Comparison of mathematical and artificial neural network models for inhibition of fuel oil ash under high temperature corrosion

Abstract

This research reports the results of literature data of mass loss tests of high temperature corrosion inhibition of steel in different concentration ratios of MgO, Al₂O₃ and SiO₂ to corrosive fuel ash of V₂O₅ in the temperature range of 550–590°C and time range of 8–100 h. Analysis focused on determining optimum mathematical equation and artificial neural network (ANN) architecture in order to gain good prediction properties. Three mathematical equations and five ANN architectures were suggested. A computer aided program was used for developing these models. Results show that polynomial mathematical equation and multilayer perceptron are able to accurately predict selected data with high correlation coefficients.

Keywords:

• Corrosion
• Steel
• High temperature
• Inhibitor
• Fuel ash

Acknowledgement

This work was supported by Diyala University, Chemical Engineering Department, which is gratefully acknowledged.