Fault detection of bearings in a drive reducer of a hot steel rolling mill

Abstract

Defective bearings can jeopardize the good functioning of rotating machinery. In this work we employ multivariate statistical techniques to monitor a drive reducer in a hot steel rolling mill, with the aim of detecting incipient defects associated to rolling bearings. Several vibration signals are measured and processed for this purpose, as well as the current absorbed by the motor driving the mill. A normal condition reference model is first constructed and deviations from it are detected by monitoring T2 statistics. Classical bearing defect models are employed to test the fault detection capabilities of the method.