High-Fidelity Adjoint-based Aircraft Shape Optimization with Aeroelastic Trimming and Engine Coupling

Abstract

A high-fidelity adjoint-based aerodynamic shape optimization process has been set up that engages the Reynolds-averaged Navier-Stokes solver TAU and couples with the finite element solver Nastran and the thermodynamic engine cycle analysis tool GTlab-Performance. A so-called CAD-ROM, which is a reduced order model representation of a parametric computer-aided design model, is used as shape parameterization. The process enables a gradient-based aeroelastic optimization of full aircraft configurations with powered engines. Therein, a trimming module ensures that each optimization iteration is feasible by imposing an equilibrium of forces and moments. The process is embedded within FlowSimulator, an environment dedicated to efficient multi-physics simulations on high-performance computing clusters. The shape sensitivities are consistently derived by coupling the discrete flow, mesh and structural adjoints and by accounting for a trim-correction of the objective functional. The optimization process has been demonstrated for a cruise range optimization of a generic transport aircraft model. The range can be increased by approximately five percent within the first six optimization iterations by adapting the wing shape defined by a 126-dimensional wing section Bspline parameterization.