SVM‐based PQ disturbance recognition system. Issue 2 (20th November 2017)

• Record Type: Journal Article
• Authors: Huang, Jiansheng; Jiang, Zhuhan; Rylands, Leanne; Negnevitsky, Michael
• Is Part Of: IET generation, transmission & distribution. Volume 12:Issue 2(2018)
• Page Start: 328
• Page End: 334
• Publication Date: 2017-11-20
• Subjects: support vector machines -- power supply quality -- power grids -- discrete wavelet transforms -- power engineering computing; SVM‐based PQ disturbance recognition system -- control systems -- communication networks -- electrical components -- power distribution systems -- waveform generator -- SVM classifiers -- SVM‐based techniques -- DWT -- discrete wavelet transform; Electric power production -- Periodicals; Electric power transmission -- Periodicals; Electric power distribution -- Periodicals; 621.3105
• DOI: 10.1049/iet-gtd.2017.0637
• Languages: English
• ISSNs: 1751-8687

Abstract

Abstract : The quality of power delivered by modern electricity grids is of interest as disturbances to power quality (PQ) have the potential to cause malfunction of control systems, interfere with communication networks, increase power losses and reduce the life of electrical components. It is, therefore, necessary to determine if there are PQ disturbances in a grid, and if so what forms these disturbances take. On the basis of site measurements at power distribution systems, a waveform generator is designed to emulate 11 types of PQ disturbances as well as harmonics, and a prototype for recognising these undesirable disturbances is presented. The first step is to use the discrete wavelet transform (DWT) to extract the most representative transients at different time spans from the original waveform. The second step is to use the output of the DWT to construct two sets of classifiers, which can recognise the types of disturbances present. Non‐linear support vector machine (SVM)‐based techniques are exploited for this step. Case studies are carried out to verify the prototype system. Simulations show that the SVM classifiers developed can achieve superior performance in recognising PQ disturbances compared with conventional counterparts.