Sheath induced voltage prediction of high voltage cable based on artificial neural network

This paper aims to propose an Artificial Neural Network (ANN) model for voltage prediction in cable sheath of combined overhead-cable line under lightning condition. To this end, the effect of different parameters, including tower footing resistance, sheath ground resistance, a kind of ground connection of sheath on the maximum induced voltage of cable sheath in 132 kV combined line are investigated using EMTP/RV software. It is assumed, in this study, that lightning strike to the Guard wire and back-flashover occurred and/or lightning strike to the overhead line. With these results in mind, the proposed model is designed with ten inputs data and four outputs data. The validation of the model indicates that the absolute values of relative errors between induced voltages of simulation and prediction are less than 8%. This indicates high efficiency of ANN technique in the maximum induced voltage prediction of cable sheath under lightning surge.