Methods of photovoltaic fault detection and classification: A review

Abstract

Photovoltaic (PV) fault detection and classification are essential in maintaining the reliability of the PV system (PVS). Various faults may occur in either DC or AC side of the PVS. The detection, classification, and localization of such faults are essential for mitigation, accident prevention, reduction of the loss of generated energy, and revenue. In recent years, the number of works of PV fault detection and classification has significantly increased. These works have been reviewed by considering the categorization of detection and classification techniques. This paper improves of the categorization of methods to study the faulty PVS by considering visual and thermal method and electrical based method. Moreover, an effort is made to list all potential faults in a PVS in both the DC and AC sides. Specific PV fault detection and classification techniques are also enumerated. A possible direction for research on the PV fault detection and classification, such as quantum machine learning, internet of things, and cloud/edge computing technologies, is suggested as a guide for future emerging technologies.