Online fault detection and identification for an isolated PV system using ANN
A. AALLOUCHE, Hamid Ouadi
Abstract
In this paper, the problem of modeling, detecting, and identifying faults using an artificial neural network (ANN) for an isolated photovoltaic (PV) system, is addressed. The considered PV system consists of four parallel strings with four modules per string. The healthy state PV model is established by a neural network, where meteorological data (solar irradiance and ambient temperature) are considered as input variables and the coordinates (current and voltage) of maximum power point as targets. Several defects’ types are considered in this work: faults associated with the PV generator as well as faults linked to the boost converter. To classify the considered defects, two indicators are proposed. In fact, the detection and fault identification are ensured by a second ANN which compares in real-time the measured data with those delivered by the healthy PV model. The performance of the proposed diagnostic system has been validated by simulation using MATLAB-Simulink.