An advanced finite-element package, tailored to the static and dynamic analysis of radial impellers has been produced. Two families of new elements, one for thin and thick plates and the other for thin and thick shells, have been derived and proved to perform very well within a wide range of structural thicknesses. Static and dynamic economical solvers, two- and three-dimensional mesh generation and plotting, sectorial symmetric analysis, steady state response, transient response, and other programs are part of the large number of facilities available in the package. The finite-element package has been validated by solving a large number of simple case studies and comparing the package results with those obtained from analytical solutions. Two different radial impeller, experimental validation tests have been carried out, the first being the dynamic analysis of a radial impeller using the time averaged holographic technique, and the second the measurement of the steady-state stresses by means of a strain-gauge/slip ring assembly for a rotating impeller. The experimental results have been shown to be in good agreement with those obtained from the package.

A computational and physical modelling study is made of the removal of inclusions from liquid steel by use of a novel form of hydrocyclone in which a solid conical core that replaces the conventional vortex finder acts as a guide to the spiralling liquid flow and acts also as a capture surface for disentrained inclusions. In preliminary investigations, an inviscid computational model is derived that is found to be effective in outlining the general behaviour of specific hydrocyclone flows when tested against published experimental results. The more generally applicable commercial CFD code Fluent is likewise tested, from which it is shown that, among the turbulence models available, the anisotropic turbulence typical of spiralling hydrocyclone flows requires a form of Reynolds stress model for effective computation. The conventional k-c model is found to be misleading. On this basis, mathematical modelling and optimal computational design of hydrocyclones containing an axial conical solid core show that the separation efficiency of the cyclone is profoundly enhanced by the presence of a core, and that by use of a particle tracking model effective centripetal migration of inclusion particles in steel will occur towards the core. Experiments with a water model of computed optimal cyclone designs provided effective validation of the numerical study. Photographically active particles of neutral density were tracked by a novel stroboscopic technique which permitted bi-directional observation revealing instantaneous velocity, spatial position and spiral angle. Using populations of low density particles having the same spectrum of Stokes velocity as inclusions in the size range 35 to 150 microns found typically in liquid steel, sampling by Coulter counter showed that effective separation to the core surface of particles down to an equivalent size of 30 microns was achieved. In a final step, a pilot cyclone design for use with steel was established and water model tests at full scale showed that stable cyclone flow and discharge are achievable with gravity feed to the cyclone.

Software processes relate to the sequence of steps that must be carried out by humans to pursue the goals of software engineering. In order to have an accurate representation of what these steps actually are, software processes can be modelled using a process modeling language (PML). Some PMLs simply support the specification of the steps, while others enable the process to be executed (or enacted). When enacted, software processes can provide guidance, automation and enforcement of the software engineering practices that are embodied in the model. Although there has been much fruitful research into PMLs, their adoption by industry has not been widespread. While the reasons for this lack of success may be many and varied, this thesis identified two areas in which PMLs may have been deficient: human dimension issues in terms of support for awareness and visualisation; and support for addressing management and resource issues that might arise dynamically when a process model is being enacted. In order to address some of these issues, a new visual PML called Virtual Reality Process Modelling Language (VRPML) has been developed and evaluated. Novel features have been introduced in VRPML to include support for the integration of a virtual environment, and dynamic creation and assignment of tasks and resources at the PML enactment level. VRPML serves as a research vehicle for addressing our main research hypothesis that a PML, which exploits a virtual environment, is useful to support software processes for distributed software engineering teams.

Models of the writing process are used to design software tools for writers who work with computers. This thesis is concerned with the construction of a model of fiction writing. The first stage in this construction is to review existing models of writing. Models of writing used in software design and writing research include behavioural, cognitive and linguistic varieties. The arguments of this thesis are, firstly, that current models do not provide an adequate basis for designing software tools for fiction writers. Secondly, research into writing is often based on questionable assumptions concerning language and linguistics, the interpretation of empirical research, and the development of cognitive models. It is argued that Saussure's linguistics provides an alternative basis for developing a model of fiction writing, and that Barthes' method of textual analysis provides insight into the ways in which readers and writers create meanings. The result of reviewing current models of writing is a basic model of writing, consisting of a cycle of three activities - thinking, writing, and reading. The next stage is to develop this basic model into a model of fiction writing by using narratology, textual analysis, and cognitive psychology to identify the kinds of thinking processes that create fictional texts. Remembering and imagining events and scenes are identified as basic processes in fiction writing; in cognitive terms, events are verbal representations, while scenes are visual representations. Syntax is identified as another distinct object of thought, to which the processes of remembering and imagining also apply. Genette's notion of focus in his analysis of text types is used to describe the role of characters in the writer's imagination: focusing the imagination is a process in which a writer imagines she is someone else, and it is shown how this process applies to events, scenes, and syntax. It is argued that a writer's story memory, influences his remembering and imagining; Todorov's work on symbolism is used to argue that interpretation plays the role in fiction writing of binding together these two processes. The role of naming in reading and its relation to problem solving is compared with its role in writing, and names or signifiers are added to the objects of thought in fiction writing. It is argued that problem solving in fiction writing is sometimes concerned with creating problems or mysteries for the reader, and it is shown how this process applies to events, scenes, signifiers and syntax. All these findings are presented in the form of a cognitive model of fiction writing. The question of testing is discussed, and the use of the model in designing software tools is illustrated by the description of a hypertextual aid for fiction writers.

No description available.

No description available.

The credibility of the scientific methodology of mathematical models and their adequacy to form the basis of public policy decisions has frequently been challenged. Skeptical views of the scientific value of modelling argue that there is no true model of an ecological system, but rather several adequate descriptions of different conceptual basis and structure. The purpose of this work was to first advance the Bayesian calibration of process-based models for guiding the water quality criteria setting process in Hamilton Harbour, Ontario, Canada. The analysis suggests that the water quality targets for total phosphorus and chlorophyll a concentrations will likely be met, if the recommendation for phosphorus loading at the level of 142 kg day-1 is achieved. My dissertation also examines how the Bayesian approach can effectively support the decision making process by synthesizing the predictions of different models developed for the same system. The model averaging approach consolidates the finding that the existing total phosphorus goal is most likely unattainable. The discrepancy between the chlorophyll a predictions of the two models pinpoints the need to delve into the dynamics of phosphorus in the sediment-water column interface. This work also aims to examine statistical formulations that explicitly accommodate the covariance among the process error terms for various model endpoints. The analysis suggests that the statistical characterization of the model error can be influential to the inference drawn by a modelling exercise. Finally, my dissertation challenges the capacity of the ecological foundation of eutrophication models to predict the role of nutrient regeneration. It shows that the recycled nutrients can be significant drivers in low as well as in high-productivity ecosystems depending on the period of the year examined. My dissertation also discusses several prescriptive guidelines that should be helpful towards a structural and methodological improvement of eutrophication modelling.

Models of aqueous metal speciation and ecotoxicity have become commonplace due to their ability to estimate metal behaviour. This study evaluated commonly used aqueous geochemical speciation and ecotoxicity models with application to a mine impacted lake in northern Manitoba. The geochemical speciation model Winderemere Humic Aqueous Model (WHAM) was compared with Diffusive Gradients in Thinfilm (DGT) measurements of zinc and copper. DGT measurements in the water column corresponded well with WHAM-estimated Zn2+, Cu2+ was off by up to 100x. Additional metal, either from small organic bound species or dissolution of metal sulphides from resuspended sediment, served to improve model estimates. The single metal Biotic Ligand Model (BLM) predicted acute toxicity to Daphnia magna attributable to copper but not zinc, at low pH (3.55 – 5.5). Comparison of results did not show a significant difference between the single and mixture BLMs, suggesting a non-interactive effect on metal toxicity for measured water chemistry.

The credibility of the scientific methodology of mathematical models and their adequacy to form the basis of public policy decisions has frequently been challenged. Skeptical views of the scientific value of modelling argue that there is no true model of an ecological system, but rather several adequate descriptions of different conceptual basis and structure. The purpose of this work was to first advance the Bayesian calibration of process-based models for guiding the water quality criteria setting process in Hamilton Harbour, Ontario, Canada. The analysis suggests that the water quality targets for total phosphorus and chlorophyll a concentrations will likely be met, if the recommendation for phosphorus loading at the level of 142 kg day-1 is achieved. My dissertation also examines how the Bayesian approach can effectively support the decision making process by synthesizing the predictions of different models developed for the same system. The model averaging approach consolidates the finding that the existing total phosphorus goal is most likely unattainable. The discrepancy between the chlorophyll a predictions of the two models pinpoints the need to delve into the dynamics of phosphorus in the sediment-water column interface. This work also aims to examine statistical formulations that explicitly accommodate the covariance among the process error terms for various model endpoints. The analysis suggests that the statistical characterization of the model error can be influential to the inference drawn by a modelling exercise. Finally, my dissertation challenges the capacity of the ecological foundation of eutrophication models to predict the role of nutrient regeneration. It shows that the recycled nutrients can be significant drivers in low as well as in high-productivity ecosystems depending on the period of the year examined. My dissertation also discusses several prescriptive guidelines that should be helpful towards a structural and methodological improvement of eutrophication modelling.

Models of aqueous metal speciation and ecotoxicity have become commonplace due to their ability to estimate metal behaviour. This study evaluated commonly used aqueous geochemical speciation and ecotoxicity models with application to a mine impacted lake in northern Manitoba. The geochemical speciation model Winderemere Humic Aqueous Model (WHAM) was compared with Diffusive Gradients in Thinfilm (DGT) measurements of zinc and copper. DGT measurements in the water column corresponded well with WHAM-estimated Zn2+, Cu2+ was off by up to 100x. Additional metal, either from small organic bound species or dissolution of metal sulphides from resuspended sediment, served to improve model estimates. The single metal Biotic Ligand Model (BLM) predicted acute toxicity to Daphnia magna attributable to copper but not zinc, at low pH (3.55 – 5.5). Comparison of results did not show a significant difference between the single and mixture BLMs, suggesting a non-interactive effect on metal toxicity for measured water chemistry.