Brazilian Journal

of Chemical ISSN 0104-6632

Printed in Brazil
Engineering www.abeq.org.br/bjche

Vol. 30, No. 03, pp. 627 - 641, July - September, 2013

THE EFFECT OF THE INTRODUCTION OF

AN EXIT TUBE ON THE SEPARATION
EFFICIENCY IN A CYCLONE
J. Cernecky1 and K. Plandorova2
1
Technical University in Zvolen, Faculty of Environmental and Manufacturing Technology,
Department of Environmental Technology, T. G. Masaryka 24, 960 53 Zvolen, Slovakia, Europe.
E-mail: [email protected]
2
SIGMA SLOVAKIA, Zvolenska cesta 14 / 1769 974 01 Banska Bystrica, Slovakia, Europe.

(Submitted: February 22, 2012 ; Revised: August 14, 2012 ; Accepted: August 25, 2012)

Abstract - The paper shows the analysis of the effect of introduction of an exit tube into the cylindrical part
of a cyclone on the flow velocity, pressure losses and, above all, the separation efficiency. We made
measurements and carried out CFD simulations for three levels of the exit tube introduction. The ratios of the
depth of the exit tube introduction to the cyclone diameter (Hp/D) were 0.4, 0.475 and 0.89 at three velocity
settings of 8 m.s-1, 13 m.s-1 and 15 m.s-1. The Reynolds Stress Model (RSM), which was compatible with the
experimental results, was used for numerical solutions. The efficiency of cyclone separation was explored on
a sample of oak sawdust. The efficiency of cyclone separation increased with the geometric size of the
particles, the inlet velocity and a deeper introduction of the exit tube.
Keywords: Cyclone; Separation; Computational fluid dynamics; Oak sawdust.

INTRODUCTION towards the bottom conical part, where the separated

material does not usually descend any further. Under
Cyclones are technical devices which are used for the influence of centrifugal forces and the negatively
separating loose matter from carrier gas with the inclined plane of a lateral surface of a cone, the
utilization of gravitational, inertial and centrifugal particles are conversely forced to ascend against the
forces. A simple construction, possible operating newly incoming material, with which it combines
reliability, low investment and operating costs and and acquires volume. The fact that the particles leave
low energy consumption are the basic qualitative the hopper is due to the increase in volume, slowing
aspects of the cyclone. Low separation of soft down their rotation by the friction against the lateral
particles, abrasive effects and adhesive power of surface, gravity force and the pressure of other
particles are among to the disadvantages. The total material entering the separator. The conical shape of
cyclone efficiency is influenced by factors such as the bottom part of the separator is necessary because
the flow rate of a heterogeneous mixture through the of gradual closing of the central ascending vacuum
cyclone, the properties and concentration of loose flow, which is produced by rotation together with a
matter, cyclone design and its setting, total change of medium flow direction nearer to a
separation, fractional separation, cyclone resistance, separator axis. This implies that, on the axis of
deterioration and other factors. The heterogeneous rotation, it is not possible to have a big outlet
mixture enters the cylindrical part of a separator and, diameter because of reverse drag of the separated
after that, it is put into a downward rotary movement particles.
*To whom correspondence should be addressed
628 J. Cernecky and K. Plandorova

The works of Lapple (1951) and Barth (1956) are A literature survey showed that the vortex length
considered to be the first studies dealing with cyclone can be important for predicting the separation
models. The studies were simple and provided efficiency, especially in short cyclones, but for
acceptable results with a limited range of parameters. longer cyclones the turbulence does not have to
Later, those theories were improved by Smith (1962) reach the cone top. If the length of a cyclonic
and Muschelknautz (1970). In their theories they cylinder increases, the separation efficiency in-
considered more parameters such as surface creases to a certain value and then it begins to
roughness, area of distribution of particles, shape and decrease; it is also similar in an overflow pipe. By
size of particles, pressure losses and others. The lengthening the height of the cyclonic cylinder or by
efficiency of vortex separators depends on the shortening the length of the overflow pipe we can
regime, which is also influenced, apart from the lessen the pressure drop to a certain extent (Zhu and
already mentioned parameters, by inlet velocity, Lee, 1999). The definition and composition of the
temperature and viscosity, which are introduced in pressure drop in the cyclone were analysed in the
the work of Avci and Karagoz (2003). This work paper of Chen and Shi (2006). The pressure drop in
investigated the cyclone flow patterns introducing the cyclone includes spreading losses at the inlet into
different shapes and diameters of exit tube simulated the cyclone and outlet spinning losses caused by the
at diverse flow rates. Lim et al. (2004) examined friction near the walls of the cyclone and the
experimentally the effect of vortex finder shape on dynamic energy of the diffused gas in the overflow
the characteristics of the collection efficiency at pipe. The surface resistance or losses caused by
different flow rates. Elsayed et al. (2010) evaluated friction have a substantial influence on the vortex
the flow patterns and performance of a cyclone with length; with decreasing surface resistivity the length
different vortex finder diameters to examine the of turbulence in the cyclone increases. Numerical
effect of the vortex finder diameter on the studies of separable properties of different concentra-
performance and velocity profile. In the experimen- tions of particles in the cyclone at the inlet were
tal work of Hoffman et al. (1995) it was proved that investigated in the paper of Oian et al. (2007). The
the inlet velocity has a significant influence on the effects of inlet concentration of particles on the
length of turbulence depending on cyclone geometric tangential velocity, particle size, pressure drop and
dimensions. The paper of Lidén and Gudmundsson separation efficiency in the cyclone were deter-
(1996) described the use of semi-empirical polyhedra mined. The impacts of cone dimensions on the
for separation efficiency depending on the operating cyclone performance were examined in the study of
conditions and cyclone dimensions. The length of Xiang et al. (2000).
turbulence in the cyclone increased with increasing The mathematical expression for three-dimen-
inlet velocity. In the work of Martignoni et al. sional flow in the cyclone is complex and has not
(2007), the impact of cyclone geometry was studied been explored in depth. The calculations of physical
by means of creating a symmetrical double inlet and laws are dependent on the attention, time and effort.
a cyclone model with a spiral outlet compared with a Therefore, it was necessary to examine these theories
common cyclone with a tangential inlet. It was on the computer by means of polyhedra and to
verified that the total separation efficiency increases compare the results obtained with experimental data
with decreasing pressure drop in both cyclone or theoretical knowledge. For a better idea of flow
shapes, but in a cyclone with shaped symmetrical and physical actions occurring in the cyclone, CFD
inlets a bigger pressure drop was recorded than in the (Computational Fluid Dynamics) models are utilized
cyclone with the spiral outlet. The results of the at present. The results of numerical calculations and
experimental study of Kim, Lee and Kuhlman (2001) gas flow simulations in the typical Lapple cyclone
provide information about the efficiency of particle are described in the paper of Wang et al. (2006). The
separation for a modified inner surface of the body accuracy of the numerical solution of a flow
of cyclones with spiral, circumferential and vertical simulation in the cyclone was achieved using the
grooves. The estimation of separation efficiency RSM model. The CFD flow simulations completed
dependent on the concentration of particles in axial with properties of heat transfer in the cyclone with a
flow in a cyclone dust collector was published by tangential inlet were published by Karagoz and Kaya
Ogawa (1999). Under the centrifugal force solid (2007). The publication of Griffiths and Boysan
particles slide along the cone surface and fall into a (1996) focused on the comparison of performance of
dust hopper, where some particles are trapped and three small cyclone types with the use of CFD data
some are blown away by secondary flow back to the from experiments and three empirical theories. Gas
cyclone body. flow and a solid component in the cyclone are

Brazilian Journal of Chemical Engineering

 The Effect of the Introduction of an Exit Tube on the Separation Efficiency in a Cyclone 629

described in the paper of Meier and Mori (1998). In where ρ is the density, u is the mean velocity, x is the
this work, the CFD model was based on an Eulerian position, t is the time, p is the pressure, μ is the
approach for both phases with the condition of a 3D dynamic viscosity, u′i u′j is the Reynolds stress tensor,
symmetrical domain, the mathematical model was
compiled for the utilization of the k-ε turbulent g is the acceleration of gravity and Fi is the external
model. body force which arises due to the interaction with
The possibility of using CFD techniques enables dispersed solid particles.
one to make a proposal of more effective separation This equation of motion now represents the
conditions without the need of conducting expensive averaged velocity values (or main flow). Turbulence
experiments. Currently, CFD simulations have rela- is represented by means of the “Reynolds stresses”
_____
tively wide application for relatively precise mathe- ρ u′i u′j . FLUENT describes the effect of main flow
matical approximation of flow, while the arduousness
of its performance and time restraints in solving and the Reynolds stresses by means of the above
concrete physical problems must not be forgotten. mentioned models. The Reynolds Stress Model
includes the calculation of particular Reynolds
stresses by means of a transport differential equation
TURBULENT MODEL DESCRIPTION in the form:

The use of CFD (computational fluid dynamics) ∂ u′i u′j ∂ u′i u′j
+ uk =
for numerical calculations of gas flow in the cyclone ∂t ∂x k
has increasingly been applied. Three models are ( LTM ) ( )
Cij
usually used for the simulation in the cyclone: the
k-ε model, Algebraic Stress Model (ASM) and ∂ ⎡ p′ ∂ ⎤
Reynolds Stress Model (RSM). Wang et al. (2006) −
∂x k ⎣
(
⎢ u′i u′ju′k +
ρ
) ( )
δkju′i + δki u′j − υ
∂x k
( )
u′i u′j ⎥
found that the standard k-ε turbulence model is ⎦
inadequate to simulate the flow with swirling motion Dij ( ) (3)
because it leads to excessive turbulence viscosity and
unrealistic tangential velocity. Therefore, the accu- ⎡ ∂u j ∂u ⎤ p′ ⎡ ∂u ∂u j ⎤
− ⎢ u′i u′k + u′i u′k i ⎥ + ⎢ i + ⎥
racy of the numerical solution can be improved by ⎣ ∂x k ∂x k ⎦ ρ ⎢⎣ ∂x j ∂x i ⎥⎦
using the Reynolds Stress Model (RSM), which was
also utilized in this paper for the simulation of
( )
Pij Φ ij( )
physical actions in the cyclone by using the ⎡ ∂u ∂u i ⎤
FLUENT commercial program. −2 υ ⎢ i ⎥ + Su ,
In turbulent flow, which is accompanied by ⎣ ∂x k ∂x k ⎦
pulsation and fluctuation, the total velocity equals ( )
εij
the sum of the central and fluctuation components of
velocity: where the left two terms are the local time derivative
of stress (LTM) and the convective transport term
u i = ui + u i ' . (1) (Cij), respectively. The right five terms are: Dij is the
stress diffusion term, Pij is the shear production term,
By substituting this expression into the equation Фij is the pressure- strain term, εij is the dissipation
of motion we get the averaged Navier-Stokes term and Su is the source term.
equations of motion in the form: For better expression, the individual members are
approximated for the purpose of closing the set of
∂ ∂ simultaneous equations from this differential equa-
∂t
( ρu i ) +
∂x j
(
ρu i u j ) tion. The Reynolds stresses are consequently substi-
tuted into the equation of momentum transfer.
Generally, the following equations are solved by
∂ ⎛ ⎡ ∂u i ∂u j ⎤ ⎛ 2 ∂u j ⎞ ⎞ using the RSM model:
= ⎜μ ⎢ + ⎥−⎜ μ ⎟⎟ (2)
∂x j ⎜⎝ ⎣⎢ ∂x j ∂x i ⎥⎦ ⎝⎜ 3 ∂x j ⎠⎟ ⎟⎠ ƒ 6 transport equations in the form (3),
ƒ 3 transport equations for the mean component of
∂p ∂
−
∂x i
+ ρg + Fi +
∂x j
(
−ρ u′i u′j ,) velocity,
ƒ continuity equation,

Brazilian Journal of Chemical Engineering Vol. 30, No. 03, pp. 627 - 641, July - September, 2013
630 J. Cernecky and K. Plandorova

ƒ transport equation for the rate of dissipation ε.

Generally, this system solves 12 equations. In
comparison with the previous models, a dispropor-
tionally greater arduousness is placed on the
numerical model (Launder et al., 1975).
The equation of motion of a small particle,
including the effects of nonlinear drag and gravita-
tional forces, is given by:

du iP 3νC D Re p
dt
=
4d 2S
( )
u i − u iP + gi , (4)

dx i
u iP = . (5)
dt

where u iP is the velocity of the particle and xi is its

position, d is particle diameter, S is the ratio of
particle density to fluid density, gi is the acceleration
of gravity and according to Raoufi et al. (2008) the
drag coefficient CD is given as:
Figure 1: A cyclone model with the main dimensions
24
CD = , for ReP < 1, (6)
Re P Table 1: Basic dimensions of a cyclone model
(mm)
24 1
CD = (1 + Re 2/3
P ) , for 1 < ReP < 400, (7) D dp dk b1 h1 Hv Hk Hp Hc dm Hm
Re P 6
200 65 85 40 90 262 520 178 782 150 110

where ReP is the particle Reynolds number defined A set of measurements was carried out for three
as: positions of introduction of an exit tube into the
inner part of the cyclone at a barometric pressure of
d u j − u Pj 1 025 hPa. The exit tube is in the shape of a cylinder
Re P = . (8) with a length of 300 mm and was designed in a way
ν so that it is possible to set the necessary depth of
introduction of the tube in the cyclone. The first
measurements were made at a ratio of Hp/D = 0.89,
CONDITIONS OF THE EXPERIMENT AND Figure 2 (a). The second measurements were performed
CFD SIMULATION at a ratio of Hp/D = 0.475, just below a bottom wall
of the inlet cyclone pipeline. The third measurements
Figure 1 shows a designed cyclone model with a were carried out at a ratio of Hp/D = 0.4, Figure 2 (c).
tangential inlet and Table 1 shows basic cyclone A five-port dynamic probe clamped in a special
dimensions. The cyclone model is placed in an holder was used for these measurements. The probe
experimental overpressure system for the loose was calibrated on a calibration device before the
matter separation of diverse composition and measurements (Polansky and Stech, 2011). A sample
concentration. This device was used for experimental of oak sawdust with a density of 670 kg.m-3 and
purposes with an inlet concentration of particles of dampness of 6.5% was used for finding out the
0.01 kg. N-1.m-3. In terms of air flow it is an air- cyclone separation efficiency.
technical system where the transport air is drawn These experiments were carried out to verify the
from the room to a pipe system and from the numerical model. The inlet air velocity as well as the
cyclone separator transported back to the ambient particle velocity were set with a control damper on
environment. the intake opening of the fan at the values of 8 m.s-1,

Brazilian Journal of Chemical Engineering

 The Effect of the Introduction of an Exit Tube on the Separation Efficiency in a Cyclone 631

13 m.s-1 and 15 m.s-1. Three types of CFD models it is 92,606 CFD cells and for Hp/D = 0.89 it is
were created. They represented three different 111,661 CFD cells. Due the use of the polyhedral
examples of the introduction of an exit tube into the grid, which is of high quality, the grid optimization
cylindrical part of the cyclone. Surfaces were was not done during the computation.
covered with a hexagonal grid of tetra-hybrid type, In order to calculate the particle trajectories the
Figure 3, which was converted to a grid of the poly- following parameters of DPM (Discrete phase
hedral type in the FLUENT program. A computa- modelling) were set:  the total flow rate – 0.0029 kg/s,
tional domain for the variant of Hp/D = 0.4 comprises the particle diameter – 1 mm, uniform particle distri-
79,393 CFD cells, for the variant of Hp/D = 0.475 bution and particle material oak.

(a) (b) (c)

Figure 2: The introduction of an exit tube into the cylindrical part of the cyclone

Figure 3: A CFD cyclone grid

Brazilian Journal of Chemical Engineering Vol. 30, No. 03, pp. 627 - 641, July - September, 2013
632 J. Cernecky and K. Plandorova

RESULTS AND DISCUSSION With an increase in inlet velocity the dynamic

pressure in the cyclone decreases. The experimental
Figure 4 shows a static pressure at the velocity of results coincide well with the numerical ones,
15 m.s-1 and three examples of the introduction of an although they are a bit higher.
exit tube. The B-B section represents an inlet plane Figure 6 shows the comparison of numerical and
for all measurements and its distance is 218 mm calculated tangential velocities in the cylindrical
from the top part of the cyclone cap. The static cyclone part at the inlet velocity of 15 m.s-1 to the
pressure is the greatest along the inner wall of the cyclone. The results of the CFD simulation are in
lateral surface of the cyclone. On the cyclone axis agreement with the calculated ones. The flow pattern
the static pressure has lower values than on the walls indicates quasi-forced and free vortex flow. The
and they are dependent on the vortex core and the turbulence on the cyclone axis agrees approximately
introduction of an exit tube. with the axis of cyclone geometry just at this
Figure 5 shows the relation between the dynamic measurement.
pressure and the inlet gas velocity into the cyclone at The axial and tangential velocities were measured
Hp/D = 0.475. experimentally for three velocity settings at the ratios

(a) (b) (c)

Figure 4: Contours of static pressures at the velocity of 15 m.s-1 and at Hp/D of: (a) 0.89; (b) 0.475; (c) 0.4.

Figure 5: A comparison of numerical and calculated Figure 6: A comparison of numerical and calculated
dynamic pressure. tangential velocities.

Brazilian Journal of Chemical Engineering

 The Effect of the Introduction of an Exit Tube on the Separation Efficiency in a Cyclone 633

of Hp/D = 0.4, 0.475 and 0.89. Figures 7 – 9 show vortex along the cyclone axis and in a closed
the contours of tangential velocity for the introduc- container.
tion of an exit tube at Hp/D = 0.4, 0.475 and 0.89 In all three cases of the introduction of an exit
and for three settings of the inlet velocity of 8 m.s-1, tube, the vortex core is in the shape of a twisted
13 m.s-1 a 15 m.s-1. From the figures it is obvious cylinder and is not axially completely symmetrical,
that, with increasing inlet velocity, the vortex core in especially in the conical part of the cyclone towards
the whole cyclone section is changed. The dominant the container of trapped particles. The axis of the
velocity in the cyclone is a tangential velocity forced vortex is not identical in time with the
component. The value of the tangential velocity geometric cyclone axis, which is in the shape of
equals to zero on the wall and in the centre of field a curve. With increasing inlet velocity the length of
flow. The highest tangential velocity is achieved by turbulence in the cyclone increases.
sucking from the inlet pipeline; then the velocity is Figure 10 shows the contours of tangential velocity
decreased by the vortex gas motion downwards for the introduction at three inlet velocities. From the
along the cyclone wall. Negative values of the figures it can be seen that, with increasing inlet
tangential velocity are achieved in a created, forced velocity, the tangential velocity rises proportionally.

(a) (b) (c)

Figure 7: Contours of tangential velocity for the ratio of Hp/D = 0.4 at the inlet velocity of: (a) 8 m.s-1;
(b) 13 m.s-1; (c) 15 m.s-1.

(a) (b) (c)

Figure 8: Contours of tangential velocity for the ratio of Hp/D = 0.475 at the inlet velocity of: (a) 8 m.s-1;
(b) 13 m.s-1; (c) 15 m.s-1.

Brazilian Journal of Chemical Engineering Vol. 30, No. 03, pp. 627 - 641, July - September, 2013
634 J. Cernecky and K. Plandorova

(a) (b) (c)

Figure 9: Contours of tangential velocity for the ratio of Hp/D = 0.89 at the inlet velocity of: (a) 8 m.s-1;
(b) 13 m.s-1; (c) 15 m.s-1.

Figure 10: Contours of tangential velocity for the ratios of Hp/D = 0.4, 0.475 and 0.89.
Figures 11 – 13 show the contours of axial velocity sample of oak sawdust with a density of 670 kg. m-3
for Hp/D = 0.89 and at the inlet velocity of 8 m.s-1, at three velocities and at three settings of an exit
13 m.s-1 and 15 m.s-1. In the figures it is possible to tube, Figure 15. The particle sizes of the sample used
see that the axial velocity near the cyclone wall were 1 mm, 0.6 mm and 0.125 mm and the separation
achieves negative values and towards the cyclone period was 60 s. From the measurement results it can
centre these velocity values turn into positive ones. be said that the separation efficiency decreased with
Higher values of the axial velocity were recorded in the size of the particles. The next important factor by
the centre of the cylindrical part than in the conical which the separation efficiency can be influenced is
cyclone part. The cyclone hopper has negative values separation velocity and the introduction of an exit
of axial velocity in the whole cross-section. The tube. With increasing inlet velocity the separation
highest axial velocity is achieved near the wall of an efficiency increased for all three settings of an exit
exit tube. tube. At Hp/D = 0.4 the separation was lower because
Figure 14 shows the contours of axial velocity for the particles were dragged from the trajectory of
the introduction at three inlet velocities. swirling motion into the exit tube where they
The separation efficiency was investigated on a escaped from the cyclone as light ashes.

Brazilian Journal of Chemical Engineering

 The Effect of the Introduction of an Exit Tube on the Separation Efficiency in a Cyclone 635

(a) (b) (c)

Figure 11: Contours of axial velocity for the ratio of Hp/D = 0.4 at the inlet velocity of: (a) 8 m.s-1;
(b) 13 m.s-1; (c) 15 m.s-1.

(a) (b) (c)

Figure 12: Contours of axial velocity for the ratio of Hp/D = 0.475 at the inlet velocity of: (a) 8 m.s-1;
(b) 13 m.s-1; (c) 15 m.s-1.

(a) (b) (c)

Figure 13: Contours of axial velocity for the ratio of Hp/D = 0.89 at the inlet velocity of: (a) 8 m.s-1;
(b) 13 m.s-1; (c) 15 m.s-1.

Brazilian Journal of Chemical Engineering Vol. 30, No. 03, pp. 627 - 641, July - September, 2013
636 J. Cernecky and K. Plandorova

Figure 14: Contours of axial velocity for the ratios of Hp/D = 0.4, 0.475 and 0.89.

Figure 15: The efficiency of cyclone separation for oak sawdust at the ratios of Hp/D = 0.4, 0.475 and 0.89.

Brazilian Journal of Chemical Engineering

 The Effect of the Introduction of an Exit Tube on the Separation Efficiency in a Cyclone 637

In Figure 16 there is a simulation of the trajectory velocity at the inlet into the cyclone and the
of oak particles with a size of 1 mm coloured downwards along the conical cyclone part the
according to the velocity size for Hp/D = 0.475 at velocity decreases. The fact that the velocity
the inlet velocity of 15 m.s-1; the time step was 0.2 s decreases is due to the abrasion of particles on the
and the time stop was 10 s. From the figure it is cyclone wall, the action of gravity and also the
possible to see that the particles achieve the highest weight of newly arriving particles in the cyclone.

Figure 16: Particle trajectories coloured by particle time (s).

Brazilian Journal of Chemical Engineering Vol. 30, No. 03, pp. 627 - 641, July - September, 2013
638 J. Cernecky and K. Plandorova

The following figure (Figure 17) shows a graphical for the fractions of 1 mm, 0.6 mm and 0.125 mm, parti-
dependence of the fractional separation by particle size cle sizes that occur mainly in the wood-processing indus-
for Hp/D = 0.475 for the flow velocities of 8, 13 and try. A comparison of the experimental results with the
15 m.s-1. The separation was investigated experimentally outputs of CFD simulations can be seen in the Figure 18.

Figure 17: A curve of fractional separation for the ratio

of Hp/D = 0.475 for three flow velocities.

Brazilian Journal of Chemical Engineering

 The Effect of the Introduction of an Exit Tube on the Separation Efficiency in a Cyclone 639

Figure 18: The efficiency of cyclone separation for diverse flow velocities, particle sizes and ratios of Hp/D.

CONCLUSIONS 5) CFD simulations were used in the study of the

physical phenomena occurring in the cyclone. The
1) From the experimental results and the CFD numerical results presented for axial and tangential
simulations it can be said that the introduction of an velocity are in good agreement with the experimental
exit tube has a significant influence on the air flow ones.
and pressure losses in the cyclone and on the 6) The position of the exit tube and the particle
separation efficiency. By introducing an exit tube size influenced greatly the cyclone separation. At the
into a cylindrical cyclone at the ratio of Hp/D = 0.89, smaller introduction of the tube, the separation was
the vortex core was not created symmetrically along lower because the particles were dragged from the
the whole central cyclone axis as it was at the ratios trajectory of swirling motion into the exit tube where
of Hp/D = 0.475 and 0.4. they escaped from the cyclone as light ashes. In all
2) Deep introduction of an exit tube into a measurements it was found that the particles with a
cylindrical cyclone causes higher fluid flow velocity smaller size had smaller separation than the particles
in the separation space of the cyclone, which is with a bigger diameter. With increasing inlet velocity
manifested in higher pressure cyclone losses. the separation efficiency increases and pressure
3) In all three cases of the introduction of an exit losses increase. The purpose of geometrical ratio
tube, the vortex core is in the shape of a twisted optimization in a vortex separator is to ensure the
cylinder and is not completely axially symmetrical, highest separation efficiency at the lowest pressure
especially in the part of the conical cyclone towards losses. Based on the results obtained from the
the dust hopper. The axis of the forced vortex is not experimental measurements, the generalized optimi-
temporally coincident with the geometric cyclone zation of the separation of dust particles in the wood-
axis, which is in the shape of a curve. With processing industry does not lead to explicit conclu-
increasing inlet velocity, the length of turbulence in sions. In the real operation of the wood-processing
the cyclone is increased. CFD simulations were industry the process of dust particle separation is
applied for particle sizes of 0.002 mm, 0.005 mm, influenced by many factors (blunting of cutting tools,
0.125 mm, 0.6 mm and 1 mm. The curve of particle shape, humidity of processing material,
fractional separation was also created from CFD environmental humidity, etc.). In terms of the
simulations. It is obvious from the graph that the optimal geometrical ratios in the experimental
separation decreases upon reducing the particle size. separator we focused mainly on the balancing of the
Apart from the particle size impact, separation was influence of the ratio of Hp/D on the separation
also influenced by the geometrical ratios of the efficiency. From the experimental and CFD results it
particular parts of the vortex separator. can be concluded that with the increase in Hp/D the
4) With increasing inlet velocity to the cyclone separation efficiency also increases, which was
the tangential velocity increases proportionally and reflected mainly in the dust particles with larger
this leads to an increase in the separation efficiency diameters (1 mm and 0.6 mm). For smaller particles,
in the cyclone. a more significant influence on the separation
Brazilian Journal of Chemical Engineering Vol. 30, No. 03, pp. 627 - 641, July - September, 2013
640 J. Cernecky and K. Plandorova

process was environment humidity, which was Kraft (BWK), 8, 1 (1956).

reflected in the outputs from the experimental Elsayed, Kh. and Lacor, C., The effect of vortex
measurements as an influence of small particle finder diameter on cyclone separator performance
agglomeration. It can be concluded that the optimal and flow field. ECCOMAS CFD Conf., J. C. F.
ratio of Hp/D under our conditions is 0.89. Pereira and A. Sequeira (Eds.) Lisbon (2010).
Griffiths, W. D. and Boyson, F., Computational Fluid
Dynamics (CFD) and empirical modeling of the
NOMENCLATURE performance of a number of cyclone samplers.
Journal of Aerosol Science 27, No. 2, 281 (1996).
CD drag coefficient dimensionless Hoffman, A. C., Jonge, R., Arends, H. and Hanrats,
d particle diameter m C., Evidence of the natural vortex length and its
Fi external body force N/m3 effect on the separation efficiency of gas cyclones.
g acceleration of gravity m/s2 Filtration & Separation, 32, No. 8, 799 (1995).
p′ dispersion pressure Pa Chen, J. and Shi, M., A universal model to calculate
Rep particle Reynolds number dimensionless cyclone pressure drop. Powder Technology, 171,
S ratio of particle density to dimensionless No. 3, 184 (2007).
fluid density Karagoz, I. and Kaya, F., CFD investigation of the
t time s flow and heat transfer characteristics in a tangential
u′ dispersion velocity m/s inlet cyclone. International Communications in Heat
u time average velocity in the m/s and Mass Transfer, 34, No. 9-10, 1119 (2007).
axial direction Kim, H. T., Lee, K. W. and Kuhlman, M. R.,
velocity of the particle m/s Exploratory design modifications for enhancing
u ip cyclone performance. Journal of Aerosol Science,
v velocity m/s 32, No. 10, 1135 (2001).
x position m Lapple, C. E., Processes use many collector types.
Chemical Engineering, 58, 144 (1951).
Greek Letters Launder, B. E., Reece, G. J. and Rodi, W., Progress
in the development of a Reynolds-stress turbulent
μ dynamic viscosity Pa.s closure. Journal of Fluid Mechanics, 68, No. 3,
ν kinematic viscosity m2/s 537 (1975).
ρ density kg/m3 Lidén, G. and Gudmundsson, A., Semi-empirical
modeling to generalise the dependence of cyclone
Subscripts collection efficiency on operating conditions and
cyclone. Journal of Aerosol Science, 28, No. 5,
i, j, k 1,2,3 853 (1996).
p particle Lim, K. S., Kim, H. S. and Lee, K. W., Characteristics
of the collection effciency for a cyclone with
different vortex finder shapes. Journal of Aerosol
ACKNOWLEDGMENTS Science, 35, No. 6, 743 (2004).
Martignoni, W. P., Bernardo, S. and Quintani, C. L.,
The contribution was created within the KEGA Evaluation of cyclone geometry and its influence
project no. 027TUZVO-4/2011 funded by the Ministry on performance parameters by computational
of Education, Science, Research and Sport of the fluid dynamics (CFD). Brazilian Journal of
Slovak Republic. Chemical Engineering, 24, No. 1, 83 (2007).
Meier, H. F. and Mori, M., Gas-solid flow in
cyclones: The Eulerian approach. Computers &
REFERENCES Chemical Engineering, 22, No. 1, 641 (1998).
Muschelknautz, E. and Krambrock, W., Aerodyna-
Avci, A. and Karagoz, I., Effects of flow and geo- mische Beiwerte des Zyklonabscheiders aufgrund
metrical parameters on the collection efficiency neuer und veberssertter Messungen. Chemie-
in cyclone separators. Journal of Aerosol Science, Ingenieur-Technik, 42, No. 5, 247 (1970). (In
34, No. 7, 937 (2003). German).
Barth, W., Design and layout of the cyclone separator Ogawa, A., Estimation of collection efficiency
on the basis of new investigations. Brennstoff- depended on feed particle concentration for axial

Brazilian Journal of Chemical Engineering

 The Effect of the Introduction of an Exit Tube on the Separation Efficiency in a Cyclone 641

flow cyclone dust collector. Journal of Thermal Smith, J. L., An experimental study of the vortex in
Science, 8, No. 3, 143 (1999). the cyclone separator. Journal of Basic
Polansky, J. and Stech, J., Calibration of a five-port Engineering, 84, No. 4, 602 (1962).
probe – in Czech. ZCU Plzen (2011). Wang, B., Xu, D. L., Chu, K. W. and Yu, A. B.,
Qian, F. and Wu, Y., Effects of the inlet section Numerical study of gas-solid flow in a cyclone
angle on the separation performance of a cyclone. separator. Applied Mathematical Modelling, 30,
Chemical Engineering Research and Design, 87, No. 11, 1326 (2006).
No. 12, 1567 (2009). Xiang, R., Park, S. H. and Lee, K. W., Effects of
Raoufi, A., Shams, M., Farzaneh, M., and Ebrahimi, cone dimension on cyclone performance. Journal
R., Numerical simulation and optimization of of Aerosol Science, 32, No. 4, 549 (2001).
fluid flow in cyclone vortex finder. Chemical Zhu, I. and Lee, K. W., Experimental study on small
Engineering and Processing, 47, No. 1, 128 cyclones operating at high flowrates. Journal of
(2008). Aerosol Science, 30, No. 10, 1303 (1999).

Brazilian Journal of Chemical Engineering Vol. 30, No. 03, pp. 627 - 641, July - September, 2013