Research On The Application Of Agent Technology In The Smart Grid

Now, many of fields and technology of smart grid have become research hot point.About the research of generation side, the work of the wind and solar complementary systemis a new attempt, at the same time adopting the Agent technology to accomplish thecoordination control of the wind and solar complementary system also will be an effectivemethod. About the research of the power distribution side, the structure of the distributionnetwork optimization is an eternal theme. A lot of artificial intelligence algorithms emerged inlarge numbers. Adopting Agent technology to improve the characteristic of some algorithmsis reasonable idea. The following is the main work about my paper:1. At first, about the wind and solar complementary system, I research the principle ofoperation deeply. Adopting the idea of breaking up the whole into parts to research the windpower subsystem and photovoltaic subsystem respectively. There are two kinds of controlstrategies about two subsystems that is the maximum power point tracking control strategyand load power tracking control strategy. Building the simulation model in matlab to test thesimulation effect. For accomplishing the coordination control, designing a multi-Agentsystem, I will abstract the output unit that is wind power subsystem, photovoltaic subsystem,storage battery subsystem to wind power Agent, photovoltaic Agent, storage battery Agentand so on. At the same time I also designed a variety of control strategies Agent that ismaximum power point tracking control Agent of wind power subsystem and photovoltaicsubsystem, load power tracking control Agent of wind power subsystem and photovoltaicsubsystem. At last, designing a centering control Agent for completing the coordinationcontrol function successfully. These Agents adopted the design philosophy of layering andused the JADE to develop. We can observe the process of Agents’ coordination using thesniffer on the JADE.2. About the research of the network reconfiguration algorithm, I use the learningcharacteristic of Agent to improve the binary particle swarm optimization algorithm(PSO). Iwill abstract the particle to become an Agent and design a multi-Agent system. Using thecharacteristic of Agent can make every particles in the algorithm cooperate and compete withother particles in the neighborhood. By the way can the algorithm maintain the diversity ofparticles of PSO, meanwhile the way also can solve the problem of locally optimal solution inPSO.