Combined Fault Classification and Location Scheme for TCSC-compensated Transmission Lines using Machine Leaming Techniques

This article proposes the hybrid framework for detection, classification and location of transmission line fault. The proposed technique utilizes symmetrical component analysis, fast walsh-hadamard transform, bagged decision tree ensemble classifier and Gaussian process regression for fault diagnosis task. Fast Walsh-Hadamard Transform (FWHT) has been utilized to estimate the effective features from the faulty signals. A modified 3-machine 9-bus WSCC TCSC compensated line is considered for study and different types of fault and system scenarios is simulated. The simulated test result confirms that the proposed scheme is accurately classify the faults with $\lt 1\%$; of location error.