chemical engineering research and design 9 0 ( 2 0 1 2 ) 21292134

Contents lists available at SciVerse ScienceDirect

Chemical Engineering Research and Design

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

Development of a new type high-efcient inner-cone

hydrocyclone

Lixin Zhao a, , Minghu Jiang a , Baorui Xu a , Baojun Zhu b

a Northeast Petroleum University, Daqing 163318, Heilongjiang, China
b PetroChina Liaohe Oileld Company, Panjin 124010, Liaoning, China

a b s t r a c t

Performance of a hydrocyclone can be inuenced by many factors, such as structural type, geometric parameters
and operation parameters, among which the structural type plays a very important role. The separation principle
of a typical hydrocyclone was introduced. Focused on a gasliquid separation, numerical simulation of the typical
hydrocyclone was carried out. The gas phase fraction distribution was analysed. It is shown that the separation effect
was not satisfactory. A revising idea was thus proposed and developed step by step, so a new inner-cone hydrocyclone
(ICH) was designed. The inner-cone structure was thought to provide a more stable ow eld for phase separation.
It functions like a gas carrier that is benecial for radially separated gas congregating on and growing into larger
gas bubbles. It also produces an upward pushing force to gas bubbles, so as to enhance the gasliquid separation
performance, although the ICH has lower inner tangential velocities than the typical hydrocyclone. Numerical sim-
ulation and experimental study veried the analysis. Pressure drop of the ICH is much lower than that of the typical
hydrocyclone; and the ICH has a wider scope for the change of liquid split ratio or inlet gasliquid ratio. Development
of the ICH would provide a new thought for the design of other separators.
2012 The Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.

Keywords: Hydrocyclone separator; Gasliquid separation; Numerical simulation; Experimental study; Inner-cone
hydrocyclone

1. Introduction The separation performance of hydrocyclones can be

inuenced by many factors (Jiang et al., 2002), and the
Hydrocyclone is a kind of multi-phase separator widely used hydrocyclonic separation mechanism is not yet under-
in petrochemical, mining and many other industries (Bradley, stood thoroughly. The research work is quite complicated
1965). It performs a separation function by utilising the den- although its structure is relatively simple (Slack et al., 2004;
sity difference among different immiscible phases (Svarovsky, Dyakowski and Williams, 1993; Nowakowski et al., 2004). Many
1984; Zhao et al., 2010). Hydrocyclone has many advantages researchers studied on the effects of geometric and operat-
such as higher separation efciency, smaller separator body, ing parameters on the pressure characteristics and separation
no moving parts inside, faster separation process, and so on. performance of hydrocyclones (Chu et al., 2000; Pasquier and
The hydrocyclones can be classied into several categories Cilliers, 2000; Ni, 2003; Belaidi and Thew, 2000; Zhao et al.,
according to the types of separating media. For two phase sep- 2008).
aration, there are gasliquid, gassolid, liquidsolid, oilwater, This paper mainly introduces the design idea and pro-
solidsolid hydrocyclones, etc.; and for three-phase separa- cess of a new type high-efcient hydrocyclone, inner-cone
tion, there are gasliquidsolid, gasoilwater, oilwatersolid, hydrocyclone (ICH), based on a gasliquid separation analy-
gassolidsolid hydrocyclones, etc. (Zhao and Li, 2006). sis. Experiments were performed to investigate the separation

Corresponding author at: College of Mechanical Science and Engineering, Northeast Petroleum University, 199 Fazhan Rd., High-Tech
Development Zone, Daqing 163318, Heilongjiang, China. Tel.: +86 459 6503853; fax: +86 459 6503853.
E-mail address: lx [email protected] (L. Zhao).
Received 19 December 2011; Received in revised form 6 March 2012; Accepted 20 May 2012
0263-8762/$ see front matter 2012 The Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.cherd.2012.05.013
2130 chemical engineering research and design 9 0 ( 2 0 1 2 ) 21292134

effect and pressure loss of the ICH. The design idea would also and moves upwards to overow outlet. For the segment of
be applied into the design of other kinds of separators. underow tube, its main function is to maintain a stable ow
for liquid drainage, and almost no separation process happens
there. Finally, the feed ow is divided into the overow which
Nomenclature carries most of the gas may still mixed with a part of liquid
phase, and the underow which contains most of the liquid
a centrifugal acceleration (m/s2 ) and some gas bubbles normally.
Fl ratio of overow liquid owrate to inlet liquid In order to enhance the separation performance of
owrate (%) hydrocyclones, much work has been done. Most of the
Qil inlet liquid owrate (m3 /h) research focused on changing the diameter, length, or cone
Qug underow gas owrate (m3 /h) angle of each part of standard hydrocyclone. Chu stud-
r radius (mm) ied the effect of different inlet pipes, vortex nders, cone
Rgl ratio of gas owrate at STP condition to liquid structures, underow pipes, and central insertions on hydro-
owrate at inlet cyclone performance (Chu et al., 2000). Noroozi studied
vt tangential velocity (m/s) the effects of different inlet chamber designs (exponen-
p pressure drop (MPa) tial, conical, quadratic polynomial body prole and standard
pio overow pressure drop (MPa) design) on hydrocyclone separation efciency (Noroozi and
piu underow pressure drop (MPa) Hashemabadi, 2011).
g dynamic viscosity of gas (mPa s) In addition, some researchers revised the structural style
l dynamic viscosity of liquid (mPa s) of hydrocyclones for the sake of improving the separa-
g density of gas (kg/m3 ) tion performance for specic immiscible mixtures. A revised
l density of liquid (kg/m3 ) hydrocyclone, air-sparged hydrocyclone (ASH), was investi-
gated by many researchers (Nieuwoudt et al., 1990; Beeby and
Nicol, 1993; Yalamanchili and Miller, 1995; Chu et al., 1996). In
the ASH, air is sparged radially through a porous cylinder and
2. Background dispersed by the high velocity of slurry in swirl ow, coun-
tercurrent to the froth phase rising axially in the centre of
2.1. Separation principle of a typical hydrocyclone the cylinder, so as to enhance oilwater separation efciency.
Nieuwoudt also carried out the study on the effect of different
Normal hydrocyclone, or standard hydrocyclone, is a typical underow congurations (with a bafe, a pedestal, or other
single-cone or double-cone separator, as shown in Fig. 1. Tak- insert section) on ASH separation performance (Nieuwoudt
ing a gasliquid separation for example, hydrocyclone is a et al., 1990). Some insert parts did enhance the separation per-
static separator based on centrifugal separation in a vortex formance. All these designs and research provide useful ideas
generated within the body. The feed ow, usually entering tan- for hydrocyclone structural optimisation design.
gentially into the hydrocyclone with a relatively high speed, In this research, the aim is to develop a high efcient
revolves around the inner wall of the hydrocyclone and moves gasliquid separator. The requirement is that zero gas exists
forward driven by the following feed ow. Basically there within the underow.
is much turbulent ow in the vortex cavity segment, and
an effective centrifugal force is formed preliminarily. When
moving into the main separation part, cone segment, the cen- 2.2. Performance of a typical hydrocyclone
trifugal force plays a key role. In that part, the relatively heavier
phase, liquid, moves outwards to the wall and simultaneously Initially a typical single-cone hydrocyclone, HC1, was used for
moves downwards to underow outlet, while the other phase, carrying out numerical simulation analysis. The main dimen-
gas, is forced to move inwards to the hydrocyclone centre sions of the HC1 are: separator diameter is 100 mm; separator
length 460 mm; inlet-tube diameter 12 mm; overow-tube
diameter 10 mm; overow-tube intrusion length 40 mm; cone
1 angle 30 ; underow-tube diameter 20 mm.
The adopted computational uid dynamics (CFD) soft pack
was FLUENT. Reynolds Stress Model (RSM) was employed as
2
a turbulent model. Based on the requirement of the actual
3 5 application, the simulated parameters were set up as follows:
liquid density l = 1.25 103 kg/m3 , liquid dynamic viscosity
l = 1.5 103 kg/m s; gas density g = 8.189 102 kg/m3 , and
6
its viscosity g = 8.411 106 kg/m s. Operating parameters for
the numerical simulation are as follows: inlet liquid owrate
Qil = 4 m3 /h; inlet gasliquid ratio Rgl = 0.20; liquid split ratio
4 Fl = 20%.
Boundary conditions were set up as follows: (1) Inlet con-
dition: gas and liquid phases are mixed uniformly; the inlet
(a) a single-cone hydrocyclone (b) a double-cone hydrocyclone mixture is assumed to ow vertically to the inlet cross sec-
tion; (2) outlet condition: outlet ows are assumed to be fully
1-overflow tube; 2-tangential inlet(s); 3-cone segment;
4-underflow tube; 5-large cone segment; 6-small cone segment developed; (3) solid wall condition: no slip is assumed. Inde-
pendency check of meshing has been done before carrying out
Fig. 1 Structural sketch of typical hydrocyclones. the numerical simulations.
 chemical engineering research and design 9 0 ( 2 0 1 2 ) 21292134 2131

Fig. 2 Gas phase fraction contour of HC1.

Gas phase fraction distribution contour is shown in Fig. 2.

It could be found that the gas phase was separated radially
from the mixture, while much gas still congregated near the Fig. 3 A testing photo of HC1.
underow. It must inuence the escape of the separated gas
phase from the overow outlet.
spokes were very thin. So the scheme shown in Fig. 4b was
Experimental study was carried out at an indoor labora-
accepted for further consideration.
tory. The experiments adopted a mixture of air and tap water
In theory the needle would push the radially separated gas
as testing mixture. Clearly, the density difference between air
if it was installed rightly and the gas core was just at the centre
and tap water is smaller than that of the two phases in the
as well. But in fact it is difcult to full the requirements. In
actual application described above, so if the separation of air
addition, the needle, due to its tiny diameter, would be shaken
and tap water was solved successfully, the actual requirement
would be handled well. In the experiments, larger gasliquid
ratio values were adopted so as to further verify and investi-
gate the separators.
Fig. 3 shows a testing photo of HC1. It could be found that
air was separated from water and driven to the centre, so an
air core was formed at the hydrocyclone centre. In the vortex
cavity segment the ow eld was not clear enough, while in
the cone segment the separation effect was much improved.
But the air core extended to the underow outlet. It results in
some air escaped within underow, and the separation effect
was much inuenced.

3. Design idea and procedure of an ICH

1 3
2 1
In order to design a high-efcient gasliquid separation hydro-
cyclone, initially a hydrocyclone with an upright needle at the
centre of the underow tube was considered as an optimis- 3
ing hydrocyclone scheme (as shown in Fig. 4). The needle was
designed for providing an additional upward pushing force to 1 2 1
the radially separated gas so as to hopefully enhance the sep-
aration efciency. Two ways could be done for fastening the (a) with wheel spokes (b) with a tangential outlet
needle at the centre. One was to use wheel spokes (Fig. 4a);
the other was to adopt a tangential underow outlet and x 1-upright needle; 2-spokes; 3-tangential outlet
the needle at the bottom centre (Fig. 4b). It was thought that
the wheel spokes would destroy nearby ow eld even if the Fig. 4 Revised hydrocyclones with needle.
2132 chemical engineering research and design 9 0 ( 2 0 1 2 ) 21292134

(a) enlarge the needle diameter (b) change to a cone needle Fig. 7 Liquid-ow path-line contrast between HC1 and the
ICH.

original hydrocyclone structure; the outer cone structure of

hydrocyclone was eliminated, i.e., the outer wall of hydrocy-
clone was revised to a cylinder shape (Fig. 5d).
Inside the new ICH, the vortex cavity, i.e., the part between
inlets and the inner-cone point keeps the same with original
HC1. The annular cross-section area between the inner cone
and the cylindrical outer wall decreases from top to bottom of
the inner cone. This structure can accelerate the mixed uid
and play an important role for phase separation, just like what
the cone segment of HC1 does.
Numerical simulation and experimental study were per-
formed so as to investigate the design.

4. Numerical simulation analysis and

(c) enlarge the inner cone (d) a new ICH
experimental investigation
Fig. 5 Design procedure of the ICH.
4.1. Numerical simulation analysis
when the hydrocyclone is operating, and then lead to a poor
performance of hydrocyclone separation. In order to improve Numerical simulation results of the ICH are shown in Figs. 68.
it, a feasible way is to enlarge the diameter of the needle and Some results are compared with those of HC1.
that of the underow tube as well, as shown in Fig. 5a. Con- Fig. 6 shows the tangential velocity distributions at the
sidering the effect of a larger needle on nearby ow eld, it mid cross section of the two hydrocyclones. Tangential veloc-
should be sharpened to a cone shape (Fig. 5b). Based on it, we ity was thought to be one of the main factors that inuence
continued to enlarge the whole cone, not just the diameter, hydrocyclone separation. Normally it was believed that higher
but also the length (Fig. 5c). During this process, diameter of tangential velocities are benecial for making a better hydro-
the underow tube was enlarged furthermore, and cone angle cyclone separation effect, while it can be found that the
of the hydrocyclone cone segment was decreased as well. Fol- tangential velocity values of the ICH are lower than those of
lowing the trend of this idea, a totally new design scheme, an HC1.
ICH, emerged, i.e., the inner cone was enlarged continuously Fig. 7 shows the ow path lines of the two hydrocyclones.
till its top reached the upper section of the cone segment in It indicates that liquid path lines of the separators are of the
shape of helixes. The helix angle of the path lines inside HC1
increases with movement of the liquid from inlet to under-
ow outlet, while the path-line helix angle of the ICH keeps
nearly the same throughout the separator length, which can
remain a more stable ow eld and provide a better condition
for phase separation. It is surely believed that the design of the
tangential underow outlet in the ICH can make uid main-
tain a relatively longer and more effective revolving movement
inside the separator rather than that in HC1. Furthermore, the
separator length can also be shortened, and pressure drop of
the separator can be effectively reduced.
Gas fraction simulation result of the ICH is shown in Fig. 8.
It indicates that the gas was separated pretty well inside the
ICH. The radially separated gas congregated at the centre of
Fig. 6 Contrast of tangential velocity distributions the ICH. It can also be found that the separation performance
between HC1 and the ICH (mid cross section). was much improved comparing with the result of HC1 (Fig. 2).
 chemical engineering research and design 9 0 ( 2 0 1 2 ) 21292134 2133

Fig. 9 A testing photo of the ICH.

0.35

0.30

0.25
p (MPa)

0.20
Fig. 8 Gas phase fraction contour of the ICH.
0.15

0.10
pio of HC1
Therefore, it is thought that there must be other factors that piu of HC1
are more important than tangential velocity for inuencing 0.05 pio of ICH
separation performance of the ICH. piu of ICH
0
The main difference of the new ICH from HC1 is the inner- 2.0 2.5 3.0 3.5 4.0
cone structure. The inner-cone structure can provide a solid
3
gas carrier which is benecial for radially separated gas gath- Qil (m /h)
ering on and growing into larger gas bubbles. Although the
Fig. 10 Relationship between inlet liquid owrate and
design can hardly change the radial velocities that the gas
pressure drops.
bubbles have, it can surely enlarge the gathering space for
gas bubbles due to the structural change from a circular
owrate is 3.5 m3 /h, while for HC1, the overow pressure drop
cross-section area in the cone segment of HC1 to an annular
was 0.30 MPa and underow 0.23 MPa. The values of the ICH
cross-section area in the inner-cone segment of the new ICH.
were lower than those of HC1 by 36.7% and 30.4%, respectively.
The inner-cone structure can also provide an upward pushing
It indicates that the energy-saving effect of the new ICH is
force to the radially separated gas at the separator centre, so
prominent.
as to enhance separation performance of the ICH.
Fig. 11 shows the relationship between liquid split ratio Fl
and underow gas owrate Qug with a gasliquid ratio of 1.0
4.2. Experimental investigation

0.14
Experimental study was then carried out in order to investi-
gate actual separation performance of the new ICH, and make 0.12
a contrast with the results of HC1. 0.10
Qug (m3/h)

A testing photo of the ICH is shown in Fig. 9. It could be

0.08
found that the air mainly concentrated above the inner-cone
part, almost no air existed below the cone point, let alone in 0.06
the tangential underow tube. The separation effect seemed 0.04 HC1
to be very satisfactory by phenomenon observation. ICH
0.02
Fig. 10 shows the relationship between inlet liquid owrate
Qil and pressure drop p with a liquid split ratio Fl of 10% and 0
0 5 10 15 20 25
a gasliquid ratio Rgl of 1.0. It indicates that, for the ICH, with Fl (%)
the rise of inlet liquid owrate, both overow pressure drop
pio and underow pressure drop piu increased. The overow Fig. 11 Relationship between liquid split ratio and
pressure drop was 0.19 MPa and underow 0.16 MPa when the underow gas owrate.
2134 chemical engineering research and design 9 0 ( 2 0 1 2 ) 21292134

0.06 respectively when with a 3.5 m3 /h owrate, 10% liquid split

0.05 ratio and 1.0 gasliquid ratio.
(4) The inner-cone hydrocyclone does enhance the separation
Qug (m3/h)

0.04
effect theoretically and practically. It has a wider suitable
0.03 range for the change of liquid split ratio or gasliquid ratio.
HC1 (5) The inner-cone design would provide a new idea for hydro-
0.02
ICH
cyclone design.
0.01

0 Acknowledgement
0 0.5 1.0 1.5 2.0
Rgl
We would like to express our thanks for the sponsorship of
Fig. 12 Relationship between gasliquid ratio and the Program for New Century Excellent Talents in Heilongjiang
underow gas owrate. Provincial University, China (Contract No. 1155-NCET-003).

and an inlet liquid owrate of 3 m3 /h. It could be found that, for

References
the ICH, the underow gas content kept zero if only the liquid
split ratio was larger than 2%, while in HC1 the gas could be
Beeby, J.P., Nicol, S.K., 1993. Concentration of oil-in-water
detected at underow outlet only except the single point of 7%
emulsion using the air-sparged hydrocyclone. Filtr. Sep. 30,
split ratio. 140, 141146.
Fig. 12 shows the relationship between gasliquid ratio Rgl Belaidi, A., Thew, M.T., 2000. Drop size effects on a de-watering
and underow gas owrate Qug with an inlet liquid owrate hydrocyclone. In: Proceedings of the Vortex Separation, York,
of 3 m3 /h and a liquid split ratio of 3%. It shows that, for the England, pp. 119129.
ICH, the underow gas content was totally zero even if the Bradley, D., 1965. The Hydrocyclone. Pergamon Press, London.
Chu, L.Y., Luo, Q., Yu, R.H., 1996. Concentration and classication
gasliquid ratio reached 1.6, while for HC1, the gas content at
characteristics in a modied air-sparged hydrocyclone (ASH).
underow outlet increased with the rise of gasliquid ratio;
Int. J. Miner. Process. 48, 7393.
when gasliquid ratio was higher than 0.7, the gas appeared Chu, L.Y., Chen, W.M., Lee, X.Z., 2000. Effect of structural
at underow outlet even if the liquid split ratio Fl was 10%. It modication on hydrocyclone performance. Sep. Purif.
illustrates that the new ICH has a very large gasliquid ratio Technol. 21, 7186.
range for keeping higher separation efciency. Dyakowski, T., Williams, R.A., 1993. Modelling turbulent ow
It is well known that a larger tangential velocity can pro- within a small diameter hydrocyclone. Chem. Eng. Sci. 48,
11431152.
duce a stronger centrifugal separation force, which makes gas
Jiang, M.H., Zhao, L.X., Wang, Z.C., 2002. Effects of geometric and
bubbles move towards the centre more easily. However, other operating parameters on separation performance for
factors, including a stable ow eld for phase separation, a hydrocyclones. In: Proceedings of the 12th International
sufcient space for radially separated gas congregating at the Offshore and Polar Engineering, vol. I, Kitakyushu, Japan, pp.
centre, an upright pushing force for the radially separated 102106.
gas to move upwards, etc., can also enhance the gasliquid Ni, L.Y., 2003. Effect of intermittent ow on hydrocyclones
separation. It is these factors that can be utilised to design oilwater separation efciency. Chin. Filter Sep. 13, 2224.
Nieuwoudt, D.J., Van Deventer, J.S.J., Reuter, M.A., Ross, V.E., 1990.
hydrocyclones. The design of the ICH improves hydrocyclone
The inuence of design variables on the otation of pyrite in
separations by successfully employing an inner-cone struc- an air-sparged hydrocyclone. Miner. Eng. 3, 483499.
ture. Noroozi, S., Hashemabadi, S.H., 2011. CFD analysis of inlet
This new ICH design has been granted an Invent Patent chamber body prole effects on de-oiling hydrocyclone
in China. The study focused on gasliquid separation, but it is efciency. Chem. Eng. Res. Des. 89, 968977.
surely believed that the new design idea would be also applied Nowakowski, A.F., Cullivan, J.C., Williams, R.A., Dyakowski, T.,
2004. Application of CFD to modelling of the ow in
to the design of other two-phase separation hydrocyclones,
hydrocyclones. Is this a realizable option or still a research
even three-phase separation hydrocyclones.
challenge? Miner. Eng. 17, 661669.
Pasquier, S., Cilliers, J.J., 2000. Sub-micron particle dewatering
5. Conclusions using hydrocyclones. Chem. Eng. 80, 283288.
Slack, M.D., Del Porte, S., Engelman, M.S., 2004. Designing
(1) Simulation and experimental results indicated that the automated computational uid dynamics modelling tools for
hydrocyclone design. Miner. Eng. 17, 705711.
gas content could not be removed thoroughly just by using
Svarovsky, L., 1984. Hydrocyclones. Holt, Rinehart and Winston,
a typical hydrocyclone.
London.
(2) The design of a tangential underow outlet in the inner- Yalamanchili, M.R., Miller, J.D., 1995. Removal of insoluble slimes
cone hydrocyclone (ICH) can make uid maintain a from potash ore by air-sparged hydrocyclone otation. Miner.
relatively longer and more effective revolving movement Eng. 8, 169177.
inside the hydrocyclone rather than that in the typical Zhao, L.X., Li, F., 2006. Centrifugal Separation Technology. Press of
hydrocyclone. The inner-cone structure also provides a gas Northeast Forestry University, Harbin.
Zhao, L.X., Jiang, M.H., Wang, Y., 2008. Experimental study of a
carrier that is benecial for the radially separated gas gath-
hydrocyclone under cyclic ow conditions for ne particle
ering on and growing into larger gas bubbles, and forms separation. Sep. Purif. Technol. 59, 183189.
an upward pushing force to gas bubbles, so as to enhance Zhao, L.X., Jiang, M.H., Li, F., 2010. Experimental study on the
hydrocyclone separation performance. separation performance of air-injected de-oil hydrocyclones.
(3) For the ICH, the overow and underow pressure drop val- Chem. Eng. Res. Des. 88, 772778.
ues were lower than those in HC1 by 36.7% and 30.4%,