Separation and Purification Technology 74 (2010) 271–279

Contents lists available at ScienceDirect

Separation and Purification Technology

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

Solid/liquid separation performance of hydrocyclones with different cone

combinations
Qiang Yang, Hua-lin Wang ∗ , Yi Liu, Zhi-ming Li
State-Key Laboratory of Chemical Engineering, East China University of Science and Technology, Shanghai 200237, PR China

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

Article history: Hydrocyclones used for solid–liquid separation are usually composed of a single cone. In this paper, we
Received 8 March 2010 designed hydrocyclones with two cone combinations for solid–liquid separation and studied the flow
Received in revised form 7 June 2010 field and separation performance. Simulation and experimental results showed that when the second
Accepted 20 June 2010
cone remained unchanged, the angle change of the first cone had significant effect on the value of three-
dimensional velocities, flow split, separation efficiency, energy consumption, and separation sharpness,
Keywords:
but little effect on the distribution of pressure and that of three dimensional velocities, the capacity and
Hydrocyclones
cut size. The bigger of the first cone’s angle, the smaller of the flow split and the higher the separation
Solid–liquid separation
Cone combinations
efficiency, the stronger of the centrifugal force, and the more the small particles in underflow. The smaller
Flow fields the angle change between the two cones, the larger the sharpness of the grade efficiency curve, i.e., the
hydrocyclone is more suitable for the classification process.
© 2010 Elsevier B.V. All rights reserved.

1. Introduction
Hydrocyclones are widely used for particle separation, classi-
fication and thickening in many solid–liquid processes for their
advantages such as simple structures, robust separation devices
with no moving parts, small physical size of unit, and low installa-
tion and maintenance costs. In the recent years, the hydrocyclones
have been widely applied in mineral processing [1], food [2],
petrochemical [3], electro-chemical [4], textile and pulp [5], envi-
ronmental [6], biological [7] and other industries.
The investigations on the design and performance of hydrocy-
clones have been continued for several decades. It is well known
that the structure of the hydrocyclone and the operating param-
eters affect the hydrocyclone separation performance. For the
established solid–liquid separation system, the structure of hydro-
cyclone is the determinant factor. To get satisfactory separation
performance, a great number of investigations have been made
concerning on the structure and the geometric dimension scale of
hydrocyclones [8]. Some efforts have been made to improve certain
performance indices of conventional hydrocyclones by introduc-
ing special structural modifications. For example, a solid core was
inserted into the hydrocyclone to improve separation efficiency and
reduce energy loss by stabilizing the tangential flow and canceling
the air core [9], a diffuser type of vortex finder was introduced
to reduce the energy loss by recovering the velocity head [10],
a hydrocyclone with small diameter was used for separate small
particles [11], and a hydrocyclone with spiral inner surface was
developed to improve the separation sharpness by interrupting
the boundary layer flow [12], etc. Also, some brand new struc-
tural hydrocyclones have been studied recently. For example, some
new structures like new types of inlet, vortex finder, cone parts,
underflow pipes, central insertions were designed to improve the
separation efficiency, separation sharpness, cut size, capacity, flow
split and energy loss efficiency of the hydrocyclone [13–16]. Chu
et al. studied the effect of structural modification on hydrocyclone
performance more comprehensively [17]. However, up to now, few
experimental and simulation results have been obtained on the
effect of hydrocyclone cone combinations on the solid–liquid sep-
aration performance and the flow field inside hydrocyclone yet.
In this study, hydrocyclones with two cone combinations will be
designed for solid–liquid separation, and the Computational Fluid
Dynamics (CFD) method was used to simulate the flow fields inside
hydrocyclones with different cone combinations, also the effect of
hydrocyclone cone combinations on the separation performance
will be experimentally studied.

2. Simulation parts
∗ Corresponding author at: State-Key Laboratory of Chemical Engineering, East
China University of Science and Technology, 130 Meilong Road, Shanghai 200237,
PR China. Tel.: +86 21 64252748; fax: +86 21 64251894.
Three hydrocyclones with diameter of 100 mm were
E-mail addresses: yangq2005@mail.ecust.edu.cn, tianhei 2005@hotmail.com designed in this study. The hydrocyclone geometry is
(H.-l. Wang). shown in Fig. 1. They have the same type of inlet, outlet

1383-5866/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.seppur.2010.06.014
272 Q. Yang et al. / Separation and Purification Technology 74 (2010) 271–279

of simulation of hydrocyclones using the Reynolds stress model

Nomenclature (RSM) have been reported to be suitable for modeling the flow in
the hydrocyclone in recent years [18–20]. Therefore, the Reynolds
de inlet equivalent diameter of the hydrocyclone (m) stress model (RSM) of the hydrocyclone internal turbulent flow
E separation capability of the hydrocyclone, simulation is used in this paper. Gambit, which is the main prepro-
E = (1 − (go /gi )) × 100( % ) cessor of Fluent, is used to create geometry, meshing and specifying
F flow split of the hydrocyclone, F = Qu /Qi × 100 % (%) boundary types of hydrocyclone. The created geometry and mesh-
go , gi the concentration of solid in the overflow and feed ing are shown in Figs. 2 and 3. Boundary types are considered as
(mg/L) inlet mass flow rate, overflow and underflow outlet pressure. The
G (d) modify grade efficiency, G (d) = (G(d) − F)/(1 − F) (%) solid weight concentration of feed slurry was 4 wt%. Solid particles
pi , po the pressure in the feed and overflow of the hydro- with different sizes varying between 1 and 60 ␮m were injected
cyclone (MPa) from the feed inlet boundary zone along the surface. Using Fluent
pio pressure drop of inlet and overflow of the hydrocy- the created geometry by Gambit can be read and simulation is done.
clone pio = pi − po (MPa)
Qu , Qi individually the volume flow of the underflow and
feed (m3 /h) 3. Experimental parts
(Re)e inlet Reynolds number of the hydrocyclone, (Re)e =
de ve / 3.1. Hydrocyclones
ve inlet velocity of the hydrocyclone, ve = 4Qi /de2
Three hydrocyclones are designed as that shown in Fig. 1.
Greek letters
 liquid medium kinetic viscosity (Pa s)
 liquid medium density (kg/m3 ) 3.2. Materials and methods

and geometric dimension scale except the cone combina-
tions.
The multiphase flow in a hydrocyclone is quite complicated. In
this work, pressure and velocity fields of the hydrocyclones with
different cone combinations were numerically simulated using
FLUENT computational fluid dynamics (CFD) software. Some works
The feed was a mixture of solids and water. The solid was from
the flue gas desulphurization (FGD) system of a power plant. The
suspended solids were composed of CaSO4 ·2H2 O with a purity of
90.15%, CaCO3 with a purity of 2.55%, and fly ash with a purity
of 1.17%. The density of solids was 2.39 g/cm3 (at 20 ◦ C) and the
mean size of particle was 23.954 ␮m. The particle size distribution
is illustrated in Fig. 4, which was measured by a laser granulometer
(Master 2000).

Fig. 1. Structure diagram of designed hydrocyclone in the work. (D = 100 mm;  2 = 10◦ ; Do /D = 0.35; Du1 /D = 0.25; Du2 /D = 0.14; L/D = 1.18; H/D = 0.7; Wi /D = 0.16; Hi /D = 0.25;
Li1 /D = 1.6; Li2 /D = 1  1 = 20◦ , 10◦ and 8◦ for hydrocyclones A, B and C)
 Q. Yang et al. / Separation and Purification Technology 74 (2010) 271–279 273

Fig. 4. Particle size distribution of solids.

The schematic diagram of the experimental apparatus is illus-

trated in Fig. 5. The feed was pumped directly from the stirring tank
to the hydrocyclones. The overflow and underflow samples that
have been separated by the hydrocyclones were taken from the
relevant pipelines. The flow meter and pressure gauge recorded
the flow rate and the pressure of the inlet and overflow of each
hydrocyclone. The experimental equipment is illustrated in Fig. 6.
Because the objective of this study is to understand the effects
of different cone combinations on the hydrocyclone performance,
the characteristics of feed slurry and the operating parameters
were the same for each hydrocyclone in the experiments. The solid
weight concentration of feed slurry was same with simulation in
all the experiments. The inlet pressure e of the hydrocyclone was
changed from 0.06 MPa to 0.22 MPa. The underflow and overflow
pressures were zero. The flow rates of the feed slurry and overflow
of hydrocyclone were measured simultaneously. Samples were
simultaneously taken from both the feed and underflow, and then
the solid weight concentrations and the particle size distributions
of the samples were measured.

Fig. 2. The geometries of hydrocyclones. (a) Hydrocyclone A; (b) Hydrocyclone B;

(c) Hydrocyclone C.

Fig. 3. The mesh generation of the hydrocyclone.

Fig. 5. Schematic diagram of the experimental apparatus. (1) Stirring tank; (2)
pump; (3) flow meter; (4) pressure gauge; (5) hydrocyclone.
274 Q. Yang et al. / Separation and Purification Technology 74 (2010) 271–279

4. Results and discussion

4.1. Distribution of pressure profiles

Pressure profiles had been simulated using FLUENT computa-

tional fluid dynamics (CFD) software. The axis and the wall pressure
profiles of hydrocyclones are shown in Fig. 7. According to the axial
pressure profile, it can be found that the pressure in the cylindrical
section is the highest, followed by that in the first cone section and
that in the second cone section near the underflow outlet, while
the lowest one is the pressure in the overflow section. The isobars
are “V”-like type, descending from the wall to the axis. According to
the wall pressure profile, the isobars go down spirally rather than
be perpendicular to the z-axis of hydrocyclones, which shows that
the pressure in the equal-height areas axisymmetric to the central
axis in the single-inlet hydrocyclone is not the same. The compari-
son of the axial and wall surface pressure profiles of hydrocyclones
shows the cone combination has no significant influence on the
Fig. 6. Experimental equipment.
distribution of pressure, but really has effect on the value of pres-
sure drop in hydrocyclones. The pressure drop in hydrocyclone A is

Fig. 7. Axial and wall surface pressure profiles of hydrocyclones. (a) Axial (left) and wall surface (right) pressure profiles of hydrocyclone A; (b) axial (left) and wall surface
(right) pressure profiles of hydrocyclone B; (c) axial (left) and wall surface (right) pressure profiles of hydrocyclone C.
 Q. Yang et al. / Separation and Purification Technology 74 (2010) 271–279 275

significantly higher than that in hydrocyclones B and C, and that in

hydrocyclone B is slightly higher than that of hydrocyclone C. The
results show that the bigger the angle of the first cone is, the higher
the pressure drop will be.
The axis center pressure profiles of hydrocyclones are shown in
Fig. 8. According to the axis center pressure profiles, the pressure of
the axis center near the underflow is the highest, and declines from
the underflow to the overflow area before the minimum pressure
occurring in the cross-section of the vortex finder and the cylin-
drical section, and then increases gradually. The variation of the
pressure is found mainly in the cone section, and increases with
the cone angle. The negative pressure occurs in the cross-section of
the vortex finder and the cylindrical section in hydrocyclone A.

4.2. Distribution of three-dimensional velocities

The position of vortex finder tip is defined as Z = 0. The four-
dimensional velocity distributions in the position Z = −50, −150,
−250 and −400 of the hydrocyclones are studied separately. The
simulation results are described as follows.
4.2.1. Axial velocity
The axial velocity profiles in different cross-sections are shown
in Fig. 9. The axial velocity in each cross-section basically shows a
symmetric distribution, changing gradually from negative to pos-
itive in the direction from the wall to the centre. It can be seen
that the fluid along the wall moves to the underflow outlet, and
the downward velocity increases with its radial distance from
the hydrocyclone axis. The fluid around the axis moves to the
underflow outlet, and the upward velocity increases with its radial
distance from the hydrocyclone wall. For hydrocyclone A, because
of the great variation of cone angle, the axial velocity in the same
cross-section also changes greatly. In the first cone, the maximum
Fig. 8. Z-axis pressure curves of hydrocyclones.
downward velocity in radial section of hydrocyclone A is higher
than that in radial section of hydrocyclones B and C, while there is
little difference among the maximum upward velocity in the axis of
three hydrocyclones, as shown in Fig. 10. In conclusion, cone angle
has a certain influence on the particle velocity down the wall. The
particle velocity down the wall of the hydrocyclone with a small
cone angle is higher than that of the hydrocyclone with a larger
cone angle, but the combination of cone angle has no significant
influence on the upward velocity of the particles in the central axis.
4.2.2. Radial velocity
The radial velocity profiles in different cross-sections are shown
in Fig. 11. The radial velocity in the cross-section along the Z-axis
is not symmetrically distributed around the axis due to the influ-
ence of single inlet. It increases gradually from wall to axis and

Fig. 9. Axial velocity distributions of hydrocyclones.

276 Q. Yang et al. / Separation and Purification Technology 74 (2010) 271–279
Fig. 10. Axial velocity curves at central axis.
decreases sharply in the proximity of central axis. The variation of
cone angle has large influence on the radial velocity. The absolute
value of radial velocity increase with the value of cone angle, which
means the centrifugal force, also increases with the cone angle. The
combination of large-angle cone and small-angle cone in a hydrocy-
clone is usually used to enhance the separation performance when
the density difference of the two phases is too small.
4.2.3. Tangential velocity
Tangential velocity is the most important one among the three-
dimensional velocities, which is one of the key indexes of the
separation factor by determining the value of centrifugal force. It
is because the tangential velocity is greater in value than the other
Fig. 11. Radial velocity distributions of hydrocyclones.
two velocities, and what is more important is that, the centrifu-
gal force deriving from the tangential velocity is the prerequisite
for the separation of the two or more phases in hydrocyclones.
Fig. 12 gives out the tangential velocity profiles in different cross-
sections of hydrocyclones. The tangential velocity increases in the
same direction with the reduction of radius from wall to axis. In the
area near to the wall, the tangential velocity reaches its maximum
and decreases with the reduction of radius until it reaching to zero
in the central axis.
The cone combination has great influence on the value of tan-
gential velocity. The maximum tangential velocity of hydrocyclone
A is significantly higher than that of hydrocyclones B and C, and that
of hydrocyclone B is slightly higher than that of hydrocyclone C.
Thus, the variation of cone angles of a hydrocyclone influences the
tangential velocity profile in that the tangential velocity increases
with the angle of the first cone.
4.3. Capacity and flow split
The capacity and flow split were investigated under experimen-
tal method. The capacity is the volume flow of the feed, which
is directly related to the inlet pressure of the solid–liquid hydro-
cyclone. It is an important factor for a hydrocyclone. When the
inlet pressure is maintained constantly and the separation perfor-
mance of hydrocyclone could meet the technological requirements,
the larger the capacity of hydrocyclone the better. Under normal
operating conditions, the capacity has a direct relationship to the
pressure drops of inlet and overflow across the solid–liquid hydro-
cyclone. The relationship between the two pressure drops is also
important and can be used for control purposes.
The effect of different cone combinations on the capacity is
shown in Fig. 13. From Fig. 13, it can be seen that the volume flow of
the feed increased with the increase of inlet pressure or the pressure
 Q. Yang et al. / Separation and Purification Technology 74 (2010) 271–279 277

Fig. 12. Tangential velocity distributions of hydrocyclones.

drop between inlet and overflow. The hydrocyclones with different
cone combinations had basically the same capacities. The capacity
was mainly related to the structural size of the hydrocyclone.
The relationship between pressure drop and the flow split of
hydrocyclones is shown in Fig. 14. It can be seen that the flow split
decreases with increasing the pressure drop. When the inlet pres-
sure was larger than 0.2 MPa, the decrease of flow split became
gentle as the pressure drop increased. This indicates that the
increase of inlet flow has divided equally to the overflow and
underflow outlet by a certain percentage. The cone combinations
influenced the flow split of the hydrocyclone. When the angle of
the second cone remained unchanged, the larger the angle of the
first cone, the smaller the split of hydrocyclone.
4.4. Removal performance
The removal performance was also investigated under experi-
mental method.
4.4.1. Separation efficiency
The inlet Reynolds number is a key factor and has significant
effect on the separation efficiency of hydrocyclones. Fig. 15 shows
the effect of inlet Reynolds number on the separation efficiency
of hydrocyclones. It can be seen that the separation efficiency
increases first and then decreases as the inlet Reynolds number
increases. When the inlet Reynolds number is close to 85,000, the
separation efficiency of hydrocyclone B first reaches maximum. The

Fig. 14. Relationship between flow split and pressure drop of inlet to over let in
Fig. 13. Relationship between inlet flow rate and pressure drop of hydrocyclones. hydrocyclones.
278 Q. Yang et al. / Separation and Purification Technology 74 (2010) 271–279
Fig. 15. Relationship between the inlet Reynolds number and the separation effi-
ciency of hydrocyclones.
separation efficiency of hydrocyclone C reaches maximum when
the inlet Reynolds number is close to 105,000, and the separa-
tion efficiency of hydrocyclone A reaches maximum when the inlet
Reynolds number is close to 110,000. The maximum values of sepa-
ration efficiency of hydrocyclones A, B, and C are 80%, 88% and 83%,
respectively.
The different cone combinations have significant influence on
the separation efficiency of hydrocyclones. On the one hand, the
changes of two cone angles have the influence on high separation
efficiency area of the hydrocyclone. The effect degree of the changes
of two cone angles on the high separation efficiency area is in an
order as follows: hydrocyclone B > hydrocyclone C > hydrocyclone
A. The larger the two cone changes, the smaller the high separa-
tion efficiency area is. It is because every hydrocyclone with fixed
angle of cone has maintained high separation efficiency area. The
change of the cone angles results in the change of fluid flow inside
the hydrocyclone, and consequently the change of the high sep-
aration efficiency area. The larger the change of cone angle, the
smaller the overlap of high separation efficiency area is. On the
other hand, the change of two-cone angle affects energy consump-
tion in the hydrocyclone at the maximum separation efficiency.
The energy consumption in the hydrocyclone is directly related
to the inlet Reynolds number of the hydrocyclone. The bigger the
inlet Reynolds number, the higher the energy consumption. The
effect degree of the change of two cone angles on the energy con-
sumption is in an order as follows: hydrocyclone A > hydrocyclone
C > hydrocyclone B. The angles of the two cones change, the flow
field inside the hydrocyclone changes. The larger the flow field
inside hydrocyclone changes, the higher the energy consumption
as a result.
4.4.2. Grade efficiency
The grade efficiency curves of hydrocyclones are shown in
Fig. 16. It can be seen that the grade efficiency curves are “S”-like
type. The sharpness of grade efficiency curve of hydrocyclone A is
significantly lower than that of hydrocyclones B and C, and that of
hydrocyclone B is slightly larger than that of hydrocyclone C. The
cut size of hydrocyclones A, B, and C are almost the same. That is,
the cone combination has little effect on the cut size.
The smaller the change of the angle between the two cones,
the greater the sharpness of the grade efficiency curve; and vice
versa. Higher sharpness of the grade efficiency curve is preferable
in the classification process. The cone combination has impact on
the separation efficiency of different particle sizes. Hydrocyclone A
Fig. 16. Grade efficiency curves of hydrocyclones.
has significantly larger grade efficiency for the particles with size
smaller than 15 ␮m compared with hydrocyclones B and C. Because
the angle of the first cone of hydrocyclone A is much larger than that
of hydrocyclones B and C, and the tangential velocity is much bigger
than that of hydrocyclones B and C according to the simulation
results, there are more small particles go into the next cone by the
role of centrifugal force in hydrocyclone A. Therefore, it is better to
use larger changes of two cones in the design of hydrocyclone, if
focused on the uniformity of particle size separation process.
5. Conclusions
Simulate and experimental study of hydrocyclones with differ-
ent cone combinations for solid–liquid separation and flow field
inside hydrocyclones show that, the angle change of the cone
has significant effect on the value of three-dimensional velocities,
flow split, separation efficiency, energy consumption, and separa-
tion sharpness, but little effect on the distribution of pressure and
that of three-dimensional velocities, the capacity and cut size. For
solid–liquid separation, when the angle of the second cone does
not change, the bigger the angle of the first cone, the smaller the
split, the higher separation efficiency area, the bigger the pressure
drop, and the stronger the centrifugal force. The smaller the angle
of two cones changes, the larger the sharpness of the grade effi-
ciency curve, and the less the energy consumption; and vice versa.
If the angle of the first cone is much larger than that of the second
cone, more small particles will be separated into the underflow.
Acknowledgments
We would like to express our thanks for the sponsorship of the
New Century Excellent Talents in University (NCET-08-0777) and
Innovational Fund (0805H117000).
References
[1] G.I. Schwerzler, Recycling of glaze waste through hydrocyclone separation,
Powder Technol. 160 (2005) 135–140.
[2] L.C. Dickey, M.F. Dallmer, E.R. Radewonuk, N. Parris, M. Kurantz, J.C. Craig
Jr., Hydrocyclone separation of dry-milled corn, Cereal Chem. 74 (5) (1997)
676–680.
[3] G.A.B. Young, W.D. Wakley, D.L. Taggart, S.L. Andrews, J.R. Worrell, Oil–water
separation using hydrocyclones—an experimental search for optimum dimen-
sions, J. Petrol. Sci. Eng. 11 (1994) 37–50.
[4] N. Dhamo, Electrochemical oxidation of cyanide in the hydrocyclone cell, Waste
Manage. 16 (1996) 257–261.
[5] J.R. Wood, M. Grondin, A. Karnis, Characterization of mechanical pulp fines with
a small hydrocyclone. 1. The principle and nature of the separation, J. Pulp Paper
Sci. 17 (1991) J1–J5.
 Q. Yang et al. / Separation and Purification Technology 74 (2010) 271–279
[6] L.-Y. Chu, W.-M. Chen, X.-Z. Li, P.-K. Liu, Y. Du, M.-G. Lei, Y.-L.
Liu, Hydrocyclone–flocculation–filtration separation technique for waste
treatment—development and industrial application, in: Proceedings of the Sev-
enth World Filtration Congress, vol. I, Budapest, Hungary, May 20–23, 1996,
1996, pp. 354–358.
[7] J.J. Cilliers, S.T.L. Harrison, The application of mini-hydrocyclones in the con-
centration of yeast suspensions, Chem. Eng. J. 65 (1997) 21–26.
[8] L. Svarovsky, Hydrocyclones, Holt, Rinehart & Winston, Eastbourne, UK, 1984.
[9] J. Xu, Q. Luo, J. Qiu, Studying the flow field in a hydrocyclone with no forced
vortex. Part II: turbulence, Filtr. Separat. 27 (1990) 356–359.
[10] E.G. Arato, Reducing head or pressure losses across a hydrocyclone, Filtr. Sep-
arat. 21 (1984) 181–182.
[11] Z.-S. Bai, H.-L. Wang, S.-T. Tu, Removal of catalyst from oil slurry by hydrocy-
clone, Sep. Sci. Technol. 44 (2009) 2067–2077.
[12] L.-Y. Chu, Q. Luo, Hydrocyclone with high separation sharpness, Filtr. Separat.
31 (1994) 733–736.
[13] K. Rietema, Performance and design of hydrocyclones. IV: Design of hydrocy-
clones, Chem. Eng. Sci. 15 (1961) 320–325.
279
[14] J. Xu, Q. Luo, J. Qiu, Research on the preseparation space in hydrocyclones, Int.
J. Miner. Process. 31 (1991) 1–10.
[15] L. Zhao, M. Jiang, Y. Wnag, Experimental study of a hydrocyclone under cyclic
flow conditions for fine particle separation, Sep. Purif. Technol. 59 (2008)
183–189.
[16] I.J. Lin, Hydrocycloning thickening: dewatering and densification of fine par-
ticulates, Separat. Sci. Technol. 22 (1987) 1327–1347.
[17] L.-Y. Chu, W.-M. Chen, X.-Z. Lee, Effect of structural modification on hydrocy-
clone performance, Sep. Purif. Technol. 21 (2000) 71–86.
[18] K. Udaya Bhaskar, Y. Rama Murthy, M. Ravi Raju, J.K. Sumit Tiwari, N. Srivas-
tava, Ramakrishnan, CFD simulation and experimental validation studies on
hydrocyclone, Miner. Eng. 20 (1) (2007) 60–71.
[19] W.K. Evans, A. Suksangpanomrung, A.F. Nowakowski, The simulation of the
flow within a hydrocyclone operating with an air core and with an inserted
metal rod, Chem. Eng. J. 143 (1–3) (2008) 51–61.
[20] L.-Y. Wang, Z.-C. Zheng, Y.-X. Wu, J. Guo, J. Zhang, C. Tang, Numerical and exper-
imental study on liquid–solid flow in a hydrocyclone, J. Hydrodyn. Ser. B 21 (3)
(2009) 408–414.