MODELING AND PREDICTION OF HIGH TEMPERATURE OXIDATION OF STEELS USED IN PETROLEUM REFINERY HEATERS

This paper implements Response Surface Methodology (RSM), with rotatable Central 
Composite Design (CCD), in modeling and prediction of high-temperature oxidation of steels used in 
petroleum refinery heaters. A mathematical quadratic model was developed for modeling and prediction 
of the relationship between the weights gained of the oxidized steel and condition parameters, namely 
chromium percentage, temperature and exposure time with a range of, 0-5wt%, 450-500°C, and 0-40hrs 
respectively. Also, the surface plots and response surface contour plots were constructed for investigating 
the simultaneous effect of the condition process parameters on the weights gained from the oxidized steel 
and determining the optimum values of the parameter conditions. The Mean Absolute Error (MAE) between 
the actual and predicted weights gained from the oxidized steel values was calculated to test the validity of 
the developed mathematical model. The value of the (MAE) was 0.42 which indicates that the developed 
mathematical model have good agreement with the actual. The statistical analysis of variance (ANOVA) was 
also utilized to analyze the effect of the investigated parameter condition, this analysis results demonstrated 
that chromium weight percent and temperature have the most significant effect on the weights gained 
oxidation of steels used in petroleum refinery heaters followed by exposed time.