(608d) Exploring Chemistry into Heavy Hydrocarbons: Automatic Fragment Modeling | AIChE

(608d) Exploring Chemistry into Heavy Hydrocarbons: Automatic Fragment Modeling

Authors 

Wang, Y. T. - Presenter, Massachusetts Institute of Technology
Green, W., Massachusetts Institute of Technology
Heavy hydrocarbons are the starting material for lighter hydrocarbons which have more commercial value; often this conversion is accomplished by pyrolysis. Modeling techniques for the pyrolysis of light hydrocarbons (e.g. steam-cracking of naphtha, also processes in fuel-rich combustion) are very advanced, allowing us to create detailed kinetic models that give very helpful predictions that aid in understanding, design, and optimization of those systems. It would be very helpful if similar modeling techniques could be applied to heavy hydrocarbons. However, as the number of carbon atoms in a molecule increases, the number of possible isomers grows very rapidly, and soon gets intractable for finite computation resources. Not even mention the potential number of possible reactions can occur in such systems may be explosive and cause limitation to existing predictive tools. Although some lumping methods have been developed for estimation and simulation of heavy residues, the ability of extrapolation is uncertain and the estimated parameters might only be accurate at certain conditions. To overcome the difficulties mentioned above, a framework called Auto-Fragment Modeling is proposed. This approach considers pieces of the large molecules as independent fragments which react independently. By defining different fragments, the molecules are treated as different structures where some sharing same ones and can have similar reactions occurring; this approximation overcomes the combinatorial explosion, making it possible to model heavy hydrocarbons using methods previously developed for light hydrocarbons. The open-source software Reaction Mechanism Generator (RMG) software package has been modified to construct the corresponding kinetic models. This framework is validated by some examples, which show promising results by capturing important chemistry and apparent product profiles. With the working idea of fragmentation, it is possible to further elucidate our understanding heavy hydrocarbon chemistry.