<p dir="ltr">The most predominantly used rotating electric machines today are the radial and axial flux varieties (denoted RFM and AFM, respectively). There is another category of machines called the transverse flux machines (denoted TFMs) which are best suitable for high torque low speed applications such as in wind energy conversion systems, ship propulsion systems, and other direct drive applications. In this work, a design methodology based on a magnetic equivalent circuit (MEC) model for a three-phase stacked transverse flux machine is presented. Using this MEC model, an optimization-based design paradigm is created. Finite element analysis is used to validate a design obtained from the proposed algorithm. </p>

The Finite Element Tearing and Interconnecting (FETI) and its variants are probably the most celebrated domain decomposition algorithms for partial differential equation (PDE) scientific computations. In electromagnetics, such methods have advanced research frontiers by enabling the full-wave analysis and design of finite phased array antennas, metamaterials, and other multiscale structures. Recently, closer scrutiny of these methods have revealed robustness and numerical scalability problems that prevent the most memory and time efficient variants of FETI from gaining widespread acceptance. This work introduces a new class of FETI methods and preconditioners that lead to exponential iterative convergence for a wide class of problems, are robust and numerically scalable. First, a two Lagrange multiplier (LM) variant of FETI with impedance transmission conditions, the FETI-2λ, is introduced to facilitate the symmetric treatment of non-conforming grids while avoiding matrix singularites that occur at the interior resonance frequencies of the domains. A thorough investigation on the approximability and stability of the Lagrange multiplier discrete space is carried over to identify the correct LM space basis. The resulting method, although accurate and flexible, exhibits unreliable iterative convergence. To accelerate the iterative convergence, the Locally Exact Algebraic Preconditioner (LEAP), which is responsible for improving the information transfer between neighboring domains is introduced. The LEAP was conceived by carefully studying the properties of the Dirichlet-to-Neumann (DtN) map that is involved in the sub-structuring process of FETI. LEAP proceeds in a hierarchical way and directly factorizes the signular and near-singular interactions of the DtN map that arise from domain-face, domain-edge and domain-vertex interactions. For problems with small number of domains LEAP results in scalable implementations with respect to the discretization. On problems with large domain numbers, the numerical scalability can only be obtained through ``global'' preconditioners that directly convey information to remotely separated domains at every DDM iteration. The proposed ``global" preconditiong stage is based on the new Multigrid FETI (MG-FETI) method. This method provides a coarse grid correction mechanism defined in the dual space. Macro-basis functions, that satisfy thecurl-curl equation on each interface are constructed to reduce the size of the coarse problem, while maintaining a good approximation of the characteristic field modes. Numerical results showcase the performance of the proposed method on one-way, 2D and 3D decomposed problems, with structured and unstructured partitioning, conforming and non-conforming interface triangulations. Finally, challenging, real life computational examples showcase the true potential of the method.

Metamaterials are a growing area of interest in the electromagnetics community due to their highly uncommon wave-material interaction characteristics, and they can be modeled using transmission line (TL) based networks. From verification of negative refraction to modeling more complex devices such as invisibility cloaks and field rotators, TL metamaterials offer a tangible solution to modeling novel devices in 1-D, 2-D, and 3-D structures. While currently available TL metamaterials allow for a predictable and easily manufactured network, the need for periodic, regular grids make current TL metamaterials sub-optimal for devices with curved boundaries or realization on curved surfaces. Our work presents the theory and application of TL metamaterials on irregular, nonperiodic grids for modeling 2-D electromagnetic phenomena in TE polarization, allowing for accuracy in curved device boundary modeling and significantly increased adaptability in potential application to curved surfaces. Based on an irregular grid obtained using an unstructured surface mesher, irregularly-shaped individual cells are related to local medium parameters to represent an overall device. The design method is validated using simple scattering problems with known analytical solutions and simulation data through lumped-element circuit- network simulations. The design process is then applied to more complex devices such as the Luneburg lens and field rotator. Capabilities and limitations of this technique are tested and explored. A microstrip based version of this method is subsequently developed and investigated using circuit and full-wave simulation data as well as experiment of a printed-circuit realization.

This thesis covers recent advances in the adjoint variable method for the sensitivity estimations through time-domain electromagnetic simulations. It considers both frequency-independent and frequency-dependent response functions, and at the same time, provides a novel adjoint treatment for addressing dispersive sensitivity parameters in the material constitutive relation. With this proposed adjoint technique, response sensitivities with respect to all N sensitivity parameters can be computed through at most one extra simulations regardless of the value of N. This thesis also extends the existing adjoint technique to estimate all N^2 second-order sensitivity entries in the response Hessian matrix through N additional simulations. All adjoint sensitivity techniques presented in this thesis are numerically validated through various practical examples. Comparison shows that our produced adjoint results agree with those produced through central finite-difference approximations or through exact analytical approaches. / Dissertation / Doctor of Engineering (DEng)

Fluorescence microscopy has played a pivotal role in the realm of biological and biomedical research, allowing researchers to delve into the intricacies of living organisms at the cellular and molecular levels. By using fluorescent probes, one can visualize specific molecules and structures within cells, fundamentally transforming our comprehension of biology and medicine. However, fluorescence microscopy faces its own set of challenges, namely, photobleaching and photodamage. Photobleaching involves the irreversible loss of fluorescence signal during imaging, while photodamage results in harmful effects on cells. Both severely limit fluorescence signal and observation time. Although remedies exist to mitigate these problems, most of them rely on chemical approaches. In this dissertation, to address these issues, I investigated two optical approaches that exploit control of light either in space or time. Firstly, I developed multiline scanning confocal microscopy (mLS) with a digital micro-mirror device. This method provides programmable patterns of the illumination beam as well as the detection slit. Through experimental results and optical simulations, I assessed the depth discrimination of mLS under different optical parameters and compared it with a multipoint system such as spinning disk confocal microscopy (SDCM). Surprisingly, under the same illumination duty cycle, I found that mLS offers better optical sectioning than SDCM. Importantly, the parallelized line illumination showed a much lower photobleaching rate compared to single-line scanning microscopy, while their optical sectioning capabilities remained similar. I applied this technique to visualize heterogeneous mouse epiblast stem cells, a challenging task in imaging. Secondly, I delved into low photobleaching rate two-photon microscopy (2PM). 2PM inherently provides excellent optical sectioning due to its nonlinear effects, making it suitable for high-resolution imaging within biological tissues. However, the high peak power of ultrafast pulses has always been associated with severe photobleaching, posing a longstanding challenge. I found that controlling the repetition rates of ultrafast lasers is a potential strategy to enhance photostability. Specifically, I used repetition rates lower or higher than 80 MHz in 2PM and conducted systematic experiments to investigate how optical parameters such as wavelengths, excitation powers, and pulse schemes can influence the photobleaching kinetics of fluorescent proteins and organic fluorophores. This thesis embarks on a journey to explore innovative strategies and methodologies aimed at reducing photobleaching while maintaining high-quality imaging in the realm of fluorescence microscopy.

Large area fiber array telescopes are a relatively modern development in the long history of telescope design and effectively create a single large equivalent aperture by combining many smaller unit telescopes at a fraction of the cost. In this study a spectrograph optical design is demonstrated that utilizes photonic lanterns in the input fiber feed for the Large Fiber Array Spectroscopic Telescope (LFAST) design concept, where photonic lanterns provide a simplified approach to the fiber feed originally proposed for LFAST. Conservation of etendue is applied to derive the relationship between slit size and photonic lantern ratio, and classical echelle spectrograph designs are explored. A high-resolution spectrograph for LFAST is shown to not be feasible due to the large number of input fibers that result in a large required cross-dispersion within the constraint of practical and cost-effective optical component sizes. An alternative approach of splitting the slit length and replicating spectrographs is explored, and an optimal choice of four spectrographs with a photonic lantern ration of four is determined. Even with this approach, detailed spectrograph component designs show that the spectrograph does not meet the resolution requirements and is fundamentally limited by design constraints imposed by component sizes. Alternative layouts and design decisions to improve optical performance are surveyed for future design research, and include white pupil optics, echelle grating design considerations, spectral arm splitting, and use of reflective and catadioptric systems with aspheric surfaces. While these show promise for improving the spectrograph performance, the design is demonstrated to still be challenging and will likely require innovative approaches to meet the high-resolution design goals.

Modern RF and millimeter-wave communication links call for high-efficiency front end systems with high output power and high linearity to meet minimum transmission requirements. Advanced modulation techniques, such as orthogonal frequency-division multiplexing (OFDM) require a large power amplifier (PA) dynamic range due to the high peak-to-average power ratio (PAPR). This thesis provides the analysis, design, and experimental verification of a high-efficiency, high-linearity S-band Doherty power amplifier (DPA) based on the Class F PA. Traditional Class F PAs use harmonically tuned output matching networks to obtain up to 88.4% power-added efficiency (PAE) theoretically, however the amplifier experiences poor linearity performance due to switched mode operation, typically yielding less than 30dB C/I ratio [1]. The DPA overcomes this linearity limitation by using an auxiliary amplifier to boost output power when the amplifier is subject to a high input power due to its limited conduction cycle. The DPA also provides improved saturated output power back-off performance to maintain high PAE during operation. The DPA presented in this thesis optimizes PAE while maintaining linearity by employing harmonically tuned Class F amplifier topology on a primary and an auxiliary amplifier. A Class F PA is first designed and fabricated to optimize output network linearity – this is followed by a DPA design based on the fabricated Class F PA. A GaN HEMT Class F PA and DPA operating at 2.2GHz are implemented with the PAs measuring 40% and 45% PAE respectively while maintaining a 30dB carrier-to-intermodulation (C/I) ratio on a two-tone test. The PAE is characterized at maximum 21dBm input power per tone and 20MHz tone spacing. When subject to a single 24dBm continuous wave input tone, the Class F PA and DPA output 37dBm and 35.5dBm respectively. The PAs presented in the thesis provide over 30dB C/I ratio up to 21dBm input tones while maintaining over 40% PAE suitable for base station applications.

京都大学 / 新制・課程博士 / 博士(工学) / 甲第24618号 / 工博第5124号 / 新制||工||1979(附属図書館) / 京都大学大学院工学研究科電気工学専攻 / (主査)教授 大村 善治, 教授 松尾 哲司, 教授 小嶋 浩嗣 / 学位規則第4条第1項該当 / Doctor of Philosophy (Engineering) / Kyoto University / DFAM

Due to the ever-increasing demands of highly integrated optical devices in imaging, spectroscopy, communications, and so on, there is a compelling need to design and characterize novel compact photonic components. The traditional approaches to realizing compact optical devices typically result in large footprints and sizable optical thicknesses. Moreover, they offer few degrees of freedom (DOF), hampering on-demand functionalities, on-chip integration, and scalability. This thesis will address the design and development of ultracompact diffractive devices for imaging and spectroscopy, utilizing advanced machine learning techniques and optimization algorithms. I first present the inverse design of ultracompact dual-focusing lenses and broad-band focusing spectrometers based on adaptive diffractive optical networks (a-DONs), which combine optical diffraction physics and deep learning capabilities for the inverse design of multi-layered diffractive devices. I designed two-layer diffractive devices that can selectively focus incident radiation over well-separated spectral bands at desired distances and also optimized a-DON-based focusing spectrometers with engineered angular dispersion for desired bandwidth and nanometer spectral resolution. Furthermore, I introduced a new approach based on a-DONs for the engineering of diffractive devices with arbitrary k-space, which produces improved imaging performances compared to contour-PSF approaches to lens-less computational imaging. Moreover, my method enables control of sparsity and isotropic k-space in pixelated screens of dielectric scatterers that are compatible with large-scale photolithographic fabrication techniques. Finally, by combining adjoint optimization with the rigorous generalized Mie theory, I developed and characterize functionalized compact devices, which I called "photonic patches," consisting of ~100 dielectric nanocylinders that achieve predefined functionalities such as beam steering, Fresnel zone focusing, local density of states (LDOS) enhancement, etc. My method enables the inverse design of ultracompact focusing spectrometers for on-chip planar integration. Leveraging multiple scattering of light in disordered random media, I additionally demonstrated a novel approach to on-chip spectroscopy driven by high-throughput multifractal (i.e., multiscale) media, resulting in sub-nanometer spectral resolution at the 50×50 µm²-scale footprint.

Reflector design stemmed from the need to shape the light emitted by candles or lamps. Over 2,000 years ago people realized that a mirror shaped as a parabola can concentrate light, and thus significantly boosts its intensity, to the point where objects can be set afire. Nowadays many applications require an accurate control of light, such as automotive headlights, streetlights, projection displays, and medical illuminators. In all cases light emitted from a light source can be shaped into a desired target distribution with a reflective surface. Design methods for systems with rotational and translational symmetry were devised in the 1930s. However, the freeform reflector shapes required to illuminate targets with no such symmetries proved to be much more challenging to design. Even when the source is assumed to be a point, the reflector shape is governed by a set of second-order partial non-linear differential equations that cannot be solved with standard numerical integration techniques. An iterative approach to solve the problem for a discrete target, known as the method of supporting ellipsoids, was recently proposed by Oliker. In this research we report several efficient implementations of the method of supporting ellipsoids, based on the point source approximation, and we propose new reflector design techniques that take into account the extent of the source. More specifically, this work has led to three major achievements. First, a thorough analysis of the method of supporting ellipsoids was performed that resulted in two alternative implementations of the algorithm, which enable a fast generation of freeform reflector shapes within the point source approximation. We tailored the algorithm in order to provide control over the parameters of interest to the designers, such as the reflector scale and geometry. Second, the shape generation algorithm was used to analyze how source flux can be mapped onto the target. We derived the condition under which a given source-target mapping can be achieved with a smooth continuous surface, referred as the integrability condition. We proposed a method to derive mappings that satisfy the integrability condition. We then use these mappings to quickly generate reflector shapes that create continuous target distributions as opposed to reflectors generated with the method of supporting ellipsoids that create discrete sets of points on the target. We also show how mappings that do not satisfy the integrability condition can be achieved by introducing step discontinuities in the reflector surface. Third, we investigated two methods to design reflectors with extended sources. The first method uses a compensation approach where the prescribed target distribution is adjusted iteratively. This method is effective for compact sources and systems with rotational or translational symmetry. The second method tiles the source images created by a reflector designed with the method of supporting ellipsoids and then blends the source images together using scattering in order to obtain a continuous target distribution. This latter method is effective for freeform reflectors and target distributions with no sharp variations. Finally, several case studies illustrate how these methods can be successfully applied to design reflectors for general illumination applications such as street lighting or luminaires. We show that the proposed design methods can ease the design of freeform reflectors and provide efficient, cost-effective solutions that avoid unnecessary energy consumption and light pollution.