The application of artificial neural networks to the diagnosis of induction motor bearing condition using Hilbert-based bispectral analysis

Three approaches based on Hilbert-based bispectral vibration analysis are investigated as vibration signal preprocessing techniques for application in the diagnosis of a number of induction motor rolling element bearing conditions. The bearing conditions considered are a normal bearing and bearings with outer and inner race faults. The vibration analysis methods investigated are based on the fusion Hilbert transform and the bispectrum, the bispectrum diagonal slice and the summed bispectrum. Selected features are extracted from the vibration signatures so obtained and these features are used as inputs to artificial neural networks trained to identify the bearing conditions. The results obtained show that the diagnostic system using a supervised multi-layer perceptron type neural network is capable of classifying bearing conditions with high success rate, particularly when applied to the Hilbert-based summed bispectrum signatures.