The hybrid model of domain knowledge, symbolic regression and neural networks for multiaxial fatigue life prediction

IF 6.8 2区材料科学 Q1 ENGINEERING, MECHANICAL

International Journal of Fatigue Pub Date : 2025-08-23 DOI:10.1016/j.ijfatigue.2025.109246

Zhanguang Zheng , Cheng Lin , Jun Yang , Dongyang Chen , Liping Jiang

{"title":"The hybrid model of domain knowledge, symbolic regression and neural networks for multiaxial fatigue life prediction","authors":"Zhanguang Zheng , Cheng Lin , Jun Yang , Dongyang Chen , Liping Jiang","doi":"10.1016/j.ijfatigue.2025.109246","DOIUrl":null,"url":null,"abstract":"<div><div>To further enhance the prediction performance of multiaxial fatigue life based on Physics Informed Neural Network (PINN), Domain Knowledge guided Symbolic Regression-Neural Network (DKSR-NN) framework is proposed. At first, domain knowledge is employed to guide symbolic regression in generating expressions that possess both physical interpretability and high predictive accuracy. These expressions are then incorporated as physics-informed loss functions within the PINN architecture. This integration significantly improves the model’s accuracy and stability, especially under conditions of limited fatigue data. At last, the proposed method is validated by comparisons with different machine learning on AZ61A magnesium alloy, TC4 titanium alloy, and sintered porous iron. The results demonstrate that the DKSR-NN framework is better than PINN using critical plane models as physical loss constraints, DKSR and pure data-driven machine learning methods. This will provide a prospect for multiaxial fatigue life prediction.</div></div>","PeriodicalId":14112,"journal":{"name":"International Journal of Fatigue","volume":"202 ","pages":"Article 109246"},"PeriodicalIF":6.8000,"publicationDate":"2025-08-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"International Journal of Fatigue","FirstCategoryId":"88","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S0142112325004438","RegionNum":2,"RegionCategory":"材料科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"ENGINEERING, MECHANICAL","Score":null,"Total":0}

引用次数: 0

Abstract

To further enhance the prediction performance of multiaxial fatigue life based on Physics Informed Neural Network (PINN), Domain Knowledge guided Symbolic Regression-Neural Network (DKSR-NN) framework is proposed. At first, domain knowledge is employed to guide symbolic regression in generating expressions that possess both physical interpretability and high predictive accuracy. These expressions are then incorporated as physics-informed loss functions within the PINN architecture. This integration significantly improves the model’s accuracy and stability, especially under conditions of limited fatigue data. At last, the proposed method is validated by comparisons with different machine learning on AZ61A magnesium alloy, TC4 titanium alloy, and sintered porous iron. The results demonstrate that the DKSR-NN framework is better than PINN using critical plane models as physical loss constraints, DKSR and pure data-driven machine learning methods. This will provide a prospect for multiaxial fatigue life prediction.

查看原文本刊更多论文

多轴疲劳寿命预测的领域知识、符号回归和神经网络混合模型

为了进一步提高基于物理信息神经网络（PINN）的多轴疲劳寿命预测性能，提出了领域知识引导的符号回归神经网络（DKSR-NN）框架。首先，利用领域知识指导符号回归生成既具有物理可解释性又具有高预测精度的表达式。然后将这些表达式合并为PINN体系结构中的物理通知损失函数。这种集成显著提高了模型的准确性和稳定性，特别是在疲劳数据有限的情况下。最后，通过AZ61A镁合金、TC4钛合金和烧结多孔铁的不同机器学习对比，验证了所提方法的有效性。结果表明，使用临界平面模型作为物理损失约束、DKSR和纯数据驱动的机器学习方法，DKSR- nn框架优于PINN。这为多轴疲劳寿命预测提供了前景。

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

International Journal of Fatigue 工程技术-材料科学：综合

CiteScore

10.70

自引率

21.70%

发文量

619

审稿时长

58 days

期刊介绍： Typical subjects discussed in International Journal of Fatigue address: Novel fatigue testing and characterization methods (new kinds of fatigue tests, critical evaluation of existing methods, in situ measurement of fatigue degradation, non-contact field measurements) Multiaxial fatigue and complex loading effects of materials and structures, exploring state-of-the-art concepts in degradation under cyclic loading Fatigue in the very high cycle regime, including failure mode transitions from surface to subsurface, effects of surface treatment, processing, and loading conditions Modeling (including degradation processes and related driving forces, multiscale/multi-resolution methods, computational hierarchical and concurrent methods for coupled component and material responses, novel methods for notch root analysis, fracture mechanics, damage mechanics, crack growth kinetics, life prediction and durability, and prediction of stochastic fatigue behavior reflecting microstructure and service conditions) Models for early stages of fatigue crack formation and growth that explicitly consider microstructure and relevant materials science aspects Understanding the influence or manufacturing and processing route on fatigue degradation, and embedding this understanding in more predictive schemes for mitigation and design against fatigue Prognosis and damage state awareness (including sensors, monitoring, methodology, interactive control, accelerated methods, data interpretation) Applications of technologies associated with fatigue and their implications for structural integrity and reliability. This includes issues related to design, operation and maintenance, i.e., life cycle engineering Smart materials and structures that can sense and mitigate fatigue degradation Fatigue of devices and structures at small scales, including effects of process route and surfaces/interfaces.