COMPOSITES 2023

An assessment on the effect of draping of 3D woven composites utilising a numerical and empirical material characterisation approach

  • Topalidis, Ioannis (Bristol Composite Institute (BCI))
  • El Said, Bassam (Bristol Composite Institute (BCI))
  • Thompson, Adam J (Bristol Composite Institute (BCI))
  • Keulen, Jerome (Eindhoven University of Technology)
  • Hallett, Stephen R (Bristol Composite Institute (BCI))

Please login to view abstract download link

3D woven composites ability to produce near net-shape preforms minimizes the labor steps, thereby fullfiling the industry needs for production time and cost reductions. In addition, these materials exhibit desirable mechanical characteristics, such as enhanced delamination resistance and energy absorption, a consequence of the complex internal yarn architecture. However, this distinct yarn architecture is prone to geometric irregularities induced during the manufacturing stages, which can substantially undermine the structures’ integrity. Of interest in this study is in-plane shear deformations and irregularities arising from draping the preform to conform to a curved mold. The underlined dependency of the material performance on mesoscopic features of the internal yarn architecture imposes the need to include key details of this length-scale in the numerical characterisation models. A set of unit cell, meso-scale models of the compacted, as-woven material geometry are generated using an in-house kinematic model. An energy regularised, smeared crack formulation predicts material failure in the yarn and matrix material independently. The progressive nature of failure in this class of materials demands careful consideration of the stress relief mechanisms, encompassed in the material models by an appropriate unloading regime. Three shear angles are selected, 0°, 8° and 16°, to be studied under tensile loading. Experimental verification is provided by tests on VARTM manufactured coupons. The empirical data 2D strain distribution for each case is compared to the DIC measurements, while the empirical stress-strain curves serve to further validate the model results in the elastic region. High fidelity computed-tomography scans on interrupted test specimens aid a deeper comprehension of the dominant damage modes and further verify the numerical model. Behavioural patterns due to the material pre-shearing are identified and the overall capabilities and shortcoming of the proposed meso-scale framework are highlighted.