A New Adaptive Method to Identify TA Saturation Using Grille Fractal

In this paper, the different singularity characteristics of differential current of transformer differential protection under inner-zone and out-zone faults are analyzed while the current transformer is saturated. Then, using grille fractal, a new approach to prevent maloperation of transformer differential protection is proposed. Combining the proposed approach with an adaptive generalized morphological filter, the singularity characteristic of the signal can be fully reserved. Meanwhile, noises and interferences can be restrained. It is not necessary for the proposed approach to accurately determine the moment when fault occurs and differential current appears. Using relative magnitude of smooth domain of maximal value and that of minimal value and by means of setting an appropriate threshold, the inner fault of the transformer can be reliably identified and the identification is not affected by inrush current. Besides, the evolved fault can be reliably identified. Results of dynamic simulation prove that the proposed approach is effective and feasible.