Comparative scaling analysis of two different sized pilot-scale fluidized bed reactors operating with biomass substrates

Cardoso, João; Silva, Valter; Eusébio, Daniela; Brito, Paulo; Hall, M.J.; Tarelho, Luís

http://hdl.handle.net/10400.26/27953

Use this identifier to reference this record.

Name:	Description:	Size:	Format:
energy_scale_up.pdf		5.11 MB	Adobe PDF	Download

Send Feedback

Authors

Abstract(s)

This paper presents a comparative scaling analysis of two different sized pilot-scale fluidized bed reactors operating with biomass substrates. A multiphase Eulerian-Eulerian 2-D mathematical model was implemented, coupled with in-house user-defined functions (UDF) built to enhance hydrodynamics and heat transfer phenomena. The model validation was attained by comparison to experimental data gathered from both reactors. A grid refinement study was carried out for both geometries to achieve an appropriate computational domain. Hydrodynamics was deeply studied for both reactors concerning the scale-up effect. Mixing and segregation phenomena, solid particle distribution and biomass velocity were matters of great concern. Results showed that UDF implementation successfully minimized deviations and increased the model’s predictability. The largest deviations measured between experimental and numerical results for syngas composition were of about 20%. Solids mixing and segregation was found to be directly affected by the particles size, density, and superficial gas velocity, with the larger reactor revealing improved mixing ability. Improved mixing occurred for smaller particles size ratio (dbiomass ¼ 3 mm), smaller particles density ratio (rbiomass ¼ 950 kg/m3), and higher dimensionless superficial gas velocities (U0=Umf¼3.5). The larger unit showed an increase in near-wall velocity, lateral dispersion, and bubble size. As for the smaller reactor, higher velocities were obtained at the center region due to a more pronounced wall boundary layer. Similarities were found between the two reactors regarding the bubble distribution, dimensionless average bed pressure drop and biomass velocity vector profiles when dimensionless parameters were employed.