Diagnostique des systèmes photovoltaïques par les techniques d’intelligence artificielle

Abstract

The performance of a solar panel is not limited in terms of its design and construction materials, but it is greatly affected by faults that can disrupt or at least minimize its performance. In order to deal with these faults, it is important to identify them as soon as they appear. Several techniques and methods have been proposed. A literature review of recent diagnostic methods has enabled us to propose in this work a diagnostic method based on the use of the value of the short-circuit current and the fill factor as input parameters. In addition to the short-circuit current chosen by the majority of studies as input data for fault diagnosis and detection, this work proposes the use of a new criterion, the fill factor, in order to refine the diagnosis of the various faults and make it more reliable. Diagnosis will be carried out in two stages: the first one is based on threshold detection, in which fault identification is carried out simply by considering the threshold and consequently the signature of each symptom, while the second stage uses artificial intelligence techniques, particularly for cases presenting the same fault signatures. At the end of this work, a simplified fault diagnosis method can be proposed, this method is based on the use of the value of the short-circuit current and the fill factor using artificial intelligence techniques. This methodology enables us to effectively diagnose the presence of faults on photovoltaic panels.