ANN-Based Resynchronization Control for Virtual Synchronous Generator

To address the problem of loss of synchronization (LOS) with grid-forming converter, an artificial neural network (ANN)-based resynchronization control method is proposed in this work. The approach uses a pre-trained ANN to predict the stability boundary and applies a state limiter to guide system states back into stable regions, ensuring reliable synchronization under various conditions. A multi-task software architecture enables realtime control and neural network inference on embedded controllers with limited resources. The method also integrates online grid impedance estimation, and is validated through both simulation and hardware-in-the-loop (HIL) experiments using a TI 280049C microprocessor. Results demonstrate that the proposed strategy achieves effective and real-time resynchronization control, supporting practical application of GFM converters.