Comparison of ANN and ANFIS Methods for the Voltage-Drop Prediction on an Electric Railway Line

Railway electrificationsystems are designed with regard to the operating data and design parameters.The minimum voltage rating required by traction during the operation should beprovided. The maximum voltage drop on a line determines the minimum tractionvoltage. This voltage should be maintened within certain limits for thecontinuity of operation. In this study, the maximum voltage drop generated viatraction was determined using artificial neural network (ANN) and adaptiveneuro-fuzzy inference system (ANFIS) for a 25-kV AC-supplied railway. Thevoltage drop on line was calculated with regard to the operating data using ANNand ANFIS. ANN and ANFIS were explained, and the results were compared. TheLevenberg–Marquardt (LM) algorithm was used for the ANN model. The LM algorithmis preferred because of the speed and stability it provides for the training ofANNs. The data created for one-way supply status were examined for simulation.