[1]刘芳宁,孙瑞侠,王 越,等.doi: 10.3969/j.issn.1001-3849.2026.02.011基于可解释性机器学习的铸造铝硅合金成分-工艺[J].电镀与精饰,2026,(02):94-104.
　LIU Fangning,SUN Ruixia,WANG Yue,et al.Interpretable machine learning-based modeling of composition-processing interaction in cast Al-Si alloys[J].Plating & Finishing,2026,(02):94-104.
点击复制

doi: 10.3969/j.issn.1001-3849.2026.02.011基于可解释性机器学习的铸造铝硅合金成分-工艺()

《电镀与精饰》[ISSN:1001-3849/CN:12-1096/TG]

卷:
期数:
2026年02
页码:
94-104
栏目:
出版日期:
2026-02-28

文章信息/Info

Title:
Interpretable machine learning-based modeling of composition-processing interaction in cast Al-Si alloys
作者:
刘芳宁; 孙瑞侠; 王 越; 赵 晋
耦合机制建模刘芳宁,孙瑞侠,王 越,赵 晋(中国航发北京航空材料研究院,北京 100095)
Author(s):
LIU Fangning;  SUN Ruixia;  WANG Yue;  ZHAO Jin
(Beijing Institute of Aeronautical Materials, Beijing 100095, China)
关键词:
机器学习; 极限梯度提升; 沙普利加和解释; 局部可解释模型无关方法; 铝硅合金
Keywords:
machine learning;  XGBoost;  SHAP;  LIME;  Al-Si alloy
分类号:
TG136, TQ050.4
文献标志码:
A
摘要:
铝硅系铸造合金因其良好的高温性能与铸造成形能力，被广泛应用于航空发动机热端部件。为深入理解成分与工艺参数对力学性能的耦合影响，并提升传统经验模型的预测能力，提出了一种基于可解释性机器学习的建模方法。研究基于378组实验数据，构建了涵盖Si、Cu、Mg、Mn、Zn等成分变量及固溶处理方式、冷却方式、时效温度等工艺参数的综合数据集。分别训练并对比了极限梯度提升(XGBoost)、随机森林回归(RF)、支持向量回归(SVR)和K-最近邻回归(KNR) 4种模型，并通过交叉验证评估其在抗拉强度与延伸率预测中的性能表现。结果表明，极限梯度提升模型在3项性能指标上均优于其他模型，其中抗拉强度的决定系数达0.87，平均绝对误差(MAE)为10.34?MPa，均方根误差(RMSE)为7.99?MPa；延伸率的决定系数为0.95，平均绝对误差为0.4%，均方根误差为0.65%。进一步通过SHAP方法分析特征贡献及交互作用发现，Mg和Cu的强化效应均受时效温度调控，表现出与传统材料学中析出强化机制一致的非线性规律。结合局部可解释模型无关方法(LIME)进行局部解释验证后，模型在个体样本层面亦展现出良好的稳定性与一致性，进一步确认了本研究在构建可解释性预测模型及揭示成分-工艺协同机制方面的有效性。
Abstract:
Cast Al-Si alloys are widely used in hot-section components of aircraft engines due to their excellent high-temperature performance and castability. To better understand the coupled effects of composition and processing parameters on mechanical properties and improve the predictive capability of traditional empirical models, this study proposes an interpretable machine learning-based modeling approach. A dataset comprising 378 experimental records was compiled, incorporating compositional variables such as Si, Cu, Mg, Mn, and Zn, as well as processing parameters including solution treatment method, cooling method, and aging temperature. Four regression models, including extreme gradient boosting (XGBoost), random forest (RF), support vector regression (SVR) and K-nearest neighbors regression (KNR) were trained and evaluated using cross-validation to assess their performance in predicting ultimate tensile strength (UTS) and elongation (EL). Results show that the XGBoost model outperform others across all 3 evaluation metrics. Specifically, the coefficient of determination ( R2) for UTS reaches 0.87, with a mean absolute error (MAE) of 10.34?MPa and a root mean square error (RMSE) of 7.99?MPa; for EL, the R2 is 0.95, with an MAE of 0.4% and an RMSE of 0.65%. Further analysis using shapley additive explanations (SHAP) method reveal that the strengthening effects of Mg and Cu are both regulated by aging temperature, exhibiting nonlinear trends that align with known precipitation strengthening mechanisms in traditional materials science. Local interpretable model-agnostic explanations (LIME) method is also applied for local interpretability verification, demonstrating strong consistency and stability at the individual sample level. These results confirm the effectiveness of the proposed interpretable modeling approach in both property prediction and revealing the underlying composition–process synergy in cast Al-Si alloys.

参考文献/References:

[1].禹露, 许婄鑫, 罗志翔. A356铝合金的表面改性与耐磨及耐蚀性能研究[J]. 电镀与精饰, 2024, 46(4): 29-37.
[2].ZHU L, ZHANG Y, LUO Q, et al. Effectively refining Al-10Si alloy via Al-Ti-Nb-B refiner with Nb2O5[J]. Journal of Materials Science & Technology, 2024, 176(9): 204-210.
[3].吴忠. 汽车用ZL101A铝合金表面Ni-Co合金镀层的硬度[J]. 电镀与精饰, 2018, 40(11): 5-9.
[4].YUN M, JUNG I-H. Development of a rapid solidification model for additive manufacturing process and application to Al-Si alloy[J]. Acta Materialia, 2024, 265: 119638.
[5].戴圣龙, 张坤, 杨守杰, 等. 先进航空铝合金材料与应用[M]. 北京: 国防工业出版社, 2012: 143-144.
[6].薛利文. 高强塑性Al-Si合金微观组织调控及强韧化机制[D]. 长春: 吉林大学, 2024.
[7].CAI Q, MENDIS C L, CHANG I T H, et al. Effect of short T6 heat treatment on the microstructure and the mechanical properties of newly developed die-cast Al-Si-Mg-Mn alloys[J]. Materials Science and Engineering: A, 2020, 788: 139610.
[8].邱谨, 曹杨婧, 罗晓梅, 等. A356铝合金挤压铸造凝固过程及热处理对其性能的影响[J]. 特种铸造及有色合金, 2024, 44(2): 189-197.
[9].BAIRD S G, LIU M, SAYEED H M, et al. Data-driven materials discovery and synthesis using machine learning methods[G]//Reedijk J, Poeppelmeier K R. Comprehensive Inorganic Chemistry III. Amsterdam: Elsevier, 2023: 3-23.
[10].FENG X, WANG Z, JIANG L, et al. Simultaneous enhancement in mechanical and corrosion properties of Al-Mg-Si alloys using machine learning[J]. Journal of Materials Science & Technology, 2023, 167(36): 1-13.
[11].LU Z, KAPOOR I, LI Y, et al. Machine learning driven design of high-performance Al alloys[J]. Journal of Materials Informatics, 2024, 4: 19.
[12].ARUNA D M, PRAKASH C P S, CHINNANNAVAR R P, et al. An informatic approach to predict the mechanical properties of aluminum alloys using machine learning techniques[C]// Proceedings of the 2020 International Conference on Smart Electronics and Communication (ICOSEC). Trichy: IEEE, 2020: 1576-1581.
[13].FU K, ZHU D, ZHANG Y, et al. Predictive modeling of tensile strength in aluminum alloys via machine learning[J]. Materials, 2023, 16(22): 7236.
[14].LUNDBERG S M, LEE S I.A unified approach to interpreting model predictions[J]. Advances in Neural Information Processing Systems, 2017, 30: 4765-4774.
[15].姜磊. 数据驱动的新一代高强高韧铝合金研究[D]. 北京: 北京科技大学, 2024.
[16].RIBEIRO M T, SINGH S, GUESTRIN C. “Why should I trust you?”: Explaining the predictions of any classifier[C] // Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. San Francisco: ACM, 2016: 1135-1144.
[17].张福全, 蒋安定, 何海琪. 发动机缸盖用铸造铝硅合金的成分优化与组织性能研究[J]. 湖南大学学报(自然科学版), 2013, 40(12): 84-89.
[18].蒋安定. 铸造铝硅合金成分优化及其组织性能研究[D]. 长沙: 湖南大学, 2013.
[19].李飞龙, 苏广才, 黄万鹏, 等. 改良铸造铝硅镁合金的成分优化设计[J]. 铝加工, 2011(3): 33-37.
[20].华雄飞. 微合金化对全铝发动机高强度铸造Al-Si合金性能和组织的影响[D]. 镇江: 江苏大学, 2010.
[21].雷健, 郑宇新, 袁序弟, 等. 轿车发动机缸盖铸态铝合金成分优化设计[J]. 特种铸造及有色合金, 2001(3): 24-25, 2.
[22].郑戬. 新型高强、高耐腐蚀铸造铝合金的研究[D]. 上海: 上海交通大学, 2015.

更新日期/Last Update: 2026-02-09