Date of Award

12-1997

Document Type

Thesis

Degree Name

Master of Science

Department

Department of Electrical and Computer Engineering

First Advisor

Peter J. Collins, PhD

Abstract

The optimized design methodology implemented in this research employs a design methodology, namely Response Surface Methodology (RSM), that is relatively new to the electromagnetics field. The design methodology used a full factorial designed experiment, a Hybrid Finite Element Method (HFEM) analysis code, a cubic spline empirical model, an Analysis of Variance (ANOVA) test, and a simplex optimization code to successfully measure, predict, analyze, and optimize average antenna gain and antenna input impedance for a cavity backed microstrip antenna with dielectric overlays. Based on HFEM results from the full factorial designed experiment, a cubic spline empirical model successfully predicted average antenna gain and input impedance values for variable combinations within the experiment variable limits. The ANOVA test determined the impact each design variable introduced on the average antenna gain and indicated the areas of greatest concern during manufacturing and testing phases. Using simplex optimization, an optimized, realizable antenna design with minimum average antenna gain above the antenna bore sight and a target input impedance of 50 ohms was located within two minutes that agreed within approximately 4.0 percent to HFEM values determined from design validation tests, while measured results from the constructed optimized antenna design closely replicated the predicted antenna patterns and the predicted input impedance value. The design methodology eliminated numerous antenna builds by applying straight forward RSM techniques and saved countless man hours. The significant time savings found using this optimized design methodology demonstrates the power of RSM and motivates its application to other electromagnetic design problems.

AFIT Designator

AFIT-GE-ENG-97D-04

DTIC Accession Number

ADA335573

Download

Share

COinS