Agricultural University of Athens
Ανάπτυξη συστημάτων διαχείρισης ενέργειας που ενσωματώνουν υπολογιστική νοημοσύνη για το σχεδιασμό και έλεγχο αυτόνομων μικροδικτύων
Abstract
dc:descriptionThe increase of human population, urbanization and modernization have led to one of the most notable issues of the worldwide agenda, which is the outstanding growth of global energy demand. Today, millions of people are still left without basic energy services and most of these people reside in remote areas, which are not interconnected to any major electrical power system network. Taking into consideration that the problem of electrification is mainly addressed by the use of fossil fuel power stations which have a significant environmental impact, renewable energy generation is an alternative, sustainable and environmentally friendly solution. The rapid use of renewable energy has led to a change in the operation of electrical grids and the use of hybrid renewable energy systems, including storage systems and generators form the microgrids. Microgrids are usually small-scale, low-voltage, integrated energy systems, which can call for sufficient load to warrant local generation and distribution and can address the grid congestion problems as far as high introduction of distributed renewable energy generators are concerned. Microgrids can operate at either grid-connected or islanded (autonomous) mode and their primary aim is to cover all the needs present in the area that they serve. Another major issue of the worldwide agenda is the availability of drinking water. There are many areas in the world, such as islands, coastal regions and arid main lands, which although have a significant shortage in potable water supply, their solar and wind potential is very high. With the fact that around 97% of the earth’s water is salty, desalination process emerged as an attractive solution to the water scarcity problem. A prominent challenge associated with desalination is its high energy consumption, which is often covered largely by fossil fuels. Therefore, the combination of desalination plants with renewable energy sources (RES) have been identified as a sustainable and economically viable solution compared to conventional fossil fuel-powered desalination plants. Moreover, the consumption of fossil fuels is increasing due to the use of fuels for transportation. Hydrogen is an energy carrier the use of which has been rising constantly throughout recent years and it can be used as a fuel for transportation, as well as a medium to long-term energy storage. A hydrogen subsystem is usually composed of a water electrolyzer, which produces the hydrogen, a hydrogen storage tank which acts as energy storage, a refuelling system which supports the dispensing hydrogen into fuel cell vehicles and a fuel cell which produces electricity from the stored hydrogen. The desalination and hydrogen systems consume electrical energy and hence can both constitute subsystems of a microgrid. As the microgrids become more complex, advanced energy storage systems and corresponding advanced energy management systems are required for optimal operation. The concept behind this thesis is the design and the development of energy management systems based on distributed intelligence for the design, control and optimum operation of microgrids. A brief description of decentralized energy management systems is presented, including their advantages instead of the use of centralized energy management systems. In addition, the multi agent systems approach in electrical power systems is mentioned. Then, the computational intelligence techniques that can be integrated into a multi agent system are presented. A decentralized energy management system based on multi-agent system for the control and management of an autonomous microgrid is presented. In the proposed system of intelligent agents, each unit of the network is controlled by its own processing unit, which is able to act as an independent agent. Every agent communicates and cooperates with other agents, thus forming an intelligent distributed computing system. The decisions concerning the energy production and storage, the production and storage of potable water and the production and utilization of hydrogen as fuel are made by this multi-agent system. Every agent receives information from sensors or through other agents and determines the operation point of the particular controlled unit, which correspond to full or partial load of the electrolyzer, fuel cell and desalination unit. The intelligent agents make decisions using the Fuzzy Cognitive Maps theory. The proposed decentralized energy management system is afterwards compared techno-economically to a centralized energy management system, based on a combined Petri Nets – Fuzzy Cognitive Maps approach. The results of the comparison revealed that the decentralized energy management system utilize the available devices better, which in turn led to a lower net present cost for a 20 year investment period and the managed microgrid has much higher chances of partial operation in cases when malfunctions occur at different parts of it. Another important issue that has been addressed is the determination of the optimal configuration of a renewable energy desalination system so that the system meets the water load demand at the minimum cost. Initially, a seawater reverse osmosis desalination plant was experimentally investigated at the laboratory of Agriculture University of Athens. Then, five different configurations for a seawater desalination system powered by photovoltaics were effectuated and compared in order to define the optimum technical and economic configuration of system components for autonomous mode operation. The results showed that the optimum system is the system that includes a small capacity conventional energy storage system (lead acid batteries) and an energy management system that incorporates computational intelligence techniques. An experimental investigation of a PV powered small-scale SWRO desalination system which employs hydraulic energy recovery, based on a DC microgrid concept was designed and tested in the laboratory. The DC microgrid incorporates a short term electric energy storage in the form of hybrid capacitors and a short term hydraulic energy storage in the form of pressure vessels. A multi-agent decentralized energy management system which relies on intelligent agents and employs Fuzzy Cognitive Maps for its implementation, allowing for efficient operation of the desalination system under variable conditions (variable feed pressure) was designed and experimentally investigated in order to maximize the fresh water production, optimizing the use of power plants and storage units. Finally, after being tested, both through simulation and experimentation, the beneficial application of multi agent systems in microgrids has been discussed to solve the energy management problem in distributed energy management systems by exploiting a theoretical framework relying on the cooperation among the agents, then the use of competing intelligent agent was examined. In several cases, an interactive operation among multiple agents in an energy system appears very often. In these cases, the goals of each agent are influenced by the actions of the remaining agents, while it is often possible that cooperation of the agents cannot be established, because they have competitive goals or different strategies which influence agents’ actions. On this basis, a multi-agent decentralized energy management system for a microgrid was designed and the energy management problem was formulated through the application of game theory, in order to model the set of strategies between two players/agents, as a non-cooperative power control game or a cooperative one, according to the level of the energy produced by the renewable energy sources and the energy stored in the battery bank, for the purpose of accomplishing optimal energy management and control of the microgrid operation. The Nash equilibrium was used to compromise the possible diverging goals of the agents by maximizing their preferences. The results obtained from the proposed energy management system, demonstrate that the application of game theory based control, in autonomous microgrids, can led to a higher efficiency, fewer power losses and a lower cost system.
Degree
thesis:*- Grantor dc:publisher
- Agricultural University of Athens
- Year dc:date
- 2019
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Karavas, Christos-Spyridon
- Καραβάς, Χρήστος-Σπυρίδων
Subjects
dc:subject × 26- Μικροδίκτυα
- Ανανεώσιμες πηγές ενέργειας
- Αφαλάτωση
- Συστήματα διαχείρισης ενέργειας
- Αποκεντρωμένα συστήματα διαχείρισης ενέργειας
- Υπολογιστική νοημοσύνη
- Ευφυείς πράκτορες
- Πολυπρακτορικά συστήματα
- Ασαφείς γνωστικοί χάρτες
- Θεωρία παιγνίων
- Βελτιστοποίηση με θεωρία σμηνών
- Microgrids
- Renewable energy sources (RES)
- Desalination
- Energy management systems
- Decentralized energy management systems
- Computational intelligence
- Intelligent agents
- Multi - agent systems
- Fuzzy cognitive maps
- Game theory
- Particle swarm optimization
- Επιστήμες Μηχανικού και Τεχνολογία
- Επιστήμη Ηλεκτρολόγου Μηχανικού, Ηλεκτρονικού Μηχανικού, Μηχανικού Η/Υ
- Engineering and Technology
- Electrical Engineering, Electronic Engineering, Information Engineering
Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/46279