Date of Award

Summer 2018

Document Type


Degree Name

Master of Science (MS)


Computational Modeling & Simulation Engineering


Modeling and Simulation

Committee Director

Sachin Shetty

Committee Member

Yuzhong Shen

Committee Member

Hong Yang


Critical infrastructure such as a Bulk Power System (BPS) should have some quantifiable measure of resiliency and definite rule-sets to achieve a certain resilience value. Industrial Control System (ICS) and Supervisory Control and Data Acquisition (SCADA) networks are integral parts of BPS. BPS or ICS are themselves not vulnerable because of their proprietary technology, but when the control network and the corporate network need to have communications for performance measurements and reporting, the ICS or BPS become vulnerable to cyber-attacks. Thus, a systematic way of quantifying resiliency and identifying crucial nodes in the network is critical for addressing the cyber resiliency measurement process. This can help security analysts and power system operators in the decision-making process. This thesis focuses on the resilience analysis of BPS and proposes a ranking algorithm to identify critical nodes in the network. Although there are some ranking algorithms already in place, but they lack comprehensive inclusion of the factors that are critical in the cyber domain. This thesis has analyzed a range of factors which are critical from the point of view of cyber-attacks and come up with a MADM (Multi-Attribute Decision Making) based ranking method. The node ranking process will not only help improve the resilience but also facilitate hardening the network from vulnerabilities and threats.

The proposed method is called MVNRank which stands for Multiple Vulnerability Node Rank. MVNRank algorithm takes into account the asset value of the hosts, the exploitability and impact scores of vulnerabilities as quantified by CVSS (Common Vulnerability Scoring System). It also considers the total number of vulnerabilities and severity level of each vulnerability, degree centrality of the nodes in vulnerability graph and the attacker’s distance from the target node. We are using a multi-layered directed acyclic graph (DAG) model and ranking the critical nodes in the corporate and control network which falls in the paths to the target ICS. We don't rank the ICS nodes but use them to calculate the potential power loss capability of the control center nodes using the assumed ICS connectivity to BPS. Unlike most of the works, we have considered multiple vulnerabilities for each node in the network while generating the rank by using a weighted average method. The resilience computation is highly time consuming as it considers all the possible attack paths from the source to the target node which increases in a multiplicative manner based on the number of nodes and vulnerabilities. Thus, one of the goals of this thesis is to reduce the simulation time to compute resilience which is achieved as illustrated in the simulation results.