This study presents an improved ghost-cell immersed boundary approach to represent a solid body in compressible flow simulations. In contrast to the commonly used approaches, in the present work ghost cells are mirrored through the boundary described using a level-set method to farther image points, incorporating a higher-order extra/interpolation scheme for the ghost cell values. In addition, a shock sensor is in- troduced to deal with image points near the discontinuities in the flow field. Adaptive mesh refinement (AMR) is used to improve the representation of the geometry efficiently. The improved ghost-cell method is validated against five test cases: (a) double Mach reflections on a ramp, (b) supersonic flows in a wind tunnel with a forward- facing step, (c) supersonic flows over a circular cylinder, (d) smooth Prandtl-Meyer expansion flows, and (e) steady shock-induced combustion over a wedge. It is demonstrated that the improved ghost-cell method can reach the accuracy of second order in L1 norm and higher than first order in L∞ norm. Direct comparisons against the cut-cell method demonstrate that the improved ghost-cell method is almost equally accurate with better efficiency for boundary representation in high-fidelity compressible flow simulations. Implementation of the improved ghost-cell method in reacting Euler flows further validates its general applicability for compressible flow simulations.

In order to achieve results that are reliable when using the finite element method one has to use an acceptable element mesh with respect to the shape and size of the elements. As a help to produce an acceptable mesh there are quality criteria that must be fulfilled in most pre-processors. One objective with this thesis is to perform a sensitivity study that can be used as a basis for a Mesh guideline for chassis parts which is requested from engineers at Volvo 3P. The software used in the sensitivity study is ANSA as pre-processor, Nastran as solver and Metapost as post-processor. In the first part of the sensitivity study three different models are used for studying quality criteria such as aspect ratio, skewness, mid point alignment, mid point deviation and element size. Solid elements of second order, which are used in the three models, can be generated in two ways, which constitutes another part of the sensitivity study. They may either be generated from the beginning or can be converted from first order elements. This means geometrically that if second order elements where generated from the beginning the element mesh would follow the shape of the component in a better way compared to the other method. Recently a pre- and post-processing program called SimLab was introduced on the market. Since SimLab supports geometry import from several CAD-systems without loss of feature information, the automatic element mesh generation is supposed to be better as the mesh generator has access to more information concerning the geometry. An evaluation of SimLab is the second major objective of the thesis. More specifically, the evaluation concerns the possibility of using the software at Volvo 3P. Results show a surprising insensitivity regarding the criteria and that the method of generating second order elements from the beginning is to be preferred. SimLab is a new program with big potential and the conclusion is that it is possible to use it at Volvo 3P.

Computational Fluid Dynamics (CFD) is a main field that contributes to the development of high efficiency aircraft. CFD accuracy depends on the flow solver and the meshing of the geometry, and while it is doable to determine why a certain solver is more accurate than another, it is much more difficult to discern why two meshes produce differently accurate solutions. A framework is presented to evaluate the performance or ``goodness" of a mesh and to compare meshes. The framework is composed of quantifiable mesh parameters which define a mesh, and three performance measures: functional accuracy, their order of convergence, and their behaviour under the adjoint correction method. Although it seems that the relationships between parameters and results are not trivial, there are trends from which optimal mesh parameters are deduced. The H topology performs best, and the most important parameters are related to spacings and cell quality around the aerofoil leading edge.

Computational Fluid Dynamics (CFD) is a main field that contributes to the development of high efficiency aircraft. CFD accuracy depends on the flow solver and the meshing of the geometry, and while it is doable to determine why a certain solver is more accurate than another, it is much more difficult to discern why two meshes produce differently accurate solutions. A framework is presented to evaluate the performance or ``goodness" of a mesh and to compare meshes. The framework is composed of quantifiable mesh parameters which define a mesh, and three performance measures: functional accuracy, their order of convergence, and their behaviour under the adjoint correction method. Although it seems that the relationships between parameters and results are not trivial, there are trends from which optimal mesh parameters are deduced. The H topology performs best, and the most important parameters are related to spacings and cell quality around the aerofoil leading edge.

Service providers’ goals include providing reliable connections with the minimum allocated resources over a shared-mesh path restoration scheme in WDM networks. However, in some cases, the requested parameters in an SLA are beyond the capacity of the network, and the connection is typically blocked. To give the customer a chance to choose another provider, or in the case of having only one provider, to comply with the provider’s network capacity, new SLA-based architectures and mechanisms are required to be introduced to provide better service to prioity-aware shared mesh WDM networks. To achieve this goal, the dissertation’s contributions focus on three main characteristics of the network design: i) A dynamic SLA negotiation infrastructure to negotiate and propagate crucial SLA parameters, ii) Path attributes which can provide a better picture of network resources and status and are suitable to be propagated by the negotiating system, and iii) Algorithms benefiting from the path attributes to improve the blocking probability and resource utilization of the network. To fulfill the first goal of the contributions, a dynamic SLA negotiation mechanism for both intra and inter-domain communications using OSPF and BGP protocols is proposed. Link attributes via intra-domain, and new proposed TE path attributes through inter-domain mechanisms are advertised. Several novel path constraints and attributes are proposed which are dynamically updated and propagated through the network over the connections provisioning process period to satisfy the second objective of the contributions in this dissertation. The path availability, holding time, SLA violation risk, and path risk factor are the important characteristics of the proposed path attributes. As the third goal considered for the contributions, novel priority-aware algorithms and SLA-based mechanisms are proposed to improve the network performance for different traffic types of various priority classes. The algorithms and mechanisms proposed in this thesis take advantage of the new path attributes and SLA negotiation infrastructure to better serve high-priority connection requests at the lowest cost. The mechanisms and network architectures proposed in this work are a solution for the high-priority requests that normally cannot be accommodated as they violate the best availability offered by service providers.

With continually growing availability of high performance computing resources, the finite element methods (FEM) are becoming increasingly more efficient and practical research tools. In the domain of computational biomechanics, FEMs have been successfully applied in investigation of biomedical problems that include impact and fracture mechanics of bone, load transmission through the joints, feasibility of joint replacements, and many others. The present research study was concerned with the development of a detailed, anatomically accurate, finite element model of the human hand and wrist. As a first step in this direction, we used a publically available database of wrist bone anatomy and carpal kinematics to construct a multi-body statistical shape model (SSM) of the wrist. The resulting model provides an efficient parameterization of anatomical variations of the entire training set and can thus overcome the major shortcoming of conventional biomechanical models associated with limited generalization ability. The main contributions of this work are: 1) A robust method for constructing multi-body SSM of the wrist from surface meshes. 2) A novel technique for resampling closed genus-0 meshes to produce high quality triangulations suitable for finite element simulations. Additionally, all techniques developed in the course of this study could be directly applied to create an equivalent model of the tarsus.

Wireless networks are less reliable than wired networks because channels are “exposed” to the surrounding environment that is susceptible to interference and noise. To minimize losses of data due to collisions, wireless networks need a mechanism to regulate the access on the transmission medium. Medium Access Control (MAC) protocols control access to the shared communication medium so that it can be used efficiently. In this thesis, we first describe the collision-controlled Dual Prohibition Multiple Access (DPMA) protocol [45]. The main mechanisms implemented in DPMA, such as binary dual prohibition, power control, interference control, and support for differentiated services (DiffServ), are presented in detail. We conducted a thorough simulation study on DPMA protocol from several aspects. First, we conduct simulations to observe the effects of binary competition number (BCN), unit slot length and safe margin on the performance of DPMA. Secondly, the DiffServ capability of DPMA is demonstrated through simulation results. Finally, we compare the DPMA protocol with the CSMA/CA protocol and find that DPMA with optimal configuration has better performance than CSMA/CA under both low and high network density. / Thesis (Master, Electrical & Computer Engineering) -- Queen's University, 2007-09-28 16:25:02.515

Mémoire numérisé par la Division de la gestion de documents et des archives de l'Université de Montréal

A novel electrode was constructed and tested in a bio-battery. This configuration consisted of a composite electrode with immobilized bacteria (Escherichia coli K-12) in the anode and a composite electrode with immobilized Carbon Nanoparticles (CNP) and algae (Chlorella vulgaris/Scenedesmus sp.) suspended in the cathode. The composite electrode consisted of three parts: a 304L stainless steel mesh base, an electro-polymerized layer of pyrrole, and an electro-polymerized layer of methylene blue. The bacteria were immobilized on the anode electrode using a technique incorporating CNP and a Teflontm emulsion. The anode and cathode electrodes were tested separately in conjunction with chemical cathodes and anodes respectively. The composite electrode with immobilized bacteria was tested in a bioanode setup. The cathode chamber of the cell contained a potassium ferricyanide and buffer solution with a graphite electrode. Factors affecting electrode performance, such as Teflontm and carbon nanoparticle concentration, were investigated to find optimum values. The maximum power density generated by the composite electrode with immobilized bacteria and a chemical cathode was 378 mW/m2. This electrode configuration produced approximately 69% more power density and 53% more current density than composite electrodes with bacteria suspended in solution. Electrochemical Impedance Spectroscopy analysis determined that a significant portion of the bio-battery’s resistance to charge transfer occurred at the surface of the anode and this resistance was significantly lowered when using immobilized bacteria (51% lower than bio-batteries with suspended bacteria). Similarly, biocathodes containing composite electrodes coated with CNP were tested using two algae species, Chlorella vulgaris and Scenedesmus sp., suspended in solution. This electrode configuration was compared with composite electrode without CNP coating. The anode chamber contained potassium ferrocyanide solution with a graphite counter electrode. The composite electrode with CNP produced approximately 23% more current density than composite electrode without CNP. A complete bio-battery was designed using a composite electrode with immobilized bacteria anode and a CNP coated composite electrode with algae suspended in the cathode. EIS analysis showed that the resistance was higher in the biocathode than in the bioanode and a significant portion of the ohmic resistance was contributed by the membrane.

This dissertation presents a multilevel algorithm to solve constant and variable coeffcient elliptic boundary value problems on adaptively refined structured meshes in 2D and 3D. Cacheaware algorithms for optimizing the operations to exploit the cache memory subsystem areshown. Keywords: Multigrid, Cache Aware, Adaptive Mesh Refinement, Partial Differential Equations, Numerical Solution.