Pulsating flows in a Helmholtz pulse combustor tailpipe were numerically simulated by a commercial CFD software package, FLUENT. The effects of ambient temperature on the characteristics of the pulsating tailpipe flows were studied. Two study cases, with high and low levels of ambient temperature, were simulated with compressible flow equations. An additional case, with high ambient temperature, was simulated with incompressible (temperature-dependent density) flow equations. Results showed that the effect of ambient temperature on the mean temperature profile in the tailpipe was limited to the distance where the ambient fluid traveled into the tailpipe during the period of flow reversal. In this region, the amplitude of mass flow rate oscillation significantly increased, due to higher density associated with low ambient temperature. The overall effects of cooler ambient temperature included an increase in mean pressure at the entrance of the tailpipe and a decrease in the magnitude of velocity amplitude profile along the tailpipe. Interestingly, the mean velocities along the tailpipe, even at the tailpipe exit, were not affected by the cooler ambient air. The mean velocity at the exit corresponded to the higher temperature of fresh fluid from upstream, which was not affected by the ambient temperature, driven out of the tailpipe in each oscillation cycle. The linear acoustic theory with appropriate assumptions could be used to calculate the magnitude of the profiles of velocity amplitude along the tailpipe as a fair approximation, at least for the study cases in this thesis.

This thesis is concerned with the evolution of implicit or explicit surfaces. The first part of this thesis addresses three problems in fluid simulation: advection, thin film, and the volume error. First, we show that the back and forth error compensation and correction (BFECC) method can significantly reduce the dissipation and diffusion. Second, thin film is hard to simulate since it has highly complex liquid/gas interface that requires high memory and computational costs. We address this difficulties by using cell centered octree grid to reduce memory cost and a multigrid method to reduce computational cost. Third, the volume loss is an undesired side effect of the level set method. The known solution to this problem is the particle level set method, which is expensive and has small but accumulating volume error. We provide a solution that is computationally effective and can prevent volume loss without accumulation. The second part of this thesis is focused on filtering a triangle mesh to produce a mesh whose details are selectively reduced or amplified. We develop a mesh filter with a rational transfer function, which is a generalization from previously developed mesh filters. In addition, we show that the mesh filter parameters can be computed from the physical size of mesh feature.

The natural phenomena simulation in computer graphics is commonly achieved by the procedural methods or the physics model. However, these approaches are hard to directly approach the visual experience. On the other hand, the image reconstruction works can provide the outcome based on real images but lack of interactivity and efficiency on using image resource. For solving these drawbacks, we propose a novel method that enhances the fire simulation effect using the visual learning of image features and generates continuous animations by integrating with procedural methods. We first obtain the dynamics of fire contour by binarization and edge detection. The information extracted from images is gathered into a set of feature data called fire profile. To generate a long sequence of fire animation from a short clip of fire video, we propose two approaches of visual learning to utilize fire profile to produce continuous animation. One is to use the fire image to setup a color value lookup table which contains the average color value of the fire spatial divisions; the other is to design a state machine for describing fire wiggling movement that can generate effects based on user¡¦s input. During the rendering stage of 3D visualization, we set up the fire volume which connecting the feature points of two cross-views by the cubic spline. Then the edge points found on the fire volume can be used as the contour points of the supplementing slices and generate these supplements inside the planned fire volume to formulate a complete fire effect. The proposed method can raise not only the visual reality but also the interactive ability compared with the existing work.

Machinery investments represent a substantial portion of agribusiness firms’ costs. Because of high machinery costs, variable profit margins, and increasing competition, agribusiness managers continually seek methods to maintain profitability and manage risk. One relatively new method is jointly owning and sharing machinery. Contract design issues to enhance horizontal linkages between firms through machinery sharing are addressed. Specifically, costs and depreciation sharing between two firms entering into a joint machinery ownership contract are examined. Two, two-player models, a Nash equilibrium game theoretical model and an applied two-farm simulation model are used to determine impacts of machinery sharing on firms engaged in machinery sharing. The Nash equilibrium model determines theoretical optimal sharing rules for two generic firms. Using the Nash equilibrium model as the basis, the two-farm simulation model provides more specific insights into joint harvest machinery sharing. Both models include contractual components that are uniquely associated with machinery sharing. Contractual components include penalty payment structure for untimely machinery delivery and the percentages of shared costs paid and depreciation claimed paid by each firm. Harvesting windows for each farm and yield reductions associated with untimely machinery delivery are accounted for within the models. Machinery sharing can increase the NPV of after tax cash flows and potentially reduce risk. Sharing will, however, not occur if own marginal transaction costs and/or marginal penalty costs associated with untimely machinery delivery are too large. Further, if the marginal costs of sharing are small relative to own marginal net benefits, sharing will not occur. There are potential tradeoffs between the percentage of shared costs paid and the percentage of shared depreciation claimed depending on each farms’ specific tax deductions. Harvesting window overlaps help determine the viability of machinery sharing. Farms may be better off sharing larger, more efficient machinery than using smaller machinery even when harvest must be delayed. Percentages of shared costs, depreciation, and tax deductions have important tax implications that impact the after tax cash flows and should be considered when negotiating machinery sharing contracts.

The three marketing strategies, buying a put option, cash sale at harvest, and cash sale in June after December harvest, are simulated for six representative irrigated and dryland cotton farms in West Texas. Each marketing strategy is ranked using the net cash income probability distribution for the representative farms using stochastic efficiency with respect to a function (SERF). SERF rankings were consistent across dryland and irrigated farms. The buying of a put option was found to be the marketing strategy that produced the highest certainty equivalent (CE) for normal risk averse decision makers. Cash sale at harvest followed by cash sale in June marketing strategies were ranked second and third, respectively. A sensitivity analysis increased the national baseline price used in the model by 45 percent. Cash sale at harvest then consistently became the highest ranked marketing strategy followed by buying a put option and then cash sale in June. The research found that if a strike price and premium that covered the production costs of the representative farm was available during the pre-harvest period, the decision maker may have the ability to increase utility by hedging with the put option.

Foot and mouth disease (FMD) is a highly contagious viral infection that affects all Artiodactyls (cloven-hoofed) species. The United States has been free of FMD since 1929, and the entire population of cloven-hoofed species is therefore susceptible to FMD virus infection. In the face of an outbreak, it is crucial that appropriate control measures be applied rapidly to control the disease. However, in most cases decisions on mitigation strategies must be made with little current or empirical data and in the context of political, economic and social pressures. Disease spread models can be used to evaluate the design of optimal control strategies, for policy formulation, for gap analysis and to develop and refine research agendas when disease is not present. This research project is designed to investigate the potential role of wildlife (deer) in the transmission and spread of FMD in an extensive livestock management system in southern Texas. The spread of FMD was simulated in white tailed deer populations using a Geographic Automata model. Past research has focused primarily on modeling the spread of FMD in livestock populations. There has been limited research into the potential role of wildlife in the spread and maintenance of FMD, specifically in the United States and using a spatial modeling approach. The study area is a nine-county area located in southern Texas, bordering Mexico. It is a region of concern for the introduction of foreign animal diseases, particularly through the movement of wild and feral animal species. It is both a strategic location and is generally representative of the many similar eco-climatic regions throughout the world. It is an ideal model landscape to simulate FMD incursions. In this research project, the potential spread of FMD is simulated based on various spatial estimates of white tailed deer distribution, various estimates of critical model parameters (such as the latent and infectious periods), seasonal population variability and in the face of potential pre-emptive mitigation strategies. Significant differences in the predicted spread were found for each group of simulations. The decision-support system developed in the studies described in this dissertation provide decision-makers and those designing and implementing disease response and control policy with information on the potential spread of a foreign animal disease incursion with a likely wildlife reservoir. Use of such a decision-support system would enhance the disease incursion preparedness and response capacity of the United States.

Reservoir geomechanics is a production induced phenomena that is experienced in large number of fields around the world. Hydrocarbon production changes the pore pressure which in turn alters the in-situ stress state. For reservoirs that are either stress sensitive or where rock is soft and unconsolidated, stresses have appreciable effect on rock properties like porosity and permeability. Anisotropic and isotropic permeability changes affect flow direction and movement of flood front thereby influencing well performance and reservoir productivity. Coupling of geomechanical calculation with multi-phase flow calculation is needed to make prudent predictions about the reservoir production and recovery. The post processing tools provided with the simulators cannot monitor flood front movement and fail to capture important information like flow directionality and dominant phase in a flow. Geomechanical simulation is combined with streamline tracing to aid in better understanding of the reservoir dynamics through visualization of flow patterns in the reservoir. Streamline tracing is a proved reservoir engineering tool that is widely used by industry experts to capture information on flood movement, injector-producer relations and swept area. In the present research, we have incorporated total velocity streamlines and phase streamlines for coupled geomechanical simulation and compared the results with streamline tracing for conventional reservoir simulator to explain geomechanics behavior on reservoir flow processes in a more detailed and appealing manner. Industry standard simulators are used for coupled geomechanical simulation and conventional simulation and streamline tracing has been done through in-house tracing code. The research demonstrates the benefits and power of streamline tracing in visualizing flow patterns through work on two cases; first, a synthetic case for studying water injection in a five spot pattern and second, a SPE 9th comparative study. The research gives encouraging results by showing how geomechanics influences reservoir flow paths and reservoir dynamics through visualization of flow. The streamlines captures flow directionality, information regarding appearance and disappearance of gas phase and the connectivity between injector and producer.

Fully atomistic molecular dynamics (MD) simulations were used to predict the properties of EPON 862 cross-linked with curing agent DETDA, a potentially useful epoxy resin for future applications of nanocomposites. The properties of interest were density (at nearambient pressure and temperature), glass transition temperature, bulk modulus, and shear modulus. The EPON molecular topology, degree of curing, and MD force-field were investigated as variables. The range of molecular weights explored was limited to the oligomer region, due to practical restrictions on model size. For high degrees of curing (greater than 90%), the density was found to be insensitive to the EPON molecular topology and precise value of degree of curing. Of the two force-fields that were investigated, cff91 and COMPASS, COMPASS clearly gave more accurate values for the density and moduli as compared to experiment. In fact, the density predicted by COMPASS was in excellent agreement with reported experimental values. However, the bulk and shear moduli predicted by simulation were about two times higher than the corresponding experimental values.

The goals of this research include developing and identifying software technologies, which facilitate the use of buildings described in Building Information Modeling (BIM) tools in both a simulation and visualization. The study focused on the development of a tool to fulfill the visualization needs of a Visual Energy Use System. To accomplish this, the study identified an open BIM file standard, the Industry Foundation Classes (IFC). The study also identified a video game based 3D virtual environment, the Doom 3 Engine. A tool developed during the study, IFCtoMAP, converts IFC data into the .MAP file format understood by the Doom 3 Engine. Finally, the study identified the IFCtoIDF utility, which translates IFC data into a format understood by the building energy simulation program EnergyPlus. Data from the Building Information Modeling tool Revit Building exports to the .IFC file format, which in turn drives the two conversion utilities IFCtoMAP and IFCtoIDF. The output of the IFCtoIDF tool consists of an .IDF file that EnergyPlus uses to perform an energy simulation. The output of the IFCtoMAP tool consists of a .MAP file, which the Doom 3 game engine uses to display three dimensional first person perspective visualization. The result of the study was the successful creation of an automated tool that converts building geometry found in .IFC files into the .MAP file format understood by Doom 3 game engine. This document details the methods employed by the IFCtoMAP software along with a brief discussion of the IFCtoIDF conversion utility.

Dose calculation methodology was developed for irradiation treatment of complex-shaped foods. To obtain satisfactory electron beam irradiation of food products, a strict process control is required to ensure that the dose delivered to all parts of the treated product falls within some specified range. The Monte Carlo electron transport simulation and computer tomography (CT) scan technology were used to predict the dose distribution in complex shaped foods, an apple phantom composed of paraffin wax, chloroform, and methyl yellow, and a chicken carcass. The Monte Carlo code used was successfully tested against the experimental data, resulting in less than 5% discrepancy between the simulated and measured data. For 1.35 MeV electron beam simulation of apple phantom, tilting and axial rotation ensures dose distribution of the entire surface of the phantom, even reaching the critical regions of the apple stem and calyx ends. For 1 and 5 MeV X-ray simulations, both depth-dose curves show exponential attenuation after a build-up region. The depth to peak for the former is shorter than that of the latter. For 1.35 MeV electron beam simulation of a chicken carcass, dose adsorption occurred up to 5-7 mm deep, resulting in surface irradiation of the carcass. For 10 MeV electron beam simulation, the doses within the carcass reached a peak of 1.2 times the incident dose with increasing depth. Two-sided X-ray (5 MeV) irradiation significantly improved the dose uniformity ratio, from 2.5 to 1.8. A web-based integrated system was developed for data manipulation and management for irradiation treatment of foods. Based on CT scan, three dimensional geometry modeling was used to provide input data to the general Monte Carlo N-Particle (MCNP) code. A web-based interface provided the on-line capability to formulate input data for MCNP and to visualize output data generated by MCNP. The integrated Matlab and Matlab Web Server programs automatically functions through the steps and procedures for data input and output during simulation. In addition, a database having D10 values (decimal reduction value), food nutrition composition, and qualities was integrated into the dose planning system to support food irradiation treatment.