Artificial Neural Network Modeling for Al-Zn-Sn Sacrificial Anode protection of Low Carbon Steel in Saline Media

Abstract

This work presents the artificial neural network(ANN) modeling for sacrificial anode cathodic protection of low carbon steel using Al-Zn-Sn alloys anodes in saline media. Corrosion experiments were used to obtain data for developing a neural network model. The Feed forward Levenberg-Marquadt training algorithm with passive time, pH, conductivity,% metallic composition used in the input layer and the corrosion potential measured against a silver/silver chloride(Ag/AgCl) reference electrode used as the target or output variable. The modeling results obtained show that the network with 4 neurons in the input layer, 10 neurons in the hidden layer and 1 neuron in the output layer had a high correlation coefficient (R-value) of 0.850602 for the test data, and a low mean square error (MSE) of 0.0261294. 9