Fault diagnosis of bearings and gears using impulse-induced phase discontinuity

Abstract

The vibration data from the defective bearings and gears are complex and challenging to comprehend due to the presence of strong machinery vibrations. A fault in a bearing or a gearbox introduces periodic impulses in the vibration signal. These impulses are difficult to detect at the early stage of fault due to their weak strength at the early stage. These impulses are buried in background machinery vibrations. The present study focuses on developing a zero frequency filter (ZFF) and instantaneous frequency (IF)-based method for detecting the impulses in the vibration signal due to the faults. The vibration signal is filtered through a ZFF to make the signal near monotone and to preserve the impulse information as a local disturbance, resulting in phase discontinuity at the impulse location. These disturbances are small and difficult to observe in the presence of background machinery vibration from the ZFF output. The IF is obtained from the phase of the monotone signal to enhance the weak impulse information. The combined approach of ZFF and IF enables the detection of weak impulses at the early stage of faults. The working of the algorithm is tested with the bearing and gearbox datasets. The present algorithm successfully detected the naturally grown incipient-stage faults in bearings and distinguished clearly between healthy and faulty gearboxes. Therefore, the proposed technique indicates potential applications in the condition monitoring of bearings and gears.