Multiobjective optimization of an axisymmetric supersonic-inlet bypass-duct splitter via surrogate modeling

Jacob C Haderlie, Purdue University

Abstract

The aerodynamic performance of an axisymmetric supersonic engine inlet is optimized via Kriging surrogate models by changing the splitter geometry that separates the engine and super-bypass duct regions. The splitter length and parameterized splitter leading-edge shape served as variables, while the total pressure recovery at the core aerodynamic interface plane (AIP) and the peak radial distortion intensity were the two objectives for the presented problem. Due to the discontinuous nature of the peak radial distortion at the aerodynamic interface plane, this objective was resolved using the ϵ-constraint approach, which is a subset of the bounded objective function method, for multiobjective optimization. The splitter leading edge was parameterized using the Class function Shape function Transformation (CST). The WIND-US Reynolds-averaged Navier-Stokes (RANS) computational fluid dynamics (CFD) solver conducted the high-fidelity analyses of splitter shapes sampled using an optimal Latin hypercube (OLH) design of experiments (DOE) plan. The Kriging model construction then used these CFD solutions to build the surrogates. The resulting approximate multiobjective, multipoint optimization problem was solved at cruise and takeoff conditions.

Degree

M.S.E.

Advisors

Crossley, Purdue University.

Subject Area

Aerospace engineering

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server
.

Share

COinS