Corrosion prediction of iron-based metals by the prophet model

Abstract

This paper proposes the prophet model for corrosion prediction of iron-based materials. The corrosive degree of iron-based materials under different conditions is assessed though the weight loss method and the experimental results are fed to the proposed model as the training data. It indicated that HT200, Q345B, and 45 steel are easily and rapidly corroded in the periodic immersion condition. The predicted results from the Prophet model are compared with the experimental results to evaluate the performance of the prediction model. The highest RMSE is 2.35×10-4 and the largest MAPE is 2.90×10- 4%. That means the proposed model is a reliable machine learning model for corrosion prediction of iron-based materials.