Molecular Modeling of Coprocessing Biomass Fast Pyrolysis Oil in Fluid Catalytic Cracking Unit

Abstract

Integration of renewable sources into transportation fuels production system through FCC units in an oil refinery has gained increased attention. For better understanding of the effect of reaction conditions, blending ratios and feed properties on product yields and qualities, kinetic modelling of FCC units is necessarily required. In this paper, a novel framework for molecular-level modelling of co-processing biomass pyrolysis oil with VGO in an oil refinery FCC unit is developed, which includes molecular-level characterisation of biomass pyrolysis oil and VGO feed blends, synthesis of large-scale and complex reaction network, molecularlevel kinetic modelling and parameter estimation. The rule "same type of reactions has similar activation energies" is employed to significantly reduce the number of kinetic parameters. The kinetic parameters in the proposed model are estimated using a hybrid solution algorithm combining deterministic and stochastic optimisation methods. The computational results demonstrate an overall good agreement between measured and predicted yields using the developed kinetic model for VGO: FPO blending ratio, C/O ratio and reaction temperature of 95:5, 5 and 530 °C, respectively. PONA composition in each layer of product stream (e.g. gasoline, diesel, gasoil, etc.) as well as oxygen compounds compositions and oxygen content are also successfully predicted. The proposed framework can be easily to be extended for modelling of other refinery processes and creates potentials for rigorous simulation and optimisation of refinery operations in order to achieve maximisation of refinery profit or better product quality control.

Usefulness of Lumped Kinetic Modeling

“…The drawbacks of detailed modeling are that the simulation is generally computationally expensive and requires significant amounts of time and processor power to run. Subsequentially, high computation times of single model runs significantly affect the computational burden for the creation of those systems and the determination of unknown parameters …”

“…Subsequentially, high computation times of single model runs significantly affect the computational burden for the creation of those systems and the determination of unknown parameters. 30 Lumped kinetic modeling (LKM) is a technique which attempts to reduce the complexity of a detailed model while preserving certain physical and chemical relationships. This is achieved by grouping together several components and elementary reactions into an apparent reaction scheme.…”