Hardware-in-Loop Implementation of ANN Based Differential Protection of Transformer

Abstract

This paper presents an efficient Artificial Neural Network (ANN) approach for discriminating the internal faults from the non-internal faults in a transformer. The wavelet transform is a powerful tool for analyzing transient conditions because of its ability to extract information both in time and frequency domain simultaneously. Simulation of the differential protection scheme of a transformer to obtain various operating conditions is done using MATLAB/SIMULINK taking 1 cycle of data window (20 msec.). Different operating conditions such as normal, internal fault, external fault, switching inrush, and over fluxing are analyzed and processed to obtain certain statistical parameters of wavelet coefficients at the different decomposition levels. Authors have used Arduino Uno ATmega328P platform for hardware implementation of ANN architecture. Results indicate that overall classification accuracy is found to be 95.63 %.