Comparison between finite volume and lattice-Boltzmann method simulations of gas-fluidised beds: bed expansion and particle–fluid interaction force

Effect of granular temperature and solid concentration fluctuation on the gas-solid drag force: A CFD test

“…For a gas velocity of 0.3 m/s, the values of H calculated using equations (7) and (10) are 15.8 mm and 15.5 mm, respectively, both of which are in a good agreement with the DNS result (H=15.7 mm); For a gas velocity of 0.24 m/s, the values of H calculated using equations (7) and (10) are 14.6 mm and 13.2 mm, respectively, the former is very close to the DNS result (H=14.4 mm), whereas the latter is much less than the DNS result. However, it should be noted that (i) Third et al (Third et al, 2016) did not report the values of e and e w , the values used here are only reasonable guesses. Furthermore, there is a particle size distribution of 5% in the DNS of Third et al (Third et al, 2016), but it is monodisperse in present study.…”

“…However, it should be noted that (i) Third et al (Third et al, 2016) did not report the values of e and e w , the values used here are only reasonable guesses. Furthermore, there is a particle size distribution of 5% in the DNS of Third et al (Third et al, 2016), but it is monodisperse in present study. Therefore, the comparison between the results of present study with Third et al (Third et al, 2016) is less rigorous than the comparison with the one of Kriebitzsch et al ; (ii) After simulating the case by Kriebitzsch et al (Kriebitzsch et al, 2013b), we found that the model of Wylie et al (Wylie et al, 2003) has essentially no effect on the simulation result and the underlying reason is that the granular temperature only has a very minor effect on the interphase drag force for the studied range of granular temperature and voidage.…”

“…Since the model has been extensively used, we do not present the details, which can be found in the documentation of ANSYS Fluent. It is interesting to note that previous studies have compared the DNS results with CFD-DEM results (Kriebitzsch et al, 2013aThird et al, 2016), however, present study has selected the two-fluid model as the working platform, this is because the study of Kriebitzsch et al(Kriebitzsch et al, 2013a) have shown that how to implement the effects of granular temperature into the CFD-DEM simulation of gas-solid flow is still not clear. However, there is no any ambiguity when implementing the effects of granular temperature and/or solid concentration fluctuation into the two-fluid model.…”

“…We note that (i) Kriebitzsch et al (2013b) reported the main difference between the results obtained from both DNS and CFD-DEM method via the bed expansion characteristics, we therefore reported the CFD results of H only; (ii) due to the restitution coefficient used is unity, which means there is no energy dissipation during particle-particle and particle-wall collisions, therefore, there is no bubbling structure formed (the results are not shown here for brevity), although the superficial gas velocity is roughly two times of the minimum fluidization velocity, this is in agreement with the conclusion made from both DNS and CFD-DEM method ; (iii) The value of H is calculated as in present study, where is the number of time step used to calculated the mean value ( , meaning 10 seconds have been used), is the total number of particles used in simulations, N is the number of computational cells, is the volume of cell i, is the volume of particle, is the height of the center of cell i and is the solid concentration. Further simulations have been carried out to compare with the DNS results of Third et al (Third et al, 2016), the relevant parameters are summarized in table 2. For a gas velocity of 0.3 m/s, the values of H calculated using equations (7) and (10) are 15.8 mm and 15.5 mm, respectively, both of which are in a good agreement with the DNS result (H=15.7 mm); For a gas velocity of 0.24 m/s, the values of H calculated using equations (7) and (10) are 14.6 mm and 13.2 mm, respectively, the former is very close to the DNS result (H=14.4 mm), whereas the latter is much less than the DNS result.…”

“…One of the conclusions is that modification of currently-used drag correlations is needed, to consider the effects of the fluctuation of state variables (Benyahia, 2008;Kriebitzsch, 2011;Kriebitzsch et al, 2013aKriebitzsch et al, , 2013bTang, 2015;Tang et al, 2016b). Models or correlations have also been proposed (Wylie et al, 2003;Zhang and Reese, 2003;Tang et al, 2016b;Third et al, 2016), however, no CFD test has been carried out except that Zhang and Reese has tested the effect of granular temperature on the hydrodynamics of circulating fluidized beds (Zhang and Reese, 2003). To this end, CFD simulations are performed to study the effects of solid concentration fluctuation and granular temperature on the interphase drag force and the resultant CFD results, where the results of DNS are assumed to be perfect and are worked as the benchmark data.…”