Files in this item



application/pdfCHATTOPADHYAY-THESIS-2020.pdf (2MB)
(no description provided)PDF


Title:A fast state-estimation-based data integrity threat detection approach for combined AC-DC bulk power systems
Author(s):Chattopadhyay, Abhiroop
Advisor(s):Gross, George
Department / Program:Electrical & Computer Eng
Discipline:Electrical & Computer Engr
Degree Granting Institution:University of Illinois at Urbana-Champaign
Subject(s):state estimation
data integrity attacks
measurement prioritization
Abstract:The challenges encountered in modern power system operations have increased as electricity grids have become more geographically widespread and complex. Some modern bulk grids now combine AC and DC subsystems to effectively serve their loads. Communications and controls for such combined grids, in particular, have increasingly become more challenging. System disturbances in such combined grids have the potential to cascade and affect much larger portions of the grid. These challenges have only been exacerbated by the deepening penetrations of renewable energy resources, such as wind and solar. A key operational concern of such combined grids is the ability of the system operators to continually maintain situational awareness in their operations. The response times within which corrective control actions must be dispatched under contingencies have shortened considerably. In an attempt to ensure the intact transmission of an ever larger amount of information for combined AC-DC grids so as to maintain situational awareness and promptly dispatch control actions, the electric power industry has increased the deployment of cyberphysical, microprocessor-based devices in system monitoring and control. These devices provide system operational information to the system operator over digital channels with latencies much lower than those using conventional copper wire analog signals. But, with this more intense reliance in power system monitoring and control on communication channels, cybersecurity has become an additional concern. The possibility that entities/individuals with malicious intent can gain access to these communication channels and are able to alter operational commands is a fact of life. The types of cyber attacks that threat agents can perform are varied and include false data injection and data integrity attacks, spoofing and denial of service. While it is advisable to include information technology-based intrusion detection/prevention techniques to parse and verify the syntax of protocol messages, effective use of the physical characteristics of the power grid provides alternative, physics-based detection methods. So far, physics-based detection methods have mostly focused on AC system applications. Some investigations have been conducted on combined AC-DC systems, which have focused primarily on microgrids so as to restrict the applications to low- and medium-voltage systems. In this report, we propose and investigate a physics-based approach to threat detection in bulk combined AC-DC grids via the use of a rapid, approximate state estimation scheme. We specifically investigate data integrity attacks, which aim to corrupt the active power dispatch commands on the HVDC lines in these combined bulk AC-DC grids. The state-estimation based scheme we propose requires the determination of the system state estimates at sufficiently frequent time intervals to allow the performance of consistency checks between the approximate injections computed from the estimate of the state with respect to that of the power flow that corresponds to the true power order transmitted for implementation. We obtain gains in computational speed in the proposed approximate state-estimation-based approach to make it capable to track the changes in state with adequate accuracy for detection purposes. For this purpose, we use the power transfer distribution factors (PTDFs) as the criterion for measurement prioritization to produce a reduced subset of prioritized measurement. In addition, we impose a judiciously specified limit on the number of iterations in the state estimation to meet the time response requirements. These two modifications, combined with effectively implemented sparsity-techniques, result in a robust approach for the detection of the class of cyber threats considered in this report. The contribution of this report lies in the use of the widely used power system state estimation tool to develop a simple, practical physics-based approach to data integrity attack detection specifically for use in combined bulk AC-DC grids. We advantageously use the incorporation of this PTDF-based measurement prioritization feature into the conventional AC state estimation extensively deployed in modern EMSs to create a detection scheme for the cyber threats considered in this report. We demonstrate the effective deployment of this state-estimation-based approach with results from case studies on a representative 2470-bus synthetic combined AC-DC test system that is based on the U.S. part of the WECC interconnection with the California-Oregon Pacific DC intertie. In our simulation studies, we are able to detect the corruption of a 920 MW power order command to within 5 % of its true value. The implementation of this corrupted power order is detected within a 30-second time period with the prioritized measurement subset to contain the measurements associated with less than 2 % of the total number of lines in the system. The results discussed have provided insights into the performance of the heuristic procedures and a basis for the appropriate choices of the tunable parameters of the state estimation scheme. We discuss the computational and accuracy aspects and provide bounds on the extent to which an attacker can corrupt power orders that the scheme successfully detects. We observe, for instance, that the accuracy of the approach is more sensitive to our choice of the prioritized measurements than the limit on the number of iterations. We also share our insights on the deployment aspects of this approach by a system operator of a physical combined AC-DC bulk power grid.
Issue Date:2020-06-29
Rights Information:Copyright 2020 Abhiroop Chattopadhyay
Date Available in IDEALS:2020-10-07
Date Deposited:2020-08

This item appears in the following Collection(s)

Item Statistics