Comparative merits of acceleration and acoustic emission data in the preventive maintenance of antifriction bearings

It is an ages long art that engineers and machine operators get clues from the sound produced by any machine on whether the machine is operating properly or not. Drivers listen to the sound of their engine and doctors listen to the heart beats and pulmonary sounds. In machinery, well being of anti-friction bearings is extremely important as a bearing breaking into pieces may generate irreparable damage. Although life time of antifriction bearings are estimable from the working conditions and statistical data, these are never accurate. Therefore it is very important to be able to detect a bearing moving towards a failure. Old bearings have greater clearance due to wear, yet, fatigue generates pittings of various sizes on the tracks where rolling elements move. These effects generate vibration and sound. Actually vibration and sound are almost synonymous in this respect, as sound is generated by the vibration of the machine. Sound and vibration recordings can be done by microphones and accelerometers, but the task of differentiation of the relevant data from the others is great. Some of the vibration certainly decays in the material of the machine components, whereas sound they produce may be beyond the threshold of the microphones. Both microphones and accelerometers add their own dynamic problems onto the recorded signal. This dissertation makes a survey on the use of acceleration and acoustic emission data in this respect. Then data taken from a test rig which contains a purposefully indented ball bearing under load and how they interpret the known fault is examined by using routines from Labview® Toolbox. Lastly, a comparison of the advantages and disadvantages of both are presented.