Development of Cloud Movement Prediction Method for Solar Photovoltaic System

Variability of power generation due to the prevalence of cloud cover over solar photovoltaics (PV) power plants is a challenge faced by grid operators and independent system operators (ISOs) in the integration of solar energy into the grid. Solar forecasts generated through ground-based sky imaging systems are useful for short-term cloud motion predictions. However, the cost of sky imaging systems currently available in industries is relatively high. Hence, a ground-based camera system utilizing a simple webcam is proposed in this study. The proposed method can produce predictions with high levels of accuracy. Forecasts were generated through video analysis using MATLAB for the computation of cloud motion predictions. The image processing involved in the implementation of the proposed system is based on the detection of cloud regions in the form of a cluster of white pixels within individual frames and tracking their motion through comparison of subsequent frames. This study describes the techniques and processes used in the development of the proposed method, along with the evaluation of performance through analysis of the results. The predictions were carried out over multiple time horizons. The time horizons selected include 5, 10, 15, 20, 25, and 30 s. The overall results computed showed promising accuracy levels above 94.60%, which makes it adequate for generating reliable forecasts.