Two-Stage Hybrid-Filtering Based Fault Detection & Classification method for Active Distribution Networks

Active distribution networks (ADNs) are the modern power networks that are cultivated due to the widespread dispersion of renewable energy resources (RERs) near consumer territory. However, faults detection and classification is an issue in such ADNs owing to the low current level during faults, and bidirectional power flows. This paper establishes a new fault detection and classification method for ADNs. In the first stage, a discrete Kalman filter (DKF) is implemented on measured current signals for noise-less state estimations. In the second stage, the second-order low pass filter (SOLPF) is implemented to the per phase current signature to take out the desired filtered features (DFF). Furthermore, the DFF of the current signal is squared, and then the exponential is taken to compute the single-phase fault detection & classification index (SPFD&CI). If the SPFD&CI of any phase is more than a constant threshold level the associated phase is deliberately faulty. Moreover, due to phase segregation, the fault categorization is autonomous. The suggested approach is tested in MATLAB/Simulink firmware on the ADN's tested. The results show that, in various cases, the suggested technique detects and classifies all varieties of fault conditions with less than 1/2 a cycle.