Rural electrification serves as a tool to reach several of the United Nations Sustainable Development Goals (SDGs). However, challenges occur related to increased electricity access, involving financial, informational, technical and regulatory factors. Another issue concerning electrification is the potential climate effects that increased electricity production causes. To minimise emissions and energy losses, usage of renewable energy sources is a possible solution as well as increased decentralised electricity production. Microgrids enable both of these parts and therefore serve as a great solution to sustainably meet the increased demand of electricity. Therefore, the main purpose of this study is to investigate the potential of microgrids to sustainably contribute to an increased access to electricity. The study focuses on the electrical power situation in Thailand. The literature study covers the political strategies regarding rural electrification and implementation of renewable energy sources in Thailand’s electrical power system, and how the country is currently developing according to these strategies. The case study consists of a cost and emission analysis related to a pilot microgrid project in northern Thailand. The Provincial Electricity Authority (PEA) in Thailand initiated and built the microgrid project called Maesariang Microgrid, in MaeSariang District in the northern province Mae Hong Son. The purpose of the project was to improve reliability and quality of the power delivered to Mae Sariang District. The microgrid consists of a solar power plant, a hydro power station, diesel generators and a battery energy storage system (BESS). The results from the case study shows that implementation of the Maesariang Microgrid, in addition to power reliability and quality, resulted in both reduced cost of electricity generation (COEG) with 3 percent and carbon dioxide (CO2) emissions with 11 percent. Furthermore, emissions per generated electricity unit has reduced from 380 kg CO2 per MWhto 340 kg CO2 per MWh. By using a system model covering the current microgrid in Mae Sariang, three alternative microgrid compositions with different shares of installed renewable energy capacity were simulated in three scenarios. In these scenarios, the amount of electricity generated from the solar power plant, diesel generators and BESS varied in order to analyse the effects on emissions and COEG. The analysis displayed that an increased capacity of solar power installed, both with the current size of the BESS capacity, and with an increased BESS capacity, resulted in lower COEG and emissions. It also showed that usage of less installed solar power capacity and increased usage of diesel generators would reduce CO2 emission sper year compared to using the local grid. However, with the current price of diesel fuel and electricity from the local grid, it would result in a higher COEG. / Elektrifiering av landsbygden fungerar som ett verktyg för att nå flera av FN:s mål för hållbarutveckling. Däremot uppstår utmaningar relaterade till ökad tillgång till elektricitet, som involverar både finansiella, informationsmässiga, tekniska och regulatoriska faktorer. Ett ytterligare problem som elektrifiering medför är de potentiella klimateffekter som ökad elproduktion orsakar. För att minimera utsläpp och energiförluster är användning av förnybara energikällor samt ökad decentraliserad elproduktion en möjlig lösning. Microgrids möjliggör båda dessa delar och fungerar därför som en utmärkt lösning för att på ett hållbart sätt möta den ökade efterfrågan på el. Huvudsyftet med denna studie är därför att undersöka microgrids potential att på ett hållbart sätt bidra till ökad tillgång till elektricitet. Studien fokuserar på elsituationen i Thailand. Litteraturstudien täcker landets politiska strategier för landsbygdens elektrifiering och implementering av förnybara energikällor i det thailändska elkraftsystemet, samt hur landet för närvarande utvecklas enligt dessa strategier. Fallstudien består av en kostnads- och utsläppsanalys relaterad till ett microgridpilotprojekt i norra Thailand. Provincial Electricity Authority (PEA) i Thailand initierade och byggde microgridprojektet Maesariang Microgrid i Mae Sariang-distriktet, i den norra provinsen MaeHong Son. Syftet med microgridprojektet var att förbättra tillförlitligheten och kvaliteten på elektriciteten som levereras till Mae Sariang-distriktet. Microgridet består av en solpark, ett vattenkraftverk, dieselgeneratorer och ett batterienergilagringssystem. Resultaten från fallstudien visar att implementeringen av microgridet i Mae Sariang, förutom strömtillförlitlighet och kvalitet, har lett till både minskad kostnad för elproduktion med 3 procent och CO2-utsläpp med 11 procent. Vidare har utsläppen per genererad elenhet minskat från 380 kg CO2 per MWh till 340 kg CO2 per MWh. Genom att använda en systemmodell som omfattar det nuvarande microgridet i Mae Sariang, simulerades tre alternativa microgridsammansättningar med olika andelar installerad förnybar energikapacitet i tre scenarion. I dessa scenarion varierar mängden el som genereras från solparken, dieselgeneratorer och batteriet för att analysera hur utsläpp och kostnad förelproduktion påverkas. Analysen visade att en ökad kapacitet av installerad solenergi, både med nuvarande storlek på batteriet, och med en ökad batterikapacitet, resulterade i lägre kostnader för elektricitet och utsläpp. Den visade också att användning av mindre installerad solenergikapacitet och ökad användning av dieselgeneratorer skulle minska CO2-utsläppen per år jämfört med att använda det lokala nätet. Dock skulle det resultera i högre kostnader för elektricitet med det nuvarande priset på dieselbränsle och el från det lokala nätet.

Cuba has vast natural resources for domestic renewable energy generation, but their energy mix is heavily dominated by fossil fuels. This contributes to a high dependence on expensive oil imports and has led to significant generation shortfalls, which in turn has resulted in extensive power outages and serious fuel crises. Additionally, large amounts of CO2 emissions are generated from power generation based on oil or gas. George Washington is a small industrial town in the Villa Clara province in Cuba that frequently experiences these problems. It holds a rum factory, a sugar mill, and a small residential area containing 710 households. The implementation of a microgrid utilizing the available solar, wind, and biomass potential could work to simultaneously reduce the town's dependence on energy imports, increase the renewable electricity share, and increase self-sufficiency of the electricity demand, enabling the industries and residential area to access energy services even when the national grid is not delivering power. By examining different potential microgrid configurations in HOMER Pro, an optimal system was decided based on cost parameters such as CAPEX and NPV, the self-sufficiency share of the electricity consumed, and the available potential to utilize domestic natural resources as well as the available workforce able to operate such a system. Because of Cuba's difficulties in accessing investment capital, a low CAPEX, high self-sufficiency index, and a high NPV was considered the best possible system. The scenario that best correlated with this outcome was the Middle Road scenario. By considering the area limiations of George Washington, one model run of the Middle Road scenario produced a system with additional solar PV (2.9 MW) and wind capacity (9.2 MW) paired with the already existing 6 MW of bagasse-fired CHP capacity in the sugar mill and 688 kW of solar PV capacity. It had a low investment cost of $34 million USD, a high NPV at $112 million USD, and a self-sufficiency index at 91.33%. Another model run of the Middle Road scenario that didn't take avaliable area into consideration produced a microgrid with an additional 43.1 MW of wind capacity. This model run had an NPV of $292 million USD, an investment cost of $79 million USD, and a self-sufficiency index of 94%. By implementing more capacity than this in the 100% Self-sufficient scenario, the self-sufficiency index reached a maximum of 100%, but had a lower NPV at $282 million USD, and a much higher investment cost of $1.324 billion USD. These scenarios only used biomass, solar, or wind energy for microgrid electricity generation, and therefore only consumed fossil fuels when importing electricity from the grid. / Kuba har en stor mängd naturliga resurser för att generera förnybar energi, men deras energimix idag domineras av fossilt bränsle. Landet är beroende av att importera dyr olja, vilket bidrar till en otillräcklig inhemsk energiproduktion samt många timmars strömavbrott och svåra bränslebrister. Stora mängder CO2-utsläpp genereras även när olja eller gas används för kraftproduktion. George Washington är en liten industriell by som ligger i provinsen Villa Clara, i Kuba, och som ofta får erfara dessa problem. I byn finns det en romfabrik, en sockerkvarn och ett litet bostadsområde som består av 710 hem. Installationen av ett microgrid som utnyttjar lokal solenergi, vindenergi samt biomassa kan minska byns beroende av importerad energi, öka andelen förnybar energi samt öka självförsörjningen av elbehovet. Ett sådant microgrid skulle möjliggöra byns tillgång till viktiga energitjänster även när det nationella nätverket inte har möjlighet att leverera elektricitet. Genom att undersöka flera olika microgridkonfigurationer i mjukvaruverktyget HOMER Pro valdes ett optimalt system baserat på parametrarna CAPEX, NPV, självförsörjningsgraden av den konsumerade elektriciteten, potentialen att använda sig av de lokala naturresurserna samt tillgängligheten av arbetskraft för att kunna driva ett sådant nätverkssystem. På grund av de begränsade tillgångarna till investeringskapital i Kuba så blev ett lågt CAPEX, hög självförsörjning samt ett högt NPV viktiga parametrar för att utse det bästa möjliga systemet. Det scenario som genererade system som bäst stämde överens med dessa egenskaper är Middle Road-scenariot. För att undersöka potentialen hos systemet och samtidigt ta hänsyn till den begränsade landtillgången i George Washingtons närområde så kördes en av systemsimuleringarna av Middle Road-scenariot med en areabegränsning i HOMER Pro. Detta resulterade i ett system med ytterligare 2.9 MW kapacitet från solpaneler, 9.2 MW vindkraft tillsammans med de redan existerande 6 MW av bagasse-drivna turbiner i sockerkvarnen samt de 688 kW av solpaneler som är installerade på romfabrikens tak. Systemet har en investeringskostnad (CAPEX) på $34 miljoner USD, ett högt NPV på $112 miljoner USD och ett självförsörjningsindex på 91.33%. När systemsimuleringen av Middle Road inte tog hänsyn till tillgänglig landyta så blev resultatet att det bästa systemet hade ytterligare 43.1 MW av vindkraft. Detta system har ett NPV på $292 miljoner USD, en investeringskostnad på $79 miljoner USD och ett självförsörjningsindex på 94%. Genom att implementera en högre kapacitet i 100% Self-sufficient-scenariot så blev resultatet ett självförsörjandeindex på 100%, men samtidigt ett lägre NPV på $282 miljoner USD och en mycket högre investeringskostnad på $1.324 miljarder USD. I dessa scenarion så används biomassa, solenergi samt vindenergi för generering av elektricitet i microgridet och konsumtion av fossilt bränsle sker endast när elektricitet importeras från det nationella elnätverket.

This project is aimed at developing a model for an off-grid microgrid by studying the wind and solar resource at Östergarnsholm, an island to the east of Gotland, Sweden. For the required consumption, a study and analysis of Herrvik’s local demand has been made which is a village close to the island and the possibility for implementing the microgrid to supply electricity to the inhabitants of Herrvik has been examined. Renewable resource is paired with batteries for mitigating intermittency and using a generator running on biofuel is proposed during hours with no wind and sun.The data for wind and solar resource is obtained from Department of Earth Sciences, Uppsala University and household load is estimated for a typical detached Swedish house from a study funded by International Energy Agency. Microgrid is modelled and simulated in MATLAB and the simulation results show that the proposed microgrid can provide 100% availability for local load. Results also show that not only the inhabitants of the village can have their independent electricity supply for the whole year, but the excess electricity can also be used to desalinate the seawater at Herrvik’s desalination plant for around 90% of the time during the year.

A hybrid ac/dc microgrid is a more complex but practical network that combines the advantages of an AC and a DC system. The main advantage of this network is that it connects both alternating current and direct current networks via an interlinking converter (IC) to form a unified distribution grid. The hybrid microgrid (HMG) will enable the direct integration of both alternating current (AC) and direct current (DC) distributed generators (DGs), energy storage systems (ESS), and alternating current and direct current (DC) loads into the grid. The alternating current and direct current sources, loads, and ESS are separated and connected to their respective subgrids primarily to reduce power conversion and thus increase overall system efficiency. As a result, the HMG architecture improves power quality and system reliability. Planning a hybrid microgrid entails estimating the capacities of DGs while taking technical, economic, and environmental factors into account. The hybrid ac-dc microgrid is regarded as the distribution network of the future, as it will benefit from both ac and dc microgrids. This thesis presents a general architecture of a hybrid ac-dc microgrid, which includes both planning and design. The goal of the Hybrid ac-dc microgrid planning problem is to maximise social welfare while minimising total planning costs such as investment, maintenance, and operation costs. This configuration will assist Hybrid microgrid planners in estimating planning costs while allowing them to consider any type of load ac/dc and DER type. Finally, this thesis identifies the research questions and proposes a future research plan.

Master of Science / Department of Architectural Engineering and Construction Science / Fred Hasler / This paper discusses a brief history of electricity, specifically alternating current (AC) and direct current (DC), and how the current standard of AC distribution has been reached. DC power was first produced in 1800, but the shift to AC occurred in the 1880’s with the advent of the transformer. Because the decisions for distribution were made over 100 years ago, it could be time to rethink the standards of power distribution. Compared to traditional AC distribution, DC microgrids are significantly more energy efficient when implemented with distributed generation. Distributed generation, or on-site generation from photovoltaic panels, wind turbines, fuel cells, or microturbines, is more efficient when the power is transmitted by DC. DC generation, paired with the growing DC load profile, increases energy savings by utilizing DC architecture and eliminating wasteful conversions. Energy savings would result from a lower grid strain and more efficient utilization of the utility grid. DC distribution results in a more reliable electrical service due to short transmission distances, high service reliability when paired with on-site generation, and efficient storage. Occupant safety is a perceived concern with DC microgrids due to the lack of knowledge and familiarity in regards to these systems. However, with proper regulation and design standards, building occupants never encounter voltage higher than 24VDC, which is significantly safer than existing 120VAC in the United States. DC Microgrids have several disadvantages such as higher initial cost due, in part, to unfamiliarity of the system as well as a general lack of code recognition and efficiency metric recognition leading to difficult certification and code compliance. Case studies are cited in this paper to demonstrate energy reduction possibilities due to the lack of modeling ability in current energy analysis programs and demonstrated energy savings of approximately 20%. It was concluded that continued advancement in code development will come from pressure to increase energy efficiency. This pressure, paired with the standardization of a 24VDC plug and socket, will cause substantial increases in DC microgrid usage in the next 10 years.

En raison de l'augmentation des coûts d'énergie et des préoccupations environnementales telles que les empreintes de carbone élevées, les systèmes de la production d'électricité centralisée se restructurent pour profiter des avantages de la production distribuée afin de répondre aux exigences énergétiques toujours croissantes. Les microgrids sont considérés comme une solution possible pour déployer une génération distribuée qui inclut des ressources énergétiques distribuées DERs (Distributed Energy Resources)(e.g, solaire, éolienne, batterie, etc). Dans cette thèse, nous traitons les défis de la gestion d'énergie dans un microgrid industriel où les charges énergétique sont constituées de processus industriels. Notre plan consiste à diviser la gestion de l'énergie du microgrid en deux parties: la production et la demande d’énergie.Du côté de la production d'énergie, les défis incluent la modélisation des générations de puissance et le lissage des fluctuations des DER. Pour modéliser les générations de puissance, nous proposons un modèle basé sur les concepts de service courbé de Network Calculus. En utilisant cet outil mathématique, nous déterminons une quantité minimale de puissance que les DERs peuvent générer; leur agrégation nous donnera une production d'énergie totale dans le microgrid. Après cela, s'il existe un déséquilibre entre la production et la demande d'énergie, nous proposons des stratégies différentes pour minimiser les coûts d'approvisionnement énergétique. Sur la base des données réelles de la consommation d'énergie d'un site industriel situé en France, des économies significatives peuvent être réalisées en adoptant ces stratégies. Dans cette thèse, nous étudions également comment atténuer les effets des fluctuations de puissance des DERs en conjonction avec des systèmes de stockage d'énergie. Pour cela, nous proposons un algorithme de lissage gaussien et nous le comparons avec des algorithmes de lissage trouvés dans l'état de l'art. Nous avons trouvé que l'algorithme proposé utilise de batterie de moins de taille à des fins de lissage par rapport à d'autres algorithmes. À cette fin, nous sommes également intéressés à étudier les effets de la gamme admissible des fluctuations sur les tailles de la batterie.Du côté de la demande, l'objectif est de réduire les coûts de l'énergie grâce aux approches de gestion de la demande DSM (Demand Side Management) telles que Demand Response (DR) et Energy Efficiency. Comme les processus industriels consomment énormément, une petite réduction de la consommation d'énergie en utilisant les approches DSM pourrait se traduire par des économies cruciales. Cette thèse se concentre sur l'approche DR qui peut profiter des prix variables de l'électricité dans le temps pour déplacer les demandes énergétiques des heures de pointe aux heures creuses. Pour atteindre cet objectif, nous comptons sur un modèle basé sur la théorie de file d'attente pour caractériser les comportements temporels (arrivée et départ des tâches) d'un système de fabrication. Après avoir défini les processus d'arrivée et de départ de tâches, une fonction d'utilisation efficace est utilisée pour prédire le comportement de la machine dans un domaine temporel et qui peut afficher son statut (allumé/éteint) à tout moment. En prenant le statut de chaque machine dans une ligne de production comme une entrée, nous proposons également un algorithme de planification DR qui adapte la consommation d'énergie d'une ligne de production aux deux contraintes de puissance disponibles et de taux de production. L'algorithme est codé à l'aide d’une machine d’état fini déterministe (Deterministic Finite State Machine) dans laquelle les transitions d'état se produisent en insérant une tâche à l'entrée du tapis roulant (on peut aussi avoir des transitions sans insertion de taches). Nous définissons des conditions pour l'existence d’un planificateur réalisable et aussi des conditions pour accepter positivement des demandes DRs / Due to increased energy costs and environmental concerns such as elevated carbon footprints, centralized power generation systems are restructuring themselves to reap benefits of distributed generation in order to meet the ever growing energy demands. Microgrids are considered as a possible solution to deploy distributed generation which includes Distributed Energy Resources (DERs) (e.g., solar, wind, battery, etc). In this thesis, we are interested in addressing energy management challenges in an industrial microgrid where energy loads consist of industrial processes. Our plan of attack is to divide the microgrid energy management into supply and demand sides.In supply side, the challenges include modeling of power generations and smoothing out fluctuations of the DERs. To model power generations, we propose amodel based on service curve concepts of Network Calculus (NC). Using this mathematical tool, we determine a minimum amount of power the DERs can generate and aggregating them will give us total power production in the microgrid. After that, if there is an imbalance between energy supply and demand, we put forward different strategies to minimize energy procurement costs. Based on real power consumption data of an industrial site located in France, significant cost savings can be made by adopting the strategies. In this thesis, we also study how to mitigate the effects of power fluctuations of DERs in conjunction with Energy Storage Systems (ESSs). For this purpose, we propose a Gaussian-based smoothing algorithm and compare it with state-of-the-art smoothing algorithms. We found out that the proposed algorithm uses less battery size for smoothing purposes when compared to other algorithms. To this end, we are also interested in investigating effects of allowable range of fluctuations on battery sizes.In demand side, the aim is to reduce energy costs through Demand Side Management (DSM) approaches such as Demand Response (DR) and Energy Efficiency (EE). As industrial processes are power-hungry consumers, a small power consumption reduction using the DSM approaches could translate into crucial savings. This thesis focuses on DR approach that can leverage time varying electricity prices to move energy demands from peak to off-peak hours. To attain this goal, we rely on a queuing theory-based model to characterize temporal behaviors (arrival and departure of jobs) of a manufacturing system. After defining job arrival and departure processes, an effective utilization function is used to predict workstation’s (or machine’s) behavior in temporal domain that can show its status (working or idle) at any time. Taking the status of every machine in a production line as an input, we also propose a DR scheduling algorithm that adapts power consumption of a production line to available power and production rate constraints. The algorithm is coded using Deterministic Finite State Machine (DFSM) in which state transitions happen by inserting a job (or not inserting) at conveyor input. We provide conditions for existence of feasible schedules and conditions to accept DR requests positively.To verify analytical computations on the queuing part, we have enhanced Objective Modular Network Testbed in C++ (OMNET++) discrete event simulator for fitting it to our needs. We modified various libraries in OMNET++ to add machine and conveyor modules. In this thesis, we also setup a testbed to experiment with a smart DR protocol called Open Automated Demand Response (OpenADR) that enables energy providers (e.g., utility grid) to ask consumers to reduce their power consumption for a given time. The objective is to explore how to implement our DR scheduling algorithm on top of OpenADR

[pt] O Objetivo dessa dissertação é propor um conjunto de indicadores e métricas para avaliar e mensurar a eficiência das subestações de um microgrid. No conceito institucional o qual a pesquisa se insere, considera-se que o esforço proporcionará a empresa em questão avaliar a eficiência e inteligência das suas subestações em seu microgrid. A pesquisa pode ser considerada aplicada, descritiva e metodológica. Quanto aos meios de investigação, foi usado o método de construção de indicadores para avaliação da eficiência energética aliada a dois métodos de apoio à decisão: AHP (Analytic Hierarchy Process) para definição dos pesos dos critérios classificatórios e a técnica TOPSIS (Technique for Order of Preference by Similarity to Ideal Solution) para hierarquização dos indicadores propostos por categoria dos requisitos de eficiência energética e inteligência, a partir dos graus atribuídos pelos especialistas. Destacam-se como resultados: a ferramenta de seleção e classificação de subestações e um conjunto consistente de indicadores de eficiência e inteligência. Como resultados, espera-se obter a melhoria da eficiência das subestações de um microgrid e poder acompanhar o processo de melhoria contínua. / [en] The aim of this study is to propose a set of indicators and metrics to evaluate and measure the efficiency and intelligence on microgrid electric substations. It was used a methodology to stablish indicators for evaluating the energy efficiency combined with two decision support methods: AHP (Analytic Hierarchy Process) for definition the importance of criteria selected, and TOPSIS (Technique for Order Preference by Similarity to Ideal of Solution) to classify the most efficiency and intelligent substation. As a result, it is expected to obtain the improvement of efficiency of power substations of a microgrid and follow the process of continuous improvement of their intelligence. Geller (2004) emphasizes the importance to adopt energy efficiency measures, through the use of alternative energy resources. The author states that the numbness in the current energy sources will result in irreparable harm to the environment, generating major climate change .It is necessary consider an innovative energy management model capable for meeting the challenges to meet the new standards of consumption, combining environmental, social and economic aspects . Smart Grids (SG) are able to incorporate information and communication technology in all aspects of generating electricity, allowing the entrance of new suppliers in the energy network, generating a better management of distribution and consumption, optimizing efficiency, minimizing costs, and diminishing environmental impacts. Currently, the SG is in development stage, with large spaces for research and development of enhancements and applications. Countries like USA, Japan and other European Community are planning to implement this novel grid till mid-2020 to 2030 Rivera (2013). In Brazil, this trend follows a different reality, due to economic crisis and regulatory aspects. In this country, most of smart grid projects are in development research and they are in slower progress status when compared to developed countries Rivera (2013). The implementation of a smart grid, with various alternatives of energy sources, provides better management and optimal use of each alternative. So, the necessity of managing the power distribution network more efficiently as possible increases Caneppele (2011). Microgrid is a concept widely disseminated in the literature. It can be defined as an intelligent electrical network applied in a smaller scale. As example we can cite: large commercial centers, a large industries; a large condominiums, a neighborhoods, etc. The aim of this article is to propose a set of indicators and metrics to evaluate and measure the efficiency and intelligence on microgrid electric substations. The methodology proposed will support mathematically the decision maker in the energy management of a huge television studio, which can be considered a microgrid. The research methodology followed a procedural structure analysis based on literature review focused on central themes of research: (i) substations (ii) energy efficiency (iii) microgrids (iv) multiple criteria decision methods (MCDM), trying to select the best methods to be considered in the context of energy efficiency of substations in microgrids. The model used in this work considered a hybrid multi-criteria decision-making (MCDM) method, putting together two methods widely used in literature: AHP (Analytic Hierarchy Process), and (TOPSIS Technique for Order Preference by Similarity to Ideal of Solution). AHP method was proposed by Saaty (2000). It is a multi-criteria support tool for decision making, which reveals the wide number of applications in current literature. Through peer comparisons, it is possible to perform the trials because the priorities calculated by the method will capture both the subjective measures, and the depth the particular criterion (or alternative) n relation to another. The second method used will be the TOPSIS. The technique is used to examine the performance of options by similarity to ideal solution. It was introduced by Hwang and Yoon Hwang (1981), in 1981, since then, it

La gestion d’énergie est un sujet de plus en plus important dans notre société. Nous faisons actuellement face à un nombre croissant de problèmes tels que l’épuisement des réserves pétrolières, le réchauffement climatique ou encore la diminution de la qualité de l’énergie (principalement due aux coupures pendant les pics de consommation). Les smartgrids sont une des solutions à ces problèmes. En ajoutant une communication bidirectionnelle et de nouvelles capacités en matière de technologies de l’information et de la communication, il est possible de créer un système autonome de gestion intelligente de l’énergie.Les travaux décrits dans ce mémoire s'intéressent particulièrement à la gestion des microgrids à l'aide de systèmes multi-agents (SMA). Les microgrids sont des réseaux de faibles puissances, composés de petits producteurs d’énergie décentralisés (éventuellement renouvelables) et de consommateurs. Ces réseaux peuvent être reliés (ou non) au réseau global ce qui ajoute à leur complexité. De par leurs complexités et leurs répartitions géographiques, les smartgrids, comme les microgrids, ne peuvent pas être gérés facilement par des systèmes centralisés. Les intelligences artificielles distribuées et plus particulièrement les SMA apparaissent comme un moyen cohérent de résoudre les problèmes liés aux smartgrids.Dans un premier temps, nous avons défini une approche mettant en oeuvre des boucles de rétroaction. Une boucle de rétroaction apparaît dans les systèmes complexes qui peuvent être définis avec plusieurs niveaux d'abstraction. Deux niveaux sont ainsi en interaction. Le niveau micro regroupe un ensemble d'agents ayant des comportements qui, une fois combinés, influeront sur l'état du système. Le niveau macro traite ces influences pour définir un nouvel état du système qui influera sur le comportement des agents du niveau micro. Cette boucle de rétroaction permet de séparer les comportements sur plusieurs niveaux.Cette approche est utilisée pour définir un problème de gestion offre-demande dans un microgrid. Ce problème permet de prendre en compte un ensemble d'objectifs qui sont actuellement traités de manière indépendante. Enfin, une application utilisant un SMA a été développée. Cette approche peut s'intégrer dans ce problème. Elle a pour but d'assurer la stabilité du réseau à tout instant grâce au contrôle de systèmes de stockage.Dans un second temps, un simulateur de réseau électrique permettant le contrôle dynamique des périphériques a été développé. Ce simulateur repose sur trois grands principes. Le premier est une modélisation à base d’agents du simulateur lui-même, pour représenter la complexité des réseaux électriques. Le second principe repose sur l’utilisation du paradigme holonique afin de prendre en compte les multiples niveaux inhérents aux réseaux électriques. Enfin, le troisième principe est inspiré du modelé influence/réaction et propose une technique qui permet de gérer les actions simultanées, éventuellement conflictuelles, au sein des SMA. / Energy management is, nowadays, a subject of uttermost importance. Indeed, we are facing growing concerns such as petroleum reserve depletion, earth global warming or power quality (e.g. avoiding blackouts during peak times). Smart grids is an attempt to solve such problems, by adding to power grids bidirectional communications and ICT capabilities in order to provide an intelligent autonomic management for the grid.This thesis focuses on the management of microgrids thanks to multiagent systems (MAS). Microgrids are low-power networks, composed of small and decentralized energy producers (possibly renewable) and consumers. These networks can be connected to the main grid or islanded, this make them more complex. Due to their complexity and their geographical distribution, smart grids and microgrids can not be easily managed by a centralized system. Distributed artificial intelligences especially MAS appear to be a solution to resolve problems related to smart grids.Firstly we defined an approach implementing feedback loops. These feedback loops exist in complex systems which can be defined with several abstraction levels. Two levels are interacting. The micro-level contains a set of agents owning behaviours that can be combined. The result of the combination imapct the state of the system. The macro-level processes these influences to define a new state of the system which will imapct the agents behaviours at the micro-level.This feedback loop separates behaviours on several levels.This approach is used to defined a demand and supply matching problem in microgrid. This problem afford to manage a set of goals which currently are independently processed. Finally, an application is developed using MAS that ensures grid stability thanks to storage systems. This application was thought to be integrated to the approach detailed above.Secondly, a grid simulator id developed. This simulator allows dynamic control of devices. It is based on three main principles.• An agent-based model of the simulator to represent the complexity of electrical networks. • The use of the holonic paradigm to take into account the multiple levels inherent to power grids. • The use of the influence/reaction model and offers a technology that can manage simultaneous actions, possibly conflicting, in MAS.

En 2007 avec la directive sur les énergies renouvelables, l’Union Européen s’est engagée à développer une économie à faible intensité de carbone. Cette directive amène à réduire les émissions de gaz à effet de serre en augmentant entre autres la partie d’énergie produite par des sources renouvelables. Le processus d’insertion massive d’énergies renouvelables dans le mix électrique européen, est d’ores et déjà un fait acquis et ses effets sont tangibles. Cependant, à côté de ses effets environnementaux bénéfiques, l’intégration à large échelle du renouvelable ne va pas sans causer des interrogations techniques et réglementaires. Par conséquent, de nouvelles stratégies de gestion du système électrique doivent être pensées et actées pour garantir un fonctionnement fiable et économiquement acceptable. Les micro réseaux sont à cet effet, un réceptacle intégrateur avec suffisamment de flexibilité pour accueillir un système de gestion capable de répondre aux exigences ci-dessus. Les travaux de cette thèse sont centrés sur la conception, le développement et l’implémentation de différentes stratégies de gestion des micro réseaux. Les algorithmes développés visent, soit à faciliter l’intégration du renouvelable à large échelle, soit à garantir un fonctionnement efficace et économique du système électrique. Une nouvelle architecture de réseau de distribution composé de micro réseaux clustérisés a été premièrement proposée. Chaque micro réseau est composé de systèmes de production à base ou non de renouvelable, des systèmes de stockage et de charges. Une stratégie de gestion énergétique optimale a été ensuite définie et développée. Cette stratégie permet de gérer la planification à court-terme et le contrôle en temps-réel des micro réseaux via un usage adéquat des sources et ce, tout en réduisant le coût du micro réseau. Un système multi-agents et l’optimisation linéaire mixte en nombres entiers ont été utilisés pour le développement et l’implémentation de cette stratégie intelligente distribuée. D’un point de vue extérieur, chaque micro réseau est vu comme une entité cohérente capable de supporter le fonctionnement du réseau principal en utilisant un ensemble de ses sources flexibles. Ainsi, que la seconde partie de cette thèse exploitera les clusters des micro réseaux et leurs propriétés pour gérer au mieux le réseau de distribution hôte. La conceptualisation technico-économique de différents mécanismes de gestion des réseaux de distribution a été abordée. Le développement d’une architecture de gestion hiérarchisée en plusieurs niveaux d’intelligence a permis de réduire la complexité du système et faciliter l’implémentation d’un réseau flexible, extensible et à fort taux de pénétration de renouvelables. Cette gestion distribuée a été possible grâce à une connaissance locale des modèles et des comportements des différentes systèmes connectés, et à un usage local des informations. Les travaux théoriques ont été ensuite testés sur une plateforme expérimentale conséquente et les résultats finaux ont corroboré les attentes de la théorie. / In 2007 with the renewable energy directive, the European Union established the development of a low-carbon economy. This directive aims to decrease greenhouse gas emissions by increasing the energy produced by renewable energy. Already today, the massive diffusion of renewable systems is tangible in the European electricity mix. However, in spite of their potential benefits, their large-scale integration leads to new technical and regulatory questions. Consequently, new management strategies need to be developed and applied in order to ensure a reliable and economical operation of the system. Microgrids are considered to be one of the most effective and flexible solutions able to meet these new needs.The main goals of this thesis are the conceptualization, development and implementation of different management strategies for microgrids. The algorithms developed aim to facilitate the massive integration of renewables and at the same time lead to an effective and economic operation of the systems. A new architecture of distribution grids based on cluster of microgrids was proposed. Each microgrid is composed of a number of renewable-based and conventional generation systems, storage systems and consumption. An optimal and distributed energy management strategy was then defined and developed. This strategy allows to manage the short-term energy management and real-time control of microgrids by using the connected sources in a smart and cost-efficient way. A multi-agent system and the mixed integer linear optimization technique were used for the implementation of this strategy.From a global point of view, each microgrid is seen as a coherent entity, which can support network operation by using its flexible and aggregated sources. Hence, the second part of this thesis aims to understand how distribution grids can exploit these cluster of microgrids and their properties. Different mechanisms for the active management of distribution grids are conceptualized from the technical and economical point of view. A new strategy based on hierarchical management of different smart levels allow to reduce the complexity of the system and to implement a more flexible and extensible system, thanks to a more local use of model knowledge and users behaviour. On the end, the theoretical work were tested on an experimental test-bed in order to show the effectiveness of the proposed theories.

Microgrids are new challenging power systems under development. This report presents a feasibility study of microgrid development. This is an essential task before implementing microgrid systems. It is extremely important to know the number and size of distributed energy resources (DERs) needed and it is necessary to compare investment costs with benefits in order to evaluate the profitability of microgrids. Under the assumption that a large number of DERs improves the reliability of microgrids an optimization problem is formulated to get the accurate mix of distributed energy resources. Uncertainty in physical and financial parameters is taken into account to model the problem considering different scenarios.  Uncertainty takes place in load demanded, renewable energy generation and electricity market price forecasts, availability of distributed energy resources and the microgrid islanding. It is modeled in a stochastic way. The optimization problem is formulated firstly as a mixed-integer programming solved via branch and bound and then it is improved formulating a two stage problem using Benders’ Decomposition which shortens the problem resolution. This optimization problem is divided in a long-term investment master problem and a short-term operation subproblem and it is solved iteratively until it reaches convergence. Bender’s Decomposition optimization problem is applied to real data from the Illinois Institute of Technology (IIT) and it gives the ideal mix of distributed energy resources for different uncertainty scenarios. These distributed energy resources are selected from an initial set. It proves the usefulness of this optimization technique which can be also applied to different microgrids and data. The different solutions obtained for different scenarios are explained and analyzed. They show the possibility of microgrid implementation and determine the most favorable scenarios to reach the microgrid implementation successfully.  Reliability is a term highly linked to the microgrid concept and one of the most important reasons of microgrid development. Thus an analysis of reliability importance is implemented using the importance index of interruption cost (  ) in order to measure the reliability improvement of developing microgrids. It shows and quantifies the reliability improvement in the system.