Solar Cell Fault Detection Using Robust Machine Learning Algorithms

In photovoltaic systems, defect detection by using Artificial Intelligence (AI) algorithms is quite mandatory for maintaining its reliability and efficiency. As the adoption of solar technology increases, the need for effective monitoring and fault detection becomes paramount to ensure optimal performance and minimize downtime. In this context, we explore various machine learning algorithms including K-nearest Neighbor, Random Forest, and AdaBoost to detect various types of faults, like string fault, string to string fault, and string to ground fault. The performance of these algorithms is evaluated based on F1score, precision, recall and specificity. Our findings indicate that our Random Forest and AdaBoost models demonstrate higher accuracy in fault detection compared to other state-of-the-art methods. This research highlights the potential of machine learning in enhancing the reliability of PV systems and contributes to the ongoing efforts to optimize solar energy systems. Further, it supports the transition towards sustainable energy solutions.