Development of symmetrical fault detection during power swing based on entropy

This paper proposes the new technique for detecting symmetrical faults occurred during power swing. The proposed technique is simulated for the protection of two machine system, distribution system with distributed generation and WSCC 9-bus system in Matlab. Symmetrical fault detection during power swing is a critical issue. Hence, in this paper, modified weighted wavelet packet entropy technique is proposed to detect symmetrical fault during power swing. This entropy criterion is applied to wavelet packet coefficients to enhance the energy of fault signals and to reduce the vector size of the wavelet packet transform coefficients. This technique provides better results irrespective of various fault locations, fault inception angle and power swing frequencies. This technique is compared with conventional scheme and the results are tabulated. Moreover, this technique is also able to detect asymmetrical faults with high impedance during power swing. Mathematical derivation support system performance through simulation.