This paper was produced for the 2019 NAFEMS World Congress in Quebec Canada

Resource Abstract

Airbus has recently created the first-ever single-piece composite centre wing box. This structural component allows to carry the fuselage with the lift generated by the wings. For this reason, stiffness and strength requirements are very high. The new single-piece design reduces manufacturing costs and –lead-time, facilitates assembly and provides equivalent or better load-carrying characteristics compared to existing, differential designs.



To manufacture the single-piece centre wing box, high-performance thermoset resins need to be cured at high temperature to reach the required properties. The polymerization reaction reduces the space taken by the monomers: chemical shrinkage. Interaction between the curing resin and the fibres prevents the full shrinkage to develop in the fibre direction, whereas in the directions perpendicular to that, the effect of the fibres is much less pronounced. This orthotropy in thermal- and chemical shrinkage makes carbon fibre reinforced polymer prone to process-induced distortion, and it is extremely challenging to manufacture a full-scale aircraft component in-tolerance and ready for assembly – especially a closed box.



The Airbus Technical Centre in charge of Manufacturing Simulation, ESCSAM, in collaboration with Digital Product Simulation (DPS), has set up a simulation-driven approach to predict these cure-induced deformations thanks to high-fidelity numerical models and provide design orientation and guidance to make part right first time.



At each increment of the non-linear simulation, material properties are homogenized from the microscale to the ply-scale, to compute the effective instantaneous orthotropic shrinkage. The gradual increase of material stiffness is taken into account in way that allows to capture the specifics of a closed structure. The properties of the contact interactions with the tooling are constantly updated to also reflect the gradual increase of material stiffness.



Spring-in predictions resulting from the simulation were used to create a pre-compensated tool geometry based using an iterative “mirror method”. The entire approach was validated experimentally before applying it as an industrial method.