Expansion of existing EM workbench for multiple computational electromagnetics codes

An "EM workbench" is described. It has a modular, code-independent architecture and a common design environment. The technique for integrating new EM codes is demonstrated with a method-of-moments code specialized for circularly symmetric feed horns, and with a new hybrid DFT-MoM (discrete Fourier transform - method of moments) code for finite-boundary planar phased arrays. The use of a common design environment for all codes takes advantage of common infrastructure and utility software, and minimizes the learning-curve penalty for using a new EM code capability.

Introduction

Modern system designs involving electronic or communications components rely heavily on numerical computer codes for synthesizing antenna designs, computing desired and undesired antenna coupling, scattering properties, etc. The existence of several well-known computer codes for electromagnetic computations, such as NEC-BSC [1], NEC-REF [2], NEW-AIR [3], and ESP [4], coupled with the increased availability of inexpensive but powerful computing resources, potentially broadens the range of designs and problems that can be feasibly addressed. Proper use of these codes can significantly shorten system design times while improving performance, and can lead to totally new insights into complex interactions. A problem that may have taken several days of computer time on a company mainframe can now be run on a single engineer's desktop PC in a few hours. However, effective use of these codes has been hampered by the following four major problems: 1

Steep, non-overlapping learning curves

2

Problem definition is non-intuitive and error-prone

3

Geometry verification is difficult

4

Validation of results is difficult