Contribution to the devlopment of a fault diagnosis technique for photovoltaic generators

Abstract

In today's world, renewable energy is becoming increasingly popular due to growing concerns about the environment and increasing demand for energy. Of all forms of renewable energy, photovoltaic energy appears to be the most promising option for global production. However, PV power plants often encounter technical faults that can negatively affect the performance of PV panels and their components. These defects include partial shading, aging, cracks in photovoltaic cells, hot spots, short circuits, open circuit faults, and others. To effectively address these challenges, an automatic fault diagnosis system is crucial. This thesis proposes three new metaheuristic techniques for modeling photovoltaic generators quickly and accurately. The third technique was used for fault diagnosis through the process of the PV generator modeling itself. This strategy adopted for fault diagnosis is based on the assumption that I-V data acquired during healthy conditions produce correct model parameters, while I-V data acquired during faulty or abnormal conditions produce modeling parameters that contain the same faults’ signatures. The results indicate that this strategy is innovative, reliable, and accurate and that the initial assumption was valid and proven.