Analysis of GRU Based Dynamic Modeling Method for Microgrid With SMES

As the deployment of superconducting magnetic energy storage (SMES), the transient behavior of grid-tied microgrid (MG) systems becomes more complex. To accurately simulate the transient behavior of distributed generation units in MG systems, researchers have developed dynamic modeling methods based on deep neural networks (DNNs). These methods use terminal data to simulate transient behavior without the necessity of system parameters. However, little attention has been paid to the influence of sample distribution on their performance. This paper proposes a novel dynamic modeling approach based on Gate Recurrent Units (GRUs) for MG systems with SMES, which takes into account the influence of sample distribution on model training. Several modeling suggestions are provided to achieve high accuracy. The effectiveness of the proposed method is verified through simulation results.