Optimization Methods For The Operation Of Intelligent Electricity Distribution And Utilization Systems Based On Game Theory

Intelligent Electricity Distribution and Utilization System(IEDUS)is an important component of strong smart grid.With the reform of the energy system and power marketing as well as the development of the automatic demand response(ADR)technology,IEDUS also faces big challenges.Conventional optimization theory based on single subject decision is difficult to deal with these changes.In this context,the game theory for multi-objective optimization of complex subjects is expected to be a powerful tool to overcome many key problems of smart grid.This dissertation,which mainly studies the energy management of Smart Building Cluster(SBC)and community energy Internet based on game theory,contains the following work:We establish a typical application framework of game theory in IEDUS,including peer-to-peer non-cooperative game,leader-follower non-cooperative game,and cooperative game models.The basic application processes of these three types of games are proposed.We analyze the various boundary conditions involved in the application of game theory to IEDUS,and point out the relationship between the game type selection and the boundary conditions.The theoretical basis of the game models used in this dissertation is also refined.We propose a practical equivalent model of shiftable load for the large number of discrete variables in the demand side response problem.The original large-scale discrete variable optimization problem is transformed int o a continuous variable optimization problem.The equivalent expressions of the optimization variables and the constraint functions are given.The recovery method of the equivalent model and the application process in the practical optimization problem are proposed.The validity of the equivalent and recovery method is verified by an example.A joint operation system framework for SBC is proposed,and the information exchange processes between the smart building cluster operator and the participating smart buildings are described.A day-ahead distributed energy management model for SBC based on peer-to-peer non-cooperative game theory is proposed,considering building-integrated PV systems and ADR,with all participating smart buildings viewed as players in the game.The price model takes into account the photovoltaic price policy and inverse power situation,and the smart buildings' cost involves many aspects,including the cost of electricity,operators' operating cost,comfort costs and PV subsidies.The existence of Nash equilibrium in the game model is proved.Furthermore,the solution method and procedure is also proposed.Finally,via a practical example,the effectiveness of the model is verified in terms of reducing the smart buildings' total cost,improving the SBC's load characteristic,and avoiding the new load peak in low price period in the time-of-use price model.Based on the leader-follower non-cooperative game theory,a day-ahead distributed energy management model for community energy Internet is proposed,with community manager viewed as leader and prosumers viewed as followers.Based on the microturbine's heat-led mode,the selection method of prosumers' trading mode,models of prosumers' and community manager 's loads and profits are put forward respectively.The existence of Stackelberg equilibrium in the game is proved,and the solving method and process are also proposed.Via a practical example,the effectiveness of the model is verified in terms of improving profits of community manager and prosumers,and optimizing system load characteristic,etc.Taking the community energy Internet as the research object,we put forward a real-time energy management model based on cooperative game theory.At the same time,we propose the realtime energy management model of community energy Internet under the traditional transaction mode as the benchmark.The model considers the uncertainties of the predicting variables for the future time,and proposes the multiscenario determination method based on the probability distribution of the prediction error,which takes into account the current time electricity cost,the future time cost expectation and the conditional value at risk in the optimization target of the prosumers' cost.We also put forward the method of profit distribution based on fairness principle and prove the stability of the coalition.The optimal model with objective of segmentation functions is transformed into a mixed integer linear optimization problem,and the implementation process of the real-time energy management model is proposed accordingly.The case study shows that the cooperative model of centralized optimization is more advantageous than the traditional model in the aspects of increasing the income of the community operator and reducing the costs of the prosumers.