Improved Prony method for online identification of low-frequency oscillations in power systems

On-line oscillation characteristic analysis algorithms are used to monitor on-line low-frequency oscillations for control of wide-area damping. The development of wide area measurement systems allow detection of low-frequency oscillations on line. This paper discribes an improved Prony method for the detection of low-frequency oscillations. That significantly improves the identification of the practical order of the input signal and estimates of the linear prediction parameters. The time complexity of the algorithm is analyzed in the paper. Numerical simulations and physical test results demonstrate that the reduced-order model is more accurate with less computations; therefore, it is useful for online applications.