Auto-cross correlation coefficients used for detecting an internal fault in three phase transformer

A new algorithm based on correlation coefficients analysis of waveforms to detect an internal fault occurs in three phase transformer is proposed. The algorithm depends on the changes in current waveform that occur during internal fault. Auto-correlation has been added to cross-correlation algorithm for enhancing the security and speed in detecting the fault within very short time. It is able to detect the fault after just 0.4 msec from incidence of fault occurrence. The scheme has been tested by using Matlab/Simulink software [1]. The test results show that the internal fault even inter-turn type can be correctly identified from external fault. The algorithm is simple and can be experimentally implemented.