Please cite as: Inż. Ap. Chem.

2015, 54, 4, 201-203

Nr 4/2015 INŻYNIERIA I APARATURA CHEMICZNA str. 201

Magdalena STEC1, Piotr Maria SYNOWIEC 1,2

e-mail: [email protected]
1 Katedra Inżynierii Chemicznej i Projektowania Procesowego, Wydział Chemiczny, Politechnika Śląska, Gliwice
2 Instytut Nowych Syntez Chemicznych, Oddział Chemii Nieorganicznej „IChN”, Gliwice

Analysis of the pressure drop calculation method impact on the accuracy of

the experimental results in the Koflo static mixer
Introduction
Nowadays, CFD (Computational Fluid Dynamics) techniques are
modern tools commonly used by engineers as so as manufacturers
and project designers. They allow for the reduction of the designing
time and what is more important, they contribute to the relieving of
investment costs. However, numerical models require validation on
the basis of experimental data. Otherwise, their value is considerably
lower because they are not supported by the real terms [Ansys Inc.,
2013]. That is why, the developed Koflo® static mixer numerical
model (described fully in the previous paper [Stec and Synowiec,
2015] used for pressure drop predictions during one-phase turbulent
flow was verified in addition to obtained experimental data
performed on a specially designed laboratory setup.

Experimental
As previously mentioned, the pressure drop is one of the basic
parameters used for assessing the validity of static mixers apply
Fig. 1. Scheme of the experimental setup intended for pressure drop measurements:
[Thakur et al., 2003]. Due to it, it is possible to make a correct 1 – water tank, 2 – cenrtifugal pomp, 3 – Koflo static mixer, 4 – storage tank,
system design and what is also really important to compare the costs 5 – valve, 6 – thermostat, 7 – mechanical agitator, FCR – flow rate automatic ad-
of energy consumption with solutions based on use of mechanically justment with registration, PRd – differential pressure measurement with registra-
tion, TIR – temperature measurement with the identification of a set point value
agitated vessels. Furthermore, it should be emphasized that pressure and registration
drops will play a key role particularly in case of turbulent flows. It is
expected that their values will be high, especially for large Reynolds compared with the pressure drops calculated from the manufacturer’s
number. Therefore, owing to the lack of a strict correlation between correlation [Koflo® Corporation]. Achieved results were presented in
the pressure drop and Reynolds number in turbulent flow regime Fig. 2 together with the experimental and numerical data (based
in accessible literature it was decided to research the subject in on k-ε model). The design, experimental and computational points
order to broaden the existing knowledge and to validate were fitted using a power function in accordance with Eq. (1).
the prepared numerical model. For this purpose, the experimental
∆p ∝ Re C (1)
study was performed. The scheme of an experimental setup is
presented in Fig.1.
The water from the tank - 1 was pumped by centrifugal pump (with
the automatic adjustment of the revolution number) - 2 to the static
mixer - 3 with a differential pressure transducer. The pressure taps
were located in the inlet and outlet connections of the mixer to
enable the measurement of pressure drop during the flow of fluid.
After the end of the experiment the water was directed to a storage
tank - 4 and then with the use of the valve - 5a to the sewage system
(or recycled to the tank - 1 with mechanical agitator - 7 and used as
a feed in the next measurement series).
The experimental setup was also equipped with two additional
temperature sensors and thermostat - 6 providing a permanent
control of the process medium and control valves - 5 which help to
preserve the occupational safety.
All of the experiments were carried out in the range of Reynolds Fig. 2. Model validation for the flow in the range of Re = 1000÷5000
numbers Re = 1000÷5000 which were corresponding to flow rates F
= 40÷200 l/h (liters per hour). The water flow rate was changed by The comparison of the estimated exponent values C were pre-
increasing the rpm of the impeller in the pump. To increase the accu- sented in the Tab. 1.
racy of the experiments the measurements were repeated four times From Fig. 2 it can be observed that pressure drop values obtained
each and the final result was averaged. The commonly existing meas- by use of numerical simulations stay in a good compatibility with
urement error was in the limit of 15 %. experimental ones and those calculated from vendor’s correlation.
What is more, the influence of Reynolds number on an analyzed
Results and discussion quantity powered by CFD is almost identical with one given by
Numerical model validation Tab. 1.Values of C exponent from Eq. 1

The experimental pressure drops were used for validation of the Manufacturer’s
Experimental CFD simulation
developed CFD model (described in details in the previous correlation
manuscript [Stec and Synowiec, 2015]. Obtained values were also 2,0 1,780 1,998
 Please cite as: Inż. Ap. Chem. 2015, 54, 4, 201-203

str. 202 INŻYNIERIA I APARATURA CHEMICZNA Nr 4/2015

manufacturer. Admittedly, with regard to the experimental results the

exponent value at Reynolds number is lower than in two other cases
(Tab. 1) but it stays within the normal range allowing for the safety
margin and it is consistent with other literature data [Thakur et al.,
2003].
New pressure drop correlation
In the basis of pressure drop results obtained from CFD computa-
tions and performed experiments, a new pressure drop correlation in
the Koflo® static mixers was developed. For this purpose, the
dimensional analysis was performed (due to well known procedure,
only the most important assumptions were showed) and next (with
the use of available computer software) equation parameters were
calculated:
Fig. 3. Validation of the developer pressure drop correlation
∆ p = f ( w , ρ ,η , d SM , L ) (2)
where: Pressure drop as a function of Newton number
∆p – pressure drop [Pa] Well known is that the equation for pressure drop can be also
dSM – inner diameter of the static mixer [m] presented in another form as a function of Newton number. The start-
L – static mixer length [m] ing point used for presenting it in that kind of form is always the
w – fluid velocity [m/s] Darcy − Weisbach equation [Çengel, Cimbala, 2006] presented
ρ – fluid density [kg/m3] below
η – fluid dynamic viscosity [Paּs]
w2  L  (9)
B
∆p = (2CRe A ) ρ  
 L  (3) 2  d SM 
Eu = Re A  
 d SM  where the first term in brackets is defined as the friction factor f,
where: which included into Eq. (9) gives:
Re – Reynolds number
ρwd SM w2  L  (10)
Re = (4) ∆p = f ρ  
η 2  d SM 
Eu – Euler number
Next, knowing that the expression f/2 represents the Newton number
By use of Euler number definition [Bayer et al., 2003] the pressure drop equation can be written as:
∆p (5)  L  (11)
Eu = ∆p = Ne w 2 ρ  
w2 ρ  d SM 
Eq. (3) may be also transformed to the following form: The transformation of Eq. (11) gave the expression for Newton
B number
 L  2 (6)
∆p = Re A   w ρ ∆p  d  (12)
 d SM 
Ne = 2  SM 
w ρ L 
Because of the fact that presented paper did not consider the which can be also written as:
influence of mixer diameter and length, it was adopted (according to
d  (13)
the literature data [Bayer et al., 2003]) that B exponent in the relevant Ne = Eu SM 
B  L 
module  L  is equal to 1. Hence, the formula for pressure drop
  and that form (Eq. 13) was subsequently used for calculations in
 dSM 
order to compare Newton number values for experimental and CFD
can be written as: computations as so as results of developed correlation (Eq. 8). The
 L  2 (7) obtained data presented as a function of Reynolds number was
∆p = CRe A   w ρ
 d SM  showed in Fig. 4.
It can be seen from Fig. 4 that the Newton number (and so the fric-
After the data adjustment, the Koflo® static mixer empirical tion factor f) in the turbulent flow computed on the basis of CFD
equation (from which the pressure drop calculations can be done) was
presented in the following form:
 L  (8)
Eu = 4,95Re − 0, 22  
 d SM 

and it can be use for Re є (1000; 5000).

Next, in the aim of checking the accuracy between the values
calculated on the basis of Eq. (8) and other given data an appropriate
graph was created (Fig. 3), showing the pressure drop as a function of
Reynolds number.
From presented Fig. 3 it can be concluded that thedeveloped cor-
relation stays in a good agreement with bothexperimental and simula-
tion data (calculated relative error was in the range of 3,4÷8,6 % and
it can be successfully used for Koflo® static mixer pressure drop
predictions. Fig. 4. Newton number as a function of Reynolds number
 Please cite as: Inż. Ap. Chem. 2015, 54, 4, 201-203

Nr 4/2015 INŻYNIERIA I APARATURA CHEMICZNA str. 203

techniques has a constant value which is consistent with the litera- Conclusions
ture references [Kumar et al., 2008] and proves the correctness of
The paper deals with the validation of the developed Koflo® static
performed analysis. Furthermore, Fig. 4 shows a very good
mixer numerical model (described in the earlier report [Stec and
agreement between the developed pressure drop correlation (Eq. 8)
Synowiec, 2015] in addition to experimental data of pressure drops
and obtained experimental data. In turn, with regard to the test
made on the built experimental setup as so as the values calculated
results, some discrepancies were observed (Newton number shows
from the manufacturer’s correlation.
a slight downward trend). However, there is an expectation that the
From the presented study it can be concluded that the pressure drop
slope will decrease with the Reynolds number increase until the mo-
values obtained by performed experiments are consistient to both:
ment when the considered dependence (Ne = f(Re)) becomes
CFD simulation results and values calculated from equation given by
linear. That is why occurred deviations may be regarded as staying
the vendor. It should be mentioned that there were some slight dis-
within the acceptable limits and were found to be correct.
crepancies (especially between the calculated from Eq. (1) exponent
It should be also mentioned that equation describing the Newton C values, however they stayed within the limits of acceptability and
and Reynolds number comparative dependence (presented in Fig. 4) were related to the commonly occurring measurement errors.
shows clearly that the friction factor f is proportional to Reynolds On the basis of obtained results, the new pressure drop correlation
number in the power of -0,22 was developed (Eq.8). The equation proposed by the authors stayed
in a good agreement with both experimental and simulation data and
f ∝ Re −0, 22 (14) it can be successfully used for Koflo® static mixer pressure drop
predictions.
That fact complies with the Nikuradse Equation [Çengel, Cimbala, Furthermore, the analysis of Newton number presented in the
2006] used for friction factor calculations during the flow in the Fig. 5, which includes different static mixers geometries shows that
turbulent regime and it also demonstrates the rightness of the per- the use of Koflo® static mixer results in obtaining the smallest
formed tests: pressure drops. That stands a very important feature which makes the
0,221 (15) Koflo® type an interesting alternative to other well known used
f = 0,0032 + 0, 237
Re constructions.
All in all, the experiments connected to pressure drops acquired
If the comparison of Newton number as a function of Reynolds
during the flow in the range of 40÷200 l/h confirmed that the numeri-
number will be made for different commercially available static
cal model of Koflo® static mixer was prepared in the proper way and
mixers [Bayer et al., 2003], adding the line corresponding to analyzed
means that CFD techniques can be successfully intended for pressure
Koflo® type (Fig. 5), it can be concluded that it shows the smallest
drop predictions. What is more, the Koflo® static mixer, unrecognized
Newton number value which results in obtaining the smallest pressure
in the literature, provides a great base to other experiments that may
drops (in comparison to considered geometries).
affirm its superiority to another, commonly used types.
This allows for the conclusion that use of Koflo® static mixer may
be a competitive solution in respect to commonly used mixer types REFERENCES
and it’s worth considering. ANSYS Inc., 2013. ANSYS Fluent Theory Guide, Release 15.0
Bayer, T., Himmler, K., Hessel, V., 2003, Don’t be baffled by static mixers.
How to select and size the correct static mixer. Chem. Eng., 110, 50-57
Berkman P.D., Calabrese R.V., 1988. The Dispersion of viscous liquids by
turbulent flow in a static mixer. AIChE J., 34, 602-609. DOI: 10.1002/
aic.690340409
Çengel, Y.A., Cimbala, J.M., 2006, Fluid Mechanics. Fundamentals and
Applications. Chapter 8 – Flow in pipes, 321-336
Koflo® Corporation advertising materials, Koflo® Static In-line Mixers
Kumar, V., Shirke, V., Nigam, K.D.P., 2008. Performance of Kenics static
mixer over a wide range of Reynolds number. Chem. Eng. J., 139, 284-
295. DOI: 10.1016/j.cej.2007.07.101
Stec, M., Synowiec, P.M., 2015. Numerical method effect on pressure drop
estimation in the Koflo® static mixer. Inż. Ap. Chem., 54, nr 2, 48-50
Thakur, R.K., Vial, Ch., Nigam, K.P.D., Nauman, E.B., Djelveh, G., 2003.
Fig. 5. The comparison of Reynolds number dependence on Newton number Static mixers in the process industry – A review. Chem. Eng. Res. Design,
for different static mixers 81, 787-826, DOI: 10.1205/026387603322302968

The scientific and technological journal

INŻYNIERI A I AP ARATURA CHEMICZNA
Chemical Engineering and Equipment
published since 1961
Journal is devoted to process calculations, construction and designing problems dealing with equipment and
devices for process industries, especially chemical, petrochemical, power and food industry, both municipal
engineering and environmental protection.
Readership consists of research workers, constructors and designers, managers and engineers.
Papers are dealing with unit operations of chemical engineering, processes and operations in such areas as
bio- and nanotechnology, biomedical engineering, recycling, process safety. Scientific research, improved
design methods, proper operating and maintenance of various apparatuses and devices are presented
considering better capacity, better use of raw materials, energy saving and environmental protection.
Papers are revised by professional referees.
Journal homepage: http://chemical-engineering-equipment.eu