This article proposes an intelligent prediction model for power network security situation that integrates fuzzy analytic hierarchy process (FAHP) and deep learning (DL), aiming to enhance the early warning and response capabilities of the electrical power systems (EPS) to network attacks. Network security situation assessment is a prerequisite for achieving effective early warning. However, in the complex and ever-changing network environment, various types of attack behaviors occur frequently, which not only exacerbates network load, but also may cause sudden failures, posing a serious threat to the stable operation of the power system. To address this challenge, this article first introduces the FAHP method, which comprehensively considers multi-level and multi-dimensional security indicators to scientifically and reasonably quantify the current security situation of the power network. On this basis, combined with the powerful temporal modeling capability of Long Short Term Memory (LSTM) networks, dynamic predictions of security situations can be made to achieve early warning of potential attack behaviors. The results indicate that our model can not only accurately evaluate the current network security status, but also achieve high-precision real-time situation prediction. This study provides a feasible technical path and theoretical support for building an active defense type power network security system.
Chen et al. (Sun,) studied this question.