Experimental and Numerical Simulation Studies of Failure Behaviour of Carbon Fibre Reinforced Aluminium Laminates under Transverse Local Quasi-static Loading

Abstract

This paper investigated the failure behaviour of medium thick carbon fibre reinforced aluminium laminates (CARALL) panels under transverse local quasi-static contact crush experimentally and numerically. Four types of CARALL specimens with a 3/2 configuration were manufactured by hot-pressing process using two type of aluminium alloys (2024-T3, 7075-T6 aluminium alloy) and CFRPs with different lay-up ([0°/90°/0°]3, [45°/0°/-45°]3). Three dimensional Hashin progressive failure model in quadratic strain form with damage evolution laws defined by subroutine VUMAT based on ABAQUS was used to simulate the composite layers. The bilinear cohesive contact model was used to predict interfacial failure. Aluminium alloys were defined by Johnson-Cook model. Numerical predictions composing of load-deflection curves and failure mode were matched closely with experimental results. The results of the analysis shown that delamination in CFRP/Al interfaces was the initial damage of the tested CARALL panels. Matrix failure occurred after the delamination closely. The maximum load bearing was determined by fracture of aluminium alloy. Fibre breakage was responsible for complete failure of specimens.

A Review on Finite-Element Simulation of Fibre Metal Laminates

“…Different types of elements are used to model fibre-metal laminates. The most common approach is the three-dimensional solid model [6,19,20,22,27,29,38,[51][52][53][54][55] with usage of elements like C3D8R (eight-node linear brick, reduced integration, hourglass control) in Abaqus or SOLID184 /185/186 in ANSYS. Authors in [30] tested the usage of higher-order elements (C3D20R) but from the perspective of time, it is not necessary as accuracy gain is negligible in comparison to additional computational cost.…”

“…For most popular metals in FMLs such as aluminium and magnesium, the plastic behaviour is important and has a big impact on the response of the whole material. Different types of metal material models, including plasticity, are used: bilinear, isotropic hardening (multilinear) [8,27,30,67] with rate dependence [52], and with the Johnson-Cook constitutive model [6,10,17,18,51,64,68] also in simplified form [69]. Researchers use strain rate-dependent models (also including hardening) are used by researchers for different problems, including lowvelocity impact, which is a very popular test.…”

“…In composite materials, general-purpose failure criteria capable of determining failure in a few modes are common: Hashin criterion in Abaqus [6,8,53] or Ansys [46] and Chang-Chang criterion in LS-Dyna [21,33,69]. It should be noted that the Hashin criterion is available in Abaqus only for twodimensional models (including shells) so researchers prepare user-defined subroutines for solid 3D models [9,20,28,51], 75. In addition, other criteria are used that only predict damage initiation, such as Puck [46], Tsai-Wu [46], or Tsai-Hill [46].…”

“…More information about friction in models can be found in further sections. Examples of cohesive surfaces as a way to model interfaces can be found in: [6,17,25,51,52].…”

“…Similarly to standard materials simulations, the mesh is often finer in the region of applied boundary conditions or mechanical notches and coarser in other regions to save computational costs [24]. The limitation of computational capabilities and the decrease in simulation time are factors that influence the rationale for mesh size [51]. However, a too coarse mesh may be the reason for result errors.…”