基于物理-信息关联分析的智能电网安全检测方法

Attackers can avoid the information network and power system security methods to attack smart grids infromation system, tamper the data and crash the system, using the vulnerabilities and connection of the information network and power system. In our research, a micro smart grid is established in our lab to investigate the security problems in smart grids. The research shows that it is difficult for the information network and power system security methods to protect smart grids, since the communication protocols and system architechture of smart grids are difference. Thus, three contents are proposed in this proposal: 1) a traffic flow model for smart grids and an entropy-based traffic monitoring method; 2) a distributed state estimation-based method for the bad power data detection; 3) a traffic abnormal weighted method for bad power data identification and locating, and the traffic abnormal and bad power data are fused to detect smart grids attacks..This proposal targets on the latest problem with high theoritical value. The traffic features of information network and physical characteristics of power system are fused to monitor the smart grids, based on the idea of the Cyber-Physical System and the interdisciplinary study of power system, computer network and system engineering.

在智能电网中，攻击者利用信息网络和电力系统的缺陷及两者间的耦合关联，躲避传统信息网络和电力系统的安全监控，入侵智能电网信息系统，进行数据篡改、系统破坏等恶意行为。通过搭建智能微电网并进行攻击测试，发现由于通信协议、系统结构等因素，导致传统信息网络和电力系统的安全监控技术在智能电网环境中的使用受到限制。为此，本项目重点研究：1）智能电网通信流量建模方法和基于信息熵的异常流量检测方法；2）基于分区状态估计的智能电网异常电力数据检测方法；3）基于流量异常检测结果加权的电力系统不良数据辨识方法，关联信息网络和电力系统的检测结果实现对智能电网的安全监控。.本项目具有良好的前瞻性和理论价值，通过综合运用电力系统、计算机网络、系统工程多个学科的理论方法，从物理-信息融合系统的角度，解决智能电网中信息网络与电力系统的监测与关联分析问题，实现对智能电网的安全监控。

本项目重点研究：1）智能电网通信流量建模方法和基于信息熵的异常流量检测方法；2）基于分区状态估计的智能电网异常电力数据检测方法；3）基于流量异常检测结果加权的电力系统不良数据辨识方法，关联信息网络和电力系统的检测结果实现对智能电网的安全监控。.相关研究成果先后在IEEE Trans. on Smart Grid、IEEE Trans. on Fuzzy Systems等高水平期刊以及ACM SIGCOMM、IEEE SmartGridComm、ACM SAC等国际顶级会议上发表论文15篇。获得发明专利5项，软件著作权2项。