A supervised contrastive learning method with novel data augmentation for transient stability assessment considering sample imbalance

Yaodi Huang, Yunpeng Song, Zhongmin Cai

Abstract: Highlights•Propose a novel balanced supervised contrastive learning method for TSA.•Present a novel data augmentation method in the contrastive learning framework.•Effectiveness of method is verified on imbalanced datasets from two power systems.•High generalization capability across various deep learning methods.