Predictive modeling of 3D textile composites using realistic micromechanical representations

Abstract

An innovative and efficient progressive damage model, informed by coupon tests, and using minimal, readily available material parameters, is proposed in this paper to comprehensively predict the tensile, compressive and shear stress–strain response and strengths of a 3D woven angle interlock composite within 5% of test data. Realistic, high-fidelity microgeometries, with process-induced defects, and devoid of geometric assumptions, are constructed using the digital element approach. Excellent correlations with test micrographs are observed for yarn cross-section areas, aspect ratios and crimp ratios, that are on average within 5.5%, 1.9% and 1% of the micrograph measurements respectively. The mesoscale predictions capture the detailed failure mechanisms under each loading condition. Analytical stiffness expressions are derived in terms of the constituent material elastic properties and architecture geometry, which also predict the stiffnesses within 10% error margin. A robust workflow is established that combines the high-fidelity microgeometry generation, automated conformal finite element model creation and comprehensive property prediction capabilities to provide a holistic solution.