Liquid Impregnation of Fiber Fabrics
Impregnated fiber fabrics - Generation of structures and analysis of mechanical and fluid dynamic properties
Resin injection is carried out to manufacture composite components by impregnating a textile semi-finished product with a liquid, thermosetting resin. The impregnation process is largely responsible for the component quality and, therefore, a precise understanding of the impregnation process is of great importance.
Simulations of filling at component level are established tools for process optimization, but only meaningful if the impregnation properties of the textile are known at micro- (or fiber) as well as at meso-level (or roving). For this, permeability may be measured in elaborate tests. A simulation at micro- and meso-level in GeoDict can significantly reduce this experimental effort. For this purpose, we have developed a workflow together with the Leibniz Institute for Composite Materials (IVW) as part of the Math2Composites project and implemented it in a GeoApp included in GeoDict:
- Import and analysis of a µCT scan of a glass fiber fabric.
- Determination of the permeability of the individual rovings.
- Generation of a tissue model in GeoDict, which becomes the digital twin of the glass fiber fabric by compression.
- Calibration of the digital twin to permeability measurements.
What were the project results?
- Using µCT scan data, textiles are represented faithfully in GeoDict.
- Compaction of textiles is precisely simulated by the mechanics simulation in GeoDict.
- Permeabilities are predicted with good accuracy by means of flow simulations in GeoDict.
What do these results mean for GeoDict users?
- A simple GeoApp is available to determine the permeability of the textile.
- All findings have been incorporated into the development of GeoDict.
- Scientific benefit from the journal publications of our partners.
In GeoDict, over 25 file formats are supported and are implemented within the ImportGeo-Vol module.
- Using the ImportGeo-Vol module, all desired data are loaded with just one click, and the desired voxel length is individually adjusted.
- 3D scans are segmented quickly and easily, and the image quality is subsequently improved using appropriate filters.
- Desired parameters, such as the selection of materials or the segmentation threshold, are edited even after segmentation and the structure is aligned or rotated accordingly.
- In GeoDict, the generated structure is displayed in overall and in section views, as well as in 2D or 3D, i.e. in detail from all sides and angles.
µCT scans are therefore quickly and easily loaded into GeoDict and mapped as a gray-scale image. Transparency and gray scale range are also adjustable. The gray scale image is enhanced and adjusted using a variety of image filters available in GeoDict. The quality of the image is thus optimally enhanced, making it a snap to segment in the next step.
In addition, contrast, brightness, or possible flickering are corrected.
During segmentation, a material ID is assigned to the fibers and a structure is obtained that can be used for any required calculation or simulation. Since GeoDict works directly on voxel structures, no meshing of the fibers is necessary and simulations are performed directly on the segmented structure.
Using the data sheets of the original scrim, the various fabric layers are created using the FiberGeo module. The different directions of the layers (0°and 90°), which alternate in the layer structure, are taken into account. After the structure of the individual layers has been created, the mechanical properties are calculated on both layers through a mechanical simulation using the ElastoDict module. These computations are an important part of the subsequent simulations on the entire fabric.
The mechanical simulations are performed directly on the µCT-scans of the uncompressed samples without the need for mesh generation, by using the FFT-based voxel solver FeelMath integrated in the ElastoDict module. In addition, this solver allows the simulation of large compressions of fibers, which is often a challenge with standard FE-based solvers.
The properties of the fabric are calculated directly and easily after the structure generation. Using the FlowDict module, the permeability tensor is obtained, as well as the values for the mean flow velocity at a given pressure drop, the estimated Reynolds number, and the Gurley value in the selected calculation directions.
Graphs for pressure, speed, and convergence are included in the results.
Normally, a fiber fabric is inserted into the prepared mold and compacted. This experimentally-performed compacting is simulated easily in GeoDict with the mechanical solver in the ElastoDict module. A dedicated GeoApp has been developed by Math2Market to make the handling as simple as possible. This GeoApp has a variety of common workflows and additionally, allows for the possibility of creating own workflows and saving them as Python scripts. Thus, any workflow may be repeated or modified easily and quickly.
From the data generated in Part 1 and Part 2, the entire fiber architecture is then displayed in GeoDict. The number of layers, the distances between the layers, and the height and width of the glass fibers are input for the GeoApp to obtain an exact image of the architecture. The GeoApp may also compress the generated structure, which allows the real process of textile compaction to be simulated realistically.
In addition, if more data is needed, a variety of simulations may be performed. In this case, a flow simulation for permeability determination was performed on the compressed glass fiber architecture. For example, the injection process of the epoxy resin during the RTM process may be exactly simulated.
The compressed fibers are clearly seen in the µCT-scan and are reproduced as a compressed 3D structure using GeoDict. Using the digital model, the effects on the compressed structure are directly simulated and computed.
Our cooperation partner IVW carried out the permeability measurements on the fiber fabric. The results obtained were calibrated to 50% compression and subsequently validated. The calibrated model can also be successfully transferred to 60% compression in the following. The comparison of experimental and simulation results shows only a slight deviation of 27%.
We would like to thank our partners at the Leibniz Institute for Composite Materials (IVW) for the excellent and pleasant collaboration, and for providing the µCT scans and experimental data.
The project "Math2Composites" was funded by the ZIM program of the German Federal Ministry for Economic Affairs and Energy (BMWi),
Funding code: ZF40523110EB6
- Schmidt T. et al., A Novel Simulative-Experimental Approach to Determine the Permeability of Technical Textiles, Key Engineering Materials, vol. 809, Trans Tech Publications, June 2019, 1662-9795.
- Rimmel O. et al., Modeling transverse micro flow in dry fiber placement preforms, Journal of Composite Materials, Nov. 2019, 1691-1703, https://doi.org/10.1177/0021998319884612.