Date of Award

3-1998

Document Type

Thesis

Degree Name

Master of Science

Department

Department of Operational Sciences

First Advisor

James W. Chrissis, PhD

Second Advisor

Edward A. Pohl, PhD

Abstract

The analysis of fuel air mixing in a scramjet is often accomplished either with Computational Fluid Dynamics (CFD) algorithms or through experimental research. These approaches, while accurate and reliable, are extremely expensive and thus not well suited for use with conventional design optimization methods. In this investigation, Variable Complexity Modeling (VCM) is used to significantly reduce the number of complex, expensive analyses required to optimize the design of a scramjet fuel injection array. A design problem formulation for a lateral transverse injection array is developed and a VCM approach to design optimization is conducted in two stages. Initially, a simplified analysis model is used to provide relatively inexpensive predictions of design fuel air mixing characteristics. A parametric analysis is conducted to explore the design region, and a preliminary optimal design is found using both Sequential Quadratic Programming and a Genetic Algorithm. In the second stage, response surface methodology is supplemented with preliminary stage information to minimize the number of expensive analyses required to finalize the design. It is shown that only 25 design evaluations are required to develop a near optimal design.

AFIT Designator

AFIT-GOR-ENS-98M-18

DTIC Accession Number

ADA342233

Download

Share

COinS