Non-axisymmetric aero-engine nacelle design by surrogate-based methods

For many aerodynamic design tasks, a key challenge is the balance between the non-linearity of the transonic flow, the inherent 3D nature of the geometry and flow field, the computational cost and the level of accuracy. Within an optimisation process this is further compounded by the high degrees of freedom to define the geometry and the requirement for 3D computations at both design and off-design conditions. An example of this is the design of compact nacelles for future civil aero-engines which will have larger bypass ratios than current in-service architectures. This paper presents an approach for the design and optimisation of 3D drooped and scarfed non-axisymmetric nacelles. To reduce the computational expense, a range of surrogate-based adapted methods are investigated. Relative to the conventional approach of full numerical simulations in the optimisation loop, the adapted method identifies an acceptable design space with a 65% reduction in the total computational cost. Overall, this demonstrates a useful approach for reducing the time and cost of high-dimensional, aerodynamic design problems.