Data assimilation for three-dimensional flow monitoring in non-flat composite structures during vacuum-assisted resin transfer molding: A numerical study

摘要

This study set out to investigate the applicability of data assimilation to the monitoring of the flow of non-planar 3D shapes through the integration of visual observations with a stochastic numerical simulation of the resin flow in a vacuum-assisted resin-transfer molding process. First, we investigated the effect of the image resolution on the flow-monitoring estimation performance. By means of numerical experiments using images of various resolutions, we verified the applicable range of the image resolution in the proposed technique. Furthermore, the proposed technique was applied to 3D structures such as curved surfaces or ribbed structures. Using the visual observation values, the impregnation distances obtained from the image data were calculated in combination with the shape of the molding. The results confirmed that our proposed method is capable of successfully estimating the 3D impregnation behavior and permeability fields. (C) 2017 Elsevier Ltd. All rights reserved.