Zhu Jiran

Waveform Complexity Analysis of Differential Current Signal to Detect Magnetizing Inrush in Power Transformer

A method to discriminate between magnetizing inrush and internal fault in power transformer is proposed and evaluated in this paper. First, the differential current signal in power transformer is scrutinized and its intrinsic features during magnetizing inrush conditions are introduced. Secondly, an adaptive generalized morphological filter is designed to suppress the noise in differential current signal. At last, the complexity of signal waveform is estimated applying the grille fractal. As the waveform complexity is much bigger in magnetizing inrush than in fault condition, the proposed method can reliably and accurately identify transformer inrush from faults. Extensive simulations validate the merit of this method for various conditions.

Fault Location Based on Correlative Characteristic Values in Resonant Grounded System

A method for fault location based on correlative characteristic values (CCVs) in resonant grounded system is presented. In this method, zero-sequence current signals of two different data windows are extracted for each feeder terminal unit (FTU) after a single-phase-to-ground fault occurs. Then, their correlation coefficient which is called CCV of each FTU is calculated. Furthermore, CCVs of all FTUs in the faulty feeder are uploaded to control centre. For healthy section, two CCVs of two FTUs at both sides of the section are almost the same while two CCVs of two FTUs, for faulty section, have a great difference. Thus, faulty section can be located. Simulation results under different fault conditions show that the method is effective and feasible.

The research and application of grounding fault locating method based on EMD and waveform similarity in non-solidly earthed network

In order to solve the fault location problem of single phase grounding fault in non-solidly earthed distribution network, the constitution, principle and application condition of a new fault locating method are presented in this paper. This method is realized by the distribution automation system (DAS), feeder monitoring unit has been deployed to acquire fault zero sequence current waveform and send the waveform to DAS master station. The locating workstation deployed on master station to get intrinsic mode function (IMF) component from zero sequence current waveform by using empirical mode decomposition (EMD) algorithm, then calculated similarity coefficient of IMF component by analyzing waveform similarity algorithm, finally located fault section according to similarity coefficient. The actual operation results verified that the locating method is reasonable and effective.