Este trabalho propõe um modelo hierárquico para coleta de dados de monitoramento em sistemas distribuídos. Seu objetivo é proporcionar a análise on-line do comportamento de sistemas e programas distribuídos. O meio escolhido para realizar essa análise foi a visualização. Inicialmente é apresentada uma contextualização sobre monitoramento de sistemas distribuídos. Também são abordados aspectos específicos ao monitoramento de Grid. Após, é analisado um conjunto de ferramentas de monitoramento. Então tem-se a apresentação do modelo proposto. Esse é composto por coletores locais, por uma hierarquia de agregadores e por clientes. É utilizado o modelo push de transmissão de dados e há um mecanismo de subscrição aos coletores. Foi implementado um protótipo do modelo de coleta proposto, que foi utilizado na implementação de um protótipo de ferramenta de monitoramento on-line. Nessa, os dados coletados são fornecidos ao DIMVisual, que é um modelo de integração de dados para visualização. Para visualização, o protótipo utiliza a ferramenta TRIVA, que recebe os dados integrados como entrada. Essa ferramenta foi modificada para gerar uma visualização que é atualizada de maneira on-line. Também foram realizados experimentos para avaliar o tempo necessário para enviar mensagens com diferentes hierarquias e configurações dos coletores. Além disso, foi avaliada a capacidade de o cliente implementado processar os dados recebidos, gerando sua visualização. / This work proposes a hierarchical model for collecting monitoring data from distributed systems. Its goal is to allow the on-line analysis of the behavior of distributed systems and applications. The means we chose to perform this analysis is to generate a visualization of the collected information. In the beginning of this dissertation we present an overview of the monitoring of distributed systems. Aspects that are specific to the monitoring of Grid systems are also reviewed. Next, we have an analysis of a set of monitoring tools. Then we present the proposed model, which is composed by local collectors, an hierarchical structure of aggregators and clients. A push data transmission model is used in the model and it also has a subscription mechanism. A prototype monitoring tool was implemented, integrating the data collection model with DIMVisual and TRIVA. The former is a data integration model whose output is formatted to be used as input for a visualization tool. The later is a visualization tool which, in the prototype, receives the integrated data from DIMVisual. TRIVA generates a visualization of the received information, which is updated in an on-line fashion. In order to evaluate the model, we performed a set of experiments using the prototype. One of the experiments measured the time spent to send data though different hierarchies. In these tests we have also varied the quantity and the configuration of the collectors. In another experiment we evaluated the capacity of the client to process the received data.

A Distributed Multi-Input Multi-Output (DMIMO) system consists of many transceivers coordinating themselves into a "virtual antenna array" in order to emulate MIMO capabilities. In recent years, the field of research investigating DMIMO Communications has grown substantially. DMIMO systems offer all of the same benefits of standard MIMO systems on a larger scale because arrays are not limited by the physical constraint of placing many antennas on a single transceiver. This additional benefit does come at a cost, however. Since nodes are distributed and run from independent clock signals and with unknown geometry, each one must its own obtain channel state information (CSI) to the target nodes. In existing DMIMO architectures, array nodes depend on feedback from target nodes to properly synchronize. This means that target nodes must be cooperative and are responsible for the overhead calculating and transmitting CSI feedback to each node in the array. Within this work, we develop a set of techniques for Retrodirective Distributed Antenna Arrays. Retrodirective arrays have traditionally been used to direct a beam towards a target node, but the work in this thesis seeks to develop a more generalized definition of retrodirectivity. By our definition, a retrodirective array is one that acquires CSI to one or more intended targets simply by listening to the incoming transmissions of those targets; the array may subsequently use this information to do any number of typical MIMO tasks (i.e., beamforming, nullforming, spatial multiplexing, etc.). We explore two primary techniques: i) distributed beamforming and ii) distributed nullforming. Beamforming involves focusing transmitted power towards a specific target node and nullforming involves directing transmissions of array nodes to cancel one another at a specific target node. We focus on these techniques because they can be thought of as basic building blocks for more sophisticated DMIMO techniques. We first develop the theory for retrodirective arrays. Then, we present an architecture for the implementation of this theory. Specifically, we focus on the pre-synchronization of the array, which involves use of a master/slave architecture and a timeslotted message exchange among the array nodes. Finally, developing algorithms to make these arrays both robust and scalable is the focus of this thesis.

A fundamental issue in many distributed computing problems is the need for nodes to distinguish themselves from their neighbors in a process referred to as symmetry breaking. Many well-known problems such as Maximal Independent Set (MIS), t-Ruling Set, Maximal Matching, and (\Delta+1)-Coloring, belong to the class of problems that require symmetry breaking. These problems have been studied extensively in the LOCAL model, which assumes arbitrarily large message sizes, but not as much in the CONGEST and k-machine models, which assume messages of size O(log n) bits. This dissertation focuses on finding upper and lower bounds for symmetry breaking problems, such as MIS and t-Ruling Set, in these congested models. Chapter 2 shows that an MIS can be computed in O(sqrt{log n loglog n}) rounds for graphs with constant arboricity in the CONGEST model. Chapter 3 shows that the t-ruling set problem, for t \geq 3, can be computed in o(log n) rounds in the CONGEST model. Moreover, it is shown that a 2-ruling set can be computed in o(log n) rounds for a large range of values of the maximum degree in the graph. In the k-machine model, k machines must work together to solve a problem on an arbitrary n-node graph, where n is typically much larger than k. Chapter 4 shows that any algorithm in the BEEP model (which assumes 'primitive' single bit messages) with message complexity M and round complexity T can be simulated in O(t(M/k^2 + T) poly(log n)) rounds in the k-machine model. Using this result, it is shown that MIS, Minimum Dominating Set (MDS), and Minimum Connected Dominating Set (MCDS) can all be solved in O(poly(log n) m/k^2) rounds in the k-machine model, where 'm' is the number of edges in the input graph. It is shown that a 2-ruling set can be computed even faster, in O((n/k^2+ k) poly(log n)) rounds, in the k-machine model. On the other hand, using information theoretic techniques and a reduction to a communication complexity problem, an \Omega(n/(k^2 poly(log n))) rounds lower bound for MIS in the k-machine model is also shown. As far as we know, this is the first example of a lower bound in the k-machine model for a symmetry breaking problem. Chapter 5 focuses on the Max Clique problem in the CONGEST model. Max Clique is trivially solvable in one round in the LOCAL model since each node can share its entire neighborhood with all neighbors in a single round. However, in the CONGEST model, nodes have to choose what to communicate and along what communication links. Thus, in a sense, they have to break symmetry and this is forced upon them by the bandwidth constraints. Chapter 5 shows that an O(n^{3/5})-approximation to Max Clique in the CONGEST model can be computed in O(1) rounds. This dissertation ends with open questions in Chapter 6.

Oregon utilities are replacing their portfolios of traditional fossil fuel generation with renewable generating sources. Stepping away from carbon-producing energy will leave a deficit of on-demand power, resulting in decreased reliability. To overcome these technical challenges, utilities must maximize the use of their present dispatchable resources. One such resource is the Portland General Electric (PGE) Dispatchable Standby Generation Program (DSG), which is an aggregated 105 MWs of distributed generation (DG). These resources are brought on-line when there is a critical need for power. Resources are added to the program if a transfer trip scheme is in place or a modeling study reveals that the feeder load is at least three times the generator capacity. If the load-to-capacity ratio were lower, more assets could be added to the DSG program. To investigate the impacts of lowering the DG load-to-capacity ratio on existing distribution feeders, we use Open-Source Distribution System Simulator (OpenDSS). We modeled the Oxford Rural feeder by converting a utility CYME database to instantiation files using several MATLAB programs. A MATLAB control program varies the load-to-capacity ratio of the OpenDSS feeder model and monitors the generator behavior immediately following a fault. We analyzed the results to determine the ideal load-to-capacity ratio that prevents unintentional islanding. The results show that the instantaneous (50) relay element settings dictate both the minimum load-to-capacity ratio and the maximum DG capacity. The present three-to-one ratio is very conservative and can be reduced. Additional dispatchable resources include a five MW battery-inverter system currently used as grid-back up. The battery is grid-tied to a 12.4 kV feeder making it an ideal candidate for conservation voltage reduction (CVR). Using the same feeder model, we investigated the effects of lowering the system voltage to the allowable minimum using injections of reactive power. A lower system voltage reduces the load at peak times. Conversely, increasing the voltage prevents generation conflicts. To determine the benefit of CVR by VAr-injection on the Oxford Rural feeder, we created a MATLAB optimization program to output the optimal feeder voltage for reduced system power. We use a Simulink feedback model to determine the appropriate reactive power needed to achieve the voltage change. We analyze the system model to reveal that the feeder is ideal for CVR but the system capacity must be increased to achieve the maximum power reduction.

As global climate change concerns, technological advancements, and economic shifts increase the adoption of electric vehicles, it is vital to study how best to integrate these into our existing energy systems. Electric vehicles (EVs) are on track to quickly become a large factor in the energy grid. If left uncoordinated, the charging of EVs will become a burden on the grid by increasing peak demand and overloading transformers. However, with proper charging control strategies, the problems can be mitigated without the need for expensive capital investments. Distributed control methods are a powerful tool to coordinate the charging, but it will be important to assess the trade-offs between performance, information privacy, and computational speed between different control strategies. This work presents a comprehensive comparison between four distributed control algorithms simulating two case studies constrained by dynamic transformer temperature and current limits. The transformer temperature dynamics are inherently nonlinear and this implementation is contrasted with a piece-wise linear convex relaxation. The more commonly distributed control methods of Dual Decomposition and Alternating Direction Method of Multipliers (ADMM) are compared against a relatively new algorithm, Augmented Lagrangian based Alternating Direction Inexact Newton (ALADIN), as well as against a low-information packetized energy management control scheme (PEM). These algorithms are implemented with a receding horizon in two distinct case studies: a local neighborhood scenario with EVs at each network node and a hub scenario where each node represents a collection of EVs. Finally, these simulation results are compared and analyzed to assess the methods’ performance, privacy, and processing metrics for each case study as no algorithm is found to be optimal for all applications.

In this dissertation, we propose a novel distributed approach to the control of generators in the electric grid. Specifically, we consider the problem of the optimal economic dispatch of generator; we present a simple, distributed algorithm, which adjusts the power-frequency set-points of generators to correct for power imbalances arising from generation and load fluctuations. In this algorithm each generator independently adjusts its real-power output based on its estimate of the aggregate power imbalance in the network; such as an estimate can be independently obtained by each generator through local measurements of the frequency deviation on the grid. Eventually, over the course of network operation, the distributed algorithm achieves the equal-marginal-cost power allocation among generators while driving the power imbalance exponentially to zero. In the absence of power losses, we prove the eventual optimality of the distributed algorithm under mild assumptions (strict convexity and positivity of cost functions) and present simulation results to compare its performance with traditional (centralized) dispatch algorithms. Furthermore, we present numerical simulation results that show that the distributed algorithm performs well even in the presence of power losses and other constraints. We argue that distributed control methods are especially attractive for electric grids with smart meters and other advanced capabilities at the end node and grids with high penetration of alternative energy generators and we identify interesting open problems for future work in this area.

In the last decade, spatially distributed hydrological models have rapidly advanced with the widespread availability of remotely sensed and geomatics information. Particularly, the areas of calibration and evaluation of spatially distributed hydrological models have been attempted in order to reduce the differences between models and improve realism through various techniques. Despite steady efforts, the study of calibrations and evaluations for spatially distributed hydrological models is still a largely unexplored field, in that there is no research in terms of the interactions of snow and water balance components with the traditional measurement methods as error functions. As one of the factors related to runoff, melting snow is important, especially in mountainous regions with heavy snowfall; however, no study considering both snow and water components simultaneously has investigated the procedures of calibration and evaluation for spatially distributed models. Additionally, novel approaches of error functions would be needed to reflect the characteristics of spatially distributed hydrological models in the comparison between simulated and observed values. Lastly, the shift from lumped model calibration to distributed model calibration has raised the model complexity. The number of unknown parameters can rapidly increase, depending on the degree of distribution. Therefore, a strategy is required to determine the optimal degree of model distributions for a study basin. In this study, we will attempt to address the issues raised above. This study utilizes the Research Distributed Hydrological Model (HL-RDHM) developed by Hydrologic Development Office of the National Weather Service (OHD-NWS). This model simultaneously simulates both snow and water balance components. It consists largely of two different modules, i.e., the Snow 17 as a snow component and the Sacramento Soil Moisture Accounting (SAC-SMA) as a water component, and is applied over the Durango River basin in Colorado, which is an area driven primarily by snow. As its main contribution, this research develops and tests various methods to calibrate and evaluate spatially distributed hydrological models with different, non-commensurate, variables and measurements. Additionally, this research provides guidance on the way to decide an appropriate degree of model distribution (resolution) for a specific water catchment.

Governments and utility organizations around the world have mandated and provided incentives for new distributed renewable energy generation (DREG) capacity, and market projections indicate strong growth in distributed renewable energy generation installations in the coming years. New distributed renewable energy generation utilities, by definition, will be primarily located in built environments near consumers; these utilities are often planned and designed by landscape architects, yet no evidence-based, distributed renewable energy generation research is explicitly done by landscape architects or recognizes the role landscape architects play in planning and designing these spaces. The research and analysis provided by this study indicates that distributed renewable energy generation lacks a strong foundation as an independent concept which could benefit from clear broad phraseology linked to organized sub-terms/phrases for specific forms of DREG, that there has been some research done on topics familiar to landscape architects, that more needs to be done to meet important research questions and recommendations already posed, and that landscape architects are positioned to contribute to future distributed renewable energy generation research.

[No Abstract]

Garbage collectors (GCs) automate the problem of deciding when objects are no longer reachable and therefore should be reclaimed, however, there currently exists no automated process for the design of a correct garbage collector. Formal models exist that prove the correctness of individual GCs; more general models describe a wider range of GCs but do not prove their correctness or provide a concrete instantiation process. The lack of a formal model means that GCs have been designed in an ad-hoc manner, published without proof of correctness and with bugs; it also means that it is difficult to apply experience gained from one implementation to the design of another. This thesis presents Surf, an abstract model of distributed garbage collection that bridges the gap between expressibility and specificity: it can describe a wide range of GCs and contains a proof of correctness that defines a list of requirements that must be fulfilled. Surf’s design space and its requirements for correctness provide a process that may be followed to analyse an existing collector or create a new GC. Surf predicts the abstract behaviour of GCs; this thesis evaluates those predictions in light of the understood behaviour of published GCs to confirm the accuracy of the model. A distributed persistent implementation of the Train Algorithm is created as an instantiation of Surf and the model is used to analyse progress in the GC and drive the design of a partition selection policy that provides a lower bound on progress and therefore reduces the GC’s complexity to completeness. Tests with mesh data structures from finite element analysis confirm the progress predictions from Surf. Published GCs cluster mostly in one corner of the Surf design space so this thesis explores the design of a GC at an unoccupied design point: the Tram Algorithm. Analysis via Surf leads to the prediction that Trams are capable of discovering topology in the live object graph that approximately identifies the strongly connected components, permitting O(1) timeliness that is unique to the Tram Algorithm. / Thesis (Ph.D.) -- University of Adelaide, School of Computer Science, 2008