A Transient-Extracting Transform Assisted Intelligent Fault Detection and Classification Approach for UPFC Installed Transmission Line

Abstract

The frequent changes in current and voltage pro-file during different operational modes of unified power flow controller (UPFC) limit the performance of traditional distance protection of transmission lines (TLs). This paper looks into the effect of the UPFC on the TL protection and propose a new fault detection and classification (FDC) technique to support the relay. The sample-to-sample difference of the transient-extracting transform (TET) operator of modal component of current for detection of fault, first. Later, the one-cycle post fault TET operators of three-phase currents and ground current are utilized as features for various classifier like back propagation neural network (BPNN), support vector machine (SVM), decision tree (DT) and random forest (RF) for classification of fault. This protection technique is tested by generating various cases during fault in a standard two-bus MATLAB/Simulink model and validated in OPAL-RT based real-time simulator. The low detection time and high accuracies of different classifier indicate the reliability of TET feature based FDC technique for such crucial TL.