Enhanced nonlinear least squares for power system frequency estimation with phase jump immunity

摘要

Nonlinear least squares (NLLS) fitting is a method that can be used for estimating power system signal parameters. NLLS algorithms posit a model for the measured signal and then fit the model to the data by finding the parameters that minimize the sum of squared error between the measurements and model predictions. Here, we focus on the particular problem of frequency estimation and show that the NLLS algorithm is flexible enough to be made immune to harmonic distortion, DC offset, and phase imbalances. Tools for tuning the algorithm for desired noise/bandwidth performance are presented. Finally, a new method for estimating frequency in the presence of phase jumps is presented and tested. The new algorithm is capable of estimation in the presence of phase and amplitude jumps with near-zero error. This performance is shown to be vastly superior to two benchmark algorithms.