Probabilistic transient stability assessment of power systems with renewable energy using a bounded-Transformer embedding fault severity rules

The increasing penetration of renewable energy introduces significant uncertainty into modern power systems, posing an unpredictable challenge to their stable operation. However, conventional transient stability assessment methods only provide a deterministic result, ignoring the effects of uncertainties. This paper proposes a probabilistic transient stability assessment method based on a Bounded-Transformer model, which can consider the effects of uncertainty induced by renewable energy and the impacts of assessment errors. First, a Bounded-Transformer model is designed to automatically predict the most probable transient stability index (TSI), along with its upper and lower bounds. The upper and lower bounds of TSI can indicate the potential error induced by model uncertainty, thereby improving the reliability of the assessment results. Second, physical fault severity rules (FSRs) are introduced into the training process, enabling the model to learn the inverse relationship between fault clearing time and stability margin without requiring additional labeled samples. Finally, by combining the proposed model with Monte-Carlo sampling, an aggressive stability probability, a conservative stability probability, and a common stability probability can be derived for the reference of operators. The effectiveness of the proposed method is verified in a modified IEEE 39-bus power system, a modified WECC 179-bus power system, and a realistic provincial power system.