Search

Moving boundary problems are often encountered in engineering and science. One of the methods to solve such problems numerically is the volume of fluid (VOF) method. The VOF method uses an additional field variable/to track the interface using the Eulerian fixed grid. In this work, the VOF method is combined with the Continuum surface force model. The governing flow equations for Newtonian fluids in incompressible flow in two dimensions are solved using the pressure-based algorithm. A linear interface reconstruction algorithm is developed and improved by interpolations. In order to validate the code, the problem of the convection of a two-fluid channel flow with a cylindrical interface is given as an example.

This paper demonstrates that neuroevolution is an effective method to determine an optimal neural network topology. I provide an overview of the NeuroEvolution of Augmenting Topologies (NEAT) algorithm, and describe how unique characteristics of this algorithm solve various problem inherent to neuroevolution (namely the competing conventions problem and the challenges associated with protecting topological innovation). Parallelization is shown to greatly speed up efficiency, further reinforcing neuroevolution as a potential alternative to traditional backpropagation. I also demonstrate that appropriate parameter selection is critical in order to efficiently converge to an optimal topology. Lastly, I produce an example solution to a medical classification machine learning problem that further demonstrates some unique advantages of the NEAT algorithm.

The world's oceans provide the basis for life on the planet. One microscopic algae, the coccolithophores, and Emiliania huxleyi in particular, is a major source of carbon drawdown in the context of the global carbon cycle and account for a significant amount of the primary production in oceanic ecosystems. We know that the oceans are packed with marine viruses and they have an important role in the rise and fall of plankton populations but current mathematical models do not accurately account for virus-host interactions when predicting plankton blooms. Therefore I am using model optimization and comparison techniques to evaluate current models and reassess how the virus-host system should be described. Analysis of in situ blooms and mesocosm data has revealed previously unreported trends occurring over the course of an E. huxleyi bloom. Inclusion of virus-host interactions in models describing the calcifying coccolithophore E. huxleyi may give insight into global carbon cycling, as large-scale blooms are known to draw down significant amounts of atmospheric carbon.

The main objective of this research was to create a two dimensional and three dimensional Environmental Fluid Dynamics Code (EFDC) model using Aberdeen Pool of the Tennessee-Tombigbee Waterway for the purpose of determining the differences in their ability to address sediment transport. These objectives were reached in the results with comparisons of water levels, sediment concentrations, shear stress, and bed change. The models produced very similar results for the majority of the sediment transport throughout both models with the overall trend being deposition except in the upper limits of the Tombigbee River. The main differences between the two models are produced from the 2D model being depth averaged and the 3D being able to transport sediment vertically. The results show the 2D model tends to erode less and deposit more whereas the 3D model tends to follow the same pattern except for less deposition with more erosion.

The Upper Stage Reaction Control System provides in-flight three-axis attitude control for the Ares I Upper Stage. The system design must accommodate rapid thruster firing to maintain proper launch trajectory and thus allow for the possibility to pulse multiple thrusters simultaneously. Rapid thruster valve closure creates an increase in static pressure, known as waterhammer, which propagates throughout the propellant system at pressures exceeding nominal design values. A series of development tests conducted at Marshall Space Flight Center in 2009 were performed using a water-flow test article to better understand fluid characteristics of the Upper Stage Reaction Control System. A subset of the tests examined the waterhammer pressure and frequency response in the flight-representative system and provided data to anchor numerical models. This thesis presents a comparison of waterhammer test results with numerical model and analytical results. An overview of the flight system, test article, modeling and analysis are also provided.