Automated image-based geometric modelling of thick laminates with three-dimensional ply wrinkles

Ply wrinkles are commonly induced during thermal-mechanical manufacturing process of laminated composite structures or components. Fibre misalignment due to ply wrinkles compromises the structural integrity of a composite structure. The damage criticality can be assessed numerically based on non-destructive testing results. Geometric modelling is usually required and it is time-consuming if complex wrinkles have to be incorporated into the finite element model. This study presents an automated image-based three-dimensional geometric reconstruction method for thick laminates with complex ply wrinkles. The geometry of a thick laminate is reconstructed based on the limited inspection images from the sides of laminates. The two-dimensional geometric reconstruction of the detected wrinkled plies is done using a multiple Gaussian functions model or linear interpolation, while the intermediate plies are defined by distance-based interpolation. Based on the reconstructed two-dimensional geometries, the three-dimensional ply wrinkles are modelled via an inverse distance weighting interpolation method. To verify the proposed method, a thick glass fibre laminate with artificially induced wrinkles is manufactured and the cross-sectional geometries of the reconstructed model are compared well with and the realistic inspection data. Progressive failure analysis of a two-dimensional and a three-dimensional reconstructed model are compared in terms of damage progression and load versus nominal strain response to highlight the relevance of modelling three-dimensional wrinkles to obtain more accurate and efficient damage criticality assessments.

Automated summary

An automated image-based three-dimensional geometric reconstruction method is presented for thick laminates with complex ply wrinkles, which can accurately and efficiently assess the damage criticality of laminated composite structures. The method reconstructs the geometry based on limited inspection images, using multiple Gaussian functions model or linear interpolation for wrinkled plies and distance-based interpolation for intermediate plies. The proposed method is verified through comparison with realistic inspection data, demonstrating its accuracy and practicality.