Online monitoring system is usually used in different environments and thus is being exposed to different noise levels with characteristics not completely known. Therefore, digital processing algorithms are preferable than analogue filters to extract partial discharge (PD) pulses. The wavelet transform (WT) is a powerful tool for signal analysis and filtering due to its multiresolution characteristics. However, this technique does not work properly for filtering purposes in real time when undistorted transmission of signal is required, if wavelet function, number of scales and threshold value algorithm are applied improperly. In this paper, proper wavelet parameters selection are analyzed, aiming at improving the sensitivity of PD signal extraction and making PD pulse unaltered after filtering. This is because of that reduction of the amplitude and distortion of the PD pulse can lead to a reduction in the identification accuracy and location accuracy.
展开▼
