Doctoral Dissertations

Permanent URI for this collection

https://hdl.handle.net/10919/11041

Browse

Now showing 1 - 20 of 17541

US Foreign Policy Towards National Movements: Impact of Joint Combat Operations, Affective Trust, and Identity
Barwari, Delovan Fattah (Virginia Tech, 2025-02-14)
This study explores US foreign policy toward national movements (NMs), focusing on Kurdish groups across all parts of Kurdistan: Syria, Iraq, Iran, and Turkey. It investigates the central question of why the US views certain Kurdish NMs as strategic allies while labeling others as terrorists. The PKK and PYD—sister organizations sharing the same ideology and leader—serve as a prime example of this inconsistency: the PKK is designated a terrorist organization, while the PYD has emerged as a key US partner. Similarly, Iraq's Kurdish ruling parties were initially placed on the third-tier terrorist list, only to later become Washington's most reliable allies in Iraq. The study reveals that this discrepancy is mainly due to the impact of joint combat operations. Driven by US strategic interests, these operations strengthen ties with NMs partners. Positive joint operations, free of insider attacks, are instrumental in building rational trust that evolves into affective trust over time. This trust elevates them to in-group status, fostering a shared identity. The affective bonds forged during these combat experiences shape policy makers' perceptions, further reinforcing these relationships. Furthermore, diplomatic engagements in the post-combat phase complement this process, deepening trust and enabling the US and NMs to address challenges collaboratively while advancing broader strategic objectives.
Trilemma in Optimization for Time-critical Cyber-Physical Systems: Balancing Optimality, Generality, and Scalability
Wang, Sen (Virginia Tech, 2025-02-13)
The increasing complexity of time-critical Cyber-Physical Systems (CPS) presents significant challenges in designing optimization algorithms that balance generality, scalability, and performance. Traditional approaches often compromise one or more of these properties: general metaheuristic algorithms lack scalability and performance guarantees, while problem-specific methods sacrifice generality for improved efficiency or optimality. However, due to the NP-hard nature of many real-time scheduling and optimization problems, it is highly unlikely to design optimization algorithms that are simultaneously general, scalable, and optimal. Therefore, this dissertation addresses these challenges by developing novel optimization frameworks tailored for time-critical CPS and try to improve the trade-off among the three factors. The first contribution focuses on general and scalable optimization techniques, introducing frameworks such as NORTH, which operates with black-box schedulability constraints while achieving very good scalability and reasonably good performance. Additionally, another optimization framework targets at general robotic working environments by performing dynamic resource allocation. It demonstrates 20–50\% improvements in safety-performance metrics with low computational overhead. The second contribution advances domain-specific optimization techniques by relaxing the general requirements. For instance, flexible Logical Execution Time (LET) optimization achieves significant improvements in end-to-end latency, time disparity, and jitter by leveraging symbolic operations and efficient exploration of solution spaces. Similarly, a novel scheduling approach for DAG-based task models minimizes worst-case end-to-end latency and time disparity through 1-opt solutions with polynomial runtime complexity, achieving up to 40\% performance gains over existing methods. These contributions push the boundaries of generality, scalability, and optimality in real-time systems optimization, providing practical solutions to complex scheduling and resource allocation problems. The proposed frameworks are validated through extensive experimental studies, demonstrating their applicability and impact across a range of real-world scenarios.
Effects of ambient temperature on mechanisms of pathogen transmission in house finches (Haemorhous mexicanus)
Richards, Sara Teemer (Virginia Tech, 2025-02-13)
Ambient temperature is an important abiotic factor shaping the process of pathogen transmission because of its effects on hosts, pathogens, and interactions between them. However, most experimental studies demonstrating the effects of temperature on transmission remain correlative and often exclude endothermic taxa, which modify behavior and energy allocation strategies in colder environments in ways that could increase pathogen spread. Additionally, because many endotherms serve as important reservoirs for zoonotic diseases and are facing conservation threats due to disease, understanding how temperature influences transmission in these systems has downstream relevance to human and wildlife health. In this dissertation, I use three laboratory experiments to determine how temperature affects several mechanisms of transmission in a naturally occurring songbird-pathogen system. House finches (Haemorhous mexicanus) are small songbirds that rely on bird feeders to meet thermoregulatory demands during winter. However, interactions with other birds at the feeder and contact with contaminated feeder surfaces are important sources of transmission of the bacterial pathogen Mycoplasma gallisepticum (MG). These interactions likely contribute to the fall and winter outbreaks of mycoplasmal conjunctivitis, a disease characterized by severe conjunctival swelling and changes in behavior in house finches. In my first experiment, I simulated infection in house finches to determine how temperature (warm versus cold) affected contact-relevant sickness behaviors, and in turn, the potential for transmission. I found that ambient temperature had a complex effect on some but not all contact-relevant sickness behaviors in this system, which could have key implications for downstream pathogen spread. Next, I investigated how ambient temperatures influenced another mechanism of transmission, the viability and pathogenicity of MG harbored on bird feeder surfaces. I found that MG remained viable and pathogenic to birds significantly longer when incubated on feeder surfaces at colder versus warmer temperatures. In my final chapter, I determined how temperature influenced the pairwise-transmission of MG from an experimentally-inoculated "donor" bird to its susceptible "receiver" bird cagemate. Here I examined how temperature influenced host infectiousness and estimated exposure dose, as well as the behaviors of both sick and healthy birds. I found that donor birds in colder temperatures were slower to recover from infection, and thus remained infectious for longer, compared to donor birds in warmer temperatures. I also found that receiver birds had more contacts with bird feeders and higher estimated doses of MG in colder temperatures. Despite evidence suggesting that MG transmission could be more successful in colder versus warmer temperatures, overall transmission success did not differ by temperature treatment. My work highlights the complex and non-uniform effects of temperature on aspects of the MG transmission process and suggests ways that temperature could have major implications for seasonal disease dynamics in this system. More broadly, my dissertation provides a framework for testing how different abiotic factors could influence the spread of other directly-transmitted diseases, which will be needed now more than ever in the face of global climate change.
Educational Leadership Impact on Early Career Teacher Retention: Making Meaning of School Principal and Classroom Educator Perceptions
Riganti, Heather Victoria (Virginia Tech, 2025-02-07)
Teacher retention, specifically early career teacher retention, is a prominent issue facing educational leaders. This dissertation addresses the impact high early career teacher attrition has on student academic achievement, establishes the purpose of my qualitative research study, and clarifies the problem of study. Integrated into Chapter 1 is a conceptual framework that outlines leadership and non-leadership factors impacting early career teacher retention. The literature review in Chapter 2 examines current teacher attrition trends as well as practices and policies implemented to improve teacher retention. Literature included in the review is peer reviewed and published between 2012 and 2024. Articles produced from the search criteria were derived from the Virginia Tech remote library's Education Research Complete from EBSCO host, ERIC from EBSCOhost, Educators Reference Complete from Gale, and Teacher Reference Center from EBSCOhost in addition to articles referred to me by Virginia Tech faculty. The connection between leadership practices or behaviors and new teacher retention is supported by current literature. Teacher retention is discussed as a global and national issue as well as an issue facing educational leaders in the state of Virginia. Monetary and non-monetary costs of teacher attrition are discussed in addition to the monetary and non-monetary factors impacting teacher retention. School leadership and the connection to mentorship literature are analyzed to determine the connection between leadership behaviors and new teacher retention. Using the identified research questions in Chapter 1, I outlined my research methodology and framework for my qualitative research study in Chapter 3. The purpose of this study was to investigate the influence educational leadership has, specifically building level principals, on early career teacher retention decisions. The study specifically focused on the secondary level in a medium sized school division in southwest Virginia. I wanted to gain a better understanding of how principals at the secondary level positively contribute to the retention of early career teachers within their schools. New teacher, beginning teacher, novice teacher, and early career teacher are used interchangeably throughout this dissertation. Chapter 4 presents an analysis of individual semi-structured early career teacher interviews and principal interviews. Chapter 5 presents implications and meanings generated from this research study. Meanings made and implication from early career teacher interviews and principal interviews are presented in this study. This study has value in its potential to inform school policy makers, drive future leadership practices, or influence the practice of future school leaders.
Muscle Loading Treatments for Achilles Tendinopathy
Easley, Dylan Cole (Virginia Tech, 2025-02-07)
Tendinopathies are common, painful, and debilitating injuries that can be challenging to treat. Current treatment methods are limited to surgery, nonsteroidal anti-inflammatory drugs, dry needling, and injectable therapeutics, platelet rich plasma and corticosteroids. Unfortunately, these existing treatments display poor long-term outcomes and have an increased risk of reinjury. Additionally, the healing mechanism for injured tendons forms scar tissue which is characterized by disrupted extracellular matrix rather than complete injury resolution. These structural changes impact the mechanical properties of tendon, reducing their capacity to transfer and store energy, making them inferior to uninjured tendons. The reduced mechanical properties increase the risk of rupture, exacerbating this debilitating disease and decreasing quality of life. Physical therapy (eccentric loading) decreases the symptoms of tendinopathy and restores Achilles tendon functionality. However, the mechanism by which these mechanical stimulations induce healing is poorly understood. There is a clinically relevant motivation to better understand the healing cascade in response to eccentric exercises. We aim to identify and characterize the effects of eccentric rehabilitative muscle loading on the Achilles tendon and gastrocnemius muscle complex using our preclinical TGF-ß1-induced murine model of Achilles tendinopathy. To accomplish our objective, we tested three muscle loading magnitudes (50%, 75%, and 100% body weight), over three treatment durations (1, 2, and 4 weeks) to determine their effects on tendon healing. Age-matched injured/untreated and naïve groups accompanied each loading magnitude and duration period. The functional biomechanical properties, morphological adaptations, transcriptomic response, and muscle strength of the Achilles tendon were assessed. Injured/untreated tendons had a significantly increased cross-sectional area compared to naïve and all loading groups at 2 and 4 weeks. Maximum stress and elastic modulus of injured/untreated tendons were significantly lower compared to naïve and all loading groups after 4 weeks. Gastrocnemius muscle strength was maintained over time as loading magnitude increased. Force output was lower after 2 weeks at 100% body weight loading compared to the naïve group, then recovered to naïve levels after 4 weeks. Histological findings included increased cross-sectional area, matrix disorganization, and increased cellular density of injured/untreated tendons. The transcriptomic evaluation revealed several patterns of expression among exercised groups. Biological processes associated with exercised groups revealed genes responsible for inflammation, extracellular matrix organization, and cell to cell signaling. Overall, eccentric muscle loading improved tendon geometry and material properties compared to naïve levels and improved muscle strength over time. Morphological evaluation also showed improvements in cross-sectional area, and collagen orientation, and cell appearance after 2 and 4 weeks of eccentric loading. Similarly, the transcriptomic changes showed an effect from exercise and upregulation of genes essential for extracellular matrix organization, inflammatory regulation, and cell to cell signaling.
Personalized and Adaptive HVAC Control Strategies in Grid-Interactive Buildings
Meimand, Mostafa Ebrahimi (Virginia Tech, 2025-02-06)
Efficient control of HVAC (Heating, Ventilation, and Air Conditioning) systems is crucial for balancing demand and supply of energy in buildings, particularly during peak demand pe-riods. This dissertation aims to address three research gaps. First, previous research effortshave focused on decreasing energy consumption over peak time while considering comfort asa fixed range of temperatures or using generic indices for a population rather than focusingon individual thermal preferences. In response to this gap, a novel occupant-centric con-trol strategy is proposed to minimize energy costs while prioritizing personalized comfort.The proposed controller is tested in a simulation environment under different contextualconditions and in a real-world testbed. Second, another challenge of the existing HVACsystem controllers is finding the right balance between energy cost and occupant comfort inco-optimization formulations. The proposed balance should be adapted to different environ-ments. To address this challenge, an evolutionary Reinforcement Learning (RL) approachis introduced that enables the system to learn and adapt the trade-off coefficient betweenenergy and comfort optimization, enhancing the system's adaptability to different environ-mental and contextual conditions. Third, existing load flexibility models mainly considerspace-related factors and often overlook individual preferences. In the final phase, we shiftour focus from spaces to people and examine how current load flexibility models may affectindividual thermal comfort. Also, we devise a feature to predict load-shedding potentialbased on user properties. The performance of these three frameworks/models is assessedthrough a comprehensive uncertainty quantification analysis, taking into account the di-versity in occupants' preferences and the number of individuals present. Furthermore, theproposed approaches are compared with benchmark controllers from existing literature in asimulated environment. To validate their feasibility, a real-world experiment in an apart-ment unit as a practical test-bed is conducted. This research aims to improve the energyefficiency of HVAC systems, improve overall comfort experience, and evaluate the effect ofindividual comfort based on the current load flexibility models.
Robust and Data-Driven Uncertainty Quantification Methods as Real-Time Decision Support in Data-Driven Models
Algikar, Pooja Basavaraj (Virginia Tech, 2025-02-05)
The growing complexity and data in modern engineering and physical systems require robust frameworks for real-time decision-making. Data-driven models trained on observational data enable faster predictions but face key challenges—data corruption, bias, limited interpretability, and uncertainty misrepresentation—which can compromise their reliability. Propagating uncertainties from sources like model parameters and input features is crucial in data-driven models to ensure trustworthy predictions and informed decisions. Uncertainty quantification (UQ) methods are broadly categorized into surrogate-based models, which approximate simulators for speed and efficiency, and probabilistic approaches, such as Bayesian models and Gaussian processes, that inherently capture uncertainty into predictions. For real-time UQ, leveraging recent data instead of historical records enables more accurate and efficient uncertainty characterization, making it inherently data-driven. In dynamical analysis, the Koopman operator represents nonlinear system dynamics as linear systems by lifting state functions, enabling data-driven estimation through its applied form. By analyzing its spectral properties—eigenvalues, eigenfunctions, and modes—the Koopman operator reveals key insights into system dynamics and simplifies control design. However, inherent measurement uncertainty poses challenges for efficient estimation with dynamic mode and extended dynamic mode decomposition algorithms. This dissertation develops a statistical framework to propagate measurement uncertainties in the elements of the Koopman operator. This dissertation also develops robust estimation of model parameters, considering observational data, which is often corrupted, in Gaussian process settings. The proposed approaches adapt to evolving data and process agnostic— in which reliance on predefined source distributions is avoided.
Quilted Narratives: Patchworking Rural Appalachian Cultural Influence and Identities through the Storied Experiences of Women Educational Leaders
Rippey, Leanna Blake (Virginia Tech, 2025-02-04)
This qualitative study used narrative inquiry interviews of a purposeful sample of three rural Appalachian women educational leaders in the South Central Appalachian region to determine how their sociocultural backgrounds or Appalachian identity structures influenced their leadership practices in K-12 educational settings. Using the metaphor of quilting, my research study considered the storied narratives of rural Appalachian women educational leaders as patches for a quilt, taking each story and sewing them together to see what commonalities and complexities exist that guide leadership in educational settings. The research questions for this study were grounded in the participants' narratives and explored the following two concepts: 1. How might storied experiences or sociocultural influences impact the leadership and decision-making of rural Appalachian women educational leaders? 2. How does the intersection of Appalachian identities and gender influence rural Appalachian women educational leaders in their leadership practice and decision-making? Participants were asked a series of questions in a narrative interview session, followed by follow-up questions asked via email to further their narrative interview responses. Common themes included the importance of relationships and the community, their educational experiences, leadership aspirations, and the challenges these Appalachian women leaders encountered in pursuing their leadership roles. Evidence from the study supported that their sociocultural backgrounds and Appalachian identities influenced rural Appalachian women educational leaders in how they lead in their schools. Like threads running throughout the quilt, the findings suggest that the Appalachian women educational leaders interviewed for this study have common experiences from their sociocultural backgrounds that influence how they lead. The findings from this study provided a foundation to further this research into how sociocultural backgrounds and other identity structures are carried with educational leaders and created a need for reflection on these backgrounds to help shape leaders into effective practitioners in their settings.
Boundarywalkers: conceptualizing the dynamics of equitable science between Indigneous and Western knowledge
Johnson, Cheri Lynn (Virginia Tech, 2025-02-04)
Through interviews with Western educated scientists who also identify as Indigenous tribal members, this study seeks to understand how two knowledge systems, Indigenous Knowledge Systems (IKS) and Western Science Knowledge Systems (WSKS) can cogenerate knowledge. Interviewees for this study, as primary sources, contributed to the construction of a new concept, the boundarywalker framework, that conceptualizes the dynamics of equitable science between the knowledge systems through several key principles that promote inclusivity, mutual respect, and collaborative knowledge cogeneration. By bridging these distinct epistemic worlds, the boundarywalker framework facilitates equitable dialogue and challenges structural injustices that have historically marginalized Indigenous perspectives. This study addresses the challenges and possibilities of equitable knowledge cogeneration, where both knowledge systems contribute uniquely while retaining their distinct values. Through in-depth analysis of boundarywalker practices, the research highlights two foundational principles: epistemic pluralism, which fosters the coexistence and mutual enrichment of diverse worldviews, and epistemic revolution, which seeks to democratize the frameworks of knowledge production by centering marginalized voices and advancing reflexive justice. This involves a continuous re-evaluation of inclusivity, ensuring research methodologies and priorities are shaped by Indigenous perspectives. The boundarywalker framework offers a pathway toward an ethical, pluralistic, and reflexive science, where IKS and WSKS cogenerate knowledge as equal partners. Through epistemic pluralism and democratized framesetting, the boundarywalker framework envisions a science that honors diverse epistemologies, advances sustainable knowledge production, and strengthens the resilience of both scientific and Indigenous communities.
Using Alternative Data Visualization Formats to Impact Residents Energy Estimation of Household Appliances
James, Joseph Andrew (Virginia Tech, 2025-02-03)
Data visualization has the power to portray an informative message when designed with the end user in mind. Energy data visualizations must be tailored to the resident's energy, graphical, and data literacy level. A resident's energy, graphical, and data literacy level depicts their understanding and life experience with energy. Current utility companies standardize data visualization formats for all customers, regardless of their literacy level. My aim for this dissertation is to evaluate how data visualization mediums (2D chart types and virtual reality visual aids) aid residents when reading, working with, analyzing, and arguing energy consumption data of household appliance pairs. The data visualization chart types explored include the area, bar, and circular column charts. The visual aids displayed in the virtual environment explored include color coding, electricity flow, and the power meter. The energy data of the household appliances is embedded within the visual aids without displaying energy metrics. The household appliances include lighting (LED vs incandescent bulb), cooking (air fryer and stove), and heating appliances (heat pump and space heater). The participants included 32 graduate students from Virginia Tech engineering programs. Results from the study showed that some participants had a hard time interpreting axis unit metrics energy such as watts, watt*minutes, and kWhs in all three 2D chart types. If participants could not read and work with the units on charts, their ability to analyze and argue about the energy data was diminished quickly. In addition, when participants were interacting with the visual aids, researchers discovered that the power meter was the easiest to convey because it provided participants with a way to qualitatively and quantitatively answer the questions presented by the questionnaire. This dissertation provides insights for researchers, utility companies, and policymakers to move away from standardized data visualizations and utilize alternative visuals for reading, working with, analyzing, and arguing residential energy consumption data. Researchers can utilize the dissertation insights to explore other data visualization mediums that have the potential to convey energy insights. Utility companies can begin implementing these alternative data visualizations in pilot programs to test their effectiveness with the public. And lastly, policymakers can enforce utility companies to prioritize customer literacy levels when administering utility bills.
The Perceptions of Elementary Teachers on Induction Programs and Retention
Brock, Bernette Dywanda (Virginia Tech, 2025-02-03)
Unveiling Origins and Dynamics of Fecal Indicator Bacteria in an Urban Creek
Alvi, Dongmei (Virginia Tech, 2024-12-03)
Urban waterways are highly vulnerable to bacterial contamination, which presents significant risks to public health and water quality. Common methodologies typically measure the total concentration of fecal indicator bacteria (FIB) but are unable to address the complex sources of contamination contributing to the overall bacterial load. This study established chip-based digital polymerase chain reaction (cdPCR) techniques for microbial source tracking to unveil the origins of Escherichia coli (E. coli). Along with a simultaneous analysis of physicochemical water quality indicators, an assessment was conducted using host-associated genetic markers that indicate fecal sources from humans (HF183/BacR287), ruminants (Rum2Bac), dogs (DG3), and birds (GFD) in the lower portion of Rock Creek River (RCR) in the District of Columbia, United States. Stream samples were collected twice a month (n = 24) and after rain events (n = 6) from three sites along the RCR in the district area that feature a mix of highly developed urban areas and park surface regions. Approximately 50% of the stream samples (n = 96) were found to be impaired, exceeding the district's single sample maximum assessment level (410 MPN/100 ml) for E. coli. Herein, we adopted a multi-scale characterization of the relationship of cultural E. coli with host-associated markers, the regression with in-stream physiochemical constituents, the distinction between sampling sites, and the correlation with sizeable land cover categories. In Chapter 1, a comprehensive overview of MST methods is presented. This chapter summarizes the development of MST, categorizes common MST techniques into library-dependent versus library-independent and culture-dependent versus culture-independent groups, and provides a brief history of the advancements in molecular instrumentation used for culture-independent methods. In Chapter 2, consistently elevated E. coli levels were observed at all sites during wet weather, highlighting the substantial impact of storm runoff on water quality deterioration. Among the four molecular markers tested, HF183/BacR287, which indicates human-associated contamination, was particularly prevalent, with the highest frequency found in one of the tributaries. The second marker, derived from avian sources (GFD), showed a moderate to low frequency across the sites. Detection of the ruminant- and dog-specific markers was sporadic at all three sites. Correlation and regression analyses involving E. coli, molecular markers, and physicochemical constituents revealed significant statistical relationships. Notably, turbidity and flow were useful indicators for quickly assessing bacterial contamination. These findings emphasize the importance of reducing microbial contributions from runoff in watershed areas to urban streams during wet weather. The methods and findings of this study are expected to assist stormwater management and regulatory agencies in developing best management practices (BMP) to protect the water quality of urban streams. In Chapter 3, a strong association of E. coli with low-intensity developed land was established, but this association to forested areas at smaller spatial scales. The HF183/BacR287 marker exhibited similar trends, reinforcing its role as a reliable indicator of E. coli contamination sources. This study highlights the value of MST markers in identifying sources of microbial contamination. It provides important insights for managing water quality across various land cover types and changing weather conditions. In Chapter 4, the scalability of cdPCR to cell equivalents was investigated. By transforming scaled cdPCR DNA copies, the study revealed that 3,153 DNA copies per 100 mL of human-associated HF183BacR287 corresponded to the same regulatory threshold as cultured E. coli, enabling direct comparison between cdPCR and Colilert methods for contamination detection. This approach highlights the potential of cdPCR as a complementary tool to traditional methods in MST studies, offering a more detailed and efficient approach for water quality monitoring and management. In Chapter 5, a summary of the results is presented, and a perspective of future research direction is proposed.
AI Methods for Anomaly Detection in Cyber-Physical Systems: With Application to Water and Agriculture
Sikder, MD Nazmul Kabir (Virginia Tech, 2025-02-03)
In today's interconnected infrastructures, Cyber-Physical Systems (CPSs) play a critical role in domains including water distribution, agricultural production, and energy management. Modern infrastructures rely on a network of cyber-physical components—mechanical actuators, electrical sensors, and internet-connected devices—to supervise and manage operational processes. However, the increasing complexity and connectivity of these systems amplify their vulnerability to cyberattacks, necessitating robust cybersecurity measures and effective Outlier Detection (OD) methods. These methods are essential to prevent infrastructure failures, reduce environmental waste, and mitigate damages caused by malicious activities. Existing approaches often lack the integration of multiple operational metrics and context-driven techniques, hampering their effectiveness in real-world scenarios. In large CPSs—comprising hundreds or thousands of sensors, actuators, PLCs, IoT devices, and complex Control and Protection Switching Gear (CPSG)—the challenge of ensuring data quality, security, and reliability is costly. Cyberattacks frequently appear as outliers or anomalies in the data and are launched with "minimum perturbation," making their detection significantly challenging. This dissertation proposes a novel framework, multiple pipelines, and AI-based methods to develop context-driven, data-driven, and assurance-focused OD solutions. Emphasis is placed on water and agricultural systems, illustrating the proposed framework's effectiveness, particularly through enhanced decision-making, operational efficiency, and cybersecurity measures. A comprehensive survey of OD methods that employ Artificial Intelligence (AI) techniques establishes the foundational understanding of OD. This survey underscores that successful OD depends on domain knowledge, contextual factors, and assurance principles. Synthesizing these insights, the dissertation leverages synthetically generated SCADA data and GAN-produced poisoned data, as well as real-world SCADA data from Wastewater Treatment Plants (WWTPs), to identify outliers and address critical problems—such as forecasting tunnel wastewater overflows under extreme weather conditions—by applying Recurrent Neural Network (RNN)-based Deep Learning (DL) methods. Additionally, an AI-based decision support tool is introduced to detect anomalies in complex plant data and optimize operational set-points, thereby aiding Operation and Maintenance (OandM) in Water Distribution Systems (WDSs). Similarly, in Agricultural Production Systems (APSs), which traditionally rely on reactive policies and short-term solutions, integrating advanced AI-driven OD methods provides farmers with timely, data-informed decisions that account for contextual changes resulting from outlier events. Machine Learning (ML) and DL methods measure associations, correlations, and causations among global and domestic factors, aiding in the accurate prediction of agricultural production. This contextual awareness helps manage policy, optimize resource utilization, and support precision agriculture strategies. The main contributions of this dissertation include introducing a novel framework that integrates OD techniques with AI assurance and context-driven methodologies in CPSs; developing multiple pipelines and DL models that enhance anomaly detection, forecasting accuracy, and proactive decision support in WDSs and APSs; and demonstrating measurable improvements in cybersecurity, operational efficiency, and predictive capability using real-world and synthetic data. These efforts collectively foster more trustworthy and sustainable CPSs. Experimental results are recorded, evaluated, and discussed, revealing that these contributions bridge the gap between complex theoretical constructs and tangible real-world applications.
User-centered evaluations of multi-modal building interfaces
Kianpour rad, Simin (Virginia Tech, 2025-01-31)
In the evolving landscape of building systems and human-building interaction (HBI), the complexity of building interfaces has significantly increased, posing both challenges and opportunities for enhancing energy consumption, indoor environmental quality (IEQ), and building services. This dissertation, titled "User-centered Evaluation of Multi-modal Building Interfaces," delves into the realm of HBI by focusing on the user's experience and perception of multimodal building control interfaces, particularly the various visual modalities of Connected Thermostats (CTs). This body of work aims to support CTs' ongoing adoption, expansion, and performance through a user-centered perspective. The research is motivated by the observation that the design process in the current building industry often overlooks a human-centered approach, leading to a disconnection between occupants' needs and building interface design. This misalignment not only results in user dissatisfaction but also leads to a missed opportunity in leveraging smart building technologies to enhance building performance for achieving climate change mitigation goals. This research attempts to address the main identified gaps in the literature and AEC industry concerning 1) human interaction and perception of multimodal CT interfaces,2) the scarcity of knowledge in the field of human-computer-building interaction (HCBI) regarding the user study methods, 3) the exiting highly non-standard practices in the design of building interfaces. This research highlights 1) the necessity of a multimodal interaction approach, 2) robust mixed-methods User Experience (UX) summative evaluation studies, and 3) the need for standardization in HCBI. This body of work is grounded in the Technology Acceptance Model (TAM) and Human Information Processing (HIP) theories, aiming to foster the adoption of connected building controls with a special focus on usability by suggesting best practices in design and research. The methodology comprised three-step mixed-methods summative evaluation studies designed using a funnel approach to answer the general question: "How do users interact with connected thermostats, and how do these interactions inform our understanding of human-building interaction?": 1) The first and broadest study leveraged texting mining big data of user reviews to identify the general themes and patterns that affect the UX and acceptance of CTs. 2) The second study employed mixed-methods lab experiments to further focus on usability, being recognized as the most determining factor in the adoption of CTs in the first study. This study investigated human interaction with three of the most prevalent modalities of CTs: the Fixed Visual Display (FVD), the phone app, and the web portal. 3) The third study investigated human interaction with a specific visual aspect of UI of FVD and phone app modalities, the interface icons, with the goal of providing some data-driven guidelines for their standardization. Throughout the three studies, the dissertation employed and evaluated some novel and established HCI summative user evaluation methods, including a grounded theory approach for text mining and analyzing user-generated content, eye-tracking think-aloud protocol and contextual inquiry, A/B testing and NASA TLX and SUS surveys to evaluate UX, usability and mental workload. The dissertation outlined three discrete contributions: 1) It bridged some of the well-established UX research methods into HCBI and highlighted the potential of knowledge in the HCI field, 2) Provided guidance for human-centered design of multimodal building interfaces through identifying the main strengths, weaknesses, opportunities, and threats in UX of CTs, 3) Informed the standardization of UI of multimodal building interfaces.
Aerial Cadastral and Flood Assessment for Disaster Risk Management in Appalachia
Whitehurst, Daniel Scott (Virginia Tech, 2025-01-30)
As natural disasters have continued to become more prevalent in recent years, the need for effective disaster management efforts has become even more critical. Flooding is an extremely common natural disaster which can cause significant damage to homes and other property. Using low-cost drones, 3D cadastre models can be created and combined with flood models to quantify individual building risk before, during, and after flood events. As severe flooding devastated areas nearby to Virginia Tech, the need for accurate flood risk quantification became evident. In this work, we focused on the Appalachian area of the United States for flood modeling. The unique terrain of this area coupled with increasing major weather events has lead to devastating flooding in the area. In particular, we focused on an area in Southwest Virginia, Hurley, due to a devastating flood event in 2021 as well as its proximity to Virginia Tech. Digital Elevation Models from before the flood and available weather data are used to perform simulations of the flood event using HEC-RAS software. These were validated with measured water height values and found to be very accurate, with errors as low as 2 percent. After this, simulations are performed using the Digital Elevation Models created from drone imagery collected after the flood, and we found that a similar rainfall event on the new terrain would cause even worse flooding, with water depths between 29% and 105% higher. Simulations like these could be used to guide recovery efforts as well as aid response efforts for any future events. After this, a major flood event in 2022 shifted our focus to an area in Eastern Kentucky. The terrain in this area has been affected by significant surface coal mining, which became a focus due to the limited amount of research into the impacts of surface coal mining on flooding. Through the digitization of historical topographic maps, pre-mining terrain and land cover is compared to the current landscape with respect to runoff and flood potential. Additionally, multiple mine reclamation methods, including the regrowth of forest, grassland, or shrubland, were looked at to reduce the risk of major flooding in the future after mining has been completed. SWAT simulations showed a significant increase, as large as high as 55.8 percent, in surface runoff from the coal mining in the area. HEC-RAS simulations showed localized increases in flooding resulting from mine lands, with some areas seeing an increase of over 2 feet of water depth. Mine reclamation methods show the potential to reduce the amount of surface runoff, by as much 1 foot of water depth, although these ideal scenarios still do not reach pre-mined levels. While the impact which surface mining has had on the environment can not be fully reversed, significant improvements can be made to prevent future flooding in these areas. After these flood case studies, the water depth modeling is combined with high-resolution cadastre data to produce accurate flood risk assessments for the community and property level.
Structural Characterization and Material Property Evaluation in Polymer-Derived SiOC Ceramics and Ceramic Nanocomposites
Rau, Advaith Valliyur (Virginia Tech, 2025-01-30)
The field of advanced ceramics will experience significant growth in the upcoming decade to address increasing demands for multifunctional temperature- and corrosion-resistant materials for aerospace, energy, and electronics sectors. Polymer-derived ceramics (PDCs) and specifically polymer-derived silicon oxycarbide (SiOC) are a promising and attractive material class to accommodate the need for novel ceramics with tailorable compositions and material properties. SiOC is a unique member of the PDC family as the polymer precursor route is the predominant fabrication and synthesis method. As the composition and properties of the evolved SiOC ceramic can be tuned by polymer chemistry and choice of additives, a variety of multifunctional SiOC ceramics have been prepared with additional electric, magnetic, or structural characteristics. However, SiOC microstructures have been difficult to resolve as the amorphous matrix that shows nanoscale heterogeneity has not been rigorously characterized due to instrument and detector limitations. Therefore, understanding phase evolution in SiOC is critical for further development and commercial application of SiOC ceramics and ceramic composites. This work focuses on fabrication and characterization of novel SiOC ceramics and ceramic nanocomposites to examine phase formation and functional material properties imbued by reinforcement phases. In particular, SiOC ceramic 2D nanocomposites will be fabricated with montmorillonite (MMT – a naturally-occurring clay comprised of stacked 2D nanosheets) and Ti3C2Tx 'MXene' (a rapidly growing class of two-dimensional transition metal carbides/nitrides) to create nanostructured ceramic composites. Phase formation, porosity, and electrical conductivity will be analyzed to demonstrate attractive multifunctional capabilities of SiOC ceramics. In addition, thermodynamic modeling and advanced electron microscopy techniques will be utilized to better understand the locally-ordered amorphous SiOC matrix. The results and findings from this work will be among the first reported in the SiOC system and address limitations in the current state of knowledge.
Machine Learning-Driven Uncertainty Quantification and Parameter Analysis in Fire Risk Assessment for Nuclear Power Plants
Sahin, Elvan (Virginia Tech, 2025-01-27)
Fire poses a critical risk to the safe operation of nuclear power plants (NPPs), with electrical cabinet and liquid spill fires being among the most frequent and challenging scenarios to address. Traditional fire risk assessment models often lack precision due to complex physics and inherent uncertainties, especially in predicting the heat release rate (HRR) — a key parameter for assessing fire severity. This dissertation presents an innovative framework that integrates machine learning (ML) models, particularly neural networks and tree-based algorithms, with uncertainty quantification (UQ) techniques to enhance fire modeling and risk assessment in NPPs. The framework is applied to electrical enclosure cabinets and spill fires that represent about 50% of challenging fire scenarios in NPPs. By leveraging extensive experimental datasets, this study develops ML models that capture the influence of critical fire parameters on HRR, enabling more accurate predictions of fire behavior. Key features are evaluated to establish their influence on peak HRR. Advanced UQ tools, including Monte Carlo sampling and sensitivity analysis are applied to quantify uncertainties and identify parameters with the greatest impact on model output variability. The resulting ML-driven insights allow for a refined understanding of fire dynamics, guiding experimental planning and uncertainty reduction efforts. For electrical enclosure fires, the models highlight the importance of cable surface area, heat release rate per unit area of the cable, ignition source heat release rate, ventilation area, and cabinet volume in determining peak HRR. Sensitivity analysis revealed that HRRPUA is the most significant parameter. For spill fires, the models underscore the significance of substrate thermal conductivity and slope, ignition delay time, and fuel properties, showing that fuel amount and properties are key in fixed quantity spills, while fuel discharge rate and properties are most influential in continuous spills.
Physical Layer Data Integrity Attacks and Defenses in Cyber-Physical Systems
Mohammed, Abdullah Zubair (Virginia Tech, 2025-01-24)
Loss of data integrity in a safety-critical cyber-physical system (CPS), such as healthcare or intelligent transport, has a severe impact on its operation that can potentially lead to life-threatening consequences. This work investigates the vulnerability of CPS to physical-layer data integrity attacks and proposes countermeasures to enhance system resilience. Software-based cybersecurity approaches may not be efficient in mitigating threats aimed at the physical layer, leaving CPS particularly susceptible to manipulation through methods that exploit hardware vectors such as electromagnetic interference and data transmission medium. This work begins with a focus on using intentional electromagnetic interference (IEMI) to manipulate data and further explores other physical layer characteristics that can be exploited to conduct physical-layer attacks across various CPS environments. In the first phase of the research, the use of IEMI to induce controlled bit flips in widely used serial digital communication protocols is examined. In contrast to state-of-the-art IEMI attacks that use a narrow-band sinusoid as an attack signal, a complex, wideband, rectangular waveform is designed to improve the attack success rate from less than 50% to 75%. Further, the vulnerabilities of printed circuit board (PCB) traces to IEMI in highly safety-critical applications, such as electric vehicle (EV) charging, is addressed. On PCBs, IEMI attacks exploit the signal-carrying traces, that act as unintentional antennas under an adversarial electromagnetic field. Experiments demonstrated that such attacks are more challenging due to the PCB's structure but are still feasible with sufficient attacker power. A suite of passive countermeasures is evaluated, including differential signaling, via-fencing, and optical fiber interconnects, along with a novel multiplexer-based defense that dynamically modifies signal paths to evade detection. Each countermeasure is extensively evaluated and ranked based on its effectiveness, and adaptive attack strategies are analyzed to address potential future threats. In the IoT domain, this work presented a preliminary investigation on a novel "wireless spiking" technique on smart locks, that enables attackers to bypass standard security measures and unlock/lock with no physical contact. Using IEMI, the control circuitry is manipulated to unlock devices remotely. The methodology, involving hardware reverse engineering and attack point identification, is presented, which applies to other IoT devices in smart home environments. In the field of automotive cybersecurity, bit manipulation attacks targeting the Controller Area Network (CAN) bus are investigated. By exploiting its transmission line nature, these attacks challenge the fundamental assumptions of the CAN's physical layer and are capable of inducing bidirectional bit flips, from recessive to dominant (R→D) and significantly difficult dominant to recessive (D→R). The flips are further made undetectable to CAN's standard error-checking mechanisms. These attacks are simulated and validated in both lab and real-world vehicle environments. Finally, a defense mechanism for vehicle identification security in intelligent transportation systems using device fingerprinting is proposed. This approach utilizes inductive loop detectors (ILD) to capture unique electromagnetic signatures of vehicles, achieving up to 93% accuracy in identifying their make, model, and year. The ILD-based technique secures access control in automated systems and provides a cost-effective, drop-in solution for existing infrastructure, mitigating risks such as unauthorized vehicle impersonation and charging station exploitation. This work establishes a systematic framework for understanding, detecting, and defending against physical-layer data integrity attacks in CPS. Through the development of novel attack vectors and robust countermeasures, this research enhances the field of CPS security, emphasizing the need for comprehensive defenses that extend beyond conventional software-based approaches.
Multiscale Peridynamics Analysis of Nanocomposites and Energetic Materials Using Nonlocal and Local Interface Models
Genckal, Neslihan (Virginia Tech, 2025-01-24)
Interface modeling is a critical aspect in any multi-material system modeling. Even a small change in the interface model may lead to significant changes in material behavior of the microscale, and these changes may transfer up to higher scales influencing the strain and stress fields, and damaging behavior in the macroscale material. This work focuses on the effects of different interface models in nanocomposites composed of carbon nanotubes in polymer matrix materials and their applications as nanocomposite binders in energetic materials. These material systems include materials that span multiple scales from nano to macroscale, and thus require a detailed multiscale analysis. A hierarchical multiscale framework is employed here, where the effective material properties from subscales are obtained by solving the subscale boundary value problem. The information obtained from the subscale simulations are transferred up to higher scales to be used as input properties. A nonlocal continuum mechanics framework known as peridynamics is used to perform the computational simulations. Peridynamics uses integro-differential equations for conservation laws instead of partial differential equations as in the classical continuum mechanics. This makes it possible for peridynamics to inherently account for nonlocal effects such as damage initiation, crack growth, and crack branching without any modifications such as element deletion, adaptive mesh refinement, using enrichment functions and so on, which are commonly used in other numerical methods. Peridynamics is a particle-based method where the particles are allowed to interact with other particles within their horizon which serves as a cut-off distance for forming particle-to-particle bonds and therefore defines the extent of nonlocality. Peridynamics has different formulations regarding the bond interactions. A bond-based peridynamics framework is used here. A verified and validated in-house code is used for the simulations. The simulations for the carbon nanotube and nanofiber-based nanocomposites, and for nanocomposite bonded energetic materials start from the microscale and range up to the macroscale. For only the carbon nanotube-polymer nanocomposites, the interfaces include the CNT-polymer interfaces. For the energetic materials, the interfaces consider the CNT-polymer interfaces in the microscale and the grain-nanocomposite binder interfaces in the mesoscale. Peridynamics, being a nonlocal continuum mechanics method, by default will have nonlocal interfaces. The material systems investigated in this work first use different nonlocal interfaces in peridynamics which consider the bond between two particles at the interface to be connected in series or in parallel. The nonlocal interface model in peridynamics makes it challenging to control the interface properties and leads to fuzzy interfaces, i.e. interfaces of finite thickness. In this work, a local cohesive interface model is implemented in the peridynamics framework. Cohesive zones were originally used for modeling the growth of cracks by introducing cohesive forces that hold the crack surfaces together, thereby removing the stress singularity problem in linear elastic fracture mechanics. The idea of cohesive zones are applied to peridynamics interfaces, which introduces locality into the nonlocal framework. This interface model does not only remove the nonlocality at the peridynamics interfaces, but it leads to a higher fidelity interface model that is controllable by the user. The differences between the nonlocal and local interfaces are studied in detail in different scales and for different material systems. Implementing a local model into a nonlocal framework brings some challenges, namely obtaining and calibrating the cohesive interface properties for the materials used, the numerical problems with material interpenetration in extreme compression, and very small time steps that are required to resolve the material response. Some remedies are proposed for the problems encountered. The cohesive zone model used in this work can have different functional forms in normal and tangential direction to reflect differences in opening mode and frictional sliding behaviors.
Discontinuous Galerkin Studies of Collisional Dynamics in Continuum-Kinetic Plasma
Rodman, John Morgan (Virginia Tech, 2025-01-24)
Numerical investigations of collisional physics have historically been impeded by the issue of computational expense. While the continuum-kinetic Vlasov-Maxwell-Fokker-Planck system is well-established in theory and has been used as the basis for many approximate fluid equations, simulations utilizing the distribution function are relatively uncommon, due primarily to the high dimensionality of the problem. However, advances in numerical methods are steadily making these models more accessible. In this work, we utilize the Gkeyll framework, which applies a novel, highly efficient discontinuous Galerkin (DG) finite element method to the Vlasov-Maxwell-Fokker-Planck system. We first investigate the Rayleigh-Taylor (RT) instability in a neutral gas in regimes of finite collisionality which are inaccessible to the fluid codes that are traditionally applied to this instability. Utilizing a spatially constant, finite collision frequency, we demonstrate the ability of the Vlasov-Boltzmann model to approach the fluid result at high collision frequency while also accessing a regime of intermediate collisionality in which the RT instability deviates greatly from classic fluid behavior. We then extend upon this finding by choosing a collision frequency that varies spatially, resulting in new dynamics with asymmetric diffusion affecting the development of the RT instability. Having demonstrated the utility of collisional kinetic modeling even in the simple case of a neutral gas with a basic collision operator, we transition to development and implementation of a fully-conservative, recovery-based DG algorithm for the full nonlinear Rosenbluth/Fokker-Planck collision operator (FPO). Details of the novel recovery scheme for the cross-derivatives and conservation enforcement are presented, and we show that the scheme converges and exhibits stability criteria as expected. Finally, the FPO is applied to test cases that demonstrate the importance of accurate handling of the velocity-dependent collision frequency as compared to an approximate model.

Browse

Recent Submissions