Multiple phenomena occurring at the microscopic scale affect the final mechanical performance of composite parts manufactured through processes involving impregnation of dry fibers, such as resin transfer molding. Formation of fiber-poor areas in specific locations or air entrapment within the resin are issues that commonly arise during the impregnation. Such challenges have motivated the use of numerical simulations to understand the manufacturing processes better and to optimize the process design. However, the limitation imposed by their computational cost has encouraged the use of machine learning (ML) to replace them. Thus far, the state of the art has focused on predicting the permeability of fiber-reinforced microstructures. We expand the limits by proposing an ML-based surrogate for microscale steady-state velocity prediction of a fluid flowing through a fibrous microstructure. This model, inspired by the U-net architecture, takes as input the image representation of fiber-reinforced composite microstructures. It subsequently outputs the resin velocity field around the fibers based on prescribed boundary conditions. Those results are further used to estimate the permeability of the microstructures, thus encompassing previous works. We describe in this work the computational pipeline of our approach, starting from generation of the ground truth data to the optimization of the UNet hyperparameters.@en

This study presents the first steps of a benchmarking exercise on image processing of composite materials. Employing three distinct imaging protocols and five different image processing algorithms, the research explains the variability in capturing microstructural features of common micrograph dataset. Results highlight the sensitivity of different methods to factors like illumination inhomogeneities and pixel density, influencing the accuracy and consistency of obtained results. By comparing methodologies from different researchers in a blind format, the study identifies strengths and limitations, laying the groundwork for future benchmarking activities. Moving forward, this research sets the stage for standardized protocols and guidelines, aiming to enhance the reproducibility and reliability of microstructural analysis in composite materials. Such efforts are crucial for advancing material design and development, ultimately creating tailored composite materials with enhanced performance and functionality.@en

The transverse permeability of roving/tow-based fiber reinforcement is of great importance for accurate flow modeling in the pultrusion process. This study proposes an experimental approach to characterize the roving-based fiber beds' permeability under different compaction conditions. The experimental permeability results of thick roving-based preforms were reported and compared with the permeability values of roving-based preforms in the literature. A representative preform was infused under vacuum conditions. Its thickness was varied to replicate the different compaction values observed in permeability tests. Micrographs were then collected from it and analyzed to highlight the microscale transformations caused by processing/compaction on the fiber arrangement. The analysis revealed that compaction resulted in the reorganization of filaments along the direction of the applied compaction. Overall, the uniformity of the spatial filament distribution, i.e., the homogeneity within the fibrous domain, increased with increasing compaction. Furthermore, the microstructural analysis demonstrated transverse anisotropy within the tested domains, indicating that the obtained permeability results represented an upper boundary. In addition to the experimental analyses, various transverse permeability models, which were developed based on recently introduced statistical descriptors of fiber distribution, were evaluated by using the statistical descriptors extracted from the analyzed cross-sections. Among these models, the one correlating the second neighbor fiber distance with apparent permeability exhibited good agreement with the experimental results. Highlights: Transverse permeability measurement of a roving-based reinforcement was presented. The influence of compaction on the microstructure was investigated at the filament level. Filament distribution in a pultruded profile was analyzed by using statistical descriptors. The results of the experiments and the models in the literature were compared. The correlation between microstructural features and apparent permeability was discussed.

@en

Transverse permeability is a key parameter to construct accurate process modelling frameworks for composite manufacturing processes whereby dry reinforcements are impregnated by a resin. The transverse permeability of fibrous structures is strongly correlated with the spatial fiber distribution which is inevitably irregular through the cross section. This irregularity and possible resin channel formation through the cross section cause complex permeability field. In this study, we analysed the correlation between the microstructural features and the apparent transverse permeability. Multiple regions of interest were collected from a pultruded part’s cross section that exhibited apparent resin rich layers. Apparent transverse permeability values of these Regions of Interest (ROIs) were calculated by using OpenFOAM. The principal components of the transverse were calculated, and the statistical descriptors were captured for the corresponding ROIs. The principal permeability component analysis showed that principal directions are in-line with the visual predictions. The correlation between the permeability components and the statistical descriptors shows some traces of the capability of predicting the transverse permeability and the principal components based on the microstructural features. A larger dataset and more elaborate statistical descriptors are needed for accurate prediction of the flow behaviour and the apparent permeability components.@en

We propose a methodology to monitor the progressive saturation of a non-translucent unidirectional carbon fabric stack through its thickness by means of X-ray radiography and extract the dynamic saturation curves using image analysis. Four constant flow rate injections with increasing flow speed were carried out. These were simulated by a numerical two-phase flow model for both capillary and viscous leading flow conditions. The hydraulic functions describing pressure and relative permeability versus saturation were determined by fitting the saturation curves using a heuristic optimization routine. As the fluid velocity increases and the flow regime at the flow front shifts from capillary to hydrodynamically driven, the resulting capillary pressure curves for a given saturation level are shifted to higher values, from negative to positive. These as well as the capillary pressure calculated from the pressure drop within the unsaturated region of the fabric correlate well with a corresponding change in the averaged dynamic contact angle.

@en

Three different J-integral formulations to derive the experimental translaminar toughness of composites from compact tension tests with a large-scale fracture process zone are implemented and discussed. They improve the existing approaches by taking advantage of stereo-digital image correlation to acquire full-field displacement fields. A field fitting procedure based on robust and efficient piecewise cubic smooth splines addresses noise-related issues reported in previous studies. Additionally, the paper proposes a novel crack tip extraction procedure to report the energy release rate as a function of the crack increment, even if knowledge of the crack tip is not required for the proposed J-integral method. The three methods are discussed in light of a parametric study conducted on synthetic and experimental data, including artificially noisy data. The study reveals that the proposed J-integral methods are suitable for translaminar toughness evaluation of a wide range of materials without the need for restrictive assumptions. However, variations in propagation values were observed when applied to experimental data. Finally, guidelines are drawn to chose the most suitable parameters for the algorithms that are proposed as a Python package.

@en

Permeability of fibrous microstructures is a key material property for predicting the mold fill times and resin flow path during composite manufacturing. In this work, we report an efficient approach to predict the permeability of 3D microstructures from deep learning based permeability predictions of 2D cross-sections combined via a circuit analogy. After validating the network's predictions in 2D and extending it to 3D, we investigate its capabilities for handling images of various sizes obtained from virtual and real microstructures. More than 90% of 2D predictions is within ± 30% of their counterparts obtained via flow simulations, similarly for 3D transverse permeability predictions, while in 3D case computational time is reduced from several thousands of seconds to less than 10 s. This work provides a robust and efficient framework for characterizing the permeability of fibrous microstructures and paves the way for extending this capability to estimate the permeability of fabric mesostructures.

@en

Knowledge of permeability of fibrous microstructures is crucial for predicting the mold fill times and resin flow path in composite manufacturing. Herein we report a method to rapidly predict the permeability of 3D fibrous microstructures. Our method relies on predicting the permeability of 2D cross-sections via deep neural networks and extending this capability to 3D microstructures via circuit analogy as a means of reduced order modeling. Approximately 50% of the permeability predictions of 2D cross-sections have 10% or less deviation from the permeability results obtained via flow simulations in Geodict. Computational time required for predicting the permeability of 3D microstructures is reduced from hours to less than 10 seconds. This framework enables fast and accurate prediction of micro-permeability and serves as the first building block towards prediction of fabric mesostructures’ permeability via deep learning based methods. @en