Information and Diagnostics Systems

Development of an Artificial Neural Network to Predict the Corrosion Rate in Carbon Steels Under Atmospheric Conditions

Michelle Nuñez Morales1, Ildefonso Esteban Pech Pech1, Gustavo Domínguez Rodríguez1, Oscar de Jesús May Tzuc2, Francisco Eduardo Corvo Perez1

1. Universidad Autónoma de Campeche, Centro de Investigación en Corrosión, México
2. Autónoma de Campeche, Facultad de Ingeniería, México

https://doi.org/10.15407/msse2025.01.256

PDF DocumentPDF

Keywords

corrosion rate, artificial neural network (ANN), data mining

Abstract

Atmospheric corrosion of carbon steel under atmospheric conditions is a complex, nonlinear process. It involves a large number of interacting and varying factors: material composition, location of exposure, temperature, relative humidity, wet-dry patterns, amount of precipitation, concentration of main pollutants, etc. The most important variables were chosen from a data mining that was carried out with experimental results that were documented worldwide. The resulting statistical correlations served as the basis for the modelling of the artificial neural network (ANN). Although the modelling of the ANN is still being improved, the results show a R-square value of 0.94244 indicating that the ANN model has a good fit for the predicted corrosion rate. A global sensitivity analysis (GSA) and the development of a Graphical User Interface are the final purpose of this study.

References

  1. Tran N. L., Nguyen T. H., Phan V. T., & Nguyen D. D. A Machine Learning‐Based Model for Predicting Atmospheric Corrosion Rate of Carbon Steel // Adv. Mater. Sci. Eng. – 2021. – 1. – № https://doi.org/10.1155/2021/6967550
  2. Morcillo M., Chico B., D´ıaz I., Cano H., De la Fuente D. Atmospheric corrosion data of weathering steels. A review // Corros. Sci. – 2013. – 77. – P. 6–24. https://doi.org/10.1016/j.corsci.2013.08.021
  3. Morcillo M., Atmospheric corrosion in Ibero-America: the MICAT project in Atmospheric Corrosion ASTM International // West Conshohocken, PA, USA. –
  4. Yan L., Diao Y., Lang Z. & Gao K. Corrosion rate prediction and influencing factors evaluation of low-alloy steels in marine atmosphere using machine learning approach // Sci. Technol. Adv. Mater. – 2020. – 21(1). – P.359-370. https://doi.org/10.1080/14686996.2020.1746196
  5. Peralta Arboleda B.A. Desarrollo de una red neuronal artificial para la predicción de la tendencia corrosiva e incrustante en tuberías de producción de petróleo // Bachelor’s thesis, Quito. – 2020
  6. Bolaños-Rodríguez E., Pérez-Rojas A., Vega-Cano G. Y. & González-Islas J. C. Monitoreo de Corrosión Atmosférica de Acero en el Sureste Mexicano Aplicando Redes Neuronales Artificiales // Construyendo el Software del Futuro. – 2014.