Chemical Engineering Science, 1972, Vol. 27, pp. 39-44. Pergamon Press.

Printed in Great Britain

Scale-up method for power consumption of agitators in the

creeping flow regime
F. RIEGER and V. NOVAK
Department of Chemical Engineering, Fakulta strojni &UT, Technickit 4, Praha 6,
Czechoslovakia

(Received 24 March 197 1)

Abstract-A method enabling the determination of power consumption of industrial agitators in a

non-Newtonian fluid on the basis of measurements on a model device is presented. The applicability
of the method is limited to creeping flows in the vessel. The suggested method was experimentally
confirmed and is suitable mainly for mixing of highly viscous non-Newtonian fluids.

THEORY Dv
pDt= v.r-vp. (6)
IN THIS paperwe shall deal with mixing of purely
viscous incompressible fluids. These fluids are
expected to obey the rheological equation of In order to obtain the velocity profile we must
state solve Eqs. (5) and (6) along with the rheological
Eq. (1) and the following boundary conditions
T=~)A ( 1) (written in cylindrical coordinates)

where the non-Newtonian viscosity T is a func- ve=2qrN, vz=v,.=O on the agitator

tion of the invariants of the rate of deformation
surface (7a)
tensor A. For incompressible fluids the first
invariant I vanishes and we have only the v = v = v = 0
on the inner vessel
dependence on the second and third invariants II r ’ ’
and III surface (7b)

?j = q(II, III) (2) rrz = Arz + r& = Aez + 0, v, = 0

where on the fluid surface (7~)
II=A:A (3) General solution of Eqs. (l), (5)-(7) and de-
termination of the velocity profile in the vessel
III = det A. (4) has not been published even for Newtonian fluid
due to geometrical complexity of agitators.
In order to obtain the velocity field in the vessel
Therefore we shall try only to analyse Eqs. (l),
we should solve a system of partial differential
(5)-(7) with use of the following dimensionless
equations which consists of the equations of
quantities
continuity and motion. Continuity equation for
the incompressible fluid has the form
v* = v/Nd (84
v.v=o. (5) V* = dV (gb)

Neglecting body forces the equation of motion p* = plpN2d2 (8~)

in terms of stress becomes 71*=315i (gd)
39
 F. RIEGER and V. NOW&K

A” = A/N @d Reynolds number in the creeping flow regime

is so small that inertial and pressure forces in the
II* = II/N2 (8f) equation of motion (10) are negligible with
III* = III/N” (8g) respect to viscous forces and Eq. (10) becomes
r* = r/d (W V” . [q* (N211*, N3111*)A*] = 0. (12)
S* = S/d2 (89
From the last equation it follows that the
D
-=-- ID
Dt* NDt’ @j) dimensionless velocity v* in a creeping flow in
the vessel will depend only on dimensionless
position and dimensionless numbers containing
Equation of continuity (5) transforms, making
rheological constants of the mixed fluid and the
use of (8), to the form
agitator speed N.
Having established the velocity field in the
V”.v”=O (9)
vessel we can determine the power consumption
of the agitator. Power consumption of a ditferen-
Inserting (1) for the stress to the Eq. (6) and
tial area of agitator surface dS may be calculated
using (8) we get
as a scalar product of the force acting on this
area and its velocity v. The total power consump-
pNd2 Dv*
- = V” . [r)* (N211*, N3111*)A*] tion of the agitator is then obtained on inte-
r, Dt* grating, over the whole agitator surface, Sy.
_pNdZV*p*_ Doing this we obtain the expression
(10)
r)
P=j&v.dF=&/.r.dS. (13)
Boundary conditions (7) on using (8) become
v: = 2)* Making use of Eq. (8a) and introducing the
vB = 2m*, z = 0 on the agitator
dimensionless area S* defined as (8i), we obtain
surface (1 la)

v; = VQ = v: = 0 on the inner vessel

surface (1 lb) From the discussion outlined above it is clear
that the velocity v* in creeping flows depends
A,*,+ 0, A& + 0, v,* = 0 on the fluid
only on dimensionless coordinates, rheological
surface. (1 lc) constants of mixed fluid and agitator speed. If
we substitute dimensionless variables (8) in
Using Eqs. (9)-(11) we arrive at the con- Eqs. (1) and (2), we obtain the following expres-
clusion that dimensionless velocity v* is a sion
function of dimensionless position, modified
Reynolds number (Nd2p/5j) and dimensionless T = q(N211*, N3111*)NA*. (1%
numbers containing rheological parameters of
mixed fluid and the agitator speed N. From the From the last equation we may conclude that
boundary conditions (11) it is evident that partic- stress in a given position is influenced only by
ular form of this function will depend on the the constants of the rheological equation of the
agitator shape, relative size of the vessel and the fluid and agitator speed N. Integrating Eq.
agitator and rheological behaviour of the mixed (14) we find PIN@ to be dependent only on
fluid. rheological constants of mixed fluid and agitator
40
 Scale-up method for power consumption of agitators

speed N. The form of this function will depend

on rheological behaviour of fluid and geometrical
characteristics of the agitator and the vessel.
From this theoretical analysis we can make an -
important conclusion that for a given fluid the 7
L
quantity. P/Nd3 will depend only on agitator
speed N and the form of relationship

y&i=f(N) (16) 0

will be the same for all geometrically similar

mixed systems. Relation between P/Nd3 and N
is analogical to the relationship between APD/
-
4L and 8Q/?rD3 which is used in connection with
flow of non-Newtonian fluids in tubes [ 1I. - 7’ -
1

APPLICATION
J
0, = I.1 d
i
The Eq. (16) was confirmed experimentally in II =I.6d
a system consisting of a screw agitator with
draught tube (Fig. 1). Two diameters of screws, Fig. 1. Screw agitator with a draught tube.
(60 and 94 mm), were used in the measurements.
Experiments were carried out with a 4 per cent in Fig. 2. The same screw agitators were used
solution of CMC in water, which exhibited for mixing of paint; the rheological behaviour of
pseudoplastic behaviour. Experimental results which can be described with the Bingham model.
of these measurements are shown graphically Results of these experiments are shown in Fig. 3.

N. I/s
Fig. 2. P/N8 vs. N for the pseudoplastic fluid in creeping flow regime. Screw agitator with draught tube,
od=!Mmm,Ad=6Omm.

41
 F. RIEGER and V. NOV6K

N, I/S
Fig. 3. P/N8 vs. N for the Bingham plastic in creeping flow regime. Screw agitator with a draught tube,
od=60mm,Ad=94mm.

Experimental results with both fluids are in very creeping flow does not apply. If the graphical
good agreement with Eq. (16). form of function (16) is such that it is very dilii-
On the basis of preceding analysis we suggest cult to recognize transition from creeping flow
the following method for the determination of regime, (this is the case with mixing of Bingham
power consumption of agitators operating in the plastics), then it is necessary to use the two
creeping flow regime: dimensions of model agitators. If the curve for
(1) Obtain power measurements with a given both dimensions is the same, the condition of
fluid at a given temperature on a geometrically creeping flow is fulfilled.
similar model. (3) Power consumption of industrial agitator
(2) Use experimental results for calculation of may be calculated from values of P/Nd3 corres-
P/Nd3. On the basis of graphical form of relation- ponding to a given agitator speed N. If we intend
ship (16) decide if the creeping flow condition is to apply the above method for power calculation,
satisfied; e.g. for a power-law fluid the graphical the condition of creeping flow in the industrial
form of function P/Nd3 vs. N in logarithmic co- equipment must be satisfied. For this reason the
ordinates is a straight line with slope n. Typical inertial forces have to be negligible. As out-
form of function (16) for standard turbine agitator lined above, the measure of relative magnitude of
in a pseudoplastic fluid is shown in Fig. 4. inertial forces is expressed by the value of the
In the creeping flow regime the relation is Reynolds number. In a power law fluid mixing
expressible by the two straight lines. Below the Reynolds number is, according to Calderbank
agitator speed 1.5 l/set the behaviour of the fluid and Moo Young[2], proportional to the product
can be described with power-law model with flow Nzvnd2. The magnitude of product N2-“dZ for
index n = O-8. Between 1.5 and 7 revolutions per the transition point from creeping region can be
second we can characterize the fluid behaviour determined on the model device. If the transition
with power-law with n = O-63. At a speed speed on model of agitator diameter dM was Nnn,
higher than 7 llsec the power begins to rise due the transition in the industrial device of diameter
to action of inertial forces and the condition of d will be at speed N
42
 Scaleup method for power consumption of agitors

Fig. 4. Typical form of function P/Nd3 vs. N for the pseudoplastic fluid. Standard disc style turbine.

described in this paper may also be used for

(17) more general simple fluids, but we have not
experimental evidence of the applicability of this
CONCLUSION method for highly viscoelastic fluids.
If the assumption of creeping flow for a fluid
in the vessel of a mixing device can be applied, NOTATION
the ratio of power of industrial agitator and d diameter of agitator
model is proportional to the third power of ratio F force
of their diameters at the same agitator speed. flow index of power-law fluid
This rule is applicable to the mixing of all purely ; speed of agitator
viscous fluids without respect to their rheological pressure
behaviour. The above method for the deter- ; power consumption of agitator
mination of agitator power is suitable mainly for S area
low speed agitators such as screw, helical ribbon V velocity
and anchor, because only such types of agitators r, 8, z cylindrical coordinates
are efficient mixing devices in the creeping flow I, II, III invariants of the rate of deformation
regime [3,4]. If the above method is used for tenzor
determination of power, the rheometric measure-
ment is not necessary. This is an important Greek symbols
advantage of the present method, because usually A rate of deformation tenzor
it is difficult to obtain the rheological properties 7 non-Newtonian viscosity
at very low rates of deformation occurring in the 11 characteristic viscosity
vessel. For the estimation of agitator power con- V Hamilton’s operator
sumption it is necessary to carry out measure- p density
ments on the model. The theoretical procedure 7 stress tenzor
REFERENCES
[l] SKELLAND A. H. P., Non-Newtonian Flow and Heat Transfer. Wiley, New York 1967.
[2] CALDERBANK P. H. and MOO-YOUNG M. B., Trans. Instn &em. Engrs 196139 26.

43
 F. RIEGER and V. NOVAK

[3] NOVAK V. and RIEGER F., Trans. Instn hem. Engrs 1969 47 335.
[4] NOVAK V. and RIEGER F., The influence of the asymmetrical position of helical screw agitator on power consumption
and mixing time, III Congress CHISA, MarihkC L&z& 1969.

R&um&Les auteurs presentent une methode permettant la determination de la consommation

d’tnergie d’agitateurs industriels dans un fluide non-Newtonien en s’appuyant sur des mesures faites
sur un appareil de modele reduit. La methode n’est valable que darts le cas d’ecoulements avec agita-
tion lente. La m&rode proposee a ttt? contirmCe experimentalement et elle est valable principalement
pour les melanges de fluides non-Newtoniens t&s visqueux.

Zusammenfassung- Es wird ein Verfahren fiir die Bestimmung des Leistungsverbrauchs technischer
Riihmr in einer nicht-Newtonschen Fltissigkeit auf Grundlage von Messungen an einer Modellvor-
richtung dargelegt. Die Anwendbarkeit des Verfahrens ist auf Kriechstriimungen in dem Behalter
begrenzt. Die vorgeschlagene Methode wurde versuchsmksig bestiitigt und eignet sich in erster Linie
fiir die Mischung hochviskoser nicht-Newtonscher Fliissigkeiten.

44