A Study on Optimal Geometry of Scramjet 2D Intake Considering Viscous Boundary Layer and Transition Phenomena

A study of geometry optimization is conducted for a double compression-ramp type two-dimensional scramjet intake, considering the presence of a viscous boundary layer and its transition phenomena. The research is carried out in the following order: defining the optimization problem, generating data points based on the design of the experiment (DOE), obtaining intake performances using RANS analysis, generating and analyzing the surrogate model of response surface using Kriging/Deep Neural Network (DNN), finding the optimal solution based on the desirability function, and evaluating the performance of the resulting optimal geometries. The optimization is performed respectively based on the analyses using the fully turbulence model and the hypersonic transition model to evaluate the influence of the boundary layer transition phenomenon on the optimized geometry and intake performance. The optimal geometries have a smaller ramp angle compared to the baseline geometry to compromise the increase in the effective ramp angle due to the effect of displacement thickness of boundary layer on the ramp surface. Differences in the flow/shock-wave structure and optimal geometries are identified depending on the application of the transition model, and no significant differences in results appear depending on the surrogate model employed. The optimal geometries show improved intake performance compared to the baseline geometry, not only at the design point but also at off-design conditions.