An Investigation into Faults of PV system using Machine Learning: A Systematic Review

Due to the ongoing maintenance required for consistent generation efficiency, the rapid expansion of photovoltaic (“PV”) energy generation as an alternative to conventional fossil fuel generation has spurred the search for more efficient and dependable operating and maintenance processes. As a result, the current study first discusses the many failure modes that all photovoltaic systems face, as well as the critical integrated detection techniques and technologies. Machine learning (ML) has demonstrated the functional potential to discover and categorize “PV” defects among the many methodologies. A thorough investigation of the integration of ML and “PV” system fault detection methods is undertaken in this article. The ability of the ML approach to managing a non-linear connection, identifying characteristics with identical signatures, and their online application has made FDD in “PV” systems more appealing in recent years. According to the survey, the most often used ML approaches in recent literature are SVM and ANN. In addition, the review study identifies key implementation issues and makes recommendations for further research.