Artículos de revistas
Comparative 2D and 3D analysis on the hydrodynamics behaviour during biomass gasification in a pilot-scale fluidized bed reactor
Fecha
2019-02-01Registro en:
Renewable Energy, v. 131, p. 713-729.
1879-0682
0960-1481
10.1016/j.renene.2018.07.080
2-s2.0-85053185402
Autor
Polytechnic Institute of Portalegre
Angra dos Reis Campus
University of Aveiro
Universidade Estadual Paulista (Unesp)
Institución
Resumen
2D and 3D simulations were carried out to predict the whole gasification process behaviour in a pilot-scale bubbling fluidized bed reactor. Special concern for the complex hydrodynamics phenomena within the fluidized bed was undertaken. The implemented multiphase Eulerian-Eulerian mathematical model was validated by comparison to experimental gasification runs and fluidization curves gathered from the pilot-scale fluidized bed. Appropriate 2D and 3D computational domains were achieved by applying a mesh sensitivity study. Solids distribution within the fluidized bed, mixing and segregation phenomena and binary mixture heat transfer were comparatively studied for both configurations. 3D simulations showed improved predicting performance with the experimental results. Also, 3D simulations presented improved segregation degree, while 2D simulations showed improved mixing index, alongside with a tendency to underestimate the reactor heat transfer behaviour. Main findings point to a general good agreement with some close resemblances in the solids distribution between the 2D and 3D simulations whenever quantitative values were considered, while in absolute terms larger discrepancies were seen. The bed expansion was misrepresented at higher superficial gas velocities to a great extent by the 2D configuration. Moreover, it was found that higher superficial gas velocity will induce higher differences between both configurations. Lastly, both configurations successfully described the general tendencies, however, 2D simulations are appropriate every time accuracy is not demanding, whereas 3D simulations should be considered for accurate predictions.