Evaluation of denoising performance indices for noisy partial discharge signal based on DWT technique

摘要

The purpose for this investigation is to choose the best mother wavelet and decomposition level. Online partial discharge (PD) always occur in insulation system and may disturb the perform of current transmission. There are two type of noise that involve, such as discrete spectral interference and white noise. Wavelet transform is used to discriminate the noisy signal and able to capture the actual PD signal. Selection of mother wavelet and decomposition level is the important aspect in the denoising technique. MATLAB is used to create the denoising algorithm. The wavelet that used in this paper is Daubechies, Coiflets, Symlets, BiorSplines and Reverse Bior. Mean square error has been calculated as the performance evaluation. The best mother wavelets are db7, coif3, sym5, bior3.5, rbio6.8 and level 4 for decomposition level.