An alternative bearing fault detection strategy for vibrating screen bearings

Abstract

The transformation of the mining machinery industry is oriented towards the adoption of highly sophisticated technologies, with the aim of enhancing operational efficiency. Specifically, there is a focus on improving the nature and quality of operations carried out, such as vibrating screen. Vibrating screens, being one of the pivotal machines, have gained extensive utilization in mining processing plants. However, the demanding working conditions can lead to potential failures in certain components of vibrating screens. Bearings are one of the components that are susceptible to faults. Moreover, owing to the unique working environment and structure of vibrating screens, the signals extracted from defective bearings in vibrating screens exhibit greater complexity and distinctiveness, including (i) vibration interference related to the unbalanced shafts, (ii) cyclic impulses produced by the damaged bearing, (iii) non-cyclic random impulsive noise owing to pieces of coal falling into the deck, and (iv) inevitable noise from environmental conditions. To address this issue, an alternative procedure for bearing fault detection from vibrating screens served in the coal industry is proposed in the work. The strategy for bearing fault signature extraction covers three main steps: (a) a new fast resampled iterative filtering (FRIF) method is utilized to remove the vibration interference produced by the unbalanced shafts and simultaneously eliminate large amounts of background noise; (b) a signal processing method based stable distribution is designed to cancel the random impulses (impulsive noise); and (c) an improved energy operator demodulation technique is proposed to identify the bearing fault characteristics from the purified signal. The proposed strategy is validated using simulated and real tests.