Machine Fault Detection using Instantaneous Frequency

This paper present a novel technique to analyze acoustical signal emitted from rotating machinery by means of instantaneous frequency (IF). IF is defined as the derivative of phase with respect to time from the analytic signal which one of its nice feature is the ability to track a highly time varying parameter. In this work, we recorded the sound pattern of the vibration machine simulator in Pusdiklat Migas Cepu with some simulated damage condition determined beforehand. From the results show that the IF representation shows higher clarity for tracking fundamental frequency of acoustical sound of rotating machinery compared to spectrogram which suffers from Heisenberg's uncertainty trade-off. For unbalance condition with 7.8gr of mass, the most dominant fundamental frequency at 2600 rpm, is at 94.68 Hz (15.41 dB SPL). The more unbalance mass is added, the higher machines rotation will be, and consequently the higher value of the amplitudes of the fundamental frequency as well. In the misalignment condition, the dominant of fundamental frequency occur in the vertical misalignment condition at 1160 rpm in 199,9 Hz is 12,07 dB. Keywords : rotating machinery, instantaneous frequency, fundamental frequency, sound pattern

Nama : Yuniar Rahmadana, Wiratno Argo Asmoro, Dhany Arifianto
Email : yuniar_ rahmadana@yahoo.com, arifianto@ieee.org
Kategori : POWER ELECTRONIC & ELECTRIC DRIVE
Institusi : Department of Engineering Physics, Faculty of Industrial Technology Sepuluh Nopember Institute of Technology Campus ITS - Sukolilo Surabaya 60111