Parameter estimation of power system signals using SVD

Abstract

This paper presents an estimation technique for amplitude, frequency, phase and damping of a signal embedded in noise. This signal considered in this paper belongs to that of a power system, where frequency deviations in the neighbourhood of the nominal value occur due to dynamic oscillations or faults. The off-nominal frequency, amplitude and damping are estimated using a damped sinusoidal model, and a Taylor series expansion technique around the nominal value. Singular value decomposition(SVD) technique is used to solve the set of over-determined and under-determined equations for the instantaneous data samples, and the relevant computer simulation results are presented.