A New Technique to Identify Internal Faults and Inrush Currents Using Morphological Gradient with Adaptive Filtering

This paper presents a novel technique for distinguishing between transformer inrush currents and internal fault currents. On the basis of basic morphological transforms and their combined modes, transient current signals can be extracted meanwhile a variety of noises can be greatly suppressed by use of the novel morphological gradient (NMG) with adaptive weighting. Based on different characteristics of energy spectrum analysis of extracted transients between the short circuit current and the inrush current, the transformer protection criterion is proposed. This method can avoid the symmetrical inrush current and the operating time of the scheme is less than one power frequency cycle with very high accuracy. The results of dynamic simulation verify the feasibility of the proposed method.