Abstract: In view of the large variation of catalytic reforming feed properties, a new method of Delta-Base data generation by variable step size calculation and molecular level mechanism model was proposed. The method calculates the change rate of the absolute value of product mass fraction with the change of the absolute value of feed property when the feed property is disturbed, and divides the range of the rate of change into several equal parts. According to the different range of the change rate value obtained from each calculation, the variable disturbance step was selected for the next step of calculation, and then the midpoint value of each variable disturbance in the calculation process was taken to verify the product quality score, thus, the accuracy of the calculation results was ensured while the calculation efficiency was improved. The results showed that the relative error of each product quality fraction calculated by this method was less than 2.5%, the fitting equations of different product quality fractions could be obtained by fitting the calculated data, and the three-stage Delta-Base data of aromatic potential content variation could be generated by fitting the equations. Compared with the product prediction results, when the aromatic potential content of the feed was 42%, it was found that the average deviation between the predicted value of the Delta-Base data and the calculated value of the fitting equation was 0.16%, which showed that the method had excellent accuracy, and the Delta-Base data generated by the molecular level mechanism model can better reflect the real molecular reaction process.

Key words: catalytic reforming, molecular level model, linear programming, Delta-Base data