Application of vibration signal analysis and artificial intelligence methods in fault identification of rolling element bearings

Abstract

Paper deals with the automation of fault identification in roller element bearings by the means of artificial neural networks and vibration signal analysis implementation. Several types of roller element bearing faults in combination with several stages or rotor imbalances were tested. Both frequency and time domain vibration features are used as inputs to fault classifiers. Several architectures of multilayer perception artificial neural network were applied.