Machinery diagnostics is an important research area that requires inter-disciplinary expertise. Typically, machines operate in complicated, uncertain and varying environments. A good diagnostic methodology is expected to work accurately in tough conditions. Early and accurate fault diagnosis of rolling element bearings is important for improving mechanical system reliability and performance. Signals, even from experimental setups, let alone from industrial machinery monitoring and control, can be interfered by environmental noise. A worst-case condition is impulsive noise, caused by external stimulations, e.g. hydraulic hummers; it can cause strong and complicated interferences and it can lead to erroneous conclusion concerning the performance of the ball bearing. On the other hand, avoiding external stimulations by employing optimised machinery mounting, although vital, is not always possible in an industrial environment. The aim of this project is to analyse and assess the mounting of the experimental device to estimate non-linear signal distortions due to external impulse stimulations. Therefore, two signal processing methods are employed, the Zhao-Atlas- Marks time-frequency distribution (ZAMD) and the Hilbert Huang transform (HHT) to analyse non-stationary signals obtained from a ball bearing testing setup. Comparative results are provided by applying the two methods on vibration signals derived by the testing setup under an external stimulation.
|Number of pages
|Journal of the Balkan Tribological Association
|Published - 1 Jan 2015