Multi-branch convolutional neural network with generalized shaft orbit for fault diagnosis of active magnetic bearing-rotor system

Fault diagnosis based on vibration signals in active magnetic bearing-rotor systems is an important research topic. However, it difficult to obtain discriminative features represent faults due the nonlinear and non-stationary characteristics of diverse sources failures. Hence, this paper proposes a novel end-to-end learning mechanism multi-sensor data fusion learn fault representation structural bearings. Taking five displacement sensors bearing as signal sources, generalized shaft orbits are constructed converted into discrete 2D images. Based these images, multi-branch convolutional neural network designed achieve high types. The experiments performed rig supported by bearings, effectiveness proposed algorithm verified, proving suitability cases with changing rotating speeds sample lengths.