Machine Learning based PV Fault Detection, Classification, and Localization using Solar Cell Parameters

Abstract

In recent years a lot of work is done regarding energy generation from photovoltaics (PV). But still the efficiency of this power is far less than the other energy resources. By Increasing the performance of the PV system, not only reduces the overall cost but also enhances the life span of the PV system. Previously many efforts have been put in this direction. This research presents the detection, classification, and localization of partial shadow (PS) fault, inter string line-to-line (ISLL) fault, intra string line-to-line (ESLL) fault, line to ground (LG) fault and open circuit (OC) fault through machine learning Gaussian process regression (GPR). Solar cell’s parameters for one diode model (ODM) are utilized for detection, classification. In addition, defining fault location is as important to detect them. Fault location is also determined in this work.