Wear particle analysis-utilization of quantitative computer image analysis: A review

This paper provides a general overview of developments and progress in quantitative computer image analysis as applied to wear particle identification/classification technology, over the last two decades. Since many technical disciplines are involved in this 'infant-stage' technical area, an attempt is made to put into perspective mechanical failure prediction/diagnosis and prevention through quantitative wear particle morphological analysis. The problems experienced with applying conventional wear particle analysis methods in machinery condition monitoring, notably the employment of wear debris morphological diagnostic systems, revealed that it is not prudent to rely solely on human interpretation in the analysis of 'filtergram' slides. This has highlighted the need for improving the provision of 'intelligent' objective methods for performing this type of analysis. In this paper, some of the developments reported in the literature relating to progress made with wear particle image analysis are reported and examined as a basis for establishing improved methods of diagnostic analysis.