In this thesis, the problem of partitioning a network into interconnected sub-networks is addressed. The goal is to achieve a partitioning which satisfies a set of specific engineering constraints, imposed in this case, by the requirements of the decomposed state-estimation (DSE) in electrical power systems. The network-partitioning problem is classified as NP-hard problem. Although many heuristic algorithms have been proposed for its solution, these often lack directness and computational simplicity. In this thesis, three new partitioning techniques are described which (i) satisfy the DSE constraints, and (ii) simplify the NP-hard problem by using the natural graph properties of a network. The first technique is based on partitioning a spanning tree optimally using the natural property of the spanning tree branches. As with existing heuristic techniques, information on the partitioning is obtained only at the end of the partitioning process. The study of the DSE constraints leads to define conditions of an ideal balanced partitioning. This enables data on the balanced partitioning to be obtained, including the numbers of boundary nodes and cut-edges. The second partitioning technique is designed to obtain these data for a given network, by finding the minimum covering set of nodes with maximum nodal degree. Further simplification is then possible if additional graph-theoretical properties are used. A new natural property entitled the 'edge state phenomenon' is defined. The edge state phenomenon may be exploited to generate new network properties. In the third partitioning technique, two of these, the 'network external closed path' and the 'open internal paths', are used to identify the balanced partitioning, and hence to partition the network. Examples of the application of all three methods to network partitioning are provided.

The performance of various designs of outdoor noise barrier has been investigated using numerical modelling and full scale experiments. The numerical modelling has been performed using a two-dimensional boundary element method. The model has been extended to allow the efficient simulation of barrier arrangements on ground having two distinct impedance values and cross-sections incorporating cuttings. It has been reported previously that the performance of a plane screen can be enhanced by adding a device to the top of the barrier to induce destructive interference. Full scale modelling and boundary element simulations have been performed on one such commercially available device. It has been shown that, taking the height increase into account, the major contribution to the improved performance is the presence of two diffracting edges rather than any interference effects generated. It is known that the performance of a single barrier is degraded following the introduction of a barrier on the opposite side of a source. Boundary element simulations of such parallel arrangements have been performed. Modifications have been proposed to reduce the over-estimation of multiple reflections within the model, together with a method for converting predictions to the equivalent point source values. Sound absorptive, tilted and median barriers have been shown to be effective in reducing the degradation. A multiple-edge barrier configuration is known to offer improved screening performance over a plane screen. Reported in-situ measurements have suggested the behaviour to be influenced by site geometry. Boundary element calculations have been performed to identify a more efficient variant of the device. The results suggest the addition of an inclined base panel to be most effective. The boundary element model has been used to investigate the effect of shape and surface treatment upon railway noise barriers. The model has been adapted to allow the use of dipole sources characteristic of railway noise. The cross-section of the rolling stock has been shown to affect the performance of rigid barriers. If the upper edges are coincident, the results suggest that simple absorptive barriers provide better screening than tilted designs. The addition of multiple edges further enhances performance.

Evolution of measured profiles of densities, temperatures and velocities in the edge pedestal region between successive ELM (edge-localized mode) events are analyzed and interpreted in terms of the constraints imposed by particle, momentum and energy balance in order to gain insights regarding the underlying evolution of transport processes in the edge pedestal between ELMs in a series of DIII-D discharges. The data from successive inter-ELM periods during an otherwise steady-state phase of the discharges were combined into a composite inter-ELM period for the purpose of increasing the number of data points in the analysis. These composite periods were partitioned into sequential intervals to examine inter-ELM transport evolution. The GTEDGE integrated modeling code was used to calculate and interpret plasma transport and properties during each interval using particle, momentum, and energy balance. Variation of diffusive and non-diffusive (pinch) particle, momentum, and energy transport over the inter-ELM period are examined for discharges with plasma currents from 0.5 to 1.5 MA and inter-ELM periods from 50 to 220 ms. Diffusive transport is dominant for ρ< 0.925, while non-diffusive and diffusive transport are very large and nearly balancing in the sharp gradient region 0.925 <ρ <1.0. Transport effects of ion orbit loss are significant for ρ > 0.95, and are taken into account. During the inter-ELM period, diffusive transport increases slightly more than non-diffusive transport, increasing total outward transport. Both diffusive and non-diffusive transport have a strong inverse correlation with plasma current. Weakening the electromagnetic pinch may increase outward particle transport, and enable control over the rebuilding of the edge pedestal between ELMs.

<p>A quantum point contact (QPC) is a quasi-one dimensional electron system, for which the conductance is quantized in unit of $2e^2/h$. This conductance quantization can be explained in a simple single particle picture, where the electron density of states cancels the electron velocity to a constant. However, two significant features in QPCs were discovered in the past two decades, which have drawn much attention: the 0.7 effect in the linear conductance and zero-bias-anomaly (ZBA) in the differential conductance. Neither of them can be explained by single particle pictures.</p><p>In this thesis, I will present several electron correlation effects discovered in asymmetric QPCs, as shown below:</p><p>The linear conductance of our asymmetric QPCs shows conductance resonances. The number of these resonances increases as the QPC channel length increases. The quantized conductance plateau is also modulated by tuning the gate voltage of the QPCs. These two features, observed in the linear conductance, are ascribed to the formation of quasi-bound states in the QPCs, which is further ascribed to the electron-correlation-induced barriers. </p><p>The differential conductance for long channel QPCs shows the zero-bias-anomaly for every other linear conductance resonance valley, suggesting a near even-odd behavior. This even-odd law can be interpreted within the electron-correlation-induced barrier picture, where the quasi-localized non-zero spin in the quasi-bound state (Kondo-like) couples to the Fermi sea in the lead. For a specific case, triple-peak structure is observed in the differential conductance curves, while the electron filling number is still even, suggesting a spin triplet formation at zero magnetic field.</p><p>Small differential conductance oscillations as a function of bias voltage were discovered and systematically studied in an asymmetric QPC sample. These oscillations are significantly suppressed in a low in-plane magnetic field, which is completely unexpected. The oscillations are washed out when the temperature is increased to 0.8K. Numerical simulation, based on the thermal smearing of the Fermi distribution, was performed to simulate the oscillation behavior at high temperatures, using the low temperature data as an input. This simulation agrees with the oscillations off zero-bias region, but does not agree with the temperature evolution of the structure near zero-bias. Based on the above oscillation characteristics, all simple single particle pictures were carefully considered, and then ruled out. After exhausting all these pictures, we think these small oscillations are related to novel electronic and spin correlations.</p> / Dissertation

In the past few years, some experimental investigations have been conducted to verify seismic behaviour of fiber reinforced polymer reinforced concrete (FRP-RC) beam-column joints. Those researches were mainly focused on exterior beam-column joints without lateral beams. However, lateral beams, commonly exist in buildings, can significantly improve seismic performance of the joints. Moreover, the way the longitudinal beam bars are anchored in the joint, either using headed-end or bent bars, was not adequately addressed. This study aims to fill these gaps and investigate the shear capacity of FRP-RC exterior beam-column joints confined with lateral beams, and the effect of beam reinforcement anchorage on their seismic behaviour. Six full-scale exterior beam-column joints were constructed and tested to failure under reversal cyclic loading. Test results showed that the presence of lateral beams significantly increased the shear capacity of the joints. Moreover, replacing bent bars with headed-end bars resulted in more ductile behaviour of the joints.

Suburban Centres were established in the Toronto region as the population dispersed beyond the city’s borders. Intended as a set of delivery points for municipal services and concentrations of commercial and social program serving local suburban residents, government policy and market forces are now encouraging these centres to accumulate a greater range of program, and absorb a significant share of population growth. They have a mandate to orient new residents toward improved public transit routes as a relief for overburdened road infrastructure, but their fundamental role as a suburban downtown requires continued accessibility by car. The structure of the suburbs is fixed, dominated by the car as the primary element of an extensive mobility system that has generated its own spatial protocols and building typologies. The morphology of older urban areas was developed in response to the parameters of streetcar service and human abilities, and also shows a resistance to change. The two mobility systems co-exist, each with their own associated territories, creating an intermodal metropolis. In suburban centres, the intensive urban mobility extends into the reach of the suburban territory, creating a threshold condition that requires a hybrid morphology to serve both. The design adopts Scarborough Centre as a test site, proposing a morphology that accommodates urban and suburban mobility by embracing the suburban planning paradigm that separates vehicle traffic from public space. The interaction between the two networks is managed to create variations in accessibility characteristics that determine programmatic distribution. The public realm is compartmentalized into differentiated spaces that support a highly permeable pedestrian network integrated with the central transit station. The proposal allows Scarborough Centre to expand its public space network without compromising its function as a highly accessible suburban downtown.

In network design, a communication network is optimized for a given set of parameters like cost, reliability and delay. This study analyzes network design problem using Genetic Algorithms in detail and makes comparison of different approaches and representations. Encoding of a problem is one of the most crucial design choices in Genetic Algorithms. For network design problem, this study compares adjacency matrix representation with list of edges representation. Also, another problem is defining a fair fitness function that will not favor one optimization parameter to the other. Multi-objective optimization is a recommended solution for such problems. This study describes and compares some of those approaches for different combinations in network design problem.

Two approaches were developed for detecting earthquake damaged buildings from post-event aerial photographs using shadow analysis and perceptual grouping. In the first approach, it is assumed that the vector boundaries of the buildings are not known a priori. Therefore, only the post-event aerial photographs were used to detect the collapsed buildings. The approach relies on an idea that if a building is fully damaged then, it will not generate a closed contour. First, a median filter is applied to remove the noise. Then, the edge pixels are detected through a Canny edge detector and the line segments are extracted from the output edge image using a raster-to-vector conversion process. After that, the line segments are grouped together using a three-level hierarchical perceptual grouping procedure to form a closed contour. The principles used in perceptual grouping include the proximity, the collinearity, the continuity and the perpendicularity. In the second approach, it is assumed that the vector boundaries of the buildings are known a priori. Therefore, this information is used as additional data source to detect the collapsed buildings. First, the edges are detected from the image through a Canny edge detector. Second, the line segments are extracted using a raster-to-vector conversion process. Then, a two-level hierarchical perceptual grouping procedure is used to group these line segments. The boundaries of the buildings are available and stored in a GIS as vector polygons. Therefore, after applying the perceptual grouping procedure, the damage conditions of the buildings are assessed on a building-by-building basis by measuring the agreement between the detected line segments and the vector boundaries.

A method for detecting the buildings from high resolution aerial images is proposed. The aim is to extract the buildings from high resolution aerial images using the Hough transform and the model based perceptual grouping techniques.The edges detected from the image are the basic structures used in the building detection procedure. The method proposed in this thesis makes use of the basic image processing techniques. Noise removal and image sharpening techniques are used to enhance the input image. Then, the edges are extracted from the image using the Canny edge detection algorithm. The edges obtained are composed of discrete points. These discrete points are vectorized in order to generate straight line segments. This is performed with the use of the Hough transform and the perceptual grouping techniques. The straight line segments become the basic structures of the buildings. Finally, the straight line segments are grouped based on predefined model(s) using the model based perceptual grouping technique. The groups of straight line segments are the candidates for 2D structures that may be the buildings, the shadows or other man-made objects. The proposed method was implemented with a program written in C programming language. The approach was applied to several study areas. The results achieved are encouraging. The number of the extracted buildings increase if the orientation of the buildings are nearly the same and the Canny edge detector detects most of the building edges.If the buildings have different orientations,some of the buildings may not be extracted with the proposed method. In addition to building orientation, the building size and the parameters used in the Hough transform and the perceptual grouping stages also affect the success of the proposed method.

No description available.