Graduate Thesis Or Dissertation


Partitioning Techniques to Improve Resilience and Cyber Security in Electric Grid Infrastructure Public Deposited

Downloadable Content

Download PDF


Attribute NameValues
  • With the increasing penetration of renewable energy integration and potential threats from natural disasters, modern electric power system have been significantly exposed to both small- and large-scale disturbances. Traditional performance metrics including indices of reliability, stability, and security are no longer adequate to describe a power network. Infrastructure properties like resilience (sometimes also called resiliency) frequently appear in current related literature. Generally, resilience means that a power system demonstrates certain abilities in three different stages depending on the scale of a disturbance. In the first stage, the electric network should have the ability to absorb the impact of an event for example with the protective relaying system providing a quick and precise reaction. The event could be random faults on transmission or distribution lines, loss of generation or load, bus failures, or even extreme weather conditions and severe natural disasters like earthquake and tsunami. During the second stage, instead of extending the failed area and becoming a cascading outage, the electric network should have the ability to survive and move the entire system into another steady-state operating point. In the end, the electric network should have the ability to recover the lost areas and return to the original operation mode. Timely evaluation of power system resilience using proper clustering techniques is critical when considering both potential system failures and nature disasters. Intentional islanding operation is a solution to mitigate those threats because a proper islanding scheme has the potential to stop the propagation of cascading outages. On the other hand, the affected areas for a natural disaster can differ substantially from those areas that can be affected by a relay-corrected fault. The islanding operation mode will isolate the affected areas from others in which the power balance and dynamic stability properties are adequate. Therefore, adequate partitioning of a large power grid is paramount with the goal of positioning different areas of the grid at good starting points for survival and reconnection. Moreover, with the rapid development of computational technologies in society, the modern electric power system has already become a cyber-physical infrastructure that serves as backbone for the public infrastructure. Cyber security has also become a significant concern due to the increasing trend of cyber attacks. Thus, improving system resilience from the cyber perspective is another critical thrust for this research. We propose security metrics for the assessment and monitoring of the security status and risk exposure of a cyber-physical power grid. Motivated by the above challenge, we propose a new clustering algorithm based on the sensitivity between reactive power injection with phase voltage changes. We also illustrate a new index that evaluates the goodness of these clustering results. Combining these with previously formulated indices, we apply this algorithm to a variety of use case to show clustering results with higher fidelity. Finally, we propose a set of cyber-attack graph based metrics that are helpful in order to detect protection vulnerabilities for the cyber-physical system.
Resource Type
Date Issued
Degree Level
Degree Name
Degree Field
Degree Grantor
Commencement Year
Academic Affiliation
Rights Statement
Peer Reviewed



This work has no parents.

In Collection: