Microgrid Coordination Control Research Based On Multi-Agent System

With advantages of less pollution, higher energy efficiency, saving power transmission and distribution resources and operation cost, less transmission line loss and so on, microgrid received widespread attention. Microgrid and the main power grid of joint operation mode make safer and reliable power supply. Multi-agent system has good autonomy, adaptability, coordination and sociality and it will ensure the control of micro power grid, thus make the operation of power grid more stable.Firstly, the definition and advantage of microgrid and multi-agent system have been introduced, and analyzes the latest research status of microgrid and multi-agent system. Then, the related contents of distributed generation have been introduced, including the basic principle and mathematical model of the wind turbines, photovoltaic cells, micro turbine and fuel cells.Secondly, using the multi-agent system to realize the smooth transition of microgrid operation modes.The multi-agent system contains one upper coordination level agent and three lower execution level agent. The coordination level agent connect main distribution network agent and execution level agent, overall coordinating the agent controlling and using the inference module and algorithm moduleto optimize the operation of the whole microgrid. According to the requirement of the distributed power grid and the detecting result of islanding detection agent and connecting-grid detection agent, coordination level agent can determine whether microgrid parallel operation. Islanding detection agent uses wavelet neural network method to detect island. At the same time, the main power grid faults will be detected, and communicate with the coordination level agent to convey the test results or receive instructions. Control level agent uses different control methods in different operating modes by receiving coordination level agent instruction. Control methods include P/Q control method when interconnecting to the power grid, the off-grid V/f control method and pre-synchronization control method. The connecting-grid detection agent runs the power grid interconnection test before microgrid interconnection, and calculate the amplitude difference, phase difference and frequency difference between the microgird voltage and the main power grid, then conveys test results to the coordination level agent. Using the MATLAB/Simulink to model and run the real-time simulation by S-function, then using the multi-agent system to realize the smooth transition of the microgrid’s operation modes.Finally, introduced the economical model of wind turbine, photovoltaic cells, micro turbine, fuel cells and storage battery. Using the multi-agent system to realize microgrid’s economical optimization operation with particle swarm optimization (pso) algorithm. According to the requirements of load and the output power of each distributed generations, the coordination level agent uses pso algorithm and reasonably distributes the load to the trickle source, so as to minimize the cost of power generation. According to the real-time electricity price and load value of the 48 time periods in a summer day, considering fuel cost, operation maintenance cost, gas emissions overhead expenses, cooling earnings ect. Runs the example simulation on the MATLAB and maximizes the profit of the microgrid.