Abstract
This paper describes a new automatic method of detecting flaws in ball bearings and classifying them into four categories: normal, flaw on the inner ring, flaw on the outer ring, and flaw on the balls. Detection and classification are currently carried out mainly aurally by inspectors who listen to a vibration signal obtained by a vibration pick-up in an Anderon meter, which is attached to the outer ring of a ball bearing while the inner ring rotates at a uniform speed. There have been some proposals for the automatic detection and classification of flaws which are based on the analysis of the periodicity of vibration pulses excited by flaws. When flaws are slight, however, it is hard to detect them by recently proposed methods because of the following two reasons:(A) the power of the vibration due to flaws is less than that of noise, and (B) the period of flaw pulses is irregular. We have developed a new method, though, by which slight flaws can be detected. We have applied this new method in the flaw detection of small-sized ball bearings: flaws were classified with an accuracy rate of greater than 97% into the above four categories.
