On-line parameter identification of power plant characteristics based on phasor measurement unit recorded data using differential evolution and bat inspired algorithm

• Published in: IET science, measurement & technology Vol. 9; no. 3; pp. 376 - 392
• Main Authors: Rashidi, Farzan, Abiri, Ebrahim, Niknam, Taher, Salehi, Mohammad Reza
• Format: Journal Article
• Language: English
• Published: The Institution of Engineering and Technology 01.05.2015
• Subjects: bat inspired algorithm; differential evolution algorithm; dynamic modelling; evolutionary computation; frequency domain analysis; hybrid method; maximum likelihood estimation; measurement errors; measurement noise; measurement uncertainty; model uncertainty; nonlinear optimisation problem; nonlinear programming; online power system model parameter estimation; particle swarm optimisation; phasor measurement; phasor measurement unit; power plant characteristics; power system identification; power system parameter estimation; power system planning; power system security; power system security assessment; power system stability; hybrid method; particle swarm optimisation; nonlinear programming; power system security assessment; maximum likelihood estimation; power system identification; model uncertainty; online power system model parameter estimation; measurement uncertainty; power system planning; power system stability; power system parameter estimation; measurement errors; measurement noise; dynamic modelling; frequency-domain analysis; nonlinear optimisation problem; power system security; frequency domain analysis; evolutionary computation; phasor measurement unit; differential evolution algorithm; power plant characteristics; phasor measurement; bat inspired algorithm
• ISSN: 1751-8822; 1751-8830
• DOI: 10.1049/iet-smt.2014.0022

Parameter estimation and dynamic modelling of power systems and their components are basis of design, planning and stability or security assessment in power systems. This study considers the estimation of power system model parameters by a global identification framework based on the maximum-likelihood principle. The proposed framework is formulated as a non-linear optimisation problem, which is solved by a hybrid method based on the bat-inspired algorithm and differential evolution method. The combination of these algorithms makes the hybrid method faster and it obtains closer to the global minimum than a pure global method. Since noise and model uncertainties are inherent parts of system identification, the effect of these factors on the performance of the proposed identification framework are studied. Results based on synthetic data in frequency domain show that the estimated parameters are close to the correct values even in the presence of significant measurement noise and considerable uncertainties.