CFD simulations to study bed characteristics in gas-solid fluidized beds with binary mixtures of Geldart B particles: II quantitative analysis

Hydrodynamics of fluidized beds with binary mixtures of particles is important in many industrial applications. The binary particles are generally in the Geldart particle range. In our earlier work, (Part I) of this work simulations were carried out and qualitative analysis was presented. Quantitative predictions of gas velocity and particle velocity profiles have been presented in the present work, which is Part II of the two-part work on computational fluid dynamics (CFD) simulations of binary fluidized beds. It was observed that the dynamics of the bed vary for different binary mixtures and are a strong function of superficial velocity and bed height. Mixing and segregation in beds for two different initial bed heights and six different binary mixtures and superficial velocities have been identified. Segregation is prominent for binary mixtures with 20 wt.% and 80 wt.% of large particles, whereas mixing is observed in 40 wt.% and 60 wt.% large particle mixtures. Bypassing of gas near the walls is prominently seen for 60 wt.% large particles with gas velocities as high as 5 m/s. Time-averaged axial particle volume fractions have been observed to be lower in the dilute phase with large undulations in the middle whenever the bed is well mixed for central axial profiles. The axial volume fraction profiles also confirm the mixing and segregation for the 40 wt.% and 20 wt.% composition of large particles for the operating conditions considered for the study. Bed height expansion is linear until a certain superficial velocity with the increase or decrease depending on the superficial velocity or bed height of operation. Furthermore, correlations for minimum fluidization velocity and pressure drops from the literature have been compared with experimental results. The simulated data have been considered for the development of a correlation for minimum fluidization velocity. The predicted results match experimental data with a 10%-15% deviation.