Smart Grid Scheduling Strategy Based On Multi Agent Consensus Theory

Smart grid is the development trend of the future power grid. Smart grid scheduling strategy is one of the focuses of smart grid research. Due to the unique structure and components of smart grid, it is difficult to realize the effective scheduling of smart grid in the centralized scheduling method which is used in the traditional power grid. The smart grid system is a kind of multi-agent system. Each component of smart grid can be used as a single agent in a multi-agent system, such as generator, flexible load, energy storage unit. Therefore, considering the application of multi agent system, the consensus algorithm is applied to the smart grid dispatching. This thesis is devoted to the study of the smart grid scheduling problem by using multi agent system. The main contents of the thesis are as follows:(1) After the flexible load is connected to the smart grid, the significance of the economic dispatch in smart grid is that the generator and flexible load meet certain constraints, which makes the whole system to maximize the social welfare. The economic dispatch of the system is realized by the consensus algorithm of multi agent system. Firstly, this problem is solved by using the "Leader-Follower" consensus algorithm, and then the economic dispatch of the smart grid is designed based on the design of a fully distributed consensus algorithm.(2) With the rapid development and industrialization of the new energy storage technology, integrated mass storage unit that smooth the renewable power fluctuations becomes possible. By using a fully distributed algorithm, the power balance control of the power grid is realized by the real-time power allocation of the energy storage unit. Because of the internal resistance of battery energy storage system, the resistance will affect the efficiency of storage battery. The paper design a distributed optimization scheduling strategy which considered the internal resistance of the battery energy storage unit.In this paper, the verification of each distributed strategy is realized through the MATLAB simulation analysis. The effectiveness of the distributed consensus algorithm is verified by simulation results.