Zhijie Nie – Preliminary Exam

About the event

Degree:  Ph.D. Electrical & Computer Engineering

Advisor: Dr. Anurag Srivastava

Abstract:

Traditionally, the Energy Management System (EMS) has been deployed at the control center in suggesting preventive and corrective control actions at a minute scale and the local automated control at substations at a much faster timescale. Deployment of Intelligent Electronic Devices (IEDs) such as Phasor Measurement Units (PMUs) plays an essential role in addressing the system situational awareness. Remedial action schemes (RAS) have been implemented to automate coordinated control actions for the adverse operational violations. The supporting cyber system is highly dependent on the integrity of communication networks and devices, but local cyber situational awareness is limited. Cyber vulnerabilities in such devices and networks pose risks for critical applications such as RAS. A network intrusion detection system (IDS) is an essential security mechanism to prevent devices from malicious attacks.

In this dissertation, the real-time PMU measurement data has been utilized to develop RAS for minimum wind curtailment. The developed RAS applications provide automated control actions to eliminate system violations using DC and AC Optimal Power Flow (OPF) method implemented with fault-tolerant distributed computing algorithms. The RAS applications are simulated and tested in the real-time hardware-in-the-loop testbed. Machine-learning-based IDS has been introduced to identify the cyber anomalies at the substation level. Additionally, an adaptive framework is proposed in this dissertation to not only detect the cyber-attack behaviors by using network packets but also to conduct incremental learning to enhance the detection ability with changing network characteristics and new device vulnerabilities. Developed algorithms have been validated in a cyber-physical test system for enhanced cyber-physical resiliency using a set of assessment metrics.