Single line to ground fault detection and location in medium voltage distribution system network based on neural network

The aim of this project was to detect and locate the single ground failure lines that occurs in medium voltage (MV) networks on the transmission lines (TL). Compared with anther faults, single line-to-ground (SLG) is the most frequent. The neural network (NN) algorithm was advanced in order to discover and locate SLG faults. The network is simulated through simulated numerous defects at various locations, as well as changing earth resistance from (or 100 Ω) to TL to gather all of the data. In the electromagnetic transients’ program (EMTP) program software, the existing fault have been measured. In addition, the waves were evaluated by utilize MATLAP's fastfourier-transform to calculate the waves of top three of them, On the MV network are fifty hundred faults are simulated all data in the neural network at MATLAB were trained and examined to improve the NN algorithm according to this data. Comparing all the simulated location faults that have been applied with those all locations detected in the NN algorithm, the overall error between them has been found to be very low and not to exceed 0.7. The Simulink circuit was created from this algorithm and checked in order to predict each failure could occur in the future in the MV network.