An algorithm for power system disturbance monitoring

This paper presents a spectral analysis algorithm for monitoring the characteristics of the damped oscillating "modes" which are set up after a disturbance in an electric power distribution system. This monitoring is necessary so that one can detect early whether any modes are "inverse damped" (i.e. exponentially growing), and hence likely to create instability in the power system. In this application it is necessary to be able to resolve modes which are very closely spaced in frequency, and which may be imbedded in substantial noise. The proposed algorithm is Fourier based, and is an extension of a method proposed by the author. The extension enables resolution of very closely spaced multiple modes and has good noise performance. The scheme is tested on simulated signals and on a real power system example.