Power system frequency estimation using Supervised Gauss-Newton algorithm

A supervised Gauss-Newton (SGN) algorithm for power system frequency estimation is presented in this paper. Taking the signal amplitude, the frequency and the phase angle as unknown parameters, the Gauss-Newton algorithm is applied to estimate the frequency for high accuracy. Meanwhile, a recursive DFT method and a zero-crossing method are used to compute the amplitude and the frequency roughly, so that the parameters can be initialized properly and the updating steps can be supervised for fast convergence of Gauss-Newton iterations. With this combined approach, both high accuracy and good tracking speed can be achieved for power system fundamental frequency estimation.