Please wait a minute...
金属学报  2026, Vol. 62 Issue (9): 1606-1614    DOI: 10.11900/0412.1961.2024.00302
  研究论文 本期目录 | 过刊浏览 |
基于随机森林算法与数据增强策略双驱动的多尺度疲劳裂纹扩展预测
刁圣轩, 肖金雍, 陈永保, 单康中, 杨杰()
上海理工大学 能源与动力工程学院 上海市动力工程多相流动与传热重点实验室 上海 200093
Multi-Scale Fatigue Crack Propagation Prediction Based on the Dual Drive of Random Forest Algorithm and Data Augmentation Strategy
DIAO Shengxuan, XIAO Jinyong, CHEN Yongbao, SHAN Kangzhong, YANG Jie()
Shanghai Key Laboratory of Multiphase Flow and Heat Transfer in Power Engineering, School of Energy and Power Engineering, University of Shanghai for Science and Technology, Shanghai 200093, China
引用本文:

刁圣轩, 肖金雍, 陈永保, 单康中, 杨杰. 基于随机森林算法与数据增强策略双驱动的多尺度疲劳裂纹扩展预测[J]. 金属学报, 2026, 62(9): 1606-1614.
Shengxuan DIAO, Jinyong XIAO, Yongbao CHEN, Kangzhong SHAN, Jie YANG. Multi-Scale Fatigue Crack Propagation Prediction Based on the Dual Drive of Random Forest Algorithm and Data Augmentation Strategy[J]. Acta Metall Sin, 2026, 62(9): 1606-1614.

全文: PDF(1949 KB)   HTML
摘要: 

为了实现多尺度疲劳裂纹扩展过程的准确预测,并增强预测模型的可解释性和泛化能力,本工作选用304奥氏体不锈钢为研究对象,首先对K近邻(KNN)、支持向量机回归(SVR)和随机森林(RF)三种算法的疲劳裂纹扩展预测能力进行了对比,筛选出预测性能最优的算法。随后,选择最优算法,分别基于长裂纹、短裂纹和多尺度裂纹三种疲劳裂纹扩展模型添加数据增强策略进行数据增强,以提高疲劳裂纹扩展预测的准确性。最后,基于添加数据增强策略的双驱动模型框架,在不同载荷下进行多尺度疲劳裂纹扩展预测。结果表明,与SVR和KNN算法相比,RF算法更适合于疲劳裂纹扩展预测,其基于多尺度疲劳裂纹扩展模型的数据增强效果明显优于其余两种模型。基于RF算法与数据增强策略的双驱动模型具有很好的泛化能力,且相比于基于RF算法的纯数据驱动模型,双驱动模型预测结果更加准确。

关键词 : 多尺度疲劳裂纹扩展,  随机森林算法,  数据增强策略,  双驱动模型    
Abstract:

Data-driven methods based on machine learning have been employed to predict fatigue crack propagation. However, existing studies have largely overlooked the multi-scale nature of this process. Relying solely on macroscopic data for long-crack prediction often fails to capture the complete crack growth process, potentially resulting in non-conservative predictions. Moreover, purely data-driven models often lack interpretability, exhibit limited generalization capabilities, and struggle to adhere to physical laws. The challenge of integrating modeling with reasonable interpretations, driven by both data and mechanisms, remains a significant issue for researchers. In this study, we selected 304 austenitic stainless steel as the research object. To identify the algorithm with the best predictive performance, firstly, the fatigue crack propagation prediction capabilities of three algorithms were compared: K-nearest neighbor (KNN), support vector machine regression (SVR), and random forest (RF). The most effective algorithm and implemented data augmentation strategies based on three fatigue crack propagation models were selected, namely, long cracks, short cracks, and multi-scale cracks, to enhance prediction accuracy. Finally, using a dual-drive model framework that incorporated the data augmentation strategy, predictions of multi-scale fatigue crack propagation under different loads were conducted. Compared to the SVR and KNN algorithms, the results indicated that the RF algorithm had a lower root mean square error (RMSE) value and higher coefficient of determination (R2), making it more suitable for predicting fatigue crack propagation. Under a load of 370 MPa, the prediction accuracy for the training set ranked in the order of RF > KNN > SVR. By contrast, the accuracy for the test set was in the order of RF > SVR > KNN. The multi-scale fatigue crack propagation model effectively captured the entire process of crack growth, whereas the long- and short-crack models accurately represented only parts of it. Data enhancement based on the multi-scale crack model demonstrated significantly better results than the other two models, with increases in accuracy of 25.76% and 71.74% for the training and test sets, respectively. The dual-drive model based on the RF algorithm and data enhancement strategy exhibited strong generalization capabilities. Under a load of 350 MPa, the RMSE values for the training and test sets were 0.046 and 0.111, respectively, and R2 reached 0.995 and 0.961. Under a load of 330 MPa, the RMSE values for the training and test sets were 0.171 and 0.081, respectively, and the R2 reached 0.911 and 0.721. Finally, compared with the pure data-driven model based on the RF algorithm, the predictions from the dual-drive model were found to be significantly more accurate.

Key words: multi-scale fatigue crack propagation    random forest algorithm    data augmentation strategy    dual drive model
收稿日期: 2024-08-28     
ZTFLH:  TG111.8  
基金资助:国家自然科学基金项目(52375154);国家自然科学基金项目(52311530067)
通讯作者: 杨 杰,yangjie@usst.edu.cn,主要从事重要装备寿命及可靠性等研究
Corresponding author: YANG Jie, professor, Tel: (021)55272320, E-mail: yangjie@usst.edu.cn
作者简介: 刁圣轩,男,2001年生,硕士生
图1  不同应力下304不锈钢的多尺度疲劳裂纹扩展实验数据[9]
图2  数据的训练与测试流程
图3  数据增强流程
ModelParameterValue
Long crackC1.615 × 10-10
m2.58
Short crackA1.000 × 10-10.49
B1.19
Multi-scale crackC3.630 × 10-10
q1.37
表1  三种疲劳裂纹扩展模型中的参数[36]

Load / MPa

Number of experimental datasetNumber of data augmentation dataset
3306118
3505015
3704514
表2  数据集数量
图4  三种算法训练集训练结果和测试集预测结果对比
图5  三种算法的均方根误差(RMSE)及决定系数(R2)对比
图6  三种疲劳裂纹扩展模型对训练集和测试集的数据增强效果
图7  基于三种模型数据增强后RMSE及R2对比
图8  350 MPa下基于双驱动模型和随机森林(RF)算法模型的多尺度疲劳裂纹扩展预测结果
图9  330 MPa下基于双驱动模型和RF算法模型的多尺度疲劳裂纹扩展预测结果
图10  350和330 MPa下基于双驱动模型和RF算法模型的RMSE及R2
[1] Xu W, Huang M H, Wang J L, et al. Review: Relations between metastable austenite and fatigue behavior of steels [J]. Acta Metall. Sin., 2020, 56: 459
doi: 10.11900/0412.1961.2019.00399
[1] 徐 伟, 黄明浩, 王金亮 等. 综述: 钢中亚稳奥氏体组织与疲劳性能关系 [J]. 金属学报, 2020, 56: 459
[2] Wei L S, Liu H L, Tang C G, et al. Investigation of novel type of cylindrical lithium-ion battery heat exchangers based on topology optimization [J]. Energy, 2024, 304: 131886
doi: 10.1016/j.energy.2024.131886
[3] Wang H J, Li B, Gong J G, et al. Machine learning-based fatigue life prediction of metal materials: Perspectives of physics-informed and data-driven hybrid methods [J]. Eng. Fract. Mech., 2023, 284: 109242
doi: 10.1016/j.engfracmech.2023.109242
[4] Irwin G R. Analysis of stresses and strains near the end of a crack traversing a plate [J]. J. Appl. Mech., 1957, 24: 361
doi: 10.1115/1.4011547
[5] Paris P, Erdogan F. A critical analysis of crack propagation laws [J]. J. Basic Eng., 1963, 85: 528
doi: 10.1115/1.3656900
[6] Forman R G, Kearney V E, Engle R M. Numerical analysis of crack propagation in cyclic-loaded structures [J]. J. Basic Eng., 1967, 89: 459
doi: 10.1115/1.3609637
[7] Pearson S. Initiation of fatigue cracks in commercial aluminium alloys and the subsequent propagation of very short cracks [J]. Eng. Frac. Mech., 1975, 7: 235
doi: 10.1016/0013-7944(75)90004-1
[8] Okazaki M, Yada T, Endoh T. Surface small crack growth behavior in low-cycle fatigue at elevated temperature and application limit of macroscopic crack growth law [J]. Nucl. Eng. Des., 1989, 111: 123
doi: 10.1016/0029-5493(89)90284-7
[9] Deng G J, Tu S T, Wang Q Q, et al. Small fatigue crack growth mechanisms of 304 stainless steel under different stress levels [J]. Int. J. Fatigue, 2014, 64: 14
doi: 10.1016/j.ijfatigue.2014.01.027
[10] Ye S, Zhang C C, Zhang P Y, et al. Fatigue life prediction of nickel-based GH4169 alloy on the basis of a multi-scale crack propagation approach [J]. Eng. Fract. Mech., 2018, 199: 29
doi: 10.1016/j.engfracmech.2018.05.023
[11] Agrawal A, Choudhary A. Perspective: Materials informatics and big data: Realization of the “fourth paradigm” of science in materials science [J]. APL Mater., 2016, 4: 053208
[12] Schleder G R, Padilha A C M, Acosta C M, et al. From DFT to machine learning: Recent approaches to materials science—A review [J]. J. Phys.: Mater., 2019, 2: 032001
[13] Zhao S K, Jiang C, Long X Y. Remaining useful life estimation of mechanical systems based on the data-driven method and bayesian theory [J]. J. Mech. Eng., 2018, 54(12): 115
doi: 10.3901/JME.2018.12.115
[13] 赵申坤, 姜 潮, 龙湘云. 一种基于数据驱动和贝叶斯理论的机械系统剩余寿命预测方法 [J]. 机械工程学报, 2018, 54(12): 115
doi: 10.3901/JME.2018.12.115
[14] Liu Y L, Niu C, Wang Z, et al. Machine learning in materials genome initiative: A review [J]. J. Mater. Sci. Technol., 2020, 57: 113
doi: 10.1016/j.jmst.2020.01.067
[15] Wang H X, Zhang W F, Sun F Q, et al. A comparison study of machine learning based algorithms for fatigue crack growth calculation [J]. Materials, 2017, 10: 543
doi: 10.3390/ma10050543
[16] Younis H B, Kamal K, Sheikh M F, et al. Prediction of fatigue crack growth rate in aircraft aluminum alloys using optimized neural networks [J]. Theor. Appl. Fract. Mech., 2022, 117: 103196
doi: 10.1016/j.tafmec.2021.103196
[17] Zafar M H, Younis H B, Mansoor M, et al. Training deep neural networks with novel metaheuristic algorithms for fatigue crack growth prediction in aluminum aircraft alloys [J]. Materials, 2022, 15: 6198
doi: 10.3390/ma15186198
[18] Mortazavi S N S, Ince A. An artificial neural network modeling approach for short and long fatigue crack propagation [J]. Comput. Mater. Sci., 2020, 185: 109962
doi: 10.1016/j.commatsci.2020.109962
[19] Zhang J Q, Zhu J C, Guo W, et al. A machine learning-based approach to predict the fatigue life of three-dimensional cracked specimens [J]. Int. J. Fatigue, 2022, 159: 106808
doi: 10.1016/j.ijfatigue.2022.106808
[20] Barbosa J F, Correia J A F O, Júnior R C S F, et al. Fatigue life prediction of metallic materials considering mean stress effects by means of an artificial neural network [J]. Int. J. Fatigue, 2020, 135: 105527
doi: 10.1016/j.ijfatigue.2020.105527
[21] Gan L, Wu H, Zhong Z. Fatigue life prediction considering mean stress effect based on random forests and kernel extreme learning machine [J]. Int. J. Fatigue, 2022, 158: 106761
doi: 10.1016/j.ijfatigue.2022.106761
[22] Yang J Y, Kang G Z, Liu Y J, et al. A novel method of multiaxial fatigue life prediction based on deep learning [J]. Int. J. Fatigue, 2021, 151: 106356
doi: 10.1016/j.ijfatigue.2021.106356
[23] Liu F, Chen H F, Yang J, et al. Application of physics-informed machine learning methods in buckling design of axially compressed cylindrical shells [J]. Thin-Walled Struct., 2024, 200: 111963
doi: 10.1016/j.tws.2024.111963
[24] Zhang X C, Gong J G, Xuan F Z. A physics-informed neural network for creep-fatigue life prediction of components at elevated temperatures [J]. Eng. Fract. Mech., 2021, 258: 108130
doi: 10.1016/j.engfracmech.2021.108130
[25] Karpatne A, Atluri G, Faghmous J H, et al. Theory-guided data science: A new paradigm for scientific discovery from data [J]. IEEE Trans. Knowl. Data Eng., 2017, 29: 2318
doi: 10.1109/TKDE.2017.2720168
[26] Karniadakis G E, Kevrekidis I G, Lu L, et al. Physics-informed machine learning [J]. Nat. Rev. Phys., 2021, 3: 422
doi: 10.1038/s42254-021-00314-5
[27] Yang J Y, Kang G Z, Kan Q H. Rate-dependent multiaxial life prediction for polyamide-6 considering ratchetting: Semi-empirical and physics-informed machine learning models [J]. Int. J. Fatigue, 2022, 163: 107086
doi: 10.1016/j.ijfatigue.2022.107086
[28] Jiang L F, Hu Y N, Liu Y X, et al. Physics-informed machine learning for low-cycle fatigue life prediction of 316 stainless steels [J]. Int. J. Fatigue, 2024, 182: 108187
doi: 10.1016/j.ijfatigue.2024.108187
[29] Jiang L F, Hu Y N, Li H, et al. A cGAN-based fatigue life prediction of 316 austenitic stainless steel in high-temperature and high-pressure water environments [J]. Int. J. Fatigue, 2025, 190: 108633
doi: 10.1016/j.ijfatigue.2024.108633
[30] Jain A K, Duin R P W, Mao J C. Statistical pattern recognition: A review [J]. IEEE Trans. Pattern Anal. Mach. Intell., 2000, 22: 4
doi: 10.1109/34.824819
[31] Cover T, Hart P. Nearest neighbor pattern classification [J]. IEEE Trans. Inform. Theory, 1967, 13: 21
doi: 10.1109/TIT.1967.1053964
[32] Cortes C, Vapnik V. Support-vector networks [J]. Mach. Learn., 1995, 20: 273
doi: 10.1023/A:1022627411411
[33] Smola A J, Schölkopf B. A tutorial on support vector regression [J]. Stat. Comput., 2004, 14: 199
doi: 10.1023/B:STCO.0000035301.49549.88
[34] Breiman L. Random forests [J]. Mach. Learn., 2001, 45: 5
doi: 10.1023/A:1010933404324
[35] Xie J X, Su Y J, Xue D Z, et al. Machine Learning for Materials Research and Development [J]. Acta Metall. Sin., 2021, 57: 1343
doi: 10.11900/0412.1961.2021.00357
[35] 谢建新, 宿彦京, 薛德祯 等. 机器学习在材料研发中的应用 [J]. 金属学报, 2021, 57: 1343
doi: 10.11900/0412.1961.2021.00357
[36] Lu R S, Yang J, Wang J, et al. Probabilistic damage tolerance assessment method based on the multi-scale crack growth model [J]. Eng. Fract. Mech., 2023, 285: 109297
doi: 10.1016/j.engfracmech.2023.109297
[1] 王晨充, 徐伟. 综述:合金设计中物理模型与人工智能的集成与发展[J]. 金属学报, 2025, 61(4): 541-560.
[2] 赵鹏, 谢光, 段慧超, 张健, 杜奎. 两种高代次镍基单晶高温合金热机械疲劳中的再结晶行为[J]. 金属学报, 2023, 59(9): 1221-1229.
[3] 张禄, 余志伟, 张磊成, 江荣, 宋迎东. GH4169高温合金热机械疲劳循环损伤机理及数值模拟[J]. 金属学报, 2023, 59(7): 871-883.
[4] 韩冬, 张炎杰, 李小武. 短程有序对高层错能Cu-Mn合金拉-拉疲劳变形行为及损伤机制的影响[J]. 金属学报, 2022, 58(9): 1208-1220.
[5] 潘复生, 蒋斌. 镁合金塑性加工技术发展及应用[J]. 金属学报, 2021, 57(11): 1362-1379.
[6] 兰亮云, 孔祥伟, 邱春林, 杜林秀. 基于多尺度力学实验的氢脆现象的最新研究进展[J]. 金属学报, 2021, 57(7): 845-859.
[7] 王金亮, 王晨充, 黄明浩, 胡军, 徐伟. 低应变预变形对变温马氏体相变行为的影响规律及作用机制[J]. 金属学报, 2021, 57(5): 575-585.
[8] 周红伟, 白凤梅, 杨磊, 陈艳, 方俊飞, 张立强, 衣海龙, 何宜柱. 1100 MPa级高强钢的低周疲劳行为[J]. 金属学报, 2020, 56(7): 937-948.
[9] 徐伟,黄明浩,王金亮,沈春光,张天宇,王晨充. 综述:钢中亚稳奥氏体组织与疲劳性能关系[J]. 金属学报, 2020, 56(4): 459-475.
[10] 王常帅,郭莉莉,唐丽英,周荣灿,郭建亭,周兰章. GH984G合金在700 ℃水蒸气中的氧化行为[J]. 金属学报, 2019, 55(7): 893-901.
[11] 彭剑,高毅,代巧,王颖,李凯尚. 316L奥氏体不锈钢非对称载荷下的疲劳与循环塑性行为[J]. 金属学报, 2019, 55(6): 773-782.
[12] 张哲峰, 刘睿, 张振军, 田艳中, 张鹏. 金属材料疲劳性能预测统一模型探索[J]. 金属学报, 2018, 54(11): 1693-1704.
[13] 谭梅林, 王常帅, 郭永安, 郭建亭, 周兰章. Ti/Al比对GH984G合金长期时效过程中γ′沉淀相粗化行为及拉伸性能的影响[J]. 金属学报, 2014, 50(10): 1260-1268.
[14] 谭梅林 王常帅 郭永安 郭建亭 周兰章. Ti/Al比对GH984G合金长期时效过程中γ′沉淀相粗化行为及拉伸性能的影响[J]. 金属学报, 0, 0(0): 0-0.
[15] 聂德福 赵杰 张俊善. 一种估算结构钢室温蠕变的方法[J]. 金属学报, 2011, 47(2): 179-184.