Chinese Journal of Chemical Engineering, 16(5) 686692 (2008)

Calculation of Metzner Constant for Double Helical Ribbon Impeller

by Computational Fluid Dynamic Method*

ZHANG Minge ()1, ZHANG Lhong ()1,**, JIANG Bin ()1,2, YIN Yuguo
()1 and LI Xingang ()1,2
1
School of Chemical Engineering and Technology, Tianjin University, Tianjin 300072, China
2
National Engineering Research Centre of Distillation Technology, Tianjin 300072, China

Abstract Using the multiple reference frames (MRF) impeller method, the three-dimensional non-Newtonian
flow field generated by a double helical ribbon (DHR) impeller has been simulated. The velocity field calculated by
the numerical simulation was similar to the previous studies and the power constant agreed well with the experi-
mental data. Three computational fluid dynamic (CFD) methods, labeled I, II and III, were used to compute the
Metzner constant ks. The results showed that the calculated value from the slop method (method I) was consistent
with the experimental data. Method II, which took the maximal circumference-average shear rate around the impel-
ler as the effective shear rate to compute ks, also showed good agreement with the experiment. However, both
methods suffer from the complexity of calculation procedures. A new method (method III) was devised in this paper
to use the area-weighted average viscosity around the impeller as the effective viscosity for calculating ks. Method
III showed both good accuracy and ease of use.
Keywords computational fluid dynamic, double helical ribbon impeller, non-Newtonian fluid, Metzner constant

1 INTRODUCTION where ks, originally noted to be a proportional coeffi-

The mixing of liquids by mechanical agitation is
one of the most commonly used operations in chemi-
cal and food industries as well as in polymerization
applications [1, 2]. Of all close clearance impellers, the
helical ribbon impeller, due to its high mixing effi-
ciency [3] and its unique feature of producing axial
flow in a stirred tank [4], is widely chosen in mixing
highly viscous Newtonian fluids and non-Newtonian
fluids. In industrial applications, power consumption
is one of the most important factors concerning de-
signers [5-8]. For Newtonian fluids, theoretical and
experimental studies [9, 10] have shown that the power
number Np is inversely proportional to the Reynolds
number Re in the laminar flow:
N p Re = K p
where Re is defined as
Nd 2
Re =
Given a certain type of impeller, Re can be cal-
culated based on the fluid properties and the rotation
speed of the impeller, then Np can be determined from
the power curve. However, for non-Newtonian fluids,
fluid viscosity varies along with shear rate. Metzner
and Otto [11] defined Rea based on apparent (or effec-
tive) viscosity a:
Nd 2
Rea =
a as the slop method. The ks value can be obtained from
and assumed that the effective shear rate eff is pro-
portional to the rotation speed of the impeller:
eff = ks N (4)
cient, is referred to as the Metzner constant. For shear
thinning non-Newtonian fluids, the apparent viscosity
a can be expressed by a power law model:
n 1
a = K psu eff (5)
The power curve of non-Newtonian fluids ob-
tained using the Metzner-Otto method is consistent
with that of Newtonian fluids, which explains the fact
that the Metzner-Otto method is a widely used method
to design impellers for non-Newtonian fluids applica-
tions [12, 13]. As a result, the ks value has become a
key factor to predict power consumption.
ks is a function of impeller geometry. Its value is
normally obtained experimentally using the Metzner
method and Rieger-Novak method [10, 14, 15]. The
Metzner method defines the Ren and Kpn by
(1)
d 2 N 2 n
Ren = (6)
K psu
(2)
K pn = N p Ren = K p ksn 1 (7)
where Kpn is a function of n [8], so that
1
K pn n 1
ks = (8)
K p
The ks value can be directly calculated from Eq.
(8), so this method is also referred to as direct calcula-
tion of ks.
(3) The Rieger-Novak method is denoted sometimes
the slope of the straight line resulting from the plot of
lnKpn versus ( 1 n ), based on the linearized Eq. (8):
ln K pn = ln K p (1 n ) ln ks (9)

Received 2007-11-11, accepted 2008-07-10.

* Supported by the Natural Science Foundation of Tianjin (07JCZDJC02600).
** To whom correspondence should be addressed. E-mail: zhanglvh@tju.edu.cn
 Chin. J. Chem. Eng., Vol. 16, No. 5, October 2008
The two aforementioned methods are both based
on the Metzner concept and can obtain ks through
measuring the power input of mixing in a stirred tank.
From Eqs. (4) and (5), one can notice that the ks value
can be acquired directly and easily using the calcu-
lated effective shear rate or the effective viscosity in a
mixing process. However, no explicit methods for the
calculation of effective shear rate or effective viscosity
have been given since Metzner and Otto brought about
the concept of effective shear rate in 1957 [11]. Fortu-
nately, with the development of computer technology
and computational fluid dynamic (CFD) method, it is
feasible now to simulate the detailed mixing flow
fields and determine the various performance parame-
ters [16-19]. Shekhar and Jayanti [20] took the circum-
ference-averaging shear rate at mid-height of the DHR
impeller calculated by CFD as the effective shear rate
to determine ks, however, the calculating steps of the
method were complex. The calculating procedure of
the slop method (experimental method) was also te-
dious. So a cost-effective and time-saving method to
predict ks through the calculation of effective viscosity
is proposed and compared with the above two meth-
ods in this work.
2 MATHEMATIC MODELS
2.1 Flow equations
In this research, the flow field generated by a
double helical ribbon (DHR) impeller was studied.
Assuming that the fluids were continuous and incom-
pressible non-Newtonian fluids and the mixing proc-
ess was carried out under a constant temperature.
Numerical simulations were run in a steady state cal-
culation mode. The mass and momentum conservation
equations are:
v = 0 (10)
1 a 2
(v )v = f P + v (11)
The governing equations were solved using the com-
mercial flow solver Fluent 6.2 (ANSYS Inc.). The
finite volume method and the second-order modified
scheme were applied to discrete the control equations
to algebra equations.
2.2 Rheological model
For shear-thinning non-Newtonian fluids with no
or negligible elasticity, some researchers [21-23] used
the power law equation as the rheological model.
Convenient for engineering calculations with suffi-
cient accuracy, the power law equation is simple and
easy to use. The power law model is given by
n 1
a = K psu  n 1 = K psu 0.5tr( )2 (12)
where 0.5tr( )2 is the second shear rate tensor in-
variant.
In this study, several fictitious power law fluids
were investigated according to Kelly and Gigas [24].
687
The consistency factor Kpsu was given as 10 Pasn, and
the flow behavior index n of the non-Newtonian fluids
was set between 0.52 1 .
3 NUMERICAL METHOD
3.1 Calculation model
It has been difficult to simulate the flow field in a
stirred tank because of the complex interactions be-
tween moving blades and stationary tank wall and
baffle plates. Two methods, the multiple reference
frames method (MRF) and the sliding mesh method,
have been developed to solve the problem without
requiring any empirical data [25]. The MRF method
can be applied when the relative position between the
impeller and baffle does not significantly affect the
flow field due to weak interactions, whereas the slid-
ing mesh method can be used in the case of strong
interactions. The sliding mesh model requires more
time for calculations [26].
No baffle was used in the studied tank in this pa-
per. Therefore, the MRF model was chosen to simu-
late the flow field generated by DHR impeller in the
stirred tank.
3.2 Mixer configuration and mesh generation
As shown in Fig. 1, the mixer consisted of a
mixing tank of 101 mm in both diameter (D) and
height (H). The tank was equipped with a DHR im-
peller with both a diameter and height of 96 mm.
Other important dimensional parameters of the impel-
ler are shown in Table 1. The impeller was located 2.5
mm above the bottom of the vessel. In order to use the
MRF model, the calculation area is segmented into
two zones, impeller area (Zone 1) and tank wall area
(Zone 2). Steady-state calculations were performed
with a rotating reference frame in Zone 1 and a sta-
tionary reference frame in Zone 2. The unstructured
tetrahedral mesh was generated in impeller region and
structured hexahedral mesh in tank wall region. Adaptive
Figure 1 The structural sketch of stirred tank
Table 1 The dimensional parameters of DHR impeller
d/m s:d L:d w:d c:d
0.096 1.0 1.0 0.1 0.025
688 Chin. J. Chem. Eng., Vol. 16, No. 5, October 2008

(a) X = 0 (b) Z = 50 mm (c) Near the impeller tip

Figure 2 Illustration of a computational mesh at the X0 plane of the stirred tank, at the Z50 mm of the stirred tank and
near the impeller tip

mesh refinement had been applied to the edge of the
impeller where the higher shear strain rate was pro-
duced. Grid independence was verified by demon-
strating that additional requirement on mesh size near
the impeller surface did not change the calculated power
number by more than 2% (Table 2). From Table 2, it
can be seen that the maximal mesh size of 1 mm nor-
mal to the impeller surface (Grid 3) is proper. The
generated mesh is shown in Fig. 2, where X, Y, Z are
the three directions of the Cartesian coordinate system
which make the bottom center of the stirring shaft as
the coordinate origin. Interpolation was used at the
interface of two zones as the solutions progressed.
Table 2 Mesh independence and selection
(Kpsu10 Pasn, n0.8, N2.5 rs 1)
4 RESULTS AND DISCUSSION
4.1 Characteristics of flow field
Flow patterns enhanced by DHR have been stud-
ies by many investigators, such as Bourne and Butler
[27], Carreau et al. [6]. However, Delaplace et al. [4]
did a survey on existing literature and found that, in
spite of the different DHR impeller geometries, the
primary circulation patterns are approximately the
same, namely the liquid between the blades and the
wall flows downwards, inwards along the bottom,
upwards in the core near the shaft and radically out-
wards near the liquid surface of the tank. It showed
contrary flow pattern when the impeller rotated re-
versely. From Fig. 3 (a), we can see the basic flow

Max. mesh size NP defined Relative patterns on the Y = 0 plane of the stirred tank simu-
Total lated in this work is consistent with the conclusion in
Grid normal to impeller in Eq. (14) deviation for
meshes
surface/mm by CFD power number
1 342756 1.5 44.45 -
2 789779 1.2 45.44 2.23%
3 1515566 1.0 44.90 1.19%
4 1984809 0.8 44.36 1.21%

3.3 Boundary conditions

No slip boundary conditions were imposed at the

solid walls of the tank and the impeller, while the free
surface at the top of the vessel was treated as a flat,
shear free boundary. In this study, considering that the (a) Calculated by CFD in this work (Kpsu10 Pasn,
n0.8, N2.5 rs1)
fluid flow in the stirred tank is in the laminar region,
the free surface assumption is realistic. The rotational
speeds of the impeller under the stirring condition
with various fluid properties were reported in Table 3.

Table 3 Description of the fluids and

rotational speed of the impeller
Kpsu/Pasn
1
No. n N/rs Ren
1 10 0.52 1 0.92
2 10 0.6 1.5 1.63
3 10 0.7 2 2.27
4 10 0.8 2.5 2.77 (b) Calculated by Yao et al. [28] (Newtonian fluid,
20 Pas, 1400 kgm3, N0.25 rs1)
5 10 0.9 3 3.09
Figure 3 Velocity vector map in Y0 plane of the stirred
6 10 1 3.5 3.23 tank
 Chin. J. Chem. Eng., Vol. 16, No. 5, October 2008
Ref. [4], and there is a secondary circulation near the
impeller region which is similar to the calculation re-
sults of Yao et al. [28] with different geometrical di-
mensions [Fig. 3(b)].
4.2 Power consumption
To verify the simulation model quantitatively, we
compared experimental results by Rieger and Novak
[10] with simulation predictions on the same experi-
mental equipment.
For the non-Newtonian fluids, the power con-
sumption P and the power number Np of the mixing
process were calculated from Eqs. (13) and (14):
P = 2N (13)
P
Np =
N 3d 5
where can be obtained through data-processing by
Fluent 6.2 after calculation convergence:
= i ( p )i Ai ri (15)
where the summation in Eq. (15) is carried over all the
volumes having the impeller as one of the boundary.
Then the mixing power constant Kpn of the
non-Newtonian fluids can be calculated according to
Eqs. (6) and (7). The mixing power constant calcu-
lated from the CFD was compared with the experi-
mental data of non-Newtonian fluids in the DHR im-
peller system, as shown in Fig. 4. It can be seen that
two sets of data satisfactorily agreed with one another
with an average error of 6.32%, a value acceptable for
engineering design.
Figure 4 Experimental vs. CFD-computed power constant
Kpn for DHR impeller
(Kpsu10 Pasn, N1-3.5 rs1)
CFD-computed; experimental [10]
4.3 CFD solution of ks
Previous research [29] has proven that the
Metzner number ks of weak non-Newtonian fluids
with the flow behavior index greater than 0.4 for heli-
cal ribbon impeller is related to the structure and di-
mensions of the impeller, but that it is independent of
the consistency factor Kpsu and the flow behavior in-
dex n. In this study, ks was a constant given that the
fluids were weak non-Newtonian fluids and the di-
mensions of the helical ribbon impeller used in the
experiment were not changed. Therefore, three
689
post-simulation-processing methods are discussed on
such a basis to determine ks.
4.3.1 Method I-slop method
Wang et al. [30] found that, when ks is independ-
ent of the flow behavior index n, the slope method,
which can avoid exponential operation, can be applied
to predict ks more accurately than direct calculation
method. In this paper, the power consumption was
acquired through CFD data-processing and ks was
predicted through the slope method. The curve of
lnKpn versus (1 n ) for DHR impeller was plotted in
Fig. 5. It was determined by the line slope that the
value of ks was 39.6.
(14)
Figure 5 The relationship between lnKpn and ( 1 n ) of
DHR impeller
(N1-3.5 rs1)
4.3.2 Method II
When conducting numerical simulation to predict
the Metzner number ks of DHR impeller in pseudo
plastic fluid, Shekhar and Jayanti [20] drew the curve
of circumferentially averaged local shear rate versus
the radial profile at the mid-height of the impeller to
find the maximal value under each rotational speed.
Based on this, the curve of the maximal shear rate
value versus rotational speed was plotted. As a result,
the value of ks was obtained through measuring the
slope of the line. In this work, the same method was
adopted to predict ks for analysis. The curve of the
maximal circumferentially-averaged shear rate and
rotational speed was shown in Fig. 6. The line slope,
known as the ks value, was 33.6. The curve of circum-
ferentially averaged shear rate versus radius at each
rotational speed was omitted.
Figure 6 The relationship between maximal circumfer-
ence-averaging shear rate in the middle-height of impeller
and rotational speed (Kpsu10 Pasn, n0.52-1)
It can be seen that two drawing steps were nec-
essary to obtain ks. In addition, the maximal values of
690 Chin. J. Chem. Eng., Vol. 16, No. 5, October 2008
shear rate under each rotational speed need to be read
out from the first curve to plot the second one in this
method.
4.3.3 Method III
Due to the convenience of the Fluent software to
calculate area-weighted average viscosity on the im-
peller, the method to predict ks through calculating
area-weighted average viscosity as the effective vis-
cosity was investigated in this paper. The feasibility of
this method was analyzed and described in the next
paragraph.
After the numerical simulation converges, the
area-weighted average viscosity can be obtained di-
rectly through calculating the facet values of viscosity
on the DHR impeller in Fluent software
1 X0 section in the stirred tank for the non-Newtonian
av = j j | A j | (16)
A
Here, the area-weighted average viscosity was
taken as the effective viscosity. Then the effective
shear rate can be calculated using Eq. (5). The curve
of the effective shear rate versus the rotational speed
was plotted in Fig. 7. The results proved that the ef-
fective shear rate had a linear relationship with the
speed of rotation (R20.996), a relationship that was
assumed by Metzner and Otto [11]. So this method
names the slope value, 34.8, as the value of ks. This
method is clear and concise according to the Metzner
and Otto assumption.
Figure 7 The relationship between effective shear rate and
rotational speed
(Kpsu10 Pasn, n0.52-1)
Simple and clear, this method only needed one
calculation step and one plotting procedure to predict
the value of ks. Table 4 compares the values of ks pre-
dicted by the three aforementioned methods with the
experimental data. It can be seen that the ks values
calculated using method I and method II agreed with
the experimental data with calculation errors of 7.8
Table 4 Comparison of ks values by CFD
solution and experiment
Method ks R2
Rieger and Novak [10] 36.73 1.45 -
method I 39.6 0.990
method II 33.6 0.986
method III 34.8 0.996
and 8.5%, respectively. Yet calculation procedures of
these two methods were quite complicated. The
method III, however, utilized the area-weighted aver-
age viscosity as the effective viscosity in the mixing
process and thus was easy to use. Also, the calculation
error was only 5.4.
In some literature [15, 31], the effective shear rate
and the effective viscosity are also referred to as the
average shear rate and the average viscosity, respec-
tively. Therefore, the effective viscosity and the effec-
tive shear rate, to some extent, are close to the average
viscosity and the average shear rate in the entire
stirred tank during the mixing process. However, ac-
cording to Metzner and Otto [11], the effective shear
rate must also be proportional to the rotational speed.
Fig. 8 illustrates the distribution of viscosity in the
fluid (Kpsu10 Pasn, n0.52) at the rotation speed of
1 rs 1. Results showed that the viscosity in the stirred
tank ranged from 0.39 to 8.96 Pas with only a little
change. Therefore, the area-weighted average viscos-
ity around the impeller, to some extent, can represent
the average viscosity in the whole stirred tank. Also
the effective shear rate calculated from the
area-weighted average viscosity in Eq. (5) was pro-
portional to the rotational speed and thus the
area-weighted average viscosity around the DHR im-
peller can be taken as the effective viscosity. There-
fore, ks predicted by method III was consistent with
the experimental data. Similarly, the maximal circum-
ferential-average shear rate on the DHR impeller is the
average shear rate in the motion region of the DHR
impeller (method II ), so it is not only approximated to
the average shear rate in the whole stirred tank, but
also proportional to the rotational speed. Thus ks pre-
dicted by method II agree with the experimental data,
too. The difference is that the post treatment steps of
method II is more complicated.
Figure 8 Apparent viscosity distribution on X0 section
of the agitated tank for non-Newtonian fluid (Kpsu10 Pasn,
n0.52, N1 rs1)
It seems that Method III might be used to pre-
dicted ks value for any close-clearance impeller in
laminar region based on the analysis in this study,
such as anchors, spirals, etc. However, for the other
impellers or in the turbulent region, more effort is
 Chin. J. Chem. Eng., Vol. 16, No. 5, October 2008 691

needed in order to draw a definitive conclusion on this a apparent(or effective) viscosity, Pas
subject. av area-weighted average viscosity, Pas
j facet viscosity, Pas
fluid density, kgm 3


5 CONCLUSIONS torque, Nm
Subscripts
Using the CFD method, three-dimensional simu- i, j surface element
lation of the flow field generated by a DHR impeller
in non-Newtonian fluids was investigated. The veloc- REFERENCES
ity field calculated from the numerical simulation was
similar to the previous studies. The power constant
Kpn of non-Newtonian fluids with a flow behavior in- 1 Ryan, D.F., Janssen, L.P.B.M., van Dierendonck, L.L., Circulation
dex n which is higher than 0.52 agreed well with the time prediction in the scale-up of polymerization reactors with heli-
cal ribbon agitators, Chem. Eng. Sci., 43 (8), 1961-1966 (1988).
experimental data.
2 Kumaresan, T., Joshi, J.B., Effect of impeller design on the flow
Three methods were used to calculate the
pattern and mixing in stirred tanks, Chem. Eng. J., 115 (3), 173193
Metzner constant for post-simulation data processing.
(2006).
The ks values calculated from the slope method 3 Shervin, C.R., Raughley, D.A., Romaszewski, R.A., Flow visuali-
(method I) and method II agreed fairly well with the zation scaleup studies for the mixing of viscoelastic fluids, Chem.
experimental data. However, the post-data processing Eng. Sci., 46 (11), 2867-2873 (1991).
were more complicated than that of method III. The 4 Delaplace, G., Leuliet, J.C., Relandeau, V., Circulation and mixing
area-weighted average viscosity around the DHR im- times for helical ribbon impellers. Review and Experiments, Exp.
peller, close to the average viscosity in the entire Fluids, 28 (2), 170-182 (2000).
stirred tank in some degree and the shear rate calcu- 5 Niedzielska, A., Kuncewicz, C., Heat transfer and power consump-
lated from power law equation is proportional to the tion for ribbon impellers. Mixing efficiency, Chem. Eng. Sci., 60
rotational speed of the impeller, can be used as the (8/9), 2439-2448 (2005).
effective viscosity in the stirred tank to predict ks 6 Carreau, P.J., Chhabra, R.P., Cheng, J., Effect of rhecological prop-
(method III). This method generated satisfactory erties on power consumption with helical ribbon agitators, AIChE
simulation results and simplified the calculation. The J., 39 (9), 1421-1430 (1993).
calculation error of method III is 5.4%, also lower 7 Yap, C.Y., Patterson, W.I., Carreau, P.J., Mixing with helical ribbon
than that of methods I and II. Therefore, the method agitators (III) Non-Newtonian fluids, AIChE J., 25 (3), 516-521
III can be safely applied for DHR impeller to engi- (1979).
neering calculations. 8 Brito-De La Fuente, E., Choplin, L., Tanguy, P.A., Mixing with
helical ribbon impellers: Effect of highly shear thinning behavior
and impeller geometry, Chem. Eng. Res. Des., 75, 45-52 (1997).
NOMENCLATURE 9 Novak, V., Rieger, F., Homogenization with helical screw agitators,
Trans. Inst. Chem. Eng., 47 (10), 335-340 (1969).
10 Rieger, F., Novak, V., Power consumption of agitators in highly
A surface area of the impeller, m2
viscous non-Newtonian liquids, Trans. Inst. Chem. Eng., 51, 105111
Ai projected surface area of the surface element i, m2
(1973).
c impeller clearance, m
11 Metzner, A.B., Otto, R.E., Agitation of non-Newtonian fluids,
D stirred tank diameter, m
AIChE J., 3 (1), 3-10 (1957).
d impeller diameter, m
12 Sestak, J., Zitny, R., Houska, M., Anchor-agitated systems: Power
body force, ms 2

f
input correlation for pseudoplastic and thixotropic fluids in equilib-
H stirred tank height, m
rium, AIChE J., 32 (1), 155-158 (1986).
i surface element
13 Tanguy, P.A., Lacroix, R., Bertrand, F., Choplin, L., Brito de la
Kp power constant for Newtonian fluids
Fuente, E., Finite element analysis of viscous mixing with a helical
Kpn power constant for non-Newtonian fluids
ribbon-screw impeller, AIChE J., 38 (6), 939-944 (1992).
Kpsu consistency factor, Pasn
14 Castell-Perez, M.E., Steffe, J.F., Evaluating shear rates for power
ks Metzner constant
law fluids in mixer viscometry, J. Text. Stud., 21, 439-453 (1990).
L impeller height, m
15 Thakur, R.K., Vial, C., Djelveh, G., Labbafi, M., Mixing of com-
impeller rotational speed, rs 1

N
plex fluids with flat-bladed impellers: Effect of impeller geometry
Np power number
and highly shear-thinning behavior, Chem. Eng. Process., 43 (10),
n flow behavior index
1211-1222 (2004).
P hydrodynamic press, Pa
16 Wang, Z., Mao, Z.S., Yang, C., Shen, X.Q., Computational fluid
P pressure difference, Pa
dynamics approach to the effect of mixing and draft tube on the pre-
R2 correlation coefficient
cipitation of barium sulfate in a continuous stirred tank, Chin. J.
Re Reynolds number ( Re = Nd 2 / )
Chem. Eng., 14 (6), 713-722 (2006).
Rea apparent Reynolds number ( Rea = Nd 2 / a )
17 Zhang, Y.H., Yang, C., Mao, Z.S., Large eddy simulation of liquid
Ren modified Reynolds number ( Ren = d 2 N 2n / K psu )
flow in a stirred tank with improved inner-outer iteration algorithm,
ri radial coordinate, m
Chin. J. Chem. Eng., 14 (3), 321-329 (2006).
s pitch, m
18 Cui, B., Zhu, Z., Zhang, J., Chen, Y., The flow simulation and ex-
velocity, ms 1

v
perimental study of low-specific-speed high-speed complex cen-
w blade width, m
trifugal impellers, Chin. J. Chem. Eng., 14 (4), 435-441 (2006).
 shear rate, s 1

19 Min, J., Gao, Z.M., Large eddy simulations of mixing time in a

eff effective shear rate, s 1

stirred tank, Chin. J. Chem. Eng., 14 (1), 1-7 (2006).

viscosity, Pas
20 Shekhar, S.M., Jayanti, S., Mixing of pseudoplastic fluids using
692 Chin. J. Chem. Eng., Vol. 16, No. 5, October 2008

helical ribbon impellers, AIChE J., 49 (11), 2768-2772 (2003).
21 Jiang, S., Fan, W., Zhu, C., Ma, Y., Li, H., Bubble formation in
non-Newtonian fluids using laser image measurement system, Chin.
J. Chem. Eng., 15 (4), 611-615 (2007).
22 Iranshahi, A., Heniche, M., Bertrand, F., Tanguy, P.A., Numerical
investigation of the mixing efficiency of the Ekato Paravisc impel-
ler, Chem. Eng. Sci., 61 (8), 2609-2617 (2006).
23 Xu, T., Zhu, Q., Chen, X., Li, W., Equivalent cake filtration model,
Chin. J. Chem. Eng., 16 (2), 214-217 (2008).
24 Kelly, W., Gigas, B., Using CFD to predict the behavior of power
law fluids near axial-flow impellers operating in the transitional flow
regime, Chem. Eng. Sci., 58 (10), 2141-2152 (2003).
25 Deglon, D.A., Meyer, C.J., CFD modelling of stirred tanks: Nu-
merical considerations, Minerals Eng., 19 (10), 1059-1068 (2006).
26 Lane, G.L., Schwarz, M.P., Evans, G.M., Comparison of CFD
methods for modelling of stirred tanks, In: Proceedings of the 10th
Europ. Conf. on Mixing, Delft, the Netherlands, 273-280 (2000).
27 Bourne, J.R., Butler, H., An analysis of the flow produced by heli-
cal ribbon impellers, Trans Inst. Chem. Eng., 47, 11-17 (1969).
28 Yao, W., Mishima, M., Takahashi, K., Numerical investigation on
dispersive mixing characteristics of MAXBLEND and double heli-
cal ribbons, Chem. Eng. J., 84, 565-571 (2001).
29 Delaplace, G., Guerin, R., Leuliet, J.C., Chhabra, R.P., An analyti-
cal model for the prediction of power consumption for shear-thinning
fluids with helical ribbon and helical screw ribbon impellers, Chem.
Eng. Sci., 61 (10), 3250-3259 (2006).
30 Wang, J.J., Feng, L.F., Gu, X.P., Wang, K., Hu, C.H., Power con-
sumption of inner-outer helical ribbon impellers in viscous Newto-
nian and non-Newtonian fluids, Chem. Eng. Sci., 55 (12), 2339-2342
(2000).
31 Montante, G., Mostek, M., Jahoda, M., Magelli, F., CFD simula-
tions and experimental validation of homogenisation curves and
mixing time in stirred Newtonian and pseudoplastic liquids, Chem.
Eng. Sci., 60 (8/9), 2427-2437 (2005).

BOOKS FROM ELSEVIER (www.elsevierdirect.com)

UV Coatings
Basics, Recent Developments and New Applications
By Schwalm
Product Type: Hardcover
Price: $205.00
Subject Area: Chemistry & Chemical Engineering - Chemical Engineering

Since UV curing (light induced polymerisation of multifunctional oligomers) is a very ecoefficient and energy saving curing method,
the growth rates of UV curable coatings are in the range of 10% per year. The typical UV coatings are solvent free (100% solids),
thus helping the industry and the environment to reduce significantly VOC (volatile organic compounds). Recently, the automotive
industry has discovered that UV cured coatings are very scratch resistant, which stimulated very extensive work into the development
of UV coatings for automotive applications. Since UV curing is very universal, also other systems besides the 100% solid (typical)
UV coatings are developed, like waterbased UV- , UV powder and Dual cure (UV and thermal) systems. UV Coatings contains an
overview of the technology, the curing process including the equipment necessary, the raw materials (resins, diluents, photoinitiators)
used, the advantages and drawbacks of this fast emerging technology, as well as proposed technical solutions to tackle the disadvan-
tages. Structure-property relationships will be given, especially regarding the mechanical properties of coatings as well as scratch
resistance, mainly dealing with automotive performance criteria. The main part of the book will deal with new developments, like
water-based UV coatings, UV powder coatings and dual cure systems, cured by UV and thermal energy, which have been developed
to cure the coating on three dimensional substrates in shadow areas. The main applications of UV Coatings will be described, starting
with the classical ones on temperature sensitive substrates, like wood, paper and plastics, where the UV curable coatings are already
well established.