(737e) Multi-Scale Simulation of Steam-Oxygen Blown Bubbling Fluidized Bed Biomass Gasification

Bubbling fluidized bed biomass gasification (BFBBG) is seen as a promising technology for the conversion of lignocellulosic biomass to power and fuels. During BFBBG, biomass is injected into a high temperature fluidized bed (700-900 ^○C), where it rapidly devolatilizes forming char as well as a variety of other combustion products, light gases and condensable compounds known as tar. The porous char also reacts with steam and carbon dioxide to form syngas (hydrogen, carbon monoxide) which after cleanup can be upgraded to liquid fuels using Fischer Tropsch synthesis.  Because of the high levels of heat/mass transfer and the thermal inertia of the bed material, fluidized beds can be used to process a wide variety of feedstocks with minimal preparation.  Despite these advantages, the presence of condensable tars in the product syngas (2-50 g/Nm³)[1],  low carbon conversion, and scale-up complexity, remain as major hurdles to further implementation.  Because of the high capital expense of instrumented reactors, there has been increased interest in applying detailed simulation in order to improve physical understanding as well as aid in scale-up and optimization.

Recently, the multi-fluid (Eulerian-Eulerian) model (MFM) has been adopted to simulate reacting fluidized bed biomass gasification and pyrolysis systems. It describes the gas and solid phases as inter-penetrating continua and a number of closure relations are required for inter-phase transfers and boundary conditions. Compared to other computational fluid dynamic (CFD) approaches, the multi-fluid approach is highly scalable while still giving detailed spatio-temporal description of the motion of the gas and solid phases.  However, even with state of the art computing capabilities, simulation times achievable for reacting two dimensional [2] or three dimensional [3] fluidized bed system have been short (less than 60 seconds). Meanwhile, the average solids residence time of reacting char particles is on the order of 10² to 10³ seconds. It can be expected and has been experimentally observed that the time for the char to reach steady state inventory may take several multiples of this (hours to tens of hours depending on the bed size and conditions [4].  It is important to describe the steady state char inventory because it directly affects the rates of gasification and combustion therefore dictates the carbon content of the major gas species and overall gasification efficiency. Due to this mis-match in physical and simulation time-scales, MFM type simulations cannot be used to accurately predict the steady-state major gas specie output of fluidized bed gasifiers without some a priori knowledge of the bed carbon inventory.

The aim of this work is the development of a multi-scale simulation which combines an initial model to compute/estimate steady state char inventory using this output as an input for a shorter (30sec) transient MFM simulation. In the first part of this presentation, the formulation of a standalone, transient, isothermal single particle model accounting for external mass transfer, intraparticle diffusion, and reaction of the combustion and gasification of a single batch of particles is presented.  The results are validated against published experimental data of batch fluidized bed combustion experiments performed with char produced from pelletized and non-pelletized feedstocks [5] . Through appropriate residence time averaging, the single particle model is extended to describe a realistic continuously fed system. In the second part, the particle model predictions for steady state char inventory are utilized as an input for the MFM model.  The multi-scale, steady state simulation are compared with experimental data from a 4 inch diameter steam-oxygen blown bubbling fluidized bed gasifier using varying levels of oxygen input and temperature.

Sources Cited:

[1]           Milne T., Evans RJ, Abatzoglou N. Biomass gasifier “tars”: Their nature, formation and conversion. Boulder, CO: National Energy Technology Laboratory (NETL); 1998.

[2]           Gerber S, Behrendt F, Oevermann M. An Eulerian modeling approach of wood gasification in a bubbling fluidized bed reactor using char as bed material. Fuel 2010;89:2903–17. doi:doi: 10.1016/j.fuel.2010.03.034.

[3]           Zhang N, Lu B, Wang W, Li J. 3D CFD simulation of hydrodynamics of a 150 MWe circulating fluidized bed boiler. Chem Eng J 2010;162:821–8. doi:10.1016/j.cej.2010.06.033.

[4]           Valin S, Ravel S, Guillaudeau J, Thiery S. Comprehensive study of the influence of total pressure on products yields in fluidized bed gasification of wood sawdust. Fuel Process Technol 2010;91:1222–8. doi:10.1016/j.fuproc.2010.04.001.

[5]           Ammendola P, Chirone R, Ruoppolo G, Scala F. The effect of pelletization on the attrition of wood under fluidized bed combustion and gasification conditions. Proc Combust Inst 2013;34:2735–40. doi:10.1016/j.proci.2012.06.008.