Hoje em dia, desenvolvedores de sistemas embarcados enfrentam uma crescente complexidade de projeto, tanto nas aplicações quanto nas plataformas usadas para executá-las. O uso de plataformas complexas faz com que os engenheiros precisem fazer escolhas não-triviais, e muitas vezes contra-intuitivas durante a fase de projeto. Para permitir que os projetistas gerenciem esta complexidade, o uso de metodologias baseadas em modelos tem atraído atenção, e dentro deste contexto, a metodologia HIPAO2 está sendo desenvolvida dentro da UFRGS. Dentre os problemas que os engenheiros precisam enfrentar, o mapeamento entre tarefas e processadores em sistemas multiprocessados heterogêneos é um problema NP-completo, onde o espaço de projeto rapidamente se torna grande demais para que seja explorado satisfatoriamente de maneira manual. Este trabalho detalha a extensão das ferramentas que suportam a metodologia HIPAO2, de maneira a incluir facilidades de Exploração de Espaço de Projeto semi-automática para a solução deste problema. A ferramenta proposta faz uso de um algoritmo genético multiobjetivo para evidenciar tradeoffs existentes no projeto, e algoritmos de análise de aplicações modeladas como synchronous dataflow para avaliar possíveis mapeamentos sem um custo computacional proibitivo. / Designers of today’s embedded systems are faced with increasing complexity both in the applications being developed and the platforms they run on. The use of complex platforms means that the engineers need to make non-trivial and many times non-intuitive decisions during the design phase. To help developers work with this complexity, model-driven techniques are gaining attention, and in this context, the HIPAO2 model-driven engineering methodology is being developed at UFRGS. Among the problems that designers must solve, the task-to-processor mapping in heterogeneous multiprocessor systems is an NP-complete problem and the design space will quickly become too large to be explored adequately by humans. This work details the extension of the tools that support HIPAO2 to include semiautomatic Design-Space Exploration capabilities for the mapping problem. The proposed tool includes the use of a multiobjective genetic algorithm to make tradeoffs explicit to the designers; it also uses synchronous dataflow analysis algorithms to evaluate potential alternatives with a reasonable computational cost.

Un micro-drone à voilure tournante est l’appareil aérien optimal pour assister un rover d’exploration à lanavigation sur la planète Mars. Toutefois, les écoulements qu’il rencontre sont compressibles à très faiblenombre de Reynolds, ce qui constitue un domaine de l’aérodynamique inédit et quasiment inexploré à cejour. L’objectif de la thèse est de comprendre, simuler et recréer expérimentalement les phénomènes aérodynamiquesliés au régime inédit des écoulements martiens pour concevoir un système propulsif performant.Après avoir validé les outils de simulation numérique, le comportement instationnaire des écoulements estétudié sur des géométries 2D et 3D. L’écoulement est dominé par la viscosité : les couches limites laminairessont épaisses et le décollement a beaucoup d’influence sur son comportement très instationnaire.Par la suite, plusieurs millions de géométries de profil sont évaluées par un processus d’optimisation basésur un code 2D stationnaire. Les profils optimisés sont fortement cambrés (entre 5.5% et 7%) et de faibleépaisseur relative (e/c ∼ 2%). Le bord d’attaque et le bord de fuite sont très cambrés pour permettrerespectivement l’adaptation à l’écoulement incident et la fixation du point de décollement de la couchelimite. À partir des géométries de profils, l’ensemble du système propulsif est optimisé par intégration despolaires 2D. La théorie des éléments de pale permet de déterminer rapidement les configurations les plusperformantes aérodynamiquement. Et une méthode de sillage libre permet l’optimisation de rotors isoléset de systèmes propulsifs coaxiaux. Les rotors ont des solidités et des vrillages importants, ce qui rappelleles formes d’hélices marines. Les simulations Navier-Stokes 3D mettent en évidence la tridimensionnalitédes écoulements sur la pale, elle est fortement corrélée avec la solidité du rotor et le vrillage de bout depale. La rotation stabilise la couche limite et donne lieu à un décollement stable au bord d’attaque pourcertaines géométries. Le dévrillage en bout de pale permet de stabiliser le tourbillon et de diminuer la perteinduite. Un banc de mesure est placé dans un caisson dépressurisé pour estimer les efforts de poussée et decouple générés par les rotors optimisés en conditions aérodynamiques martiennes. Les essais permettentde valider les tendances d’estimation des codes de simulation ainsi que les processus d’optimisation. Laconfiguration bi-rotors coaxiaux, en comparaison avec une configuration à deux rotors adjacents, permetun gain d’encombrement de moitié pour une perte sur la puissance de seulement 15%. C’est la configurationla plus adaptée pour un micro-drone en atmosphère ténue. Un système propulsif coaxial optimisé (dediamètre 30 cm) permettrait de sustenter un micro-drone d’environ 400 grammes en conditions nominalessur la planète Mars. / A micro-rotorcraft is the most suited aerial vehicle for rover navigation assistance on Mars. The martianatmosphere’s density, being hundred times lower than on Earth, requires the micro-drone to hover at highrotational speed. Hence, flows on the blade are both compressible and at very low Reynolds number (fewthousands). It constitutes a new aerodynamic domain to be explored. The purpose of the dissertation isto understand, simulate and experimentally duplicate the aerodynamic phenomena in a view to design anefficient propulsion system. After a phase of validation of the simulation tools, the flows’ unsteady behavioris studied on 2D and 3D geometries. Wall flow is highly viscous : laminar boundary layers are thick andtheir separation has a huge influence on its unsteadiness. Then, several millions of airfoil geometries areevaluated by an optimization process based on a steady 2D solver. As final result, the optimized airfoildisplays a highly cambered shape (between 5.5% and 7%) with low relative thickness (t/c ∼ 2%). Leadingand trailing edges are strongly cambered, allowing proper incoming flow adaptation and late boundarylayer separation. Based on this airfoil geometry, rotor shapes are optimized by two methods. Blade elementtheory provides a quick investigation of the most aerodynamically efficient configurations. And a free wakesolver is applied for the final design of isolated rotors and coaxial dual-rotors. Optimized shapes exhibitimportant twist and solidity, evoking marine propellers. 3D Navier-Stokes simulations highlight the flow’sthree-dimensional mechanisms on the blade, which are highly correlated to the rotor’s solidity and twist.Rotation stabilizes the blade’s boundary layer and a stable leading edge separation is observed in somecases. Blade tip twist reduction diminishes the tip vortex and the induced loss. A thrust and torquemeasurement setup is placed in a depressurized tank for the evaluation of optimized rotors in martianatmospheric conditions. Performed experiments confirm the solvers’ trend and validate the design process.Therefore, chosen coaxial dual-rotors configuration provides an important size gain for a low correspondingpower loss (∼ 15%) compared to adjacent dual-rotors. Coaxial dual-rotors are the optimal configuration fora micro-rotorcraft in low-pressure atmosphere flight conditions. Such propulsion system (with a diameterof 30 cm) could lift a 400 grams micro-rotorcraft in hover on the planet Mars.

As systems become more complex, there are increasing demands for improvement with respect to attributes such as performance, dependability, and security. Optimization is defined as theprocess of making the most effective use of a set of resources with respect to some attribute. Existing optimization techniques, however, have two fundamental limitations. They target individual parts of a system without considering the potentially significant global picture, and they are designed to improve a single attribute at a time. These limitations impose significant restrictions on the kinds of optimization possible, the effectiveness of the techniques, and the ability to improvethe optimization process itself.This dissertation presents holistic system optimization, a new approach to optimization based on taking a broad view of a system. Unlike current approaches, holistic optimizations consider different kinds of interactions at multiple levels in a system, and target a variety of metrics uniformly. A key component of this research has been the use of proven compiler techniques to ensure transparency, automation, and correctness. These techniques have been implemented in Cassyopia, a software prototype of a framework for performing holistic optimization.The core of this work is three new holistic optimizations, which are also presented. The first describes profile-directed static optimizations designed to improve the performance of eventbased programs by spanning boundaries that separate code that raises events from handlers that field them. The second, system call clustering, improves the system call behavior of an entire program by grouping together calls that can be executed in a single boundary crossing. In thiscase, the optimization spans kernel and user address spaces. Finally, authenticated system calls optimize system security through a novel implementation of an efficient system call monitor. This example demonstrates how the new approach can be used to create new optimizations that not only span address space boundaries but also target attributes such as dependability. All of these optimizations involve the application of standard compiler techniques in non-traditional contexts and demonstrate how systems can be improved beyond what is possible using existing techniques.

Electric power is a basic requirement for present day life and its various economic sectors. To satisfy the ever-increasing needs for electricity, the number of generating units, transmission lines and distribution systems is rising steadily. In addition, electric power systems are among the most complex industrial systems of the modern age. Beside complexity, the generation of electric power is a main source of gaseous emissions and pollutants. The planning and operation of electric power systems must be done in a way that the load demand is met reliably, cost-effectively and in an environmentally responsible manner. Practitioners strive to achieve these goals for successful planning and operations utilizing various optimization tools. It is clear that the objectives to be satisfied are mostly conflicting. In particular, minimizing the fuel cost and the gaseous emissions are two conflicting and non-commensurate objectives. Therefore, multi-objective optimization techniques are employed to obtain trade-off relationships between these incompatible objective functions in order to help decision makers take proper decisions. In this thesis, two main power system operation problems are addressed. These are the economic load dispatch (ED) and the short-term hydro-thermal generation scheduling (STHTS). They are treated first as single-objective optimization problems then they are tackled as multi-objective ones considering the environmental aspects. These problems, single and multi-objective, are nonlinear non-convex constrained optimization problems with high-dimensional search spaces. This makes them a real challenge for any optimization technique. To obtain the optimal or close to optimal solutions, a modified bacterial foraging algorithm is proposed, developed and successfully applied. The bacterial foraging algorithm is a metaheuristic non-calculus-based optimization technique. The proposed algorithm is validated using diverse benchmark optimization examples before implementing it to solve the problems of this thesis. Various practical constraints are considered in the different cases of each problem. These include transmission losses, valve-point effects for both the ED and the STHTS problems and water availability and reservoir configurations for the STHTS problem. In all cases the optimal or near-optimal solution is obtained. For the multi-objective optimization cases, the Pareto optimal solution set that shows the trade-off relationship between the conflicting objectives is successfully captured.

This thesis has improved Baltazar's methodology for potential energy savings estimation from retro-commissioning/retrofits measures. Important improvements and discussions are made on optimization parameters, limits on optimization parameter values, minimum airflow setting for VAV systems, space load calculation, simulation of buildings with more than one type of system, AHU shutdown simulation, and air-side simulation models. A prototype computer tool called the Potential Energy Savings Estimation (PESE) Toolkit is developed to implement the improved methodology and used for testing. The implemented methodology is tested in two retro-commissioned on-campus buildings with hourly measured consumption data. In the Sanders Corps of Cadets Center, the optimized profiles of parameter settings in single parameter optimizations can be explained with engineering principles. It reveals that the improved methodology is implemented correctly in the tool. The case study on the Coke Building shows that the improved methodology can be used in buildings with more than one system type. The methodology is then used to estimate annual potential energy cost savings for 14 office buildings in Austin, TX with very limited information and utility bills. The methodology has predicted an average total potential savings of 36% for SDVAV systems with electric terminal reheat, 22% for SDVAV systems with hot water reheat, and 25% for DDVAV systems. The estimations are compared with savings predicted in the Continuous Commissioning assessment report. The results show it may be helpful to study the correlation by using generalized factors of assessment predicted energy cost savings to estimated potential energy cost savings. The factors identified in this application are 0.68, 0.66, and 0.61 for each type of system. It is noted that one should be cautious in quoting these factors in future projects. In the future, it would be valuable to study the correlation between measured savings and estimated potential savings in a large number of buildings with retrocommissioning measures implemented. Additionally, further testing and modifications on the PESE Toolkit are necessary to make it a reliable software tool.

This research proposes an efficient reliability modeling and simulation methodology in power systems to include photovoltaic units, wind farms and storage. Energy losses by wake effect in a wind farm are incorporated. Using the wake model, wind shade, shear effect and wind direction are also reflected. For solar modules with titled surface, more accurate hourly photovoltaic power in a specific location is calculated with the physical specifications. There exists a certain level of correlation between renewable energy and load. This work uses clustering algorithms to consider those correlated variables. Different approaches are presented and applied to the composite power system, and compared with different scenarios using reliability analysis and simulation. To verify the results, reliability indices are compared with those from original data. As the penetration of renewables increases, the reliability issues will become more important because of the intermittent and non-dispatchable nature of these sources of power. Storage can provide the ability to regulate these fluctuations. The use of storage is investigated in this research. To determine the operating states and transition times of all turbines, Monte Carlo is used for system simulation in the thesis. A conventional power system from IEEE Reliability Test Systems is used with transmission line capacity, and wind and solar data are from National Climatic Data Center and National Renewal Energy Laboratory. The results show that the proposed technique is effective and efficient in practical applications for reliability analysis.

Hoje em dia, desenvolvedores de sistemas embarcados enfrentam uma crescente complexidade de projeto, tanto nas aplicações quanto nas plataformas usadas para executá-las. O uso de plataformas complexas faz com que os engenheiros precisem fazer escolhas não-triviais, e muitas vezes contra-intuitivas durante a fase de projeto. Para permitir que os projetistas gerenciem esta complexidade, o uso de metodologias baseadas em modelos tem atraído atenção, e dentro deste contexto, a metodologia HIPAO2 está sendo desenvolvida dentro da UFRGS. Dentre os problemas que os engenheiros precisam enfrentar, o mapeamento entre tarefas e processadores em sistemas multiprocessados heterogêneos é um problema NP-completo, onde o espaço de projeto rapidamente se torna grande demais para que seja explorado satisfatoriamente de maneira manual. Este trabalho detalha a extensão das ferramentas que suportam a metodologia HIPAO2, de maneira a incluir facilidades de Exploração de Espaço de Projeto semi-automática para a solução deste problema. A ferramenta proposta faz uso de um algoritmo genético multiobjetivo para evidenciar tradeoffs existentes no projeto, e algoritmos de análise de aplicações modeladas como synchronous dataflow para avaliar possíveis mapeamentos sem um custo computacional proibitivo. / Designers of today’s embedded systems are faced with increasing complexity both in the applications being developed and the platforms they run on. The use of complex platforms means that the engineers need to make non-trivial and many times non-intuitive decisions during the design phase. To help developers work with this complexity, model-driven techniques are gaining attention, and in this context, the HIPAO2 model-driven engineering methodology is being developed at UFRGS. Among the problems that designers must solve, the task-to-processor mapping in heterogeneous multiprocessor systems is an NP-complete problem and the design space will quickly become too large to be explored adequately by humans. This work details the extension of the tools that support HIPAO2 to include semiautomatic Design-Space Exploration capabilities for the mapping problem. The proposed tool includes the use of a multiobjective genetic algorithm to make tradeoffs explicit to the designers; it also uses synchronous dataflow analysis algorithms to evaluate potential alternatives with a reasonable computational cost.

Hoje em dia, desenvolvedores de sistemas embarcados enfrentam uma crescente complexidade de projeto, tanto nas aplicações quanto nas plataformas usadas para executá-las. O uso de plataformas complexas faz com que os engenheiros precisem fazer escolhas não-triviais, e muitas vezes contra-intuitivas durante a fase de projeto. Para permitir que os projetistas gerenciem esta complexidade, o uso de metodologias baseadas em modelos tem atraído atenção, e dentro deste contexto, a metodologia HIPAO2 está sendo desenvolvida dentro da UFRGS. Dentre os problemas que os engenheiros precisam enfrentar, o mapeamento entre tarefas e processadores em sistemas multiprocessados heterogêneos é um problema NP-completo, onde o espaço de projeto rapidamente se torna grande demais para que seja explorado satisfatoriamente de maneira manual. Este trabalho detalha a extensão das ferramentas que suportam a metodologia HIPAO2, de maneira a incluir facilidades de Exploração de Espaço de Projeto semi-automática para a solução deste problema. A ferramenta proposta faz uso de um algoritmo genético multiobjetivo para evidenciar tradeoffs existentes no projeto, e algoritmos de análise de aplicações modeladas como synchronous dataflow para avaliar possíveis mapeamentos sem um custo computacional proibitivo. / Designers of today’s embedded systems are faced with increasing complexity both in the applications being developed and the platforms they run on. The use of complex platforms means that the engineers need to make non-trivial and many times non-intuitive decisions during the design phase. To help developers work with this complexity, model-driven techniques are gaining attention, and in this context, the HIPAO2 model-driven engineering methodology is being developed at UFRGS. Among the problems that designers must solve, the task-to-processor mapping in heterogeneous multiprocessor systems is an NP-complete problem and the design space will quickly become too large to be explored adequately by humans. This work details the extension of the tools that support HIPAO2 to include semiautomatic Design-Space Exploration capabilities for the mapping problem. The proposed tool includes the use of a multiobjective genetic algorithm to make tradeoffs explicit to the designers; it also uses synchronous dataflow analysis algorithms to evaluate potential alternatives with a reasonable computational cost.

Motivated by the role of decarbonizing the electric power sector to mitigate climate change, I assess the economic and environmental merits of three key technologies for decarbonizing the electric power sector across four chapters in this thesis. These chapters explore how adding flexibility to power plants equipped with carbon capture and sequestration (CCS) affects system costs and carbon dioxide (CO2) emissions, how grid-scale electricity storage affects system CO2 emissions as a power system decarbonizes, and how distributed solar photovoltaic (distributed PV) electricity generation suppresses wholesale electricity prices. In each chapter, I address these questions through a combination of power system optimization, statistics, and techno-economic analysis, and tie my findings to policy implications. In Chapter 2, I compare the cost-effectiveness of “flexible” CCS retrofits to other compliance strategies with the U.S. Clean Power Plan (CPP) and a hypothetical stronger CPP. Relative to “normal” CCS, “flexible” CCS retrofits include solvent storage that allows the generator to temporarily eliminate the CCS parasitic load and increase the generator’s net efficiency, capacity, and ramp rate. Using a unit commitment and economic dispatch (UCED) model, I find that flexible CCS achieves more cost-effective emissions reductions than normal CCS under the CPP and stronger CPP, but that flexible CCS is less cost-effective than other compliance strategies under both reduction targets. In Chapter 3, I conduct a detailed comparison of how flexible versus normal CCS retrofits affect total system costs and CO2 emissions under a moderate and strong CO2 emission limit. Given that a key benefit of flexible CCS relative to normal CCS is increased reserve provision, I break total system costs into generation, reserve, and CCS capital costs. Using a UCED model, I find that flexible CCS retrofits reduce total system costs relative to normal CCS retrofits under both emission limits. Furthermore, 40-80% of these cost reductions come from reserve cost reductions. Accounting for costs and CO2 emissions, though, flexible CCS poses a trade-off to policymakers under the moderate emission limit, as flexible CCS increases system CO2 emissions relative to normal CCS. No such trade-off exists under the stronger emission limit, as flexible CCS reduces system CO2 emissions and costs relative to normal CCS. In Chapter 4, I quantify how storage affects operational CO2 emissions as a power system decarbonizes under a moderate and strong CO2 emission limit through 2045. In so doing, I aim to better understand how storage transitions from increasing CO2 emissions in historic U.S. systems to enabling deeply decarbonized systems. Additionally, under each target I compare how storage affects CO2 emissions when participating in only energy, only reserve, and energy and reserve markets. Using a capacity expansion (CE) model to forecast fleet changes through 2045 and a UCED model to quantify how storage affects system CO2 emissions, I find that storage quickly transitions from increasing to decreasing CO2 emissions under the moderate and strong emission limits. Whether storage provides only energy, only reserves, or energy and reserves drives large differences in the magnitude, but not the direction, of the effect of storage on CO2 emissions. In Chapter 5, I quantify a benefit of distributed photovoltaic (PV) generation often overlooked by value of solar studies, namely the market price response. By displacing high-cost marginal generators, distributed PV generation reduces wholesale electricity prices, which in turn reduces utilities’ energy procurement costs. Using 2013 through 2015 data from California including a database of all distributed PV systems in the three California investor owned utilities, we estimate historic hourly distributed PV generation in California, then link that generation to reduced wholesale electricity prices via linear regression. From 2013 through 2015, we find that distributed PV suppressed historic median hourly LMPs by up to $2.7-3.1/MWh, yielding avoided costs of up to $650-730 million. These avoided costs are smaller than but on the order of other avoided costs commonly included in value of solar studies, so merit inclusion in future studies to properly value distributed PV.

With the fast growth of data-intensive distributed systems today, more novel and principled approaches are needed to improve the system efficiency, ensure the service quality to satisfy the user requirements, and lower the system running cost. This dissertation studies the design issues in the data-intensive distributed systems, which are differentiated from other systems by the heavy workload of data movement and are characterized by the fact that the destination of each data flow is limited to a subset of available locations, such as those servers holding the requested data. Besides, even among the feasible subset, different locations may result in different performance. The studies in this dissertation improve the data-intensive systems by exploiting the data storage location flexibility. It addresses how to reasonably determine the data placement based on the measured request patterns, to improve a series of performance metrics, such as the data access latency, system throughput and various costs, by the proposed hypergraph models for data placement. To implement the proposal with a lower overhead, a sketch-based data placement scheme is presented, which constructs the sparsified hypergraph under a distributed and streaming-based system model, achieving a good approximation on the performance improvement. As the network can potentially become the bottleneck of distributed data-intensive systems due to the frequent data movement among storage nodes, the online data placement by reinforcement learning is proposed which intelligently determines the storage locations of each data item at the moment that the item is going to be written or updated, with the joint-awareness of network conditions and request patterns. Meanwhile, noticing that distributed memory caches are effective measures in lowering the workload to the backend storage systems, the auto-scaling of memory cache clusters is studied, which tries to balance the energy cost of the service and the performance ensured. As the outcome of this dissertation, the designed schemes and methods essentially help to improve the running efficiency of data-intensive distributed systems. Therefore, they can either help to improve the user-perceived service quality under the same level of system resource investment, or help to lower the monetary expense and energy consumption in maintaining the system under the same performance standard. From the two perspectives, both the end users and the system providers could obtain benefits from the results of the studies. / Graduate