Characterization and multi-scale modeling of the transverse compression of thick RTM-processed uni-directional samples

Chevalier, Jérémy;Morelle, Xavier;Camanho, Pedro;Lani, Frédéric;Pardoen, Thomas
(2017) 11th International Conference on Advanced Computational Engineering and Experimenting — Location: Vienna, Austria (3.July.2017)

Files

abstract_ACEX2017_VIP034.pdf
  • Open Access
  • Adobe PDF
  • 9.92 KB
Download

Details

Authors
Abstract
Bottom-up multi-scale approaches of composites involve a first scale-transition from the constituents at the microscopic scale to the behavior of the yarn and the ply at the mesoscopic scale. This contribution presents the results of a bottom-up approach which aims at a high fidelity prediction of the non-linear transverse deformation of the unidirectional plies for a composite processed by RTM. Most of the so-called uni-directional laminates produced by RTM are actually quasi-UD architectures. Hence, a low volume fraction of reinforcement is present in the in-plane transverse direction, hindering the first scale transition from the characterization of RTM epoxies. To overcome this limitation, an original processing route was developed for the manufacturing of true uni-directional samples by RTM. A specific mould was designed and the cure cycle of RTM6 was adapted to allow for curing thick samples without the risk of temperature overshoot. Yarns were unstitched from non-crimp fabrics and were introduced in the mould. Cubic samples were machined out of the UD slab and subjected to uniaxial transverse compression. Representative volume elements (RVE) of the UD material were built in 2D and 3D, based on a random generation algorithm. The interface between each fibre and the resin was modeled using cohesive zone elements. The deformation behavior of the RTM6 was assumed to be strain-rate, temperature and pressure dependent, using the constitutive model identified in [1]. As a first approximation, the fibres were considered as linear elastic. The explicit finite element simulations highlight the lack of accuracy of micro-mechanical analyses to predict the behavior of composites when the response is mainly dictated by the matrix. Following this observation, the in-situ behavior of the matrix and the constraining effect of the surrounding fibers were investigated using nanoindentation, digital image correlation and by manufacturing and testing UD specimens with a wide range of different carbon fiber contents. Both factors were shown to significantly influence the mechanical response of the matrix within the composite. As a matter of fact, well-identified and validated models for the mechanical behavior of bulk epoxy resins are not necessarily sufficient to obtain satisfactory predictions of its behavior when surrounded by fibers in a composite.
Affiliations

Citations

  • APA
  • Chicago
  • FWB

Chevalier, J., Morelle, X., Camanho, P., Lani, F., & Pardoen, T. (2017). Characterization and multi-scale modeling of the transverse compression of thick RTM-processed uni-directional samples. 11th International Conference on Advanced Computational Engineering and Experimenting, Vienna, Austria. https://hdl.handle.net/2078.5/49428