Abstract
Tribological materials play a critical role in high-end fields such as aerospace, rail transportation, manufacturing processes and energy equipment. However, their friction and wear behaviors are highly complex, governed by the coupled effects of multi-scale and multi-physics interactions. Traditional research approaches, which rely heavily on experiments and empirical knowledge, are costly, time-consuming, and often insufficient to uncover the underlying mechanisms. With the rapid development of artificial intelligence (AI), machine learning (ML) has emerged as a powerful tool in tribology research, leading to the introduction of the concept of triboinformatics. This review systematically summarizes the methodological framework of ML in tribology, covering data acquisition, feature selection, model selection and optimization, performance evaluation and validation, as well as result interpretation. Key methodological advances, including data augmentation, physicsinformed approaches, and interpretability techniques, are emphasized and systematically discussed. Representative applications demonstrate that data-driven interpretable learning, multi-source data fusion, and physicsinformed ML (PIML) models can enhance prediction accuracy and robustness while revealing fundamental tribological mechanisms. Moreover, ML has been widely applied to the design and performance optimization of novel tribological materials, accelerating the materials development cycle. Finally, the future prospects of triboinformatics are outlined, highlighting high-quality database construction, physics-driven integration, interpretability enhancement, and small-sample learning as key research trends. This review aims to provide a systematic methodological reference for tribology research and to promote the paradigm shift from experience-driven to data-mechanism integrated approaches.
| Original language | English |
|---|---|
| Pages (from-to) | 243-270 |
| Number of pages | 28 |
| Journal | Journal of Manufacturing Processes |
| Volume | 159 |
| Early online date | Jan 2026 |
| DOIs | |
| State | Published - 15 Feb 2026 |
Funding
This work was supported by the National Natural Science Foundation of China (Grant No. 12272392, No. 11790292) , the Strategic Priority Research Program of the Chinese Academy of Sciences (Grant No. XDB0620301) , the Youth Innovation Promotion Association of the Chinese Academy of Sciences (No. 2022020) and Yip Kee Sun Founda-tion (Grant No. U2441207) .
Indexed by
- EI
- SCIE
- Scopus
Fingerprint
Dive into the research topics of 'Machine learning in tribology: A review on framework, case studies, and future perspectives'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver