Novel friction stabilization technology for surface damage conditions using machine learning

• Published in: Tribology international Vol. 180; p. 108280
• Main Authors: Murashima, Motoyuki, Yamada, Takazumi, Umehara, Noritsugu, Tokoroyama, Takayuki, Lee, Woo-Young
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01-02-2023
• Subjects: Active friction control; Artificial intelligence; Genetic algorithm; Machine learning; Morphing surface; Active friction control; Genetic algorithm; Artificial intelligence; Machine learning; Morphing surface
• ISSN: 0301-679X; 1879-2464
• DOI: 10.1016/j.triboint.2023.108280

The surface damage is a serious cause of failure in tribosystems. In the present paper, we propose a new damage avoidance method that combines a contact position control system (e.g., morphing surface) and artificial-intelligence-based control (e.g., genetic algorithm: GA) to achieve stable friction and long life of sliding surfaces. In the case of the single-damage condition, the GA sequentially avoided contact with the damaged position, and finally complete damage avoidance was achieved. In the multiple-damage condition, we confirmed that learning by GA effectively stabilized friction, although the learning time was longer. In summary, the contact-position control method should provide new capabilities on real machine surfaces where unexpected damage occurs.
•A new active friction control technology was developed to avoid contact with surface damage.•The avoidance was realized using contact-position control surface (e.g., morphing surface).•Damaged position on the mating surface was detected using artificial intelligence.