Performance evaluation of two progressive damage models for composite laminates under various speed impact loading

• Published in: International journal of impact engineering 2020-09, Vol.143, p.103615, Article 103615
• Main Authors: Sridharan, Sriraghav, Pankow, Mark
• Format: Article
• Language: English
• Subjects: Composites; Composites Failure; Computer simulation; Damage assessment; Damage localization; Finite element analysis; Finite element method; Graphite fiber reinforced plastics; Graphite-epoxy composites; Impact; Impact damage; Impact loads; Laminates; Layers; Mathematical models; Performance evaluation; Progressive damage; Projectiles; Simulation; Velocity
• ISSN: 0734-743X; 1879-3509
• DOI: 10.1016/j.ijimpeng.2020.103615

•Comparison of two computational models at different impact speeds to understand the effects of model choice on results.•Through thickness fiber crush is a dominant failure mode under high velocity impact.•Indentation simulations using solid elements are prone to hour glassing.•Rise in hourglass energy in solid element results in decrease of contact force.•Even models developed for high rates don't accurately capture the full failure response of the material systems. The focus of this paper is to explore two different finite element modeling strategies to understand how well they can be used to characterize different types of impacts. In this paper, finite element numerical simulations of both low velocity and high velocity projectile impact on composite laminates have been carried out. Low velocity impact was based on 3 mm thick 24 ply unidirectional graphite epoxy laminates. High velocity impacts were performed on a laminated S2 Glass with SC-15 resin, since large amounts of material data, and experimental and computational results are available for validation. Two separate progressive material damage models were investigated in the commercial finite element codes Abaqus/CAE and LS-Dyna. The Abaqus/CAE models used a VUMAT subroutine while MAT 162 was chosen due to its wide adoption for impact problems in LS-Dyna. For high velocity projectile impact simulations MAT 162 captures the correct velocity but misses the spread of damage with localization. While the Abaqus/CAE model underperformed due to the lack of modeling the through thickness shear type failures. For Low velocity impact simulations the Abaqus/CAE model showed a good correlation with experimental data while MAT 162 struggled. The modeling limitation imposed by MAT 162 restricting the model to only reduced integration solid elements was found to induce a mesh instability making the model prone to hourglassing. This was specifically a challenge in modeling laminates with a greater number of plies or more representative of structures which might be subjected to impact, which was not an issue at the higher energy.