FAULT DETECTION IN ROLLING ELEMENT BEARINGS USING VIBRATION AND ACOUSTIC EMISSION SIGNALS Sami Al-Sulti, B. Samanta * , K. R. Al-Balushi and S. A. Al-Araimi Department of Mechanical and Industrial Engineering, College of Engineering Sultan Qaboos University, PO Box 33, PC 123, Muscat, Sultanate of Oman. ABSTRACT A study is presented for fault detection in rolling element bearings using vibration and acoustic emission (AE) signals. The acquired signals of a rotating machine with normal and defective bearings are analysed using different signal processing techniques. The features obtained from the original and the processed signals are used for detection of bearing condition. The features include statistical, spectral, cepstral and time-spectral parameters of the acquired and the preprocessed signals. The procedure is illustrated through the experimental vibro-acoustic signals of a rotating machine for different types of bearing faults. Several signal processing techniques with both vibration and AE signals are considered. The results present a comparative study of the signals and the signal processing techniques for detection of different types of faults in rolling element bearings. KEYWORDS: Signal processing, Vibration, Acoustic Emission, Rolling Element Bearings. 1. INTRODUCTION Machine condition monitoring is gaining importance in industry because of the need to increase reliability and to decrease the possibility of production loss due to machine breakdown. The use of vibration signals is quite common in the field of condition monitoring of rotating machinery. By comparing the signals of a machine running in normal and faulty conditions, detection of faults like mass unbalance, rotor rub, shaft misalignment, gear failures and bearing defects is possible. These signals can also be used to detect the incipient failures of the machine components, through the on-line monitoring system, reducing the possibility of catastrophic damage and the down time [1-7]. There are several conventional methods for detection and identification of machine faults. The methods use representation signals in time or frequency domains, which are valid for stationary signals. However, during the initial stages of fault development, the vibration signals may not be stationary thus limiting the applications of time or frequency domain methods. To overcome the non- stationary nature of the signals, time-frequency methods like short-time Fourier transform (STFT), Wigner-Ville distribution (WVD) have been proposed for machine condition monitoring and • Corresponding author Tel: +968 515355 Fax: +968 513453; E-mail address: samantab@squ.edu.om