Algorithm comparison.

With the rapid development of smart grids, power grid systems are becoming increasingly complex, posing significant challenges to their security. Traditional network intrusion detection systems often rely on manually engineered features, which are not only resource-intensive but also struggle to handle the diverse range of attack types. This paper aims to address these challenges by proposing an automated DDoS attack detection algorithm using the Informer model. We introduce a windowing technique to segment network traffic into manageable samples, which are then input into the Informer for feature extraction and classification. This model captures both the temporal dependencies and global attention information in the traffic data. Experimental results on the CICIDS-2018 dataset demonstrate the effectiveness of our approach, showing significant improvements in detection accuracy and efficiency. Our findings suggest that the proposed method offers a promising solution for real-time intrusion detection in complex power grid environments.