This thesis includes the modification and optimization to an electric field mill based on micromachining technology. The sensor was originally designed to overcome the disadvantages of the conventional macroscopic field mill. Although it achieved all these listed above with a minimum detectable field strength at 42 V/m, some drawbacks are yet to be settled. In order to overcome these drawbacks, modifications are studied and put forward. Metal coating of the sensor surface could lead to a good electrical grounding that addresses the charging problem. Non-resonant working design was established with a shutter displacement around 5μm. Finite element simulations were set up to look into the optimizations of the structure parameters of the sensor, focusing on the shutter and electrodes. Moreover, the fabrication process was also studied with attempts of each step carried out in the NSFL of University of Manitoba.

Mobile Ad-hoc Wireless Networks (MANETs) present a new paradigm in which to realize a variety of communication technologies and services. The use of stochastic event-based simulation is a common approach to modelling MANET operations as part of the engineering process. To improve observations many simulations are often averaged together to produce estimations of MANET operation; however, to be statistically meaningful start-up transients must be removed, and only ergodic data averaged. These statistical issues of stationarity and ergodicity are often approached in an ad-hoc manner, if at all. This thesis presents a formal method to address these two statistical issues and applies it to the problem of quantifying MANET operation under different physical-layer jamming strategies. This demonstration illustrates the complex nature of MANET operation and the need for rigorous statistical analysis as part of the engineering process. / Graduate

No description available.

Buildings expend vast quantities of energy, which has a detrimental impact on the environment. Buildings systems are often oversized to cope with possible extreme environmental conditions. Building simulation provides an opportunity to improve building thermal design, but the available tools are typically used in combination in order to overcome their individual deficiencies. Two such tools, often used in tandem are computational fluid dynamics (CFD) and dynamic thermal modelling (DTM). DTM provides a coarse analysis, by considering external and internal thermal conditions over a building (including its fabric) over time. CFD is usually used to provide steady state analysis. Boundary conditions typically in the form of surface temperatures are manually input from DTM into CFD. CFD can model buildings dynamically, but is not commonly used, since solving for hugely different time constants of solid and air pose significant limitations, due to data generated and time consumed. A technique is developed in this study to tackle these limitations. There are two main strands to the research. DTM techniques had to be incorporated into CFD, starting from first principles of modelling heat transfer through solid materials. These were developed into employing the use of functions such as the 'freeze flow' function (FEF) and the 'boundary freeze' function (BFF) in combination with a time-varying grid schedule to model solids and air simultaneously. The FFF pauses the solution of all governing equations of fluid flow, except temperature. The BFF can be applied to solid boundaries to lock their temperatures whilst all other equations are solved. After extensive research the established DTM-CFD Procedure eventually used the FEF and BFF with transient periods and steady state updates, respectively. The second strand of research involved the application of the DTM-CFD Procedure to a typical office space over a period of 24-hours. Through inter-model comparisons with a fully transient simulation, the DTM-CFD Procedure proved to be capable of providing dynamic thermal simulations 16.4% more efficiently than a typical CFD code and more accurately than a typical DTM code. Additional research is recommended for the further improvement of the DTM-CFD Procedure.

The application of a wide range of computational methods to several problems in zeolite chemistry is explored in this thesis. Symmetry-constrained lattice energy minimisations have been performed on a series of pure silica polymorphs using the shell model for silicates and quantitative agreement is found between the experimental and calculated structures. The computed lattice energies of the silicas are found to be between 8 and 20 kJmol^-1 less stable than quartz. The energies are found to be directly dependent on the densities of the structures and show good agreement with a recent calorimetric study. A new forcefield for aluminophosphates based on the shell model has been obtained by fitting to the structure and properties of berlinite and lattice energy minimisation calculations have been carried out on a series of aluminium phosphate polymorphs. The experimental structures are reproduced to a reasonable accuracy, especially in cases where high quality crystallographic data are available on calcined structures. In cases where experimental methods give conflicting results regarding the space group symmetry, calculated structures having lower symmetry than those observed in the crystallographic studies are suggested. An approximately linear dependence of lattice energy on density is again observed; the computed lattice energies are found to span a range of 11.7 kJmol^-1 higher than berlinite, which compares to an experimentally determined range of 9.7 kJmol^-1. Proton binding calculations have been performed on the structure of H-SAPO-37 to determine the most favourable binding proton sites. The calculations correctly reproduce the sites which have the highest fractional occupancies in a crystallographic study. Molecular dynamics simulation has been used to study the diffusion of xenon in ferrierite and zeolite-L. It was found that at 298K and a loading level of 1.33 atoms per unit cell, diffusion down the tenring channel in ferrierite is a more facile process than down the wider twelve-ring channel in zeolite-L (D=8.90xl0^-9 m²s^-1 for ferrierite versus 1.78xl0^-9 for zeolite-L). This effect can be rationalised by consideration of the effect of channel shape on the diffusion pathway. Under the same conditions, the interaction energy was calculated to be more favourable for ferrierite (ΔU=-25.7 kJmol^-1 versus -20.0 kJmol^-1). A new forcefield for the interaction of hydrocarbons and aromatics with siliceous zeolites was fitted to thermochemical and crystallographic data. The forcefield successfully reproduced the crystallographically determined positions of pyridine and propylamine in siliceous ferrierite and dodecasil-3C. In addition, quantum mechanical calculations were used to fit a forcefield for the interaction of benzene with cation-containing zeolites. Molecular dynamics calculations were used to study the transport of benzene in siliceous faujasite. The coupling of lattice vibrations to the benzene molecule was found to enhance the mobility (for example, at 298K, D=0.11xl0^-9m²s^-1 with a fixed lattice compared to D=0.31xl0^-9m²s^-1 with aflexible lattice). Two diffusion regimes were observed corresponding to intra- and inter-cage benzene mobility which correlate well with hypothetical hopping pathways. Analogous pathways for benzene in cation-containing zeolites have shown that cation sites act as traps for the benzene in Na-X and Na-Y, which reduce the mobility compared to the siliceous case. In Na-X, the pathways are further modified by the addition of extra cation sites that act to reduce the hopping activation energy and therefore enhance the diffusion. This behaviour is consistent with observed trends in experimentally determined diffusivities.

This thesis examines various properties of galaxies and clusters within hierarchical models of structure formation. A simple model based on analytical scaling relations is applied to X-ray observations of clusters at low and high redshift, in an attempt to constrain cosmological parameters from their evolutionary properties. It is found that the density parameter, Ωo cannot be constrained using the data alone. Two independent constraints on the slope of the linear power spectrum, infer values of Ωo < 0.7 at 95 per cent confidence. The remainder of this thesis concentrates on the method of cosmological simulation, a self- consistent approach to the modelling of structure in the Universe. A parameter-space study is performed for the simplest model of galaxy formation: the radiative cooling of baryons within the cores of dark matter haloes. It is found that the properties of the galaxies in the simulations are insensitive to the range of parameters studied, with the exception of those that affect the cooling rate of the gas. For modest resolution and reasonable choices of physical parameters, the amount of baryons in galaxy material is around a factor of 2 too high. An investigation is then performed for including the effects of star formation and energy from supemovae (feedback) within cosmological simulations, to reduce the amount of gas that cools. The star formation rate is driven by the minimum density for which the stars formation occurs and, for high star formation efficiencies, is limited by the cooling rate of the gas. A successful model for feedback is found to require the prevention of reheated gas from cooling for a short period of time, as an attempt to mimick the properties of a multiphase medium. Finally, preliminary results are presented for simulations of a galaxy cluster, including the effects of radiative cooling, star formation and feedback. The properties of the cluster are found to vary significantly between models with and without feedback, due to the feedback reducing the star formation rate by reheating gas that cools.

This thesis presents the results of research performed into the simulation of electrical power systems using a set of microprocessors operating in parallel , The uses and methods of simulation on analog and single processor computers are discussed as well as on multiple processor machines . It then considers various methods already used in the field of simulation for both the dynamic and network sets of equations in detail and the problems of using them on parallel processors . Several possible methods of parallel simulation are proposed and the best of these developed into a detailed algorithm for simulating both the dynamic and network portions of the power system .The different types of multiprocessor system are looked at , both in terms of physical configuration and the type of hardware used to implement the different types of system .The problems inherent in parallel computing are discussed and a form of multiprocessor, suitable for the simulation algorithm, is then developed taking these problems Into account. The hardware is developed using widely available hardware and the algorithm Is then Implemented upon this hardware .The results obtained using the simulator show that the proposed system provides a more economical solution, both in terms of the time taken in producing results and in the cost of the system, when compared with a conventional single processor computing system such as a mini computer.

No description available.

A computer model was constructed that allowed two vessels involved in a possible collision situation to take collision avoidance action following the "International Regulations for Preventing Collisions at Sea". The mariners’ actions were modelled by the concepts of the domain and the RDRR (Range to Domain/Range-rate). The domain was used to determine if a vessel was threatening and the RDRR to determine the time at which a vessel should give-way to a threatening target. Each vessel in the simulation had four domains corresponding to the type of encounter in which the vessel was involved. Values for the time at which a vessel manoeuvres and the domain radii were determined from an analysis of high quality cine films of the radar at H.M. Coastguard at St. Margaret's Bay, Dover. Information was also taken from simulator exercises set up on the Polytechnic radar simulator. The two ship encounter was then developed to become the multi-ship encounter and eventually was able to model over 400 vessels over a two day period through a computer representation of the Dover Strait. A further development included a computer graphical representation of a radar simulator running in real-time, and which allowed a mariner to navigate one of the vessels using computer control. A validation of the computer model was undertaken by comparing the simulated results with those observed from the cine films. Following the validation several examples of the computer model being used as a decision support system were included.

No description available.