PINNs-Driven Transient Estimation in Power Systems with the Second-Order Kuramoto Model

This paper presents the application of Physics-Informed Neural Networks (PINNs) to the second-order Kuramoto Model for enhancing transient estimation in modern power systems. Amidst growing end-use electrification and integration of distributed resources, PINNs offer a novel approach to state estimation, addressing the limitations of traditional methods with improved computational efficiency and nonlinear dynamics handling. The second-order Kuramoto Model is employed to analyze the impact of nodal admittance matrix parameters on the transient stability of a Single Machine Infinite Bus (SMIB) system, with cross-validation studies using CloudPSS platform extending these findings to the IEEE 39-Bus 10-Machine system. Results reveal that these parameters significantly influence system synchronization capability and stability, with an optimal coupling strength enhancing synchronicity, while extreme values can lead to instability.