Fault Diagnosis of Electric Drive Systems Based on Dynamic Independent Component Analysis and A Hidden Markov Model

Due to the rapid development of high-speed trains, the fault diagnosis of electrical drive systems is a hot issue. A fault diagnosis method for electrical drive systems based on dynamic inner independent component analysis and the hidden Markov model is proposed. First, the offline data of the electric drive systems under different operating states are processed by using the dynamic inner independent component analysis algorithm. The dynamic features in the offline data are extracted by establishing equations that maximize the non-Gaussianity of the predicted latent variables. Second, these dynamic features are used to train a Hidden Markov Model to represent various states of the electric drive systems. Finally, the online data are injected into the Hidden Markov Model for fault diagnosis, and the experimental results on the electric drive systems simulation platform show the effectiveness of the method.