Independent Microgrid Transient Stability Evaluation Model Based on a Message Passing Graph Neural Network

In isolated microgrids, virtual synchronous generators (VSGs) are utilized to enhance frequency stability. However, the presence of these generators can lead to undesirable levels of voltage or current during transient angle instability. The current Lyapunov transient stability analysis method is challenging to utilize due to the complexity of acquiring the energy function. Furthermore, existing data-driven microgrid transient stability methods do not easily adapt to changes in topology. Inaccuracies may also be present in existing stability analysis methods, as they fail to account for nonlinearities such as saturation and dead zone. To address these issues, we propose a microgrid transient stability analysis method based on a message-passing graph neural network (MPNN). When the topology changes, the information bus and transmission line are aggregated and updated through the message-passing mechanism in the graph neural network, and we enhance the approach by introducing current saturation as an additional feature to the transient stability analysis model. We present simulation results for a microgrid test case to validate the effectiveness of the proposed method, and the evaluation accuracy is maintained above 98%.