Fault Diagnosis of Turnout Switch Machine Based on Brain-Inspired Intelligence

Abstract

As the core equipment to control the running direction of railway vehicle, the turnout switch machine, its failure will lead to a series of problems. At present, more and more researchers conduct intelligent re-search on its fault diagnosis. In this paper, aiming at different types of faults during the operation of turnout switch machines, a spiking neural networks model based on apoptosis mechanism (BASNNs) is established as a classifier for fault diagnosis. The model uses neuron competition rules and intelligent unsupervised learning algorithms to update the network. The system uses the non-contact measurement of the sound signal as the input, and uses the combination of empirical mode decomposition (EMD) and wavelet packet decomposition energy entropy (WPDE) to extract features of the sound data. Experimental results show that the proposed method has better data recognition ability than traditional ANNs with the same number of layers.