Prediction of Atmospheric Corrosion of Ancient Door Knockers via Neural Networks

Abstract

The importance of door knockers persuades us to anticipate the atmospheric corrosion through Neural Network (NN) which is validated by data originated from literature. NNs are used in order to anticipate the effective parameter on bronze atmospheric corrosion including the ambient temperature, exposition time, relative humidity, PH, SO₂ concentration as an air pollutant and also metal's precipitations. As these factors are extremely complicated, exact mathematical language of the diverse metals corrosion are not comprehended. The results of this study showed that SO₂ concentration as an air pollutant and time of exposition are the fundamental effects on corrosion weight loss of bronze.