A fairly priced, unfitted spline image-based model and its first use for data assimilation in mechanics of materials - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2021

A fairly priced, unfitted spline image-based model and its first use for data assimilation in mechanics of materials

Résumé

Recent advances in micro-tomography provide a unique opportunity to model and characterize the behavior of complex micro-structures. On the one hand, the acquired volumetric images can be used to build representative finite element Digital Image-Based (DIB) models [1, 2]. On the other hand, the use of image-based kinematic measuring tools such as image registration (or more commonly named Digital Image Correlation (DIC) in experimental solid mechanics) offers the opportunity to estimate displacements at the heart of complex materials. However, DIC is still limited for volumetric image data due to the almost absence of texture. In this work, two main contributions are performed: (i) construction of a fairly priced, unfitted, spline DIB model [3] and (ii) development of an isogeometric, mechanically sound, regularized DIC algorithm. Regarding (i), a special care is taken to analyze the numerical approximation error while taking into account the intrinsic geometry error resulting from the image generation process (sampling + quantization). This study thus breaks with the usual practice in the field where the modeling error and the convergence of numerical methods are most of the time assessed with geometries deemed to be exact [1, 2]. Concerning (ii), we show that it is possible to estimate displacements and their derivatives within the strut thickness of a mechanically stressed microstructure by combining the mechanical regularized DIC scheme of [4] with the developed fairly priced DIB model. No classical method of the literature has been able to correctly estimate local strains in such situations. Synthetic as well as experimental image data will be considered to assess the performance of our method in both the linear and in non-linear regimes.
Fichier non déposé

Dates et versions

hal-03282365 , version 1 (09-07-2021)

Identifiants

  • HAL Id : hal-03282365 , version 1

Citer

Ali Rouwane, Robin Bouclier, Jean-Charles Passieux, Jean-Noël Périé. A fairly priced, unfitted spline image-based model and its first use for data assimilation in mechanics of materials. YIC 2021 - VI ECCOMAS Young Investigators (virtual) Conference, Jul 2021, Valencia, Spain. ⟨hal-03282365⟩
39 Consultations
0 Téléchargements

Partager

Gmail Facebook X LinkedIn More