Fault Location Prediction in Power Transmission Lines Using an Artificial Neural Network Model

Energy transmission lines are an important element that ensures the sustainability of existing living conditions and the uninterrupted need for electricity. Therefore, it is of great importance to locate short circuit faults that may occur in transmission lines and to intervene in these faults immediately. In this study, a fault location study is carried out in a power system designed as a real line model. Firstly, a 478.9 km long transmission line with three transpositions was created using the EMTP/ATP program. The study starts from a point close to the beginning of the line until the first transposition, and a three-phase ground fault is simulated in the system at certain intervals. The input current and voltage values of the line are then taken. The obtained data were analyzed by applying Modal Transformation to the current and voltage signals in the fault condition. Thus, the occurrence of high-frequency harmonics of the fault condition is characterized. Afterwards, in order to locate the fault in the system, Fast Fourier Transform (FFT) spectra were obtained using MATLAB software. These spectra were trained with Artificial Neural Networks (ANN) using fault data analyzed at 74 different points, and the fault location was determined with 99% accuracy.