The rising interest in fan performance in the presence of total pressure distortion, a topic of fundamental interest for integrated airframe/engine architectures, has led to increased research in ground based testing environments. Included in these studies is the generation of simulated total pressure distortion profiles using wire mesh screens. Although the inlet duct development of total pressure distortion patterns has been studied in a historical context when distortion effects on engine performance were first of interest, these were typically simplified experimental studies for low-speed flows. To aid in the understanding of total pressure distortion development approaching a transonic fan face, a series of experiments were conducted to detail the development of such a profile downstream of the screen plane in the absence and presence of fan effects. Presented is an extensive experimental set to detail (1) the evolution of a screen-generated total pressure distortion profile as it develops in a constant diameter inlet duct and (2) the effect that a single stage transonic fan has on the distortion development. Included is a detailed analysis of the distortion profile characteristics for increasing development length, and the behavioral changes of the profile when fan blockage is present near the screen plane. Recommendations are made regarding the placement of total pressure distortion screens relative to the fan face, and insights are given into the expected profile evolution. This work is a contributing part of an ongoing systematic investigation of fan performance when subjected to screen-generated total pressure inlet distortion. / Master of Science

<p>Packet transmission scheduling plays a key role in Quality of Service (QoS) support for real-time traffic and efficient radio resource utilization in a wireless mesh network (WMN). It is a highly complicated problem due to the fact that any scheduling decision at one mesh access point (AP) may affect the scheduling decisions in the entire network. The strict delay requirement of real-time applications makes the scheduling problem even more challenging.</p> <p> In this thesis, the packet transmission scheduling problem for real-time constant-bit-rate (CBR) traffic in a WMN is first formulated as a standard integer linear programming problem, which takes into consideration both the multihop packet transmission delay and timeline coordinations of the mesh APs. The objective is to efficiently utilize the radio resources, subject to available bandwidth of the mesh APs, co-channel interference, and packet transmission latency requirement.</p> <p>Two heuristic schemes, namely AP-based scheduling (ABS) and connection-based scheduling (CBS) schemes, are then proposed to support real-time CBR traffic. ABS makes scheduling decisions on a per-AP basis. Scheduling decisions at APs with a higher traffic load are determined before those at APs with a lower traffic load. ABS achieves close-to-optimum capacity but may go through multiple iterations before reaching a feasible solution. CBS makes scheduling decisions on a connection-by-connection basis. It gives a higher priority to connections with more hops. In CBS, connections with a lower priority can only use resources remaining from serving all higher priority connections. CBS requires much lower complexity than ABS while achieving capacity performance slightly lower than ABS.</p> <p>We extend the proposed ABS and CBS scheduling schemes for supporting real-time variable bit rate (VBR) traffic in a WMN. By combining the concept of effective bandwidth and the proposed scheduling schemes, both delay and packet loss performance of the VBR traffic can be effectively satisfied. The scheduling schemes are further extended for supporting real-time traffic in a WMN with multi-radio APs.</p> <p>All the scheduling decisions are done at the time when new connection requests arrive and the results are used to make admission control decisions. In this sense, the work in this thesis is for both packet transmission scheduling and admission control for real-time traffic in WMNs.</p> / Thesis / Doctor of Philosophy (PhD)

<p> Wireless mesh networks are now being used to deploy radio coverage in a large variety of outdoor applications. One of the major obstacles that these networks face is that of providing the nodes with electrical power and wired network connections. Solar powered mesh nodes are increasingly used to eliminate the need for these types of connections, making the nodes truly tether-less. In these types of networks however, the cost of the energy collection and storage components can be a significant fraction of the total node cost, which motivates a careful selection of these resources.</p> <p> This thesis focusses on key issues relating to the deployment and operation of solar powered wireless mesh networks. First, the problem of provisioning the mesh nodes with a suitable solar panel and battery configuration is considered. This is done by assuming a bandwidth usage profile and using historical solar insolation data for the desired deployment location. A resource provisioning algorithm is proposed based on the use of temporal shortest-path routing and taking into account the node energy-flow for the target deployment time period. A methodology is introduced which uses a genetic algorithm (GA) to incorporate energy-aware routing into the resource assignment procedure. Results show that the proposed resource provisioning algorithm can achieve large cost savings when compared to conventional provisioning methods.</p> <p> During post-deployment network operation, the actual bandwidth profile and solar insolation may be different than that for which the nodes were originally provisioned. To prevent node outage, the network must reduce its workload by flow controlling its input traffic. The problem of admitting network bandwidth flows in a fair manner is also studied. A bound is first formulated which achieves the best max/min fair flow control subject to eliminating node outage. The bound motivates a proposed causal flow control algorithm whose operation uses prediction based on access to on-line historical weather data. The results show that the proposed algorithm performs well when compared to the analytic bound that is derived for this problem.</p> <p> Finally, as user traffic evolves, the network resources need to be updated. This problem is considered using a minimum cost upgrade objective. A mixed integer linear programming (MILP) formulation is derived to obtain a lower bound on the network update cost. A genetic algorithm is used to determine practical cost-effective network resource upgrading. The results show that the proposed methodology can obtain significant cost savings.</p> / Thesis / Doctor of Philosophy (PhD)

<p> This thesis details work done in the development of the phase field model which allows simulation of elasticity with diffuse interfaces and the extension of a thin interface analysis developed by previous authors to study non-dilute ideal alloys. These models are coupled with a new finite difference adaptive mesh algorithm to efficiently simulate a variety of physical systems. The finite difference adaptive mesh algorithm is shown to be at worse 4-5 times faster than an equivalent finite element method on a per node basis. In addition to this increase in speed for explicit solvers in the code, an iterative solver used to compute elastic fields is found to converge in O(N) time for a dynamically growing precipitate, where N is the number of nodes on the adaptive mesh. A previous phase field formulation is extended such as to make possible the study of non-ideal binary alloys with complex phase diagrams. A phase field model is also derived for a free energy that incorporates an elastic free energy and is used to investigate the competitive development of solid state structures in which the kinetic transfer rate of atoms from the parent phase to the precipitate phase is large. This results in the growth of solid state dendrites. The morphological effects of competing surface anisotropy and anisotropy in the elastic modulus tensor is analyzed. It is shown that the transition from surfaceenergy driven dendrites to elastically driven dendrites depends on the magnitudes of the surface energy anisotropy coefficient (E4 ) and the anisotropy of the elastic tensor (β) as well as on the super saturation of the particle and therefore to a specific Mullins-Sekerka onset radius. The transition point of this competitive process is predicted from these three controlling parameters. </p> / Thesis / Doctor of Philosophy (PhD)

The rapid advances in networking technology have enabled large-scale deployments of online video streaming services in today's Internet. In particular, wireless Internet access technology has been one of the most transforming and empowering technologies in recent years. We have witnessed a dramatic increase in the number of mobile users who access online video services through wireless access networks, such as wireless mesh networks and 3G cellular networks. Unlike in wired environment, using a dedicated stream for each video service request is very expensive for wireless networks. This simple strategy also has limited scalability when popular content is demanded by a large number of users. It is desirable to have a robust wireless access environment that can sustain a sudden spurt of interest for certain videos due to, say a current event. Moreover, due to the mobility of the video users, smooth streaming performance during the handoff is a key requirement to the robustness of the wireless access networks for mobile video users. In this dissertation, the author focuses on the robustness of the wireless mesh access (WMA) environment for mobile video users. Novel video sharing techniques are proposed to reduce the burden of video streaming in different WMA environments. The author proposes a cross-layer framework for scalable Video-on-Demand (VOD) service in multi-hop WiMax mesh networks. The author also studies the optimization problems for video multicast in a general wireless mesh networks. The WMA environment is modeled as a connected graph with a video source in one of the nodes and the video requests randomly generated from other nodes in the graph. The optimal video multicast problem in such environment is formulated as two sub-problems. The proposed solutions of the sub-problems are justified using simulation and numerical study. In the case of online video streaming, online video server does not cooperate with the access networks. In this case, the centralized data sharing technique fails since they assume the cooperation between the video server and the network. To tackle this problem, a novel distributed video sharing technique called Dynamic Stream Merging (DSM) is proposed. DSM improves the robustness of the WMA environment without the cooperation from the online video server. It optimizes the per link sharing performance with small time complexity and message complexity. The performance of DSM has been studied using simulations in Network Simulator 2 (NS2) as well as real experiments in a wireless mesh testbed. The Mobile YouTube website (http://m.youtube.com) is used as the online video website in the experiment. Last but not the least; a cross-layer scheme is proposed to avoid the degradation on the video quality during the handoff in the WMA environment. Novel video quality related triggers and the routing metrics at the mesh routers are utilized in the handoff decision making process. A redirection scheme is also proposed to eliminate packet loss caused by the handoff.

The creation of 3D models of bone from CT images has become popular for surgical planning, the design of implants, and educational purposes. Software is available to convert CT images into 3D models of bone, however, these can be expensive and technically taxing. The goal of this project was to create an open-source and easy-to-use methodology to create 3D models of bone and allow the user to interact with the model to extract desired regions. The method was first created in MATLAB and ported to Python. The CT images were imported into Python and the images were then binarized using a desired threshold determined by the user and based on Hounsfield Units (HU). A Canny edge detector was applied to the binarized images, this extracted the inner and outer surfaces of the bone. Edge points were assigned x, y, and z coordinates based on their pixel location, and the location of the slice in the stack of CT images to create a 3D point cloud. The application of a Delaunay tetrahedralization created a mesh object, the surface was extracted and saved as an STL file. An add-on in Blender was created to allow the user to select the CT images to import, set a threshold, create a 3D mesh model, draw an ROI on the model, and extract that region based on the desired thickness and create a new 3D object. The method was fully open-sourced so was inexpensive and was able to create models of a skull and allow the segmentation of portions of that mesh to create new objects. Future work needs to be conducted to improve the quality of the mesh, implement sampling to reduce the time to create the mesh, and add features that would benefit the end-user. / Thesis / Master of Applied Science (MASc) / The creation of 3D models of bone from CT images has become popular for education, surgical planning, and the design of implants. Software is available to convert CT images into 3D models but can be expensive and technically taxing. The purpose of this project was to develop a process to allow surgeons to create and interact with models from imaging data. This project applied a threshold to binarize a set of CT images, extracted the edges using a Canny Edge detector, and used the edge pixels to create a 3D point cloud. The 3D point cloud was then converted to a mesh object. A user interface was implemented that allowed the selection of portions of the model and a new 3D model to be created from the selection. The process can be improved by improving the quality of the mesh output and adding features to the user interface.

Surface mesh generation is a critical component of the mesh generation process. The objective of the described effort was to determine if a combination of constrained Delaunay triangulation (for triangles), advancing front method (for quadrilaterals), curvature-based refinement, smoothing, and reconnection is a viable approach for discretizing a NURBS patch holding the boundary nodes fixed. The approach is significant when coupled with recently developed geometry specification that explicitly identifies common edges. This thesis describes the various techniques used to achieve the above objectives. Application of this approach to several representative geometries demonstrates that it is an effective alternative to traditional approaches.

The primary objective of this study is to develop a sliding interface method for simulations involving relative rotational grid motion suitable for unstructured grid topologies. The present method alleviates computationally expensive grid deformation, remeshing, and hole cutting procedures. Rotational motion is accomplished by rigidly rotating a subdomain representing the moving component. At the subdomain interface boundary, the faces along the interfaces are extruded into the adjacent subdomain to create new volume elements and provide a one-cell overlap. These new volume elements close the control volumes for the nodes on the interface surface and allow a flux to be computed across the subdomain interface. An interface flux is computed independently for each subdomain. The values of the solution variables and other quantities for the nodes created by the extrusion process are found by interpolation. The extrusion is done so that the interpolation will maintain information as localized as possible. A parallel implementation of the neighbor search is used to find the extruded points in the adjacent subdomain. The method has been implemented in a parallel, node-centered finite volume, high-resolution viscous flow solver. The method does not impose any restrictions on the subdomain interface aside from the axisymmetric limitation required for rotational motion. In addition, the grid on the subdomain interface is arbitrary. The boundary surfaces between the two subdomains can have independent grids from one another. They do not have to connect in a one-to-one manner and there are no symmetry or pattern restrictions placed on the surface grid. A variety of numerical simulations were performed on several small-scale model problems to examine conservation of the interface flux. Overall flux conservation errors were found to be comparable to that for fully connected and fully conservative simulations. In addition, excellent agreement was obtained with both theoretical and experimental results. Three large-scale applications were also used to validate the method and highlight some of the advantages of the sliding interface method compared to the current state-of- the-art for unstructured grid applications. This sliding interface method requires no geometric modifications and has significantly shorter run times Furthermore, there were no apparent adverse effects on the numerical solutions by not strictly enforcing flux conservation at the subdomain boundary.

No description available.

No description available.