No description available.

This study applies the Sacramento Soil Moisture Accounting Model (SAC-SMA)model to the Upper Black Creek Basin, Mississippi and attempts to improve operational lumped hydrologic model performance. The SAC-SMA is a lumped continuous soil moisture model which is typically calibrated continuously over time to all ranges in flow observed during the life of the gauge except when anthropogenic influences warrant historical data irrelevant. This study shows that persistent land use signatures are evident in the historical data indicating a shorter period of record for calibration is appropriate. This study also quantifies the error introduced to the operational model by inputting radar-derived precipitation estimates during forecast operations while Thiessen gauge weighted estimates are used to calibrate model parameters. Radar derived precipitation was used to calibrate the SAC-SMA model parameters for a shorter period of record than that used in the current operational set. The correlation coefficient improved 5 percent from 86 percent to 91 percent.

We performed atomistic modeling to study structural and mechanical properties of materials. We used density functional theory (DFT) for all the studies presented and constructed a method for quickly optimizing semi-empirical modified embedded atom method (MEAM) potentials. In our first study, we show that the reconstruction model in the literature for GaSb(001) is not predicted to have the lowest surface reconstruction energy. A modification was proposed that improves the energy. The second study tries to validate a crystal structure for the Φ Phase of Al-Mg-Zn. The third study deals with plain carbon steel, including some microalloying of Vanadium and vacancy assisted diffusion of Fe in Cementite(Fe3C). In the fourth study, we show a method for optimizing a MEAM potential. The code written is specific to hexagonal closed packed structures and was applied to a Magnesium potential.

In the software engineering process some tasks of software engineers are to design software documents, analyze the documents, and comprehend component relationships within software diagrams. Those diagrams represent the software architecture which models the structure, behavior, relationships, and constraints among system components while ignoring implementation detail. In the software lifecycle, the system is implemented from the software architecture and errors and mistakes caused from a lack of comprehension or incorrect comprehension could cause engineers to incorrectly design the system. These errors can be defined as lapses, slips, or lack of understanding and fall into three categories: skill, rule, and knowledge errors. The Gestalt principles of organization, from the cognitive science domain, deal with how humans perceive the world around them. This dissertation seeks to identify whether the Gestalt principles of continuity, similarity of size, proximity, and similarity of name affect comprehension of the Unified Modeling Language (UML) class diagrams. Diagram comprehension is measured by response time and subject accuracy on questions and the mental workload perceived by subjects while answering questions related to the diagrams. The research hypotheses are diagrams that utilize the Gestalt principles of continuity, similarity of size, proximity, and similarity of name will have faster response times, higher accuracy, and lower mental workload scores than diagrams that do not use the Gestalt principles. The results of the research indicate that the Gestalt principle of proximity helped ease diagram comprehension. Through the use of this design principle, the Gestalt principle of continuity is applied because line crossings, line bends, and line length are minimized. Subjects were prone to make more errors on knowledge based questions that dealt with system understanding and UML semantics than skill and rule questions that dealt with system connections and UML syntax. These results provide software designers heuristics that can lead to better diagram design and identifies software engineering tasks that can lead to more errors.

Polyhydroxyalkanoates (PHAs) are biodegradable plastic synthesized by microorganisms from renewable carbon resources and they are promising substitutes for conventional fossil-fuel-based plastics due to their similar physical properties. Pure cultures of particular microorganisms are commonly used for industrial PHA production but high production costs due to requirements of sterile conditions and refined substrates hinder the mass production of PHAs. Thus, model development for PHA production by microbes is essential to investigate the PHA formation and microbial metabolisms for enhanced productivity and PHA contents. In the present study, a comprehensive numerical model has been developed and calibrated for the non-growth associated PHA production process by Cupriavidus necator. The model parameters were calibrated with 8 selected experimental studies and the simulation results show good agreement with experimental data. Two methods were used to conduct sensitivity analysis: the simple method and the overall relative sensitivity analysis method. Maximum specific residual biomass growth rate was the most sensitive parameter. The calibrated model was used to investigate fed-batch feeding strategies that optimize PHA accumulation by limited nutrient feeding in the PHA production phase. The simulation results showed limited phosphorous feeding accumulated more PHA than limited nitrogen feeding. The optimal feeding strategy was determined to be limited phosphorous feeding at 5% of initial phosphorous during the PHB production phase, yielding simulated 226.0 g/L PHB at the end of the 168-hour operation. / Thesis / Master of Applied Science (MASc)

A new constitutive equation has been developed for the flow of blood through capillaries. Pressure drop and volume flow data of Haynes and Burton and Merrill et al. have beer utilized in this development for a range of radii from 57.04 micra to 747.4 micra and a hematocrit range of 8.8% to 82.5%. A comparison has been made with the Casson equation used by Merrill and Pelletier and the advantage of this new equation over the Casson equation has been verified. The usual assumption of no-slip-at-the-wall has been verified to be valid, up to a hematocrit level of 39.3%. / Thesis / Master of Engineering (ME)

No description available.

Throughout literature, one finds where numerous methodologies and models have been developed to predict the effect of surface roughness on a flat surface. Many of the models utilize a drag coefficient as one of the necessary parameters. In urban settings with groups of buildings, the drag coefficient on an individual obstacle would be determined by parameters like wind direction and the relative positioning of a building, in addition to Reynolds number and shape. Computational experiments were performed to simulate the fluid flow around a single row and two rows of “cube” obstacles. Based on dimensional analysis, the drag coefficient was formulated as a function of four input variables. The effect of these input variables on the drag coefficient was individually studied. Finally, using the central composite design method and the numerically obtained experiment data, a second-order mathematical model was devised for the drag coefficient as a function of the four input variables.

<p>We developed a mathematical model to investigate the transmission dynamics of SARS-CoV-2, the virus responsible for the COVID-19 pandemic. Our main model is built on the Susceptible-Exposed-Infected-Recovered framework to account for the unique characteristics of COVID-19. In a particular case of the main model, we assessed the optimal allocation of resources to mitigate the spread of the virus. Additionally, we expanded the main model to include another vaccination compartment to explore the strategic distribution of vaccinations. Our findings provided insights into the management of COVID-19 and could guide evidence-based decision-making for public health authorities.</p>

Under realistic scenarios, data are often incomplete, asymmetric, or of high-dimensionality. More intricate data structures often render standard approaches infeasible due to methodological or computational limitations. This monograph consists of four contributions each solving a specific problem within model-based clustering. An R package is developed consisting of a three-phase imputation method for both elliptical and hyperbolic parsimonious models. A novel stochastic technique is employed to speed up computations for hyperbolic distributions demonstrating superior performance overall. A hyperbolic transformation model is conceived for clustering asymmetrical data within a heterogeneous context. Finally, for high-dimensionality, a framework is developed for assessing matrix variate normality within three-way datasets. All things considered, this work constitutes a powerful set of tools to deal with the ever-growing complexity of big data / Dissertation / Doctor of Science (PhD)