The calculation of chemical structures and physical properties is the basis of petroleum refining optimization. In this article, a method to build an average molecule model for hydrocarbons and petroleum fractions is developed. 1 H-NMR, molecular weight, and elemental composition were used as input parameters to construct a single molecular model that represents the average chemical structure. The average molecules were constructed by assembling the average building blocks, which are a set of predefined structural fragments, covering typical hydrocarbon and heteroatom functional groups in petroleum systems. After applying a group contribution method to the derived molecule, the bulk property of a sample could be directly calculated. The method was validated by being applied to various model compounds (including paraffins, cycloalkanes, aromatics, and heteroatom-containing species), where it successfully predicted the average building blocks, unit sheets number, and physical properties. The application of this method to petroleum fractions was demonstrated.

This work built a molecular-level kinetic model for hydrocarbon catalytic cracking, incorporating the catalyst acidity as the parameter to estimate reaction rates. The ndecane and 1-hexene co-conversion catalytic cracking process was chosen as the studying case. The molecular reaction network was automatically generated using a computer-aided algorithm. A modified linear free energy relationship was proposed to estimate the activation energy in a complex reaction system. The kinetic parameters were initially regressed from the experimental data under several reaction conditions. On this basis, the product composition was evaluated for three catalytic cracking catalysts with different Si/Al. The Bronsted acid and Lewis acid as the key catalyst properties were correlated with kinetic parameters. The built model can calculate the product distribution, gasoline composition, and molecular distribution at different reaction conditions for different catalysts. This sensitive study shows that it will facilitate the model-based optimization of catalysts and reaction conditions according to product demands.

We built a molecular-level kinetic model for hydrocarbon catalytic cracking, incorporating the catalyst acidity as the parameter to estimate the reaction rates. The n-decane and 1-hexene co-conversion catalytic cracking process was chosen as the studying case. The reaction network was automatically generated with a computer-aided algorithm. A modified linear free energy relationship was proposed to estimate the activation energy in a complex reaction system. The kinetic parameters were initially regressed from the experimental data under various reaction conditions. On this basis, the product composition was evaluated for three catalytic cracking catalysts with different Si/Al. The Bronsted acid and Lewis acid as the key catalyst properties were correlated with the kinetic parameters. The built model can calculate the product distribution, and molecular composition at different reaction conditions for different catalysts. The sensitive study shows that it will facilitate the model-based optimization of catalysts and reaction conditions according to product demands.