Bi-Directional Gated Recurrent Unit Approach for Detecting and Classifying High Impedance Fault in Power Distribution

In the recent days, detecting the High Impedance Faults (HIF) is an extremely difficult task because it is unpredictable, asymmetric, and nonlinear. Due to the size of the fault current is typically much lower than the normal load current, these faults are typically undetectable and impossible to isolate using traditional over-current approaches. Therefore, this research introduced a novel technique name called Bi-directional Gated Recurrent Unit (Bi-GRU) for power distribution-related HIF detection and isolation. The devised technique makes use of voltage as well as data from current sensors. In order to clarify the procedure, many types of algorithms were utilized during the detecting phase. For classification purposes, the algorithm Bi-GRU is applied on IEEE 13- distribution of node networks to test this method's ability to recognize and detect HIF with both damaged and unbroken wires. The suggested Bi-GRU provides fast and flawless output when troubleshooting high impedance issues. Bi-GRU considerably increases input processing and enhances classification accuracy. From the overall analysis, suggested Bi-GRU combines the forward and reverse GRUs in reverse order, which allows it to complete the operation more quickly and accurately than existing Morphological Fault Detector by achieving 95.63 % accuracy rate.