zyxwvutsrqp

zyxwvutsrqponmlk
zyxwvuts
zyxwvutsrqp
( E i j k t - CijkJ presented here is equal to the E used by T = absolute temperature

zyxwvu zy
Lu and Wang ( 3 ) . This has the advantage that in the
absence of quaternary data, the corresponding &jkl may
be set to zero to obtain a better approximation than by
setting (Eijkl - C i j k l ) to zero.

NOTATION
x
y
= mole fraction
= activity coefficient

LITERATURE CITED
1. Benedict, M., C. A. Johnson, E. Solomon, and L. C. Rubin,

zyxwvutsrqpon
Trans. Am. Inst. Chem. Engrs., 41, 371-392 (1945).

zyxwvutsrqponmlkjihgfed
zyxwvut
a = interaction coefficient 2. Jordan, D., J. A. Gerster, A. P. Colburn, and K. Wohl,
A = binary two- and three-suffix coefficient Chem. Eng. Progr., 46, 601-613 (December, 1950).
C = ternary coefficient
3. Lu, C.-H., and Y.-L. Wang, Znd. Eng. Chem. Fundamentals,
3, 271-272 (August, 1964).
D = binary four-suffix coefficient 4. Marek, J., Chem. Listy, 47, 739 (1953); Collection Czech.
E = quaternary coefficient Chem. Commun., 1 9 , l ( 1954).
A G= ~ molar
~ excess Gibbs free energy 5. Redlich, O., and A. T. Kister, Znd. Eng. Chem., 40, 345-348
i, i, k, 1, m, n, p = index representing components ( 1948).
n = moles 6. Wohl, K., Trans. Am. Inst. Chem. Eng., 42, 215-249 (April,
P =pressure 1946).
R = gas constant 7. - , Chem. Eng. Progr., 49, 218-219 (April, 1953).

A Generalized Method for Predicting the Minimum

Fluidization Velocity
C. Y. WEN and Y. H. YU
West Virginia University, Morgantown, West Virginio

In a recent article Narsimhan (18) presented a gen- from the experimental data of the present investigation
eralized expression for the minimum fluidization velocity ( 2 3 ) , emf for spherical particles can be shown to vary
by extending the correlation proposed by Leva, Shirai, from 0.36 to 0.46. Different average values of emf have
and Wen (14) into intermediate and turbulent flow re-
ions. Based on a similar approach by employing the
&ed-bed pressure drop equation of Ergun ( 7 ) , an expres-
sion for the minimum fluidization velocity quite different
from that of Narsimhan has been obtained ( 2 3 ) .
It is the purpose of this communication to compare
these two correlations and to examine the validity and
applicability of each.
The generalized expression given by Narsimhan con-
sists of three equations [Equations ( 6 ) , (9), and ( 11) in
his communication (18)1.
The correlation obtained by Wen and Yu ( 2 3 ) can be
represented by
( N R ~=)d ~( 3 ~3 . 7 ) ' + 0.0408 N G -~ 33.7
For nonspherical particles, the particle diameter dp is de-
fined as the equivalent diameter of a spherical particle
with the same volume. As an approximation, the particle
diameter may be calculated from the geometric mean of
the two consecutive sieve openings without introducing
zyxwvu
(1)

V Van Heerden, et a l . ( Z Z )

serious errors ( 2 6 ) . 0 Fancher and Lewis (9)

The major differences between the two correlations are o.ol - Narsimhan's correlation I
the minimum fluidization voidage emf and the shape factor 0.001 aooz 0.004 aoi aoz 03
4%.
1. Narsimhan considered that for spherical dp (in.)
emf has the value of 0.35 and is independent of trrticles
e parti-
cle diameter, provided that the wall effect can be neg- 986 emf
Fig. 1. Correlation of voidage shape factor function .-
lected. From the literature data (16, 20, 2 4 ) , as well as (1 -emf)2

Page 610 A.1.Ch.E. Journal May, 1966

 002 zyxwvutsrqponmlk
zyxwvutsrqponm
’

zyxwvutsr 0
I I I I I / I

zyxwvutsrqponml
zyxwvutsrqp
001
0008
V

0.006

0.004

zyxwvutsrqpon Io
o LIYO, e i 01
S h i m (21)
(151

van Heerden, et 0 1 (221

Brownell and Kot2 (51

zyxwvutsrqponmlkjih
0 Foochar and Lewis ( 9 )

zyxwvutsrqp
Norsrmhan’r Correlation
0001
05 06 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4

(I - Emf)
+s

(1 --Emf)
Fig. 2. Correlation of voidage shape factor function .-
@S

zyxwvutsrqpo
been reported, such as, 0.386 ( 2 1 ) , 0.40 ( 4 ) , or 0.42
( 2 3 ) .Therefore, the value of 0.35 used by Narsimhan for
spheres seems to be too small.
2. For nonspherical particles, Narsimhan considered
that Emf depends on the particle diameter if it is less than
0.02 in. On the other hand, if greater than 0.02 in., he
considered that Emf is independent of particle diameter.
From Figures 1 and 2, Narsimhan’s correlations for
small particles are seen to be valid only for a limited
10

Fig. 4.
Id3
n
:’-
00

I
II o2 10’
10’
I

I I
I

Ndmf
1
10
0
1
I
II ‘
0
‘
0
:
1

A generalized correlation for minimum fluidization velocity.

IIo3
I

I
I
lo“
lo“

number of the experimental data points. Owing to the

lack of literature data, the validity of Narsimhan’s correla-
tion for large particles cannot be tested. Narsimhan used
only seven data points to establish his correlation. 284 points appearing in Figure 4 are computed. The re-
It is believed that if the wall effect can be neglected, sults are given in Table 1. Equation (1) gives an overall
emf should depend only on the shape factor regardless of standard deviation of 34% and an average deviation of
the diameter of the particle. In Figure 3, E m f are plotted & 25% based on 284 points available in the literature RS
vs. +s for spherical as well as nonspherical particles. Al- shown in Figure 4.An overall standard deviation of 46%
though some scattering may be observed, a general trend and an average deviation of -t 34% are obtained from
is seen to exist between E m f and +s. The following two Narsimhan’s correlations based on 267 data points tested.
approximate relations are obtained which cover the d, The difference in the number of data points tested for the
range from 0.002 to 1.97 in., emf from 0.385 to 0.935, ds validity of the two correlations is due to the lack of infor-
from 0.136 to 1.0, and with a particle diameter to column mation pertaining to the shape factor. The proposed cor-
diameter ratio ranging from 0.000807 to 0.25. relation represented by a single relation in Equation (1)
does not require this information and covers the widest
( 1 - ~ m f ) / ( d J s 2 E m f 3 ) = 11 (2) ( N R e ) ,,,f range, from 0,001 to 4000, heretofore attempted.
l/(+scmy3) = 14 (3) The advantages of Equation (1) over the Narsimhan’s
correlations may be summarized as follows: ( 1 ) Equation
In order to compare the validity of Equation (1) and (1) is considerably simpler. ( 2 ) Equation (1) gives
that of Narsimhan’s correlation, the deviations based on greater accuracy. ( 3 ) Equation ( 1 ) may be expressed in
a convenient graphical form such as shown in Figure 4
for a rapid estimation of the minimum fluidization veloc-
ity.
0 Shirai (21) 0 Fan(8a)
v Van Haerden, at al.(22) 0 Brownell and K a t z ( 5 )
m Oman and Watson (19) 0 Fancher and Lewis (9)

TABLE1. COMPARISON
OF EQUATION
( 1) AND
NARSIMHAN’S
CORRELATIONS

Standard devi- Standard devi-

No. of ations based ations based on
Particle data on Narsimhan’s Equation ( 1),
characteristics points correlations, % %

Spherical 55 43.4 34.1

Nonspherical
_ _ _ _ _ -* 14 d p < 0.02 in. 203 38.6 37.6
dp > 0.02 in. 9 135.3 21.3
Overall standard deviation: 46” 34f
* Calculation based on 267 data points.
t Calculation based on all the 284 data points.

Vol. 12, No. 3 A.1.Ch.E. Journal Page 61 1

 zyxwvutsrqpon
zyxwvutsr
zyxwvutsr
zyxwvutsrqpon
ACKNOWLEDGMENT 6. Brownell, L. E., H. S. Dombrowski, and C. A. Dickey,
ibid., 46,415 (1950).

zyxwvutsrq
zyxwvut
This work was supported in part by the National Science 7 . Ergun, S., and A. A. Oming, lnd. Eng. C b m . , 41, 1179
Foundation Grant GP-560. (1949).

zyxwv
8. Fan, L. T., and C . J. Swartz, Can. J . Chem. Eng., 37, 204
(1959).
NOTATION 8a. Fan, L. T., Kansas State Univ. BuU., 43, 4 (1959).
‘tp L 9. Fancher, G. H., and J. A. Lewis, lnd. Eng. Chem., 25,

zyxwv
= particle diameter,
1139 (1933).
f( ) = function of ( ) 10. Fetterman, C. P., M.S. thesis, Univ. Washington, Seattle,
g = acceleration due to gravity, L/02 (1958).
G,f = minimum fluidization flow rate, M/L20 11. Furukawa, J., T. Ohmae, and I. Ueki, Chem. High Poly-
Nca = Galileo number = 4 3 pf(ps - pf)g/pz mers, Tokyo, 8(2), 111 (1951).
N R e = particle Reynolds number = ( 4p f V / p ) 12. Johanson, L. N., Unclassified Rept. HW-52891, General
( N R e ) m f = particle Reynolds number at onset of fluidiza- Electric Co. (September, 1957).
tion 13. Kelly, V. P., M.S. thesis, Univ. Idaho, Moscow (1958).
N R e t = particle Reynolds number at terminal falling 14. Leva, Max, T. Shirai, and C. Y. Wen, Genie Chim., 75(2),
velocity 33 (1956).
V 15. Leva, Max, M. Weintraub, M. Grummer, M. Pollchik, and
= superficial fluid velocity, L/0
H. H. Storch, U . S . Bureau Mines Bull. 504 (1951).
pf = fluid density, M/L3 16. Lewis, W. K., E. R. Gilliland, and W. C. Bauer, Ind. Eng.
ps = particle density, M/L3 Chem., 41, 1104 (1949).
p = fluid viscosity, M/LB 17. Miller, C . O., and A. K. Logwinuk, ibid., 43, 1220 (1951)
emf = minimum fluidization voidage 18. Narsimhan, G., A.1.Ch.E. J., 11, No. 3, 550 (1965).
4s = 19. Oman, A. O., and K. M. Watson, Natl. Petrol. News, 36,
surface area of sphere having the same volume as particle R795 (1944).
20. Shannon, P. T., Ph.D. thesis, Illinois Inst. Technol., Chi-
surface area of particle cago ( 1961).

zyxwvutsrqp
’

21. Shirai, T., “Fixed Bed, Fluidized Bed, and the Fluid Re-
sistance. Heat Transfer and Mass Transfer of a Single
LITERATURE CITED Particle: Res. Rept., Tokyo Inst. Technol., Japan ( 195z).
22. Van Heerden, C., A. P. P. Nobel, and D. W. Van Krevelen.
1. Baerg, A.. 1. Klassen, and P. E. Gisher, Can. J. Res., F28, Chem. Eng. Sci., 1, No. 1, 37 ( 1951).
287 [igso j. 23. Wen, C. Y., and Y. H. Yu, Chem. Eng. Progr. Symposium
2. Berl., E., Catalog of Ditt and Frees, Wiesbaden (1930). Ser. No. 62, 62 ( 1966).
3. Blake, F. C., Trans. Am. Inst. Chem. Engrs., 14, 415
( 1922). 201 (1948).
-
24. Wilhelm, R. H., and M. Kwauk, Chem. Eng. Progr., 44.
Y

4. Bransom, S. H., and S. Pendse, lnd. Eng. Chem., 53, 575 25. Yagi, S., I. Muchi, and T. Aochi, Chem. Eng. (Japan),
(1961). 16.307 (1952).
5. Brownell, L. E., and D. L. Katz, Chem. Eng. Progr., 43, 26. Yn‘, Y. H., M.S. thesis, West Virginia Univ., Morgantown
537 (1947). ( 1965).

Scale-Up of Residence Time Distributions

D. S. AMBWANI and R. J. ADLER
Case Institute of Technology, Cleveland Ohio

Residence time distributions are valuable for under- cases of practical interest, it is necessary to determine the
standing the performance of many continuous flow sys- distributions experimentally. This study investigates the
tems and for formulating mathematical models of such possibility of determining residence time distributions of
systems. The residence time distribution Y (X) referred process systems from experimental tests performed on suit-
to in this communication is defmed as: Y(X)dX is the ably scaled laboratory models.
fraction of the inflowing (outflowing) stream which will
spend (has spent) a time between X and X +dX in the CRITERIA FOR MODELS
system. For ideal cases such as plug flow, perfectly stirred
vessels, laminar flow, etc., the residence time distribu- As early as 1953 the conditions under which residence
tions may be obtained analytically. However, in many time distributions of large systems can be predicted from

Page 612 A.1.Ch.E. Journal May, 1966