Research on stability control of deep learning algorithms in soft fault diagnosis technology for photovoltaic inverters

Aiming at the problems of complex circuits and short fault duration in photovoltaic (PV) inverters, a fault diagnosis algorithm based on Self-Organizing Map (SOM) and Back Propagation (BP) neural networks is proposed. Simulations are conducted using Simulink software to model the faults of PV inverters. Based on the collected relevant parameters, the SOM-BP neural network is employed for fault diagnosis. The simulation results show that the proposed SOM-BP neural network achieves high accuracy rates in both the training set and the test set, demonstrating a high accuracy in fault diagnosis. Effectively reduce operation and maintenance costs, ensure the stable operation of PV power plants, and help the sustainable development of PV power generation industry.