A spectral algorithm for extracting power system modes from time recordings

The authors present a new technique for extracting power system modes contained in a recorded time signal. The technique is based on spectral analysis and uses the Z-transform identification algorithm of Corinthios. It permits identification of the damping factor and the frequency of the modal signal components and it is computationally attractive due to efficient fast Fourier transform algorithms. Signals containing multiple modes, which may be relatively close or unstable, can be handled by the technique. The theory behind the spectral algorithm is outlined, and several examples are presented.<
>