A review on signal processing techniques utilized in the fault diagnosis of rolling element bearings

Rolling element bearings play a crucial role in the functioning of rotating machinery. Recently, the use of diagnostics and prognostics methodologies assisted by artificial intelligence tools such as artificial neural networks, support vector machines etc. have increased for assessing the health of the rolling element bearings. The effectiveness of these approaches largely depends upon the quality of features extracted from the bearing signals. Keeping this in mind, the authors have presented the various signal processing methods applied to the fault diagnosis of rolling element bearings with the objective of giving an opportunity to the examiners to decide and select the best possible signal analysis method as well as the excellent defect representative features for future application in the prognostic approaches. The review article first quotes some of the condition monitoring tools used for rolling element bearings and then the importance of signal processing methods in diagnosis and prognosis of rolling element bearings. Next, it discusses the various signal processing methods and their diagnostic capabilities by dividing them into three stages: first stage corresponding to the articles published before the year 2001, second stage refers to the articles published during the period 2001-2010 and lastly the third stage pertains to the articles issued during the year 2011 to till date. To focus more on the recent developments in the signal processing methods, the third stage has been partitioned further into several sections depending upon the methodology of signal processing. Their relative advantages and disadvantages have been discussed with regard to the fault diagnosis of rolling element bearings. (C) 2015 Elsevier Ltd. All rights reserved.