Hybrid Experimental and Numerical Characterization of the 3D Response of Woven Polymer Matrix Composites

Abstract

The need for advanced material models to simulate the deformation, damage, and failure of polymer matrix composites under impact conditions is becoming critical as these materials are gaining increased usage in the aerospace and automotive industries. The purpose of this work is to characterize carbon epoxy fabrics for composite material models that rely on a large number of input parameters to define their nonlinear and 3D response; e.g. elastic continuum damage mechanics models or plasticity damage models [1, 2]. It is challenging to obtain large sets of experimental stress-strain curves, therefore, careful selection of physical experiments that exhibit nonlinear behavior is done to significantly reduce the cost of generating threedimensional material databases. For this work, plain weave carbon fabrics with 3k and 12k tows are manufactured by VARTM. Testing is done using MTS hydraulic test frames and 2D digital image correlation (DIC) to obtain experimental stress-strain curves for in-plane tension and shear as well as transverse shear. For cases where actual experimental data is either not available or difficult to obtain, the required model input is virtually generated using the NASA Glenn developed Micromechanics Analysis Method/Generalized Method of Cells (MAC/GMC) code. A viscoplastic polymer model is calibrated and utilized to model the matrix constituent within a repeating unit cell (RUC) of a plain weave carbon fiber fabric. Verification and validation of this approach is done using MAT213, a tabulated orthotropic material model in the finite element code LS-DYNA, which relies on 12 input stress-strain curves in various coordinate directions [2]. Based on the model input generated by the micromechanics analyses in combination with available experimental data, a series of coupon level verification and validation analyses are carried out using the MAT 213 composite model.