(122f) CFD Study of Lignocellulosic Biomass Gasification in a Fluidized Bed Gasifier: A Comparison of Eulerian and Lagrangian Representations of the Biomass Fuel

Thermochemical conversion of biomass to drop-in ready fuels and chemicals offers an attractive option to directly displace petroleum consumption and effectively utilize the energetic and chemical resources available in waste lignocellulosic materials as well as dedicated energy crops. Further, thermochemical conversion technology has the benefit of being governed by a shared set of chemical pathways, such that depending on the desired product(s) reactor conditions (temperature, pressure and availability of oxidant) and design parameters (geometry, freeboard height, etc) can be optimized for many different products from the production of hydrogen-rich syn-gas (gasification) to oxygenated bio-oil (fast pyrolysis).

In this work a Computational Fluid Dynamic (CFD) investigation of biomass gasification is undertaken utilizing the open-source code, MFiX, with a particular focus on the general applicability of the modeling framework to the multitude of reactive regimes in the conversion process. The proper representation of the pyrolysis of the virgin biomass feedstock is tantamount in predictively modeling the gasification process, furthermore it has been demonstrated that the pyrolysis process of biomass is controlled by both chemical and particle-scale transport processes. In CFD modeling of fluidized systems the gaseous species are represented in an Eulerian framework and the solid phases can be represented as either an Eulerian fluid or as a collection of individual Lagrangian particles. When the reacting particles are represented as an Eulerian continuum individual particle-scale transport phenomena are neglected, whereas in a Lagrangian representation this can be directly modeled.

In this paper an Eulerian-Eulerian-Eulerian (EEE) and an Eulerian-Eulerian-Lagrangian (EEL) representation of the gases, silica bed material and fuel (biomass/char) respectively are compared. The trade-offs between these two representations - the benefits of added model fidelity versus computational complexity and cost - are analyzed. Additionally, the mixing dynamics are considered in the context of reactor design considerations such as the location of the fuel and oxidant injection ports.