TL;DR
This paper investigates how the "draping effects" (waviness, gapping, shearing) during composite part manufacturing alter the local fiber volume content (FVC) of unidirectional non-crimp fabrics (UD-NCF). The authors developed experimental tools and simulation methods to accurately predict these FVC changes, crucial for reliable structural performance prediction of composite components.
What Problem Does This Paper Address?
The structural integrity of fiber-reinforced polymer (CoFRP) parts is highly dependent on two key factors: fiber orientation and fiber volume content (FVC). While predicting fiber orientation during the complex "draping" process (forming 2D fabrics into 3D shapes) has seen progress with simulation tools, accurately determining the local FVC has remained a challenge. Draping can introduce various deformations like waviness, gapping, and shearing, which alter the fabric's structure and, consequently, its FVC. These changes are critical because they directly impact the material's stiffness and strength. Existing methods for measuring local FVC are often laborious (e.g., non-destructive testing), making draping simulations a more practical approach. However, accurately modeling these draping effects and their impact on FVC, especially for UD-NCF which deforms differently than woven fabrics, requires robust methods. This paper aims to bridge this gap by providing a comprehensive approach to experimentally and numerically determine the local FVC under various draping conditions.
Methods
The researchers employed a multi-pronged approach:
- Experimental Fabrication: They developed specialized tools to create fabric samples exhibiting specific draping effects:
- Waviness: A modified slider mechanism was used to induce sinusoidal waves with controlled amplitude and wavelength.
- Gapping: A similar mechanism, but with spacer strips, was used to create gaps between fiber rovings.
- Shearing: A shear frame was employed to apply pure shear deformation.
- A powder binder was used to "freeze" these deformations in the preforms.
- Experimental FVC Determination: The FVC of the deformed samples was determined experimentally by measuring their areal weight and assuming a constant laminate thickness.
- Analytical Modeling: Analytical equations were derived to predict FVC changes for each draping effect (waviness, gapping, shearing) based on geometric assumptions and material properties.
- Numerical Simulation: A macroscopic material model, previously developed by Schirmaier et al., was utilized within ABAQUS to simulate the draping process. This model accounts for large shear and transverse strains characteristic of UD-NCF. The simulation setup involved applying specific boundary conditions to replicate the experimental draping effects. The FVC was then calculated from the simulated element area changes.
- Interdependency Analysis: The paper also investigates how these draping effects interact at arbitrary strain states, focusing on the influence of perpendicular strain on FVC.
Datasets
Not explicitly discussed in terms of "datasets" in the machine learning sense. The paper relies on:
- Experimental Data: Fabric samples manufactured with controlled waviness, gapping, and shearing.
- Material Properties: Characterization of the Zoltek PX35 UD-NCF (areal weight, fiber density).
- Simulation Outputs: Strain components and element area changes derived from the FEA simulations.
The paper does not introduce a new benchmark dataset but rather generates experimental data to validate its proposed methods.
Key Insights and Ideas
- Draping Effects Significantly Alter Local FVC: The core finding is that waviness, gapping, and shearing are not just geometric distortions but directly and measurably change the local fiber volume content. This change can be substantial (5-13% in their experiments).
- Stitching Yarn is a Key Influencer: The polyester stitching yarn in UD-NCF plays a crucial role, significantly influencing deformation behavior and sometimes causing unintended undulations, especially in wavy and gapped samples.
- Perpendicular Strain is a Strong Indicator of FVC Change: The study highlights that the strain component perpendicular to the fiber direction is a critical factor in determining FVC changes, particularly when draping effects occur simultaneously.
- Simulation Offers Localized FVC Prediction: While analytical and experimental methods provide global FVC values, draping simulations are essential for predicting localized FVC variations in complex, curved composite parts.
- Macroscopic Model Sufficiency (with caveats): The macroscopic material model used is capable of capturing the essential membrane behavior and FVC changes, provided mesoscopic effects like significant roving deformation or out-of-plane undulations are minimal.
Key Results
- Experimental Validation: The developed experimental tools successfully created samples with controlled draping effects.
- Good Correlation: The experimental results showed good correlation with both analytical predictions and numerical simulation results for FVC.
- Waviness: FVC increased with increasing amplitude-to-wavelength ratio (A/λ). Deviations between prediction and experiment increased beyond A/λ = 0.1 due to complex fabric behavior.
- Gapping: FVC decreased with increasing gap width. Analytical and simulation results showed good correlation with experiments, with slight deviations for larger gaps.
- Shearing: FVC increased with shear angle. The stitching yarn had less influence here compared to waviness and gapping.
- Simulation Accuracy: For uniform draping effects (gapping, shearing), simulations provided reliable FVC predictions. For more complex deformations like waviness, deviations increased, particularly at higher A/λ ratios.
Limitations and Future Work
- Mesoscopic Effects: The macroscopic simulation model struggles to capture fine-scale mesoscopic effects like significant roving deformation or out-of-plane undulations, which can lead to deviations from experimental results, especially in wavy patterns.
- Stitching Yarn Complexity: The behavior of the stitching yarn, particularly its influence on inducing secondary undulations, is a significant factor that can limit the accuracy of predictions for certain draping effects.
- Future Work: The authors suggest further experimental tests on infiltrated specimens with uniformly and non-uniformly distributed draping effects to validate the continuous virtual process chain for predicting structural performance. They also imply that refining the simulation to better capture mesoscopic effects could improve accuracy.
Summary updated