Scramjet Intake Design Based on Exit Flow Profile via Global Optimization and Deep Learning toward Inverse Design

Scramjet engines are a promising airbreathing propulsion technology for hypersonic flights in high-speed transportation and access-to-space. One of the significant technical challenges lies in fuel mixing and combustion due to the short residence time inside the engine, and shock waves caused by the intake are useful to enable instantaneous fuel mixing and ignition. The present study has been conducted to gain preliminary insights into the intake design that offers a desirable exit flow profile, aiming to facilitate the utilization of shock waves for mixing and combustion. A design approach that employs optimization and deep learning techniques has been developed in this paper in order to enable flexible design exploration. The intake geometries with the required exit flow profiles and compression ratio have been explored in an inverse manner by means of global optimization employing genetic algorithms and sequential quadratic programming with multiple initial points. For the reduction of computational cost in optimization processes, computational fluid dynamics simulations have been effectively replaced by deep learning-based flow prediction. This paper discusses the feasibility of the approach as well as the inverse design of scramjet intakes that fulfill the requirements for the exit flow profile, based on the results obtained.