Abstract:In order to locate the fault of MMC sub module accurately and quickly, this paper proposes a fault location method of MMC sub module based on PSO-BP. Firstly, the characteristics of MMC sub module open circuit fault are analyzed. According to the wavelet analysis theory, the wavelet packet node energy of capacitor voltage of each sub module of MMC bridge arm is extracted. Then the normalized wavelet packet node energy of capacitor voltage of each sub module is input into BP neural network for training. By coding the sub module of bridge arm, the problem of fault location of sub module is transformed into a problem of classification and recognition, By comparing the expected output of BP neural network with the actual output, the fault block can be located, and the initial weights and thresholds of BP neural network are optimized by particle swarm optimization algorithm, so that the fault location is faster and more accurate. Finally, the simulation results based on MATLAB / Simulink show that the method can quickly and accurately locate the faulty sub module in MMC bridge arm.