Date of Award

2025-08-01

Degree Name

Master of Science

Department

Electrical and Computer Engineering

Advisor(s)

Michael P. McGarry

Abstract

Electric power systems have become one of our most critical infrastructures as we've grown dependent on electricity for everyday tasks. Ensuring power systems provide reliable service is a priority that can be affected by disturbance events. A common event is transmission line outages, where a line in the system becomes disconnected due to varying forms of physical damage. If an outage isn't detected in time, other lines in the system may overload, causing cascading failures that leave many customers without power. Therefore, having a power system that can automatically detect outages is crucial for reliability, as it promotes real-time response and recovery. Various methods for line outage detection have been developed over the years, focusing on issues such as minimum deployment of Phasor Measurement Units (PMU) and detecting outages with partial data. Furthermore, machine learning has gained traction for its improvements in line outage detection. In this thesis, we developed two machine learning-based methods for detecting transmission line outages using K-Nearest Neighbors. Our first method, we showed the potential of Line Outage Distribution Factors (LODF) as a feature or data observation point selection tool. Identifying critical observation points through LODF enables the detection of outages with limited data by monitoring power flow from one transmission line while accounting for load uncertainty to estimate the status of another line. Our second method introduces a new set of factors called Line Outage Impact Factors (LOIF), a modified version of the LODF, which solves some concerns we have when using LODF for feature selection. Instead of showing the distribution of power from an outage line, LOIF shows the change in power flow of a line due to the outage of another. We develop a feature-selection method that focuses on determining the outages that provide significant and distinct changes in power flow.

Language

en

Provenance

Received from ProQuest

File Size

67 p.

File Format

application/pdf

Rights Holder

Daniel Flores

Download

Share

COinS