Predicting grain growth kinetic in steels using machine learning and XAI for mechanical properties

Selim Demirci, Durmuş Özkan Şahin, Sercan Demirci, Mehmet Masum Tünçay and Moataz M Attallah

PLOS ONE, 2026, vol. 21, issue 1, 1-24

Abstract: Understanding and controlling grain growth kinetics in steels is crucial for optimizing mechanical properties during thermomechanical processing. However, traditional empirical models often fail to account for the complex, nonlinear interactions between alloying elements and processing parameters. In this study, we introduce a novel machine learning (ML) based framework that predicts austenitic grain growth behaviour directly from chemical composition and process conditions, utilizing a comprehensive dataset of 1039 experimentally validated samples. Among various algorithms tested, the XGBoost model demonstrated exceptional predictive capability, achieving an R2 value of 0.9728 after hyperparameter optimization. Feature selection methods (Pearson correlation, CfsSubset, ReliefF) and SHAP-based explainable AI analyses were employed to identify the most influential parameters, revealing temperature, initial grain size, and holding time as dominant factors. Experimental validation was conducted on 316L stainless steel samples annealed at 1100 °C. The predicted grain sizes showed strong agreement with experimental measurements, and the observed hardness variations followed the expected Hall–Petch behaviour. This study demonstrates the first integrated ML and experimental approach for predicting grain growth kinetics in steels, offering a powerful tool for alloy design and process optimization. Future work will extend this framework to additional process variables and alloy systems.

Date: 2026
References: View references in EconPapers View complete reference list from CitEc
Citations:

Downloads: (external link)
https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0341053 (text/html)
https://journals.plos.org/plosone/article/file?id= ... 41053&type=printable (application/pdf)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:plo:pone00:0341053

DOI: 10.1371/journal.pone.0341053

Access Statistics for this article

More articles in PLOS ONE from Public Library of Science
Bibliographic data for series maintained by plosone ().