An Investigation On Toplogy Identification Of MicroGrid Based On MAS

Microgrid is a complex dynamic network with variable topologies and consists of a variety of distributed generations, energy storage devices, loads and control devices. The operating characteristics of the DG including wind turbines and photovoltaic generator will change with the climate. The energy storage and power supply functions among the energy storage device nodes lead to the change of connection characteristics. In addition, the failure of equipment can also change the microgrid topology. At the same time, the microgrid with flexible extensible nature can not only run with interconnection of grid, but also run itself, which can also cause the microgrid topology change.The structure of microgrid is different from the structure of the transmission networks and the distribution network, leading to the difference in topology analysis process. Microgrid topology analysis should be based on the changing topology, and is the basis for a number of application such as failure diagnosis, energy management, trend analysis. Therefore, network topology calculation need to be further improved, as well as a good network topology identification algorithms.To overcome the problem and enlarge the applicable area, the main contents of this dissertation are as follows:(1) The paper firstly analyzes the structure of microgrid. An control strategy is putted forward based on synovial control for the inverter which is importance connection interface of microgrid. In order to improve the rate of microgrid topology identification, microgrid is descripted based on the knowledge of graph theory. The mathematical model of microgrid is expressed by adjacency matrix notation.(2) The strategy of microgrid topology identification based on MAS is putted forward. The Agents with different functions are designed. The system architecture of MAS is designed for topology identification of microgrid. In addition, the improved ant colony algorithm is proposed for MAS communication to find the "blackboard" Agent, contributing to the acquisition of the shortest distance, less time-consuming and high efficiency for each Agent communication in the total system.(3) The paper simulates three methods of depth-first search method, breadth-first search method, intelligent MAS method based on MATLAB and analyzes the advantages and disadvantages of these three methods.(4) At last, in order to validate the algorithms, the laboratory of microgrid is constructed and the experiment is verfied.