Research On Control Algorithms Based On Multi-agent Technology Of Microgrid

In recent years, research on Microgrid is hot spot and emphasis of researching on power system at home and abroad, and control is one of the key issues to be resolved in Microgrid in the actual operation. This thesis research concentrates on the control of microgrid and expanded. The thesis discusses the development of Microgrids at home and abroad, the concept and structure of Microgrid, as well as some of the key technologies of Microgrid. Multi-agent technology has the autonomy, social, reaction, coordination, and has a strong reasoning ability, and self-organization and learning ability, can solve the problem of microgrid control. Therefore the paper proposed the Microgrid control of multi-agent technology, to design and discribe the fucntion of multi-agent, to describe the specific function of microgrid of agent types, and also discusses the communication model of agent and workflow of the multi-agent control. As the multi-microgrids have much micro power and can change difficult, the system not only has a single Agent interaction with the environment, as well as the interaction between the Agent and the Agent. Based on this, the paper gives a learning algorithm which is true of mixed environment for dynamic multi-agent. The algorithm was originally based on the Q learning algorithm was improved to apply to a mixed environment of learning between the various Agent. The proposed algorithm is mainly an increase in the original algorithm on the other variables that Agent and the corresponding local time limits. Fanally the IEEE 9 nodal system as a basic framework emulates the micro power control, and the result shows that multi-agent technology based on micro-grid control framework is effective in the management of micro source.