Optimization of Extractive Pretreatment of Naphtha Cracker Feedstock Using Response Surface Methodology

This study has examined the extractive dearomatization of naphtha cracker feedstock. By considering the solvent-to-feed ratio (2:1-4:1), solvent water content (0-10 wt.%), and extraction temperature (30-60 ºC) as major operating variables and the solvent capacity, solvent selectivity, and the yield and the degree of paraffinicity of the dearomatized fraction as the response of experiments, the priority of variables in the prediction models was determined. The analysis of variance and a standard multilayer perceptron artificial neural network showed that in extractive dearomatization of the blended naphtha cut with sulfolane solvent, the priority of variables follows the trend of solvent water content, solvent-to-feed ratio, and temperature. Also, by optimizing the operation, using the penalty function approach, and defining six different single- and multi-objective functions, the optimum values of the solvent water content, solvent-to-feed ratio, and extraction temperature were found as 9.98%, 3.86:1, and 32.6 oC, respectively. Finally, the maximum values for the degree of paraffinicity and the yield of the dearomatized fraction were obtained as 94.36 and 95.95%, respectively.