Artificial neural network-based virtual synchronous generator dual droop control for microgrid systems

In microgrid systems, traditional virtual synchronous generator (VSG) face challenges in adaptively converting the virtual inertia J and damping coefficient D following system changes. This can lead to significant system frequency and voltage deviations, resulting in unreasonable power distribution. This paper proposes an artificial neural network (ANN)-based VGS dual droop control strategy tailored for microgrid systems. The study initially analyzes the influence of moment of inertia and damping on VSG, and subsequently establishes adaptive rules through angular frequency deviation and its change rate. By harnessing ANN algorithm, virtual coefficients are adjusted under different working conditions. This ensures stable f-p and Q-U dual droop control, minimizes frequency and voltage deviations, and facilitates optimal power allocation. The effectiveness of the proposed strategy is verified through MATLAB simulations.