An adaptive fault detection method based on atom sparse and evidence fusion for the small current to ground system. (March 2018)

• Record Type: Journal Article
• Authors: Wang, Xiaowei; Wei, Xiangxiang; Yang, Dechang; Gao, Jie; Sun, Xue; Song, Guobing
• Is Part Of: Transactions of the Institute of Measurement and Control. Volume 40:Number 5(2018)
• Page Start: 1579
• Page End: 1592
• Publication Date: 2018-03
• Subjects: Fault detection -- matching pursuit algorithm -- correlation analysis -- fault trust function -- energy entropy -- D-S evidence theory; Automatic control -- Periodicals; Measuring instruments -- Periodicals; Commande automatique -- Périodiques; Mesure -- Instruments -- Périodiques; 681.2
• DOI: 10.1177/0142331216686402
• Languages: English
• ISSNs: 0142-3312

Abstract

An adaptive fault detection method for the small current to ground system is proposed based on the atom sparse and evidence fusion theory. Firstly, the two cycles of transient zero-sequence current are obtained when occur fault, which is decomposed by the matching pursuit algorithm, and the iterations are set to four, then, according to the correlation analysis, the best three atoms of maximum correlation are chosen and which are sorted by the correlation coefficients. Secondly, we obtained the determine fault measure (DFM) values by calculating the atomic energy entropy, and selected the atomic fault trust (FT) function by revising the DFM. Finally, the FT values are integrated by the D-S( Dempster–Shafer) evidence theory, then the fault comprehensive trust values are gained, the selection results are output. Simulation results show that the method of fault line detection is accurate and reliable.