Fixed-frequency Current Control Method of Islanding Micro-grid Based on Improved Neural Network

Abstract

In the islanding micro-grid operation mode, due to the lack of support from the large power grid, the voltage of the bus and each node in the network is completely supported by the cooperation of the micro-grid inverters in the grid. Therefore, the control performance of the micro-grid inverter determines the quality of the power supply voltage. This paper proposes an improved AC and DC islanding micro-grid fixed frequency current control method with variable topology. First, the neural network algorithm improved by particle swarm optimization is used as the basis to optimize the coordinated compensation control of the micro-grid to obtain the fitness value of the objective function under the positive and negative sequence potentials. Then, a control strategy for the improved AC/DC hybrid micro-grid is proposed, and the decoupling and coordinated control strategy for the output compensation of the dual filters in the islanding mode is designed. Finally, simulations and experiments verify the improved AC/DC hybrid microgrid with variable topology and achieve the goal of constant-frequency current control.