Detection of under-lubricated ball bearings using vibration signals

Abstract

Reduction in the lubricant level of a bearing will reduce the operating life of the bearing. This condition is a concern in lightly loaded bearings, such as cooling fans in electronics applications, whose life is dependent on the service life of the lubricant. If these bearings are manufactured with a sub-optimal amount of lubricant, the degradation of lubricant as well as the wear processes in the bearing can be accelerated, thereby leading to early failures of the fans. Measurements were carried out on bearings containing varying amounts of grease, ranging from none to the nominal amount specified by the manufacturer. Features were extracted from vibration signals that were obtained using an accelerometer mounted on the cooling fan. Measurements were performed on fans operated at various temperatures and speeds. The changes observed in the vibration signals as a function of the operating speed and temperature were utilized to develop features which enable the classification of the bearings according to the amount of grease in the bearing. Finally, a classification method for the detection of under-lubricated bearings was developed using the dependence between the features, temperature, and operating speed. This method can be used as a rapid method for acceptance testing of bearings.