Time Domain Analysis of Sound Signals for Bearing Damage Identification

Muhammad Aswin (Dept. of Mechanical Engineering, Brawijaya, University Malang)
ING Wardana (Dept. of Mechanical Engineering, Brawijaya, University Malang)
Yudy S. Irawan (Dept. of Mechanical Engineering, Brawijaya, University Malang)
Hadi Suyono (Dept. of Electrical Engineering, Brawijaya University, Malang)

Proceeding of the Electrical Engineering Computer Science and Informatics
Vol 1: EECSI 2014

Time domain analysis requires less computational time compared to the frequency domain. Analysis is performed directly on the signal without any conversion at all. This paper describes high-frequency signal analysis on sound produced by rotated bearing. Three bearing conditions - normal, damaged, and badly damaged - was chosen to obtain the characteristics of high frequency sound. From the entire spectrum of the recorded sound, the higher frequency range looks very different for the three conditions bearing. Phisically, more damage rotated bearing, the disturbance sound heard more loudly. Bearings were rotated at various rpm, from low to high, to ensure the similarities and differences in characteristics. Average energy and standard deviation were calculated as bearing damage indication.