Recurrence analysis of coefficient of friction oscillations in DLC-coated and non-coated Ti6Al4V titanium alloy. (January 2022)

• Record Type: Journal Article
• Authors: Łępicka, Magdalena; Grądzka-Dahlke, Małgorzata; Zaborowska, Iwona; Górski, Grzegorz; Mosdorf, Romuald
• Is Part Of: Tribology international. Volume 165(2022)
• Publication Date: 2022-01
• Subjects: Recurrence quantification analysis -- Self-organizing map -- Diamond-like-carbon -- Titanium alloy; Tribology -- Periodicals; 621.89
• DOI: 10.1016/j.triboint.2021.107342
• Languages: English
• ISSNs: 0301-679X

Abstract

Abstract: Though numerous analytical techniques are used to assess performance of surface-modified metals, identification of dominant wear modes, as well as transitions between them, is challenging. Nevertheless, the works devoted to analysis the COF oscillations are still scarce. In this paper, Recurrence Quantification Analysis, Principal Component Analysis and Self-Organizing Map neural network are proposed as tools for analysis of COF time series acquired during tribological tests of the non-coated and DLC-coated Ti6Al4V. The proposed approach can be used to recognize the dominant wear modes present during friction and the transitions between them, as well as to spot the frictional damage of the protective film. Highlights: Coefficient of friction (COF) signals, registered during a dry wear test, were analyzed with RQA, PCA and SOM. As tribological pairs, non-coated or DLC-coated Ti6Al4V vs. WC-Co ball were selected. The aim was to explore the possibilities of recurrence analysis in tribological signals. Transitions between the wear modes, as well as failure of a protective film, can be spotted with RQA, PCA and SOM. In COF analyses, SOM is far more accurate than PCA.