Fault Detection and Diagnosis in Photovoltaic Systems Using Decision Tree Method

Recently, there has been a significant increase in the popularity of solar energy for generating electricity. Technological developments have led to the global acceptance and use of photovoltaic (PV) systems. Nonetheless, these systems require meticulous monitoring and regular maintenance to enhance efficiency. PV systems are susceptible to a range of faults, spanning from temporary to permanent issues, and swiftly and affordably pinpointing these faults poses a significant challenge. This is crucial for maintaining system functionality without disrupting regular operations. Hence, implementing an effective fault detection system is imperative to mitigate damage from PV module faults and safeguard the system against potential losses. This research introduces a decision tree method, implemented through the Classification Learner in MATLAB, designed to identify and diagnose faults within PV systems. The result of trained DT models exhibited excellent fault detection and diagnosis accuracy in false alarm, partial shadow and faulty modules on the test set, with 85.40 %.