Secondary Frequency Regulation Strategy of Multi-agent Virtual Power Plant Based on Scenery Storage

The increasingly mature related technologies of multi-agent system (MAS) and virtual power plant (VPP) provide new ideas for supporting the automatic generation control (AGC) demand of the bulk power grid. In this paper, a multi-agent VPP hierarchical control strategy based on scenery storage is proposed. The upper layer is responsible for receiving the AGC requirements of the grid and using grey wolf optimizer (GWO) to solve the optimization target and output power's reference value. In order to improve the accuracy of GWO, the population members of the algorithm are optimized, the linear convergence factor is replaced, and the Levy flight strategy is adopted. After the iteration of the algorithm, the random step size is used to search the current range in a small range. The lower layer constructs the corresponding model prediction control (MPC) model based on the state space equations of wind turbine, photovoltaic and energy storage system. The output value of the upper layer is used as a reference for predictive tracking to ensure that the output power can track the reference value shortly. The simulation results show that the hierarchical control strategy based on MAS can quickly and accurately respond to AGC demand and maximize economic benefits, and the improved GWO performance is better than GWO.