An Impulsive Noise Suppression Method Based on Chi-square Distribution and Robust Local Regression Smoothing

The impulsive noise with large amplitude exists in distribution network, which brings great challenge to power system monitoring. Aiming at the issue of low measurement accuracy of phasor measurement unit (PMU) in presence of impulsive noise., a method for suppressing impulsive noise is proposed. Firstly, the spectrum threshold is determined according to the characteristic that the normalized power spectrum of background noise follows the chi-square distribution; Then, taking advantage of the fact that the amplitude of impulsive noise is much greater than background noise, the residual is used to locate impulsive noise, and the idea of robust local regression smoothing is used to suppress impulsive noise; Finally, the noise is removed according to the spectrum threshold determined in the first step. The simulation results show that this method can effectively suppress impulsive noise and improve the measurement accuracy of the phasor measurement unit.