De-Noising of Partial Discharge Signals in HV XLPE Cables by Reference Noise based on the Wavelet Transform

Abstract

Due to the problems associated with overhead lines, underground XLPE cables are increasingly being used in power systems. The main cause of deterioration in these cables is insulation failure, primarily arising from the partial discharge phenomenon. One of the main challenges in online PD detection is the presence of various noises in the environment that must be eliminated. In recent years, various types of noise with different distributions, such as impulse noises generated by power electronic devices, have been introduced into the power system. Therefore, denoising techniques should be employed to filter out the noises and interferences present in the detected PD signal. Due to the non-stationary nature of PDs, this paper suggests using the wavelet transform method, which covers both the time and frequency domains, to remove various noises from PDs. Consequently, to determine the suitable mother wavelet transform, threshold, and number of decompositions, the characteristics of PD signals occurring in the cables are investigated through experimental tests. Additionally, because different noises exist in substations, the background noise at the measurement site is recorded as a reference noise to be used in the application of the wavelet-based noise removal process. This method is examined on a sample cable, and the results are discussed. Moreover, using the suggested method, the detection of PD signals in several 20 kV substations in Iran is carried out through the use of high-frequency current transformers connected to shield wires, oscilloscopes with high-frequency bandwidth, and MATLAB software.

Keywords

• de-noising
• on-line detection
• partial discharge
• wavelet transform
• xlpe cable