Skip to Content
You are currently on the new version of our website. Access the old version .
MaterialsMaterials
  • Article
  • Open Access

9 November 2026

24 Pages

Burr Detection and Edge Condition in Ceramic Brushing of Glass and Carbon Fabric-Reinforced Composites Using Recurrence Analysis of Force Signals: A Feasibility Study

and
Department of Production Engineering, Mechanical Engineering Faculty, Lublin University of Technology, Nadbystrzycka 36, 20-618 Lublin, Poland
*
Author to whom correspondence should be addressed.

Abstract

Burrs formed during the machining of fibre-reinforced polymer composites reduce safety and product quality and are usually inspected off-process. This study assessed whether as-milled edge segments with burrs can be distinguished from burr-free chamfered segments on the basis of the force signal recorded during edge brushing, as a first step towards in-process burr detection, and how brush fibre stiffness affects the edge condition. On one edge of glass (GFRP) and carbon (CFRP) fabric-reinforced specimens, positioned at 45° to the tool axis, chamfered segments 0.1–0.4 mm deep, representing a burr-free edge, alternated with as-milled segments with burrs. The edges were brushed with four ceramic-fibre brushes of increasing stiffness (S1–S4) at constant parameters, and the force signal was analysed using recurrence plots (RP) and recurrence quantification analysis (RQA). Chamfers 0.1 mm deep did not produce unambiguous changes in the recurrence structure. For the stiff brushes (S3, S4), chamfers 0.2–0.4 mm deep produced local RP densifications consistent with their location. In the RQA, for chamfers ≥ 0.2 mm, DET (determinism) and L (average diagonal line length) mostly rose above and LAM(laminarity) and TT (trapping time) fell below the burred-edge reference levels. The chamfer width produced by brushing increased with fibre stiffness, from 588 to 1040 µm (GFRP) and from 389 to 1005 µm (CFRP). Under the tested conditions, the force signal distinguished the two edge conditions; its use for detecting burrs themselves requires further validation on edges with quantified burrs.

Article Metrics

Citations

Article Access Statistics

Multiple requests from the same IP address are counted as one view.