智能电网邻域网分级式入侵检测系统研究

Research on Hierarchical Intrusion Detection System of Smart Grid Neighborhood Area Network

近年来,智能电网逐渐成为世界电力工业的共同发展趋势。邻域网作为智能电力通信网络的中心枢纽,其安全问题越来越被研究者们所重视。目前,已应用物理控制、数据加密以及认证等技术提高邻域层安全性,但仍缺乏及时有效的检测方法来避免电网受到恶意入侵行为的威胁。针对此问题,本文设计和实现了一个高效轻量级的智能电网邻域网分级式入侵检测系统。在解决方案中,我们把邻域层网络分为三个层级。根据不同层级设备以及通信的特点,采用相对应的检测方式实现上层级对下层级的行为检测。同时,本文提出结合基于信任度的规则检测和基于SVM机器学习算法的异常检测,旨在将这两种检测技术的优点结合起来检测多种入侵。经仿真实验验证,本文提出的入侵检测方案以较低的能量消耗获得高检测率,符合预期结果。

In recent years , smar t gr id has gradual ly become the common development tr end in the whole wor ld power industry. As the center connection of smart grid communication network, the security problem of the neighborhood area network is becoming more and more important to the researchers. At present, physical con仕 ol , data encryption and authentication are used to improve the security of neighborhood, but there is still a lack of timely and effective detection method to avoid the threat of malicious intrusion. Aiming at this problem, this paper designed and implemented an efficient and lightweight hierarchical intrusion detection system for smart grid. In the solution, it divided the neighborhood layer network into three levels. According to the characteristics of different levels of equipment and communication, using the corresponding detection method to achieve the behavior detec-tion of upper level to the next level. At the same time, this paper proposed a combination of trust-based rule detec-tion and anomaly detection based on SVM machine learning algorithm, aiming to combine the advantages of these two detection techniques to detect multiple kinds of intrusion. The simulation results show that the intrusion detection scheme proposed in this paper achieves high detection rate with low energy consumption and meets the expected results.