Farzad Dalavi

MSc Thesis:

Title: Identification of Power System Oscillation Modes Characteristics with Phasor Measurement Units

Abstract: In recent years, due to increasing electrical energy demand and loading of transmission lines, inter-area oscillations by frequency in the range of 0.2 to 0.8 Hz, are the most important reason that causes oscillatory rotor angle instability. when inter-area modes are excited, the power system is divided into areas that generators at each area are coherent but they are anti-phase respect to other generator groups. The inter-area oscillations have a lower frequency and damping than local modes and they cause power oscillation at tie-lines between areas that restrict power transmission capacity. Therefore, online monitoring and control of inter-area modes are vital for secure and reliable operation of the power system. The advancement of the technology and emergence of Wide Area Measurement Systems (WAMSs) and the possibility of measure phasor quantities with Phasor Measurement Units (PMUs), can monitor the online dynamic response of power system with just data analysis with signal processing techniques. The main feature of signal processing techniques is they don't need any knowledge about the power system order or power equipment models. In this thesis, a hybrid method based on WAMS for the detection of dominant inter-area oscillation modes and evaluation generators oscillation patterns at each considerable mode is proposed. The proposed method can also detect the most efficient damping controller loops for improving the damping of dominant oscillation modes. The proposed method uses the Empirical Mode Decomposition (EMD) algorithm and Correlation Analysis (CA). In the first step, frequency components of measured signals decompose from highest to lowest frequency by the EMD algorithm and Then, calculated the instantaneous frequency characteristics of each extracted component by Hilbert Huang Transform (HHT). Power system oscillation modes are identified by comparing the frequency averages by the typical frequency range of low-frequency electromechanical oscillations (LFEOs) and finally, determine dominant oscillation modes by comparing the instantaneous energy level of oscillation modes. One feature of the EMD algorithm is can extract waveforms of oscillation modes. Therefore, by considering dominant mode waveforms as the input of correlation analysis, can determine how generators' oscillate each other by calculating the phase difference between signals. Also, efficient feedback signals can be determined by estimation relative magnitude of mode shape or right eigenvectors. Also, by comparing left eigenvectors magnitudes, efficient damping controllers determine. In other words, efficient damping controller loops detect based on observability and controllability concepts these are respectively related to right and left eigenvectors. Applicability of the proposed method in this thesis by different analyses and simulations on 9 and 39 bus IEEE test systems evaluated.