Identification and Track Improvement of Transmission Line Faults Using One Dimensional Complicated Neural Networks

Abstract This paper focuses one dimensional-convolution neural networks (1D-CNN) with a snappy adaptive architecture to prevent fault removal complications and categorization into a computational analytics system by introducing a fast and consistent algorithm for fault identification, categorization, and instructions for transmission networks. The proposed algorithm can be implemented directly to source data, eliminating the need for a separate removal step and leading in a more powerful protected approach. The proposed approach related to the 3-phase voltages and currents at the relay position within the electric power line framework is considered as feedback for the recommended 1D-CNN algorithm. In order to provide the training and test statistics for the anticipated 1D-CNN approach, a 133 kV grid connected line is simulated using MATLAB/SIMULINK software. The suggested algorithm's test precision is contrasted with two additional neural and fuzzy neural network (FNN) standard techniques. The numerical outcomes of the test show that, in comparison to other conventional methods under several fault situations, the suggested diagnostic device is fast and efficient for categorizing and controlling the discrepancy of transmission line fault with greater accuracy.