A finite deformation Cosserat continuum model for uncured carbon fibre composites

Abstract

A new three-dimensional, finite deformation Cosserat continuum model for the elastic response of uncured carbon fibre composites is presented. The new composite process model captures the bending contribution of bundles of fibres at the microscale within a mesoscale continuum description of a composite ply. This is achieved by introducing higher-order, independent rotational degrees of freedom into the continuum formulation. This paper demonstrates the inclusion of such mechanics is essential to accurately model various bending responses induced during typical composite manufacturing processes. This includes large deformation forming, finite strain consolidation and wrinkling (the formation of an unwanted defect). If such mechanics are not included, the literature demonstrates the resulting finite element solutions have a pathological dependence on the mesh size. As a result, simulations require users to fit mesh-dependent material parameters, which limits confidence in their predictive capabilities. The Cosserat continuum, which can be seen as a form of the regularised continuum model, overcomes these challenges. In particular, this paper presents details of the finite element formulation of the new continuum model within a nonlinear Taylor–Hood Cosserat Element. Implementation details of embedding this new element within the commercial code Abaqus are given, alongside a series of increasingly complex validation simulations. Notably, the examples include modelling the formation of internal fibre wrinkles and large deformation forming, which involves complex ply-to-ply and tool-to-ply contact. The paper concludes by describing: (1) how the elastic Cosserat model can be integrated into existing large deformation process models in the literature. The approach set out readily allows researchers to include the important effects of resin flow, cure kinetics and temperature distribution, not considered in this contribution, and (2) how it is envisaged that the ply scale model can be naturally scaled up to large laminate scale simulation using mathematical upscaling techniques.