Adaptive Cloud Movement Prediction for Photovoltaic Systems Using Real-Time Sensor Data and Deep Learning

Advancements in solar forecasting have been achieved through various techniques of computer vision and deep learning. This paper introduces a novel approach for cloud position prediction over photovoltaic (PV) power plants using real-time sensor data. We gather voltage drop data from sensors placed on the strings of solar panels, convert this data into matrices representing the PV system over time, and design a deep learning model to predict future cloud positions.