Effect of Mesh Resolution and Data Sparsity on the Deformation Reconstruction and Strain Sensing using iFEM Methodology
e-Journal of Nondestructive Testing · 2026-08 · 31 (8)
DOI: 10.58286/33675 ↗Kurz gefasst
Fachartikel in der Zeitschrift e-Journal of Nondestructive Testing, 2026-08. Verfasst von Hoda Mozafari feizabad und 1 weiteren Autor:innen.
Abstract (Originalsprache)
The inverse finite element method (iFEM) has emerged as a powerful computational framework for real-time shape and strain sensing of structures using sparse experimental measurements. As a purely physics-based approach, iFEM reconstructs full-field displacement, rotation, and strain distributions by minimizing a weighted least-squares functional between analytically computed and experimentally measured section strains, without requiring any prior knowledge of the material properties or external loads. This capability makes iFEM particularly attractive for structural health monitoring (SHM) applications where only limited sensor information is available. Despite its advantages, the accuracy and robustness of iFEM strongly depend on two critical parameters: the resolution of the inverse finite element mesh and the spatial density of available strain measurements. A coarse mesh may lead to insufficient representation of structural kinematics, while excessive mesh refinement without adequate sensor coverage can result in numerical instability and noise amplification. Similarly, sparse or non-optimally distributed sensors can significantly deteriorate reconstruction quality, especially for complex geometries. Although several weighting strategies have been proposed in the literature to mitigate these issues, a comprehensive evaluation of the combined effects of mesh resolution and data sparsity on iFEM performance remains largely unexplored. In this study, a systematic parametric investigation is conducted to assess how mesh discretization density and sensor availability influence the deformation reconstruction and strain-sensing accuracy of the four-node quadrilateral inverse-shell element (iQS4) formulation within the iFEM framework. Three representative benchmark cases including a clamped flat plate, a curved cylindrical shell, and a T-beam structure are examined under various loading conditions. Each configuration is analyzed using different mesh resolutions and strain sensor distributions to capture a wide range of practical SHM scenarios. The reconstructed deformation and strain fields obtained from iFEM are quantitatively compared against high-fidelity forward finite element analysis (FEA) results in terms of displacement and strain error metrics. The findings reveal that there exists an optimal balance between mesh refinement and sensor density that minimizes reconstruction error while maintaining computational efficiency. Furthermore, the study identifies critical zones where sensor placement has the greatest impact on the fidelity of the reconstructed fields. Based on these insights, practical guidelines and best practices for sensor layout design and mesh configuration in iFEM/iQS4-based SHM systems are proposed. The outcomes of this research contribute to a deeper understanding of iFEM performance limitations and provide valuable recommendations for implementing reliable and cost-effective shape and strain sensing in engineering structures.
Zitieren
APA
feizabad, H. M., & Kefal, A. (2026). Effect of Mesh Resolution and Data Sparsity on the Deformation Reconstruction and Strain Sensing using iFEM Methodology. e-Journal of Nondestructive Testing, 31(8). https://doi.org/10.58286/33675
Vancouver
feizabad HM, Kefal A. Effect of Mesh Resolution and Data Sparsity on the Deformation Reconstruction and Strain Sensing using iFEM Methodology. e-Journal of Nondestructive Testing. 2026;31(8). doi:10.58286/33675
BibTeX
@article{feizabad2026,
author = {feizabad, Hoda Mozafari and Kefal, Adnan},
title = {Effect of Mesh Resolution and Data Sparsity on the Deformation Reconstruction and Strain Sensing using iFEM Methodology},
journal = {e-Journal of Nondestructive Testing},
year = {2026},
volume = {31},
number = {8},
doi = {10.58286/33675}
}
EUVIMED