Modal identification of power system oscillation using parametric DFT technique

Fourier based technique of system identification by mode decomposition has always been non-parametric which might not be accurate for power system applications. In this paper, a parametric Discrete Fourier transform (DFT) technique is used to identify both inter-area modes and local area modes of a power system oscillation. The natural frequency and damping ratio (the poles) are obtained from fitting a curve around the dominant peaks in the DFT of the measured signal. Using the estimated poles, the amplitude and phase are obtained by fitting with the measured signal like any other parametric method. In this paper, the parametric DFT is applied for a synthetic signal with closely spaced inter-area modes with poor-damping which is validated in the presence of two levels of white Gaussian noise. The WECC 9 bus system is considered as a case study and the parametric DFT technique is compared and contrasted with the traditional Prony analysis technique.