Adaptive differential protection for power transformer based on transient signal analysis

Power transformer protection is a technical and economic task that considers initial investments, operating costs and efficiency. In this context, this paper presents an adaptive relaying scheme for percentage differential protection of three-phase power transformers. The proposed formulation uses discrete wavelet transform to extract transitory features of the differential current signal, which presents the advantage of having an adaptive threshold value. The spectral energy of the low frequency wavelet coefficient is calculated and the change in this energy is used as decision criteria to produce the Trip/Restrain signal relay and distinguish internal faults from transient inrush currents. The proposed algorithm was implemented on MATLAB® environment and had its performance evaluated through simulation of several transient events using ATP software. Test results show that the wavelet transform can be efficiently used to identify fault conditions in power transformers making the proposed differential protection approach highly reliable, fast and accurate.