College of Engineering and Computer Science

Living organisms synthesize and assemble complex bioinorganic composites with enhanced structure and properties to fulfill needs such as structural support and enhanced mechanical function. With the advent of advanced materials characterization techniques, these biomineral systems can be explored with high resolution to glean information on their composition, ultrastructure, assembly, and biomechanics. In this work, the endoskeletal features of two marine organisms are explored.
Acantharia are geographically widespread marine planktonic single-celled organisms. Their star-shaped SrSO4 endoskeleton consists of spicules emanating from a central junction, arranged to satisfy crystallochemical and spatial requirements of their orthorhombic crystal lattice. In this work, synchrotron X-ray nanotomography and deep-learning guided image segmentation methods were used to characterize the endoskeleton of 5 types of Acantharia and to extrapolate their growth mechanism. The results highlight the diverse morphology of the spicules and spicular junctions that Acantharia achieve whilst maintaining overall spatial arrangement. Fine structural features, such as interspicular interstices thought to play a role in the robustness of the overall endoskeleton, were resolved.

Flooding disasters pose a significant threat worldwide, with 2022 seeing them as the most common type of disaster. In the U.S. alone, four flooding disasters in 2023 cost more than $9.2 billion. Coastal urban areas face increasing threats from flooding disasters due to rising sea levels, changing precipitation patterns, and intensifying extreme weather events. This study focuses on Central Beach, Fort Lauderdale; the area's unique geographical, environmental, historical, and socio-economic characteristics make it a prime candidate for this analysis. The research objective is to comprehensively examine the factors contributing to water-related vulnerabilities of developed properties in Central Beach and assess localized impacts using regional models. The methodology involves developing probabilistic flood maps using GIS tools and the Cascade 2001 routing model. The flood scenarios consider groundwater table rise, extreme rainfall, high tides, storm surge, and sea level rise. Results indicate significant inundation risks, particularly for commercial and mobility infrastructure, under storm surge and sea level rise scenarios. The analysis highlights the importance of targeted mitigation efforts to protect these areas and reinforce resilience against future flooding events. The findings contribute valuable insights for policymakers, urban planners, and stakeholders, emphasizing the need for comprehensive strategies to mitigate flood risks in coastal urban areas.

El Niño Southern Oscillation (ENSO) occurrences have a well-established impact on regional hydroclimatic variability and alterations in crucial climatic factors such as temperature and precipitation. The impact of ENSO on temperature extremes can cause fluctuations in energy consumption, leading to the need for energy utilities to implement more effective management measures. This study aims to evaluate the potential impacts of El Niño Southern Oscillation (ENSO) events on local temperature patterns & extremes and residential energy usage in South Florida. The region of focus consists of three Counties: Miami-Dade, Broward, and Palm Beach. The impact of ENSO occurrences on temperature is assessed by analyzing long-term monthly average, minimum, and maximum temperature data from numerous weather stations in these counties, spanning from 1961 to 2018. The study analyzes variations of monthly electricity usage data acquired from a local power utility company (e.g., Florida Power & Light) and temperature data from 2001 to 2018. Temporal frames that align with the three phases of ENSO (namely warm, cool, and neutral) are employed to assess variations in temperature and energy consumption. Nonparametric hypothesis tests are employed to validate statistically significant variations in temperature and residential energy consumption across the stages of ENSO. This study aims to analyze the potential regional and temporal impacts of ENSO episodes on temperature and residential energy consumption in South Florida. Initial findings indicate that the non-uniform distribution of temperature, affected by El Niño and La Niña occurrences, impacts the amount of energy consumed by households in South Florida.

The aim of this dissertation is to achieve a thorough understanding and develop an algorithmic framework for a crucial aspect of autonomous and artificial intelligence (AI) systems: Data Analysis. In the current era of AI and machine learning (ML), ”data” holds paramount importance. For effective learning tasks, it is essential to ensure that the training dataset is accurate and comprehensive. Additionally, during system operation, it is vital to identify and address faulty data to prevent potentially catastrophic system failures. Our research in data analysis focuses on creating new mathematical theories and algorithms for outlier-resistant matrix decomposition using L1-norm principal component analysis (PCA). L1-norm PCA has demonstrated robustness against irregular data points and will be pivotal for future AI learning and autonomous system operations.
This dissertation presents a comprehensive exploration of L1-norm techniques and their diverse applications. A summary of our contributions in this manuscript follows: Chapter 1 establishes the foundational mathematical notation and linear algebra concepts critical for the subsequent discussions, along with a review of the complexities of the current state-of-the-art in L1-norm matrix decomposition algorithms. In Chapter 2, we address the L1-norm error decomposition problem by introducing a novel method called ”Individual L1-norm-error Principal Component Computation by 3-layer Perceptron” (Perceptron L1 error). Extensive studies demonstrate the efficiency of this greedy L1-norm PC calculator.

The atmospheric concentration of CO2 increased from 320 to 425 parts per million by volume (ppmv; 0.0425 vol.%) between 1960 and 2024. Sample CO2 reduction strategies include shifting to renewable energy sources and employing CO2 capture. CO2 capture from the air (also known as direct air capture; DAC) has recently received increased attention. CO2 has the potential to act as an asphyxiant at high concentrations, particularly in enclosed environments (e.g., spacecraft, submarines), requiring air revitalization to remove CO2. Hence, the U.S. Occupational Safety and Health Administration determined a permissible exposure limit of 5,000 ppmv CO2 (0.5 vol.%) throughout an 8-hour work shift. Considering the trace levels of CO2 and the presence of humidity in DAC and air revitalization applications, similar materials can be developed for implementation in both cases. CO2 capture involving amine-functionalized silica materials (“aminosilicas”) can achieve high CO2 uptakes at low concentrations due to high selectivity. Additionally, moisture in CO2-laden gases enhances the CO2 uptake and stability of aminosilicas. Therefore, this research investigated the potential of aminosilicas for removing CO2 from dilute streams, including DAC and air revitalization applications. Aminosilicas were produced using mesoporous silica supports with different particle sizes that were modified with tetraethylenepentamine (TEPA) or branched polyethylenimine (PEI) with different molecular weights (600, 1200, and 1800), or grafted with 3-aminopropyltrimethoxysilane (APTMS). The performance of aminosilicas was assessed to determine equilibrium CO2 adsorption capacity, adsorption kinetics, and cyclic stability.

Over the past decade, hydrogen gas generation has been a critical component toward clean energy due to its high specific energy content. Generating hydrogen gas from water is crucial for future applications, including space transportation. Recent studies show promising results using silicon nanoparticles (SiNPs) for spontaneous hydrogen generation, but most methods require external energy like high temperature or pressure. In this work, we investigated hydrogen production from SiNPs without external energy by leveraging high pH water using sodium hydroxide and optimizing the process with a microfluidic approach. When comparing the physical dispersion methods using the 0.1 mg/mL case, ultrasonic bath produced more hydrogen than magnetic stirrer. In this thesis, 0.01% dextran with pure SiNPs at concentrations of 0.1 mg/mL, 0.2 mg/mL, and 0.3 mg/mL was analyzed. The highest concentration with dextran generated at least 40% less hydrogen than silicon alone, thus dextran did not increase hydrogen gas.

The study of non-invasive techniques to analyze the propagation of corrosion in steel reinforced concrete structures proves to be a great alternative to better understanding the corrosive process of rebar and increasing its useful life. The study presented in this document examines the evolution of steel reinforced concrete corrosion over time by applying a small anodic current over four samples, one with a single rebar (16X) and three with three rebars. The rebars were interconnected to apply the anodic current and accelerate their corrosion. Galvanostatic Pulse (GP) was used. This method applies a constant current pulse to the rebar for 150 seconds while monitoring the potential of the rebars. Each rebar's corrosion current was assessed using GP measurements when no anodic current was applied, and the rebars were disconnected. Sample 16X additionally underwent ultrasonic acoustic analysis by collecting the surface and rebar echo response with a transducer and modeling the sound propagation for poroelastic media with an adapted version of the novel Biot-Stoll method.

Due to technological advancement, energy consumption and demand have been increasing significantly, primarily satisfied by fossil fuel consumption. This reliance on fossil fuels results in substantial greenhouse gas emissions, with CO₂ being the most prominent contributor to global warming. To mitigate this issue and prevent CO₂ emissions, Carbon Capture, Utilization, and Storage (CCUS) technologies are employed. Among these, the amine scrubbing method is widely used due to its high CO2 capture efficiency and regenerative ability. However, this method has drawbacks, including high toxicity, corrosion, and substantial freshwater consumption.
To develop an environmentally sustainable carbon capture solution, researchers are exploring alternatives such as the use of seawater and enhanced CO2 capture with catalysts. In this study, we analyze the catalytic performance of nickel nanoparticles (NiNPs) in seawater with carboxymethyl cellulose (CMC) polymers. Using flow-focusing geometry-based microfluidic channels, we investigated CO₂ dissolution at various concentrations of nanoparticles and CMC polymers. The objective is to optimize the concentration of nanoparticles and CMC polymers for effective CO₂ dissolution. We utilized NiNPs with diameters of 100 nm and 300 nm in CMC concentrations of 100 ml/L, 200 ml/L, and 300 ml/L. Additionally, NiNP concentrations ranging from 6 mg/L to 150 mg/L were tested for CO₂ dissolution in seawater. The results indicated that a concentration of 10 mg/L NiNPs in 100 mg/L CMC provided a CO₂ dissolution of 57%, the highest for this specific CMC concentration. At CMC concentrations of 200 ml/L and 300 ml/L, NiNP concentrations of 70 mg/L and 90 mg/L achieved CO₂ dissolution rates of 58.8% and 67.2%, respectively.

This study aims to address the unique challenges of transportation in rural and disconnected communities through innovative data-driven methodologies. The primary methods employed in this research involve Geographic Information Systems (GIS) tools and simulation techniques to model and assess the impact of flood zones on rural traffic dynamics. The study recognizes the distinct mobility patterns and limited infrastructure prevalent in rural areas, emphasizing the need for tailored solutions to manage flood-induced disruptions. By leveraging GIS tools, the study intends to spatially analyze existing transportation networks, population distribution, flood-prone areas, and key points of interest to formulate a comprehensive understanding of the local context. Simulation-based approaches using the PTV VISSIM platform will be employed to model and assess various flood scenarios and their effects on traffic flow and accessibility. This study’s outcomes aim to contribute valuable insights into improving accessibility, efficiency, and safety in transportation for these underserved areas during flood events. By combining GIS tools and simulation techniques, this research seeks to provide a robust framework for data-driven decision-making and policy formulation in the realm of rural and disconnected community mobility, particularly in the context of flood risks.

This dissertation explores one-class classification (OCC) in the context of big data and fraud detection, addressing challenges posed by imbalanced datasets. A detailed survey of OCC-related literature forms a core part of the study, categorizing works into outlier detection, novelty detection, and deep learning applications. This survey reveals a gap in the application of OCC to the inherent problems of big data, such as class rarity and noisy data. Building upon the foundational insights gained from the comprehensive literature review on OCC, the dissertation progresses to a detailed comparative analysis between OCC and binary classification methods. This comparison is pivotal in understanding their respective strengths and limitations across various applications, emphasizing their roles in addressing imbalanced datasets. The research then specifically evaluates binary and OCC using credit card fraud data. This practical application highlights the nuances and effectiveness of these classification methods in real-world scenarios, offering insights into their performance in detecting fraudulent activities. After the evaluation of binary and OCC using credit card fraud data, the dissertation extends this inquiry with a detailed investigation into the effectiveness of both methodologies in fraud detection. This extended analysis involves utilizing not only the Credit Card Fraud Detection Dataset but also the Medicare Part D dataset. The findings show the comparative performance and suitability of these classification methods in practical fraud detection scenarios. Finally, the dissertation examines the impact of training OCC algorithms on majority versus minority classes, using the two previously mentioned datasets in addition to Medicare Part B and Durable Medical Equipment, Prosthetics, Orthotics and Supplies (DMEPOS) datasets. This exploration offers critical insights into model training strategies and their implications, suggesting that training on the majority class can often lead to more robust classification results. In summary, this dissertation provides a deep understanding of OCC, effectively bridging theoretical concepts with novel applications in big data and fraud detection. It contributes to the field by offering a comprehensive analysis of OCC methodologies, their practical implications, and their effectiveness in addressing class imbalance in big data.