Fault diagnosis of cascaded inverter based on PSO-BP neural networks

Aiming at the power component open circuit faults of the cascaded inverter, the fault model is set up, and the PSO-BP neural network is used to diagnose the faults. At the same time, in order to avoid the premature convergence in the basic PSO algorithm, some mutation operations are conducted upon the particles. The wavelet decomposition is used to extract the fault characteristics for training and testing, and then the improved particle swarm algorithm is used to optimize the weights and the threshold of the BP neural network. The method can improve the convergence speed of the traditional BP algorithm and avoid trapping in local minimum easily. The simulation results show that this method has higher diagnostic accuracy and faster convergence speed. It is effective for the fault diagnosis of the cascaded inverter.