Discrimination between inrush and fault condition in transformer: a probabilistic neural network approach

Abstract

In this paper, an algorithm has been developed around the theme of the conventional differential protection of the transformer. The proposed algorithm is based on probabilistic neural network (PNN) and use of the spectral energies of detail level wavelet coefficients of differential current signal for discriminating magnetising inrush and fault condition in the transformer. Performance of the proposed PNN is investigated with the conventional backpropagation feed forward (BPFF) multilayer perceptron neural network. To evaluate the developed algorithm, relaying signals for various operating condition (i.e., inrush and fault) of the transformer, are obtained from a custom-built single-phase transformer in the laboratory. He has published more than 50 research papers in different international and national journals and conference proceedings. His main research interest includes power system planning, operation and control, intelligent methods, and its applications in power system, VLSI and embedded design, signal and image processing. He is also an IEEE, IACSIT, IE and ISTE member.