposed model also has a significant improvement with fewer deviations in the cloudy weather. There are significant reference values for the application of surface solar radiation estimation model in the accurate prediction of ultra-short-term photovoltaic power and safe operation.