目的 利用机器学习模型对铸钢的力学性能进行精准预测,避免冗杂特征的干扰,探究有效的降维策略,对铸钢件抗拉强度、屈服强度和硬度数据进行降维,实现精准预测并提高预测效率。方法 基于铸钢的“材料组织成分和热处理工艺-力学性能”数据集,系统比较了Spearman相关性分析、Adaboost重要度分析及基于遗传算法的特征选择方法3种降维策略在采用组合决策树、随机森林、极限梯度提升算法（XGBoost）、BP神经网络、Stacking模型5类机器学习算法时的模型性能。结果 屈服强度、抗拉强度和硬度的决定系数在降维策略和机器学习模型的正交实验下的最大值分别为0.9、0.94、0.902。屈服强度预测精度最优组合为Spearman特征选择法+RF模型,抗拉强度预测精度最优组合为Spearman特征选择法+Stacking模型,硬度预测精度最优组合为遗传算法特征选择法+RF模型。RF特征重要度分析结果表明,C、Mn、Mo等元素及热处理工艺对铸钢力学性能影响显著。结论 本研究系统评估了不同降维策略与机器学习算法的组合效能,合理的降维策略搭配一定的算法模型能有效提高模型预测精度,为铸钢材料设计提供了重要理论依据,对推进铸造技术的智能化发展具有实践指导意义。
Abstract
The work aims to accurately predict the mechanical properties of cast steel using machine learning models, avoid interference from redundant features, and explore effective dimensionality reduction strategies for tensile strength, yield strength, and hardness data of steel castings, thereby achieving precise predictions and improving prediction efficiency. Based on a “material composition and heat treatment process-mechanical properties” dataset of cast steel, three dimensionality reduction strategies, namely, Spearman correlation analysis, Adaboost importance analysis, and genetic algorithm-based feature selection, combined with five machine learning algorithms of Decision Tree, Random Forest, Extreme Gradient Boosting (XGBoost), BP Neural Network, and Stacking models were systematically compared in terms of model performance. The results showed that the maximum coefficients of determination for yield strength, tensile strength, and hardness reached 0.9, 0.94, and 0.902, respectively, under the orthogonal experiments combining dimensionality reduction strategies and machine learning models. The optimal combination for predicting yield strength was the Spearman feature selection method paired with the Random Forest (RF) model. For tensile strength, the best prediction accuracy was achieved using the Spearman feature selection method combined with the Stacking model. In the case of hardness, the genetic algorithm-based feature selection method coupled with the RF model yielded the highest prediction accuracy. Additionally, RF-based feature importance analysis revealed that elements such as C, Mn, and Mo, along with heat treatment processes, significantly influenced the mechanical properties of cast steel. This study systematically evaluates the performance of various combinations of dimensionality reduction strategies and machine learning algorithms. The findings demonstrate that appropriate dimensionality reduction strategies, when paired with suitable algorithmic models, can effectively enhance prediction accuracy. This provides a crucial theoretical foundation for the design of cast steel materials and offers practical guidance for advancing the intelligent development of casting technology.
关键词
铸钢 /
力学性能预测 /
机器学习 /
降维策略 /
特征重要度
Key words
cast steel /
mechanical performance prediction /
machine learning /
dimensionality reduction strategy /
feature importance
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] 朱家辉, 秦广华, 张寅, 等. 铸钢国际标准的新进展[J]. 铸造, 2023, 72(12): 1657-1662.
ZHU J H, QIN G H, ZHANG Y, et al.New Progress of International Standards for Cast Steels[J]. Foundry, 2023, 72(12): 1657-1662.
[2] 王昌龙, 黄惟一. 用神经网络法预测集装箱角铸钢件的力学性能[J]. 铸造, 2002, 51(4): 233-235.
WANG C L, HUANG W Y.Predicting the Mechanical Properties of Cast Steel Container Corners with the Neural Network Methods[J]. Foundry, 2002, 51(4): 233-235.
[3] 李建国, 贾冬生, 赵银虎, 等. 稀土Ce对32MnMoNiCu铸钢微观组织与力学性能的影响[J]. 金属热处理, 2025, 50(1): 47-52.
LI J G, JIA D S, ZHAO Y H, et al.Effect of Rare Element Ce on Microstructure and Mechanical Properties of 32MnMoNiCu Cast Steel[J]. Heat Treatment of Metals, 2025, 50(1): 47-52.
[4] 李建国, 贾冬生, 赵银虎, 等. 回火温度对铸钢32MnMoNiCuCe微观组织和力学性能的影响[J]. 新技术新工艺, 2024(12): 50-57.
LI J G, JIA D S, ZHAO Y H, et al.Effect of Tempering Temperature on Microstructure and Mechanical Properties of Cast Steel 32MnMoNiCuCe[J]. New Technology & New Process, 2024(12): 50-57.
[5] 郑宏升, 郑岩. 热处理工艺对低合金铸钢力学性能的影响[J]. 河南科技, 2021, 40(28): 46-48.
ZHENG H S, ZHENG Y.Effect of Heat Treatment Process on Mechanical Properties of Low Alloy Cast Steel[J]. Henan Science and Technology, 2021, 40(28): 46-48.
[6] 唐彩, 陈波, 范汇吉. 回火工艺对ZG30Mn铸钢组织与力学性能的影响[J]. 金属热处理, 2020, 45(2): 134-138.
TANG C, CHEN B, FAN H J.Effect of Tempering Process on Microstructure and Mechanical Properties of ZG30Mn Cast Steel[J]. Heat Treatment of Metals, 2020, 45(2): 134-138.
[7] 范汇吉, 李毅, 靳鑫, 等. 回火温度对G18NiMoCr3-6钢组织与性能的影响[J]. 金属热处理, 2017, 42(11): 163-165.
FAN H J, LI Y, JIN X, et al.Effect of Tempering Temperature on Microstructure and Mechanical Properties of G18NiMoCr3-6[J]. Heat Treatment of Metals, 2017, 42(11): 163-165.
[8] 李兵, 唐正连, 罗元元, 等. 热处理工艺对低合金高性能铸钢组织和性能的影响[J]. 材料热处理学报, 2017, 38(8): 88-95.
LI B, TANG Z L, LUO Y Y, et al.Effect of Heat Treatment Process on Microstructure and Mechanical Property of Low Alloy High Performance Cast Steel[J]. Transactions of Materials and Heat Treatment, 2017, 38(8): 88-95.
[9] 王佐铭. 正火与回火处理对低合金铸钢组织和性能的影响[J]. 铸造技术, 2017, 38(4): 843-846.
WANG Z M.Effects of Normalizing and Tempering Treatment on Microstructure and Properties of Low Alloy Cast Steel[J]. Foundry Technology, 2017, 38(4): 843-846.
[10] 李茂华, 郭二军, 冯义成, 等. Mo含量对B+钢组织和性能的影响[J]. 铸造技术, 2017, 38(3): 530-532.
LI M H, GUO E J, FENG Y C, et al.Effect of Mo Content on Microstructure and Properties of B+ Steel[J]. Foundry Technology, 2017, 38(3): 530-532.
[11] 姚丽苹, 游龙, 潘昆明, 等. Ti微合金化铸钢组织及性能研究[J]. 河南科技大学学报(自然科学版), 2024, 45(6): 38-42.
YAO L P, YOU L, PAN K M, et al.Research on Microstructure and Properties of Ti Microalloyed Cast Steel[J]. Journal of Henan University of Science & Technology (Natural Science), 2024, 45(6): 38-42.
[12] 赵江涛, 周正军, 刘斌, 等. 基于机器学习的热轧含Ti高强钢性能预测[J]. 宝钢技术, 2025(1): 50-56.
ZHAO J T, ZHOU Z J, LIU B, et al.Performance Prediction of Ti-Containing High-Strength Steel in Hot Rolling Based on Machine Learning[J]. Bao-steel Technology, 2025(1): 50-56.
[13] 汪星辰, 袁灵洋, 王鑫, 等. 基于机器学习的免热处理压铸铝合金力学性能预测[J]. 特种铸造及有色合金, 2024, 44(11): 1498-1503.
WANG X C, YUAN L Y, WANG X, et al.Mechanical Property Prediction of Non-Heat Treatment Die Casting Aluminum Alloys Based on Machine Learning[J]. Special Casting & Nonferrous Alloys, 2024, 44(11): 1498-1503.
[14] HAO C M, SUI Y D, YUAN Y R, et al.Composition Optimization Design and High Temperature Mechanical Properties of Cast Heat-Resistant Aluminum Alloy via Machine Learning[J]. Materials & Design, 2025, 250: 113587.
[15] GUO S, YU J X, LIU X J, et al.A Predicting Model for Properties of Steel Using the Industrial Big Data Based on Machine Learning[J]. Computational Materials Science, 2019, 160: 95-104.
[16] 贾冠飞. 基础制造热加工铸造专业现场及实验数据集[DB/OL]. 北京: 国家基础学科公共科学数据中心, (2024-01-18)[2025-08-15]. https://cstr.cn/16666.11.nbsdc.RJdUup0s.
JIA G F. Basic Manufacturing Thermal Processing Casting Specialty Field and Experimental Data Set[DB/OL]. Beijing: National Basic Science Data Center of China, (2024-01-18)[2025-08-15]. https://cstr.cn/16666.11.nbsdc.RJdUup0s.
[17] 刘素贞, 任佳乐, 袁路航, 等. 磷酸铁锂电池荷电状态估计的多源数据特征提取[J]. 电工技术学报, 2025, 40(11): 3349-3361.
LIU S Z, REN J L, YUAN L H, et al.Multi-Source Data Feature Extraction Method for State of Charge Estimation of LiFePO4 Battery[J]. Transactions of China Electrotechnical Society, 2025, 40(11): 3349-3361.
[18] 马振宇, 吴纬, 张威, 等. 相关系数对软件可靠性验证测试的影响分析[J]. 微电子学与计算机, 2018, 35(5): 69-73.
MA Z Y, WU W, ZHANG W, et al.Analysis of Correlation Coefficient Influence on Software Reliability Demonstration Testing[J]. Microelectronics & Computer, 2018, 35(5): 69-73.
[19] 史云峰, 韩晓明, 王利红, 等. 基于特征选择和机器学习的锂电池寿命早期预测[J]. 现代电子技术, 2024, 47(22): 90-98.
SHI Y F, HAN X M, WANG L H, et al.Lithium Battery Life Early Prediction Based on Feature Selection and Machine Learning[J]. Modern Electronics Technique, 2024, 47(22): 90-98.
[20] 谭言丹, 赵阳洋, 赵光财. 基于AdaBoost特征选择和XGBoost的帕金森病诊断[J]. 信息技术, 2020(9): 124-128.
TAN Y D, ZHAO Y Y, ZHAO G C.Parkinson’s Disease Diagnosis Based on Ensemble Learning[J]. Information Technology, 2020(9): 124-128.
[21] 耿焕同, 戴中斌, 沈阳. 基于遗传算法的特征选择方法在短时强降水预报中的应用[J]. 气象科学, 2023, 43(1): 125-133.
GENG H T, DAI Z B, SHEN Y.Application of Feature Selection Method Based on Genetic Algorithm in Short-Term Heavy Precipitation Forecast[J]. Journal of the Meteorological Sciences, 2023, 43(1): 125-133.
[22] 黄成明. 基于互信息的特征选择与加权的茶叶品质检测算法研究[J]. 信息记录材料, 2025, 26(9): 71-73.
HUANG C M.Research on Tea Quality Detection Algorithm Based on Mutual Information Feature Selection and Weighting[J]. Information Recording Materials, 2025, 26(9): 71-73.
[23] 李昕光, 王珅, 曲大义. 基于特征选择的时空融合交通流预测[J]. 电子测量技术, 2023, 46(22): 87-93.
LI X G, WANG S, QU D Y.Spatiotemporal Fusion Traffic Flow Prediction Based on Feature Selection[J]. Electronic Measurement Technology, 2023, 46(22): 87-93.
[24] 王孟超, 武川, 孟亚飞, 等. 基于机器学习模具钢大型锻件力学性能预测[J]. 精密成形工程, 2024, 16(12): 198-208.
WANG M C, WU C, MENG Y F, et al.Prediction of Mechanical Properties of Large Forgings in Mold Steel Based on Machine Learning[J]. Journal of Netshape Forming Engineering, 2024, 16(12): 198-208.
[25] 肖跃雷, 张云娇. 基于特征选择和超参数优化的恐怖袭击组织预测方法[J]. 计算机应用, 2020, 40(8): 2262-2267.
XIAO Y L, ZHANG Y J.Terrorist Attack Organization Prediction Method Based on Feature Selection and Hyperparameter Optimization[J]. Journal of Computer Applications, 2020, 40(8): 2262-2267.
[26] 杨灿, 刘磊磊, 张遗立, 等. 基于贝叶斯优化机器学习超参数的滑坡易发性评价[J]. 地质科技通报, 2022, 41(2): 228-238.
YANG C, LIU L L, ZHANG Y L, et al.Machine Learning Based on Landslide Susceptibility Assessment with Bayesian Optimized the Hyperparameters[J]. Bulletin of Geological Science and Technology, 2022, 41(2): 228-238.
[27] 陶春国, 王丽萍, 冯义成, 等. C含量对车钩用超高强铸钢组织和性能的影响[J]. 铸造技术, 2015, 36(3): 625-627.
TAO C G, WANG L P, FENG Y C, et al.Effects of C Content on Microstructure and Mechanical Properties of Ultra-High Strength Cast Steel for Coupler[J]. Foundry Technology, 2015, 36(3): 625-627.
[28] 肖利民. Mn含量对铁路货车用低温铸钢显微组织和力学性能的影响[J]. 黑龙江科学, 2019, 10(18): 8-11.
XIAO L M.Effect of Mn on the Microstructure and Mechanical Properties of Low Temperature Cast Steel for Railway Wagon[J]. Heilongjiang Science, 2019, 10(18): 8-11.
[29] 吴林才, 徐京, 杨弋涛. Mo和Nb对节镍型耐热铸钢组织和力学性能的影响[J]. 铸造技术, 2016, 37(5): 875-879.
WU L C, XU J, YANG Y T.Influence of Mo and Nb on Microstructure and Mechanical Properties of Nickel-Based Heat-Resistant Cast Steel[J]. Foundry Technology, 2016, 37(5): 875-879.
[30] 刘晓雄, 肖章玉, 马正强, 等. 热处理参数对Co3W2型耐热钢组织和力学性能的影响规律研究[J]. 大型铸锻件, 2025(3): 1-5.
LIU X X, XIAO Z Y, MA Z Q, et al.Study on the Influence Law of Heat Treatment Parameters on Microstructure and Mechanical Properties of Co3W2 Heat- resistant Steel[J]. Heavy Castings and Forgings, 2025(3): 1-5.
PDF(646 KB)
渝公网安备 50010702501719号 渝ICP备15012534号-4
