Experimental investigation of ball bearing fault diagnosis using vibration and sound signals

Bearing faults are some of the most common causes of malfunctions and damage of rotating machinery. In this paper, a laboratory test rig is used to monitor the condition of ball bearings with seeded single point defects on inner race and outer race parts. Vibration acceleration signals and sound signals are collected under different radial load and shaft speed conditions. The acquired data is analyzed using wavelet scalogram, instantaneous power spectrum, and short-time frequency plots. Extensive laboratory investigation shows that both vibration and sound signals provide high sensitivity with respect to bearing faults. It is also concluded that all three analytical techniques successfully recognized patterns relevant to each type of bearing defect.