DCNN-based virtual synchronous generator control to improve frequency stability of PV-ESS station

VSG (VSG) is proposed to complement the inertia and damping of the power grid. Compared with the synchronous machine, the VSG can flexibly adjust its inertia and damping parameters. In this paper, a data-driven adaptive control strategy for VSGs is proposed based on deep convolutional neural network (DCNN). This strategy uses the good approximation effect and learning ability of DCNN for nonlinear functions to extract the dynamic characteristics of VSG and adjust the virtual inertia and damping coefficient online. The simulation model is established in Matlab and the NN control algorithm is integrated to verify the proposed control strategy. Simulation results show that the adaptive control strategy can effectively improve the frequency dynamic performance of PV-ESS station.