Data-driven transient stability assessment of power systems with renewable energy integration and FACTS-based control strategies

With the increasing penetration of renewable energy, the dynamic uncertainty of grid transient processes has grown, posing higher demands on transient stability assessment. Existing studies are mostly based on standard IEEE test systems or fixed topologies, which fail to accurately reflect the dynamic evolution characteristics of systems with high renewable energy penetration. Moreover, systematic research on mechanisms for enhancing transient stability thresholds remains relatively scarce. To address these issues, this paper proposes an improved transient stability assessment method tailored for scenarios with renewable energy integration, and develops a framework that combines spatiotemporal modeling with control enhancement mechanisms. First, based on the IEEE 39-bus system and the NPCC 140-bus large-scale practical grid, a Flexible AC Transmission System (FACTS) device is introduced, and a targeted control strategy is designed to actively suppress power angle oscillations and voltage fluctuations, significantly improving the system’s dynamic stability. Secondly, to mitigate performance degradation due to topology changes, a model training mechanism combining transfer learning and sample augmentation is proposed. Hierarchical clustering is employed for efficient feature transfer and classification optimization, enhancing the model’s robustness and generalization ability in dynamically changing network structures. Simulation results show that the FACTS control strategy effectively improves the system’s transient stability margin under fault disturbances. The proposed Spatio-Temporal Gated Hybrid Attention Network model demonstrates superior accuracy and robustness across various topology changes, with relative errors controlled between 0.2% and 0.6%, validating the practical and engineering applicability of the method under complex operating conditions. • Expanded IEEE 39-bus and NPCC 140-bus systems with wind, solar, and energy storage. • Applied PI control, feedforward, and adaptive gain tuning in VSC; fuzzy control in STATCOM. • Proposed ST-GHAN model combining SW-DGA and MHSA for power system feature modeling. • Achieved fast adaptation and better generalization using clustering and parameter tuning.