Analysis and Modification of Fault Detection Methods in Photovoltaic Array

The reliability of the larger photovoltaic (PV) systems is very vulnerable with respect to faults. The faults may transpire in dc/ac side PV system. The detection, classification and location identification of faults are very important to mitigate the energy loss, accidents, etc. It is also required to ensure the periodic revenue generation. In this paper, different faults have been simulated and datasets are collected from it for normal and faulty conditions. From the collected datasets, relevant features were extracted and then fed into a Machine Learning (ML) algorithm. To predict and classify faults with new data, the ML model was trained with historical data. The condition of partial shading in PV arrays is also addressed and a progressive, evolutionary Maximum Power Point Tracking (MPPT) method is employed to combat it. In this work, ML algorithms are carried out for fault analysis of PV array. The developed ML algorithm is tested with few practical datasets.