UTHM Institutional Repository

Application of advanced signal processing technique for power quality disturbances detection and classification system

Mohd Noh , Faridah Hanim (2016) Application of advanced signal processing technique for power quality disturbances detection and classification system. PhD thesis, Kumamoto University.

[img]
Preview
PDF
Faridah_Hanim_Mohd_Noh.pdf

Download (8MB)

Abstract

The research focuses on power quality (PQ) monitoring systems are very much involved in the studies processing techniques applied to the monitored signals to extract valuable information for analyzing and diagnosing the cause of power quality problems. The monitoring practice of power quality includes three steps: i) detecting and extracting the features of power quality events, ii) classifying these events into known waveform categories, and iii) if the size of the monitored data is very large, compression methods will be necessary for saving these data for further analysis. This research is focused mainly on the first power quality monitoring practice step, which involved detecting and extracting PQ disturbance signal features using wavelet transform and S-transform. Wavelet transforms are multiresolution decompositions that can be used to analyze both signals and images. They describe a signal by the power at each scale and position. S-transform is a time-frequency spectral localization method which has a Gaussian window width scales inversely with frequency, while the window's height scales linearly with the frequency. This thesis proposed to apply Slantlet transform (SLT), which is another type of orthogonal wavelet transform that has been introduced by Ivan W. Selesnick in 1999. Slantlet transform with 2-scale filter banks has been used in this study to analyze single and multiple PQ disturbance signals. Several common features that being used in wavelet transform were extracted and analyzed to determined its usability in detecting and classifying PQ signals into its appropriate types. As a result, an impressive performance was displayed by the Slantlet transform, while the extracted features were effectively localized to the type of PQ disturbance. The classification method is successfully done with the application of Gaussian RBF kernel support vector machine (SVM). The performance of SLT was then tested for noise corrupted PQ disturbance signals. The result shows some reduction in the classification rate as the noise level increased. After that, Modified S-transform has been proposed to analyze PQ disturbance signal. Modified S-transform (MST) is an improved version of S-transform that has been proposed by Assous and Boashash in 2012. Ther performance of this method to analyze and extract features for PQ disturbance signals was investigated. The results exhibited a remarkable classification rate for both single and multiple PQ disturbance signals detection, as well as signal classification under noisy and unnoisy conditions.

Item Type: Thesis (PhD)
Subjects: T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK1001-1841 Production of electric energy or power. Powerplants. Central stations
Depositing User: Mr. Mohammad Shaifulrip Ithnin
Date Deposited: 13 Aug 2018 03:27
Last Modified: 13 Aug 2018 03:27
URI: http://eprints.uthm.edu.my/id/eprint/10259
Statistic Details: View Download Statistic

Actions (login required)

View Item View Item

Downloads

Downloads per month over past year