Power transformer protection scheme based on time-frequency analysis Zahra Moravej 1 * ,† , Ali Akbar Abdoos 1 and Majid Sanaye-Pasand 2 1 Department of Electrical and Computer Engineering, Semnan University, Semnan, Iran 2 School of Electrical and Computer Engineering, University of Tehran, Tehran, Iran SUMMARY This article presents a new approach for power transformer protection based on a hybrid pattern recognition scheme. The hyperbolic S-transform (HST) is a very powerful tool for analysis of the nonstationary waveforms because it is able to extract the information from transient signals simultaneously in both time and frequency domains. Magnitude, frequency, and HST contours are the main attributes obtained from the output matrix of HST. At first, differential current waveforms of different conditions such as normal, internal and external faults, inrush, and over-excitation conditions are analyzed by HST and some potential useful features are extracted from the abovementioned contours. To decrease the dimension of feature vector and to increase the classification accuracy of the proposed algorithm, the most effective features are selected by using well-known feature selection methods namely sequential forward selection, sequential backward selection, and genetic algorithm. Selected features are trained by a probabilistic neural network as an effective classifier core, which has advantages regarding learning speed and generalization capability compared with feed-forward neural network. The classification accuracy of the proposed algorithm has been used as a criterion function for the selection of the best subset features. The proposed protection scheme is evaluated for various operating conditions of three different transformers using the PSCAD/EMTDC package. Extensive simulation results show that the proposed algorithm relies only on the waveshape properties, and it is independent of the value of transformer parameters and consumed power. Copyright © 2012 John Wiley & Sons, Ltd. key words: transformer differential protection; hyperbolic S-transform; probabilistic neural networks; feature extraction; feature selection 1. INTRODUCTION Power transformers are essential and important elements of power systems. Thus, the protection of power transformers is one of the most challenging problems in the power system relaying area. The protection scheme must be designed so that the relay operates only in internal fault conditions and remains stable in external fault, magnetizing inrush and over-excitation cases. There are generally two main types of approaches for discrimination of inrush currents from internal faults, harmonic restrain (HR) and waveform identification (WI) methods. HR-based methods are based on the principle that the magnetizing inrush, and over-excitation currents are rich in second and fifth harmonic components, respectively. Therefore, many digital filtering algorithms have been presented for computation of the harmonic components of differential current waveforms such as Fourier transform, Hartley transform, Kalman filter, and so on [1]. However, in the modern power trans- formers, because of magnetic properties of the core, the second and the fifth harmonic components may decrease in transformer energization and over-excitation conditions. Moreover, sometimes large second harmonic component can appear in internal fault currents [2]. Therefore, the HR-based techniques are not very effective for the detection of internal faults in power transformers. To overcome the difficulties *Correspondence to: Zahra Moravej, Department of Electrical and Computer Engineering, Semnan University, Semnan, Iran. † E-mail: moravej.zahra@gmail.com Copyright © 2012 John Wiley & Sons, Ltd. EUROPEAN TRANSACTIONS ON ELECTRICAL POWER Euro. Trans. Electr. Power (2012) Published online in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/etep.671