Transformer Differential Protection by Online Core Modeling and Orthogonal Polynomials

By expressing the newly defined "magnetic pseudocharacteristic" in terms of orthogonal polynomials, a new method is proposed to discriminate between inrush and fault currents. In this method, the operating segment of the magnetic pseudocharacteristic is extracted over a half-cycle data window of differential current, and is then evaluated by some criteria to detect fault/inrush currents. Independence from transformer models and high accuracy are some merits of this method. The method is tested on the simulated data of wye-delta and wye-wye-connected transformers (three limb) and the results show the average fault detection delay of 16 ms (in 50-Hz base) with an accuracy above 97%. Also, for a more practical assessment, the results are compared with those of the well-known second harmonic restraint method, and this verifies greater accuracy of the proposed method. Moreover, performance of the method in noisy environments is studied. Details about efficient implementation by using fast Fourier transform algorithms are also included.