High-speed directional protection based on cross correlation of Fourier transform components of voltage and current

In this study, a new algorithm for directional protection, based on cross correlation of Fourier transform components of voltage and current for each phase is proposed. The fault is detected using k-nearest neighbour algorithm. By comparing the fifth-nearest neighbour of each sample with an adaptive threshold, the fault is detected. The adaptive threshold is derived by a moving window with length equal to 1 ms. Using Fourier transform, the fifth sinusoidal component of the measured voltage and current for each phase is extracted. Then the cross correlation of the voltage and current components is calculated. The calculated cross correlation signal is passed through a low-pass filter. By comparison of the filtered signal with a predefined threshold, fault direction is determined. It is confirmed that the fault direction can be identified in less than 8 ms with high accuracy of nearly 99%. More than 6500 scenarios are considered for different fault conditions to evaluate the proposed method. Furthermore, using some experimental results the proposed method performance is confirmed. The simulation and experimental results show that the proposed method has high speed, good accuracy, and sensitivity.