Intelligent design of Fe–Cr–Ni–Al/Ti multi-principal element alloys based on machine learning

Multi-principal element alloys (MPEAs), distinguished by their complex compositions and exceptional mechanical properties, pose significant challenges for conventional predictive approaches in mechanical property optimization. This study proposes an innovative intelligent optimization algorithm (OA)...

Full description

Saved in:
Bibliographic Details
Main Authors: Kang Xu, Zhengming Sun, Jian Tu, Wenwang Wu, Huihui Yang
Format: Article
Language:English
Published: Elsevier 2025-03-01
Series:Journal of Materials Research and Technology
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S2238785425005666
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850031800253939712
author Kang Xu
Zhengming Sun
Jian Tu
Wenwang Wu
Huihui Yang
author_facet Kang Xu
Zhengming Sun
Jian Tu
Wenwang Wu
Huihui Yang
author_sort Kang Xu
collection DOAJ
description Multi-principal element alloys (MPEAs), distinguished by their complex compositions and exceptional mechanical properties, pose significant challenges for conventional predictive approaches in mechanical property optimization. This study proposes an innovative intelligent optimization algorithm (OA) to refine feature selection in machine learning (ML) models, targeting the prediction of ultimate tensile strength (UTS) and fracture elongation (FE) in MPEAs. Comparative analysis with genetic algorithms (GA) reveals that the OA achieves high computational efficiency and improved prediction accuracy, demonstrating superior convergence rates and feature recognition capability. Additionally, low-cost experimental data from Fe–Cr–Ni–Al/Ti alloys are used to revise the model, thereby enhancing its prediction accuracy for high-cost processes. Validation experiments on Al5(Fe10Cr35Ni55)95 and Al2Ti1(Fe10Cr35Ni55)97 alloys yielded UTS prediction errors of 7.80% and 2.56%, with corresponding FE errors of 0.35% and 1.09%. These results confirm that strategic integration of experimental data within defined compositional ranges can substantially improve ML model performance.
format Article
id doaj-art-2e648758dcff4cacb15c8e5e4380d712
institution DOAJ
issn 2238-7854
language English
publishDate 2025-03-01
publisher Elsevier
record_format Article
series Journal of Materials Research and Technology
spelling doaj-art-2e648758dcff4cacb15c8e5e4380d7122025-08-20T02:58:51ZengElsevierJournal of Materials Research and Technology2238-78542025-03-01356864687310.1016/j.jmrt.2025.03.056Intelligent design of Fe–Cr–Ni–Al/Ti multi-principal element alloys based on machine learningKang Xu0Zhengming Sun1Jian Tu2Wenwang Wu3Huihui Yang4School of Materials Science and Engineering, Southeast University, Nanjing, 211189, China; School of Material Science and Engineering, Chongqing University of Technology, Chongqing, 400054, China; Suzhou Laboratory, No.388, Ruoshui Street, SIP, Jiangsu, 215123, ChinaSchool of Materials Science and Engineering, Southeast University, Nanjing, 211189, China; Corresponding author.School of Material Science and Engineering, Chongqing University of Technology, Chongqing, 400054, China; Corresponding author.Suzhou Laboratory, No.388, Ruoshui Street, SIP, Jiangsu, 215123, China; Corresponding author.Suzhou Laboratory, No.388, Ruoshui Street, SIP, Jiangsu, 215123, ChinaMulti-principal element alloys (MPEAs), distinguished by their complex compositions and exceptional mechanical properties, pose significant challenges for conventional predictive approaches in mechanical property optimization. This study proposes an innovative intelligent optimization algorithm (OA) to refine feature selection in machine learning (ML) models, targeting the prediction of ultimate tensile strength (UTS) and fracture elongation (FE) in MPEAs. Comparative analysis with genetic algorithms (GA) reveals that the OA achieves high computational efficiency and improved prediction accuracy, demonstrating superior convergence rates and feature recognition capability. Additionally, low-cost experimental data from Fe–Cr–Ni–Al/Ti alloys are used to revise the model, thereby enhancing its prediction accuracy for high-cost processes. Validation experiments on Al5(Fe10Cr35Ni55)95 and Al2Ti1(Fe10Cr35Ni55)97 alloys yielded UTS prediction errors of 7.80% and 2.56%, with corresponding FE errors of 0.35% and 1.09%. These results confirm that strategic integration of experimental data within defined compositional ranges can substantially improve ML model performance.http://www.sciencedirect.com/science/article/pii/S2238785425005666Alloy designMachine learningFeature selectionMechanical properties
spellingShingle Kang Xu
Zhengming Sun
Jian Tu
Wenwang Wu
Huihui Yang
Intelligent design of Fe–Cr–Ni–Al/Ti multi-principal element alloys based on machine learning
Journal of Materials Research and Technology
Alloy design
Machine learning
Feature selection
Mechanical properties
title Intelligent design of Fe–Cr–Ni–Al/Ti multi-principal element alloys based on machine learning
title_full Intelligent design of Fe–Cr–Ni–Al/Ti multi-principal element alloys based on machine learning
title_fullStr Intelligent design of Fe–Cr–Ni–Al/Ti multi-principal element alloys based on machine learning
title_full_unstemmed Intelligent design of Fe–Cr–Ni–Al/Ti multi-principal element alloys based on machine learning
title_short Intelligent design of Fe–Cr–Ni–Al/Ti multi-principal element alloys based on machine learning
title_sort intelligent design of fe cr ni al ti multi principal element alloys based on machine learning
topic Alloy design
Machine learning
Feature selection
Mechanical properties
url http://www.sciencedirect.com/science/article/pii/S2238785425005666
work_keys_str_mv AT kangxu intelligentdesignoffecrnialtimultiprincipalelementalloysbasedonmachinelearning
AT zhengmingsun intelligentdesignoffecrnialtimultiprincipalelementalloysbasedonmachinelearning
AT jiantu intelligentdesignoffecrnialtimultiprincipalelementalloysbasedonmachinelearning
AT wenwangwu intelligentdesignoffecrnialtimultiprincipalelementalloysbasedonmachinelearning
AT huihuiyang intelligentdesignoffecrnialtimultiprincipalelementalloysbasedonmachinelearning