Gear and bearings fault detection using motor current signature analysis

Abstract

The importance of monitoring machine health increases with the aim of reducing maintenance costs and improving availability and quality of service. In this context, the main challenge for researchers is to find a convenient and effective method of monitoring gearboxes in a non-invasive manner. Because of busy industrial environments, background noise reduces the sensitivity of vibration signals and acoustic emissions. Consequently, the Motor Current Signal Analysis (MCSA) is a promising non-intrusive alternative to vibration measurement for gearbox monitoring. On the other hand, the monitoring of bearing faults is very important in maintenance programs since these problems account for over 40% of the total amount of failures in the induction motor. Hence, this article presents the use of the Discrete Wavelet Transform (DWT) as well as the K-means clustering technique applied to bearing and gear faults indicators. The Discrete Wavelet Transform is used as a filtering method in order to extract the various frequency bands containing the defect profile. The indicators used in this work are descriptors of geometric shapes that are obtained from Clarke transformation. These indicators provide a particular shape in the presence of a mechanical fault. The results of the carried out experiments demonstrate the efficiency of the Discrete Wavelet Transform combined with the Clarke's shape indicators for bearings and gears fault classification.