Mathematical and artificial neural network models for corrosion of high temperature — High pressure boiler pipes in presence of oxygen scavengers

This research reports the consequences of mass loss estimations by utilizing of oxygen scavengers in boilers to limit the consumption rate of carbon steel tubes. The consumption rate information was chosen from the literature as function of temperature, pressure, and inhibitors level. Consumption rate increased with increasing in temperature and pressure, and diminished with inhibitor level. Hydroquinone, ascorbic acid, and ethanolamine were utilized as consumption inhibitors. The present work is focused on deciding the artificial neural network (ANN) and mathematical formulas so as to increase great expectation properties. Five scientific models and five ANN display structures were recommended. Computer aided program was utilized for building up these models. The outcomes demonstrate that the polynomial mathematical model and multi-layer discernment can precisely anticipate the deliberate information with high relationship coefficients. Multi-Layer Perceptions (MLP) 3:3-4-1:1 was the best model over than the others with higher correlation coefficient.