Optimal Stepwise Inertial Control of Wind Turbine Based on Fuzzy Control and Deep Neural Network

With the increasing installed wind power capacity, large-scale wind power integration into the grid will reduce the equivalent inertia of the power system and weaken the frequency response characteristics of the system. In order to improve the system frequency control capability, the wind turbine (WT) needs to participate in the grid frequency regulation process. The WT adopts the stepwise inertial control (SIC) to provide a step power increase after load fluctuation to participate in the grid frequency regulation, which helps to suppress the lowest frequency drop. However, the traditional SIC strategy is decoupled from the frequency of the power grid, which causes the WT cannot perceive the change of the frequency. And the serious secondary frequency drop (SFD) will occur when switching the reference power curve during the speed recovery stage. Therefore, this paper proposes an optimal stepwise inertial control of wind turbine based on fuzzy control and deep neural network(DNN) to establish the relationship between frequency state quantity, speed change quantity and SIC strategy so that the WT can dynamically adjust the output power according to the system frequency change and WT speed. The strategy can avoid the secondary frequency drop while improving the performance of the WT frequency regulation. The fuzzy control parameters are optimized by particle swarm optimization (PSO) and the optimal fuzzy control parameters are learned by DNN. The simulation results show that the proposed strategy can overcome the shortcomings of the traditional SIC strategy, effectively smooth the switching process of WT frequency regulation state and make the frequency regulation process more stable. The lowest point of the first frequency drop is better than the traditional SIC strategy and SFD is eliminated.