Automated design optimization for subsonic diffusers

An automated design optimization process for subsonic S-shaped diffusers was developed and validated. The test case for the process was the evolution of an optimal design for an S-shaped subsonic diffuser, where optimality was based on the minimization of the total pressure distortion at the diffuser exit, while sustaining total pressure recovery. Constraints placed by airframe weight, space and line-of-sight blockage of the engine face cause subsonic diffusers to be highly divergent and curved, characteristics which lead to strong secondary flows, boundary layer separation and inhomogeneity of the total pressure profile. As a part of the design process, we introduce a surface perturbation on a baseline S-shaped diffuser, and minimize an objective function, reflecting the total-pressure distortion, with respect to the perturbation height and width. Four commercial software tools - Pro/Engineer (a 3D solid modeler for the generation of diffuser geometries), GridPro (an elliptic mesh generator), GASPex (an aerodynamic simulation package) and CFSQP (a gradient based optimizer to traverse the design space automatically) - were integrated to synthesize an automated analysis and optimization method. Results indicate the development of the surface perturbation and a consequent decrease in total pressure distortion during the convergence to the optimal geometry.