The Role of Machine Learning in Tribology: A Systematic Review

The machine learning (ML) approach, motivated by artificial intelligence (AI), is an inspiring mathematical algorithm that accurately simulates many engineering processes. Machine learning algorithms solve nonlinear and complex relationships through data training; additionally, they can infer previously unknown relationships, allowing for a simplified model and estimation of hidden data. Unlike other statistical tools, machine learning does not impose process parameter restrictions and yields an accurate association between input and output parameters. Tribology is a branch of surface science concerned with studying and managing friction, lubrication, and wear on relatively interacting surfaces. While AI-based machine learning approaches have been adopted in tribology applications, modern tribo-contact simulation requires a deliberate decomposition of complex design challenges into simpler sub-threads, thereby identifying the relationships between the numerous interconnected features and processes. Numerous studies have established that artificial intelligence techniques can accurately model tribological processes and their properties based on various process parameters. The primary objective of this review is to conduct a thorough examination of the role of machine learning in tribological research and pave the way for future researchers by providing a specific research direction. In terms of future research directions and developments, the expanded application of artificial intelligence and various machine learning methods in tribology has been emphasized, including the characterization and design of complex tribological systems. Additionally, by combining machine learning methods with tribological experimental data, interdisciplinary research can be conducted to understand efficient resource utilization and resource conservation better. At the conclusion of this article, a detailed discussion of the limitations and future research opportunities associated with implementing various machine learning algorithms in tribology and its interdisciplinary fields is presented.