It is a truism that the amount of information being generated in the modern digital world is increasing at an exponential rate. This is influencing engineering as it is in other forms of business, as well as everyday life. Engineering has a significant visual dimension to it: drawings, diagrams, sketches, photographs, graphs are the everyday language of the engineer. Despite the prevalence of such visual information, the role that such information plays and how it affects, for example, how documents can be reused is an under-researched area. This thesis thus proposes the important role of visual representations and images for supporting informed decisions, in particular for complex domains such as Engineering Design. The particular context for this research is associated with in-service design knowledge and information requirements. The increasing number of actual products in-service, the requirement to create safe design solutions quickly, the amassment of service data and the importance of product services to organisational competitiveness are all increasing the information pressures upon Design teams. The pervasive nature of visual representations in Engineering Design and prevalent document information suggests that they are an important asset within document information resources. This research focusses upon the purpose of Engineering Design image utilisation for information processing, and hence supporting efficient decision making. Some of the additional challenges identified throughout this research are the immaturity of current image recognition technologies and thus limitations of automated media extraction tools for supporting Design Engineers. This is significantly contributed to by the complexity of the information media and formats that constitute design engineering information and the current knowledge management trend to capture information without clear “reuse” purpose. The methods used to conduct this research demonstrate the merits of underused techniques in design engineering such as storyboarding. This storyboarding method is used for investigating the facets of tacit knowledge and the underpinning cognitive processing of document information resources for critical Design Engineering informative content. The innovative research method developed provides a useful framework for the collection of rich data using simulated tasks. The data collection is a rich multi-stream recording of design engineers in industry conducting work based scenarios. In particular the focus is upon conducting efficient research in industrial working practices with minimal facing research time with design engineers and the rich data that can be collected from them in situ. This thesis illustrates that there are a number of pressing difficulties in reusing image media, both technical process related in nature. This is currently limiting the usefulness of valuable information resources in practice, but also significantly raises the information burden for design engineer. This thesis has attributed the value of reusing visual representations due to their important role in design engineering decisions. It has provided evidence of the intuitive and important human need for visual information to provide mental stimulation in particular for making confident design decisions. The storyboard research method has outlined an industrial data collection and decision coding framework that is reproducible and can be used to better understand human information processing, and thus supports the development of document information systems. Additional rich information utilisation patterns for design engineering document information have also been evidenced in the empirical research results provided. This thesis also provides practical industrial examples to suggest techniques that could overcome the current technological shortfalls limiting the “reuse” of visual information in documents for Design Engineers.

This dissertation presents original research into the melting process of a downward facing cryogenic solid hydrogen surface subject to a two dimensional axisymmetric jet impingement flow of superheated hydrogen vapor. The motivation for the study is to investigate concepts of storing rocket propellants as a solid and rapidly melting the solid for liquid propellant delivery to a rocket engine. The present study considers a more favorable liquid removal arrangement than prior (1970s) experiments which melted solid hydrogen at the bottom of a cryostat This is a numerical study that involves computation fluid dynamic (CFD) simulation of four distinct physical phenomena: (1) melting, (2) jet impingement heat transfer (JIHT), (3) multiphase transport, and (4) film breakup/droplet formation. The volume of fluid (VOF) method is used with the V2F turbulence model in a commercial CFD Navier-Stokes solver (FLUENT) to investigate the multiphase nature of melt transport and its interaction with the vapor stream; i.e., the phenomena relevant to effective heat transfer between the vapor and the melting interface. The goal of the research is: (1) to develop a numerical method to study the problem and (2) evaluate several simple configurations to begin investigating relevant phenomena for the purpose of enhancing melting rate. Many options exist for the vapor to interact with the solid surface. The scope of this initial research is limited to a steady jet of single phase superheated hydrogen vapor at fixed jet exit conditions (T = 525 R and Re = 11,000) at a fixed jet standoff ( H/D = 1.0). Condensation/vaporization are not considered. Although film breakup/droplet formation is a phenomenon where two dimensional features evolve into three dimensional events, this phenomenon is approximated as two dimensional to allow a computationally tractable problem for this initial study Calculations are performed validating the numerical method for melting and JIHT against known results. Validation of film breakup/droplet formation is cited in the literature. A numerical method is developed to model the four physical phenomena. Four simple configurations are evaluated and a fundamental understanding is obtained of the multiphase melt transport and vapor interaction / acase@tulane.edu

acase@tulane.edu

Anomaly detection refers to the identification of observations that are considered outside of normal. Since they are unknown to the system prior to training and rare, the anomaly detection problem is particularly challenging. Model based techniques require large quantities of existing data are to build the model. Statistically based techniques result in the use of statistical metrics or thresholds for determining whether a particular observation is anomalous. I propose a novel approach to anomaly detection using wavelet based algorithmic entropy that does not require modeling or large amounts of data. My method embodies the concept of information distance that rests on the fact that data encodes information. This distance is large when little information is shared, and small when there is greater information sharing. I compare my approach with several techniques in the literature using data obtained from testing of NASA's Space Shuttle Main Engines (SSME) / acase@tulane.edu

A method for solving the equations of compressible viscous gas dynamics is presented. It uses the combination of the random vortex method and the monotone upwind scheme for conservation laws (MUSCL). The random vortex method was designed based on the assumption of incompressibility of the fluid, and the MUSCL scheme only solves conservation laws whereas the combination of these two methods gives rise to a new method which is able to solve compressible fluid flow. The new method also enjoys the advantages of the random vortex method and the MUSCL scheme: capable of dealing with high Reynolds number flow and capable of dealing with gas flow involving strong shocks. The idea of making the combination of two methods is based on Hodge's decomposition theorem. According to this theorem, solutions to the gas dynamics equations can be decomposed into two parts, a divergence free part and a curl free part. The divergence free part, corresponding to incompressible but viscous fluid, is solved by the random vortex method. The curl free part, corresponding to inviscid but compressible fluid, is solved by the MUSCL scheme. The combination of these two methods will be referred to as the compressible vortex method. A detailed description of the random vortex method is given. The vortex sheet method is also presented which is used in conjunction with the random vortex method in creating vorticity along the boundaries of the computational domain. The MUSCL scheme is a second order extension of Godunov method, which is studied by first studying Godunov method for conservation laws. Numerical results are presented by MUSCL for both one and two dimensional flow problems. The compressible vortex method has been applied to test problems whose computational domain is a channel with a step. Results are obtained for both subsonic and supersonic flow problems / acase@tulane.edu

<p>A National Aeronautics and Space Administration computational fluid dynamics code, OVERFLOW 2, was modified to utilize a harmonic balance solution method. This modification allows for the direct calculation of the nonlinear frequency-domain solution of a periodic, unsteady flow while avoiding the time consuming calculation of long physical transients that arise in aeroelastic applications.</p><p>With the usual implementation of this harmonic balance method, converting an implicit flow solver from a time marching solution method to a harmonic balance solution method results in an unstable numerical scheme. However, a relatively simple and computationally inexpensive stabilization technique has been developed and is utilized. With this stabilization technique, it is possible to convert an existing implicit time-domain solver to a nonlinear frequency-domain method with minimal modifications to the existing code.</p><p>This new frequency-domain version of OVERFLOW 2 utilizes the many features of the original code, such as various discretization methods and several turbulence models. The use of Chimera overset grids in OVERFLOW 2 requires care when implemented in the frequency-domain. This research presents a harmonic balance version of OVERFLOW 2 that is capable of solving on overset grids for sufficiently small unsteady amplitudes.</p> / Dissertation

An efficient methodology for parameter identification is developed for general multi-degree of freedom linear or nonlinear systems in continuous time. The new methodology is based on a gradient flow algorithm and demonstrated to be useful in identifying unknown parameters for the systems defined by both linear and nonlinear differential equations. The new methodology identifies the unknown parameters by solving a system of differential equations rather than the conventional post-data fitting analysis. It is named the trajectory gradient integral flow (TGIF) algorithm. For the cases of stable, one-dimensional linear systems, the asymptotic stability of the TGIF algorithm is guaranteed in the neighborhood of the operating point. For higher order linear or nonlinear systems, certain criteria for stability are developed using the eigenvalue analysis and the Routh-Hurwitz stability criteria. A well-known system identification result is that any method works the best with non-steady, non-periodic data set that is driven by randomized inputs, however this is not an essential requirement with the TGIF algorithm. In fact, it is possible to perform efficient parameter identification with the TGIF algorithm using an unit step input or a simple sine input. Improvements over previous approaches include: (1) the new methodology is easy to apply for nonlinear systems, (2) it works well with a simple unit step or sinusoidal inputs as well as bounded (control) inputs, (3) it demonstrates a reasonable large "domain of attraction", (4) it can be applied for either "on-line" or "off-line" parameter identification processes.

In this work the problem of actuator and sensor mapping and controller design for the flexible structure control is approached as minimization of the residual deformations index (Hinfinity norm of the closed-loop disturbance - deformation path) over the set of non-destabilizing feedback controllers and over the set of possible actuator and sensor mappings. Computational load associated with this approach is reduced by restricting the search to the mapping areas where an inexpensive lower estimate of residual deformations index (derived as a part of this study) is less than the desired value of this index. Further improvement is achieved by including statistical description of the difference between the actual and the estimated performance index over the set of mappings, in order to adjust the level of the mapping acceptance/rejection in such a way that the number of rejected mappings is increased. Serial and parallel optimization procedures based on exhaustive search and genetic algorithms are discussed. These concepts and algorithms are applied to test cases of simply supported beam, the UCLA Large Space Structure, and a telescope mirror model: a hinged round plate.

In-Situ Resource Utilization (ISRU) can reduce the mass launched to low Earth orbit for a Mars Sample Return Mission by as much as 75%. Solid Oxide Electrolysis is a candidate technology for producing oxygen out of carbon dioxide. A testbed was developed to test the effect of electrode structure on electrode performance. The testbed used all metal interconnects. Electrodes of 0.5 microns (applied by evaporative deposition) and 10-12 microns thick (applied by airbrushing) were produced. The disks were tested in argon, oxygen, and carbon dioxide. The thin electrode showed deterioration in both short term and long term tests. The thick electrode showed no deterioration even over a 120 hour test. Oxygen was produced from carbon dioxide for extended durations. The experiments show the thicker electrode is needed for an oxygen production system. The electrode/electrolyte disks withstood normal handling without damage.

With an always-increasing need for higher integration, the wireless industry poses challenges regarding miniaturization and high performance circuitry. In addition, the solutions require compatibility with the rest of the design for integration and manufacturing. The present work depicts the progress toward a novel, high quality, one-pole filter-resonator operating in the 5.6-5.8 GHz range. Quality factors up to 640 are demonstrated on Silicon planar structures with volume of 177 mm³. Further size reduction yielded a volume of 24.5 mm³ and a quality factor of 186.