Undetectable delay attacks on PMU-measurement based state estimation of smart-grids
Event details
| Date | 06.02.2018 |
| Hour | 14:00 › 16:00 |
| Speaker | Marguerite Delcourt |
| Location | |
| Category | Conferences - Seminars |
EDIC candidacy exam
Exam president: Prof. Mario Paolone
Thesis advisor: Prof. Jean-Yves Le Boudec
Co-examiner: Prof. Ola Svensson
Abstract
In the setting of smart-grid linear state estimation based on Phasor Measurement Unit (PMU) measurements, it has been shown that one can mount attacks on the time synchronisation of PMUs. The result of which can have a potentially large impact on the state estimation which can cause severe damage to the grid. Typically bad data in measurements is detected and identified using techniques based on residuals. Hence, a delay attack that triggers no change in measurement residuals will pass the test and remain undetected. Another approach avoiding the use of Bad Data Detection algorithms is to use so-called robust state estimators such as the Least Absolute Value (LAV) estimator. In our setting, LAV-based state estimators's time complexity is in fact competitive with the cumulation of a non-robust state estimator and a bad data detection algorithm. Thus, they may be largely deployed on PMUs in the future in which case successful undetectable attacks would need to influence the resulting state estimation bypassing the estimator's robustness. Another interesting fact is the clear link between the possible undetectably attackable PMUs and the criticality of their measurements. Some criterion can be used to find such locations and we believe that finding critical k-tuples in Power Networks will help attackers. The three papers that will be presented will focus on the undetectable timing attack on linear state estimation, on the use of LAV-based robust state estimators in systems measured by PMUs and on the computation of critical k-tuples in Power Networks.
Background papers Undetectable Timing-Attack on Linear State-Estimation by Using Rank-1 Approximation, by Barreto S., et al.
LAV Based Robust State Estimation for Systems Measured by PMUs, IEEE TRANSACTIONS ON SMART GRID, VOL. 5, NO. 4, JULY 2014.
Computing Critical k-Tuples in Power Networks, IEEE Transactions on Power Systems ( Volume: 27, Issue: 3, Aug. 2012).
Exam president: Prof. Mario Paolone
Thesis advisor: Prof. Jean-Yves Le Boudec
Co-examiner: Prof. Ola Svensson
Abstract
In the setting of smart-grid linear state estimation based on Phasor Measurement Unit (PMU) measurements, it has been shown that one can mount attacks on the time synchronisation of PMUs. The result of which can have a potentially large impact on the state estimation which can cause severe damage to the grid. Typically bad data in measurements is detected and identified using techniques based on residuals. Hence, a delay attack that triggers no change in measurement residuals will pass the test and remain undetected. Another approach avoiding the use of Bad Data Detection algorithms is to use so-called robust state estimators such as the Least Absolute Value (LAV) estimator. In our setting, LAV-based state estimators's time complexity is in fact competitive with the cumulation of a non-robust state estimator and a bad data detection algorithm. Thus, they may be largely deployed on PMUs in the future in which case successful undetectable attacks would need to influence the resulting state estimation bypassing the estimator's robustness. Another interesting fact is the clear link between the possible undetectably attackable PMUs and the criticality of their measurements. Some criterion can be used to find such locations and we believe that finding critical k-tuples in Power Networks will help attackers. The three papers that will be presented will focus on the undetectable timing attack on linear state estimation, on the use of LAV-based robust state estimators in systems measured by PMUs and on the computation of critical k-tuples in Power Networks.
Background papers Undetectable Timing-Attack on Linear State-Estimation by Using Rank-1 Approximation, by Barreto S., et al.
LAV Based Robust State Estimation for Systems Measured by PMUs, IEEE TRANSACTIONS ON SMART GRID, VOL. 5, NO. 4, JULY 2014.
Computing Critical k-Tuples in Power Networks, IEEE Transactions on Power Systems ( Volume: 27, Issue: 3, Aug. 2012).
Practical information
- General public
- Free
Contact
- EDIC - [email protected]