Game-theoretic approach to flexible demand-side energy management for a smart neighborhood
The increasing environmental and sustainability concerns raised by energy production from fossil fuels can be mitigated by integrating renewable energy sources that can increase affordability, availability and efficiency of the power system and reduce the harmful effects on the environment. Recent developments in smart grid technologies have enabled increased interaction between energy suppliers and consumers, leading to economic benefits for both entities. Energy savings can be achieved only through the optimal allocation of resources and optimal energy usage. The utilization of small capacity renewable energy generators contributes to providing optimal resource allocation and achieving an acceptable peak average ratio. This, however, requires interaction between various entities such as renewable power generators, storage systems and the supply grid. This interaction, data sharing, and demand management will reduce the peak average ratio, grid dependency and lead to the optimal sharing of resources. Consumers can shift their power consumption times to achieve reduced power bills and increased efficiency. Local energy markets (LEM)s can play a vital role in the energy transition by facilitating the rapid proliferation of such renewable-based distributed energy resources (DER)s thereby increasing the renewable energy storage hosting capacity of the power grid. This thesis aims to develop a demand-side management framework for individual consumers and a cluster of consumers by modelling their behaviours as cooperative or non-cooperative games. This efficient energy management system based on game theory, optimizes revenue and resource allocation, increases grid reliability and has many socioeconomic benefits. The initial part of this research aims at developing a real-time price for the power procured from the energy supplier and an algorithm that predicts the consumption of appliances, based on several parameters are developed. An algorithm is developed based on game theory and the Nash equilibrium for the scheduling of consumer appliances and reduction of the peak average ratio. A round-robin gaming method is followed until all consumers do not change their strategy. Game-theoretical analysis ensures that users do not make profits if they deviate from their assigned consumption patterns. The performance of various algorithms is evaluated, and their effects on the peak average ratio (PAR) and energy costs are discussed. The effectiveness of the proposed game-theoretic optimization model is validated and compared with traditional non-game-theoretic models. Simulation results are presented for load scheduling, during summer and winter with the developed real time prices. A community energy management system (CEMS) in which multiple prosumers can share energy between peers or the Community Storage Facility (CSF), is proposed in the next work. This brings in many entities with conflicting interests and leads to a bi-level optimization. The bi-level model is devised as a game in which the CSF entity is the leader of the Stackelberg game, and the consumers are the followers. In the consumer utility function maximization problem, the intersection of the best responses in the Nash equilibrium point of the game maximizes the utility function of each consumer with respect to the strategies of other consumers. The simulations were carried out for two price scenarios from the grid: flat tariff and real time tariff. In the following work, the energy management framework is devised as an evolutionary game played between buyers to select sellers and a Stackelberg game between buyers and sellers. A prosumer with excess energy can be a seller, while an energy deficit prosumer can be a consumer. Multiple games are modelled in which consumers can choose from a list of sellers, and sellers can choose from a list of buyers. It can be proved mathematically that equilibrium can be achieved in the evolutionary game. The dynamics of the evolutionary game can be approximated iteratively. The proposed demand side management (DSM) strategies are tested in real power networks using real load data, weather data, and price data. The algorithms can effectively schedule loads and manage power allocation to achieve optimum results. The strategies have exhibited robustness to varying input conditions, and the results show that these algorithms, if implemented, can economically benefit prosumers. A comparison of these algorithms with existing strategies shows a significant improvement in terms of energy efficiency, economic benefits, and optimal resource allocation for smart buildings. The final work of this research is to develop an algorithm that coordinates the power-sharing of the appliances in a building utilizing the available resources based on operational modes selected by the consumer such as maximizing DER utilization, or priority-based modes. In both modes, the algorithm maps appliances to resources in each individual building and shares the extra power with other buildings in the community. In the DER utilization mode, less energy is shared with their peers. Abidding mechanism, initiated by the controller, is designed to showcase the functionality of the CSF to achieve a high grid demand ratio (GDR). A Message Queue Telemetry Transport (MQTT) based communication architecture is implemented through a two-way communication link between heating, ventilation, and air conditioning (HVAC)s, data aggregator and the central controller.