Doctoral Dissertations

Permanent URI for this collection


Recent Submissions

Now showing 1 - 20 of 16969
  • Essays in Transportation and Electoral Politics
    Harmony, Xavier Joshua (Virginia Tech, 2024-03-01)
    Abstract 1 – The Importance of Transportation Policies in Local Elections Building and maintaining transportation systems is one of the most important functions of local government. It is a subject that concerns local residents, jurisdictions spend a lot of money on, and local politicians use to their political advantage. This study helps us understand how transportation issues feature in local elections. Through evaluating a dataset of 542 candidates from 219 local election races from 2022, this study explores which candidates for local office are more likely to have transportation policies, what kind of content is included in these policies, and what are the factors that make including different transportation content more or less likely. The analysis primarily uses website campaign content and a mix of qualitative and quantitative methods to answer these questions. I find a variety of factors affect the inclusion of transportation issues at the local level such as variations in governance, partisanship, and regional characteristics like a jurisdiction's size and transportation behavior. It was also evident that defining transportation issues was more common than proposing transportation policy solutions. Overall, this research provides more insight into how transportation policies are included in local elections. Abstract 2 – Saliency of Transportation Policies in State Legislative Elections: The Case of Virginia Transportation systems are expensive and directly impact important issues like climate change, equity, and quality of life. However, it is not clear how important transportation policies are in state-level elections. Using the Virginia 2021 state legislative election, this research uses candidate website data, Twitter data, and data about Virginia House of Delegates districts to answer three questions: which candidates are more likely to have transportation polices, what issues or transportation modes are included, and what factors make candidates more or less likely to focus on certain issues. Using descriptive statistics, and regression methods, this research found transportation issues varied by political party with top overall issues including transportation funding as well as expanding or improving transportation systems. Public transportation was the top non-car mode. Candidates were more likely to include transportation issues if district households had higher car ownership or a lower percentage of single occupancy vehicle commuters. Finally, differences in transportation issues could be partly explained by political party, incumbency, population density, and transportation habits. These results will be helpful for understanding how state government transportation agendas change, can better inform transportation advocacy efforts, and could help transportation professionals better understand the impact of their work. Abstract 3 – Does Voting Affect the Provision of Bus Service? Inequalities in the distribution of bus services are important to understand. This chapter adds to previous literature by exploring why inequalities exist. Specifically, does voting for elected officials affect inequalities in the delivery of bus services? This study explores this question using a quantitative approach as part of a quasi-experimental research design focusing on GoRaleigh in North Carolina and the Milwaukee County Transit System in Wisconsin. The analysis provides evidence of a relationship between voting behavior and bus service. This finding is observed across cities and elections with the relationships holding even when controlling for factors associated with a bureaucratic explanation for changing bus service, like changes to population or jobs. However, the strength of the relationship can change between elections, the type of elected official, and cities. Overall, this work provides more evidence of the politics behind transit service planning, especially the political influences of voting behavior in representative democracies.
  • Long-Pulsed Laser-Induced Cavitation: Laser-Fluid Coupling, Phase Transition, and Bubble Dynamics
    Zhao, Xuning (Virginia Tech, 2024-02-29)
    This dissertation develops a computational method for simulating laser-induced cavitation and investigates the mechanism behind the formation of non-spherical bubbles induced by long-pulsed lasers. The proposed computational method accounts for the laser emission and absorption, phase transition, and the dynamics and thermodynamics of a two-phase fluid flow. In this new method, the model combines the Navier-Stokes (NS) equations for a compressible inviscid two-phase fluid flow, a new laser radiation equation, and a novel local thermodynamic model of phase transition. The Navier-Stokes equations are solved using the FInite Volume method with Exact two-phase Riemann solvers (FIVER). Following this method, numerical fluxes across phase boundaries are computed by constructing and solving one-dimensional bi-material Riemann problems. The new laser radiation equation is derived by customizing the radiative transfer equation (RTE) using the special properties of laser, including monochromaticity, directionality, high intensity, and a measurable focusing or diverging angle. An embedded boundary finite volume method is developed to solve the laser radiation equation on the same mesh created for the NS equations. The fluid mesh usually does not resolve the boundary and propagation directions of the laser beam, leading to the challenges of imposing the boundary conditions on the laser domain. To overcome this challenge, ghost nodes outside the laser domain are populated by mirroring and interpolation techniques. The existence and uniqueness of the solution are proved for the two-dimensional case, leveraging the special geometry of the laser domain. The method is up to second-order accuracy, which is also proved, and verified using numerical tests. A method of latent heat reservoir is developed to predict the onset of vaporization, which accounts for the accumulation and release of latent heat. In this work, the localized level set method is employed to track the bubble surface. Furthermore, the continuation of phase transition is possible in laser-induced cavitation problems, especially for long-pulsed lasers. A method of local correction and reinitialization is developed to account for continuous phase transitions. Several numerical tests are presented to verify the convergence of these methods. This multiphase laser-fluid coupled computational model is employed to simulate the formation and expansion of bubbles with different shapes induced by different long-pulsed lasers. The simulation results show that the computational method can capture the key phenomena in the laser-induced cavitation problems, including non-spherical bubble expansion, shock waves, and the ``Moses effect''. Additionally, the observed complex non-spherical shapes of vapor bubbles generated by long-pulsed laser reflect some characteristics (e.g., direction, width) of the laser beam. The dissertation also investigates the relation between bubble shapes and laser parameters and explores the transition between two commonly observed shapes -- namely, a rounded pear-like shape and an elongated conical shape -- using the proposed computational model. Two laboratory experiments are simulated, in which Holmium:YAG and Thulium fiber lasers are used respectively to generate bubbles of different shapes. In both cases, the predicted bubble nucleation and morphology agree reasonably well with the experimental observation. The full-field results of laser radiance, temperature, velocity, and pressure are analyzed to explain bubble dynamics and energy transmission. It is found that due to the lasting energy input, the vapor bubble's dynamics is driven not only by advection, but also by the continued vaporization at its surface. Vaporization lasts less than 1 microsecond in the case of the pear-shaped bubble, compared to over 50 microseconds for the elongated bubble. It is thus hypothesized that the bubble's morphology is determined by a competition between the speed of bubble growth due to advection and continuous vaporization. When the speed of advection is higher than that of vaporization, the bubble tends to grow spherically. Otherwise, it elongates along the laser beam direction. To test this hypothesis, the two speeds are defined analytically using a model problem and then estimated for the experiments using simulation results. The results support the hypothesis and also suggest that when the laser's power is fixed, a higher laser absorption coefficient and a narrower beam facilitate bubble elongation.
  • Volumetric Properties and Viscosity of Lubricant Oils and the Effects of Additives at High Pressure and Temperatures
    Avery, Katrina Nichole (Virginia Tech, 2024-02-26)
    This research is directed to the characterization of the thermodynamic properties and viscosity of lubricant base oils modified with polymeric additives. Several groups of mineral and synthetic base oils, including Ultra S4, Ultra S8, and poly alpha olefin PAO 4 have been studied. Among the various types of additives explored were viscosity index modifiers, polyisobutylene polymers (PIBs), and dispersants. The viscosity index modifiers are studied in terms of different polymer architectures, molecular weights, presence or absence of functional groups, and their concentrations. The dispersants are studied in terms of concentration, molecular weight, and presence of capping groups. Density data, as the basic thermodynamic data, are generated using a high-pressure variable-volume view-cell over a pressure range from 10 to 40 MPa and a range of temperatures from 298 to 398 K. The density data are then correlated with the Sanchez-Lacombe equation of state, from which key thermodynamic properties, namely isothermal compressibility, isobaric expansivity, and internal pressure are derived. These properties offer a rational approach to better understand molecular packing in lubricants under high pressure and temperature conditions which has direct impact on film formation. Viscosity determinations are carried out using a custom-designed high-pressure rotational viscometer. Data were generated in the pressure range from 10 to 40 MPa, but at temperatures ranging from 298 to 373 K as a function of shear rate up to 1270 s-1. Viscosity data were then correlated with density which provides interpretations in terms of free-volume and density scaling models. The molecular parameters produced from these correlations support the interpretation of molecular packing under high pressure and temperature conditions. The results of this study included several key findings. With regards to density, the addition of viscosity index modifiers to Ultra S4 base oil caused the density to increase, except for the addition of functionalized olefin copolymers (OCPs) which caused the density to decrease. This was true with both high and low molecular weight additives. In the case of Ultra S8 base oil, the addition of OCPs generally decreased the density, while the addition of polymethacrylates (PMAs) caused the density to increase. In terms of compressibility and expansivity, the addition of high molecular weight viscosity index modifiers to Ultra S4 base oil generally decreased both these properties. However, the compressibility increased with the addition of 5 wt % functionalized PMA and 2 wt % star styrene butadiene (SSB). Furthermore, there was less of a decrease in compressibility with the addition of functionalized additives. With the addition of low molecular weight viscosity index modifiers to Ultra S4 base oil, little change was observed in compressibility, and the expansivity decreased to a lesser degree than with the addition of high molecular weight viscosity index modifiers. Viscosity index modifiers did not alter the compressibility of Ultra S8 base oil. Compared to Ultra S4, expansivity in Ultra S8 decreased to a lesser extent. The internal pressure was observed to be lowered to a greater degree in either Ultra S4 or Ultra S8 base oil with the addition of additives with more rigid internal structures (PMA and SSB). The decrease occurred to a greater degree with the addition of the higher molecular weight versions of additives studied and/or with the incorporation of functional groups to the additives. Although density changes were often greater with the addition of additives to the Ultra S8 base oil, all other derived thermodynamic properties, including internal pressure, changed to a greater degree with the addition of additives to the lower molecular weight Ultra S4 base oil. The viscosity generally increased to varying degrees with the addition of different additives to either base oil. The addition of functionality and higher molecular weight additives led to more consistent viscosity increases at higher temperatures. At the highest viscosity isotherm tested, 373 K, the addition of viscosity index modifiers resulted in similar viscosity values in either base oil, even though the viscosity of Ultra S4 at 373 K is much lower than the viscosity of Ultra S8 at this temperature. However, at 298 K, the viscosity index modifiers increased the viscosity of the Ultra S8 base oil to much higher values than the viscosity of the Ultra S4 base oil. Model based correlations of viscosity showed that with addition of high molecular weight viscosity index modifiers to Ultra S4 base oil, the parameters that are linked to free-volume overlap and the density dependence were more sensitive to the addition of OCPs than with the addition of PMAs and SSBs. These changes were reflected in larger free-volume overlap parameters and larger density exponent values. However, with a low molecular weight addition, the resulting parameters changed more with the addition of PMAs than OCPs. Overall, the addition of polymers with more rigid architecture led to more similar changes in correlative parameters across molecular weights from that of the original base oil, while for the OCP addition, the molecular weight had more of an influence on the degree of change. With addition of viscosity index modifiers to the Ultra S8 base oil, the architecture of the additive had more of an influence on the viscosity correlation parameters as the addition of PMA led to more noticeable changes in the parameters (resulting in lower free-volume overlap parameters, and a lower density exponent) than the addition of OCP, irrespective of the molecular weight or functionality. In either base oil, the addition of PMA led to lower free-volume overlap parameters and density exponent values than the addition of OCP. In this study it was observed that the addition of functionality, or polar groups to viscosity index modifiers, led to more desirable thermodynamic and rheological property changes to the lubrication base oil. This change was more definitive with the addition of polymers with more rigid architecture, such as PMAs and SSBs in contrast to the OCPs. The study on the addition of PIBs and capped or uncapped dispersants showed little variation in the resulting density and viscosity values when added to Ultra S4 base oil. However, the compressibility in these systems generally increased while the expansivity decreased except with the addition of PIBs. The internal pressure decreased to similar levels for all additive additions, except for the lowest molecular weight PIB, in which there was little change. The study on the addition of PIBs to different base oils showed that low molecular weight PIBs had the potential to disrupt the packing of a more uniform PAO 4 base oil and change the thermodynamic properties and correlation parameters to a greater degree than with the addition of higher molecular weight PIBs. This resulted in higher compressibility and internal pressure values with the addition of low molecular weight PIB compared to the higher molecular weight PIBs. However, there was little variation in viscosity with any of the PIB additions, except for the highest molecular weight PIB.
  • Modeling the Dynamics of Liquid Metal in Fusion Liquid Walls Using Maxwell-Navier-Stokes Equations
    Murugaiyan, Suresh (Virginia Tech, 2024-02-23)
    The dissertation explores a framework for numerically simulating the deformation of the liquid metal wall's free surface in Z-pinch fusion devices. This research is conducted in the context of utilizing liquid metals as plasma-facing components in fusion reactors. In the Z-pinch fusion process, electric current travels through a plasma column and enters into a pool of liquid metal. The current flowing through the liquid metal generates Lorentz force, which deforms the free surface of the liquid metal. Modeling this phenomenon is essential as it offers insights into the feasibility of using liquid metal as an electrode wall in such fusion devices. The conventional magneto-hydrodynamic (MHD) formulation aims at modeling the situation where an external magnetic field is applied to flows involving electrically conducting liquids, with the initial magnetic field is known and then evolved over time through magnetic induction equation. However, in Z-pinch fusion devices, the electric current is directly injected into a conducting liquid. In these situations, an analytical expression for the magnetic field generated by the applied current is not readily available, necessitating numerical calculations. Moreover, the deformation of the liquid metal surface changes the geometry of the current path over time and the resulting magnetic field. By directly solving the Maxwell equations in combination with Navier-Stokes equations, it becomes possible to predict the magnetic field even when the fluid is in motion. In this dissertation, a numerical framework utilizing the Maxwell-Navier-Stokes system is explored to successfully capture the deformation of the liquid metal's free surface due to applied electric current.
  • Emitting Wall Boundary Conditions in Continuum Kinetic Simulations: Unlocking the Effects of Energy-Dependent Material Emission on the Plasma Sheath
    Bradshaw, Kolter Austen (Virginia Tech, 2024-02-23)
    In a wide variety of applications such as the Hall thruster and the tokamak, understanding the plasma-material interactions which take place at the wall is important for improving performance and preventing failure due to material degradation. In the region near a surface, the plasma sheath forms and regulates the electron and ion fluxes into the material. Emission from the material has the potential to change sheath structure drastically, and must be modeled rigorously to produce accurate predictions of the fluxes into the wall. Continuum kinetic codes offer significant advantages for the modeling of sheath physics, but the complexity of emission physics makes it difficult to implement accurately. This difficulty results in major simplifications which often neglect important energy-dependent physics. A focus of the work is on proper simulation of the sheath. The implementation of source and collision terms is discussed, alongside a brief study of the Weibel instability in the sheath demonstrating the necessity of proper collision implementation to avoid missing relevant physics. A novel implementation of semi-empirical models for electron-impact secondary electron emission into the boundary conditions of a continuum kinetic code is presented here. The features of both high and low energy regimes of emission are represented self-consistently, and the underlying algorithms are flexible and can be easily extended to other emission mechanisms, such as ion-impact secondary electron emission. The models are applied to simulations of oxidized and clean lithium for fusion-relevant plasma regimes. Oxidized lithium has a high emission coefficent and the sheath transitions into space-charge limited and inverse modes for different parameters. The breakdown of the classical sheath results in an increase of energy fluxes to the surface, with potential ramification for applications.
  • Evaluating Factors Contributing to Crash Severity Among Older Drivers: Statistical Modeling and Machine Learning Approaches
    Alrumaidhi, Mubarak S. M. S. (Virginia Tech, 2024-02-23)
    Road crashes pose a significant public health issue worldwide, often leading to severe injuries and fatalities. This dissertation embarks on a comprehensive examination of the factors affecting road crash severity, with a special focus on older drivers and the unique challenges introduced by the COVID-19 pandemic. Utilizing a dataset from Virginia, USA, the research integrates advanced statistical methods and machine learning techniques to dissect this critical issue from multiple angles. The initial study within the dissertation employs multilevel ordinal logistic regression to assess crash severity among older drivers, revealing the complex interplay of various factors such as crash type, road attributes, and driver behavior. It highlights the increased risk of severe crashes associated with head-on collisions, driver distraction or impairment, and the non-use of seat belts, specifically affecting older drivers. These findings are pivotal in understanding the unique vulnerabilities of this demographic on the road. Furthermore, the dissertation explores the efficacy of both parametric and non-parametric machine learning models in predicting crash severity. It emphasizes the innovative use of synthetic resampling techniques, particularly random over-sampling examples (ROSE) and synthetic minority over-sampling technique (SMOTE), to address class imbalances. This methodological advancement not only improves the accuracy of crash severity predictions for severe crashes but also offers a comprehensive understanding of diverse factors, including environmental and roadway characteristics. Additionally, the dissertation examines the influence of the COVID-19 pandemic on road safety, revealing a paradoxical decrease in overall traffic crashes accompanied by an increase in the rate of severe injuries. This finding underscores the pandemic's transformative effect on driving behaviors and patterns, heightening risks for vulnerable road users like pedestrians and cyclists. The study calls for adaptable road safety strategies responsive to global challenges and societal shifts. Collectively, the studies within this dissertation contribute substantially to transportation safety research. They demonstrate the complex nature of factors influencing crash severity and the efficacy of tailored approaches in addressing these challenges. The integration of advanced statistical methods with machine learning techniques offers a profound understanding of crash dynamics and sets a new benchmark for future research in transportation safety. This dissertation underscores the evolving challenges in road safety, especially amidst demographic shifts and global crises, and advocates for adaptive, evidence-based strategies to enhance road safety for all, particularly vulnerable groups like the older drivers.
  • Design, Fabrication, Characterization, and Packaging of Gallium Oxide Power Diodes
    Wang, Boyan (Virginia Tech, 2024-02-22)
    Gallium Oxide (Ga2O3) is an ultra-wide bandgap semiconductor with a bandgap of 4.5–4.9 eV, which is larger than that of Silicon (Si), Silicon Carbide (SiC), and Gallium Nitride (GaN). A benefit of this ultra-wide bandgap is the high-temperature stability due to the low intrinsic carrier concentration. Another benefit is the high critical electric field (Ec), which is estimated to be from 6 MV/cm to 8 MV/cm in Ga2O3. This allows for a superior Baliga's figure of merit (BFOM) of unipolar Ga2O3 power devices, i.e., they potentially can achieve a smaller specific on-resistance (RON,SP) as compared to the Si, SiC, and GaN devices with the same breakdown voltage (BV). The above prospects make Ga2O3 devices the promising candidates for next-generation power electronics. This dissertation explores the design, fabrication, characterization, and packaging of vertical β-Ga2O3 Schottky barrier diodes (SBDs) and P-N diodes. The power SBDs allow for a small forward voltage and a fast switching speed; thus, it is ubiquitously utilized in power electronics systems. Meanwhile, the Ga2O3 power P-N diodes have the benefit of smaller leakage current, and the diode structure could be a building block for many advanced diodes and transistors. Hence, the study of Ga2O3 Schottky and P-N diodes is expected to provide the foundation for developing a series of Ga2O3 power devices. Firstly, vertical Ga2O3 Schottky and P-N diodes with a novel edge termination (ET), the multi-layer Nickel Oxide (NiO) junction termination extension (JTE), are fabricated on Ga2O3 substrates. This multi-JTE NiO structure decreases the peak electric field (Epeak) at the triple point of device edge when the Ga2O3 diodes are reversely biased. For SBDs, BV reach 2.5 kV, the 1-D junction field reaches 3.08 MV/cm, and the BFOM exceeds 1 GW/cm2. For P-N diodes, BV reaches 3.3 kV, the junction field reaches 4.2 MV/cm, and the BFOM reaches 2.6 GW/cm2. These results are among the highest in Ga2O3 power devices and are comparable to the state-of-the-art vertical GaN Schottky and P-N diodes. Notably, all these diodes are small-area devices. Secondly, large-area (3 mm×3 mm anode size) Ga2O3 Schottky and P-N diodes with high current capability are fabricated to explore the packaging, thermal management, and switching characteristics of Ga2O3 diodes. The same ET is applied for the large-area P-N diode. The fabricated large-area P-N diodes have a turn-on voltage of 2 V, a differential on-resistance (Ron) of 0.2 Ω, and they can reach at least 15 A when measured in the pulse mode. The BV of large-area Ga2O3 P-N diodes varies due to the fabrication non-uniformity, but the best device achieves a BV of 1.6 kV, standing among the highest values reported for large-area Ga2O3 diodes. Also, the large-area Ga2O3 SBDs with similar current rating but with a FP ET are fabricated mainly for the packaging and thermal management studies. Thirdly, medium-area Ga2O3 P-N diodes with a current over 1 A and a higher yield of BV are fabricated to evaluate the JTE's capacitance and switching characteristics. The JTE accounts for only ~11% of the junction capacitance of this 1 A diode, and the percentage is expected to be even smaller for higher-current diodes. The turn-on/off speed and reverse recovery time of the diode are comparable to commercial SiC Schottky barrier diodes under the on-wafer switching test. These results show the viability of NiO JTE for enabling a fast switching speed in high-voltage Ga2O3 power devices. Fourthly, the fabricated large-area Ga2O3 diodes are packaged using silver sintering as the die attach. The sintered silver joint has higher thermal conductivity (kT) and better reliability as compared to the solder joint. Due to the low kT of Ga2O3 material, junction-side-cooled (JSC) packaging configuration is necessary for Ga2O3 devices. For the packaged device, its junction-to-case thermal resistance (RθJC) is measured in the bottom-side-cooled (BSC) and junction-side-cooled (JSC) configuration by the transient dual interface method according to the JEDEC 51-14 standard. The RθJC of the junction- and bottom-cooled Ga2O3 SBD is measured to be 0.5 K/W and 1.43 K/W, respectively. The former RθJC is lower than that of similarly-rated commercial SiC SBDs. This manifests the significance of JSC packaging for the thermal management of Ga2O3 devices. Fifthly, to evaluate the electrothermal robustness of the packaged Ga2O3 devices, the surge current capability of JSC packaged Ga2O3 SBDs are measured. The Ga2O3 SBDs with proper packaging show high surge current capabilities. The double-side-cooled (DSC) large-area Ga2O3 SBDs can sustain a peak surge current over 60 A, with a ratio between the peak surge current and the rated current superior to that of similarly-rated commercial SiC SBDs. These results show the excellent ruggedness of Ga2O3 power devices. Finally, a Ga2O3 integrated diode module consisting of four single-diode sub-modules is designed and fabricated. For many power electronics applications, high current is desired; however, for emerging semiconductors, the current upscaling is difficult by directly increasing the device area because of the limitation of heat extraction capability and the limited material/processing yield. Here we explore the paralleling of multiple Ga2O3 P-N diodes to increase the current level. For each sub-module, the JSC packaging structure is used for heat extraction, and a metal post is sintered to the anode for electric field (E-field) management. RθJC is measured to be 1 W/K for each sub-module. On-board double-pulsed test is performed for both the sub-module and the full module. The sub-module and full module demonstrate 400 V, 10 A and 150 V, 70 A switching capabilities, respectively. This is the first demonstration of Ga2O3 power module and shows a promising approach to upscale of the power level of Ga2O3 power electronics. In addition to Ga2O3 device study, a research is conducted to explore the chip size (Achip) minimization for wide-bandgap (WBG) and ultra-wide bandgap (UWBG) power devices. Achip optimization is particularly critical for WBG and UWBG power devices and modules due to the high material cost. This work presents a new, holistic, electrothermal approach to optimize Achip for a given set of target specifications including BV, conduction current (I0), and switching frequency (f). The conduction and switching losses of the device are considered, as well as the heat dissipation in the chip and its package. For a given BV and I0, the optimal Achip, Wdr, and Ndr show strong dependence on f and thermal management. Our approach offers more accurate cost analysis and design guidelines for power modules. In summary, this dissertation covers the design, fabrication, characterization, and packaging of Ga2O3 Schottky and P-N diodes, with the aim to advance Ga2O3 devices to power electronics applications. This dissertation addresses many knowledge gaps on Ga2O3 devices, including the voltage upscaling (ET), current upscaling (large-area device fabrication, packaging, and thermal management), and their concurrence (module demonstration), as well as the circuit-level switching characterizations.
  • Synthetic Solidarities: Theorizing Queer Affectivity and Trans*national/temporal Emulsification as Embodied Resistance to Global Capitalism
    Tepper, Madison Jeanette (Virginia Tech, 2024-02-20)
    This dissertation theorizes the synthesis of solidarities around queer embodied performativities as a mode of making-resistant the everyday experiences of exploitation under transnational capitalism. These solidarities, I argue, are cultivated around the affective, embodied experiences of what José Esteban Muñoz terms "queer time," which I extend to denote the ephemeral, experiential sensations of being "out of sync" with the structures and norms of capital-space-time power assemblages. I theorize "emulsion" as a heuristic for envisioning synthetic solidarities as making space and time for the importantly distinct experiences of queer spatio-temporalities of those at the various intersections of marginalized/minoritized identities to coagulate and coalesce into something new – at once remaining beautifully fragmented and becoming grotesquely amalgamated beyond distinction. I suggest that such trans-spatial/temporal/material solidarities, formed via antinormative performativities and the curation of "revolting archives," existent and not-yet-formed alike, can and indeed already do resist the totalizing and unplaceable ether of increasingly transnational capitalism across scales. This dissertation takes form and transdisciplinarity to be a part of the praxis/theory of cultivating such synthetic solidarities that confound the structures of capital-space-time. As such, I (gender)fuck with genre, and format throughout, interweaving theoretical and autotheoretical writing with prose, poetics, and altered text to create a visceral sense of disruption of spatiotemporality in not only content, but the affective experience of reading the piece itself. This dissertation thus moves across disciplines via a theoretical constellation of critical scholarship including affect theory, queer theory, (neo)Marxist theory, Black feminist theory, post- and de-colonial theory, disability theory, and transnational feminism.
  • Geochemical investigation of the co-evolution of life and environment in the Neoproterozoic Era
    Kang, Junyao (Virginia Tech, 2024-02-19)
    The co-evolution of life and the environment stands as a cornerstone in Earth's 4.5-billion-year history. Environmental fluctuations have wielded substantial influence over biological evolution, while life forms have, in turn, reshaped Earth's surface and climate. This dissertation centers on a critical period in Earth's history—the Neoproterozoic Era—when profound environmental shifts potentially catalyzed pivotal eukaryotic evolutionary events. By delving deeper into Neoproterozoic paleoenvironments, I aim at a clearer understanding of life-environment co-evolution in this crucial era. The first chapter focuses on an important juncture—the transition from prokaryote to eukaryote dominance in marine ecosystems during the Tonian Period (1000 Ma to 720 Ma). To assess whether the availability of nitrate, an important macro-nutrient, played a critical role in this evolutionary event, nitrogen isotope compositions (δ15N) of marine carbonates from the early Tonian (ca. 1000 Ma to ca. 800 Ma) Huaibei Group in North China were measured. The data indicate nitrate limitation in early Neoproterozoic oceans. Further, a compilation of Proterozoic sedimentary δ15N data, together with box model simulations, suggest a ~50% increase in marine nitrate availability at ~800 Ma. Limited nitrate availability in early Neoproterozoic oceans may have delayed the ecological rise of eukaryotes until ~800 Ma when increased nitrate supply, together with other environmental and ecological factors, may have contributed to the transition from prokaryote-dominant to eukaryote-dominant marine ecosystems. Recognizing the spatial and temporal variations in Neoproterozoic oceanic environments, the second chapter lays the groundwork for a robust stratigraphic framework for the early Tonian Period. Employing the dynamic time warping algorithm, I constructed a global stratigraphic framework for the early Tonian Period using δ13Ccarb data from the North China, São Francisco, and Congo cratons. This exercise confirms the generally narrow range of δ13Ccarb fluctuations in the early Tonian, but also confirms the presence of a negative δ13Ccarb excursion of notable magnitude (~9 ‰) at ca. 920 Ma in multiple records, suggesting that it was global in scope. This negative excursion, known as the Majiatun excursion, is likely the oldest negative excursion in the Neoproterozoic Era and marks the onset of the dynamic Neoproterozoic carbon cycle. Shifting focus to the late Neoproterozoic, the third chapter delves into the origins of Neoproterozoic superheavy pyrite, whose bulk-sample δ34S values are greater than those of contemporaneous seawater sulfate and whose origins remain controversial. Two supervised machine learning algorithms were trained on a large LA-ICP-MS pyrite trace element database to distinguish pyrite of different origins. The analysis validates that two models built on the co-behavior of 12 trace elements (Co, Ni, Cu, Zn, As, Mo, Ag, Sb, Te, Au, Tl, and Pb) can be used to accurately predict pyrite origins. This novel approach was then used to identify the origins of pyrite from two Neoproterozoic sedimentary successions in South China. The first set of samples contains isotopically superheavy pyrite from the Cryogenian Tiesi'ao and Datangpo formations. The second set of samples contains pyritic rims from the Ediacaran Doushantuo Formation; these pyrite rims are associated with fossiliferous chert nodules and do not have superheavy sulfur isotopes. For the superheavy pyrite, the models consistently show high confidence levels in identifying its genesis type, and three out of four samples were inferred to be of sedimentary origins. For the pyritic nodule rims, the models suggest that early diagenetic pyrite was subsequently altered by hydrothermal fluids and therefore shows mixed signals. The third chapter highlights the importance of pyrite trace elements in deciphering and distinguishing the origins of pyrite in sedimentary strata.
  • The Dynamics of the Impacts of Automated Vehicles: Urban Form, Mode Choice, and Energy Demand Distribution
    Wang, Kaidi (Virginia Tech, 2021-08-24)
    The commercial deployment of automated vehicles (AVs) is around the corner. With the development of automation technology, automobile and IT companies have started to test automated vehicles. Waymo, an automated driving technology development company, has recently opened the self-driving service to the public. The advancement in this emerging mobility option also drives transportation reasearchers and urban planners to conduct automated vehicle-related research, especially to gain insights on the impact of automated vehicles (AVs) in order to inform policymaking. However, the variation with urban form, the heterogeneity of mode choice, and the impacts at disaggregated levels lead to the dynamics of the impacts of AVs, which not comprehensively understood yet. Therefore, this dissertation extends existing knowledge base by understanding the dynamics of the impacts from three perspectives: (1) examining the role of urban form in the performance of SAV systems; (2) exploring the heterogeneity of AV mode choices across regions; and (3) investigating the distribution of energy consumption in the era of AVs. To examine the first aspect, Shared AV (SAV) systems are simulated for 286 cities and the simulation outcomes are regressed on urban form variables that measure density, diversity, and design. It is suggested that the compact development, a multi-core city pattern, high level of diversity, as well as more pedestrian-oriented networks can promote the performance of SAVs measured using service efficiency, trip pooling success rate, and extra VMT generation. The AV mode choice behaviors of private conventional vehicle (PCV) users in Seattle and Knasas City metropolitan areas are examined using an interpretable machine learning framework based on an AV mode choice survey. It is suggested that attitudes and trip and mode-specific attributes are the most predictive. Positive attitudes can promote the adoption of PAVs. Longer PAV in-vehicle time encourages the residents to keep the PCVs. Longer walking distance promotes the usage of SAVs. In addition, the effects of in-vehicle time and walking distance vary across the two examined regions due to distinct urban form, transportation infrustructure and cultural backgrounds. Kansas City residents can tolerate shorter walking distance before switching to SAV choices due to the car-oriented environment while Seattle residents are more sensitive to in-vehicle travel time because of the local congestion levels. The final part of the dissertation examines the demand for energy of AVs at disaggregated levels incorporating heterogeneity of AV mode choices. A three-step framework is employed including the prediction of mode choice, the determination of vehicle trajectories, and the estimation of the demand for energy. It is suggested that the AV scenario can generate -0.36% to 2.91% extra emissions and consume 2.9% more energy if gasoline is used. The revealed distribution of traffic volume suggests that the demand for charging is concentrated around the downtown areas and on highways if AVs consume electricity. In summary, the dissertation demonstrates that there is a dynamics with regard to the impacts and performance of AVs across regions due to various urban form, infrastructure and cultural environment, and the spatial heterogeneity within cities.
  • Leveraging Street View and Remote Sensing Imagery to Enhance Air Quality Modeling through Computer Vision and Machine Learning
    Qi, Meng (Virginia Tech, 2024-02-14)
    Air pollution is associated with various adverse health impacts and is identified as one of the leading risk factors for global disease burden. Further, air pollution is one of the pathways through which climate change could negatively impact health. Field studies have shown that air pollution has high spatiotemporal variability and pollutant concentrations vary substantially within neighborhoods. Characterizing air pollution at a fine-grained level is essential for accurately estimating human exposure, assessing its impact to human health, and further aiding localized air pollution policy. Air quality models are developed to estimate air pollution at locations and time periods without monitors, and these estimates are commonly used for exposure and health effects studies. Traditional land use regression [LUR] models are one of the cost-effective empirical air quality models. LUR typically relies on fixed-site measurements, GIS-derived variables with limited spatial resolution, and captures linear relationships. In recent years, innovative open-source imagery datasets and their associated features (e.g., street view imagery, remote sensing imagery) have emerged and show potential to augment or replace traditional LUR predictors. Such imagery data sources embody abundant information of natural and built environment features. Advanced computer vision techniques enable feature extraction and quantification through these extensive imagery datasets. The overarching objective of this dissertation is to investigate the feasibility of leveraging open-source imagery datasets (i.e., Google Street View [GSV] imagery, Landsat imagery, etc.) and advanced machine learning algorithms to develop image-based empirical air quality models at both local and national scale. The first study of this work established a pipeline of feature extraction through street view imagery sematic segmentation. The resulting street view features were used to predict street-level particulate air pollution for a single city. The results showed that solely using GSV-derived features can achieve comparable model fits as using traditional GIS-derived variables. Feature engineering improved model stability and interpretability through reducing spurious variables from potential misclassifications from computer vision algorithms. The second study further developed GSV-based models at national scale across multiple years. Random forest models were developed to capture the nonlinear relationship between air pollution and its impacting factors. The results showed that with sufficient street view images, GSV imagery alone may explain the variation of long-term national NO2 concentrations. Adding satellite-derived aerosol estimates (i.e., OMI column density) can significantly boost model performance when GSV images are insufficient, but the addition narrows when more GSV images are available. Our systematic assessment of the impact of image availability on model performance suggested that a parsimonious image sampling strategy (i.e., one GSV image per 100m grid) may be sufficient and most cost-effective for model development and application. Our third study explored the feasibility of combining street view and remote sensing derived features for national NO2 and PM2.5 modeling and projection at high spatial resolution. We found that GSV-based models captured both the highest and lowest pollutant concentrations while remote sensing features tended to smooth the air pollution variations. The results suggested that GSV features may have the capability to better capture fine-scale air pollution variability. The resulting air pollution prediction product may serve a variety of applications, including providing new insights into environmental justice and epidemiological studies due to its high spatial resolution (i.e., street level). Collectively, the result of this dissertation suggests that GSV imagery, processed with computer vision techniques, is a promising data source to develop empirical air quality models with high spatial resolution and consistent predictor variables processing protocol. Image-based features assisted with advanced ML approaches have the potential to greatly improve air quality modeling estimates, and successfully show comparable and even superior model performance than other modeling studies. Moreover, the ever-growing public imagery data sources are particularly promising for remote or less developed areas where traditional curated geodatabases are sparse or nonexistent.
  • Critical Elements Recovery from Acid Mine Drainage
    Li, Qi (Virginia Tech, 2024-02-13)
    The rapid development of advanced technologies has led to an increase in demand for critical elements that are essential in the manufacturing of high-tech products. Among these critical elements, manganese (Mn), cobalt (Co), and nickel (Ni) are used in the production of batteries, electronics, and other advanced applications. The demand for these elements has been growing exponentially in recent years, driven by the rise of electric vehicles, renewable energy, and other emerging technologies. However, the United States is heavily dependent on foreign sources of critical minerals and on foreign supply chains, resulting in the potential for strategic vulnerabilities to both economy and military. To address this problem and reduce the Nation's vulnerability to disruptions in the supply of critical minerals, it is important to develop critical minerals recycling technologies. A systematic study was conducted to develop a process for producing high-purity Mn, Co, and Ni products from an acid mine drainage (AMD). As major contaminants, Fe and Al in the solution were sequentially precipitated and eliminated by elevating the pH. After that, a pre-concentrated slurry containing Mn, Co, Ni, and Zn was obtained by collecting the precipitates formed in the pH range of 6.50 to 10.00. The pre-concentrated slurry was redissolved for further purification. Sodium sulfide was added into the redissolved solution to precipitate Co, Ni, and Zn selectively while retaining Mn in the solution. Almost 100% of Co, Ni, and Zn but only around 15% of Mn were precipitated using a sulfur-to-metal molar ratio of 1 at pH 4.00. The sulfide precipitate was calcined and then completely dissolved. The critical elements existing in the dissolved solution were efficiently separated using a two-stage solvent extraction process. Ultimately, Co and Ni products with almost 94% and 100% purity were obtained by sulfide and alkaline precipitation, respectively. AMD also contains rare earth elements (REEs), which can be recovered through selective chemical precipitation. REE removal improved at pH 4.0 after converting ferrous to ferric ions with H2O2. Aluminum species in the solution hindered REE adsorption on ferric precipitates, and ferrous ions reduced REE adsorption on aluminum precipitates at lower pH, but at higher pH, REE removal increased due to ferrous ion precipitation. Various tests and analyses were conducted to understand the partitioning mechanisms of REE during the precipitation process of AMD. Sulfide precipitation is crucial to separate Mn from other elements, but the presence of contaminants like Fe and Al can affect sulfide precipitation efficiency. The effects of Al3+ iii and Fe2+ on the precipitation characteristics of four valuable metals, including Mn2+, Ni2+, Co2+, and Zn2+, were investigated by conducting solution chemistry calculations, sulfide precipitation tests, and mineralogy characterizations. It was found that the ability of the valuable metals to form sulfide precipitates followed an order of Zn2+ > Ni2+ > Co2+ > Mn2+. The sulfide precipitate of Zn2+ was the most stable and did not re-dissolve under the acidic condition (pH 4.00 ± 0.05). In addition, the sulfide precipitation of Zn2+ was barely affected by the contaminant metal ions. However, in the presence of Al3+, the precipitation recoveries of Mn2+, Ni2+, and Co2+ in a solution containing all the valuable metals were noticeably reduced due to simultaneous hydrolysis and competitive adsorption. The precipitation recoveries of Ni2+ and Co2+ in solutions containing individual valuable metals also reduced when Fe2+ was present, primarily due to competitive precipitation. However, the recovery of Mn2+ was enhanced due to the formation of ferrous sulfide precipitate, providing abundant active adsorption sites for Mn species. In the solution containing all the valuable metals, Fe2+ promoted the recoveries of the valuable metals due to the higher concentration of Na2S and the formation of ferrous sulfide precipitate.
  • Physics-informed Machine Learning with Uncertainty Quantification
    Daw, Arka (Virginia Tech, 2024-02-12)
    Physics Informed Machine Learning (PIML) has emerged as the forefront of research in scientific machine learning with the key motivation of systematically coupling machine learning (ML) methods with prior domain knowledge often available in the form of physics supervision. Uncertainty quantification (UQ) is an important goal in many scientific use-cases, where the obtaining reliable ML model predictions and accessing the potential risks associated with them is crucial. In this thesis, we propose novel methodologies in three key areas for improving uncertainty quantification for PIML. First, we propose to explicitly infuse the physics prior in the form of monotonicity constraints through architectural modifications in neural networks for quantifying uncertainty. Second, we demonstrate a more general framework for quantifying uncertainty with PIML that is compatible with generic forms of physics supervision such as PDEs and closed form equations. Lastly, we study the limitations of physics-based loss in the context of Physics-informed Neural Networks (PINNs), and develop an efficient sampling strategy to mitigate the failure modes.
  • Tensile-Strained Ge/III-V Heterostructures for Low-Power Nanoelectronic Devices
    Clavel, Michael Brian (Virginia Tech, 2024-02-12)
    The aggressive reduction of feature size in silicon (Si)-based complimentary metal-oxide-semiconductor (CMOS) technology has resulted in an exponential increase in computing power. Stemming from increases in device density and substantial progress in materials science and transistor design, the integrated circuit has seen continual performance improvements and simultaneous reductions in operating power (VDD). Nevertheless, existing Si-based metal-oxide-semiconductor field-effect transistors (MOSFETs) are rapidly approaching the physical limits of their scaling potential. New material innovations, such as binary group IV or ternary III-V compound semiconductors, and novel device architectures, such as the tunnel field-effect transistor (TFET), are projected to continue transistor miniaturization beyond the Si CMOS era. Unlike conventional MOSFET technology, TFETs operate on the band-to-band tunneling injection of carriers from source to channel, thereby resulting in steep switching characteristics. Furthermore, narrow bandgap semiconductors, such as germanium (Ge) and InxGa1-xAs, enhance the ON-state current and improve the switching behavior of TFET devices, thus making these materials attractive candidates for further study. Moreover, epitaxial growth of Ge on InxGa1-xAs results in tensile stress (ε) within the Ge thin-film, thereby giving device engineers the ability to tune its material properties (e.g., mobility, bandgap) via strain engineering and in so doing enhance device performance. For these reasons, this research systematically investigates the material, optical, electronic transport, and heterointerfacial properties of ε-Ge/InxGa1-xAs heterostructures grown on GaAs and Si substrates. Additionally, the influence of strain on MOS interfaces with Ge is examined, with specific application toward low-defect density ε-Ge MOS device design. Finally, vertical ε-Ge/InxGa1-xAs tunneling junctions are fabricated and characterized for the first time, demonstrating their viability for the continued development of next-generation low-power nanoelectronic devices utilizing the Ge/InxGa1-xAs material system.
  • Novel Multitemporal Synthetic Aperture Radar Interferometry Algorithms and Models Applied on Managed Aquifer Recharge and Fault Creep
    Lee, Jui-Chi (Virginia Tech, 2024-02-09)
    The launch of Sentinel-1A/B satellites in 2014 and 2016 marked a pivotal moment in Synthetic Aperture Radar (SAR) technology, ushering in a golden era for SAR. With a revisit time of 6–12 days, these satellites facilitated the acquisition of extensive stacks of high-resolution SAR images, enabling advanced time series analysis. However, processing these stacks posed challenges like interferometric phase degradation and tropospheric phase delay. This study introduces an advanced Small Baseline Subset (SBAS) algorithm that optimizes interferometric pairs, addressing systematic errors through dyadic downsampling and Delaunay Triangulation. A novel statistical framework is developed for elite pixel selection, considering distributed and permanent scatterers, and a tropospheric error correction method using smooth 2D splines effectively identifies and removes error components with fractal-like structures. Beyond geodetic technique advancements, the research explores geological phenomena, detecting five significant slow slip events (SSEs) along the Southern San Andreas Fault using multitemporal SAR interferometric time series from 2015-2021. These SSEs govern aseismic slip dynamics, manifesting as avalanche-like creep rate variations. The study further investigates Managed Aquifer Recharge (MAR) as a nature-engineering-based solution in the Santa Ana Basin. Analyzing surface deformation from 2004 to 2022 demonstrates MAR's effectiveness in curbing land subsidence within Orange County, CA. Additionally, MAR has the potential to stabilize nearby faults by inducing a negative Coulomb stress change. Projecting into the future, a suggested 2% annual increase in recharge volume through 2050 could mitigate land subsidence and reduce seismic hazards in coastal cities vulnerable to relative sea level rise. This integrated approach offers a comprehensive understanding of geological processes and proposes solutions to associated risks.
  • Colloidal Processing, Microstructural Evolution, and Anisotropic Properties of Textured Ultra-High Temperature Ceramics Prepared Using Weak Magnetic Fields
    Shiraishi, Juan Diego (Virginia Tech, 2024-02-09)
    The texturing of ultra-high temperature ceramics (UHTCs) using weak magnetic fields is studied and developed for the first time. Textured UHTCs were prepared by magnetically assisted slip casting (MASC) in weak magnetic field (B ~ 0.5 T). Analytical calculations describing the balance of torques acting on the suspended particles suggested that texture would form at such low magnetic fields. The calculations include a novel contribution of Stokes drag arising from the inhomogeneous velocity profile of the fluid during slip casting. Experimental proof-of-concept of the theoretical calculations was successfully demonstrated. Calculations of Lotgering orientation factor (LOF) based on the intensities of the (00l) family of peaks measures by XRD revealed strong c-axis crystalline texture in TiB2 (LOF = 0.88) and ZrB2 (LOF = 0.79) along the direction of the magnetic field. Less texture was achieved in HfB2 (LOF = 0.39). In all cases, the density of the textured materials was less than that of control untextured materials, indicating that texturing hindered the densification. The findings from this work confirm the potential for more cost-effective, simple, and flexible processes to develop crystalline texture in UHTCs and other advanced ceramics and give new insight into the mechanisms of magnetic alignment of UHTCs under low magnetic fields. The microstructural evolution during slip casting and pressureless sintering is investigated. The interplay between magnetic alignment and particle packing was investigated using XRD and SEM. During MASC, the suspended particles rotate into their aligned configuration. Particles that deposit at the bottom of the mold near the plaster of Paris substrate have their alignment slightly disrupted over a ~220 μm-thick region. The aligned suspended particles lock into an aligned configuration as they consolidate, leading to a uniform degree of texturing across the entire sample height of several millimeters upon full consolidation of the particle network. If the magnetic field is removed before the particles fully consolidate, the suspended particles re-randomize their orientation. Grain size measurements done using the ASTM E112 line counting method on SEM images revealed anisotropic microstructures in green and sintered textured ZrB2 materials. Smaller effective grain sizes were observed in the direction of c-axis texture than the directions perpendicular to the texture. Grain aspect ratios of 1.20 and 1.13 were observed in materials where the c-axis texture directions were parallel (PAR) and perpendicular (PERP) to the slip casting direction, respectively. Constraint of the preferred a-axis grain growth direction in the textured materials inhibited their densification compared to the untextured material. The PERP material with the preferred grain growth direction constrained along the casting direction had smaller average grain sizes than the PAR material which contained the preferred grain growth directions in the circular plane normal to the casting direction. Compression testing suggests a trend towards higher strength and stiffness in materials with higher density. Classical catastrophic brittle failure was observed in the untextured materials, but in the textured materials some samples exhibited a multiple failure mode. The PERP material tended to exhibit superior strength and stiffness to the PAR material in the classical brittle failure mode due to the orientation of the stiffer a-axis along the loading direction and smaller average grain size in the plane normal to the loading direction in the PERP condition. In the multiple failure mode, the PAR material tended to reach higher strength values after the initial failure and reach slightly higher strains before ultimate failure due to the orientation of the compliant c-axis along the loading direction and ability of the grains elongated in the plane normal to the loading direction to rearrange themselves after initial failure(s). Regardless of density or texture condition, all ZrB2 samples survived thermal shock resistance (TSR) testing. Samples were heated to 1500°C in air, held for 30 minutes, then quenched in room temperature air. After TSR testing, oxide layers formed on the surface of the materials. The specific mass gain and oxide layer thickness tended to increase with increasing porosity and were dramatically increased when open porosity was dominant as in the CTRL 1900 condition. After TSR testing, the compressive strength and strain at failure were both higher compared to the as-sintered materials. The increases in the average compressive strength were 20%, 76%, and 57% in the CTRL, PAR, and PERP conditions, respectively. The combination of the presence of the oxide layer shifting the onset of macroscale damage to higher strain values, the dissipation of load in the more porous region near the oxide layer, and the constraining effect of the oxide layer acting against the expansion of the material contributed to reinforcement of the samples after TSR testing. The CTRL material outperformed the textured materials on average in terms of strength and stiffness due to the higher density. The results suggest that reinforcement was more effective in the PAR condition than the PERP, which may be caused by the formation of a homogenous oxide layer on the PAR while the PERP formed an anisotropic layer. The work presented in this dissertation lays the foundation for affordable, energy efficient preparation of UHTCs and other ceramic materials. Equipment costs are reduced by 3 orders of magnitude, and the operating costs and energy consumption are greatly reduced. Facilitation of the preparation of textured materials opens the door to renewed investigations into their processing and performance. This work describes in detail for the first time the relationships between processing, microstructure, and properties of a textured UHTC part, providing a model for future research. Finally, the findings in this work can be used to guide process optimization, exploration of complex shapes and microstructures, and design of manufacturing schemes to create specialty textured parts for demanding structural and functional applications.
  • Fiendish Dreams - Reverse Engineering Modern Architecture
    Heinrich, Linda Kay (Virginia Tech, 2024-02-07)
    Winsor McCay drew delightful drawings about the dreams of a Welsh rarebit fiend, 'rare bits' inspired by an overindulgence in cheese. Dreams of the Rarebit Fiend was a Saturday cartoon that appeared in the New York Evening Telegram from 1904 to 1911, psychic twin to Little Nemo in Slumberland that appeared concurrently in the Sunday Funnies of the New York Herald from 1905-1911. 'Slumberland' was a Neo-classical fantasy that closely resembled the idealized White City of the Chicago World's Fair (1893), that inspired the architecture of Coney Island's Dreamland (1905-1911), which beckoned to McCay as he drew from his house just across Sheepshead Bay in Brooklyn. The capricious side of this Architecture emerged in McCay's cartoons. A self-taught illustrator, McCay began his career in Detroit working in dime museums, worlds of wonder—filled with monsters—dioramas and sideshow performers whose livelihood depended on their ability to amaze an audience. Just this sort of rare and gifted fellow, McCay parlayed his entertaining lampoonery of Slumberland into some of the world's first animations on vaudeville. As with the Rarebit Fiend, Little Nemo's dreams were brought on by overindulgence, in his case of too many donuts or Huckleberry Pie. But, this was merely a pretense for McCay's fantastical 'dream' mode of thinking, a potentially useful body of knowledge that was simultaneously explored by Sigmund Freud, Henri Bergson and Marcel Proust, who linked the mechanisms employed by the unconscious in dreaming to those at play in wit. Architectural drawing—seen through McCay's cartoons and early animations—has a kind of 'gastronomical' alchemy that inadvertently became a treatise on the architectural imagination. Fiend and Little Nemo affected the psychic mood of early modern Architecture—its 'childhood' in the milieu of White Cities—that was both added to and commented on by Winsor McCay's pen. His cartoons portray the hidden 'flavors' of the buildings springing up a century ago. This 'other'—surreal—aspect of the White Cities, seasoned with whirling iron Ferris wheels and Flip-Flop rides, newly invented elevators and electric lights—and even fun house mirrors that made buildings suddenly seem very tall—were the ingredients that caused the fiend and Nemo to wake up, which ultimately became the culinary school of modern Architecture. McCay's 'fiendish' depictions show us that the right blend of humor and awe is a recipe for happiness.
  • Foundations of Radio Frequency Transfer Learning
    Wong, Lauren Joy (Virginia Tech, 2024-02-06)
    The introduction of Machine Learning (ML) and Deep Learning (DL) techniques into modern radio communications system, a field known as Radio Frequency Machine Learning (RFML), has the potential to provide increased performance and flexibility when compared to traditional signal processing techniques and has broad utility in both the commercial and defense sectors. Existing RFML systems predominately utilize supervised learning solutions in which the training process is performed offline, before deployment, and the learned model remains fixed once deployed. The inflexibility of these systems means that, while they are appropriate for the conditions assumed during offline training, they show limited adaptability to changes in the propagation environment and transmitter/receiver hardware, leading to significant performance degradation. Given the fluidity of modern communication environments, this rigidness has limited the widespread adoption of RFML solutions to date. Transfer Learning (TL) is a means to mitigate such performance degradations by re-using prior knowledge learned from a source domain and task to improve performance on a "similar" target domain and task. However, the benefits of TL have yet to be fully demonstrated and integrated into RFML systems. This dissertation begins by clearly defining the problem space of RF TL through a domain-specific TL taxonomy for RFML that provides common language and terminology with concrete and Radio Frequency (RF)-specific example use- cases. Then, the impacts of the RF domain, characterized by the hardware and channel environment(s), and task, characterized by the application(s) being addressed, on performance are studied, and methods and metrics for predicting and quantifying RF TL performance are examined. In total, this work provides the foundational knowledge to more reliably use TL approaches in RF contexts and opens directions for future work that will improve the robustness and increase the deployability of RFML.