Stockwell Transform based Decision Tree for Transmission Line Fault Diagnosis

The faults on transmission lines affect the normal operation of the power system and its components, which are to be diagnosed and removed as fast as possible to restrain the system stability and reliability. In this paper, a Stockwell transform based decision tree is proposed for fault diagnosis in a two-terminal transmission system. The current signals measured from generator bus are transformed into a time-frequency plane using Stockwell Transform (ST). The magnitude of the complex matrix obtained from ST is statistically analyzed. It was found that standard deviation and skewness of the maximum magnitude vector can be utilized for diagnosis of various transmission line faults. An algorithm based on such parameters is proposed using ruled decision tree for fault classification. The proposed algorithm aims at detecting and classifying the faults using half cycle information, available after the fault incidence.