Analysis of Underground Cable Fault Location Identification Using Artificial Neural Network

Abstract

Transmission outlines are the backbone of power schemes or other power utilities because they are used for power programmed or distribution. Power is distributed to end operators via overhead or underground cables. For underground cables, as time passes, the possibility of failure will increase. Now, the failure rate of older underground power cables and increasing system failure rates are seriously affecting the reliability of the system and the many losses involved. Therefore, it is clear that the necessary measures must be taken to manage the consequences of this trend. At any given length of cable, its indications of damage or failure are manifested through discrete defects. Identifying the type of defect and its location along the length of the cable is critical, this is the key to minimizing operational costs by troubleshooting while reducing long and costly patrols and speeding up repair and restoration of the power line. In this study, we propose a fault location and a recognition method that combines wavelet and neuro-diffuse technology. I developed a 100km, 50Hz transmission line model and simulated different faults and locations in MATLAB / SIMULINK, and then used some selected features of the transformed signal as input for ANN training and development. ANN technology Accurate enough to locate underground faults in the power line.