To Avoid Malfunction due to Inrush current Using Neural Network Fitting Tool

Abstract

he growing power system demands a speedy and reliable protection module for Transformer which is its one
of the important link to maintain uninterrupted power supply. The protective module should quickly initiate the tripping
during internal faults and should not allow the tripping during magnetizing inrush. The work reported in this paper
demonstrates the use of Neural Network fitting tool to be incorporated in association with differential relay operation in
the protection scheme. The ANN based algorithm is trained to recognize the waveform pattern which permits the Relay
Trip operation only during short circuit conditions due to internal faults and not for inrush current. Modelling of ANN
using feed forward Levenberg-Marquardt Back propagation techniques is used for implementation. The performance of
the Neural Network has been tested with simulated signal which proved to be the technique having accuracy and fast
classification.