Estimation of Remaining Useful Life of Slow Speed Bearings Using Acoustic Emission Signals

Various direct and indirect sensing methods for machine condition monitoring have been reported in the literature. Among these methods, acoustic emission technique is one of the effective means of monitoring rolling element bearings during industrial processes. Today, many machines use computerized classification in a wide range of applications. Further, recent developments indicate the drive towards integration of diagnosis and prognosis algorithms in future integrated machine health management systems. With this in mind, this paper concentrates on the estimation of the remaining useful life for bearings whilst in operation. To implement this, a linear regression classifier and multilayer artificial neural network model have been proposed to correlate the selected AE features with corresponding bearing wear throughout laboratory experiments. Results showed that the proposed models exhibit good prediction performance. This paper also presents the use of a new representative fault indicator, signal intensity estimator, employed for AE signals originating from natural degradation of slow speed rolling element bearings. It is concluded that the obtained results were promising and selecting this appropriate signal processing technique can significantly affect the defect identification.