Oscillation monitoring system based on wide area phasor measurements in power systems.

摘要

This dissertation presents an Oscillation Monitoring System (OMS) based on real-time wide-area measurements from Phasor Measurement Units (PMU). This OMS is designed to detect poorly-damped or negatively-damped electromechanical modes in the early stage of an oscillation event, as well as provide warning signals from normal system operating conditions when mode damping becomes insufficient for safe operation of power systems. Depending on different mathematical models of the measured data, different processing algorithms are used. The system disturbance part of the OMS is designed to monitor system events in real-time for the purpose of emergency control, while the ambient part monitors the system without any disturbances for the purpose of preventive control. These two parts are complementary to each other, constituting a complete monitoring system.;Power system responses following system disturbances contain both linear and nonlinear phenomena. Moreover, presence of noise and switching events in the measurements can upset the accuracy of results. For these reasons, we developed different crosscheck rules to avoid false alarms due to inconsistent estimations. Three signal processing algorithms are used, namely, Prony's Method, Matrix Pencil Method and Hankel Total Least Squares (HTLS) method. Results from these engines are processed using a custom developed set of rules for handling the complexities of modal analysis from real-time PMU measurements.;Ambient data are collected during normal system operations. Unlike previous methods used for ambient data processing, the modified Frequency Domain Decomposition (FDD) is able to simultaneously identify damping and mode shape of the dominant mode using several minutes of ambient data. FDD also works well for noisy measurements and correlated inputs, and it appears to be useful specifically for analyzing real-time PMU measurements. Together with the post-disturbance data processing following system events, it provides a powerful framework of an oscillation monitoring system from wide-area PMU measurements.