Journal Pre-proof

Comparative analysis of different static mixers performance by

CFD technique: An innovative mixer

M.M. Haddadi, S.H. Hosseini, D. Rashtchian, Martin Olazar

PII: S1004-9541(19)30868-7
DOI: https://doi.org/10.1016/j.cjche.2019.09.004
Reference: CJCHE 1563

To appear in: Chinese Journal of Chemical Engineering

Received date: 31 May 2019

Revised date: 3 August 2019
Accepted date: 6 September 2019

Please cite this article as: M.M. Haddadi, S.H. Hosseini, D. Rashtchian, et al., Comparative
analysis of different static mixers performance by CFD technique: An innovative mixer,
Chinese Journal of Chemical Engineering(2019), https://doi.org/10.1016/
j.cjche.2019.09.004

This is a PDF file of an article that has undergone enhancements after acceptance, such
as the addition of a cover page and metadata, and formatting for readability, but it is
not yet the definitive version of record. This version will undergo additional copyediting,
typesetting and review before it is published in its final form, but we are providing this
version to give early visibility of the article. Please note that, during the production
process, errors may be discovered which could affect the content, and all legal disclaimers
that apply to the journal pertain.

© 2019 Published by Elsevier.

 Comparative Analysis ofJournal Pre-proof
Different Static Mixers Performance by
CFD Technique: An Innovative Mixer
M. M. Haddadi a, S. H. Hosseini b* D. Rashtchian a, Martin Olazar c
a
Chemical and Petroleum Engineering Department, Sharif University of Technology, Tehran, Iran

b
Department of Chemical Engineering, Ilam University, Ilam 69315–516, Iran

c
Department of Chemical Engineering, University of the Basque Country (UPV/EHU), Sarriena s/n, Leioa, Spain.

Abstract: The flow and mixing behavior of two miscible liquids has been studied in an innovative static

of
mixer by using CFD, with Reynolds numbers ranging from 20 to 160. The performance of the new

ro
mixer is compared with those of Kenics, SMX, and Komax static mixers. The pressure drop ratio (Z-

-p
factor), coefficient of variation (CoV) and extensional efficiency (𝛼) features have been used to evaluate
re
power consumption, distributive mixing and dispersive mixing performances, respectively, in all mixers.

The model is firstly validated based on experimental data measured for the pressure drop ratio and the
lP

coefficient of variation. CFD results are consistent with measured data and those obtained by available
na

correlations in the literature. The new mixer shows a superior mixing performance compared to the

other mixers.
ur
Jo

Keywords: CFD, New Static mixer, Coefficient of Variation, Pressure drop ratio.

*
. Corresponding author: E-mail address: [email protected] (S. H. Hosseini)
 Journal Pre-proof

1. Introduction

Mixing is one of the most important unit operations in many processes, which has a major role

on the quality of the final product. Improper mixing can result in a lower quality of products, and

therefore the need for more sophisticated separation systems, which in turn increases waste

management costs. Therefore, determining mixing characteristics is extremely important,

especially in flows with low Reynolds numbers, as they involve non-homogenous and

of
imperfectly mixed flow. Mixing is carried out as a batch operation in agitated vessels and as a

continuous operation in static mixers [1]. The static mixers are composed of numerous

ro
motionless elements installed in a fixed pipe or channel. The aim of these elements is to split,

-p
rotate and then recombine the input streams. Since there are no moving parts in static mixers, the
re
only power required is the one of the external pump in order to overcome the pressure loss
lP

generated by the mixing elements and fluid movement. Some benefits of static mixers in

comparison with dynamic ones can be listed as follows: small space required for installation, low
na

maintenance costs, low power consumption, absence of moving parts, short residence times of
ur

the fluids, near plug flow behavior of the fluids, efficient heat and mass transfer, good mixing at

low shear rate, and capability for self-cleaning [2-4]. The static mixers are used in many
Jo

processing applications, such as mixing of miscible liquids [5], immiscible liquids [6,7],

dispersion of viscous liquids [8], mass transfer [9], heat transfer [10], and gas-liquid dispersion

[11,12]. These applications are available in many industries, such as polymer processing [13],

biotechnology [14], wastewater treatment [15-18], food processing [19], and petrochemistry

[20,21]. The wide range of applications of static mixers in various industries is evidence of the

importance of this equipment.

2
 Journal Pre-proof

The performance of a static mixer can be mainly specified by its mixing efficiency and pressure

drop. First, they are assessed by the pressure drop ratio (pressure drop of the mixer to pressure

drop of the empty pipe). Then mixing efficiency parameters are considered by using the

coefficient of variation, the extensional efficiency, and the residence time distribution. The

coefficient of variation characterizes the effectiveness of distributive mixing. When distributive

mixing occurs, droplets or bubbles are distributed in the mixer. This mechanism differs from

dispersive mixing, which corresponds to size reduction [22]. In order to perform distributive

of
mixing, a component A is mixed in a mainstream with another component B. Then, the

ro
concentration of component A is monitored and mapped as a function of position in a specified

-p
cross-section. Afterwards, the standard deviation of the concentration is calculated and then the
re
coefficient of variation (CoV) is specified. If component A is uniformly distributed in a cross-

section, the CoV tends to zero. The residence time distribution is sometimes used as a mixing
lP

criterion, especially when the systems include a chemical reaction. A narrower distribution leads
na

to better mixing [23]. Moreover, the extensional efficiency is introduced in order to investigate

dispersive mixing [24].

ur

In the last few decades, computational fluid dynamics (CFD) has become an essential numerical
Jo

tool in the prediction of fluid dynamics, heat transfer and also mass transfer in packed beds,

which are used in various industries [25,36]. Therefore, CFD can be successfully applied to the

design and analysis of pieces of equipment like static mixers.

Different aspects of the static mixers have been studied both experimentally and numerically. For

instance, Liu et al. [27] and Baumann et al. [28] performed numerical and experimental studies,

respectively, to measure the pressure drop in SMX static mixers. Rauline et al. [22] studied the

velocity field for creeping flow in six static mixers using three-dimensional finite volume CFD

3
 Journal Pre-proof

code. They used the numerical results of the extensional efficiency, pressure drop, mean shear

rate, stretching, and coefficient of variation to compare the performances of those mixers. They

proved that the SMX mixer was the most efficient amongst all configurations. Moreover, by

studying the Kenics mixer they showed that this mixer is more efficient when certain space is left

between the mixing elements. Zalc et al. [29,30] investigated the hydrodynamics and mixing

behavior in an SMX mixer at Reynolds number in the range of 10 -4 −100. Their results revealed

that when the injection is made at the center of the tube, the mixing efficiency is higher than for

of
off-center injection. In addition, they showed that the pressure drop ratio does not change for

ro
Reynolds numbers lower than 10. Several CFD simulations of flow behavior through Kenics

-p
mixers are found in the open literature [31-36]. Furthermore, some experimental studies have
re
also been conducted by several researchers on the flow behavior in Kenics static mixers [37,38].

However, there are few numerical studies about the Komax mixer.
lP

Most of the studies about designing novel static mixers or modifying static mixers available at
na

laboratory scale have been limited to creeping flows and laminar flows at higher flow rates, i.e.,
ur

to a low Reynolds number range, in which the influence of internals, such as blade shape,

number of elements and other design factors, have been assessed by numerical tools [34,39-42].
Jo

Regner et al. [34] studied the characterization of two static mixers, namely, Kenics KM and

Lightnin Series 45, by evaluating Z factor, helicity and the rate of striation thinning based on

CFD. The Reynolds number they used was in the 0.1-160 range. Singh et al. [39] analyzed, and

subsequently optimized, the various design parameters of SMX mixers, namely, N p, the number

of parallel cross-bars per element and, Nx, the number of cross-bars over the width of the

channel, by the mapping method (MM) under a low Re number (0.44). They obtained that the

optimum design obeys the rule of Np = (2/3)Nx – 1. To achieve the enhanced mixing
4
 Journal Pre-proof

performance of the SMX geometry, Soman and Madhuranthakam41 simulated different SMX

static mixer geometries by COMSOL Multi-physics software to investigate mixing performances

at Re numbers ranging from 30 to 100. They compared the results in terms of velocity field,

pressure field, shear rate and extensional efficiency for dispersive mixing, and the standard

deviation of the concentration for distributive mixing. They found that circular serrations in

SMX mixers reduce the pressure drop by 33–35% of that corresponding to the standard SMX

mixer. Meng et al. [41] compared (using a CFD tool) the mixing performance of Kenics and

of
Lightnin static mixers for a high-viscosity fluid at low Reynolds numbers in the 0.1-100 range

ro
and different aspect ratios, as well as different elements. They proposed a new mixing factor ƞ to

-p
evaluate the micro-mixing performance in the mixers. In another work, Meng et al. [42] studied
re
the mixing performance of a high-viscosity fluid in novel static mixers with multitwisted leaves

by CFD code FLUENT at Reynolds number in the 0.1−150 range.

lP

In addition to their use for the laminar flow, some numerical works in the open literature deal
na

with novel static mixers or modifications to the existing ones for use in turbulent flow. For
ur

instance, Meng et al. [43] simulated perforations in the standard Kenics mixer by CFD tool and

studied the impact of their size and spacing, as well as the number of segments, on the heat
Jo

transfer under the turbulent condition. They achieved outstanding results using d/W= 0.3 and

s/W= 0.6 in terms of friction loss and heat transfer rate in the modified Kenics mixer. Zhou et al.

[44] proposed a new static mixer of hollow cross disk in order to increase the turbulent reactive

mixing in the tubular pipe under turbulent flow.

Few studies have been reported in the literature regarding static mixers with complex fluids, such

as non-Newtonian and immiscible ones. In this respect, Jegatheeswaran et al. [45] studied the

intensification of non-Newtonian fluid mixing in chaotic SMX static mixers at low Reynolds
5
 Journal Pre-proof

numbers. They also reviewed process intensification by mixing non-Newtonian fluids in static

mixers, and suggested to perform CFD studies to shed light on the impact the geometry and

design parameters of static mixers have on fluid deformation and mass transfer under low

Reynolds numbers for both Newtonian and non-Newtonian fluids [46]. Mihailova et al. [47]

evaluated a Newtonian fluid (glycerol) and a non-Newtonian one (guar gum solution) in a SMX

mixer, and found that their velocity distribution is hardly affected by rheology at the same low

flow rate. Concerning the performance of immiscible fluids in static mixers, few studies have

of
been reported in the literature, and they use Eulerian-Lagrangian and Eulerian-Eulerian

ro
approaches combined with the population balance in order to study particle trajectories and

particle size distributions, respectively [48,49].

-p
re
The general goal of this study is to investigate the effects of Reynolds number and length of the
lP

mixer on the hydrodynamics and mixing behavior of brine 25% (wt) by using CFD in four static

mixers at low Reynolds numbers. In view of the brief review done above, the mixing behavior of
na

the new static mixer is firstly analyzed at low Reynolds numbers due to the high probability of
ur

having fluid inhomogeneity under those conditions. Accordingly, this study covers a wide range

of Reynolds numbers from 20 to 160 as a benchmarked to evaluate the mixing performance of

Jo

the new simple static mixer inspired by the behavior of dynamic mixers in vessels. Amongst the

different criteria, extensional efficiency, pressure drop ratio and the coefficient of variation have

been chosen to evaluate the performance of the well-known Kenics, SMX and Komax mixers

and of an innovative static mixer. The simulation results are validated by comparing the

predicted results with the existing correlations for the pressure drop ratio in a Kenics mixer and

the experimental data for the coefficient of variation in a mixer with baffle shape elements.

2. Numerical methodology
6
 Journal Pre-proof

2.1. Physical model and fluid properties

This study analyses the performance of four static mixers, namely, Kenics, SMX, Komax, and a

new static mixer proposed by the authors. Furthermore, in order to validate the CFD model,

another two other static mixers have been used, namely, a static mixer with staggered

semicircular baffle shape elements [50] and a Kenics static mixer [51] with the same geometrical

dimensions and operating conditions as those used in the mentioned four static mixers. Based on

of
the previous works [50,51], it is assumed that the tube and internal elements of the static mixers

are made of Plexiglass. Figure 1 shows the computational domains of Kenics, SMX, Komax and

ro
the new static mixer. Moreover, Figure 2a illustrates the geometry of the blades in the new

-p
mixer. More information about the static mixer with baffle shape elements can be found
re
elsewhere [50]. In Figure 1, the computational domain comprises a horizontal tube with a
lP

diameter of D and four mixing elements. Two empty tubes of 3D length are added prior and

subsequent to the mixing elements in order to provide developmental flow length and to avoid
na

creating backflow in the simulations. As shown in Figure 2b, the injection tube is a curved 3 mm
ur

diameter tube aligned with the main axis, with the injection point being located at a distance of

15 mm from the tube inlet. The detailed dimensions of the static mixers and the properties of the
Jo

fluid are set out in Tables 1 and 2, respectively. Each element of the Kenics mixer is a twisted

plate that is alternated and oriented so that each leading edge is at 90° of the trailing edge of the

next one. The mixing elements in the SMX mixer comprise inclined bars crossed to one another.

Each element is rotated 90° with respect to the foregoing element. The Komax static mixer is

manufactured from slotted plates with inclined ends. The flow channels with triangular shapes

are created by pushing the elements together in a pipe.

7
 Journal Pre-proof

of
ro
-p
re
lP
na
ur
Jo

Figure 1. Geometries of various static mixers, namely, Kenics mixer, SMX mixer, Komax mixer,
and novel mixer in different views.

The first mixing element of the new mixer consists of two rows of blades at a distance of D/2

from each other with 90° differences in angle. Each row has four blades, which form two crosses
8
 Journal Pre-proof

with an angle of 45° with respect to the tube axis at a distance of D/2 from each other. The

second element of the new mixer is the same as the first element with the difference that all

blades are rotated 90° around the pipe axis. Thus, the number of required elements to construct

the new mixer should be at least two. It is interesting to note that the design of the proposed

mixer is inspired by the dynamic mixers in stirred vessels, which include several blades at

various angles along the central axis.

of
ro
-p
re
lP
na
ur
Jo

Figure 2. (a) Geometry of the blades in the new static mixer (b) sketch of the pipeline equipped
with mixing elements.

2.2. Model equations and numerical details

The general assumptions of the model used are as follows: the flow is single-phase, three-

dimensional, steady state, incompressible, isothermal and without any reaction (without mass

production or mass consumption). The continuity, momentum, and species transport equations

9
 Journal Pre-proof

are solved by a finite volume method (FVM). The mentioned equations are expressed as follows

[52]:

∇. (𝜌𝐕) = 0 (1)

∇(𝜌𝐕𝐕) = −∇P + ∇ ∙ [𝜇∇𝐕] + 𝜌𝐠 (2)

∇ ∙ (𝜌𝐕YA ) = ∇ ∙ (𝜌DAB ∇YA ) (3)

of
where 𝜇, 𝜌, 𝐠, 𝐕, P, Y𝐴 , DAB stand for the bulk dynamic viscosity, bulk density, gravity

ro
acceleration vector, mean velocity vector, pressure, mass fraction of component A, and

molecular diffusivity respectively.

-p
Table 1. Geometrical Parameters of Static Mixers
re
parameter Kenics SMX Komax New mixer
lP

tube diameter, D (mm) 25 25 25 25

na

Aspect ratio, AR 1.5 1.5 1.5 1.5

ur

blade thickness, 𝐭 (mm) 3 3 3 3

Jo

blade width, 𝐛𝐰 (mm) - 5 - 12.5

blade length, 𝐛𝐥 (mm) - - - 12.5

Table 2. Properties of the Fluids.

parameter main liquid water tracer brine 25% (wt)

density (kg/m3) 1000 1190

viscosity (kg/(m.s)) 0.001 0.00167

10
 Journal Pre-proof

diffusion coefficient (m2/s) 1.44×10-9 1.44×10-9

Flow rate (lit/s) 0.00035-0.0026 0.000035-0.00026

Reynolds number 20-160 105-796

The governing equations are discretized on an unstructured tetrahedral mesh as algebraic

equations. The origin of the z-axis is positioned at the inlet of the tube and its direction is

opposite to the main flow. The convective terms in the governing equations are discretized by a

of
second order upwind scheme, and the SIMPLE algorithm is used for pressure-velocity coupling.

ro
Gradients are essential to compute diffusion terms and velocity derivatives. Moreover, the values

-p
of the scalar at the cell faces should be created. To solve the conservation equations and calculate
re
the gradients, a commercial ANSYS FLUENT 16.0 software is used in order to apply a Least-

Square cell-based discretization method. The convergence limit is set to a scaled residual lower
lP

than 10-5 for all the mentioned equations. The various geometries of the mixers and the
na

corresponding computational grids have been created by ANSYS DESIGN MODELER and

ANSYS MESHING, respectively. The simulations are performed on an Intel Core™ i7-6700HQ
ur

CPU@ 2.60 GHz running on Windows 10 Professional 64 bit with 12 GB of RAM.

Jo

2.3. Boundary conditions

The Reynolds number is expressed by:

̅D
𝜌V
Re = 𝜇
(4)

̅ stands for the mean velocity and D for the diameter of the tube including the mixing
where V

elements. The boundary condition at the inlet of the mixer is a fully developed velocity profile,

which has been introduced into the software using a user-defined function (UDF). According to
11
 Journal Pre-proof

previous modeling of static mixers [5,53,54], the no-slip boundary condition is utilized on the

tube wall and element surfaces. The atmospheric pressure, i.e., zero gage pressure, is utilized at

the outlet of the static mixers. Brine 25% (wt) is injected into the tube co-currently to the main

flow. The tracer (brine) velocity chosen at each Reynolds number (based on the inlet condition of

the main liquid) is that corresponding to a ratio of 0.1 between the tracer flow rate and the main

flow rate.

of
2.4. Grid independence test

ro
Grid independence test is performed to determine the proper grid size, ensuring the numerical

-p
results are independent of the grid system. The maximum values of Equisize skew for all

geometry meshes are lower than 0.79, which is evidence of the good quality of the meshes based
re
on the FLUENT criterion. In addition, the geometries and grids have been created by
lP

DesignModeler, from ANSYS productions. In order to perform the grid independence test, four

static mixers (Kenics, Komax, SMX and the novel one) with a length of 300 mm and a diameter
na

of 25 mm are used with six different grid sizes ranging from 0.5 to 3 mm (0.5 mm is chosen for
ur

the step size) at Re = 160. Tables 3 to 6 show the results of the grid independence test. As
Jo

shown in these tables, both Z-factor (related to pressure drop) and CoV (related to the

concentration at the plate located 1.5D downstream) decrease as the grid size is increased above

1.5 mm. In addition, by changing the grid size from 1 to 0.5 mm, Z-factor and CoV change by

less than 0.7 % and 0.5 %, respectively. Tables 3 to 6 summarize the key characteristics of the

grids, and they clearly show that CFD simulation time is highly dependent on the grid size.

Given the slight difference between 1 and 0.5 mm grid sizes, medium grid size (1 mm) strikes

the balance between accuracy and computational time, and has been used for further simulations.

12
 Journal Pre-proof

Table 3. Grid size, CFD simulation time, number of elements, estimated Z and CoV for the Kenics mixer at Re=160.

Grid size (mm) 0.5 1 1.5 2 2.5 3

Number of 9393720 1203476 380517 185645 115263 86139

of
elements

CFD simulation 970 55 14.1 6 3.5 2.7

ro
time (min)

Simulated CoV

Simulated Z
0.185

10.23
0.187

10.22 -p
0.15

10.27
0.07

10.14
0.04

9.95
0.02

9.85
re
lP

Table 4. Grid size, CFD simulation time, number of elements, estimated Z and CoV for the Komax mixer at
Re=160.
na

Grid size (mm) 0.5 1 1.5 2 2.5 3

Number of 9464131 1210219 384182 186300 116473 86452

ur

elements

CFD simulation 1000 56 14.2 6.1 3.6 2.7

Jo

time (min)

Simulated CoV 0.04665 0.047 0.044 0.041 0.039 0.037

Simulated Z 36.18 36.34 35.48 34.35 31.7 31.56

Table 5. Grid size, CFD simulation time, number of elements, estimated Z and CoV for the SMX mixer at Re=160.

Grid size (mm) 0.5 1 1.5 2 2.5 3

Number of 8807816 1139632 360097 180105 112442 84964

elements

13
 Journal Pre-proof

CFD simulation 750 53.5 13 5.3 3.1 2.3

time (min)

Simulated CoV 0.00695 0.007 0.0065 0.006 0.0055 0.005

Simulated Z 98.3 98.67 96.15 94.2 93.5 89.3

Table 6. Grid size, CFD simulation time, Number of elements, estimated Z and CoV for the new mixer at Re=160.

Grid size (mm) 0.5 1 1.5 2 2.5 3

Number of 9323515 1195094 378079 184912 115256 85647

of
elements

CFD simulation 950 54 14 6 3.5 2.5

ro
time (min)

Simulated CoV

Simulated Z
0.001997

63.2
0.002

63.5 66-p
0.0017 0.0015

64.2
0.0012

62
0.001

60.5
re
lP

2.5. Model validation

na

The accuracy of the numerical model is an essential aspect to be quantified prior to use it for

predicting new situations. To assess the validity of the CFD results, the simulated pressure drop
ur

ratio and the coefficient of variation are compared with the experimental data and available
Jo

correlations. The Z-factor is defined as the ratio of the static mixer pressure drop, ∆P, to the

pressure drop in the empty tube, ∆P0 :

∆P
Z = ∆P (5)
0

In this study, the pressure drop is calculated as the absolute difference between the area-weighted

average pressure at the planes located at 1.5D upstream and downstream the mixing elements.

14
 Journal Pre-proof

This is an important parameter for static mixers and provides a measure of the energy required

for the static mixer in the pipeline.

Moreover, the coefficient of variation is calculated at different points in a specified cross-section,

with the usual definition being as follows:

∑N
i=1(Yi −Ymean )
2 1
CoV = √ (6)
N−1 Ymean

of
where Yi is the local mass fraction of the tracer at the ith point, N is the number of evaluation

ro
points and Ymean is the mean mass fraction of the tracer on the cross-section of the tube.

Therefore, values of the CoV are calculated in various points of the tube cross-section following
-p
previous works [18,50]. Figure 3(a, b) compares the values computed for Z-factor in the range
re
Re = 20 − 160 and for CoV in the range Re = 70 − 480 with the data measured in the Kenics
lP

mixer [52] and in a static mixer with baffle shape elements [50], respectively. Figure 3a shows

that an increase in the Reynolds number leads an increase in pressure drop, which in turn leads to
na

an increase in Z-factor.
ur
Jo

15
 Journal Pre-proof

of
ro
-p
Figure 3. (a) Experimental and computed pressure drop ratios for various Reynolds numbers (b)
Experimental and computed CoV values for various Reynolds numbers.
re
1 |Simulationi −Experimentali |
MRE = N ∑N
lP

i=1 |Experimental i |
(7)

The mean relative error (MRE) between the data measured by Sir and Lecjaks [51] and the CFD
na

results is approximately 2.3%. Accordingly, a quite good quantitative agreement between the
ur

simulated data and the experimental ones by Sir and Lecjaks [51] is found for Z-factor. Figure 3b

compares the computed values of CoV with the corresponding experimental data provided by
Jo

Al-Atabi [50] in a static mixer with baffle shape elements for Re = 70 − 480. The MRE

between the experimental data (Al-Atabi, 2011) and the simulated results is of around 6%.

Therefore, a good agreement between the results is also attained for CoV. Overall, Figure 3(a, b)

shows that the current model is able to predict the hydrodynamics and mass transfer of brine into

water flowing through the various static mixers at different Reynolds numbers with a reasonable

accuracy. As a result, the current CFD model is suitable for studying the other static mixers.

2.6. Comparison with some existing empirical equations

16
 Journal Pre-proof

Figure 4 compares the CFD results with those for well-known available correlations developed

by Wilkinson and Cliff [55], Cybulski and Werner [19], Sir and Lecjaks [51], Shah and Kale

[56], and Grace [57] for a Kenics mixer including four elements. These empirical expressions

that are commonly used for estimating Z-factor in Kenics static mixers are listed in Table 7.

There is a considerable discrepancy in the results predicted by the different correlations for Z-

factor. This discrepancy is explained by the differences in the dimensions of the static mixers,

such as diameter and length, and because the impact of some geometrical features has been

of
ignored in the correlations. Unlike the variability predicted by the existing correlations for the

ro
Kenics static mixer, the numerical results are consistent. In the mentioned range of Reynolds

-p
numbers, the computed Z-factors have 3.6% and 8% MRE compared to the correlations by
re
Cybulski and Werner [19], and Sir and Lecjaks [51], respectively.
lP
na
ur
Jo

17
 Journal Pre-proof

of
ro
-p
re
lP

Figure 4. Comparison of CFD results with those calculated using the correlations available for
the Kenics mixer at various Reynolds numbers.
na

Table 7. Correlations for calculating Z-factor in Kenics Static Mixers.

ur

authors correlation Reynolds range

Grace [57] Z = 4.86 + 0.68√Re 𝑅𝑒 < 1000

Jo

Wilkinson and Cliff [55] Z = 7.2 + 0.031Re 𝑅𝑒 < 50

Sir and Lecjaks [51] Z = 5.34 + 0.021Re 𝑅𝑒 < 2300

Cybulski and Werner [19] Z = 5.4 + 0.028Re 𝑅𝑒 < 50

Shah and Kale [56] Z = 4 + 0.115Re0.66 0.01 < 𝑅𝑒 < 300

2.7. Numerical diffusion

One of the most important sources responsible for numerical errors in CFD is numerical

diffusion. This type of error stems from the discretization of convective terms in the governing
18
 Journal Pre-proof

equations. To investigate the numerical diffusion and select the appropriate discretization scheme

for the convective terms, five discretization schemes (First order upwind, Second order upwind,

QUICK, MUSCL and Power law) have been compared with each other. The results of this

comparison are shown in Figure 5. This figure clearly shows that except the first order upwind

and power law schemes, the other discretization schemes lead to similar results. Therefore, the

use of QUICK, MUSCL and second order upwind schemes do not lead to significant error in the

computation. The percentages have been calculated for the numerical differences between the

of
accurate scheme of second order upwind and the other schemes. The minimum and maximum

ro
differences for mass transfer are obtained by the QUICK scheme with a MRE of 0.37 % and by

-p
the first order upwind with a MRE of 78 %, respectively. The minimum and maximum
re
differences for pressure are obtained by the QUICK scheme with a MRE of 0.64 % and by the

first order upwind with a MRE of 45%, respectively. In addition, the second order upwind
lP

scheme was specifically used in many studies to simulate flow in static mixers and other systems
na

[32, 58, 59]. Accordingly, the second order upwind discretization scheme has been used in this

study for all simulations.

ur
Jo

19
 Journal Pre-proof

of
Figure 5. Radial distribution of (a) total pressure, (b) mass fraction of the brine at the outlet of

ro
Kenics static mixer with four mixing elements.

3. Results and discussion -p

re
By utilizing the numerical technique described in the previous sections, the flow behavior in the
lP

four static mixers has been evaluated and the results are discussed for various conditions.
na

3.1. Velocity profiles

ur

The contour plots of cross-sectional velocity profiles for different static mixers are shown in

Figure 6(a-d) at different locations from the inlet of the first mixing element to the end of each
Jo

mixing element at Re = 60. This figure clearly shows that the maximum values of the fluid

velocity are quite different in the mixers due to their different structure. Figure 6a shows the

contour of fluid velocity in the Kenics mixer. As observed, four velocity zones are formed at the

end of the first element. As the flow passes through the mixer, these zones combine together and

create new zones with different velocities. As expected from the mass conservation, the velocity

is proportional to the Reynolds number and is the same at the end of the first and third mixing

20
 Journal Pre-proof

elements. In the fourth element, the four velocity zones combine together again and two velocity

zones with lower pressure are created.

of
ro
-p
re
lP
na
ur
Jo

Figure 6. Cross sectional profiles of velocity at Re=60 for: (a) Kenics mixer (b) Komax mixer (c)
Proposed mixer (d) SMX mixer.
Figure 6b shows the velocity distribution in the Komax mixer. When the fluid flow enters the

first mixing element, it splits into two streams. As the cross-sectional area of each stream

decreases, the stream velocity increases. It is obvious from Figure 6b that a high velocity zone

near the wall mixer is created in each stream. This is due to the rotation of the fluid, which
21
 Journal Pre-proof

causes a centrifugal force and creates high speed zones at a distance far from the center of

rotation. This phenomenon is also observed in the Kenics mixer. The velocity profiles of the new

mixer are shown in Figure 6c. Similarly to the Komax mixer, there are two zones with high

velocity in this mixer. It is remarkable that these zones are created when the fluid passes through

the first mixing element, which is not the case in the Komax mixer where they are created after

the second mixing element. This means that the new mixer can create more disturbances

compared to the Komax one. In addition, the velocity contours of the SMX mixer are shown in

of
Figure 6d. As can be seen, the cross-bars divide the input fluid into several streams with different

ro
velocities. Furthermore, the high velocity zone in the SMX mixer is only in a region after the

first and the third mixing elements.

-p
re
lP

3.2. Pressure drop ratio

Figure 7a shows the values of Z-factor computed for different static mixers with the same
na

dimensions and four mixing elements. As expected, Z-factor increases by increasing the
ur

Reynolds number. It is clearly seen that there is a great difference in Z-factor for the different
Jo

static mixers. The Z-factor in the SMX mixer is higher than in the other mixers. This is due to

the complex geometry and the small space between the cross-bars in the SMX mixer. Moreover,

Figure 7a shows that the Z-factor for the new mixer is about 45% lower than for the SMX mixer

at all Reynolds numbers, whereas this factor is higher than those in the Kenics and Komax

mixers. In addition, the Z-factor of the new mixer has a lower slope than for the SMX mixer.

Thus, it is concluded that even for Reynolds numbers above the range studied, the Z-factor of the

new mixer is lower than that of the SMX mixer, with the trends of Z-factors being approximately

parallel in these mixers. Nevertheless, the lowest computed Z-factor corresponds to the Kenics
22
 Journal Pre-proof

mixer. In order to investigate more accurately pressure drop in the static mixers, their Z-factor

for attaining a value of CoV=0.05 have been plotted in Figure 7b. As shown in Figure 7b, when

the Reynolds number is greater than 40, the new mixer leads to a lower pressure drop compared

to the other mixers to achieve CoV=0.05. Moreover, high mixing qualities in the Kenics mixer

require high pressure drops. This difference in the pressure drop produced by the mixers is due to

the difference in the number of mixing elements required to reach a given degree of mixing.

Another point to be noted in Figure 7b is the decrease in Z-factor when the Reynolds number

of
(flow rate) is increased. In fact, an increase in flow rate increases the rate of mixing, and

ro
therefore less mixing elements are required and so pressure drop decreases.

-p
re
lP
na
ur
Jo

Figure 7. (a) Comparison of Z-factor in the different static mixers with four mixing elements, (b)
Z-factors in the static mixers for different Reynolds numbers when the target is CoV=0.05.

3.3. Coefficient of variation

23
 Journal Pre-proof

Figure 8a shows the relationship between the Reynolds number and the coefficient of variation

for the different static mixers. From this figure, it is clear that an increase in Reynolds number

leads to a reduction in CoV. At all ranges of Reynolds number, the Kenics mixer has the highest

CoV value. Moreover, for Re < 40 the new static mixer provides a higher CoV in the tube cross-

section than the SMX mixer, and therefore has a lower mixing efficiency. Nevertheless, for

Re > 40, the new mixer shows a better performance in the distributive mixing compared to the

other static mixers. Figure 8b shows the relationship between the coefficient of variation and the

of
length of the mixer at a constant Re number of 60. By increasing the mixer length, the CoV value

ro
decreases gradually. For L/D > 9 in the SMX and new mixer, CoV reaches almost a constant

-p
value. In addition, the new mixer shows the best performance in the distributive mixing
re
compared to the other mixers for all ranges of L/D. Overall, the new static mixer shows an

attractive mixing performance because pressure drop is considerably lower (about 45%) than for
lP

the SMX mixer. Therefore, to achieve a desired mixing efficiency, the new static mixer leads to
na

a reduction in pressure drop and requires a shorter length.

ur
Jo

24
 Journal Pre-proof

of
ro
-p
Figure 8. Comparison of CoV values in the different static mixers: (a) Influence of the Reynolds.
number on the coefficient of variation (b) Effect of L/D on the coefficient of variation at Re=60.
re
lP

3.4. Extensional efficiency

na

One of the most important parameters in the evaluation of dispersive mixing is extensional

efficiency, 𝛼, which is calculated as follows [24]:

ur

|γ|
𝛼 = |γ|+|ω| (8)
Jo

where |ω| and |γ| are the norms of the rate of vorticity and deformation tensors, respectively,

which are given by:

1
|γ| = [∇V + ∇V T ] (9)
2

1
|ω| = [∇V − ∇V T ] (10)
2

25
 Journal Pre-proof

where ∇V is the velocity gradient. Generally, a higher value of 𝛼 means a higher dispersive

mixing in a static mixer.

The general trend in a static mixer is that a fluid with a high extensional efficiency has both high

velocity gradient and high deformation tensor. According to the correlation shown below (Eq.

(11)) proposed by Das et al. [60], as the velocity gradient, and therefore the extensional

efficiency, is higher the static mixer may produce finer droplets in multi-phase systems.

of
2𝜎
𝑑𝑚𝑎𝑥 = 𝐶𝑎𝑐𝑟𝑖𝑡 𝛾̇ 𝜇 (11)
𝑐

ro
where 𝑑𝑚𝑎𝑥 is the maximum equilibrium droplet diameter, 𝐶𝑎𝑐𝑟𝑖𝑡 is the critical capillary

-p
number, 𝜎 is the surface tension between the dispersed and continuous phases, 𝜇𝑐 is the

continuous phase molecular viscosity and 𝛾̇ is the strain rate, which depends on the velocity
re
gradient.
lP
na
ur
Jo

26
 Journal Pre-proof

of
ro
-p
re
lP
na
ur
Jo

Figure 9. Extensional efficiency of the (a) Kenics mixer (b) Komax mixer (c) Proposed mixer (d)
SMX mixer.

Figure 9 shows the extensional efficiency over a line with coordinates (x = D/6), (y = D/6),

(−0.225 < z < −0.075 m) for various static mixers. The horizontal axis represents the number

of elements. As observed in Figure 9a, the parts between the elements in the Kenics static mixer

cover greater areas for higher values of 𝛼, which means that the parts between two mixing

element groups have a significant effect on the dispersive mixing. For the Komax mixer (Figure

9b), the maximum value of 𝛼 is observed at element 0 and element 2, which corresponds to the
27
 Journal Pre-proof

inlet of the first and third mixing elements, respectively. Three smaller peaks are observed at the

mixing elements 1, 3 and 4, which correspond to the inlet of the second and the outlet of the third

and fourth mixing elements, respectively. The wide ranges of extensional efficiency observed in

Figures 9c and 9d reveal that more complex flow patterns are predicted for the SMX and the new

mixer than those for the Kenics and Komax mixers. From the average extensional efficiency

values shown in Figure 9, it can be deduced that the new static mixer is the most efficient one

concerning dispersive mixing, with the Komax one being second, followed by the SMX and the

of
Kenics mixers. Moreover, the two different types of static mixers (the new one and the Komax

ro
mixer) have almost the same values of average extensional efficiency, but different CoV values.

-p
This means that the two mixers that have the same dispersive mixing, but not necessarily lead to
re
the same distributive mixing. The average value is not the only way to analyze this criterion.

Thus, a high extensional efficiency is required to deform a single-phase fluid or break a drop in a
lP

two-phase fluid.
na

The cumulative volumetric distributions of 𝛼 in Kenics, Komax, SMX and in the new mixer are
ur

shown in Figure 10. These percentages are the volume percentages of the mixers in which the

extensional efficiency is greater than the value in the abscissa. Thus, the extensional efficiency is
Jo

higher than 0.5 for 36% of the Kenics volume, 68% of the Komax volume, 79% of the SMX

volume, and 92% of the new mixer volume. For any value, the Kenics mixer is less efficient than

the Komax and the new mixer, since for any value of 𝛼 the cumulative volumetric percentage is

lower. On the other hand, the SMX mixer is more efficient than the Kenics mixer only if 𝛼 is

below 0.56. Moreover, it can be seen that the new mixer is more efficient than the other static

mixers, except in the range from 0.63 to 0.75, in which its efficiency is slightly lower than that of

the Komax mixer.

28
 Journal Pre-proof

of
ro
-p
Figure 10. Cumulative volumetric percentage in the different static mixers.
re
lP

3.5. Concentration distribution

The CFD results of brine concentration distribution at different cross-sections of the static mixers
na

have been evaluated before and after the mixing elements. Figure 11 shows a sample of the
ur

computed brine concentration distribution at Re = 60 and z = −0.075 m. Both liquids (water

Jo

and brine) enter the first element and split into two (Kenics and Komax) or more (SMX and new

mixer) streams along the surfaces of the elements, which promote mixing compared to the flow

at different sections of an empty tube. By analyzing the brine distribution at various cross

sections, it is found that after the third element the homogenization reaches a satisfactory level

(CoV < 0.05) for the SMX and the new mixer, whereas for the Kenics and Komax mixers the

brine was not uniformly distributed across the mixers subsequent to the fourth element. Overall,

the differences in the mixing intensity of the brine shown in Figure 11 are evidence that the

shape of the mixer geometry significantly affects distributive mixing.

29
 Journal Pre-proof

of
ro
-p
re
lP
na
ur
Jo

Figure 11. Mass fraction distribution of brine at different cross sections for different static mixers
at Re=60: (a) Kenics mixer (b) Komax mixer (c) Proposed mixer (d) SMX mixer.

30
 Journal Pre-proof

of
ro
-p
re
lP
na
ur
Jo

Figure 12. Mass fraction distribution of brine on a diametrically vertical plane in the static mixers
at Re=60: (a) Kenics mixer (b) Komax mixer (c) Proposed mixer (d) SMX mixer.
Figure 12 shows the contour plots of the mass fraction of the brine at Re = 60 along the static

mixers. As observed in Figure 12, two streams from two different boundaries (the inlet of the

main pipe and the injection tube) enter the first element, and the mixing elements change the

direction of the flow. This issue leads to full mixing of the fluids at the end of the SMX and the

new mixer.

31
 Journal Pre-proof

4. Conclusions

A three-dimensional CFD simulation was performed for different static mixers, namely, Kenics,

SMX, Komax and a novel static mixer, for Reynolds numbers ranging from 20 to 160. There are

a number of correlations in the open literature to predict the pressure drop ratio for well-known

static mixers, such as the Kenics one. These correlations have been mainly developed for laminar

and turbulent flow regimes. However, most of these correlations do not consider the geometric

of
parameters, such as the blade thickness and the number of elements, and they are therefore under

great uncertainty. To validate the numerical model, experimental data available for a static mixer

ro
with a baffle shape element (Al-Atabi [50]) and a correlation proposed for the Kenics mixer (Sir

-p
and Lecjaks [51]) have been taken as benchmarks. The pressure drop ratio and the coefficient of
re
variation predicted by the CFD model have been compared with experimental data reported in
lP

the open literature, and a good agreement (2.3% and 6% for Z-factor and CoV, respectively) has

been found. Moreover, some available correlations for Z-factor have been evaluated for a Kenics
na

mixer and the best fit was obtained for the one proposed by Cybulski and Werner [19] for
ur

Re = 20 − 160. After the validation of the CFD model, the computational method was extended

to the other static mixers in order to examine the effect operating and geometric parameters have
Jo

on the hydrodynamics and mixing performances. The velocity field, pressure drop ratio,

coefficient of variation, concentration distribution and extensional efficiency have been

determined at various conditions in different mixers. The proposed static mixer has a much better

performance than the other static mixers. Finally, the following remarks may be drawn from this

study:

1) The Z-factor generated in the new static mixer is about 45% lower than that

corresponding to the SMX mixer under the same conditions.

32
 Journal Pre-proof

2) The mixing efficiency of the new mixer is higher than those of the other static mixers.

Actually, this mixer generates a lower coefficient of variation and higher extensional

efficiency than the other mixers.

3) Two mixers with the same dispersive mixing do not necessarily lead to the same

distributive mixing.

4) The new static mixer is more efficient than the other mixers in deforming fluid (single-

phase) and breaking drops (two-phase) because it creates high extensional efficiency.

of
5) The methodology described here can be applied to simulate and design other mixers in

ro
order to evaluate their mixing performances.

-p
Finally, given that the information about use of static mixers with complex fluids is scarce, the
re
mixing performance of this new mixer will be assessed in a future study involving complex
lP

fluids, such as non-Newtonian and immiscible ones.

na

Nomenclature
ur

AR: aspect ratio, dimensionless

Jo

bl : blade length, (mm)

bw : blade width, (mm)

𝐶𝑎𝑐𝑟𝑖𝑡 : critical capillary number, (-)

CoV: coefficient of variation, dimensionless

D: tube diameter, (mm)

33
 Journal Pre-proof

DAB : molecular diffusivity, (m2 /s)

𝑑𝑚𝑎𝑥 : maximum equilibrium droplet diameter, (m)

F: flow rate ratio, dimensionless

f: friction factor, dimensionless

g: gravity acceleration, (m/𝑠 2 )

of
L: length of the mixer, (mm)

ro
N: number of evaluation points, ‒

P: total pressure, (Pa) -p

re
∆P: pressure drop in a static mixer, (Pa)
lP

∆P0 : pressure drop in an empty tube, (Pa)

na

Re: Reynolds number, ‒

ur

t: blade thickness, (mm)

Jo

𝐕: velocity, (m/s)

̅:
V mean velocity, (m/s)

∇V: velocity gradient, (1/s)

x: transversal coordinate, (mm)

Yi : local mass fraction of the injected fluid at the ith point, ‒

34
 Journal Pre-proof

Ymean : mean mass fraction of the injected fluid, ‒

y: height coordinate, (mm)

Z: pressure drop ratio, dimensionless

z: length coordinate, (mm)

Greek letters

of
α: extensional efficiency

ro
γ: deformation tensor, (1/s)

μ: viscosity, (kg/(m. s))

-p
re
ρ: density, (kg/m3 )
lP

σ: surface tension, (N/m)

na

ω: vorticity tensor, (1/s)

ur
Jo

References

[1] K.J. Myers, A. Bakker, D. Ryan, Avoid Agitation by Selecting Static Mixers, Chem. Eng.

Progress. 93 (1997) 28−38.

[2] T. Bor, The Static Mixer as a Chemical Reactor, Br. Chem. Eng. 16 (1971) 610−612.

[3] J.R. Baker, Motionless mixers stir up new uses, Chem. Eng. Progress. 87 (1991) 32−38.

35
 Journal Pre-proof

[4] R.K. Thakur, Ch. Vial, D.P. Nigam, E.B. Nauman, G. Djelveh, Static Mixers in the Process

Industries-A Review, Chem. Eng. Res. Des. 81 (2003) 787−826.

[5] G. Montante, M. Coroneo, A. Paglianti, Blending of miscible liquids with different densities

and viscosities, Chem. Eng. Sci. 141 (2016) 250−260.

[6] F. Theron, N. Le Sauze, Comparison between three static mixers for emulsification in

turbulent flow, Int. J. Multiph. Flow. 37 (2011) 488−500.

of
[7] P. Pianko-Oprych, Z. Jaworski, CFD modelling of two-phase liquid-liquid flow in a SMX

ro
static mixer, Pol. J. Chem. Technol. 11 (2009) 41−49.

-p
[8] P.D. Berkman, R.V. Calabrese, Dispersion of Viscous Liquids by Turbulent Flow in a Static
re
Mixer, AIChE J. 34 (1988) 602−609.
lP

[9] C.U. Ugwu, J.C. Ogbonna, H. Tanaka, Improvement of mass transfer characteristics and
na

productivities of inclined tubular photobioreactors by installation of internal static mixers, Appl.

Microbiol. Biotechnol. 58 (2002) 600−607.

ur

[10] R. Rakoczy, S. Masiuk, M. Kordas, P. Gradzik, The effects of power characteristics on the
Jo

heat transfer process in various types of motionless mixing devices, Chem. Eng. Process. 50

(2011) 959−969.

[11] S. Rabha, M. Schubert, F. Grugel, M. Banowski, U. Hampel, Visualization and quantitative

analysis of dispersive mixing by a helical static mixer in upward co-current gas–liquid flow,

Chem. Eng. J. 262 (2015) 527−540.

[12] F. Zidouni, E. Krepper, R. Rzehak, S. Rabha, M. Schubert, U. Hampel, Simulation of gas–

liquid flow in a helical static mixer, Chem. Eng. Sci. 137 (2015) 476−486.
36
 Journal Pre-proof

[13] S. Middleman, Fundamentals of Polymer Processing, New York, Mcgraw-Hill. (1977).

[14] E.L. Paul, Design and Scaleup of an Anchorage-dependent Mammalian Cell Bioreactor,

Ann. N. Y. Acad. Sci. 589 (1990) 642−649.

[15] Z. Qing-shi, Ozone Disinfection of Sewage in A Static Mixer Contacting Reactor System on

a Plant Scale, Ozone. Sci. Eng. 13 (1991) 313−330.

[16] N. Martin, C. Galey, Use of Static Mixer for Oxidation and Disinfection by Ozone, Ozone.

of
Sci. Eng. 16 (1994) 455−473.

ro
[17] P.B. Clancy, Treatment of Chlorinated Ethenes in Groundwater with Ozone and Hydrogen

Peroxide, Environ. Prog. 15 (1996) 187−193. -p

re
[18] J.H. Kwong, J.H. Jung, H.D. Lee, Y.S. Park, D.W. Kim, Development of a hydrodynamic
lP

static mixer for mixing chemicals in ballast water treatment systems, J. Water Process Eng. 8
na

(2015) 209−220.

[19] A. Cybulski, K. Werner, Static mixers-criteria for application and selection, Int. Chem. Eng.
ur

26 (1986) 171−180.
Jo

[20] J.P. Gingras, L. Fradette, P. Tanguy, J. Bousquet, Inline Bitumen Emulsification Using

Static Mixers, Ind. Eng. Chem. Res. 46 (2007) 2618−2627.

[21] A.S. Klett, A.M. Payne, M.C. Thies, Continuous-Flow Process for the Purification and

Fractionation of Alkali and Organosolv Lignins, ACS Sustain. Chem. Eng. 4 (2016) 6689−6694.

[22] M. Rauline, P.A. Tanguy, J.M. Le Blevec, J. Bousquet, Numerical Investigation of the

Performance of Several Static Mixers, Can. J. Chem. Eng. 76 (1998) 527−535.

37
 Journal Pre-proof

[23] E.B. Nauman, B.A. Buffham, Mixing in Continuous Flow Systems, Chem. Eng. Commun.

29 (1983) 369−370.

[24] I. Manas-Zloczower, Studies of Mixing Efficiency in Batch and Continuous Mixers, Rubber

Chem. Technol. 67 (1994) 504−528.

[25] S. Shojaee, S.H. Hosseini, A. Rafati, G. Ahmadi, Prediction of the Effective Area in

Structured Packings by Computational Fluid Dynamics, Ind. Eng. Chem. Res. 50 (2011)

of
10833−10842.

ro
[26] S.H. Hosseini, S. Shojaee, G. Ahmadi, M. Zivdar, Computational fluid dynamics studies of

-p
dry and wet pressure drops in structured packings, J. Ind. Eng. Chem. 18 (2012) 1465−1473.
re
[27] S. Liu, A.N. Hrymak, P.E. Wood, Laminar mixing of shear thinning fluids in a SMX static
lP

mixer, Chem. Eng. Sci. 61 (2006) 1753−1759.

na

[28] A. Baumann, S.A.K. Jeelani, B. Holenstein, P. Stossel, E.J. Windhab, Flow regimes and

drop break-up in SMX and packed bed static mixers, Chem. Eng. Sci. 73 (2012) 354−365.
ur

[29] J.M. Zalc, S.E. Szalai, F.J. Muzzio, S. Jaffer, Characterization of Flow and Mixing in an
Jo

SMX Static Mixer, AIChE J. 48 (2002) 427−436.

[30] J.M. Zalc, S.E. Szalai, F.J. Muzzio, Mixing Dynamics in the SMX Static Mixer as a

Function of Injection Location and Flow Ratio, Polym. Eng. Sci. 43 (2003) 875−890.

[31] D.M. Hobbs, P.D. Swanson, F.J. Muzzio, Numerical characterization of low Reynolds

number flow in the Kenics static mixer, Chem. Eng. Sci. 53 (1998) 1565−1584.

38
 Journal Pre-proof

[32] O. Byrde, M.L. Sawley, Optimization of a Kenics static mixer for non-creeping flow

conditions, Chem. Eng. J. 72 (1999) 163−169.

[33] E. Saatdjian, A.J.S. Rodrigo, J.P.B. Mota, On chaotic advection in a static mixer, Chem.

Eng. J. 187 (2012) 289−298.

[34] M. Regner, K. Östergren, C. Träġardh, Effects of geometry and flow rate on secondary flow

and the mixing process in static mixers−A numerical study, Chem. Eng. Sci. 61 (2006)

of
6133−6141.

ro
[35] H.S. Song, S.P. Han, A general correlation for pressure drop in a Kenics static mixer, Chem.

Eng. Sci. 60 (2005) 5696−5704.

-p
re
[36] E. Fourcade, R. Wadley, H.C.J. Hoefsloot, A. Green, CFD calculation of laminar striation
lP

thinning in static mixer reactors, Chem. Eng. Sci. 56 (2001) 6729−6741.

na

[37] P. Joshi, K.D.P. Nigam, E.B. Nauman, The Kenics static mixer: new data and proposed

correlations, Chem. Eng. J. 59 (1995) 265−271.

ur

[38] M. Rafiee, S. Bakalis, P.J. Fryer, A. Ingram, Study of laminar mixing in kenics static mixer
Jo

by using Positron Emission Particle Tracking (PEPT), Procedia–Food Science. 1 (2011)

678−684.

[39] M.K. Singh, P.D. Anderson, H.E.H. Meijer, Understanding and Optimizing the SMX Static

Mixer, Macromol. Rapid Commun. 30 (2009) 362–376.

[40] S.S. Soman, C.M.R. Madhuranthakam, Effects of internal geometry modifications on the

dispersive and distributive mixing in static mixers, Chem. Eng. Process. 122 (2017) 31−43.

39
 Journal Pre-proof

[41] H.-B. Meng, M.-Y. Song, Y.-F. Yu, X.-H. Jiang, Z.-Y. Wang, J.-H. Wu, Enhancement of

Laminar Flow and Mixing Performance in a Lightnin Static Mixer, Int. J. Chem. React. Eng. 15

(2016) 20160112.

[42] H. Meng, F. Wang, Y. Yu, M. Song, J. Wu, A Numerical Study of Mixing Performance of

High-Viscosity Fluid in Novel Static Mixers with Multitwisted Leaves, Ind. Eng. Chem. Res. 53

(2014) 4084−4095.

of
[43] H. Meng, J. Zhu, Y. Yu, Z. Wang, J. Wu, The effect of symmetrical perforated holes on the

ro
turbulent heat transfer in the static mixer with modified Kenics segments, Int. J. Heat Mass

Transfer 99 (2016) 647–659.

-p
re
[44] M. Zhou, D. Bai, Y. Zong, L. Zhao, J.N. Thornock, Numerical investigation of turbulent

reactive mixing in a novel coaxial jet static mixer, Chem. Eng. Process. 122 (2017) 190−203.
lP

[45] S. Jegatheeswaran, F. Ein-Mozaffari, J. Wu, Process intensification in a chaotic SMX static

na

mixer to achieve an energy-efficient mixing operation of non-newtonian fluids, 124 (2018) 1−10.
ur

[46] S. Jegatheeswaran, F. Ein-Mozaffari, J. Wu, Laminar mixing of non-Newtonian fluids in

Jo

static mixers: process intensification perspective. Rev. Chem. Eng. 2019, doi:10.1515/revce-

2017-0104.

[47] O. Mihailova, D. O’Sullivan, A. Ingram, S. Bakalis, Velocity field characterization of

Newtonian and non-Newtonian fluids in SMX mixers using PEPT, Chem. Eng. Res. Des. 108

(2016) 126-138.

40
 Journal Pre-proof

[48] Z. Jaworski, P. Pianko-Oprych, D. L. Marchisio, A.W. Nienow, CFD Modelling of

Turbulent Drop Breakage in a Kenics Static Mixer and Comparison with Experimental Data,

Chem. Eng. Res. Des. 85 (2007) 753-759.

[49] M.M. Haddadi, S.H. Hosseini, D. Rashtchian, G. Ahmadi, CFD Modeling of Immiscible

Liquids Turbulent Dispersion in Kenics Static Mixers: Focusing on Droplet Behavior, Chines J.

Chem. Eng. 2019, Submitted to Journal.

of
[50] M. Al-Atabi, Design and Assessment of a Novel Static Mixer, Can. J. Chem. Eng. 89 (2011)

ro
550−554.

-p
[51] J. Sir, Z. Lecjaks, Pressure Drop and Homogenization Efficiency of a Motionless Mixer,
re
Chem. Eng. Commun. 16 (1982) 325−334.
lP

[52] R.B. Bird, W.E. Stewart, E.N. Lightfoot, Transport Phenomena, New York. John Wiley &

Sons. (2002).
na

[53] M. Coroneo, G. Montante, A. Pagliant, Computational Fluid Dynamics Modeling of

ur

Corrugated Static Mixers for Turbulent Applications, Ind. Eng. Chem. Res. 51 (2012)
Jo

15986−15996.

[54] V. Kumar, V. Shirke, K.D.P. Nigam, Performance of Kenics static mixer over a wide range

of Reynolds number, Chem. Eng. J. 139 (2008) 284−295.

[55] W.L. Wilkinson, M.J. Cliff, An Investigation into the Performance of a Static In-Line

Mixer, In Second European Conference on Mixing. (1977).

[56] N.F. Shah, D.D. Kale, Two-Phase Gas-Liquid Flows in Static Mixers, AIChE J. 38 (1992)

308−310.
41
 Journal Pre-proof

[57] C.D. Grace, Static Mixing and Heat Transfer, Chem. Process. Eng. 52 (1971) 57−59.

[58] S.H. Hosseini, G. Ahmadi, M. Olazar, CFD simulation of cylindrical spouted beds by the

kinetic theory of granular flow, Powder Technol. 246 (2013) 303−316.

[59] C. Guenther, M. Syamlal, The effect of numerical diffusion on simulation of isolated

bubbles in a gas–solid fluidized bed, Powder Technol. 116 (2001) 142−154.

[60] M.D. Das, A.N. Hrymak, M.H.I. Baird, Laminar liquid-liquid dispersion in the SMX static

of
mixer, Chem. Eng. Sci. 101 (2013) 329−344.

ro
Captions: -p
re
Table 1. Geometrical parameters of the static mixers.
lP

Table 2. Properties of the fluids.

na

Table 3. Grid size, CFD simulation time, number of elements, estimated Z and CoV for the

Kenics mixer at Re=160

ur

Table 4. Grid size, CFD simulation time, number of elements, estimated Z and CoV for the
Jo

Komax mixer at Re=160

Table 5. Grid size, CFD simulation time, number of elements, estimated Z and CoV for the SMX

mixer at Re=160

Table 6. Grid size, CFD simulation time, number of elements, estimated Z and CoV for the new

mixer at Re=160

Table 7. Correlations of Z-factor for Kenics static mixers.

42