Prediction of electric power systems influences on pipeline systems using artificial neural networks

Electric power systems may jeopardize the integrity of the buried metal structures of pipeline systems, such as: gas pipelines, oil pipelines, ore pipelines etc, in cases of proximity between both systems, especially under conditions of single-line-to-ground fault on a transmission line, due to fault current flowing through the earthing electrode of a tower into the soil. The maximum level of coating stress voltage of the involved pipelines must comply with the value specified on ABNT NBR 16563-1:2016. In this context, this paper presents the use of an Artificial Neural Network (ANN) model developed to predict the coating stresses voltages in terrestrial pipelines that are crossing transmission lines, with changing of network input parameters: fault currents, soil resistivity, angle between transmission line and pipeline and separation distances of the adjacent towers to the crossing location. The results obtained with the ANN model developed presented an average percentage relative error of 2.92% by comparing results from Sestech software package, proving to be in good agreement with it.