Finite element–informed neural network (feinn) approach to inverse problems in 1D linear elasticity
Transport and Communications Science Journal · 21.10.2025 · 76 (9) · 1163-1172
DOI: 10.47869/tcsj.76.9.2 ↗Kurz gefasst
Fachartikel in der Zeitschrift Transport and Communications Science Journal, online erschienen am 21.10.2025. Verfasst von Khiết Nguyễn Thanh und 2 weiteren Autor:innen.
Abstract (Originalsprache)
Inverse analysis of material parameters plays a crucial role in solid mechanics problems, which remain challenging and computationally expensive with conventional numerical approaches such as the finite element method (FEM). This paper proposes a highly efficient inverse analysis framework for 1D problems based on the Finite Element Informed Neural Network (FEINN). Unlike Physics-Informed Neural Networks (PINNs), which solve the strong form of governing equations, FEINN addresses the weak form through finite discretization and incorporates Gaussian integration to compute the strain–displacement matrix, thereby significantly accelerating training and convergence. In the inverse problem, the unknown material parameters are inferred from the neural network’s output layer, while the nodal coordinates serve as input. Nodal displacements and forces are employed as constraints within the loss function. FEINN determines the material parameters by iteratively optimizing the neural network using the fmincon function in MATLAB. The proposed method demonstrates high efficiency through several benchmark cases involving both constant and spatially varying material parameters
Zitieren
APA
Thanh, K. N., Minh, P. P., & Đình, D. N. (2025). Finite element–informed neural network (feinn) approach to inverse problems in 1D linear elasticity. Transport and Communications Science Journal, 76(9), 1163-1172. https://doi.org/10.47869/tcsj.76.9.2
Vancouver
Thanh KN, Minh PP, Đình DN. Finite element–informed neural network (feinn) approach to inverse problems in 1D linear elasticity. Transport and Communications Science Journal. 2025;76(9):1163-1172. doi:10.47869/tcsj.76.9.2
BibTeX
@article{Thanh2025,
author = {Thanh, Khiết Nguyễn and Minh, Phúc Phạm and Đình, Dư Nguyễn},
title = {Finite element–informed neural network (feinn) approach to inverse problems in 1D linear elasticity},
journal = {Transport and Communications Science Journal},
year = {2025},
volume = {76},
number = {9},
pages = {1163--1172},
doi = {10.47869/tcsj.76.9.2}
}
EUVIMED