Separation and Purification Technology 168 (2016) 114123

Contents lists available at ScienceDirect

Separation and Purification Technology

journal homepage: www.elsevier.com/locate/seppur

Analysis of the effects of temperature and the share of solid and gas
phases on the process of separation in a cyclone suspension preheater
Marek Wasilewski
Faculty of Production Engineering and Logistics, Opole University of Technology, 76 Proszkowska St., 45-758 Opole, Poland

a r t i c l e i n f o a b s t r a c t

Article history: Most scientific papers discussing cyclone separators do not consider the effects of temperature or the
Received 9 February 2016 share of the solid and gas phases on the efficiency of such devices. This article analyses the effect of these
Received in revised form 12 May 2016 factors on the efficiency of solid particle separation, based on a cyclone separator that constitutes part of
Accepted 27 May 2016
a suspension preheater. The operating conditions for this device allowed for a wide range of temperatures
Available online 27 May 2016
to be taken into account. The main research procedure was preceded by exploitation research conducted
at an industrial installation. This enabled the correct reconstruction of the cyclones actual operating con-
Keywords:
ditions. The multiphase flow inside the cyclones was analysed with computational fluid dynamics (CFD).
Separation efficiency
Cyclone separator
Reynolds-averaged NavierStokes equations with the Reynolds stress turbulence model (RSM) were used
Modelling thermal and flow phenomena in the analysis.
Suspension preheater The methods used in this paper for analysing the phenomena occurring inside the device allow for the
Clinker burning conclusion that both the temperature and the concentration of solid particles have a significant effect on
the effectiveness of a cyclone separator. The presented analysis may help to optimise the cyclones used in
cyclone suspension preheaters, as well as traditional devices used solely for separation processes.
2016 Elsevier B.V. All rights reserved.

1. Introduction obtained under stable environmental conditions. The majority of

Contemporary industry makes use of many types of cyclone
separators due to their universal applicability, simple structure
and low exploitation costs. However, no universal cyclone (appli-
cable to any operating conditions) has so far been developed. This
is because some devices have a high separation efficiency; whereas
others have low exploitation costs (with a minimal pressure drop).
The subject literature includes many classifications of cyclones
depending on their structure [14].
While cyclones are primarily used as typical particulate matter
separation devices, they are also often used in technological pro-
cesses (drying, heating and decarbonisation) that usually involve
a high temperature. In such cases, special attention should be paid
to the effect of the temperature on the efficiency of the cyclone
separator. Furthermore, the share of the solid phase is usually high
when compared to the share of the gas phase, due to the fact that
the flow of the solid phase directly determines the output of a
given installation.
The proposed empirical and semi-empirical models for deter-
mining the pressure drop and the efficiency of the particulate sep-
aration were developed based on experimental data that were
E-mail address: m.wasilewski@po.opole.pl
experimental studies are also conducted under isothermic and iso-
baric conditions. The same is true for studies that use a CFD, most
of which do not take into account changes in the temperature and
pressure. In turn, exploitation data obtained from real industrial
installations that operate within a wide temperature range often
show considerable discrepancies when compared to computa-
tional studies.
Some researchers have already addressed this subject. For
example, Bohnet and Lorenz [5,6] noted a considerable effect of
temperature on the efficiency of particle separation and on a drop
in pressure. They concluded that the decrease in the pressure drop
was due to an increase in the temperature, and this could be
described by the Meissner equation. Also, the efficiency of the sep-
aration was found to decrease with an increase in the temperature.
Patterson [7] observed a similar effect of temperature on the effi-
ciency of cyclone separators. He also found that a decrease in the
gas density and an increase in the gas viscosity combined with
an increase in the temperature had a considerable effect on the
results. Gimbun and Chuah [8,9] confirmed these findings using
CFD. They also compared the computational models presented by
Shepherd and Lapple [10], Dirgo [11], Coker [12] and Casal and
Martinez [13]. They indicated that the Shepherd and Lapple model
was the most accurate for temperatures in the operating medium

http://dx.doi.org/10.1016/j.seppur.2016.05.033
1383-5866/ 2016 Elsevier B.V. All rights reserved.
 M. Wasilewski / Separation and Purification Technology 168 (2016) 114123
Nomenclature
a height of the gas inlet, m
b width of the gas inlet, m
B diameter of the cyclone lower (dust) outlet, m
CD drag coefficient
CFD computational fluid dynamics
dp diameter of a particle, m
D cyclone body diameter, m
De diameter of the cyclone gas outlet, m
DPM discrete phase model
DDPM dense discrete phase model
Dij the stress diffusion term
Fk momentum transport coefficient, t
1
g acceleration of gravity, m s
2
Gi mass flow of the particle (solid phase), kg s
1
h height of the cyclone cylindrical section, m
H total height of the cyclone, m
k turbulence kinetic energy, m2 s
2
N-S Navier-Stokes equations
P pressure, Pa
Pij the shear production term
PRESTO Pressure Staggering Option
p0 dispersion pressure, Pa
Qi volumetric flow rate of the gas phase, m3 s
1
RANS Reynolds average NavierStokes
over 500 K; and that the Coker model was best suited to tempera-
tures below 500 K. Bohnets conclusions were also confirmed in an
experimental study by Li and Chen [14]. In turn, Karagoz and Kaya
[15] investigated the heat exchange between the solid phase and
the gas phase. They noted that the heat exchange increased with
an increase in the inlet velocity on all surfaces. Gupta [16] stated
that the heat transfer coefficient increased with an increase in
the solid substance load and the inlet velocity. Other examples
[1724] offer studies aimed at determining the effect of tempera-
ture on the operation of cyclone separators.
Another important factor in particulate separation using
cyclones is the ratio between the gas and solid phases. If the solid
phase concentration is high, a secondary flux on the walls of the
cyclone will have a significant effect on its efficiency [25,26]. For
instance, Stern [27] noted that the total efficiency of the separation
increased by about 35% with a tenfold increase in the particle con-
centration. Wheeldon and Burnard [28] also found that the effect of
high particle concentration improved the efficiency of the frac-
tional separation in a cyclone. It is worth underlining the fact that
most studies conducted so far (regardless of their methodology)
have been performed with a low share of the solid phase, i.e. up
to 100 g/m3.
A cyclone suspension preheater that constitutes part of an
installation for dry clinker burning is an example of a device in
which operating conditions change within a wide temperature
range (6001200 K). The operation of a cyclone suspension pre-
heater is similar to that of a traditional suspension heat exchanger.
Modern suspension preheater systems consist of four to six stages.
Cyclones of the subsequent stages of the suspension preheater are
designed based on their functions and temperature conditions.
Cyclones used for clinker burning are used not only as separa-
tors but are also primarily used as regulators for the technological
process (decarbonisation), which means that the solid phase flux
directly determines the output. Thus, we are dealing with a chang-
ing ratio of the solid and gas phases depending on the heat effi-
ciency of the rotary kiln. Today, the share of the solid phase in
cement plants across the world ranges from 300 g/m3 to 900 g/
m3 of the gas phase.
115
Re Reynolds number
RSM Reynolds stress model
s the source term
S height of the outlet duct in the interior of the cyclone, m
SIMPLE semi-implicit method pressure-linked equations
Tfi temperature of the gas phase (i - zone of the research
system)
Tpi temperature of the particle solid phase (i - zone of the
research system)
t time, s
u gas velocity, m s
1
u0 i (j, k) fluctuating velocity to direction i (j, k)
up particle velocity, m s
1
ag angle of the cyclone inlet head, degree
DP pressure drop in a cyclone separator, Pa
d Kronecker factor
eij the dissipation term
l dynamic viscosity of gas, kg m
1s
1
Pij the pressure-strain correlation term
q density of gas, kg m
3
qp density of a particle, kg m
3
sij the Reynolds stress tensor, m2 s
2
An analysis of the subject literature leads to the conclusion that
most studies conducted so far on the development and construc-
tion of cyclone separators have focused on traditional cyclone sep-
arators; while few have addressed the effects of temperature and
the share of each phase on the efficiency of the cyclones operation.
There are few studies on the topic of cyclones used for clinker
burning. Examples of this type of research are studies [2935].
2. Materials and methods
The main aim of this study was to assess the effects of the tem-
perature and the shares of the solid and gas phases on the separa-
tion process, based on a cyclone suspension preheater used for
clinker burning. As the actual operating conditions had to be sim-
ulated as accurately as possible, a series of measurements were
conducted prior to the research on a production installation oper-
ating in a cement plant in Poland.
2.1. Explanation of the research conditions
The installation comprised a four-stage cyclone suspension pre-
heater and a rotary kiln. The first-stage of the suspension preheater
was equipped with four cyclone separators, while the subsequent
stages operated as dual systems. The installation also had an addi-
tional burning chamber, i.e. a calciner. At its maximal efficiency,
the installation processes amounts of about 3500 tonnes/day.
Table 1 presents an overview of the measured values for the tem-
peratures and pressures for each factor at each stage. The values for
first-stage cyclones are given in italic.
In order to determine the efficiency of the raw material separa-
tion by the first-stage cyclones, the mass flow rate of the raw mate-
rial, the volumetric flow rate of the gases, and the release from the
suspension preheater were measured. A mass balance was pre-
pared based on the obtained information. The collected data can
be found in Table 2.
Given that the efficiency of a cyclone suspension preheater pri-
marily depends on the first-stage cyclones, subsequent research
116 M. Wasilewski / Separation and Purification Technology 168 (2016) 114123

Table 1
Average values of the measurement data for each production stage.

No. Suspension preheater degree Gas temperature [K] Pressure drop [Pa] Raw material temperature [K]
Value Difference between degrees Value Difference between degrees Value Difference between degrees
1 Lift chamber 1211
680
2 Calciner 1108 103
1034
354 1097
3 IV 943 165
2430
1396 920 177
4 III 864 79
3364
934 817 103
5 II 700 164
4815
1451 646 171
6 I 623 76
6005
1190 340 306

Table 2
Mass balance of the installation.

Unit Value
Mean value from Mean value from Mean value from Mean value from Mean value from Mean value
measurement 1 measurement 2 measurement 3 measurement 4 measurement 5
Mass flow rate of the raw material kg/h 240,000 240,000 240,000 240,000 240,000 240,000
Rotary kiln efficiency kg/h 136,364 136,364 136,364 136,364 136,364 136,364
Volumetric flow rate of the gases m3/h 509,192 511,720 524,900 512,703 538,940 519,491
Release from the suspension preheater kg/h 22,030 24,865 23,779 24,853 24,322 23,970
Separation efficiency first-stage cyclones % 90.82 89.64 90.09 89.64 89.87 90.01

was conducted on the cyclones operating in the first-stage of the
suspension preheater. The role of the first-stage cyclones is to
retain as much of the raw material as possible inside the heat
exchanger tower (i.e. to minimise the release), which is why their
separation efficiency should be as high as possible.
The gas phase was a mixture of the exhaust gases created by
burning fuel in the rotary kiln and the process gases. In turn, the
solid phase in clinker burning is raw material whose main compo-
nent is CaCO3.
2.2. Numerical study
Computational fluid dynamics were used for the research anal-
ysis. This method is commonly used for assessing the efficiency of
cyclone separators. The subject literature indicates that the
method is effective in research conducted to determine both pres-
sure drop and separation efficiency. Some researchers [36] use the
complementary k-e RNG model as an extension of the RANS model.
The k-e RNG model provides satisfactory consistency between CFD
results and experimental results. The growing computational
power of computers has enabled the use of the RSM model, which
helps to precisely model highly anisotropic turbulent stresses. It is
currently the most frequently applied complementary model in
CFD research on cyclone separators.
The following are examples of studies that have used this
model. Wang et al. [37] applied the RSM model to analyse flows
in a Lapple cyclone. Wan et al. [38] investigated the effect of the
concentration of particles with different diameters on the effec-
tiveness of cyclones. Chu et al. [39,40] analysed the flow of solid
particles and the gaseous phase in dense medium cyclones, e.g.
cyclones used to process coal. Raoufi et al. [41] and Su et al. [42]
used the RSM model to investigate the flow in cyclones with an
unconventional structure, e.g. square cyclones. Safikhani et al.
[43] analysed the effectiveness of cyclones with three different
structures (1D3D, 2D2D and 1D2D). Their CFD results were highly
consistent with experimental results. Parihar et al. [44] proposed a
new cyclone structure based on the combined results of both CFD
and experimental research. Kepa [45,46] investigated the effect of
an additional element (a counter-cone) in the area of the lower
outlet of a cyclone. He found that the counter-cone limited the
speed at which gas swirl floes in the dust tank and prevented the
reverse movement of solid particles. Wu et al. [47] conducted a
study on the tangential velocity indicator (n) of the free vortex in
a cyclone. They proposed a new method for determining the n indi-
cator. Liu et al. [48] analysed the effectiveness of parallel cyclones
and found that such a solution leads to better separation efficiency
and a lower pressure drop compared to individual cyclones. Oh
et al. [49] characterised a particle separator that is constructed
based on the concept of a uniflow cyclone. Brar et al. [50] investi-
gated the effect of the height of a cyclone on its efficiency using a
Stairmand high-efficiency cyclone. They concluded that increasing
the height of a cyclone improves all parameters related to effi-
ciency. A study conducted by Wakizono et al. [51] led this group
to suggest installing an additional, ring-shaped element in the area
of the upper outlet of the cyclone, which, they stated, would
improve separation efficiency and reduce pressure drop. Lastly,
Houben and Pirker [52] conducted research in which they com-
bined the k-e and RSM models. They applied the k-e model during
the initial stage of simulation (in the first steps of iteration) in
order to stabilise the flow. They also proposed in the same paper
a new structural solution, i.e. a central rod, which stabilises the
vortex and prevents it from precessing around the cyclones axis
and from prematurely ending at the wall. A full review of the lit-
erature also indicates that some researchers have used the Large
Eddy Simulation (LES) model [53], which also requires highly effi-
cient computational units.
As more than 80% of the heat exchange takes place in the ducts
between the cyclones, and only about 20% takes place in the cyclones
themselves [54], the cyclone model was expanded with ducts to feed
the gases and raw material to the first-stage cyclones (Fig. 1). Most
studies presented in the subject literature do not address the feeding
method or the reactions occurring at this stage. However, this is an
important aspect, as measurements of the temperature in active
installations can often be inaccurate due to the mixing of the solid
and gas phase flows. Consequently, this research was based on input
data that could be measured more accurately (the temperature of
the raw material before it entered the gas ducts and the temperature
of the gases before they reached the raw material feeding point). In
combination with CFD, this allowed for an estimation of the temper-
atures of both phases at the inlet to the cyclone (where the raw
material had already been heated up in the gas ducts, and some of
the gas phase heat had dispersed).
The full geometric model of the first-stage of the suspension
preheater was then divided into three zones. In Zone 1, the
 M. Wasilewski / Separation and Purification Technology 168 (2016) 114123 117

continuity equation for the momentum. A second-order upwind

interpolation was used in order to determine the representative
samples of the constituent values on the surface of the control vol-
umes. Given that the flow inside the cyclone is turbulent, the RANS
model was used to model the form of this flow. Using the RANS
method required a closing hypothesis, referred to as a detailed
model, to be formulated. The aforementioned assumptions and
simplifications enabled the use of the Reynolds Stress Model
(RSM), a complex turbulent model integrated with a heat exchange
algorithm.
The presence of the solid phase was modelled using the Euler-
Lagrange method, applied through the ANSYS Fluent 14 software
as a discrete phase model (DPM). This method treats a fluid as a
continuous phase (where the calculations involve solving an N-S
equation); whereas the dispersed phase is identified through par-
ticle tracking (where the flow field is calculated).
Table 3 provides detailed information about the established
conditions for the CFD modelling.
The following assumptions were made based on the literature
[57,58] to optimise the construction of the computational mesh:

 a variable density of the computational mesh was used in order

to incorporate the areas where the highest disturbances in the
flow occurred;
 when selecting the density of the computational mesh, the min-
imal number of cells for which no significant changes in the
results occurred was taken into account;
 the maximal skewness did not exceed 0.85;
 the mean skewness did not exceed 0.25;
Fig. 1. Full geometric model of the first-stage of the suspension preheater.  the aspect ratio did not exceed 15;
 the dimensions of neighbouring cells did not differ by more
operating medium is introduced into the cyclone. Zone 2 consti- than 20%.
tutes the first-stage cyclone, and was thus the most pertinent area
Table 3
for this study. Zone 3 constitutes the ducts where the purified
Details of CFD settings.
gases are fed into the electrofilters, and this zone was omitted from
the study. Turbulence model: RANS
@ui (1)
Since research based on CFD often requires the introduction of @xi 0
h
i
@uj @ sij (2)
certain simplifications and the use of models burdened with uncer- q uj @u
@xj
i
@P

@xi
@
@xj l @ui
@xj @xi @xj
tainty, the study was divided into two stages. In the first (initial) sij
q u0i u0j (3)
stage, the temperature of both phases at the inlet to the cyclone [47]
was determined based on the temperature at the outlet of Zone Detailed model: RSM, Adopted standard numerical coefficients rk, Cl, re, Ce1, Ce2, k
1. In the second (main) stage of the study, the effect on the separa- @
qu0i u0j @x@ k quk u0i u0j Dij Pij Pij eij s (4)
@t
tion efficiency of the temperature of the gas phase and the shares h
i
@ 0 0 0 0 0
Dij
@xk qui uj uk p uj dik p ui djk
l @xk ui uj
0 0 @ 0 0 (5)
of each phase was determined. The study was conducted on a sin- h
@u
i
P
q u0 u0 j u0 u0 @ui (6)
gle cyclone (a simplification was introduced due to the fact that the ij i k @xk j k @xk


first-stage of the suspension preheater in the industrial installation @ui @u (7)
Pij p0 @xj @xij
presented in Section 2 consists of four cyclones with the same @u0i @u0i (8)
eij
2l @xk @xk
structure). The flows of the gas and the solid phases were divided
[37]
proportionally.
Two-phase flow: Discrete phase model
The research was conducted for the following inlet tempera-
18l qp d2p CD Rep (9)
tures: 330 K, 400 K, 500 K, 600 K, 700 K, 800 K and 900 K. The inlet Fk  24
qp d2p
8
velocity into the first-stage cyclone remained constant and >
< Re for Rep 6 1
24 (10)
2410;15Rep0;687 for1<Rep 61000
resulted from the real volumetric flow of the gas phase. Further- CD
>
: Rep
more, an attempt was made during Stage II of the research to 0; 44 for Rep > 1000
dp qjup
uj (11)
determine the effect of the shares of each phase on the efficiency Rep l
of the raw material separation by using three mass flow rates (in [55,56]
Zone 2): 100%, 10% and 1% of the real value. The walls of the cyclone The standard walls
CFD analysis was performed with the ANSYS Fluent 14 program function
using the finite volumes method. ANSYS Fluent 14 is a universal Pressure-Velocity Coupling SIMPLE
Terms of convergence: Continuity 10
6 for Zone 1; 10
5
tool for the numerical analysis of single-phase and multiphase flow
for Zone 2
phenomena involving chemical reactions and exchanges of heat Terms of convergence: Other 10
6 for Zone 1; 10
3
and mass. The pressure-based segregated solver method within for Zone 2
the program was selected for the analysis. Differential equations Spatial discretization: Pressure PRESTO
were solved using the Semi-Implicit Method for Pressure-Linked Spatial discretization: Momentum, Turbulent kinetic Second order
energy, Turbulent dissipation rate, Reynolds upwind
Equations (SIMPLE) algorithm to accurately determine the cou- stresses
pling of the pressure and velocity fields while maintaining the
118 M. Wasilewski / Separation and Purification Technology 168 (2016) 114123

phase temperatures were analysed for the individual diameters

3. Results and discussion
During Stage I of the research, the mean temperatures of the gas
and solid phases at the inlet to the first-stage cyclone were deter-
mined using a geometric model of the ducts feeding both phases
into the cyclone (Zone 1 in Fig. 1). The following temperatures of
the gas phase were used for the calculations: 330 K, 400 K, 500 K,
600 K, 700 K, 800 K and 900 K; and the temperature of the solid
phase before it entered the system was 330 K. The values for the
volumetric flow rate of the gases and the mass flow rate of the
raw material were defined based on the balance (Table 2). Table 4
presents the obtained results. Furthermore, mean values for solid
of the solid at the outlet from Zone I. The temperature of the solid
phase was found to decrease with an increase in particle diameter.
For instance, the mean temperature of particles with the lowest
diameter (dp = 15 lm) was 673 K. In turn, particles with the high-
est diameter (dp = 500 lm) had a mean temperature of 589 K. This
enables the conclusion that the fractional distribution of solid par-
ticles has a considerable effect on the temperature of each phase.
The obtained temperature values for the gas and solid phases
enabled the transition to the key phase of the conducted study,
involving the determination of the effects of the temperature and
the share of each phase on the particulate separation efficiency.
The values presented in Table 4 were introduced into the research

Table 4
Results of Stage I of the research.

Volumetric flow Mass flow of the Temperature of the gas Temperature of the solid Mean temperature of the gas Mean temperature of the solid
rate of the gases raw material phase at the inlet (Zone phase at the inlet (Zone phase at the inlet to the cyclone phase at the inlet to the cyclone
[m3/s] [kg/s] 1) [K] 1) [K] (Zone 2) [K] (Zone 2) [K]
Qi Gi Tf1 Tp1 Tf2 Tp2
36.1 16.7 330 330 330 330
400 330 387 383
500 330 469 458
600 330 551 536
700 330 634 611
800 330 716 681
900 330 798 756

Fig. 2. Geometry of the cyclone separator.

 M. Wasilewski / Separation and Purification Technology 168 (2016) 114123 119

system that constitutes Zone 2, i.e. the model of a cyclone. Fig. 2
presents the geometry of the studied cyclone (where D = 3.45 m;
a = 2 m; b = 0.98 m; B = 0.55 m; De1 = 2 m; De1 = 1.75 m;
H = 9.35 m; h = 5.9 m; S = 2 m; ag = 180).
Fig. 3 presents the obtained values for the total separation effi-
ciency of the solid according to the inlet temperature of the
cyclone separator. The three curves in the figure represent the
separation efficiency for the three values of the mass flow of the
solid phase.
The analysis of the obtained results indicated that the efficiency
of the solid phase separation decreased with an increase in the
inlet temperature. This effect is primarily a result of a decrease
in the density and an increase in the viscosity of the gas. Higher
separation efficiency values were obtained for the flow of raw

Fig. 3. Total separation efficiency of the solid phase according to inlet temperature.

Fig. 4. Separation efficiency of the solid phase according to particle diameter for the 660 lm fraction (Gi = 16.7 kg/s).
120 M. Wasilewski / Separation and Purification Technology 168 (2016) 114123

Fig. 5. Example trajectories with the temperature for the solid particles of each fraction (Tf2 = 634 K; Tp2 = 611 K; and Gi = 16.7 kg/s).

material amounting to 1.67 kg/s (10% of the real value) than for the
flow amounting to 0.167 kg/s (1% of the real value), which indi-
cates that the separation efficiency increased with an increase in
the solid phase load within the system. This tendency was con-
firmed by the results obtained after the flow of the solid phase
was increased further (to 16.7 kg/s; i.e. 100% of the real value). This
is most likely due to the fact that fine particles may agglomerate
into clusters, which is a common phenomenon in cyclone separa-
tors operating at a high temperature in the gaseous phase and
under a high load in the solid phase.
It is also worth noting that higher inlet temperatures corre-
sponded to a higher decrease in the efficiency, as was indicated
by comparing the sections of the curves with a temperature up
to about 700 K (exhibiting a gentle drop) with those beyond
700 K (exhibiting a rapid drop). Li and Chen [14] noticed a similar
tendency in their study. This may be the result of a sudden increase
in the viscosity of the gaseous phase within the aforementioned
temperature range.
Analysis of the separation efficiency of the individual solid
phase fractions proved necessary in order to fully investigate the
effect of temperature and the share of each phase. Such an analysis
is especially important if the share of fine particles (particles with a
diameter of less than 20 lm) is high, as it is with raw material used
for burning cement clinker. Fig. 4 shows the results of the analysis.
For clarity, the figure includes only particles with a diameter of
660 lm. For larger fractions, separation efficiency always
amounted to 100%. The obtained values indicate that the separa-
tion efficiency of the finest particles (dp = 15 lm) was of crucial
 M. Wasilewski / Separation and Purification Technology 168 (2016) 114123
Fig. 6. Temperature contours of gas inside the cyclone (Tf2 = 634 K, Tp2 = 611 K,
Gi = 16.7 kg/s).
importance, as the greatest differences due to an increase in inlet
temperature were noted for this specific diameter. Other solid par-
ticle fractions showed only slight changes.
The information provided above could be especially useful
when managing clinker burning. Those responsible for controlling
the process could use this information to adjust temperatures
inside the cyclone separator (within a limited range, of course).
Reducing the temperature will improve separation efficiency and,
consequently, enable a greater amount of the raw material to
remain inside the installation, leading to increased output. Natu-
rally, the individual technological conditions of a given installation
should be taken into account each time (e.g. the heat efficiency of
the rotary kiln, the properties of flowing materials or flow param-
eters). Actions can also be taken to decrease inlet temperatures in
traditional cyclone separators used solely for separation, and this
will improve separation efficiency as well.
Information about separation efficiency of the individual solid
phase fractions can also create a foundation for optimising separa-
tion in cyclones used for clinker burning. Analysis of separation
efficiency of particles of each diameter can affect the preparation
of raw material, i.e., grinding. Those responsible for making deci-
sions can use the conclusions presented in this article to optimise
the grinding curve in order to decrease the share of fine particles.
Decreasing the share of such particles even by a few percentage
points can bring considerable benefits considering the scale of
cement production.
Heat exchange between both phases was also analysed. The first
stage of heat exchange occurs directly beyond the inlet to a
121
Fig. 7. Temperature contours of gas at the outlet of the cyclone (Tf2 = 634 K,
Tp2 = 611 K, Gi = 16.7 kg/s).
cyclone, where particles directed into the cyclone undergo intense
heating. The particles reach the highest temperature values in the
cylindrical part and at the outlet from the cyclone. This especially
concerns low-diameter particles, as confirmed by example trajec-
tories together with the temperature value of solid particles for
each fraction (Tf2 = 634 K, Tp2 = 611 K, Gi = 16.7 kg/s), shown in
Fig. 5. This figure helps illustrate the movement of the particles
and changes in their temperature that occur due to their interac-
tion with the gas flux. Also notable is the fact that raw material
particles form spiral-shaped clusters that are transported down-
ward inside the cyclone. During this downward movement, the
particles travel along a spiral path comprising several turns near
the walls. The number of turns depends on the diameter of the par-
ticles and usually ranges from 2 to 4 (the higher the diameter, the
lower the number of turns).
Fig. 6 presents temperature contours of the gaseous phase in
the cross-section (Tf2 = 634 K, Tp2 = 611 K, Gi = 16.7 kg/s). Note that
the lowest values for temperature were observed in the conical
part. This was because, during the gaseous phase, some of its
enthalpy had dissipated before reaching the lower area. The tem-
perature in the axis of the cyclone is also low due to the gas revers-
ing its flow in the lower area, thus creating an upward vortex.
Fig. 7 presents the obtained mean value for the gas temperature
(628 K) at the outlet duct from the cyclone. This value forms the
basis for the validation of the obtained results. The mean value
was only slightly higher than the real measurement value obtained
during the research involving the industrial installation (623 K;
Table 1). This indicates that the assumptions made in the research
process were correct and that the computational process was para-
metrised appropriately.
4. Conclusions
The conducted research allows for the following conclusions to
be drawn:
 The total dust separation efficiency of the cyclone decreases
with an increase in the temperature. This is especially true for
higher temperatures (from about 700 K), where the decrease
is much more rapid when compared to lower temperatures.
122 M. Wasilewski / Separation and Purification Technology 168 (2016) 114123
 The total separation efficiency increases with an increase in the
mass flow of the solid phase.
 The main factors that affect the decrease in the separation effi-
ciency are decreases in the density and increases in the viscosity
of the gas phase.
 The decrease in the separation efficiency becomes less rapid
with an increase in the mass flow of the solid phase, especially
at higher temperatures.
 In the case of cyclones used solely for particulate separation, it
is beneficial to conduct the process at the lowest possible
temperature.
 The validation of the value of the outlet temperature leads to
the conclusion that the numerical models used in the experi-
ment were effective and that the computational and boundary
conditions were appropriately parametrised.
 The obtained values for the separation efficiency of a real mass
flow in the solid phase and the real inlet temperature in the
first-stage cyclones were slightly lower than the real values. This
may be due to the limitations of the numerical model used to rep-
resent the movement of the solid phase, as the model did not take
into account the agglomerations or the collisions between parti-
cles. In view of the above, attempts should be made in the future
to expand the model by using a hybrid Euler-Lagrange model,
which would also take into account collisions between the parti-
cles. The latter model can be applied in the Fluent program as a
dense discrete phase model (DDPM); however, it requires a com-
puter with a greater computational power.
 The conducted analysis showed that CFD research methods are
effective tools for modelling heterogeneous flows in cyclones,
which confirms the theory that the application of this method
may help to optimise the cyclones used in clinker burning, as
well as traditional devices used for separation processes.
 Detailed knowledge of the phenomena occurring inside a
cyclone suspension preheater and the conditions therein during
clinker burning may allow for the process to be intensified, and
may lead to an increase in the output and a decrease in the
energy consumption during the process.
References
[1] A.C. Hoffmann, L.E. Stein, Gas Cyclones and Swirl Tubes, Springer-Verlag, 2002.
[2] C.C. Cooper, G.C. Alley, Air Pollution Control: A Design Approach, Prospect
Heights, Waveland Press, Inc. Hinds, IL, 1994.
[3] C. William, Aerosol Technology: Properties, Behavior, and Measurement of
Airborne Particles, second ed., John Wiley & Sons, New York, 1999.
[4] J. Warych, Purification of the Waste Gases Industrial, Scientific and Technical
Publishing, 1994 (in Polish).
[5] M. Bohnet, Influence of the gas temperature on the separation efficiency of
aerocyclones, Chem. Eng. Process. 34 (1995) 151156.
[6] M. Bohnet, T. Lorenz, Separation Efficiency and Pressure Drop of Cyclones at
High Temperatures, Gas Cleaning at High Temperatures, Springer, 1993, pp.
1731.
[7] P.A. Patterson, R.J. Munz, Cyclone efficiencies at very high temperatures, Can. J.
Chem. Eng. 67 (1989) 321328.
[8] J. Gimbun, J. Chuah, T. Fakhrul-Razi, A. Choong, The influence of temperature
and inlet velocity on cyclone pressure drop: a CFD study, Chem. Eng. Process.
44 (2005) 712.
[9] T. Chuah, J. Gimbun, T.S. Choong, A CFD study of the effect of cone dimensions
on sampling aerocyclones performance and hydrodynamics, Powder Technol.
162 (2006) 126132.
[10] C. Shepherd, C. Lapple, Flow pattern and pressure drop in cyclone dust
collectors, Ind. Eng. Chem. 31 (1939) 972984.
[11] J. Dirgo, Relationships between cyclone dimensions and performance Doctoral
Thesis, Harvard University, USA, 1988.
[12] A.K. Coker, Understand cyclone design, Chem. Eng. Process. 28 (1993) 5155.
[13] J. Casal, J.M. Martinez, A better way to calculate cyclone pressure drop, Chem.
Eng. 90 (1983) 99.
[14] W. Li, J. Chen, Experimental research on cyclone performance at high
temperature, Front. Mech. Eng. China 2 (2007) 310317.
[15] I. Karagoz, F. Kaya, CFD investigation of the flow and heat transfer
characteristics in a tangential inlet cyclone, Int. J. Heat Mass Transf. 34
(2007) 11191126.
[16] P.K. Gupta, Prediction of heat transfer coefficient in the cyclone separator of a
CFB, Int. J. Energ. Res. 24 (2000) 10651079.
[17] R.J. Wakeman, Progress in Filtration and Separation, Elsevier, Amsterdam,
1981.
[18] M. Ernst, R.C. Hoke, V.J. Siminski, R. Parker, D.C. Drehmel, Evaluation of a
cyclone dust collector for high-temperature high-pressure particle control,
Ind. Eng. Chem. Proc. Des. Dev. 21 (1982) 158161.
[19] J.G. Bernard, J. Andries, B. Scarlett, B. Pitchumani, Cyclone performance at high
temperatures and pressures, in: Proceedings 5th World Filtration Congress,
Nizza, 1990.
[20] I.M. Wheeldon, G.K. Burnard, G.C. Snow, L. Svarovsky, The performance of
cyclones in the off-gas of a pressurised fluidised bed combustor, Inst. Chem.
Eng. Symp. Ser. 99 (1986) 4565.
[21] I. Gulyurtlu, I. Cabrita, C. Bordalo, Separation of particles in a cyclone at
temperatures above 400 C, in: 1. Europ. Symp. Partikelabscheidung aus
Gasen, Nurnberg, 1989.
[22] R. Parker, R. Jain, S. Calvert, D. Drehmel, J. Abbot, Particle collection in cyclones
at high temperatures and high pressures, Environ. Sci. Technol. 15 (1981) 451
458.
[23] L. Shi, D.J. Bayless, G. Kremer, B. Stuart, CFD Simulation of the influence of
temperature and pressure on the flow pattern in cyclones, Ind. Eng. Chem. Res.
45 (2006) 76677672.
[24] M. Thulasiraman, K. Pitchandi, Influence of inlet velocity of air and solid
particle feed rate on holdup mass and heat transfer characteristics in cyclone
heat exchanger, J. Mech. Sci. Technol. 29 (2015) 45094518.
[25] D. Leith, D.W. Cooper, S.N. Rudnick, Venturi scrubbers: pressure loss and
regain, Aerosol. Sci. Technol. 4 (1985) 239243.
[26] J. Abrahamson, C.G. Martin, K.K. Wong, The physical mechanisms of dust
collection in a cyclone, Chem. Eng. Res. Des. 56 (1978) 168177.
[27] A.C. Stern, K.J. Caplan, P.D. Bush, Cyclone Dust Collectors, American Petroleum
Institute, New York, 1955.
[28] I.M. Wheeldon, G.K. Burnard, Performance of cyclones in the off-gas path of a
pressurised fluidised bed combustor, Filt. Sep. 24 (1987) 178187.
[29] T. Santosh, R. Suresh, K.V. Sreenivas, U.S. Malliikrjun, CFD analysis for design
optimization of reverse flow type cyclone separator, Int. J. Mech. Prod. Eng. 1
(2011) 110123.
[30] N. Gopani, A. Bhargava, Design of high efficiency cyclone for tiny cement
industry, Int. J. Environ. Sci. Dev. 2 (2011) 350354.
[31] M. Wasilewski, J. Duda, Influence construction of suspension preheater on
energy consumption process during burning in rotary kiln, in: 14th
International Symposium Heat Transfer and Renewable Sources of Energy,
September 0609, 2012.
[32] M. Wasilewski, J. Duda, A. Duczkowska-Kadziel, Application of computational
fluid dynamics to optimization of cyclone dust separators operated in the
cement industry, Chemik 67 (2013) 985994.
[33] H. Zhang, R. Dewil, J. Degreve, J. Baeyens, The design of cyclonic pre-heaters in
suspension cement kilns, Int. J. Sustain. Eng. 7 (2014) 307312.
[34] H. Mikulcic, M. Vujanovic, M.S. Ashhab, N. Duic, Large eddy simulation of a
two-phase reacting swirl flow inside a cement cyclone, Powder Technol. 75
(2014) 8996.
[35] M. Wasilewski, J. Duda, Multicriteria optimisation of first-stage cyclones in the
clinker burning system by means of numerical modelling and experimental
research, Powder Technol. 289 (2016) 143158.
[36] C. Bhaske, Flow simulation in industrial cyclone separator, Adv. Eng. Soft. 41
(2010) 220228.
[37] B. Wang, D.L. Xu, K.W. Chu, A.B. Yu, Numerical study of gassolid flow in a
cyclone separator, App. Math. Model. 30 (2006) 13261342.
[38] G. Wan, G. Sun, X. Xue, M. Shi, Solids concentration simulation of different size
particles in a cyclone separator, Powder Technol. 183 (2008) 94104.
[39] K.W. Chu, B. Wang, A. Vince, A.B. Yu, G.D. Barnett, P.J. Barnett, CFD-DEM study
of the effect of particle density distribution on the multiphase flow and
performance of dense medium cyclone, Miner. Eng. 22 (2009) 893909.
[40] K.W. Chu, B. Wang, A.B. Yu, A. Vince, CFD-DEM modelling of multiphase flow in
dense medium cyclones, Powder Technol. 193 (2009) 235247.
[41] A. Raoufi, M. Shams, H. Kanani, CFD analysis of flow field in square cyclones,
Powder Technol. 191 (2009) 349357.
[42] Y. Su, A. Zheng, B. Zhao, Numerical simulation of effect of inlet configuration
on square cyclone separator performance, Powder Technol. 210 (2011) 293
303.
[43] H. Safikhani, M. Akhavan-Behabadi, M. Shams, M.H. Rahimyan, Numerical
simulation of flow field in three types of standard cyclone separators, Adv.
Powder Technol. 21 (2010) 435442.
[44] A.K.S. Parihar, Ch. Joshi, G. Sridhar, The performance of cyclones in producer
gas cleaning: experimental and modeling studies, Proc. Inst. Mech. Eng. (2012)
118.
[45] A. Kepa, The efficiency improvement of a large-diameter cyclone the CFD
calculations, Sep. Purif. Technol. 118 (2013) 105111.
[46] A. Kepa, The effect of a counter-cone position on cyclone performance, Sep. Sci.
Technol. 47 (2012) 22502255.
[47] J.P. Wu, Y.H. Zhang, H.L. Wang, Numerical study on tangential velocity
indicator of free vortex in the cyclone, Sep. Purif. Technol. 132 (2014)
541551.
[48] F. Liu, J. Chen, A. Zhang, X. Wang, T. Dong, Performance and flow behavior of
four identical parallel cyclones, Sep. Purif. Technol. 134 (2014) 147157.
[49] J. Oh, S. Choi, J. Kim, Numerical simulation of an internal flow field in a uniflow
cyclone separator, Powder Technol. 274 (2015) 135145.
 M. Wasilewski / Separation and Purification Technology 168 (2016) 114123
[50] L.S. Brar, R.P. Sharma, K. Elsayed, The effect of the cyclone length on the
performance of Stairmand high-efficiency cyclone, Powder Technol. 286
(2015) 668677.
[51] Y. Wakizono, T. Maeda, K. Fukui, H. Yoshida, Effect of ring shape attached on
upper outlet pipe on fine particle classification of gas-cyclone, Sep. Purif.
Technol. 141 (2015) 8493.
[52] J.J.H. Houben, S. Pirker, CFD simulations of pressure drop and velocity field in a
cyclone separator with central vortex stabilization rod, J. Appl. Fluid. Mech. 9
(2016) 487499.
[53] S.G. Bogodage, A.Y.T. Leung, CFD simulation of cyclone separators to reduce air
pollution, Powder Technol. 286 (2015) 488506.
[54] M. Masin, Present Burning Procedure of Portland Clinker, Stavivo 3, Publishers
of Technical Literature, Prague, 1966, pp. 8891.
[55] F.J. de Souza, R. de Vasconcelos Salvo, D. de Moro Martins, Simulation of the
performance of small cyclone separators through the use of Post Cyclones
(PoC) and annular overflow ducts, Sep. Purif. Technol. 142 (2015) 7182.
[56] F.J. de Souza, R. de Vasconcelos Salvo, D. de Moro Martins, Effects of the gas
outlet duct length and shape on the performance of cyclone separators, Sep.
Purif. Technol. 142 (2015) 90100.
[57] ANSYS FLUENT Theory Guide 14.0, November 2011.
123
[58] Fluid Dynamics Analysis Package, Theory Manual Revision 8.0.
After graduating from the Faculty of Mechanical Engi-
neering at Opole University of Technology, Marek
Wasilewski began work at the university and attained
his PhD in 2015. His doctoral thesis concerned the
optimisation of cyclone separators. Currently, he is
employed as an assistant within the Chair of Innovative
Technological Processes. He is the editor of two mono-
graphs and the author of 20 papers. His research inter-
ests include the separation process, thermal
engineering, heat transfer, renewable energy and opti-
misation of production processes, with a particular
focus on the production of cement.