The author of this document has limited its availability to on-campus or logged-in CSUSB users only.

Off-campus CSUSB users: To download restricted items, please log in to our proxy server with your MyCoyote username and password.

Date of Award


Document Type

Restricted Thesis: Campus only access

Degree Name

Master of Science in Computer Science


School of Computer Science and Engineering

First Reader/Committee Chair

Ghasemkhani, Amir


Phasor Measurement Units (PMUs) have been widely used in several regions to improve bulk power system situational awareness. PMUs provide utilities with enhanced monitoring and control capabilities and are regarded as one of the future's most important measuring devices for power systems. Classifying PMU events is critical for making the power transmission system more reliable. Researchers have made tremendous progress in developing robust event classification models based on machine learning techniques; however, the proposed methods are vulnerable against adversarial attacks which potentially can degrade the event classification performance significantly. In this project, we evaluate the vulnerability of event classification models against data poisoning attacks for PMU data using feature collision attack. Moreover, we will leverage real-world PMU data to generate the poisonous data and observe how the added perturbations can change the measurement data. We will explain how to knowingly generate poisoned PMU data using a feature collision attack and its negative impact on a reliable event classifier.