Research On Multi-source Joint Frequency Regulation Control Sreategy And Energy Storage Capacity Allocation Based On Improved Multi-agent Q-learning

In recent years,in order to promote the utilization rate of renewable energy and build a clean,efficient,economical and low-carbon new energy system,renewable energy led by wind power and solar energy have received attention and support from governments around the world.The proportion of forms integrated into the grid is growing rapidly.Compared with traditional units,wind and solar are easily restricted by the environment,and the output of electric energy has large fluctuation and strong randomness,moreover,problems such as "abandoning wind and light" are prone to occur,and it also brings huge challenges to the frequency modulation of the power system.Under the current situation,the traditional thermal power unit as the main frequency modulation method cannot meet the demand for high-quality frequency modulation.Therefore,this paper proposes a multi-source control strategy of joint frequency regulation based on improved multi-agent Q-learning and configures the capacity of the energy storage system.First of all,in order to make renewable energy and energy storage devices to have the ability to cooperate with traditional thermal power units to participate in primary frequency regulation,in this paper,by analyzing the power generation principles and mathematical models of wind power,solar energy and energy storage.The virtual inertia and droop control strategies of the wind turbine,the load shedding operation strategy of the solar energy and the droop control strategy of the energy storage device are designed respectively.And finally the purpose that each unit can quickly respond to the system frequency change is realized.Secondly,for the problem that the frequency deviation can be reasonably allocated to multiple types of energy sources,this paper adopts an improved multi-agent Q-learning algorithm to solve the problem.The pre-learning result is selected as the initial matrix of the algorithm and a search factor is introduced on the basis of the greedy strategy,so as to improve the algorithm computing performance.In the PSCAD/EMTDC platform,a three-machine nine-node system including wind,light and storage is built,and it is verified that this method can realize multi-source coordination and improve the frequency stability of the system under two load disturbance conditions of step and random square wave.Finally,the energy storage capacity configuration problem of the multi-source combined system is studied,considering the overall frequency regulation effect and economy of the system,and aiming at the minimum capacity of the energy storage configuration,this paper establishes an energy storage system based on frequency stability index and economic index.The multi-objective optimization model of capacity is used to solve the optimal value through particle swarm algorithm.Finally,experiments are carried out under two disturbance conditions,and the effectiveness and economy of the optimization results are verified by the frequency change curve.