Modeling Material Machining Conditions with Gear-Shaper Cutters with TiN0.85-Ti in Adhesive Wear Dominance Using Machine Learning Methods

Maciej Kupczyk; Michał Leleń; Jerzy Józwik; Paweł Tomiło

doi:10.3390/ma17225567

Modeling Material Machining Conditions with Gear-Shaper Cutters with TiN_0.85-Ti in Adhesive Wear Dominance Using Machine Learning Methods

Materials (Basel). 2024 Nov 14;17(22):5567. doi: 10.3390/ma17225567.

Authors

Maciej Kupczyk¹, Michał Leleń², Jerzy Józwik², Paweł Tomiło³

Affiliations

¹ Institute of Mechanical Technology, Poznan University of Technology, 3 Piotrowo Street, 60-965 Poznan, Poland.
² Faculty of Mechanical Engineering, Lublin University of Technology, 36 Nadbystrzycka Street, 20-618 Lublin, Poland.
³ Faculty of Management, Lublin University of Technology, 38 Nadbystrzycka Street, 20-618 Lublin, Poland.

Abstract

This paper examines the challenges of machining structural alloy steels for carburizing, with a particular focus on gear manufacturing. TiN_0.85-Ti coatings were applied to cutting tool blades to improve machining quality and tool life. The research, supported by mathematical modeling, demonstrated that these coatings significantly reduce adhesive wear and improve blade life. The Kolmogorov-Arnold Network (KAN) was identified as the most effective model comprehensively describing tool life as a function of cutting speed, coating thickness, and feed rate. The results indicate that gear production efficiency can be significantly increased using TiN_0.85-Ti coatings.

Keywords: Kolmogorov–Arnold Network; adhesive wear; carburizing steels; cutting tools; gear production; machining; predictive modeling; reactive pulse plasma method; titanium nitride coatings; tool durability.

Abstract

Grants and funding