James Jianqiao Yu
余剑峤
Home Publications Services

Lecturer

Department of Computer Science

University of York

CSE/139, YO10 5GH, UK

jqyu(at)ieee.org CV Google Scholar
Online False Data Injection Attack Detection with Wavelet Transform and Deep Neural Networks

Authors
James J.Q. Yu, Yunhe Hou, and Victor O.K. Li

Publication
IEEE Transactions on Industrial Informatics, Volume 14, Issue 7, July 2018, Pages 3271--3280

Abstract
State estimation is critical to the operation and control of modern power systems. However, many cyber-attacks, such as false data injection attacks, can circumvent conventional detection methods and interfere the normal operation of grids. While there exists research focusing on detecting such attacks in DC state estimation, attack detection in AC systems is also critical, since AC state estimation is more widely employed in power utilities. In this paper, we propose a new false data injection attack detection mechanism for AC state estimation. When malicious data are injected in the state vectors, their spatial and temporal data correlations may deviate from those in normal operating conditions. The proposed mechanism can effectively capture such inconsistency by analyzing temporally consecutive estimated system states using wavelet transform and deep neural network techniques. We assess the performance of the proposed mechanism with comprehensive case studies on IEEE 118- and 300-bus power systems. The results indicate that the mechanism can achieve a satisfactory attack detection accuracy. Furthermore, we conduct a preliminary sensitivity test on the control parameters of the proposed mechanism.