Research on Photovoltaic Arc Fault Diagnosis Model Based on Machine Learning

Photovoltaic power stations are prone to arc faults due to wind, sun, resulting in frequent occurrence of photovoltaic fires and economic losses to the photovoltaic. As a result, it is crucial to identify arc defects in solar power plants in order to guarantee their dependability and efficiency. This research proposes a multi-group string comparison method for photovoltaic fault detection. The photovoltaic system's arc fault model is established, the arc fault characteristics are obtained, and the characteristic quantities are extracted. the arc diagnosis model of photovoltaic power station is established, which can effectively identify series fault and parallel arc fault based on particle swarm optimization random forestrc. Its accuracy is higher than that of random forest recognition, the anti-interference of the diagnosis model is analyzed by the actual power station operation data. It is verified that the model is not interfered by the inverter Internal mechanisms and environmental factors.