(87b) Influence from Particle Size Distributions on the Cfd Simulations and Experiments of Bubbling Fluidized Beds

Fluidized beds have an enormous role in process industry. Good mixing ability and high contact area between the phases are among the important features of the fluidized beds. The efficiency of fluidized beds depends on bubble behaviour in the particle phase. Industrial fluidized beds in common normally use powders with size distributions. The size and size distributions of particles used in the bed may lead to different bubble behaviours. Because of that it is important to study the influence on bubble behaviour from particle size distribution in fluidized beds.

In addition to that it is important to study the influence on the simulations of fluidized beds from particle size distributions. That is because, in modelling of fluidized beds a mean particle diameter is often used, and important information about flow behaviour can therefore be lost.

A series of experiments are performed in order to check the effect from particle size distribution on bubble behaviour. A lab-scale fluidized bed which is approximated as a 2-D fluidized bed by having a depth of 0.025 m with a uniform air distributor is used along with a video camera to record the bubble behaviour in the bed. Several simulations also carried out in order to analyse the influence from the particle size distribution on the simulated results. A 2-D wire frame mesh with the same dimensions for the particle bed is used for the simulations in the commercial CFD software FLUENT 6.3. The model used is based on a multi-fluid Eulerian description of the phases. The drag model developed by Syamlal & O'Brien is used. The particle size distribution is accounted for by including multiple particle phases.

The experiments and simulations are carried out in Telemark University College, Norway. Spherical glass particles with a density of 2485 kg/m3 are considered. The mixture combinations used give a mean particle diameter of 488 µm. The superficial gas velocity is 0.134 m/s in magnitude.

The computational and experimental results are analyzed separately and compared with each other with respect to the volume fraction changes along the bed with time, particle segregation and bubble frequency. The analysis shows that the computational results vary significantly depending on the number of particle phases used and the experimental results are highly dependent on the particle size distribution used. The results from the simulations with three and four particle phases agree well with the experimental results.

The simulations and experiments show that the particle size distributions significantly influence on the bubble behaviour and particle segregation.