Systematic Approach for Finite Element Analysis of Thermoplastic Impregnated 3D Filament Winding Structures—Advancements and Validation

Jonathan Haas, Daniel Aberle, Anna Krüger, Björn Beck, Peter Eyerer, Luise Kärger, Frank HenningJournal of Composites Science 2022 · DOI

TL;DR

This paper refines and validates a systematic finite element analysis (FEA) approach for thermoplastic 3D filament wound structures (fiber skeletons). By incorporating more detailed geometric considerations and comparing meso- and macro-level modeling strategies, the authors provide a robust framework for predicting the mechanical behavior and failure of these lightweight, complex structures.

What Problem Does This Paper Address?

3D filament winding (3DSW) is a promising manufacturing technique for creating lightweight, topology-optimized composite structures. However, accurately predicting their mechanical performance, especially under tensile loads, is challenging due to their complex geometries and the intricate interactions between fibers and inserts. Previous FEA approaches often relied on simplified geometric models, which could lead to inaccuracies, particularly in predicting failure mechanisms at critical locations like fiber deflection points. This paper addresses the need for a more refined and validated FEA methodology that can accurately capture the behavior of these structures, bridging the gap between simplified models and computationally expensive detailed simulations.

Methods

The core of this work is the advancement and validation of a systematic FEA approach for 3D filament wound structures. Key methodological contributions include:

  • Enhanced Geometric Modeling: The study refines the geometric modeling of fiber skeletons by identifying and incorporating additional characteristic dimensions (e.g., local loop thickness, contact width) that were previously overlooked but found to influence mechanical behavior.
  • Meso- vs. Macro-Level Modeling: The paper systematically compares two FEA modeling scales:
  • Meso-level: Each individual winding is modeled, allowing for the consideration of interfaces like the roving-roving interface, which is crucial for capturing delamination.
  • Macro-level: Windings are grouped, neglecting the roving-roving interface to significantly reduce computational effort. This approach is investigated for its suitability when delamination is not a primary concern.
  • Material Characterization and Modeling: Experimental characterization of the structural constituents (impregnated roving, interfaces, inserts) is performed. Material models are selected and parameterized, with specific attention paid to adapting the fiber-parallel Young's modulus and tensile strength based on experimental observations of simple loop specimens.
  • Friction Modeling: The insert-roving interface, crucial for inclined loops, is characterized using a Coulomb friction model, accounting for sliding behavior.
  • Validation: The developed approach is validated through experimental tensile tests on both simple loop specimens and more complex, three-dimensional inclined loop specimens.

Datasets

The paper utilizes experimental data generated from tensile tests on:

  • Simple Loop Specimens: These were tested with varying numbers of windings (two, six, and ten) and also in a variation with artificially created fiber twist.
  • Inclined Loop Specimens: These more complex, three-dimensional structures were also subjected to tensile testing.

Additionally, friction tests were conducted on pressed PP-GF plate specimens to characterize the insert-roving interface. The material properties were derived from these experiments and supplemented by existing literature data.

Key Insights and Ideas

  • Geometric Details Matter: Even seemingly minor geometric dimensions, like local loop thickness and contact width, can significantly influence the predicted spring constant and failure load of fiber skeletons.
  • Meso vs. Macro: A Trade-off: While meso-level models offer higher fidelity, especially for capturing delamination, macro-level models provide a computationally efficient alternative when delamination is not a dominant failure mode. This allows engineers to choose the appropriate modeling strategy based on the expected behavior.
  • Material Property Adaptation is Crucial: Discrepancies between calculated and measured spring constants in simple loops highlighted the need to adapt material properties (specifically, the fiber-parallel Young's modulus) based on experimental observations. This empirical adaptation is vital for accurate FEA of these structures.
  • Failure Mechanisms Dictate Model Choice: The study clearly demonstrates that the presence and dominance of delamination significantly impact the accuracy of macro-level models. For structures prone to delamination, meso-level modeling is essential.
  • Fiber Twist's Subtle Influence: While increased fiber twist in the yarn can slightly reduce the stiffness of simple loops, its effect is less pronounced than initially expected, suggesting other factors might also contribute to observed deviations.

Key Results

  • Simple Loops:
    • Meso-models accurately reproduced the delamination-dominated failure behavior of simple loops with two windings, showing good agreement with experimental load-displacement curves.
  • Macro-models, while unable to capture delamination, provided reasonable predictions for spring constant and failure load in simple loops with six and ten windings, where delamination was less dominant.
  • Inclined Loops:
    • The macro-level FEA approach, incorporating advanced geometric modeling and adapted material properties, successfully predicted the mechanical behavior, failure location (fiber fracture at the small insert's wrapping flank), and sliding distances of inclined loop specimens.
  • The calculated total sliding distance at the inserts was within the upper range of experimental measurements, validating the friction model.
  • Recommendations: The study provides clear recommendations for measuring and modeling characteristic dimensions, guiding future FEA efforts.

Limitations and Future Work

  • Delamination Complexity: While meso-models can reproduce delamination, the paper notes that achieving perfect symmetry in delamination prediction might require further refinement, such as slightly increasing CWW on one end.
  • Material Property Simplification: The adaptation of the fiber-parallel Young's modulus is acknowledged as a temporary solution and a simplification that doesn't fully replace in-depth physical investigation.
  • Insert-Roving Interface: The friction model primarily considers static friction. The paper suggests that the smoother surface of the press-consolidated PP-GF plates used in friction tests might differ from the actual contact area of PP-GF windings, potentially influencing sliding distance predictions.
  • Future Work: The authors propose dedicating future work to investigating the influence of different process and design parameters on characteristic dimensions and applying the developed approach to optimize insert design. Further research into the underlying physical phenomena causing material property deviations is also suggested.

Summary updated

Abstract

This work aims to enhance and validate a systematic approach for the structural finite element (FE) analysis of thermoplastic impregnated 3D filament winding structures (fiber skeletons). The idealized modeling of geometrically complex fiber skeletons used in previous publications is refined by considering additional characteristic dimensions and investigating their mechanical influence. Moreover, the modeling approach is transferred from the meso- to the macro-level in order to reduce modeling and computational effort. The properties of meso- and macro-level FE models are compared using the example of simple loop specimens. Based on the results, respective application fields are defined. In the next step, the same modeling approach is applied to a more complex, three-dimensional specimen—the inclined loop. For its macro-level FE model, additional material characterization and modeling, as well as enhancements in the modeling of the geometry, are proposed. Together with previously determined effective composite properties of fiber skeletons, these results are validated in experimental tensile tests on inclined loop specimens.
BibTeX
@article{haas2022systematic,
  author = {Jonathan Haas and Daniel Aberle and Anna Krüger and Björn Beck and Peter Eyerer and Luise Kärger and Frank Henning},
  title = {{Systematic Approach for Finite Element Analysis of Thermoplastic Impregnated 3D Filament Winding Structures—Advancements and Validation}},
  journal = {Journal of Composites Science},
  year = {2022},
  doi = {10.3390/jcs6030098},
}
Figure