MODELLING AND TESTING THE CRASH BEHAVIOUR OF COMPOSITE VEHICLES COMPONENTS

Other conference contribution, 2019

In the current contribution we will present the latest developments in the project “Modelling crash behaviour in future lightweight composite vehicles – Step 2”, involving 11 Swedish partners. On the material modelling side, a fully three-dimensional model to describe fibre kinking has recently been developed. The model is physically based and considers the fibre rotation during kink-band formation under large deformations. The FE implementation of the model is straightforward which allows for easy implementation. The validation of the model for stiffness and strength shows good correlation with the experiments. The influence of initial misalignments on the stiffness is well captured, the strength defined at the onset of unstable fibre rotation, is well predicted, and, in addition, the crushing response shows very good agreement with experimental results in terms of morphology in the crushing zone, as well as in the load response.

To allow for computational efficiency, we have also developed and implemented (as a user element in LS-DYNA) an adaptive modelling strategy which allows for laminates to be initially modelled with only one element over the thickness.The user element kinematics can be adaptively enriched by introducing new degrees of freedom during the simulation to allow for more accurate stress predictions in critical regions by introducing discrete material interfaces, and for the modelling of delamination crack growth by introducing discrete crack surfaces interconnected with a cohesive zone law. In this work, special care has been taken to develop a robust method for explicit crash analysis. In the element, we also able to consider the correct intralaminar fracture toughness regularisation for various spatial discretisations.

To assess and validate the models developed in the project, we have also conducted a series of bending and crushing experiments on component level. Three-point bending tests (in total 45 beams) have been conducted for three different carbon-epoxy material systems (pre-preg and vacuum infused), two different span lengths and two different lay-ups at several impact speeds. Similarly, crushing tests have been conducted for the same material systems by crushing tubes (in total 35 tubes) at various angles, with two different lay-ups and at two different loading speeds (quasi-static and dynamic). We believe that these tests serve as a very strong basis for any crash model validation.