19304 research outputs found
Sort by
Catalytic Conversion Scalability for Industrial Carbon Management
No full textIn the field of automation, analytical instruments represent a specific subset. These instruments, including gas chromatographs, utilize catalysts to instigate the reaction of carbon monoxide (CO) and carbon dioxide (CO2) with hydrogen, thereby producing methane (CH4). This process is integral to performing concentration measurements using sensitive detectors such as Flame Ionization Detectors and Flame Photometric Detectors. Nickel alumina, a catalyst widely used in the industry, facilitates this conversion reaction at temperatures around 300-500 degrees Celsius. Raw natural gas, which contains 0.5% Hydrogen Sulfide (H2S) and 1.5% CO2, can be processed to separate H2S through a catalytic reaction using Titanium Dioxide (TiO2) modified Molybdenum Carbide (MoxC) microwave catalyst. This process generates hydrogen, which can be used for the production of methane from these gases. The resulting methane can then be transported to processing units for liquefaction. Simultaneously, the emitted CO2 can be captured at the emission sources in plants, converted, and recycled back into methane.Civil and Environmental Engineering, Department ofHonors Colleg
Synergies of Mean-field Games in Machine Learning and 6G Communication Networks
No full textMean-Field Games (MFG) are mathematical models used to analyze decision-making for large populations of agents. Initially developed in statistical physics, MFGs were later adopted in economics and mathematics. It considers scenarios where each agent's strategy has negligible impacts on the entire population. Instead, each agent interacts with an "average" effect of the whole population, termed the ``mean field." Compared with traditional game models that need to study interactions between every two agents, MFG allows for the study of individual behaviors in relation to the collective dynamics of the entire population, reducing the system complexity significantly. A reference agent makes its optimal strategy based on the Hamilton-Jacobi-Bellman (HJB) equation, and the ``mean-field" evolves based on the Fokker-Planck-Kolmogorov (FPK) equation. However, when the spatial dimension of agents increases, the computational complexity of solving the PDE systems grows exponentially due to the curse-of-dimensionality. This dissertation delves into the intersection of mean-field games (MFG) of Machine Learning (ML) and 6G Communication Networks. Our research encompasses three primary contributions. Firstly, we delve into computational solutions for MFG problems, emphasizing a G-prox Proximal-Dual-Hybrid-Gradient (PDHG) algorithm tailored for low-dimensional MFG models. Secondly, we expand the scope of MFGs to embrace higher dimensions, introducing a Generative Adversarial Network-driven solution that is capable of solving MFG problems in up to 1,000 dimensions. Our final contribution entails an exploration of MFG applications in ML and 6G Communication Networks. Firstly, within vehicle-centric Mobile Crowd Sensing (MCS) networks, MFGs efficiently modeled both cooperative and competitive behaviors, streamlining path planning and task selection while significantly reducing computational complexity. In the context of vehicular-based Multi-access Edge Computing (MEC) networks, we harnessed MFGs to introduce a rapid data offloading mechanism, achieving marked reductions in End-to-End (E2E) latency and outperforming the benchmark. Then, in the realm of Machine Learning, our MFG-inspired data augmentation strategy, adaptable to both low and high-dimensional data, yielded a notable improvement in test accuracy, setting a new benchmark for performance. Lastly, we investigate the design of incentive mechanisms for Federated Learning in satellite communications, leveraging Mean Field Evolutionary Game (MFEG) to ensure fair and efficient participation of multiple local clients
The Correlation Between Lynchings and State Sanctioned Murder
No full textLynchings in America emerged as a manifestation of fear and White supremacy, evolving into a mainstream symbol of justice, particularly in the South, where political campaigns and social norms revered such acts as lawful punishment. This racial violence peaked in the 1890s, especially in Southern states, where a significant majority of victims were Black, reinforcing a culture of intimidation and social control. As lynchings declined in the early 20th century, state-sanctioned executions surged, reflecting persistent racial biases within the justice system. Legislative efforts to address this violence, such as the Dyer Bill and subsequent anti-lynching proposals, faced significant opposition from Southern Democrats, highlighting a systemic resistance to federal intervention in local racial issues. Despite a brief period of anti-lynching advocacy in state legislatures, the legal recognition of lynching as a crime did not materialize until 1928 in Virginia. The 1930s saw a paradoxical rise in executions alongside the decline of lynchings, as the idea of the death penalty as a necessary deterrent gained traction. Attempts to pass comprehensive anti-lynching legislation stalled due to Southern filibusters, and significant federal action remained elusive for decades. Today, our justice system still shows shocking parallels between post reconstruction era lynchings and the use of state sanctioned murder as punishment.Honors CollegeHistory, Department o
Lithium-ion battery state of charge monitoring using low frequency stress waves and machine learning methods
No full textLithium-ion batteries (LIB) are widely used in modern society, from portable electronics to electric vehicles and large stationary energy storage systems, and they require appropriate management of operating conditions to maintain safety, ensure reliability, and achieve long lifespan. This requires monitoring battery state of charge (SoC) to prevent overcharge or over-discharge, which drastically reduce immediate and long-term reliability and safety. Accurate SoC measurements (within correct charge limits) are also critical to ensuring the reliability of LIB-powered systems. Although SoC cannot be measured directly, estimation methods based on voltage and current measurements exist, however their accuracy suffers when batteries are in use and under dynamic loads. Deficiencies of existing estimation systems have caused injuries, deaths, property damage, and electronic waste. Recently new estimation methods based on battery stress wave response have emerged, and while high frequency ultrasound can produce predictions with high accuracy, equipment costs and complexity favor low-frequency stress wave (LFSW) approaches. To compensate for the significant complexity of material property shifts, their relationship with electrochemical battery dynamics, and the interpretation of mixed response signals, recent research applies machine learning (ML) to enhance estimation accuracy. Three LFSW with ML approaches are evaluated, by subjecting an LIB to 54kHz Gaussian-modulated sine pulse excitations and collecting stress wave response signals while charging/discharging the LIB at 0.25C rate with a battery tester to control SoC. The Mel-Frequency Cepstral Coefficients (MFCC) from 30-160kHz were extracted from the response signals and used to train Support Vector Machine (SVM) and Back-Propagation Neural Network (BPNN) models, and compared with Multi-Rocket model performance for SoC prediction. After tuning, SVM and BPNN models predicted SoC well, with RMS errors of 11%, 8% respectively. However, the Multi-Rocket model excelled, with a prediction error of only 0.72%. Although MFCC-derived feature datasets are usable for SoC classification, the MFCC feature extraction method requires additional optimization and/or augmentation with other features for high-precision SoC regression. Comparing these shallow- and deep-learning model performances with Multi-Rocket suggests that enhancing the accuracy of LFSW with ML SoC predictions will require further enhancing feature extraction and selection processes, as well as a larger training dataset.Honors CollegeMechanical and Aerospace Engineering, Department o
Watching the clock: Reducing emergency department wait times with point-of-care testing
No full textResearch shows that overcrowding and prolonged wait times in the Emergency Department has led to a rising trend in patients leaving the ED without being seen by a physician (Abuguyan et al., 2025). Studies exploring this topic have shown that ED wait times were prolonged the most during laboratory testing, radiological investigations, and physician reassessment (Jayakumar et al., 2024). Considering the investigation and reassessment stage is the largest contributor to lengthy patient wait times, research that investigates the effectiveness of interventions such as point-of-care testing could be helpful in finding solutions that successfully reduce patient wait times in the Emergency Department (Kayuni Mtambo et al., 2024).Nursing, Andy and Barbara Gessner College o
Designing Cyclometalated Iridium(III) and Platinum(II) Complexes for Efficient Red to Near-Infrared Phosphors
No full textPhosphorescent transition metal complexes, including iridium and platinum, have gained attention due to their unique photophysical properties, which could be used for various applications such as optoelectronic technology, photocatalysis, and phosphorescent probes in biological systems. To be applied for these applications, careful control, and optimization of photophysical and electrochemical properties of these complexes are needed, which motivates studies to understand structure-property relationships in order to design efficient molecular phosphors. This dissertation is mainly focused on developing design strategies for Iridium(III) and platinum(II) complexes, focusing on the effect of electron-rich ancillary ligands on improving efficiency. Chapter 1 introduces a general photophysical background of transition metal complexes, especially cyclometalated iridium(III) and platinum(II) complexes. In Chapters 2 and 3, new designs of deep red to NIR bis-cyclometalated iridium(III) phosphors are presented with their electrochemical and photophysical properties. In Chapters 4 and 5, new designs of red emitting cyclometalated platinum(II) complexes are presented with their photophysical properties
Interpreting Nanoindentation Using Machine Learning and Modeling Capillary Pressure in Shale
No full textAs an organic-rich sedimentary formation, shale has a fine-grained structure with complex mineralogy and ultra-narrow pores. Its nano-size pores change the physical properties and fluid behavior of the formation. This dissertation addresses three major challenges characterizing shale: fracture toughness, capillary pressure, and the Young–Laplace equation (YL). It proposes a machine-learning approach and a conceptual model for analyzing images and determining the created fracture length from nanoindentation. The imaged fractures of nanoindentations form complex patterns. This dissertation analyzes over 1,500 images to characterize fractures based on color intensity using K-means clustering, showing that fracture creation is plausible when the applied load varies between 400 mN and 700 mN. The determined fracture toughness was found to be between 0.5 MPa.m0.5 and 0.7 MPa.m0.5, which is validated against the energy method for the studied shale. This dissertation also addresses the challenge of capturing capillary pressure in shale using the Brooks–Corey and Van Genuchten models. It analyzes US shale mercury injection capillary pressure measurements (MICPs) and demonstrates how the conventional models fail to capture the entry pressure and its trend. It then proposes an empirical double-log model to overcome the limitations of existing models. The proposed model honors non-zero entry pressure and accurately captures the trend, providing a simple yet significant advancement in the field. Moreover, this dissertation underscores the crucial effects of size-dependent (actual) contact angle and interfacial tension on interpreting capillary pressure measurements in shale using YL. It applies the actual properties to the MICPs of Bakken and Eagle Ford shale samples. It also shows that the actual properties are particularly significant when the throat radius is less than 10 nm. The conventional approach with fixed properties overestimates the pore–throat size from 5% to 18%, decreasing the interpreted throat radius from 10 nm to 2.5 nm. These findings highlight the necessity of considering actual properties in future studies
The Role of Emergent Vegetation in Wetland Gas Movement and Hydrodynamic Patterns
No full textWetland ecosystems are dynamic environments where vegetation, sediment, and hydrodynamic forces interact to shape critical ecological processes. However, climate change-induced factors, such as higher wind conditions, threaten to disrupt these delicate systems. This study investigates how different vegetation scenarios influence water flow and marsh gas transport under controlled conditions. Using a custom-built acrylic water tank, we simulated three vegetation scenarios: no vegetation, short grass (modeled with zip ties), and cattails. A red dye tracer was introduced to represent marsh gases, with wind effects simulated using a fan. MATLAB image analysis was employed to track dye intensity, diffusion, and spread, providing insights into flow patterns and gas dispersion.Mechanical and Aerospace Engineering, Department ofHonors Colleg
Improving pain management equity in critically ill Black adults through culturally sensitive assessment and care protocols
No full textBlack patients in critical care settings are less likely to receive consistent pain assessments and are often undertreated for pain in comparison to White patients (Jetley & Zhang, 2024). This disparity is caused by implicit bias and a lack of cultural understanding in pain evaluation (Wang & Jacobs, 2023). Research shows that using standardized tools as well as cultural competence training can help nurses recognize and address and close these gaps (Goree & Jackson, 2022). We recognize other ethnicities within critical care also experience disparities in pain management, but are focusing our research on Black patients.Nursing, Andy and Barbara Gessner College o
On High Energy All-Solid Flexible Lithium-Ion Polymer Batteries
No full textFrom a seemingly distant research on lithium solubility to the Nobel prize, 60 years in the making, lithium-ion battery has earned its well-deserving position in our daily lives. Its everchanging influence and critical integration has marked a tremendous milestone in the evolution of energy. From the rise and fall of lithium metal technology and the birth of lithium intercalation materials to the reincarnation of lithium metal as the “holy grail” electrode, emerging technologies are continuously developed to meet the demand for a smaller, denser, and power-packed energy storage device to serve a plethora of applications from portable gadgets to electric vehicles. The inevitable risk of flammability from the long-developed organic electrolytes urgently calls for replacement, where polymer electrolytes show promising potentials. Dedicated in this dissertation is a new design of lithium-ion polymer battery that can withstand rapid bending and other physical impacts such as nail penetration, folding, and cutting. The enhanced polymer electrolyte offered great stability and improved ionic conductivity with the expedited solvent-free fabrication procedure. This version of lithium-ion polymer battery was further improved with a simple tweak in recipe by the addition of fluoroethylene carbonate (FEC). An already safety-enhanced battery received a significant capacity boost thanks to the stable solid electrolyte interphase (SEI) formed by FEC. The achieved all-solid lithium-ion polymer battery was found to reach maximal performance with 10wt% FEC added due to the optimal (de)lithiation across polymer-graphite interface where desolvation of lithium-ions was key. Transitioning into practical application integration requires mechanically deformable batteries where stretchable and flexible versions are crucial. A sliding battery design, where the electrodes slide and the solid polymer electrolyte stretches, was developed with increased capacity upon stretching. In another related embodiment, a thin-film polymer battery was subjected to bending where its bending-induced voltage profile during charge and discharge was found to strongly correlate with the charge-transfer resistance at the electrode-electrolyte interfaces. This finding is a critical step toward the bigger picture of the effects of mechanical bending, which eventually will prove useful for designing a mechanically stable version of flexible/stretchable lithium-ion polymer batteries