Substantial computational time to develop high speed propulsion systems is the number one challenge facing future development of thrust producing scramjets over a wide operational envelope. Current designs times are substantial due to the need to solve the Navier-Stokes equations using computational fluid dynamics. The computational time for one point can lead to only a few design choices being run before a decision is required. In order to develop robust design a level of understanding of off design performance is also required. If a single design point requires significant resources in computational time, solving the same geometry over a number of design parameters creates a cost too great for design. To mitigate this cost, Surrogate surface modelling can be used to generate a statistical function for the design space with a sparse data set to capture the optima points as well as how the system behaves during off design performance. This paper focuses on generating surrogate surfaces using the Kriging method, employing results for both the supersonic flow and the reaction mechanisms required to accurately predict the combustion performance which requires chemical kinetics to accurately capture the low residence time inherent in supersonic combustion.
展开▼
