Research And Design Of Load Control System Based On Multi-Agent System For Smart Grid

With the global resources and environmental pressures increasing, the environmental protection, energy saving and sustainable development are required in the society. The future network demand should be able to provide a more secure, reliable, clean and high quality supply of electricity. Therefore, the different countries and organizations, such as the United States and the European Union, spontaneously proposed to set up flexible, clean, safe, economical, friendly, smart grid. Smart Grid is future direction of the grid.The development of Smart Grid make higher request for the load forecasting, monitoring and optimizing system. Many experiences of Smart Grid analyzed by overseas shows that user of power system requirement higher power stability, environmental protection and comfort level than before. This paper leads MAS (Multi-Agent System, MAS) into the load control system for the grid, enhancing the intelligence of control system to meet the needs of Smart Grid.On the basis of a large number of relevant literatures, the paper analyses the relevant technology and the condition of development of Smart Grid in China and foreign countries, and the classification of the electrical equipment for Smart Grid, and study the theory of Smart Grid. Analysis the load control system for grids.Research the method of Agent, analysis its characteristics and structure, and then focus on the existing algorithm of MAS. The method of Smart Grid load control for user-side is presented, the model for the Smart Grid load forecasting, monitoring and optimizing is established based on the MAS algorithm. Load control for Smart Grid is achieved with the MAS with distribution, and overall load information is fully used. The performance of load control system is improved.The program of load control system for Smart Grid is written in the JADE environment, MAS-based user-side load control system for Smart Grid is achieved, and the MAS feasibility of applying on load control is proved. A comparing experiment of simulation is done, and analysis of results was compared with the traditional method. Experimental results show that the proposed MAS-based load control methods for Smart Grid is superior, avoiding no solution in traditional algorithm. MAS can still set analysis according to inference in sudden accident or the case of join the new electrical equipment, while the traditional system could not recognize or failed without effective training, which shows that the MAS has better applicability.