Enhancing the Performance of the Photovoltaic Cells Employing Computer Vision

In recent years, there has been a growing interest in interdisciplinary research and large-scale economies towards adapting to renewable energy and utilizing solar power. However, several environmental factors make it necessary to provide a reliable and fault-tolerant control solution that can ensure the main objectives of power generation, even in the presence of faults. This paper aims to review the challenges of diagnosing faults in solar power systems and propose a hybrid and cloud-enabled architecture for a health monitoring system for photovoltaic (PV) farms. The proposed architecture employs both model-based and data-driven methods in a unified framework, with a focus on data privacy and easy integration into currently available cloud technologies. We propose a new 2-stage transfer learning mechanism (that utilize reinforcement learning) to increase detection accuracy. This allows for a fully autonomous fault-tolerant control solution that can detect, localize, and rectify numerous types of faults in PV systems, including shade faults.