The integration of grid-forming (GFM) converters into strong AC grids introduces new stability challenges, particularly low-frequency oscillations, emphasizing the need for real-time small-signal stability monitoring. The critical short-circuit ratio (CSCR) is a widely adopted metric for assessing small-signal stability in renewable power systems, but its dependence on grid operating conditions and converter control parameters requires online evaluation under diverse scenarios. This paper proposes an online CSCR prediction model for GFM grid-connected systems, based on a hybrid support vector regression (SVR) and particle swarm optimization (PSO) approach, enabling real-time stability margin assessment. First, a detailed 12th-order state-space model is established, incorporating both grid-side dynamics and converter control dynamics. Impedance-based sensitivity analysis is performed to examine the impact of key control parameters on CSCR. Next, an SVR-PSO prediction model is developed, trained on data generated from the state-space model. Experimental results demonstrate that the SVR-PSO model achieves superior accuracy in estimating CSCR compared to conventional methods. Using the predicted CSCR, this paper derives a small-signal stability margin index and validates its effectiveness through detailed case studies. Simulation results confirm the model’s high-fidelity prediction capability and its applicability for online stability assessment in strong grid conditions. This work provides a data-driven, computationally efficient framework for real-time stability monitoring in GFM-integrated power systems, offering practical insights for grid operators to ensure stable grid operation with high renewable penetration.
Dong et al. (Tue,) studied this question.