Authors: B. Subudhi, P.K. Ray, S.R. Mohanty, A.M. Panda
Addresses: Department of Electrical Engineering, National Institute of Technology, Rourkela 769008, India. ' Department of Electrical Engineering, National Institute of Technology, Rourkela 769008, India. ' Department of Electrical Engineering, National Institute of Technology, Allahabad, India. ' Department of Electrical Engineering, Indira Gandhi Institute of Technology, Sarang 759146, India
Abstract: This paper presents the estimation of frequency which is an important power system parameter by Extended Least Square (ELS) technique. The above technique is validated by comparing its performance with the existing techniques such as Kalman Filter (KF) and Least Mean Square (LMS) technique, etc. Using different simulation studies with signals having different signal to noise ratio values and with step change in frequency, it is observed that ELS technique outperforms over LMS and KF methods on power system frequency estimation. Initialisation of covariance matrix in KF method and complicacy due to incorporation of correlation matrix in LMS algorithm affect their convergence. But ELS algorithm becomes very simple and attractive due to the absence of covariance and correlation matrix.
Keywords: ELS technique; extended least squares; Kalman filter; LMS technique; least mean squares; covariance matrix; correlation matrix; system structure matrix; observation vector; frequency estimation; power system parameters.
International Journal of Automation and Control, 2009 Vol.3 No.2/3, pp.202 - 215
Available online: 17 May 2009 *Full-text access for editors Access for subscribers Purchase this article Comment on this article