K-means Clustering and Linear Regression Based Protection Scheme for Transmission Line

Abstract

In this paper, shunt fault is detected, classified and located in power transmission line. The synchronized Current signals at both the bus of the transmission line are used for obtaining approximate coefficients using wavelet transform. The proposed algorithm is used K-means clustering to detect and classify the faults. Fault location is estimated using linear regression method. The approximate wavelet coefficients of the both the buses are added to get resultant approximate coefficients. K-means clustering is applied on these resultant approximate coefficients to computes two centroid in a half cycle. The centroid difference (C.D) is computed, and the basis of the centroid difference the fault is detected and classified. Various case studies such as vary fault location, fault incidence angle and fault impedance to verify the robustness of the algorithm.