Scholarlink

<p>This review critically discusses the emerging area of machine learning (ML)- based prediction of tribological properties for thermal spray coatings, focusing on literature from 2019 to 2024 and supplemented by key recent publications from 2025. Due to the high dimensionality and complexity of these materials, ML is a suitable tool for providing a comprehensive analysis of process parameters, material properties, operational conditions, and coating attributes. Also, traditional experimental optimization is slow and costly, and the data is complex. The paper describes how various machine learning techniques, including artificial neural networks (ANNs), support vector machines (SVMs), and ensemble methods, have been effectively applied to predict essential attributes such as wear, friction, hardness, corrosion, and surface roughness for several types of coatings. However, ML offers a promising path for prediction, but its application requires a critical assessment to guide future research. The review presents the state of the art, highlighting the high accuracy and performance reported in the literature. Moreover, it considers critical difficulties, such as data scarcity, model explication, and multi-objective optimization. An outlook is given about physics-based model integrated ML, such as physics-informed neural networks (PINNs) and surrogate models, the use of FAIR data standards, and the synergy between ML and high-throughput experimental techniques, to guide and optimize experimental efforts and aid coating design and its performance enhancement.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A Critical Review of Tribological Property Prediction of Thermally Sprayed Coatings Using Machine Learning Approaches

  • 期刊论文

  • 发表日期：

    2025年12月8日

查看全文
  • Sourabh Mandol,
  • Peerawatt Nunthavarawong

摘要

This review critically discusses the emerging area of machine learning (ML)- based prediction of tribological properties for thermal spray coatings, focusing on literature from 2019 to 2024 and supplemented by key recent publications from 2025. Due to the high dimensionality and complexity of these materials, ML is a suitable tool for providing a comprehensive analysis of process parameters, material properties, operational conditions, and coating attributes. Also, traditional experimental optimization is slow and costly, and the data is complex. The paper describes how various machine learning techniques, including artificial neural networks (ANNs), support vector machines (SVMs), and ensemble methods, have been effectively applied to predict essential attributes such as wear, friction, hardness, corrosion, and surface roughness for several types of coatings. However, ML offers a promising path for prediction, but its application requires a critical assessment to guide future research. The review presents the state of the art, highlighting the high accuracy and performance reported in the literature. Moreover, it considers critical difficulties, such as data scarcity, model explication, and multi-objective optimization. An outlook is given about physics-based model integrated ML, such as physics-informed neural networks (PINNs) and surrogate models, the use of FAIR data standards, and the synergy between ML and high-throughput experimental techniques, to guide and optimize experimental efforts and aid coating design and its performance enhancement.