Tracking Of Non-Linearity Growth Caused By Imbalance Using Hilbert Transform

Non-linearities affect every mechanical system, sooner or later. Imbalance, misalignment, looseness, breathing cracks and Coulomb friction are some of the most common causes of non-linearities. In this paper, the Hilbert transform, a power signal processing technique, is used to detect and track non-linearities caused by imbalance. To detect non-linearity and its growth, a machine fault simulator (also referred to as a vibration simulator) was used. The simulator was divided into four regions: (a) disc drive end; (b) disc non-drive end; (c) motor drive end; and (d) motor non-drive end. Frequency domain responses were measured on the horizontal and vertical axes of all four defined regions. Frequency response functions (FRFs) were generated for the first peak frequency (1X). Succeeding harmonics, being less sensitive to detecting imbalance, were ignored. The FRFs were then processed through the Hilbert transform and the results were analysed by incorporating Hoaui's correction terms to avoid truncation errors. The results clearly demonstrate that the Hilbert transform can successfully detect non-linearity and its growth caused by imbalance.