Impact-Based Analytics of Cascaded False Data Injection Attacks on Smart Grids

Authors

Mohamadsaleh JafariFollow

Document Type

Dissertation

Degree

Doctor of Philosophy (PhD)

Major/Program

Electrical and Computer Engineering

First Advisor's Name

Mohammad Ashiqur Rahman

First Advisor's Committee Title

Committee Chair

Second Advisor's Name

Sumit Paudyal

Second Advisor's Committee Title

Co-Committee Chair

Third Advisor's Name

Kemal Akkaya

Third Advisor's Committee Title

Committee Member

Fourth Advisor's Name

Norman Munroe

Fourth Advisor's Committee Title

Committee Member

Fifth Advisor's Name

Dong Chen

Fifth Advisor's Committee Title

Committee Member

Keywords

False data injection attack, smart grid, small-signal stability, frequency stability, load-frequency control

Date of Defense

12-2-2021

Abstract

The composition of communication lines and bidirectional smart devices in smart grids makes them greatly vulnerable to cyber-attacks. False data injection attacks (FDIAs) are of the most widespread cyber-attacks severely affecting the performance of smart grids that need to be analyzed. In this context, we analyze the feasibility of launching FDIAs in power systems with high renewables leading to false relay operations (FRO). By modeling the frequency dynamics of power systems, formalizing the FRO attack as a constraint satisfaction problem, our case studies show that power grids with renewables are more susceptible to FRO attacks. Moreover, the results suggest that the inertia of synchronous generators plays a critical role in reducing the success of FRO attacks in power systems.

The minimum required time of optimal FDIAs leading to a FRO is also formalized as an optimization-based formal model. This model finds the optimal size of attacks over multiple automatic generation control (AGC) cycles to minimize the attack launch time. Thereafter, we study the impact of power system parameters including inertia, governor’s speed droop and time constant, and the attacker’s accessibility to loads on the attack success and launch time. The results demonstrate that systems with low inertia are more vulnerable to FDIAs while fewer generator protection relays are impacted by FRO in systems with higher inertia. Moreover, our studies show that combining large values of the governor’s time constants and small values of the governor’s droop can raise the time of successful attacks. On top of that, the results manifest that continuous attacks may have more severe consequences such as frequency instability in power systems while interrupted attacks can lead to less-serious impacts such as protection relays operation.

In addition, we formulate the FDIA as an optimization problem using AC power flow and SSS constraints while modeling the attacker’s limited access to measurements. We assess the proposed mechanism on the WSCC 3-machine 9-bus system, demonstrating the low possibility of successful random FDIAs in launching a destabilizing attack. However, an intelligent attacker with the grid’s knowledge can make the system unstable, even with limited access to the measurements.

Identifier

FIDC010487

ORCID

https://orcid.org/0000-0001-9339-2098

Recommended Citation

Jafari, Mohamadsaleh, "Impact-Based Analytics of Cascaded False Data Injection Attacks on Smart Grids" (2021). FIU Electronic Theses and Dissertations. 4936.
https://digitalcommons.fiu.edu/etd/4936