Intrusion Detection Model For Power Information Network Based On Multi-layer Attention Mechanism

Abstract

As the construction work of energy Internet and new power system continues to advance, the scale of power information network continues to expand, and network attacks against power systems are becoming more and more frequent. Although the existing intrusion detection technologies can defend against some network intrusions and attacks, the detection efficiency and accuracy of related algorithms need to be improved. In this paper, we propose a multi-layer attention mechanism-based intrusion detection model for power information network, which divides feature extraction into two different attention layers at feature level and slice level, and combines recurrent neural networks to fully exploit the value of historical information and current input. Experiments prove that the intrusion detection model proposed in this paper improves the detection accuracy at a lower cost, which in turn improves the security of power information network.