(560o) Computational Study on Biomass Fast Pyrolysis: Design Considerations for a Laboratory-Scale Fluidized Bed

Ramirez, E. - Presenter, Oak Ridge National Laboratory
Li, T., National Energy Technology Laboratory
Finney, C. E. A., Oak Ridge National Laboratory
Shahnam, M., National Energy Technology Laboratory
Daw, C. S., Oak Ridge National Laboratory
Fast pyrolysis is a leading candidate process for converting biomass to liquid fuels. During fast pyrolysis in bubbling bed or circulating bed reactors, biomass particles are rapidly heated through contacting with hot gases and solids, and their constituent components decompose into volatiles, ash, and char. The product vapor/gas composition, which determines the yield of fuel-compatible molecules, is highly dependent on the bubbling intensity which promotes mixing and heat and mass transfer within the biomass particles and at the particle surfaces as they transit through the reactor. Fluidized bed hydrodynamic characterization at smaller scales is a vital first step in reactor scale-up.

In this study, we simulate a 3D bubbling fluidized bed biomass fast pyrolysis reactor from a prior study [149]. This study explores operating effects on hydrodynamics and biomass conversion as the gas flow is increased through the bubbling-to-slugging transition and turbulent regime, with all the other operating variables held constant. We employ MFiX, an open-source software package supported by DOE, which utilizes a continuum (two-fluid) approach for modeling the reactor hydrodynamics. Bubbling intensity and dynamic characteristics were evaluated utilizing pressure-based measurements [http://dx.doi.org/10.1016/j.cej.2016.08.113]. A novel entropy approach based on time irreversibility is introduced to evaluate hydrodynamics.

Mixing, hydrodynamics, and pyrolysis yields are compared which show the effect of fluidizing gas and fluidization regime on biomass fast pyrolysis in bubbling bed reactors of Geldart B particles. This work highlights the importance of initial reactor design for optimizing yield. We will discuss implications on future numerical simulations and experiments based on our observations.