Real-time power quality detection and classification system

• Main Author: Abidullah, Noor Athira
• Format: Thesis
• Language: English; English
• Published: 2015
• Subjects: T Technology (General); TA Engineering (General). Civil engineering (General)
• Online Access: http://eprints.utem.edu.my/id/eprint/20608/; https://plh.utem.edu.my/cgi-bin/koha/opac-detail.pl?biblionumber=106016&query_desc=kw%2Cwrdl%3A%20Real%20time%20power%20quality%20detection%20and%20classification%20system

The increasing number of power electronics equipment contributes to the poor quality of electrical power supply and has become a vital concern to electricity users at all levels of usage. The power quality signals can affect manufacturing process, malfunction of equipment and economic losses. Thus, it is necessary to detect and classify different kind of power quality signals for rectify failures and ensure quality of power line signal. This research presents the analysis power quality signals using time-frequency distributions (TFDs) which are spectrogram, Gabor transform and S-transform for signals detection and classification. Since the signals consist of multi-frequency components and magnitude variation, the TFDs are very appropriate to be used that represent the signals, jointly, in time-frequency representation (TFR). From the TFR, parameters of the signals are estimated and then are used to identify the characteristics of the signals. Referring to IEEE Std. 1159-2009, the signal characteristics are obtained and then served as the input for signal classifier to classify power quality signals. Based on the analysis, the best TFD is identified in terms of accuracy of the signal characteristics, memory size and computation complexity of data processing and chosen for power quality signals detection and classification system. By simulating in MATLAB, the performance of the classification system is verified by generating and classifying 100 signals with various characteristics for each type of power quality signals. In addition, the system is also tested using 100 real signals which were recorded from a power line. The results show that, S-transform is the best TFD and the classification system gives 100 percent correct classification for all power quality signals. For the real signals, the system also presents 100 percent correct classification. Thus, the outcome of this research shows that the system is very appropriate to be implemented for power quality monitoring system.