Finite element methodologies for structural design and optimisation

Abstract

The development and deployment of large-scale load-bearing structures in the aeronautics industry is an endeavour that can span several decades. The complex nature of the materials involved and manufacturing processes used has historically led to iterative design stages and lengthy testing and certification procedures. In particular, composite materials, with all their promises, bring about several interesting challenges. One of these is the relatively common incidence of manufacturing defects. In the best-case scenario, these defects could mean that the component is discarded; in the worst-case scenario, it could lead to premature structural collapse. Numerical methodologies specifically tailored for the design and optimisation of large-scale composite structures enable faster, cheaper and more robust designs. Therefore, this work presents a series of finite element methodologies to address the challenges associated with large-scale composite structure development. To address the iterative nature of the design stage and the need for robust defect tolerance, this work proposes a numerical methodology to discretise a moving boundary explicitly and a topology optimisation formulation centred on the energy release rate of bodies with embedded cracks. Additionally, to address the need for virtual testing frameworks this work presents a multiscale material modelling methodology targeting modern modelling approaches and computing architectures. The methodology is applied to a full-size composite wing-box model. Finally, this work develops an optimisation methodology combining the boundary tracking capabilities, the energy release rate-based formulation and the multiscale modelling workflow to enable the optimisation of large-scale composite structures with manufacturing defects. The methodology is applied to a composite stringer run-out model with a kissing bond defect.