Efficiency Improvements to Adjoint-Based Aeroelastic Optimisations using a Trim-Corrected and Hybrid Mesh Deformation Strategy

This purpose of this research is to increase the efficiency of the aeroelastic shape optimisation process for commercial aircraft. Aeroelastic simulations capture the interaction between aerodynamic loading and structural displacements. High-fidelity aeroelastic simulations are computationally expensive, hence an adjoint-based approach to aircraft shape optimisation is the most suitable approach when large numbers of design parameters are present. The coupled nature of the fluid-structure interaction (FSI) is reflected in the resulting adjoint equations that are used to find the gradient. Previous coupled-adjoint optimisations performed in literature have used high-fidelity solvers for both computational fluid dynamics (CFD) and computational structural mechanics (CSM) while also satisfying the trim constraints within the FSI simulation. This project builds on those studies by proposing a simple yet powerful control surface parameterisation method for satisfying the trim constraints within the FSI simulation. An additional contribution of this work is an investigation into the effects that different mesh deformation algorithms have on the rate of convergence of the coupled-adjoint.
An important aspect of capturing the FSI is an effective mesh deformation strategy. The algorithm used for deforming the mesh in an FSI simulation needs to be robust to large deformations but also efficient due to the large number of times it will be required. The radial basis function (RBF) mesh deformation strategy with a data-reduction algorithm is a popular method for achieving robust and efficient deformations within FSI simulations. A key contribution of this work is the finding that the application of a data-reduction algorithm to the input field of the mesh deformation strategy has a significantly negative effect on the convergence of the coupled-adjoint whilst having only a negligible effect on the convergence of the FSI simulation. The Delaunay Graph Mapping (DGM) mesh deformation algorithm is employed to obtain faster convergence of the coupled-adjoint than the RBF approach. To increase the efficiency of optimisation process, a hybrid mesh deformation strategy is proposed by using the RBF approach within the FSI simulation and the DGM approach within the coupled-adjoint.
The gradients that are obtained via the hybrid mesh deformation approach are successfully validated. The hybrid mesh deformation strategy is then applied to two optimisation scenarios in the transonic flow region. The first is a lift constrained wing optimisation. The second is a lift and trim constrained optimisation performed on a full transport aircraft configuration. The developed trim-corrected and hybrid mesh deformation optimisation strategy is shown to demonstrate a more efficient coupled-adjoint aeroelastic shape optimisation process.