Quantitative correlation of wear debris morphology: grouping and classification

Abstract

Considerable interest has been expressed in the quantitative analysis of wear debris to minimize the subjectivity of visual assessment of debris morphology. Recent works involve the quantification of wear debris morphology using numerical parameters. In this paper, more comprehensive quantitative analysis of wear debris is performed: wear debris morphology is quantified with numerical parameters and, furthermore, quantitative correlation is performed to demonstrate how specific statistical data analysis techniques can be applied to carry out grouping and classification of debris. Grouping and classification are multivariate statistical techniques that can be used to find out morphological groups of wear debris and to classify wear debris into the predefined wear conditions, respectively. It is shown that statistical data analysis can provide systematic quantitative correlation of wear debris without subjective individual judgment.