Coke Yield Prediction Model for Pyrolysis and Oxidation Processes of Low-Asphaltene Heavy Oil

Accurate prediction of coke yield is the basis for simulation and operation of the in situ combustion process. In view of the low asphaltene content of heavy oil in China, this study established coke yield prediction models for the pyrolysis and oxidation processes of low-asphaltene heavy oil based on the coking characteristics of the main fractions, including saturates (S), aromatics (A), and resins (R). The models were verified by the experimental coke yields of heavy oils using a fixed-bed reaction system. The prediction model for the pyrolysis process included evaporation rate equations and kinetic equations for coking reactions. Besides the same parts with pyrolysis, the model for the oxidation process included the kinetic equation of coke oxidation. The evaporation rate equations were determined by fitting the mass loss rate in the evaporation range of thermogravimetric analysis (TGA) pyrolysis results with a maximum error of 6%. The kinetic parameters for the conversion of each fraction to the coke precursor were determined by the isoconversional method based on TGA results of the separated SAR fractions. Considering the interactions among fractions during oxidation, this study included mixtures of SAR fractions as the samples and obtained the mass loss results of a single SAR by subtraction. The activation energies of SAR pyrolysis range from 99 to 180 kJ/mol and those of SAR oxidation range from 92 to 130 kJ/mol. The kinetic parameters of coke from the coke precursor were obtained by temperature-programmed coking data of a prepared intermediate, with an activation energy of 177 kJ/mol in an inert atmosphere and 65 kJ/mol in an oxidizing atmosphere. The kinetic parameters of coke oxidation were derived by the isothermal kinetic method in TGA, with an activation energy of 136 kJ/mol. The experimental coke yields of Fengcheng and Hongqian heavy oils under various pyrolysis and oxidation conditions were predicted by this model with an error around 10%, indicating that this model could be successfully applied to the prediction of the coking process of low-asphaltene heavy oil under complex conditions.