Intelligent Frameworks for Solar Photovoltaic Condition Monitoring: From Classical Machine Learning to Advanced Deep Learning

The rapid and global expansion of photovoltaic (PV) systems has made maintaining their reliability and efficiency a fundamental challenge in modern industry. System components are continuously exposed to harsh environmental conditions that cause various diagnostic faults. If left undetected, these anomalies significantly reduce energy production and create severe safety hazards across the entire grid. Traditional Fault Detection and Diagnosis (FDD) methods rely heavily on physical models and static thresholds, but these conventional approaches lack the required accuracy and scalability for complex and dynamic operating conditions. Recently, the integration of Artificial Intelligence (AI) has revolutionized this engineering field. This paper presents a comprehensive evaluation of Machine Learning (ML) and Deep Learning (DL) applications for identifying and classifying anomalies in solar power plants. The primary objective is to explore the structural transition from traditional mathematical methods to modern data-driven architectures. By analyzing different fault types, this study evaluates the strengths, limitations, and real-time capabilities of intelligent networks in ensuring the stable operation of condition monitoring systems.

کلیدواژه ها:

Solar power plants ، condition monitoring ، intelligent algorithms ، anomaly classification ، data-driven architectures