Dynamic Recrystallization Grain Identification for a Duplex-Phase Titanium Alloy Based on a Machine Learning Method

IF 4 3区材料科学 Q2 MATERIALS SCIENCE, MULTIDISCIPLINARY

Metals and Materials International Pub Date : 2025-03-07 DOI:10.1007/s12540-025-01925-8

Shuai Zhang, Haoyu Zhang, Jie Sun, Liyuan Yan, Chuan Wang, Ge Zhou, Lijia Chen

{"title":"Dynamic Recrystallization Grain Identification for a Duplex-Phase Titanium Alloy Based on a Machine Learning Method","authors":"Shuai Zhang, Haoyu Zhang, Jie Sun, Liyuan Yan, Chuan Wang, Ge Zhou, Lijia Chen","doi":"10.1007/s12540-025-01925-8","DOIUrl":null,"url":null,"abstract":"<div><p>Single-pass isothermal compression experiments for a duplex-phase titanium alloy (Ti–10V–5Al–2.5Fe–0.1B alloy) at 800–920 °C and strain rate of 0.0005–0.05 s<sup>−1</sup> were carried out. A dynamic recrystallization (DRX) grain identification method based on the optimized random forest model by sparrow search algorithm (SSA-RF) was used to identify DRX grains in the microstructure and realize the prediction of the degree of DRX in the microstructure under the new parameters. To verify the predictive ability of the SSA-RF model, the Extreme Gradient Boosting (XGBoost) model is introduced to identify DRX grains, and the prediction results are compared with those of the SSA-RF model. After statistically calculating the identified DRX grains, the research results show that the average relative absolute errors between the true and predicted values for DRX fraction and average DRX grain size are 9.43% and 14.04%, respectively, which are lower than those of the XGBoost model. The SSA-RF model has a higher precision in identifying DRX grains and predicting the degree of DRX at the new process parameters. In order to realize the prediction of DRX degree under new process parameters without experimental data, new data was constructed as the predicting data for the SSA-RF model based on the average grain size and the average length–diameter ratio data predicted by the SSA-RF model. The predicted results show that the errors between the predicted and true values for average DRX grain size and DRX fraction are 4.21 μm and 15.53%, respectively. The predicted results give a good indication of the true degree of DRX.</p><h3>Graphical Abstract</h3>\n<div><figure><div><div><picture><source><img></source></picture></div></div></figure></div></div>","PeriodicalId":703,"journal":{"name":"Metals and Materials International","volume":"31 10","pages":"2931 - 2945"},"PeriodicalIF":4.0000,"publicationDate":"2025-03-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Metals and Materials International","FirstCategoryId":"88","ListUrlMain":"https://link.springer.com/article/10.1007/s12540-025-01925-8","RegionNum":3,"RegionCategory":"材料科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"MATERIALS SCIENCE, MULTIDISCIPLINARY","Score":null,"Total":0}

引用次数: 0

Abstract

Single-pass isothermal compression experiments for a duplex-phase titanium alloy (Ti–10V–5Al–2.5Fe–0.1B alloy) at 800–920 °C and strain rate of 0.0005–0.05 s⁻¹ were carried out. A dynamic recrystallization (DRX) grain identification method based on the optimized random forest model by sparrow search algorithm (SSA-RF) was used to identify DRX grains in the microstructure and realize the prediction of the degree of DRX in the microstructure under the new parameters. To verify the predictive ability of the SSA-RF model, the Extreme Gradient Boosting (XGBoost) model is introduced to identify DRX grains, and the prediction results are compared with those of the SSA-RF model. After statistically calculating the identified DRX grains, the research results show that the average relative absolute errors between the true and predicted values for DRX fraction and average DRX grain size are 9.43% and 14.04%, respectively, which are lower than those of the XGBoost model. The SSA-RF model has a higher precision in identifying DRX grains and predicting the degree of DRX at the new process parameters. In order to realize the prediction of DRX degree under new process parameters without experimental data, new data was constructed as the predicting data for the SSA-RF model based on the average grain size and the average length–diameter ratio data predicted by the SSA-RF model. The predicted results show that the errors between the predicted and true values for average DRX grain size and DRX fraction are 4.21 μm and 15.53%, respectively. The predicted results give a good indication of the true degree of DRX.

Graphical Abstract

Abstract Image

查看原文本刊更多论文

基于机器学习方法的双相钛合金动态再结晶晶粒识别

对双相钛合金（Ti-10V-5Al-2.5Fe-0.1B合金）进行了800-920℃、应变速率0.0005-0.05 s−1的单道次等温压缩实验。采用基于优化随机森林模型的麻雀搜索算法（SSA-RF）动态再结晶（DRX）晶粒识别方法，识别微观组织中的DRX晶粒，实现新参数下微观组织中DRX程度的预测。为了验证SSA-RF模型的预测能力，引入了极端梯度增强（XGBoost）模型来识别DRX颗粒，并将预测结果与SSA-RF模型进行了比较。对识别的DRX颗粒进行统计计算后，研究结果表明，DRX分数和平均DRX粒度的真实值与预测值的平均相对绝对误差分别为9.43%和14.04%，均低于XGBoost模型。在新工艺参数下，SSA-RF模型在识别DRX晶粒和预测DRX程度方面具有较高的精度。为了在没有实验数据的情况下实现对新工艺参数下DRX度的预测，基于SSA-RF模型预测的平均晶粒尺寸和平均长径比数据，构建了新的数据作为SSA-RF模型的预测数据。预测结果表明，平均DRX晶粒尺寸和DRX分数预测值与真实值的误差分别为4.21 μm和15.53%。预测结果很好地指示了DRX的真实程度。图形抽象

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

Metals and Materials International 工程技术-材料科学：综合

CiteScore

7.10

自引率

8.60%

发文量

197

审稿时长

3.7 months

期刊介绍： Metals and Materials International publishes original papers and occasional critical reviews on all aspects of research and technology in materials engineering: physical metallurgy, materials science, and processing of metals and other materials. Emphasis is placed on those aspects of the science of materials that are concerned with the relationships among the processing, structure and properties (mechanical, chemical, electrical, electrochemical, magnetic and optical) of materials. Aspects of processing include the melting, casting, and fabrication with the thermodynamics, kinetics and modeling.