Multi-agent Game Collaborative Planning of Complex Distribution Network Considering the Bounded Rationality of Wind Power Output

Abstract

A high proportion of new energy dominated by wind energy is connected to the distribution network, and its output uncertainty poses a threat to the safe and stable operation of the distribution network. This paper presents a multi-agent game collaborative planning method for complex distribution networks considering the bounded rationality of wind power output. Firstly, sort out the game relationship between nature and distribution network source-network-load side interest subjects, and establish the income models of different subjects. Then, a multi-agent game collaborative planning model of complex distribution network considering the bounded rationality of wind power output is constructed. The inner layer is the evolutionary game between nature and distribution network operators. Under the condition of considering the demand response on the user side, the best planning scheme of distribution network frame and energy storage considering the bounded rationality of wind power output is decided. The outer layer is the game between distributed generation operators and distribution network operators. Under the assumption that the game subject is completely rational, the configuration strategy of distributed generation is decided. On this basis, the idea of artificial potential field and information entropy is introduced, and an evolutionary game solution method based on stage adaptive step size is proposed. Finally, it is verified by a simulation example. The results show that the planning model constructed in this paper effectively reduces the conservatism of the traditional planning method considering the complete rationality of wind power output, and is closer to the actual situation. At the same time, the evolutionary game solution method proposed in this paper greatly reduces the number of iterations of the game and improves the convergence speed on the basis of ensuring the reliable convergence of the model.