Tracking Electromechanical Oscillations: An Enhanced Maximum-Likelihood Based Approach

Lightly damped electromechanical oscillations are major operating concerns if failed to be detected at an early stage. This paper improved the existing extended complex Kalman filter (ECKF) technique of tracking electromechanical oscillations using synchrophasor measurements. The proposed algorithm...

Full description

Saved in:
Bibliographic Details
Published inIEEE transactions on power systems Vol. 31; no. 3; pp. 1799 - 1808
Main Authors Khalid, Haris M., Peng, Jimmy C.-H
Format Journal Article
LanguageEnglish
Published New York IEEE 01.05.2016
The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
Subjects
Algorithms
Augmented Lagrangian
Convergence
Correlation
distributed estimation
inter-area oscillation modes
Kalman filters
Maximum likelihood estimation
maximum-likelihood
Monte Carlo simulation
Noise
oscillations
Oscillators
phasor measurement unit (PMU)
power system monitoring
power system stability
real-time measurements
situational awareness
synchrophasor measurement
wide-area monitoring system (WAMS)
Online AccessGet full text

Cover

More Information
Summary:Lightly damped electromechanical oscillations are major operating concerns if failed to be detected at an early stage. This paper improved the existing extended complex Kalman filter (ECKF) technique of tracking electromechanical oscillations using synchrophasor measurements. The proposed algorithm adopted a distributed architecture for estimating oscillatory parameters from local substations. The novelty lies in handling maximum likelihood (ML) to enhance the convergence property in tracking multiple modes using an expectation maximization (EM) approach. This was achieved by encapsulating the augmented Lagrangian (AL) in the maximization step of the EM algorithm, which utilized a novel ECKF-based smoother (ECKS). Performance evaluations were conducted using IEEE 68-bus system and recorded synchrophasor measurements collected from the New Zealand grid. Random noise variance test cases were generated to examine the performance of the proposed algorithm. To ensure the robustness to random noisy conditions, the algorithm was tested based on exhaustive Monte Carlo simulations. Comparisons were made with the existing Prony analysis (PA), Kalman filter (KF), and distributed EM-based FB-KLPF.
ISSN:0885-8950
1558-0679
DOI:10.1109/TPWRS.2015.2441109