open access publication

Article, 2024

Layer-by-layer reconstruction of fatigue damages in composites from thermal images by a Residual U-Net

Composites Science and Technology, ISSN 1879-1050, 0266-3538, Volume 255, Page 110712, 10.1016/j.compscitech.2024.110712

Contributors

von Houwald, Benedict [1] Sarhadi, Ali 0000-0003-1078-493X [2] Eitzinger, Christian (Corresponding author) [1] Eder, Martin Alexander 0000-0002-5306-365X (Corresponding author) [2]

Affiliations

  1. [1] Profactor (Austria)
    2. [NORA names: Austria; Europe, EU; OECD];
  2. [2] Technical University of Denmark
    3. [NORA names: DTU Technical University of Denmark; University; Denmark; Europe, EU; Nordic; OECD]

Abstract

In this paper a deep learning model is used to fully reconstruct the 3D distribution of arbitrarily shaped subsurface fatigue damages in a fiber/epoxy composite from synthetic thermal surface images. Synthetic thermal surface images (TIs) of self-heating damage hotspots are produced by thermal finite element analysis which are consequently used to train a Residual U-Net based on recent architectures designed for image segmentation. Different augmentation techniques are employed to mitigate the computational cost of generating training data through thermal finite element analysis. The Residual U-Net model accurately reconstructed – layer by layer – the ground truths and thereby enabled the quantitative assessment of location, size, shape, depth and gradient of an internal fatigue damage distribution. Moreover, the Residual U-Net achieved good predictions for comparatively small training set sizes.

Keywords

U-Net, U-Net model, analysis, architecture, assessment of location, augmentation, augmentation techniques, composition, computational cost, cost, damage, damage distribution, damage hotspots, data, deep learning models, depth, distribution, elemental analysis, fatigue, fatigue damage, fatigue damage distribution, finite element analysis, gradient, ground, ground truth, hotspots, image segmentation, images, layer, layer-by-layer reconstruction, learning models, location, model, prediction, quantitative assessment, reconstruction, reconstruction layer, residual U-Net, residual U-Net model, segments, shape, size, subsurface fatigue damage, surface images, technique, thermal finite element analysis, thermal images, training, training data, truth

Funders

  • Danish Energy Agency

Data Provider: Digital Science

LINKS
-

Matching Records in NORA

SUBJECTS
+

METRICS
+