A physics-informed neural network method for identifying parameters and predicting remaining life of fatigue crack growth

IF 5.7 2区材料科学 Q1 ENGINEERING, MECHANICAL

International Journal of Fatigue Pub Date : 2024-10-30 DOI:10.1016/j.ijfatigue.2024.108678

{"title":"A physics-informed neural network method for identifying parameters and predicting remaining life of fatigue crack growth","authors":"","doi":"10.1016/j.ijfatigue.2024.108678","DOIUrl":null,"url":null,"abstract":"<div><div>Predicting the remaining life of fatigue cracks is crucial for planning maintenance and repair strategies to prevent untoward incidents. This paper proposes a novel physics-informed neural network (PINN) method for identifying parameters and predicting remaining fatigue crack growth life (FCGL). Initially, the relationship between crack length and fatigue cycles is established through a neural network, and the gradient of fatigue cycles with respect to crack length is obtained by automatic differentiation. Subsequently, a composite loss function is designed to incorporate this gradient within the confines of physical knowledge, ensuring that the established relationship not only aligns with observed data but also adheres to physical knowledge. Furthermore, during the network training, the parameters in physical models are simultaneously updated to better conform to the individuality of the monitored subject. All predicted remaining FCGLs fall within the 1.5 times error band. Compared to purely data-driven or physics-based methods, the proposed method offers more robust and accurate predictions of remaining FCGLs.</div></div>","PeriodicalId":14112,"journal":{"name":"International Journal of Fatigue","volume":null,"pages":null},"PeriodicalIF":5.7000,"publicationDate":"2024-10-30","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/S0142112324005371","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

Predicting the remaining life of fatigue cracks is crucial for planning maintenance and repair strategies to prevent untoward incidents. This paper proposes a novel physics-informed neural network (PINN) method for identifying parameters and predicting remaining fatigue crack growth life (FCGL). Initially, the relationship between crack length and fatigue cycles is established through a neural network, and the gradient of fatigue cycles with respect to crack length is obtained by automatic differentiation. Subsequently, a composite loss function is designed to incorporate this gradient within the confines of physical knowledge, ensuring that the established relationship not only aligns with observed data but also adheres to physical knowledge. Furthermore, during the network training, the parameters in physical models are simultaneously updated to better conform to the individuality of the monitored subject. All predicted remaining FCGLs fall within the 1.5 times error band. Compared to purely data-driven or physics-based methods, the proposed method offers more robust and accurate predictions of remaining FCGLs.

查看原文本刊更多论文

用于识别疲劳裂纹生长参数和预测剩余寿命的物理信息神经网络方法

预测疲劳裂纹的剩余寿命对于规划维护和修理策略以防止意外事故至关重要。本文提出了一种新颖的物理信息神经网络（PINN）方法，用于识别参数和预测疲劳裂纹生长剩余寿命（FCGL）。首先，通过神经网络建立裂纹长度与疲劳循环之间的关系，并通过自动微分获得疲劳循环相对于裂纹长度的梯度。随后，设计一个复合损失函数，将这一梯度纳入物理知识的范围，确保建立的关系不仅与观测数据一致，而且符合物理知识。此外，在网络训练过程中，物理模型中的参数也会同时更新，以更好地符合被监控对象的个性。所有预测的剩余 FCGL 都在 1.5 倍误差范围内。与纯粹的数据驱动或基于物理的方法相比，所提出的方法对剩余 FCGL 的预测更稳健、更准确。

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.