A data-driven fault diagnosis method for photovoltaic modules

The existing fault diagnosis of photovoltaic modules mainly focuses on a single fault, while in actual operation, the impact of climate and environmental conditions often leads to multiple faults occurring simultaneously. In response to the above issues, a data-driven composite fault diagnosis method for photovoltaic modules is proposed in this paper. The simulation model of Photovoltaic system is built, and the fault characteristic parameters of photovoltaic modules under different states are extracted according to the output characteristic curves under the three states of normal, single fault and composite fault. The composite fault diagnosis model of photovoltaic modules based on hybrid kernel Extreme learning machine is established. The parameters of hybrid kernel Extreme learning machine are optimized by using adaptive strategy to improve sparrow search algorithm, The accuracy of diagnosis model of hybrid kernel Extreme learning machine is improved. The effectiveness of the proposed method is verified through a simulation example of the actual operation data of a photovoltaic power plant. The simulation results show that this method can accurately diagnose single and composite faults of photovoltaic modules, and has certain engineering application prospects.