Power Transformer Inrush Current Identification Using Relative Wavelet Energy

the magnetizing inrush current of the power transformer is a significant challenge to be successfully identified by the differential protection relay. In this paper, an approach based on wavelet coefficients energy is presented to successfully identify the inrush currents. The wavelet transform is applied to decompose the differential currents into the details coefficients in several levels of frequency bands; it is observed by inspecting different inrush and internal faults cases, a fault condition is the most probable if a remarkable change in detail coefficients is noticed. This fact is utilized in order to establish a diagnosis criterion, where the relative wavelet energy is defined as a characterizing quantity. The discrimination criteria is based on the energy variation amount, which is monitored comparing the values of the relative wavelet energy in two successive time intervals, where an inrush is detected generally when the difference is not a remarkable quantity. The reliability of the proposed approach is checked through simulation of different transient cases using MATLAB.