Aiming at the problems of high cost and long cycle in the fatigue strength assessment of metal structures under long-term loading histories, a new assessment method utilizing machine learning technology was proposed. The new method comprises three steps: structural fatigue modification, equivalent load calculation, and fatigue strength assessment. By taking four-point bending components and hole-containing plates as examples, comparing the fatigue strength utilization rates obtained from results without machine learning and those derived using the machine learning method, and by analyzing the influence of sample size on the model’s accuracy, the new method was verified to have the advantages of high accuracy and high efficiency. It is proven that this method is suitable for the fatigue strength assessment of metal structures subjected to long-term loads and multiple working conditions and significantly improves computational efficiency and reduces costs. Furthermore, the method demonstrates good generalizability, making it suitable for the fatigue strength assessment of different structures and holding considerable engineering value.
ZHOU Song, LI Ji-yuan.
Fatigue Strength Assessment Method for Longtime Service Metal Structures by Using Machine Learning Technology[J].
Science Technology and Engineering, 2026, 26(18): 7630-7640 DOI:10.12404/j.issn.1671-1815.2505190
为获得应力历史,可根据外部载荷的数量选择两种策略,如图1所示。外部载荷历史由 p个工况和每个工况中的q个外部载荷循环组成,这些工况和循环对服役部件的应力分布有显著影响。在有限元模型中,服役部件可由t个节点组成。如果p×q×t小于临界值Φ,则可直接采用有限元法(finite element method,FEM)进行等效载荷计算。有限元法对于长期载荷和多种工况条件下的结构疲劳强度评价,其计算成本高。针对这些问题,新方法在p×q×t等于或大于阈值Φ时,采用机器学习进行等效载荷计算以降低计算耗时,其计算流程如图3所示。基于机器学习技术的等效载荷计算分为3个阶段:首先,从外部载荷历史中随机抽取一些外部载荷值(Φ)来获得应力加载历史;然后利用机器学习技术建立外部载荷应力模型;最后,利用建立的模型将外部载荷历史转换为应力历史。
在新方法验证之后,根据验证结果对新方法的成本和效率等特征进行了分析。如上所述,机器学习技术中用于等效载荷计算和疲劳强度评估的样本大小阈值均设置为80。这个值大大降低了有限元计算和疲劳强度利用率的计算结果。为验证阈值80的合理性,将机器学习技术模型评估中广泛使用的3个指标:均方误差均值(mean of mean square error,MSE)、决定系数(R2)及均方根误差(root mean square error,RMSE),与样本规模的增大趋势进行对比分析。
LiuLu, DaiXin, ZhangWei, et al. Development of a heavy-load hydraulic servo loading device for landing gear fatigue test[J]. Science Technology and Engineering, 2025, 25(28): 12279-12286.
LiYuanjie, JinHaibo, XieSuijie, et al. Material parameters identification for fatigue strength assessment of additively manufactured titanium alloy components in the FKM guideline[J]. Aerospace Manufacturing Technology, 2025(2): 35-40.
ZhangJinzhou, LiuWenjia, JiShiqing, et al. Fatigue crack propagation and life prediction of welded joints under residual stress[J]. Science Technology and Engineering, 2024, 24(35): 14976-14983.
ZengLang, CaiCan, LiaoFeilong, et al. Mechanical strength and fatigue performance of offshore derricks after corrosion[J]. Science Technology and Engineering, 2024, 24(35): 14911-14920.
[9]
RennertR, VormwaldM, EsdertsA. FKM-guideline “Analytical strength Assessment”-background and current developments[J]. International Journal of Fatigue, 2024, 182: 108165.
[10]
BaiX, ZhangP, LiuS, et al. Fatigue strength prediction of large-size component through size effect measurement and determination[J]. International Journal of Fatigue, 2023, 168: 107408.
[11]
TaoZ Q, ChangY K, PanX N, et al. Numerical simulation of crack surface contacting behavior with stress-induced martensitic phase transformation in very-high-cycle fatigue regime[J]. International Journal of Fatigue, 2025, 198: 109019.
[12]
MaljaarsJ, LeanderJ, NussbaumerA, et al. Models and methods for probabilistic safety assessment of steel structures subject to fatigue[J]. Structural Safety, 2025, 113: 102446.
[13]
BaiX, ZhangP, WangQ, et al. A new dominance distribution method to select materials with higher fatigue resistance under property scatter and load uncertainty[J]. Journal of Materials Engineering and Performance, 2022, 31(12): 10248-10257.
LuYe, CaoBingyang, LiZhichao, et al. Research on the equivalent static calculation method of steel double-layer storage tank under the explosion impact load[J]. Journal of Safety and Environment, 2025, 25(3): 898-909.
[16]
LiH, XinW, XuB, et al. Delamination failure analysis of wind turbine blades based on equivalent fatigue load[J]. Ocean Engineering, 2024, 310(P1): 118642.
[17]
BaiX, ZhangP, ZhangZ J, et al. New method for determining P-S-N curves in terms of equivalent fatigue lives[J]. Fatigue & Fracture of Engineering Materials & Structures, 2019, 42(10): 2340-2353.
[18]
CoriaI, MartN I, BouzidA H, et al. Efficient assembly of bolted joints under external loads using numerical FEM[J]. International Journal of Mechanical Sciences, 2018, 142: 575-582.
[19]
LiH F, WeiJ L, LiS H, et al. Fatigue life prediction of high-speed train bearings based on the generalized linear cumulative damage theory[J]. Fatigue & Fracture of Engineering Materials & Structures, 2023, 46(6): 2112-2120.
[20]
AcharyaR, CaputoA N, NeuR W. Machine learning model for predicting the influence of crystallographic orientation on thermomechanical fatigue of Ni-base superalloys[J]. International Journal of Fatigue, 2025, 194: 108832.
[21]
SchubnellJ, FliegenerS, RosenbergerJ, et al. Data-driven fatigue assessment of welded steel joints based on transfer learning[J]. Welding in the World, 2025, 69(8): 1-16.
[22]
SrinivasanD V, MoradiM, KomninosP, et al. A generalized machine learning framework to estimate fatigue life across materials with minimal data[J]. Materials & Design, 2024, 246: 113355.
ZhouBingfeng, CaoQianqian, XieXinshui, et al. Evolution of functions and driving factors of Beijing subway stations: a machine learning XGBoost-SHAP model[J]. Science Technology and Engineering, 2025, 25(26): 11346-11357.
FuWeifang, WangHongrong, ZhangZhaomin, et al. End-to-end optimization of consumable spare parts inventory in aeronautical maintenance[J]. Science Technology and Engineering, 2025, 25(28): 12287-12295.
ZhangZitong, ShiChenyi, FuYingkun, et al. Production forecast of shale gas wells based on long short-term memory and Bayesian optimization[J]. Science Technology and Engineering, 2025, 25(27): 11569-11580.
QinSaibo, YiXianzhong, ZhangXuwen, et al. Working safety analysis of iron roughneck arm used in seabed drilling[J]. China Petroleum Machinery, 2024, 52 (10): 74-82.
DouYihua, ZhangJiaqiang, LiGuoliang, et al. Fatigue life prediction of coiled tubing based on optimized BP neural network[J]. China Petroleum Machinery, 2023, 51(10): 144-149.
ZhenChunbo, LiuShihao, ZhangAifeng, et al. Fatigue strength analysis of ship’s welded structures based on method of notch stress[J]. Journal of Shanghai Jiao Tong University, 2025, 59(4): 458-465.