Ulrich Jakobus - Academia.edu (original) (raw)

Papers by Ulrich Jakobus

Research paper thumbnail of Latest Features in Altair Feko 2022

Latest Features in Altair Feko 2022

2023 International Applied Computational Electromagnetics Society Symposium (ACES)

Research paper thumbnail of Anwendung von Basisfunktionen höherer Ordnung auf gekrümmten Netzelementen zur beschleunigten Analyse von EMV Problemen mit der Momentenmethode

Dieser Beitrag befasst sich mit einer Erweiterung des kommerziellen Feldberechnungsprograms FEKO ... more Dieser Beitrag befasst sich mit einer Erweiterung des kommerziellen Feldberechnungsprograms FEKO [FEKO (www.feko.info), EM Software & Systems - S.A. (Pty) Ltd, PO Box 1354, Stellenbosch, 7599, Südafrika] wo im Rahmen der Momentenmethode (MoM) und der beschleunigten Version der schnellen Multipolmethode (MLFMM = multilevel fast multipole method) Basisfunktionen höherer Ordnung auf gekrümmten Oberflächenelementen zum Einsatz kommen. Konkret werden hierarchische Basisfunktionen höherer Ordnung (HOBFs = higher order basis functions) gemäß [Roberto D. Graglia, Donald R. Wilton and Andrew F. Peterson, "Higher Order Interpolatory Vector Bases for Computational Electromagnetics," IEEE Transactions on Antennas and Propagation, Vol. AP-45, Nr. 3, S. 329-342, Mar. 1997; Roberto D. Graglia, Andrew F. Peterson and Francesco P. Andriulli, "Curl-Conforming Hierarchical Vector Bases for Triangles and Tetrahedra," IEEE Transactions on Antennas and Propagation, Vol. AP-59, Nr. 3, ...

Research paper thumbnail of Antenna Positioning for Bandwidth optimization Using Characteristic Mode Analysis

Antenna Positioning for Bandwidth optimization Using Characteristic Mode Analysis

2020 14th European Conference on Antennas and Propagation (EuCAP), 2020

Characteristic mode analysis is used to understand the modal behavior of antennas, and how they i... more Characteristic mode analysis is used to understand the modal behavior of antennas, and how they interact with the structure they are mounted on. While this insight can be applied in various ways to improve the design, one of the biggest challenges is often how to place the antenna on the structure to excite specific mode(s). Previous work describes a good approach but imposed certain limitations – multiple antennas were used, and a narrow frequency band was considered. For many applications an approach is needed to position a single antenna that operates in a wider frequency band.This paper will attempt to broaden the understanding of those limitations and proposes a design approach which surpasses the limitations, albeit at the cost of exciting additional modes. The approach covers positioning a single wider band antenna to excite specific modes while optimizing the antenna bandwidth. It is applied to two antenna examples. The bandwidth from the approach is verified using optimization.

Research paper thumbnail of Review of selected new features in FEKO 2018

Review of selected new features in FEKO 2018

2018 International Applied Computational Electromagnetics Society Symposium (ACES), 2018

This paper describes new features to the electromagnetic simulation software FEKO as introduced i... more This paper describes new features to the electromagnetic simulation software FEKO as introduced in the FEKO 2018 version. These include the GPU support for the RL-GO (ray launching geometrical optics) solver, a direct solver for ACA (adaptive cross approximation), and various cable modeling extensions.

Research paper thumbnail of Summary of the latest extensions to the electromagnetic field solver package FEKO

Summary of the latest extensions to the electromagnetic field solver package FEKO

2017 International Applied Computational Electromagnetics Society Symposium - Italy (ACES), 2017

This paper summarises the latest extensions to the electromagnetic field solver package FEKO. Adv... more This paper summarises the latest extensions to the electromagnetic field solver package FEKO. Advanced parallelisation options for full wave methods (MLFMM, FDTD), speed improvements for cable harness modelling, extensions to the asymptotic RL-GO solver, and improved mode tracking for CMA are discussed.

Research paper thumbnail of Recent MLFMM Improvements in FEKO

Recent MLFMM Improvements in FEKO

2018 IEEE Indian Conference on Antennas and Propogation (InCAP), 2018

The MLFMM in FEKO is continuously being extended and improved to obtain faster solution times usi... more The MLFMM in FEKO is continuously being extended and improved to obtain faster solution times using the minimum amount of memory. This paper describes automotive windscreen antenna analysis, requiring the hybridization of the MLFMM based on the surface equivalence principle, the finite element method and windscreen antenna analysis. In addition, the significant memory reduction using the latest MPI3 shared memory is shown.

Research paper thumbnail of Overview of Recent MLFMM Improvements in FEKO

Overview of Recent MLFMM Improvements in FEKO

2018 International Applied Computational Electromagnetics Society Symposium - China (ACES), 2018

The MLFMM in FEKO is continuously being extended and improved to obtain faster solution times usi... more The MLFMM in FEKO is continuously being extended and improved to obtain faster solution times using the minimum amount of memory. In this paper we present the memory savings obtained for parallel runs on clusters using the latest MPI-3 shared memory features, as well as a new tailored preconditioner to stabilise the iterative solution of the MLFMM.

Research paper thumbnail of Improving the convergence and parallel memory efficiency of the MLFMM in FEKO

Improving the convergence and parallel memory efficiency of the MLFMM in FEKO

2017 Computing and Electromagnetics International Workshop (CEM), 2017

This paper summarizes the latest extensions to the multilevel fast multipole method in the commer... more This paper summarizes the latest extensions to the multilevel fast multipole method in the commercial code FEKO: stabilisation and the use of IVIPI shared-memory programming.

Research paper thumbnail of Review of Advanced Electromagnetic Modeling Techniques in the Computer Code FEKO based on the Method of Moments with Hybrid Extensions

The aim of the present tutorial is to introduce some advanced electromagnetic modeling techniques... more The aim of the present tutorial is to introduce some advanced electromagnetic modeling techniques based on the Method of Moments (MoM) with various hybrid extensions. We are using the computer code FEKO [1] as a reference, and illustrate in the following several extensions that have been made in FEKO to the classical MoM in order to allow an efficient and fast analysis of a variety of complex electromagnetic radiation and scattering problems. The aim is not to go too much into the technical details (for the interested readers suitable references will be given), but rather to present an overview only, with a few selected application examples.

Research paper thumbnail of Benefits and Challenges of GPU Accelerated Electromagnetic Solvers from a Commercial Point of View

─ This paper discusses the benefits but also challenges of GPU accelerated electromagnetic solver... more ─ This paper discusses the benefits but also challenges of GPU accelerated electromagnetic solvers from a commercial point of view, namely using FEKO as example. Specifically, the effects of some of the complex interdependencies between different components are presented. It is shown that despite the advances made in the field of GPGPU computing, and impressive speedups for parts of a program or simplified problems, there are a number of factors to consider before these techniques can be applied to a commercial product that is expected to be robust and, most importantly, to always give trustworthy results for a wide variety of problems. Index Terms ─ Commercial Solvers, CUDA, FDTD, FEKO, FEM, GPGPU, GPU Acceleration, MoM, RL-GO, SBR.

Research paper thumbnail of Efficient Techniques for Characteristic Mode Analysis of Complex Structures

Efficient Techniques for Characteristic Mode Analysis of Complex Structures

Research paper thumbnail of New Features in Feko and WinProp 2019

Applied Computational Electromagnetics Society, 2021

This paper describes some of the latest features in the commercial electromagnetic software Feko ... more This paper describes some of the latest features in the commercial electromagnetic software Feko (including WinProp). These include the modeling of non-ideal cable shield connections, the parallel direct adaptive cross approximation (ACA) solver, edge and wedge diffraction for the ray launching geometrical optics (RL-GO) solver, and several new features related to automotive radar.

Research paper thumbnail of Introducing the Iterative Domain Green's Function Method for finite array analysis

Introducing the Iterative Domain Green's Function Method for finite array analysis

2016 IEEE/ACES International Conference on Wireless Information Technology and Systems (ICWITS) and Applied Computational Electromagnetics (ACES), 2016

The purpose of this work Is to Introduce the Iterative Domain Green's Function Method (DGFM) ... more The purpose of this work Is to Introduce the Iterative Domain Green's Function Method (DGFM) that may be used for the analysis of large, disjoint finite antenna arrays. The iterative DGFM presents an improvement over the conventional DGFM in terms of accuracy. The convergence rate of the technique will be illustrated at the hand of an example.

Research paper thumbnail of Characteristic mode analysis of electromagnetic structures

Characteristic mode analysis of electromagnetic structures

2015 European Microwave Conference (EuMC), 2015

Characteristic mode analysis (CMA) is an eigenanalysis technique that is particularly useful for ... more Characteristic mode analysis (CMA) is an eigenanalysis technique that is particularly useful for gaining insight into the radiating properties of arbitrary electromagnetic structures. This work reviews a numerically efficient implementation for solving the generalised eigenproblem that results from separating the Method-of-Moments (MoM) impedance matrix in its real and imaginary components. Furthermore, the approach undertaken to deal with the tracking of eigenvalues as a function of frequency is also explained. Finally, the usefulness of the CMA implementation and mode tracking approach is illustrated at the hand of an example.

Research paper thumbnail of Analysis of finite antenna arrays in the presence of arbitrary electromagnetic structures

Analysis of finite antenna arrays in the presence of arbitrary electromagnetic structures

2015 International Conference on Electromagnetics in Advanced Applications (ICEAA), 2015

This work considers the analysis of finite antenna arrays in the presence of arbitrary electromag... more This work considers the analysis of finite antenna arrays in the presence of arbitrary electromagnetic structures, e.g. a focal plane array illuminating a parabolic dish antenna, or an antenna array located above a finite sized ground plane. The fast solution approach applies domain decomposition where the finite array is considered as one domain and the remainder of the geometry as the other. Two Method-of-Moment (MoM) based strategies will be considered to analyse the coupling between the domains, viz. the Numerical Greens Function Method (NGF) as well as an Iterative Field Bouncing (IFB) scheme. These methods will be compared in terms of accuracy and runtime when used to analyse a bow-tie antenna array situated above a finite size ground plane.

Research paper thumbnail of Advanced EMC modeling by means of a parallel MLFMM and coupling with network theory

Advanced EMC modeling by means of a parallel MLFMM and coupling with network theory

2008 IEEE International Symposium on Electromagnetic Compatibility, 2008

The application of a parallelized multilevel fast multipole method (MLFMM) is described and illus... more The application of a parallelized multilevel fast multipole method (MLFMM) is described and illustrated through examples for the solution of large scale and complex EMC problems. Furthermore, to reduce the modeling complexity of such EMC problems, we propose a combination of field with network theory to split large problems into smaller sub-problems which can be analyzed individually and then again assembled together by means of network theory.

Research paper thumbnail of GPU-acceleration of the FEKO electromagnetic solution kernel

GPU-acceleration of the FEKO electromagnetic solution kernel

2013 International Conference on Electromagnetics in Advanced Applications (ICEAA), 2013

This paper discusses the GPU acceleration of FEKO, a versatile commercial CEM software package. S... more This paper discusses the GPU acceleration of FEKO, a versatile commercial CEM software package. Specifically, some of the challenges in accelerating complex software with many interdepen-dencies are presented. It is shown that despite the advances made in the field of GPGPU computing, and impressive speedups for parts of a program or simplified problems, there are a number of factors to consider before these techniques can be applied to a commercial product that is expected to be robust and, most importantly, to always give trustworthy results.

Research paper thumbnail of Efficient analysis of finite antenna arrays using the Domain Green's Function Method

Efficient analysis of finite antenna arrays using the Domain Green's Function Method

Proceedings of the 2012 IEEE International Symposium on Antennas and Propagation, 2012

A Method-of-Moments (MoM) based approach for solving finite antenna arrays is presented, viz. the... more A Method-of-Moments (MoM) based approach for solving finite antenna arrays is presented, viz. the Domain Green's Function Method (DGFM). The DGFM is a perturbation approach where mutual coupling between array elements is accounted for during the formulation of the Green's function for each element. The active current distribution on the entire array geometry is obtained by solving the localized problem of each element independently, leading to a significant saving in both run-time and memory usage. The method also takes into account the edge effects attributed to the finite size of the array, complex excitations with non-linear phase shift and is not limited to periodic array configurations.

Research paper thumbnail of A hybrid tracking algorithm for characteristic mode analysis

A hybrid tracking algorithm for characteristic mode analysis

2014 International Conference on Electromagnetics in Advanced Applications (ICEAA), 2014

Characteristic mode analysis (CMA) is a useful design tool that enables antenna designers to foll... more Characteristic mode analysis (CMA) is a useful design tool that enables antenna designers to follow a systematic approach for extracting the radiating properties of a structure. These properties are quantified with CMA in the form of eigenvalues and eigenvectors, i.e., the solutions of a generalised eigenvalue equation that is formulated from the Method-of-Moments (MoM) impedance matrix. One particular challenge to CMA, however, is the manner in which quantities such as eigenvalues are plotted as a function of frequency. At each discrete frequency sample, the eigenvalues are sorted according to modal significance, i.e., based on the efficiency with which they radiate. At higher frequencies, the ordering of modes may differ from that obtained at lower frequencies. A tracking algorithm will be introduced that keeps the ordering of modes as consistent as possible over frequency. The method presented is an extension of that done in [1], where a linear correlation is applied to the characteristic currents between adjacent frequency increments to obtain a ranking vector. This new hybrid technique is based on extrapolating the eigenvalue and characteristic angle curves and then applying linear correlation between the characteristic currents where needed to verify a successful mapping.

Research paper thumbnail of Accelerating the Domain Green's Function Method through adaptive cross approximation

Accelerating the Domain Green's Function Method through adaptive cross approximation

2014 International Conference on Electromagnetics in Advanced Applications (ICEAA), 2014

The Domain Green's Function Method (DGFM) is a Method-of-Moments (MoM) based domain decomposi... more The Domain Green's Function Method (DGFM) is a Method-of-Moments (MoM) based domain decomposition approach that is useful for the analysis of large, irregular antenna arrays. Mutual coupling between array elements is accounted for with the formulation of an active impedance matrix equation for each of the domains/array elements. The active current distribution on the entire array geometry is then obtained by solving these smaller matrix equations pertaining to the elements. The active impedance matrix calculation entails a summation of the MoM matrix diagonal and off-diagonal sub-matrices. For arrays containing a large number of elements this summation can lead to matrix fill times similar to that of the global MoM calculation. To mitigate this significant computational overhead, while still maintaining a sufficient degree of accuracy, the adaptive cross approximation (ACA) algorithm is applied to accelerate this part of the DGFM.

Research paper thumbnail of Latest Features in Altair Feko 2022

Latest Features in Altair Feko 2022

2023 International Applied Computational Electromagnetics Society Symposium (ACES)

Research paper thumbnail of Anwendung von Basisfunktionen höherer Ordnung auf gekrümmten Netzelementen zur beschleunigten Analyse von EMV Problemen mit der Momentenmethode

Dieser Beitrag befasst sich mit einer Erweiterung des kommerziellen Feldberechnungsprograms FEKO ... more Dieser Beitrag befasst sich mit einer Erweiterung des kommerziellen Feldberechnungsprograms FEKO [FEKO (www.feko.info), EM Software & Systems - S.A. (Pty) Ltd, PO Box 1354, Stellenbosch, 7599, Südafrika] wo im Rahmen der Momentenmethode (MoM) und der beschleunigten Version der schnellen Multipolmethode (MLFMM = multilevel fast multipole method) Basisfunktionen höherer Ordnung auf gekrümmten Oberflächenelementen zum Einsatz kommen. Konkret werden hierarchische Basisfunktionen höherer Ordnung (HOBFs = higher order basis functions) gemäß [Roberto D. Graglia, Donald R. Wilton and Andrew F. Peterson, "Higher Order Interpolatory Vector Bases for Computational Electromagnetics," IEEE Transactions on Antennas and Propagation, Vol. AP-45, Nr. 3, S. 329-342, Mar. 1997; Roberto D. Graglia, Andrew F. Peterson and Francesco P. Andriulli, "Curl-Conforming Hierarchical Vector Bases for Triangles and Tetrahedra," IEEE Transactions on Antennas and Propagation, Vol. AP-59, Nr. 3, ...

Research paper thumbnail of Antenna Positioning for Bandwidth optimization Using Characteristic Mode Analysis

Antenna Positioning for Bandwidth optimization Using Characteristic Mode Analysis

2020 14th European Conference on Antennas and Propagation (EuCAP), 2020

Characteristic mode analysis is used to understand the modal behavior of antennas, and how they i... more Characteristic mode analysis is used to understand the modal behavior of antennas, and how they interact with the structure they are mounted on. While this insight can be applied in various ways to improve the design, one of the biggest challenges is often how to place the antenna on the structure to excite specific mode(s). Previous work describes a good approach but imposed certain limitations – multiple antennas were used, and a narrow frequency band was considered. For many applications an approach is needed to position a single antenna that operates in a wider frequency band.This paper will attempt to broaden the understanding of those limitations and proposes a design approach which surpasses the limitations, albeit at the cost of exciting additional modes. The approach covers positioning a single wider band antenna to excite specific modes while optimizing the antenna bandwidth. It is applied to two antenna examples. The bandwidth from the approach is verified using optimization.

Research paper thumbnail of Review of selected new features in FEKO 2018

Review of selected new features in FEKO 2018

2018 International Applied Computational Electromagnetics Society Symposium (ACES), 2018

This paper describes new features to the electromagnetic simulation software FEKO as introduced i... more This paper describes new features to the electromagnetic simulation software FEKO as introduced in the FEKO 2018 version. These include the GPU support for the RL-GO (ray launching geometrical optics) solver, a direct solver for ACA (adaptive cross approximation), and various cable modeling extensions.

Research paper thumbnail of Summary of the latest extensions to the electromagnetic field solver package FEKO

Summary of the latest extensions to the electromagnetic field solver package FEKO

2017 International Applied Computational Electromagnetics Society Symposium - Italy (ACES), 2017

This paper summarises the latest extensions to the electromagnetic field solver package FEKO. Adv... more This paper summarises the latest extensions to the electromagnetic field solver package FEKO. Advanced parallelisation options for full wave methods (MLFMM, FDTD), speed improvements for cable harness modelling, extensions to the asymptotic RL-GO solver, and improved mode tracking for CMA are discussed.

Research paper thumbnail of Recent MLFMM Improvements in FEKO

Recent MLFMM Improvements in FEKO

2018 IEEE Indian Conference on Antennas and Propogation (InCAP), 2018

The MLFMM in FEKO is continuously being extended and improved to obtain faster solution times usi... more The MLFMM in FEKO is continuously being extended and improved to obtain faster solution times using the minimum amount of memory. This paper describes automotive windscreen antenna analysis, requiring the hybridization of the MLFMM based on the surface equivalence principle, the finite element method and windscreen antenna analysis. In addition, the significant memory reduction using the latest MPI3 shared memory is shown.

Research paper thumbnail of Overview of Recent MLFMM Improvements in FEKO

Overview of Recent MLFMM Improvements in FEKO

2018 International Applied Computational Electromagnetics Society Symposium - China (ACES), 2018

The MLFMM in FEKO is continuously being extended and improved to obtain faster solution times usi... more The MLFMM in FEKO is continuously being extended and improved to obtain faster solution times using the minimum amount of memory. In this paper we present the memory savings obtained for parallel runs on clusters using the latest MPI-3 shared memory features, as well as a new tailored preconditioner to stabilise the iterative solution of the MLFMM.

Research paper thumbnail of Improving the convergence and parallel memory efficiency of the MLFMM in FEKO

Improving the convergence and parallel memory efficiency of the MLFMM in FEKO

2017 Computing and Electromagnetics International Workshop (CEM), 2017

This paper summarizes the latest extensions to the multilevel fast multipole method in the commer... more This paper summarizes the latest extensions to the multilevel fast multipole method in the commercial code FEKO: stabilisation and the use of IVIPI shared-memory programming.

Research paper thumbnail of Review of Advanced Electromagnetic Modeling Techniques in the Computer Code FEKO based on the Method of Moments with Hybrid Extensions

The aim of the present tutorial is to introduce some advanced electromagnetic modeling techniques... more The aim of the present tutorial is to introduce some advanced electromagnetic modeling techniques based on the Method of Moments (MoM) with various hybrid extensions. We are using the computer code FEKO [1] as a reference, and illustrate in the following several extensions that have been made in FEKO to the classical MoM in order to allow an efficient and fast analysis of a variety of complex electromagnetic radiation and scattering problems. The aim is not to go too much into the technical details (for the interested readers suitable references will be given), but rather to present an overview only, with a few selected application examples.

Research paper thumbnail of Benefits and Challenges of GPU Accelerated Electromagnetic Solvers from a Commercial Point of View

─ This paper discusses the benefits but also challenges of GPU accelerated electromagnetic solver... more ─ This paper discusses the benefits but also challenges of GPU accelerated electromagnetic solvers from a commercial point of view, namely using FEKO as example. Specifically, the effects of some of the complex interdependencies between different components are presented. It is shown that despite the advances made in the field of GPGPU computing, and impressive speedups for parts of a program or simplified problems, there are a number of factors to consider before these techniques can be applied to a commercial product that is expected to be robust and, most importantly, to always give trustworthy results for a wide variety of problems. Index Terms ─ Commercial Solvers, CUDA, FDTD, FEKO, FEM, GPGPU, GPU Acceleration, MoM, RL-GO, SBR.

Research paper thumbnail of Efficient Techniques for Characteristic Mode Analysis of Complex Structures

Efficient Techniques for Characteristic Mode Analysis of Complex Structures

Research paper thumbnail of New Features in Feko and WinProp 2019

Applied Computational Electromagnetics Society, 2021

This paper describes some of the latest features in the commercial electromagnetic software Feko ... more This paper describes some of the latest features in the commercial electromagnetic software Feko (including WinProp). These include the modeling of non-ideal cable shield connections, the parallel direct adaptive cross approximation (ACA) solver, edge and wedge diffraction for the ray launching geometrical optics (RL-GO) solver, and several new features related to automotive radar.

Research paper thumbnail of Introducing the Iterative Domain Green's Function Method for finite array analysis

Introducing the Iterative Domain Green's Function Method for finite array analysis

2016 IEEE/ACES International Conference on Wireless Information Technology and Systems (ICWITS) and Applied Computational Electromagnetics (ACES), 2016

The purpose of this work Is to Introduce the Iterative Domain Green's Function Method (DGFM) ... more The purpose of this work Is to Introduce the Iterative Domain Green's Function Method (DGFM) that may be used for the analysis of large, disjoint finite antenna arrays. The iterative DGFM presents an improvement over the conventional DGFM in terms of accuracy. The convergence rate of the technique will be illustrated at the hand of an example.

Research paper thumbnail of Characteristic mode analysis of electromagnetic structures

Characteristic mode analysis of electromagnetic structures

2015 European Microwave Conference (EuMC), 2015

Characteristic mode analysis (CMA) is an eigenanalysis technique that is particularly useful for ... more Characteristic mode analysis (CMA) is an eigenanalysis technique that is particularly useful for gaining insight into the radiating properties of arbitrary electromagnetic structures. This work reviews a numerically efficient implementation for solving the generalised eigenproblem that results from separating the Method-of-Moments (MoM) impedance matrix in its real and imaginary components. Furthermore, the approach undertaken to deal with the tracking of eigenvalues as a function of frequency is also explained. Finally, the usefulness of the CMA implementation and mode tracking approach is illustrated at the hand of an example.

Research paper thumbnail of Analysis of finite antenna arrays in the presence of arbitrary electromagnetic structures

Analysis of finite antenna arrays in the presence of arbitrary electromagnetic structures

2015 International Conference on Electromagnetics in Advanced Applications (ICEAA), 2015

This work considers the analysis of finite antenna arrays in the presence of arbitrary electromag... more This work considers the analysis of finite antenna arrays in the presence of arbitrary electromagnetic structures, e.g. a focal plane array illuminating a parabolic dish antenna, or an antenna array located above a finite sized ground plane. The fast solution approach applies domain decomposition where the finite array is considered as one domain and the remainder of the geometry as the other. Two Method-of-Moment (MoM) based strategies will be considered to analyse the coupling between the domains, viz. the Numerical Greens Function Method (NGF) as well as an Iterative Field Bouncing (IFB) scheme. These methods will be compared in terms of accuracy and runtime when used to analyse a bow-tie antenna array situated above a finite size ground plane.

Research paper thumbnail of Advanced EMC modeling by means of a parallel MLFMM and coupling with network theory

Advanced EMC modeling by means of a parallel MLFMM and coupling with network theory

2008 IEEE International Symposium on Electromagnetic Compatibility, 2008

The application of a parallelized multilevel fast multipole method (MLFMM) is described and illus... more The application of a parallelized multilevel fast multipole method (MLFMM) is described and illustrated through examples for the solution of large scale and complex EMC problems. Furthermore, to reduce the modeling complexity of such EMC problems, we propose a combination of field with network theory to split large problems into smaller sub-problems which can be analyzed individually and then again assembled together by means of network theory.

Research paper thumbnail of GPU-acceleration of the FEKO electromagnetic solution kernel

GPU-acceleration of the FEKO electromagnetic solution kernel

2013 International Conference on Electromagnetics in Advanced Applications (ICEAA), 2013

This paper discusses the GPU acceleration of FEKO, a versatile commercial CEM software package. S... more This paper discusses the GPU acceleration of FEKO, a versatile commercial CEM software package. Specifically, some of the challenges in accelerating complex software with many interdepen-dencies are presented. It is shown that despite the advances made in the field of GPGPU computing, and impressive speedups for parts of a program or simplified problems, there are a number of factors to consider before these techniques can be applied to a commercial product that is expected to be robust and, most importantly, to always give trustworthy results.

Research paper thumbnail of Efficient analysis of finite antenna arrays using the Domain Green's Function Method

Efficient analysis of finite antenna arrays using the Domain Green's Function Method

Proceedings of the 2012 IEEE International Symposium on Antennas and Propagation, 2012

A Method-of-Moments (MoM) based approach for solving finite antenna arrays is presented, viz. the... more A Method-of-Moments (MoM) based approach for solving finite antenna arrays is presented, viz. the Domain Green's Function Method (DGFM). The DGFM is a perturbation approach where mutual coupling between array elements is accounted for during the formulation of the Green's function for each element. The active current distribution on the entire array geometry is obtained by solving the localized problem of each element independently, leading to a significant saving in both run-time and memory usage. The method also takes into account the edge effects attributed to the finite size of the array, complex excitations with non-linear phase shift and is not limited to periodic array configurations.

Research paper thumbnail of A hybrid tracking algorithm for characteristic mode analysis

A hybrid tracking algorithm for characteristic mode analysis

2014 International Conference on Electromagnetics in Advanced Applications (ICEAA), 2014

Characteristic mode analysis (CMA) is a useful design tool that enables antenna designers to foll... more Characteristic mode analysis (CMA) is a useful design tool that enables antenna designers to follow a systematic approach for extracting the radiating properties of a structure. These properties are quantified with CMA in the form of eigenvalues and eigenvectors, i.e., the solutions of a generalised eigenvalue equation that is formulated from the Method-of-Moments (MoM) impedance matrix. One particular challenge to CMA, however, is the manner in which quantities such as eigenvalues are plotted as a function of frequency. At each discrete frequency sample, the eigenvalues are sorted according to modal significance, i.e., based on the efficiency with which they radiate. At higher frequencies, the ordering of modes may differ from that obtained at lower frequencies. A tracking algorithm will be introduced that keeps the ordering of modes as consistent as possible over frequency. The method presented is an extension of that done in [1], where a linear correlation is applied to the characteristic currents between adjacent frequency increments to obtain a ranking vector. This new hybrid technique is based on extrapolating the eigenvalue and characteristic angle curves and then applying linear correlation between the characteristic currents where needed to verify a successful mapping.

Research paper thumbnail of Accelerating the Domain Green's Function Method through adaptive cross approximation

Accelerating the Domain Green's Function Method through adaptive cross approximation

2014 International Conference on Electromagnetics in Advanced Applications (ICEAA), 2014

The Domain Green's Function Method (DGFM) is a Method-of-Moments (MoM) based domain decomposi... more The Domain Green's Function Method (DGFM) is a Method-of-Moments (MoM) based domain decomposition approach that is useful for the analysis of large, irregular antenna arrays. Mutual coupling between array elements is accounted for with the formulation of an active impedance matrix equation for each of the domains/array elements. The active current distribution on the entire array geometry is then obtained by solving these smaller matrix equations pertaining to the elements. The active impedance matrix calculation entails a summation of the MoM matrix diagonal and off-diagonal sub-matrices. For arrays containing a large number of elements this summation can lead to matrix fill times similar to that of the global MoM calculation. To mitigate this significant computational overhead, while still maintaining a sufficient degree of accuracy, the adaptive cross approximation (ACA) algorithm is applied to accelerate this part of the DGFM.