Typescript (photocopy) / v, 169 leaves : ill. ; 31 cm. / Title page, contents and abstract only. The complete thesis in print form is available from the University Library. / Thesis (Ph.D.) Dept. of Organic Chemistry, University of Adelaide, 1983

"October 2002" / Includes bibliographical references (leaves 180-191) / v, 191 leaves : ill. ; 30 cm. / Title page, contents and abstract only. The complete thesis in print form is available from the University Library. / Thesis (Ph.D.)--University of Adelaide, Dept. of Chemistry, 2003

The problem addressed is absence of a class of objects in a finite set of objects, which is investigated by considering absence of a species and absence in relation to a threshold. Regarding absence of a species, we demonstrate that the assessed probability of absence of the class of objects in the finite set of objects given absence of the class in the sample is either exactly or approximately equal to the probability of observing a specific single object from the class of objects given the protocol for observation, where probability is interpreted as a degree of belief. Regarding absence in relation to a threshold, we develop a new estimator of the upper confidence bound for the finite population distribution function evaluated at the threshold and investigate its properties for a set of finite populations. In addition we show that estimation regarding the initial ordered value in the finite population has limited usefulness. / Graduation date: 1998

Two related techniques, based upon ion-exchange solid phase extraction, have been developed for the determination of arsenic speciation. The inorganic arsenic species arsenite (As(III)) and arsenate (As(V)) are separated by anion-exchange and detected with graphite furnace atomic absorption spectrophotometry (GFAAS) with a nickel matrix modifier. The first separation technique, which is based on a published method, utilizes a strong anion-exchange resin in a column format. The method was refined to achieve a cleaner and more rapid separation of the As species. In the second separation technique, the recently available Empore[superscript TM] anion-extraction disks are used. In both cases, rapid separations of several samples are achieved with the use of a vacuum manifold. The simplicity of the separation techniques allows them to be applied in the field which eliminates potential problems due to sample storage. In the pH range of most natural water samples (5-9), As(III) exists as a neutral species which is not retained by the resin, while As(V) exists as a monovalent or divalent species which are subsequently retained by the resin. The two arsenic species are collected in 3 to 4 fractions with As(III) appearing in the first two fractions. The As(V) species is eluted from the resin with 0.1 M HCl and collected in the last one or two fractions. Percent recoveries for each species range from 94 to 99%. The detection limit for each species with GFAAS is 2 ��g/L. The speciation techniques were used successfully in several applications. First, the resin technique was used to monitor the oxidation of As(III) by 0���, H���0���, and ��-Mn0���. The technique was also used to monitor the reduction of As(V) by Fe(II) and in solutions containing combinations of Fe(II), Fe(III), and a scorbic acid. Second, the resin technique was used to monitor the redox behavior of arsenic in soil slurries in bio-reactor systems. Upon spiking the soil slurry to a level of 500 ��g/L As(V), 80 to 90% of the As(V) was immediately adsorbed, presumably to hydrous Fe(III) oxides. In general, as conditions became more reducing, total soluble arsenic increased as a result of either abiotic or biotic reduction of the As(V) to the more soluble As(III). Third, the disk technique was applied in the field to determine arsenic speciation in creek water at Sutter Creek, Ca., where homes are built upon a large pile of mine tailings containing arsenic. In the creek water, no As(III) was detected but As(V) was detected at a level of 8 ��g/L. Fourth and finally, the resin technique was used to determine arsenic speciation when a sample of the mine tailings was placed in a reactor and combined with a soil slurry thus simulating a flooded condition. As conditions became more reducing, up to 800 ��g/L As was detected in solution with As(III) accounting for almost 90% of total soluble species. Also presented here is an investigation of zinc amalgam as a reducing agent for Cr(III) and selected redox indicators. Zinc amalgam, in a column format, also known as the classic Jones Reductor, provides an efficient means for production of Crap and reduced forms of various redox indicators. Finally, the reduction capabilities of Ti(III) citrate and zinc amalgam were compared. / Graduation date: 1996

Problems involving partial or ordinary differential equations arise in various fields of science. Therefore, the task of obtaining exact solutions of differential equations is of primary importance, and attracts high attention. The main purpose of the current thesis is the development of a Maple-based, symbolic software package for symmetry reduction of differential equations and computation of symmetry-invariant solutions. The package developed in the current thesis is compatible with and can be viewed as an extension of the package GeM for symbolic symmetry analysis, developed by Prof. Alexei Cheviakov. The reduction procedure is based on the Lie's classical symmetry reduction method involving canonical coordinates. The developed package is applicable for obtaining solutions arising from extension of Lie's method, in particular, nonlocal and approximate symmetries. The developed software is applied to a number of PDE problems to obtain exact invariant solutions. The considered equations include the one-dimensional nonlinear heat equation, the potential Burgers' equation, as well as equations arising in nonlinear elastostatics and elastodynamics.

Problems involving partial or ordinary differential equations arise in various fields of science. Therefore, the task of obtaining exact solutions of differential equations is of primary importance, and attracts high attention. The main purpose of the current thesis is the development of a Maple-based, symbolic software package for symmetry reduction of differential equations and computation of symmetry-invariant solutions. The package developed in the current thesis is compatible with and can be viewed as an extension of the package GeM for symbolic symmetry analysis, developed by Prof. Alexei Cheviakov. The reduction procedure is based on the Lie's classical symmetry reduction method involving canonical coordinates. The developed package is applicable for obtaining solutions arising from extension of Lie's method, in particular, nonlocal and approximate symmetries. The developed software is applied to a number of PDE problems to obtain exact invariant solutions. The considered equations include the one-dimensional nonlinear heat equation, the potential Burgers' equation, as well as equations arising in nonlinear elastostatics and elastodynamics.

No description available.

In a normal microarray classification problem there will be many genes, on the order of thousands, and few samples, on the order of tens. This necessitates a massive feature space reduction before classification can take place. While much time and effort has gone into evaluating and comparing the performance of different classifiers, less thought has been spent on the problem of efficient feature space reduction. There are in the microarray classification literature several widely used heuristic feature reduction algorithms that will indeed find small feature subsets to classify over. These methods work in a broad sense but we find that they often require too much computation, find overly large gene sets or are not properly generalizable. Therefore, we believe that a systematic study of feature reduction, as it is related to microarray classification, is in order. In this thesis we review current feature space reduction algorithms and propose a new, mixed model algorithm. This mixed-modified algorithm uses the best aspects of the filter algorithms and the best aspects of the wrapper algorithms to find very small yet highly discriminant gene sets. We also discuss methods to evaluate alternate, ambiguous gene sets. Applying our new mixed model algorithm to several published datasets we find that our new algorithm outperforms current gene finding methods.

An investigation of turbulent structure modification of a boundary layer for a fully developed channel flow by microbubble injection close to the upper wall was carried out using Particle Image Velocimetry (PIV). Two-dimensional velocity components in an x-y plane at Reynolds number of 5128 based on the half height of the channel and bulk velocity were measured. Microbubbles, with an average diameter of 30 ??m were produced by electrolysis and injected in the buffer layer. Different values of the void fraction were attained and used to evaluate the effects of the presence of microbubbles and their concentration within the boundary layer. A reduction in drag was observed due to the injection of microbubbles. Drag reduction augments as the value of the void fraction increases. Furthermore, increases in both the non-dimensional values of streamwise and normal turbulent intensities, normalized by the friction velocity were observed with the void fraction growth. A gradual decrease in the Reynolds shear stresses was achieved as the void fraction increases. This effect is due to a ??decorrelation?? or ??decoupling?? between the streamwise and normal fluctuating velocities. Modifications in the length and time scales due to the presence of microbubbles were detected by calculating two-point correlation coefficients in one and two dimensions and the autocorrelation coefficient at various locations within the measurement zone. Streamline length and time scales were increased. On the contrary, the normal length and time scales were decreased. The vorticity and strain rate values decreased with the injection of microbubbles. Turbulent energy production was also decreased within the boundary layer. Quadrant analysis was used to find out the contribution of the u?? and v?? fluctuating velocity components to the Reynolds stress. The presence of microbubbles reduces the contribution to the Reynolds stresses by Q4 events (sweeps), which are responsible for the production of skin friction. Vortical structure detection in the measurement area was pursued. The structure with and without the microbubble injection is compared. In this study the presence of microbubbles within the boundary layer has produced several modifications in the flow structure as well as reduction in the drag.

Thesis (M.S.)--Oregon State University, 2010. / Printout. Includes bibliographical references (leaves 67-71). Also available on the World Wide Web.