Effects of Gas Properties on the Behavior of Fluidized Beds of Small Particles

EFFECTS OF GAS PROPERTIES ON THE BEHAVIOR OF FLUIDIZED BEDS OF SMALL PARTICLES

Takami KAI and Shintaro FURUSAKI Department of Chemical Engineering, University of Tokyo, Tokyo 113

Key Words: Fluidization, Fluidized Bed, Fluidity, Bubble Frequency, Bubble Size, Pressure Fluctuations, Gas Viscosity, Gas Density The quality of fluidization of FCCparticles (dp =60/im) was studied by using six kinds of fluidizing gases in a 0.081 mdiameter column. As several investigators have described, minimumfluidization velocity was inversely proportional to gas viscosity and was almost independent of gas density. The expansion ratio of the emulsion phase was, however, dependent on gas density and increased with viscosity and density. Both measured bubble frequency and deviation of pressure fluctuations were also influenced considerably by the gas properties. The voidage of the emulsion phase was considered to be dominated by the properties of the gas flowing through the emulsion phase. Thus the apparent viscosity of the emulsion, i.e. the quality of fluidization, is influenced by those properties of gases.

properties. Introduction Several investigators have studied the bed behavior The behavior of fluidized beds at elevated tempera- of heterogeneous fluidization at elevated temperatures or pressures differs from that under ambient tures or pressures. It was found that bubble size conditions. Hong et al.l2) reported that the particle decreased with increasing temperature9'13'27'29?32) and adhesion forces increased with temperature and af- pressure.19'28) From these results gas density and fected the mean voidage of the bed at minimum viscosity are considered to affect the quality of fluidization. Piepers et al.19) reported that the in- fluidization. creasing elasticity moduluswith pressure explained In the present study, the behavior of fluidized beds the increasing bed expansion with pressure. In addi- of small powders was investigated by changing gas tion to the above causes, one of the main factors properties. For this purpose we measured Umf, the affecting fluidity seems to be a change in properties of expansion ratio of the emulsion phase, bubble fre- fluidizing gases. quency and pressure fluctuations. Six kinds of gases The relationship between minimum fluidization were used to vary the properties of fluidizing gas at velocity and gas properties has been studied exten- ambient pressure. The pressure fluctuation and bub- sively. At low values of the Reynolds number mini- ble signals were measured at ambient and elevated mumfluidization velocity Umf is almost indepen- temperatures. dent of gas density and varies inversely with vis- 1. Experimental cosity.5'7'1^ Therefore, Umf decreases with increasing temperature6'26'29) and is unaffected by the change of 1.1 Materials pressure.3'14'19'29) On the other hand, minimum bub- The solid used in the present study was a cracking bling velocity Umbchanges with gas density and catalyst (FCC) which is characterized as the group A' viscosity.140'14'23* Geldart and Abrahamsen10) have powder in Ikeda's criterion16) for good fluidization. correlated Umb with the properties of fluidizing gas The properties of the dry particles are given in Table and particles. They used the ratio Umb/Umf as a 1. parameter to judge the hydrodynamic behavior of Since fluidized beds of dry particles were affected by small powders. According to their results it is possible electrostatic charges at room temperature, FCCpar- that even powders which behave as group A8) pow- ticles containing a small amountof water in the pores ders when fluidized by air behave as group B powders were provided for the present experiments. The in- whenfluidized by some other gases, such as hydrogen. fluence of electrification was decreased by this treat- Henceit is expected that the characteristics of bub- ment.13) Also, the influence was found to be small at bling fluidized beds are also influenced by gas elevated temperatures.13) 1.2 Experimental apparatus and measuring system Received June 21, 1985. Correspondence concerning this article should be addressed to S. Furusaki. T. Kai is now at Dept. ofChem. Eng., Kagoshima Univ., Kagoshima Minimumfluidization velocities were measured in a 890. fluidized bed of a 0.051 m i.d. acrylic column. The

VOL. 19 NO. 1 1986 67 Table 1. Physical properties and size distribution of particles dp 60 ,um pb 0.46 x 103 kg/m3 Fines fraction less . _ 0. than 44/^m "

-25/mi 3.1wt% 25- 44 14.8 44- 53 12.6 53- 63 22.3 Size distribution 63- 74 22.5 74- 88 13.3 88-105 6.4 105-149 4.4 149- 0.7

Fig. 1. Viscosity correlation with minimum fluidization settled bed height above the distributor was 0.3m. velocity. Five kinds of gases were used as the fluidizing gases: nitrogen, hydrogen, helium, ethane and carbon could be explained by a dimensionless number Np: dioxide. Pressure fluctuations and bubble frequency were Np = ^gd3p(Pp - P)/fi (1) measured in a stainless steel columnat ambient and The expansion ratio was in inverse proportion to the elevated temperatures. The main part of the column parameter Np. In their experiments, however, the gas was 0.081m i.d. and 1.5m high. The settled bed properties were not changed. height was 0.6m. Fluidizing gases were air and the The relationship between (Le -Lq)/Lq and gas vis- gases noted above. cosity obtained from measurement in this study is Pressure-fluctuation signals were detected by using shown in Fig. 2. The values of the expansion ratio are a differential pressure transducer fitted to a tapping relatively small under conditions using hydrogen and 0.05 m above the distributor. The mean deviation and helium as fluidizing gases. Therefore it is considered dominant frequency of the signal were found from that not only gas viscosity but also gas density calculation. Bubble size was calculated from the influences the expansion of the emulsion phase. bubble frequency measured by an optical probe The expansion ratio in this study was found to be mounted on the central axis of the bed 0.45m above correlated with gas viscosity and density in the form the distributor. Further details on the apparatus of a power-type equation by the method of nonlinear design and measuring system were described in the least squares: previous paper.13) J±lZ^= 58o/V-08 (2) 2. Results 2.1 Minimum fluidization velocity The experimental values of (Le -Lq)/Lq are well cor- Minimum fluidization velocity was determined related by this expression, as can be seen in Fig. 3. from the usual plot of pressure drop against super- Abrahamsen and Geldart2) have also suggested a ficial gas velocity. In a region of low Reynolds relationship to predict the average emulsion-phase number, where viscous forces predominate, Umfis voidage: almost independent of gas density and inversely proportional to.gas viscosity.5'7'15* Figure 1 shows a Le _2.54po7 ~^0.10.118/ol6/i° O66exp(0.090F)\0.U8r0.043 ^ correlation between Umfand gas viscosity. The slope ^mf ap 9 \Pp P) ^mf in the figure is equal to -1. It was confirmed from The difference between Lmf and Lq is negligibly small this figure that Umfwas inversely proportional to gas and so the value of LJLmfis almost equal to that of viscosity. LJLq. The average error expressing our data ofLJLq 2.2 Expansion of emulsion phase using Eq. (3) was 6.6%. However, the parameter The equivalent height of the emulsion phase in the (Le-Lq)/Lq is so sensitive to the change ofLe that Eq. aggregative bed, Le, can be determined by extrapolat- (3) was not satisfactory to predict the value of ing the sedimentation curve to time zero.21) At gas (Le-Lq)/Lq. The predicted values from Eq. (3) are velocities above 0.04m/s, Le was almost independent twice as large as the experimental values. of gas velocity2'17'18'20* and bed diameter.17* Geldart and Abrahamsen10) have suggested that the According to the results of Morooka et al.,m the ratio Umb/Umfof group A powders could be expressed expansion ratio of the emulsion phase (Le-Lq)/Lq by

68 JOURNAL OF CHEMICAL ENGINEERING OF JAPAN Fig. 2. Expansion ratio of emulsion phase.

Fig. 4. Deviation of amplitude of pressure fluctuations at 280K.

Fig. 3. Comparison between experimental and predicted values of the expansion ratio of emulsion phase.

Fig. 5. Deviation of amplitude of pressure fluctuations at umf {pP-p)gdp 640K. They showed that the ratio Umb/Umf is a useful criterion for judging the hydrodynamic behavior of small powders. It is interesting that the dependence of (Le -Lq)/Lq on gas properties is just similr to that of Umb/Umf, while the ranges of gas velocities were quite different: Umbwas less than 0.01 m/s and Le was measured at velocities of more than 0.05 m/s. 2.3 Pressure fluctuations and bubble size Figures 4 and 5 show the relationship between the gas velocity and the average deviation of the amplitude of the pressure fluctuations for six types of gases at 280 and 640K, respectively. It is seen that the deviation depended not only on temperature but also on the kind of fluidizing gas. In both cases the Fig. 6. Variation of bubble size for six types of fluidizing deviations were very large whenbeds were fluidized gas at 280K. by hydrogen. The deviation for ethylene at 280 K was relatively small, but this did not correspond certainly to the good quality of fluidization. The fluidity might bubble size was largest and the fluidization quality be influenced by the gas adsorption of the particles. was worst when beds were fluidized by hydrogen. The Figures 6 and 7 show the relationship between gas bubble size for ethylene was also relatively large. The velocity and bubble size for six types of gases at 280 viscosity was higher for helium than nitrogen, but the and 640K, respectively. The plots show a similar bubble size was larger for helium. This maybe caused tendency to that of the deviation of amplitude. The by the lower value of the density of helium.

VOL 19 NO. 1 1986 69 Fig. 7. Variation of bubble size for six types of fluidizing gas at 640K.

Fig. 8. Relationship between the uniformity index and 3. Discussion parameter based on expansion ratio of emulsion phase. In bubble columns, mean bubble size increases with liquid viscosity under constant aeration. In fluidized beds, it has been stated that the stable bubble size increased with the apparent viscosity of a bed.n'26) Clift et al.4) suggested that the bubble stability depended strongly on the effective viscosity of the emulsion phase and that the properties of fluidizing gas had negligible influence for gas-fluidized beds. Pritchett et al.21) estimated the effective viscosity of a fluidized bed based on the particle-particle slippage and concluded that the effective viscosity decreased with increasing bed expansion. Taking account of the results of the above studies and the present work, wecan suppose that the bed behavior is influenced by the properties of fluidizing gas which flows through the emulsion phase. The Fig. 9. Relationship between bubble size and parameter effect of gas flow in the form of bubbles on fluidity is based on expansion ratio of emulsion phase. considered to be small. We can conclude that the effective viscosity of the emulsion phase, which is quality. associated with the expansion of the emulsion, is Figure 10 shows the relationship between uniform- influenced by the kind of fluidizing gas. ity index and bubble size. In our recent study13) we We will now correlate bubble size and pressure investigated the bed behavior by using four types of fluctuations with a parameter, jj°-8p°-08/pb, based on powders and air as the fluidizing gas. The results are the expansion ratio of the emulsion phase given by compared with those of the present study in Fig. 10. Eq. (2). Pressure fluctuations were expressed as the According to previous reports,16'30'31) FCC particles uniformity index1 3): such as those used in this study have optimal physical properties for good fluidization. Indeed, the quality of '^^f^ x lOO (5) fluidization was very good whena bed was fluidized by air and nitrogen, but the quality decreased when The results of the correlations are shown in Figs. 8 fluidized by hydrogen. Figure 10 shows that the bed and 9. Both the uniformity index and bubble size behavior, when the FCCparticles were fluidized by generally decreased when the fluidization quality was hydrogen, was similar to that when silica particles improved. Therefore, it is seen that quality increased (dp= \3$ jj.m) were fluidized by air. with the parameter fj°-8p°-08/pb and that the influence of gas velocity on quality was relatively small when C onclusion the quality was good. The behavior of fluidized beds The effects of gas properties on the behavior of the might be affected by various causes. It is considered fluidized bed of small powders were examined. that a change in the properties of fluidizing gases is Minimumfluidization velocity of the FCCparticles of also one of the main factors affecting fluidization 60fim diameter was independent of gas density and 70 JOURNAL OF CHEMICAL ENGINEERING OF JAPAN Literature Cited 1) Abrahamsen, A. R. and D. Geldart: Powder TechnoL, 26, 35 (1980). 2) Abrahamsen, A. R. and D. Geldart: Powder TechnoL, 26, 47 (1980). 3) Chitester, D. C, R. M. Kornosky, L. S. FanandJ. P. Danko: Chem. Eng. ScL, 39, 253 (1984). 4) Clift, R., J. R. Grace and M. E. Weber: Ind. Eng. Chem. Fundam., 13, 45 (1974). 5) Davies, L. and J. W. Richardson: Trans. Inst. Chem. Eng., 44, 293 (1966). 6) Doheim, M. A. and C. N. Collinge: Powder TechnoL, 21, 289 (1978). 7) Franz, J. F.: Chem. Eng. Prog. Symp. Ser., 62, 100 (1966). 8) Geldart, D.: Powder TechnoL, 7, 285 (1973). Fig. 10. Comparison of the fluidization quality with pro- 9) Geldart, D. and D. S. Kapoor: Chem. Eng. ScL, 31, 842 perties of particles and fluidizing gas. (1976). 10) Geldart, D. and A. R. Abrahamsen: Powder TechnoL, 19, 133 inversely proportional to gas viscosity. However, the (1978). expansion ratio of the emulsion phase under fully ll) Grace, J. R.: Can. J. Chem. Eng., 48, 30 (1970). 12) Hong, G. H., R. Yamazaki, T. Takahashi and G. Jimbo: fluidized conditions was influenced by the gas vis- Kagaku Kogaku Ronbunshu, 6, 557 (1980). cosity and density. Bubble size and the deviation of 13) Kai, T. and S. Furusaki: J. Chem. Eng. Japan, 18, 113 (1985). amplitude of pressure fluctuations increased with 14) King, D. F. and D. Harrison: Trans. Inst. Chem. Eng., 60, 26 these gas properties. The difference of fluidization (1982). 15) Kunii, D. and O. Levenspiel: "Fluidization Engineering,'' quality was caused by the change of expansion ratio Wiley, New York (1969). of the emulsion which was influenced by the gas 16) Miyauchi, T., S. Furusaki, S. Morooka and Y. Ikeda: properties. When a bed was fluidized by hydrogen, "Advances in Chem. Eng.," Vol. ll, p. 275, Academic Press, which had relatively small density and viscosity, the New York (1981). quality of fluidization was worse than that of the bed 17) Morooka, S., Y. Kato and T. Miyauchi: J. Chem. Eng. Japan, fluidized by air. Therefore, we must take account of 5, 161 (1972). 18) Morooka, S., M. Nishinaka and Y. Kato: Kagaku Kogaku, not only the particle properties but also the gas 37, 485 (1973). properties to predict bubble size and quality of fluid- 19) Piepers, H. W., E. J. E. Cottaar, A. H. M. Verkooijenand K. ization for fluidized beds of small powders. Rietema: Powder TechnoL, 37, 55 (1984). 20) Pritchett, J. W., T. R. Blake and S. K. Garg: AIChE Symp. Nomenclature Ser., 74, 134 (1978). 21) Rietema, K.: "Proc. Int. Symp. Fluidization," Eindhoven, p. db = average bubble size [m] 154 (1967). dp = average particle diameter [m] 22) Rowe, P. N., L. Santoro andJ. G. Yates: Chem. Eng. ScL, 33, F = fines fraction less than 45jim [-] 133 (1978). fp = dominant frequency of pressure fluctuations 23) Rowe, P. N., P. U. Foscolo, A. C. HoffmanandJ. G. Yates: [1/s] Chem. Eng. ScL, 37, 1115 (1982). g = gravitational acceleration [m/s2] 24) Singh, B., G. R. Rigby and T. G. Callaott: Trans. Inst. Chem. Ip = uniformity index [s] Eng., 51, 93 (1973). Le = equivalent height of emulsion phase [m] 25) Sobreiro, L. E. L. andJ. L. F. Monteiro: Powder TechnoL, 33, Lj - height of fluidized bed [m] 95 (1982). Lmf = height of bed at minimum fluidization velocity 26) Stewart, P. S. B.: Trans. Inst. Chem. Eng., 46, 60 (1968). [m] 27) Stubington, J. F., D. Barrett and G. Lowry: Chem. Eng. Res. Lq = height of settled bed [m] Des., 62, 173 (1984). T = bed temperature [K] 28) Subzwari, M. P., R. Clift and D. L. Pyle: "Fluidization," p. Vj = superficial gas velocity [m/s] 50, Cambridge Univ. Press (1978). Umb = superficial gas velocity at minimum bubbling 29) Tone, S., H. Seko, H. MaruyamaandT. Otake: /. Chem. Eng. [m/s] Japan, 1, 44 (1974). Umf = superficial gas velocity at minimumfluidization 30) Tsutsui, T. and T. Miyauchi: Kagaku Kogaku Ronbunshu, 5, [m/s] 40 (1979). 31) Yasui, T., T. Tsutsui and T. Miyauchi: Kagaku Kogaku dp = average deviation of pressure fluctuations [-] Ronbunshu, 10, 252 (1984). fi = viscosity of gas [Paà"s] 32) Yoshida, K., S. Fujii and D. Kunii: "Fluidization p = density of gas [kg/m3] Technology," Vol. 1, p. 43, Hemisphere Pub. Corp., pb = density of settled bed [kg/m3] Washington (1976). pp = density of particle [kg/m3]

VOL. 19 NO. 1 1986 71