Multi-point surrogate-based aerodynamic optimisation of a generic hypersonic waverider

Abstract

For atmospheric flight at hypersonic speeds, waverider designs offer the benefit of higher aerodynamic performance by sustaining an attached shock wave at the leading edge. Inverse design methods are typically used to generate waverider shapes by combining an a priori desired flow field with streamline-tracing starting from the rear of the vehicle. A difficulty of these inverse design methods is to perform valid parameter space explorations or design optimisation tasks, for which direct optimisation frameworks are better suited. This work builds on the direct optimisation approach introduced by Son et al. (Aerospace, Vol 9, Issue 7, 2022, 348) in which four parameters are used to define waverider shapes. An in-depth exploration of the design space is presented based on similar methods consisting of inviscid CFD-derived surrogate modelling followed by genetic algorithm-based optimisation of aerodynamic performance and internal volume. Focus is placed on open-source software release and the wide-speed problem approached through a cost function incorporating information from simulations at multiple Mach numbers.