Experimental and Numerical Investigation of Debonding in Automotive Composite Structures

Debonding between the skins and the core in a sandwich structure is a critical failure mode in automotive applications where the energy absorption capability is a key property for crashworthiness. Once debonding occurs, the load carrying capacity of a sandwich structure, and therefore its energy absorption capability, drastically decreases. Debonding between a Carbon Fibre Reinforced Polymer (CFRP) skin and a foam core depends on many factors, including skin layup, foam material properties and skin-core thickness ratio. This research investigates the effect of using various core materials, with different cell characteristics, on the interfacial strength between a cellular foam and a CFRP skin. A key finding was that foams with a coarse cellular structure favoured a high resin uptake at the interface during the manufacturing process. The role of this resin layer was well-defined with the experimental tests, where it was shown to postpone crack initiation and kinking on samples subjected to Mode I loading and to delay the collapse of foam cells under the crack tip in compression on samples under Mode II loading. The research has matched physical experiments with Finite Element (FE) modelling. Initially, the work presented in this thesis focused on gaining a deep understanding of two of the most used numerical techniques to model interfacial failure: the Cohesive Zone Method (CZM) and the Virtual Crack Closure Technique (VCCT). However, limitations of these two techniques were found when modelling skin-core debonding in sandwich specimens. Experimentally, debonding in a foam-cored pre-cracked specimen was seen to initiate at the interface but to propagate within the foam core, growing in an unstable manner under both Mode I and Mode II loading. Both CZM and VCCT were not capable of capturing the stick-slip behaviour resulting from the unstable crack propagation. Therefore, a new method was suggested in order to accurately simulate debonding; this included the modelling of foam failure with an element deletion mechanism. This was the first time a failure criterion-based method was used to model debonding in foam cored sandwich specimen. The suggested modelling technique was also validated comparing the FE predictions to the experimental results obtained performing three-point bending tests on sandwich panels with three different configurations. As a result, the work has shown that debonding in foam-cored sandwich structures can be modelled as foam failure instead of as interface failure, decreasing the computational time and the pre-processing efforts