Graphics Processing Unit (GPU) programming techniques have been applied to a range of scientific and engineering computations. In computational electromagnetics, uses of the GPU technique have dramatically increased since the release of NVIDIA’s Compute Unified Device Architecture (CUDA), a powerful and simple-to-use programmer environment that renders GPU computing easy accessibility to developers not specialized in computer graphics. The focus of recent research has been on problems concerning the Finite-Difference Time-Domain (FDTD) simulation of electromagnetic (EM) fields. Traditional FDTD methods sometimes run slowly due to large memory and CPU requirements for modeling electrically large structures. Acceleration methods such as parallel programming are then needed. FDTD algorithm is suitable for multi-thread parallel computation with GPU. For complex structures and procedures, high performance GPU calculation algorithms will be crucial. In this work, we present the implementation of GPU programming for acceleration of computations for EM engineering problems. The speed-up is demonstrated through a few simulations with inexpensive GPUs and ACEnet, and the attainable efficiency is illustrated with numerical results. Using C, CUDA C, Matlab GPU, and ACEnet, we make comparisons between serial and parallel algorithms and among computations with and without GPU and CUDA, different types of GPUs, and personal computers and ACEnet. A maximum of 26.77 times of speed-up is achieved, which could be further boosted with development of new hardware in the future. The acceleration in run time will make many investigations possible and will pave the way for studies of large-scale computational electromagnetic problems that were previously impractical. This is a field that definitely invites more in-depth studies. / This is the thesis of my Master of Applied Science work at Dalhousie University.

In the first part of this dissertation, radiative interactions with single irregular particles are simulated. We first introduce the basic method and techniques of Finite- Difference Time-Domain method(FDTD), which is a powerful method to numerically solve Maxwell's equations with high accuracy. To improve the efficiency of FDTD, we also develop a parallel FDTD code. Since FDTD can simulate light scattering by arbitrary shape and compositions, we study several radiative interaction cases for single particles in an external plane parallel light source: the surface roughness effects on the scattering, electric and magnetic energy density distribution in irregular particles, and backscattered Mueller images. We also develop an innovative and accurate method to simulate the infinitesimal electric dipole radiation from inside a particle with arbitrary shape and composition. Our research and results are very important to study light scattering by irregular particles, Raman scattering and fluorescence. In the second part of the dissertation, we study radiative interactions in an atmosphere-ocean system. By using the so called Matrix operator method, not only the radiance of the radiation field, but also the polarization of the radiation field are obtained. Given the single layer information for the atmosphere, time dependent ocean surface shapes, and the ocean with no interface, the Matrix operator method couples these three layers and provides both the radiance and polarization reaching a certain detector in the time domain, which are essential for atmospheric science and oceanography. Several simple cases are studied by this method to demonstrate its accuracy and robustness. We also show the most difficulties in this method and discuss what one need to do in future research works.

CD and DVD became the major portable and backup data storage devices because their reliability and economical cost when mass produced. As the computer technology grew, higher data storage density on CD/DVD disc was demanded. Using a shorter wavelength light source was the common technique to achieve this goal from both research and industry. However, the limit of wavelength had almost reached for applying it to optical storage. The nano-scale metallic wire polariser that was designed in this thesis provided a possible solution. This thesis introduces the method of using the nano-scale metallic wires to form a grating polariser as the data pit on CD/DVD disc. The polariser is a type of scattering polariser and could transmit one linear polarisation of the light and reject its orthogonal counter part. The designed pattern was tested by using XFDTD, an electromagnetic simulation program based on the finite difference time domain method. As the wave source was a red laser with a wavelength of 650nm, the simulation cell size was set to be 10nm. The dispersive materials were simulated by the Debye model. The electric field results were measured on X, Y, and Z components. The results were analysed by a pre-written Matlab program to find the transmission and crosstalk coefficients. The single polariser simulations showed that there are great potentials in this design. However, inter-cell crosstalk became the major problem in the polariser array simulations. The groove pattern and titanium material were used to optimise the polarisation effect. The simulation showed that a standard-sized disc with a titanium polariser array could have 5.5GB storage capacity and a 15 to 20dB inter-cell extinction ratio for an optical pickup with a red laser (650nm) and a numerical aperture (NA) of 0.6. Although the improvement is only marginal over existing optical data storage technology, there are many further researches possible to carry on such as the fabrication of the polariser arrays.

In the effort to simulate the biologically inspired continuum robot’s dynamic capabilities, researchers have been faced with the daunting task of simulating—in real-time—the complete three dimensional dynamics of the the “beam-like” structure which includes the three “stiff” degrees-ofreedom transverse and dilational shear. Therefore, researchers have traditionally limited the difficulty of the problem with simplifying assumptions. This study, however, puts forward a solution which makes no simplifying assumptions and trades off only the real-time requirement of the desired solution. The solution is a Finite Difference Time Domain method employing an explicit single step method with cheap right hands sides. The cheap right hand sides are the result of a rather ingenious formulation of the classical beam called the Cosserat rod by, first, the Cosserat brothers and, later, Stuart S. Antman which results in five nonlinear but uncoupled equations that require only multiplication and addition. The method is therefore suitable for hardware implementation thus moving the real-time requirement from a software solution to a hardware solution.

No description available.

In the field of antenna design and analysis, often the need arises to numerically extrapolate the far-zone performance of a radiating structure from its known (or assumed known) near-zone electromagnetic field. Mathematical processes developed to accomplish such a task are known in the literature as near-zone to far-zone transformations (NZ-FZTs) as well as near-field far-field (NF-FF) transformations. These processes make use of sampled near-zone field quantities along some virtual surface, viz., the transformation surface, that surrounds the radiating structure of interest. Depending upon the application, samples of the required near-zone field quantities are supplied via analytical, empirical, or computational means. Over the years, a number of NZ-FZT processes have been developed to meet the demands of many applications. In short, their differences include, but are not limited to, the following: (1) the size and shape of the transformation surface, (2) the required near-zone field quantities and how they are sampled, (3) the computational methodology used, and (4) the imbedding of various application-driven features. Each process has its pros and cons depending upon its specific application as well as the type of radiation structure under consideration. In this dissertation we put forth a new and original NZ-FZT process that allows the transformation surface along which the near-zone is sampled to be spheroidal in shape: namely a prolate or oblate spheroid. Naturally, there are benefits gained in doing so. Our approach uses a formulated eigenfunction expansion of spheroidal wave-harmonics to develop two distinct, yet closely related, NZ-FZT algorithms for each type of spheroidal transformation surface. The process only requires knowledge of the E-field along the transformation surface and does not need the corresponding H-field. Given is a systematic exposition of the formulation, implementation, and verification of the newly developed NZ-FZT process. Accordingly, computer software is developed to implement both NZ-FZT algorithms. In the validation process, analytical and empirical radiation structures serve as computational benchmarks. Numerical models of both benchmark structures are created by integrating the software with a field solver, viz., a finite-difference time-domain (FDTD) code. Results of these computer models are compared with theoretical and empirical data to provide additional validation. / Ph. D.

No description available.

Thesis advisor: Krzysztof Kempa / Nonlocal effects, the wavenumber dependence in a medium's response to external disturbance, is treated in this thesis. Numerical computation methods to include nonlocal effects in plasmonic nanostructures’ electromagnetic response are discussed, and applications of plasmonics to a few other fields are elaborated. First, a computation scheme is proposed to extend conventional finite-difference time-domain (FDTD) methods to nonlocal domain. An effective film whose response is derived from Feibelman's d-function formalism is to replace the highly non-uniform metal surfaces in simulations. It successfully produces numerical results of plasmonic resonance shift and field enhancement which agrees with the experimental data to first order. This scheme is still classical, thus very fast compared to the other first principle quantum methods such as density functional theory. Then electron's scattering rate in an effective medium with plasmonic nanostructures embedded-in, in random phase approximation, is developed, with the wavenumber dependence in the medium’s response accounted. Utilizing this calculation scheme of electron’s scattering rate, further specific applications are following. We show by simulation of the plasmonic nanostructures and calculation of the electron scattering rates that hot-electron plasmon-protection (HELPP) effects can protect the extra energy of hot electrons from being dissipated as heat. This can be a prototype of the 3rd generation solar cells. In another application, we investigate the electron polar-optical-phonon (POP) scattering in heavily-doped semiconductors when plasmonic nanostructures are embedded-in. We show that electron-POP scattering can be significantly suppressed compared to that of bulk semiconductors. In the third application, we propose the plasmonic multiple exciton generation (PMEG) scheme, with simulations and calculations, showing that the efficiency of multiple exciton generation in conventional semiconductors could be enhanced significantly with proper designed plasmonic nanostructures embedded-in or attached-adjacent. / Thesis (PhD) — Boston College, 2018. / Submitted to: Boston College. Graduate School of Arts and Sciences. / Discipline: Physics.

With the increasing demand of higher operating frequencies for electronic circuits, the printed circuit board designers face more electromagnetic radiation problems than ever. Some “rules-of-thumb” are employed to help the designers to reduce the radiation problems. The 20H rule is one of printed circuit design rules, which intends to minimize the electromagnetic radiation. This project focuses on analysis and simulation of 20H rule’s signal propagation mechanisms. The model used in the project is a 2D planar structure. The numerical electromagnetic method, Finite Difference Time Domain (FDTD) method, is used for the field computation and analysis. Simulation is based on various structures of model and different distributions of excitation sources. Analysis focuses on the signal propagation models. Field distributions and radiation patterns are visualized by mathematical software. Meanwhile, Poynting vectors are calculated to give quantitative expression. The simulation results indicate three factors, namely, operating frequency, size of PCB and separation distance that will affect the function of 20H rule. The effects of three factors are shown by comparison of specific cases in this thesis. / Singapore-MIT Alliance (SMA)

Modeling of radio wave propagation is indispensable for the design and analysis of wireless communication systems. The use of the Finite-Difference Time-Domain (FDTD) method for wireless channel modeling has gained significant popularity due its ability to extract wideband responses from a single simulation. FDTD-based techniques, despite providing accurate channel characterizations, have often employed point sources in their studies, mainly due to the large amounts of resources required for modeling fine geometrical details or features inherent in antennas into a discrete spatial domain. The underlying influences of the antenna on wave propagation have thus been disregarded. This work presents a possible approach for the efficient space-time analysis of antennas by deducing FDTD-compatible macromodels that completely encapsulate the electromagnetic behaviour of antennas and then incorporating them into a standard FDTD formulation for modeling their interactions with a general environment.