FAULT DETECTION IN PHOTOVOLTAIC SYSTEMS USING THE INVERSE OF THE BELONGING INDIVIDUAL GAUSSIAN PROBABILITY

Abstract

This article addresses the problem of fault early detection in photovoltaic systems. In the production field, solar power plants consist of many photovoltaic arrays, which may suffer from many different types of malfunctions over time. Hence, fault early detection before it affects PV systems and leads to a full system failure is essential to monitor these systems. The fields of control and monitoring of systems have been extensively approached by many researchers using various fault detection methods. Despite all this research, to early detect and locate faults in a very large photovoltaic power plant, we must, in particular, think of an effective method that allows us to do so at the lowest costs and time. Thus, we propose a new robust technique based on the inverse of the belonging individual Gaussian probability (IBIGP) to early detect and locate faults in the power curve as well as in the Infrared image of the photovoltaic systems. While most fault detection methods are well incorporated in other domains, the IBIGP technique is still in its infancy in the photovoltaic field. We will show, however, in this work that the IBIGP technique is a very promising tool for fault early detection enhancement.