Tom Cwik

A parallel electromagnetic genetic-algorithm optimization (EGO) application for patch antenna design

In this paper, we describe an electromagnetic genetic algorithm (GA) optimization (EGO) application developed for the cluster supercomputing platform. A representative patch antenna design example for commercial wireless applications is detailed, which illustrates the versatility and applicability of the method. We show that EGO allows us to combine the accuracy of full-wave EM analysis with the robustness of GA optimization and the speed of a parallel computing algorithm. A representative patch antenna design case study is presented. We illustrate the use of EGO to design a dual-band antenna element for wireless communication (1.9 and 2.4 GHz) applications. The resulting antenna exhibits acceptable dual-band operation (i.e., better than -10 dB return loss with 5.3 and 7% operating bandwidths at 1.9 and 2.4 GHz) while maintaining a cross-pol maximum field level at least 11 dB below the co-pol maximum.

Synthesis of novel all-dielectric grating filters using genetic algorithms

The feasibility of novel all-dielectric waveguide grating filters is demonstrated, using a genetic algorithm (GA) to solve for material dielectric constants and geometric boundaries separating homogeneous regions of the periodic cell. In particular, the GAs show that simple geometries (not previously reported) utilizing a small number of layers and/or gratings can be found to yield bandpass or stop-band filters with user defined linewidth. The evaluation of the fitness of a candidate design entails the solution of an integral equation for the electric field in the cell using the method of moments (MoM). Our implementation is made efficient by using only very few design frequency points and accurately approximating a given filter transfer function by a quotient of polynomials as a function of frequency. Additionally, the problem impedance matrices are conveniently represented as the product of a material independent matrix and a vector of dielectric constants, thus allowing us to fill the matrices only once. Our code has been parallelized for the Cray T3D to take advantage of the intrinsic parallelization efficiencies offered by the GAs. Solutions are illustrated for a very narrow-band single-grating transmission filter and a relatively broad-band double grating reflection filter. Additionally, a solution for a five homogeneous layers Fabry-Perot filter is also presented.

Modeling three-dimensional scatterers using a coupled finite element-integral equation formulation

Finite-element modeling has proven useful for accurately simulating scattered or radiated fields from complex three-dimensional objects whose geometry varies on the scale of a fraction of a wavelength. To practically compute a solution to exterior problems, the domain must be truncated at some finite surface where the Sommerfeld radiation condition is enforced, either approximately or exactly. This paper outlines a method that couples three-dimensional finite-element solutions interior to a bounding surface with an efficient integral equation solution that exactly enforces the Sommerfeld radiation condition. The general formulation and the main features of the discretized problem are first briefly outlined. Results for far and near fields are presented for geometries where an analytic solution exists and compared with exact solutions to establish the accuracy of the model. Results are also presented for objects that do not allow an analytic solution, and are compared with other calculations and/or measurements.

Coupling into and scattering from cylindrical structures covered periodically with metallic patches

Circular cylindrical structures covered periodically with metallic patches are considered. After an analogy to planar periodic surfaces is shown, formulations are presented for calculating induced currents on the curved surface. The equations are solved and results calculated for the specific case of periodic strips on the cylindrical surface. For a cylindrical structure a two-dimensional periodicity exists, as in a planar structure, while a spherical structure allows only a rotational periodicity. When the cylindrical structure is excited by the characteristic harmonic of the system, the spectral response of the transmitted field exhibits resonances that depend on the surface periodicity, as is known for planar structures. Since the cylindrical structure contains finite closed regions, the effects of resonances internal to the structure are seen and give additional information as compared to planar structures. >

MODTRAN on supercomputers and parallel computers

Abstract To enable efficient reduction of large data sets such as is done in the Airborne Visible/Infrared Imaging Spectrometer (AVIRIS) project at the Jet Propulsion Laboratory (JPL), a high performance version of MODTRAN is essential. One means to accomplish this is to apply the computational resources of parallel computer systems. In our present work, a flexible, parallel version of MODTRAN has been implemented on the Cray T3E, the HP SPP2000, and a Beowulf-class cluster computer using domain decomposition techniques and the Message Passing Interface (MPI) library. In this paper, porting the sequential MODTRAN to various platforms is discussed; strategies of designing a parallel version of MODTRAN are developed; detailed implementation for a parallel MODTRAN is reported, and performance data of the parallel code on various computers are presented. Near linear scaling performance of parallel MODTRAN has been obtained, and comparisons of wallclock time are made among various supercomputers and parallel computers. The parallel version of MODTRAN gives excellent speedup, which dramatically reduces total data processing time for many applications such as the AVIRIS project at JPL.

Coupling finite element and integral equation solutions using decoupled boundary meshes (electromagnetic scattering)

A method is outlined for calculating scattered fields from inhomogeneous penetrable objects using a coupled finite element-integral equation solution. The finite element equation can efficiently model fields in penetrable and inhomogeneous regions, while the integral equation exactly models fields on the finite element mesh boundary and in the exterior region. By decoupling the interior finite element and exterior integral equation meshes, considerable flexibility is found in both the number of field expansion points as well as their density. Only the nonmetal portions of the object need be modeled using a finite element expansion; exterior perfect conducting surfaces are modeled using an integral equation with a single unknown field since E/sub tan/ is identically zero on these surfaces. Numerical convergence, accuracy, and stability at interior resonant frequencies are studied in detail. >

Application of massively parallel computation to integral equation models of electromagnetic scattering

Integral equation methods are widely used in the analysis and the design of electromagnetic systems. Traditionally, the limiting parts of the simulation have been the memory required for storing the dense matrix and the computational time required for solving the matrix equation. We report on the extension of integral equation solutions to new wavelength regimes and on completion of the solution in an amount of time that is practical for engineering applications. The numerical solution of the integral equation is computed on scalable, distributed-memory parallel computers. Essential to the numerical solution was the development of a complex-valued, highly optimized, dense-matrix equation solution algorithm for scalable machines. A portion of the research outlined is the development of this production-level library routine for the solution of linear equations on parallel computers. A convenient interface, useful for integral equation solutions, among others, was specifically developed in this study. This algorithm has the conveniences offered by the sequential libraries, can be easily ported between parallel platforms, and has been placed in the public domain.

Modeling radiation with an efficient hybrid finite-element integral-equation waveguide mode-matching technique

A method is presented to model electromagnetic radiation, combining the finite-element technique in the penetrating portion of a three-dimensional (3-D) radiating structure with an integral equation on the outer surface of the computational domain, and with a waveguide mode-matching technique on a cross section of the feeding waveguide. The antenna can be of general shape, provided a portion of the surrounding medium up to a surface of revolution is chosen as part of the computational domain. This truncation scheme and the accurate source representation are more general than those existing in the literature for radiation modeling. An open-ended waveguide and a waveguide with a choke ring are analyzed and their radiation patterns are obtained and compared with available measurements.

An assessment of a Beowulf system for a wide class of analysis and design software

Abstract This paper discusses Beowulf systems, focusing on Hyglac, the Beowulf system installed at the Jet Propulsion Laboratory. The purpose of the paper is to assess how a system of this type will perform while running a variety of scientific and engineering analysis and design software. The first part of the assessment contains a measurement of the communication performance of Hyglac, along with a discussion of factors which have the potential to limit system performance. The second part consists of performance measurements of six specific programs (analysis and design software), as well as discussion about these measurements. Finally, the measurements and discussion lead to the conclusion that Hyglac is suitable for running these types of codes (in a research/industrial environment such as at JPL) and that the primary factor for determining how a given code will perform is the codes ratio of communication to computation.

Parallel Computation Applied to Electromagnetic Scattering and Radiation Analysis

Abstract We have been applying the computational power of parallel processing to the solution of large-scale electromagnetic scattering and radiation problems. Several analysis codes have been implemented on the Jet Propulsion Laboratory/California Institute of Technology Mark IIIfp Hypercubes. The first code to be implemented was the Numerical Electromagnetics Code (NEC-2) from Lawrence Livermore National Laboratory. At first we simply ported it to run in the parallel processing environment. Since that time, taking advantage of the large hypercube memory and fast computation. we have enhanced parallel NEC to permit iterative design and analysis. Three other codes, frequency domain finite elements, time domain finite difference, and frequency selective surfaces, have been largely or completely developed within this parallel processing environment. Because of the massive problem size of the typical electromagnetics problem, our work is an important influence in determining the development of hardware, syst...