Noise reduction techniques for partial discharge classification in XPLE cable joint / Noratika Mohd Zin

Noratika, Mohd Zin (2017) Noise reduction techniques for partial discharge classification in XPLE cable joint / Noratika Mohd Zin. Masters thesis, University of Malaya.

Preview

PDF (Thesis M.A) Download (3004Kb) | Preview

Abstract

Cable joints of cross-linked polyethylene (XLPE) are the weakest point in a power system and can cause insulation failures with the present of partial discharge (PD). Therefore, it is important to monitor PDs at cable joints and determine the type of the defect that exists at cable joints. Determination of the type of the defect at cable joints can reduce the repair time and maintenance cost. In this project, defect type in cable joint determination using partial discharge testing under noisy condition was carried out. Five joints of cross-linked polyethylene cable including artificial defects were created according to the conditions usually found at site. Various noise reduction techniques were applied to denoise the PD signals and the denoised PD signals were used to determine different types of defect in cable joints. The input features from different noise reduction techniques were applied to train the classifier to determine the type of the problem in the samples. Determination of the defect type was performed using Support Vector Machine (SVM) after DFT, WPT and DWT techniques. The results were compared between each noise reduction methods to evaluate the performance of the applied methods. It was found that the noise reduction technique on partial discharge signals from cable joint defects using discrete Fourier transform (DFT) yields a better accuracy than wavelet packet transform (WPT) and discrete wavelet transform (DWT). Keywords: Partial discharge measurement, support vector machine, high voltage engineering, cable insulation, signal processing

Actions (For repository staff only : Login required)