DENOISING OF ON-LINE PARTIAL DISCHARGE SIGNAL FROM HIGH-VOLTAGE ROTATING MACHINES USING STANDARD DEVIATION THRESHOLD

摘要

Recently, the increasing importance of reliable electric power supply has brought greater focus on on-line diagnosis of power equipment. A major cause of faults in high-voltage rotating machines is insulation breakdown in the stator winding. Thus, on-line diagnosis of the stator winding insulation is critical. For this, on-line detection of partial discharge is most commonly implemented by using a capacitive coupler at the stator winding. This paper proposes a denoising algorithm for the on-line partial discharge signal using a feature that considers the standard deviation of noise and the partial discharge signal. The proposed algorithm comprises a primary wavelet denoising scheme and a secondary scheme that is based on the standard deviation of white Gaussian noise. The algorithm was evaluated on a 13.2-kV-class hydro generator at Dae-chung Dam that has been operated for over 25 years. The proposed algorithm showed better performance in the evaluation than the conventional method.