Enhancing fault detection and classification in photovoltaic systems based on a hybrid approach using fuzzy logic algorithm and thermal image processing

With the rapid expansion of photovoltaic systems globally, efficient fault detection and classification have become crucial to sustaining optimal energy yield and minimizing costly downtimes. Traditional PV fault detection techniques often face limitations in scalability, precision, and the ability to diagnose various fault types. Addressing this gap, this study presents a hybrid model that leverages thermal image processing integrated with the Mamdani fuzzy logic algorithm to accurately detect and classify six critical PV fault types: partial shading, total shading, cracked cells, short circuits, activated bypass diodes, and disconnected modules. This innovative approach employs a two-phase process wherein thermal image data undergoes feature extraction for temperature and shadow analysis, followed by fuzzy logic-based classification, achieving a high accuracy rate of 98.85 %. By delivering high accuracy, the method outperforms conventional techniques, reducing misclassification in diverse operational environments. The performance of the proposed model has been validated through extensive testing, establishing a comprehensive framework for PV system diagnostics. It represents a significant advancement in renewable energy technology and contributes meaningfully to advancing sustainable energy systems.