Date of Graduation


Document Type


Degree Name

Doctor of Philosophy in Computer Science (PhD)

Degree Level



Computer Science & Computer Engineering


Qinghua Li

Committee Member

Xintao Wu

Second Committee Member

Roy A. McCann

Third Committee Member

Brajendra Nath Panda


cybersecurity, machine learning, power system, security vulnerability, vulnerability patch management


Security vulnerabilities in software pose an important threat to power grid security, which can be exploited by attackers if not properly addressed. Every month, many vulnerabilities are discovered and all the vulnerabilities must be remediated in a timely manner to reduce the chance of being exploited by attackers. In current practice, security operators have to manually analyze each vulnerability present in their assets and determine the remediation actions in a short time period, which involves a tremendous amount of human resources for electric utilities. To solve this problem, we propose a machine learning-based automation framework to automate vulnerability analysis and determine the remediation actions for electric utilities. Then the determined remediation actions will be applied to the system to remediate vulnerabilities. However, not all vulnerabilities can be remediated quickly due to limited resources and the remediation action applying order will significantly affect the system's risk level. Thus it is important to schedule which vulnerabilities should be remediated first. We will model this as a scheduling optimization problem to schedule the remediation action applying order to minimize the total risk by utilizing vulnerabilities' impact and their probabilities of being exploited.

Besides, an electric utility also needs to know whether vulnerabilities have already been exploited specifically in their own power system. If a vulnerability is exploited, it has to be addressed immediately. Thus, it is important to identify whether some vulnerabilities have been taken advantage of by attackers to launch attacks. Different vulnerabilities may require different identification methods. In this dissertation, we explore identifying exploited vulnerabilities by detecting and localizing false data injection attacks and give a case study in the Automatic Generation Control (AGC) system, which is a key control system to keep the power system's balance. However, malicious measurements can be injected to exploited devices to mislead AGC to make false power generation adjustment which will harm power system operations. We propose Long Short Term Memory (LSTM) Neural Network-based methods and a Fourier Transform-based method to detect and localize such false data injection attacks. Detection and localization of such attacks could provide further information to better prioritize vulnerability remediation actions.