Incipient fault diagnosis method for IGBT drive circuit based on improved SAE

An incipient fault diagnosis method devised for insulated gate bipolar transistor (IGBT) drive circuit based on improved stack auto-encoder (SAE) is recommended. First, the Monte Carlo method is applied to extracting the time domain response signal of the circuit under test as sample data. Then, with SAE used to extract essential features of data, the SAE is employed to extract features of sample data. Meanwhile, multi-classification relevant vector machine (RVM) is involved for fault diagnosis of the acquired features. As the structure of the hidden layer in SAE and the learning rate could exert a significant effect on the feature extraction performance, in this paper, the quantum particle swarm optimization (QPSO) algorithm is used to optimize the above parameters. As revealed by the experimental results, the improved SAE method is effective in the extraction of the essential characteristics of the incipient faults for the IGBT drive circuit. Further with this, the incipient fault multi-classification RVM of the IGBT drive circuit is capable of achieving 100% diagnostic accuracy.