CHAPTER ONE

Hydrocyclones
Chris Aldrich
Department of Mining Engineering and Metallurgical Engineering, Western Australian School of Mines,
Curtin University of Technology, Perth, WA, Australia

Contents
Nomenclature 2
1. Background 3
2. Basic Design 5
2.1 Geometry 5
2.2 Inlet Design 5
2.3 Materials of Construction 6
3. Characterization of Performance 7
3.1 Partition Curves 7
3.2 Factors Affecting the Performance of Hydrocyclones 8
4. Hydrocyclone Models 9
4.1 Fundamental and Empirical Models 9
4.1.1 Equilibrium Orbit Theory 10
4.1.2 Crowding Theory 10
4.1.3 Residence Time Theory 10
4.1.4 Turbulent Two-Phase Flow 11
4.1.5 Empirical Modelling 11
4.2 Numerical Models 12
4.2.1 Modelling Turbulence 12
4.2.2 Model Validation 13
5. Scale-up and Design 15
5.1 Scale-up 15
5.2 Hydrocyclone Networks 16
6. Monitoring and Control of Hydrocyclones 16
6.1 Control 16
6.2 Soft Sensors 16
6.3 Online Monitoring of Hydrocyclones 17
6.4 Example of Underflow Monitoring 17
7. Future Developments 19
References 20

Ó 2015 Elsevier Ltd.

j
Progress in Filtration and Separation
http://dx.doi.org/10.1016/B978-0-12-384746-1.00001-X All rights reserved. 1
2 Chris Aldrich

GLOSSARY
Cut Size diameter of particle with equal probability to report to the overflow and under-
flow of the cyclone; often a measure of the performance of the classifier.
Fish hook effect the phenomenon where partition curves in hydrocyclones do not have a
sigmoidal shape, but a shape resembling that of a fish hook.
Partition curve a graphical representation of the recovery of each particle size in the hydro-
cyclone underflow in relation to its availability in the feed; also known as a Tromp curve.
Spigot the outlet part or apex at the lower conical end of a hydrocyclone.
Vortex finder a short, removable top mounted overflow pipe in a hydrocyclone that ex-
tends a short distance into the cylindrical body part of the cyclone to prevent short-
circuiting of feed directly into the overflow.

NOMENCLATURE

Symbol Meaning SI Units

a Coefficient representing sharpness of separation e

q Cone angle of hydrocyclone 

m Viscosity of liquid Pa s
rf Density of fluid kg/m3
rs Density of solids kg/m3
A0 Cross-sectional area of particle m2
CD Drag coefficient e
Cv Volumetric concentration of solids in the feed e
Dc Diameter of hydrocyclone cylinder m
Di Diameter of inlet m
Do Diameter of overflow m
Du Diameter of underflow m
d Particle diameter m
d25 Diameter of particle with 25% probability to report to m
the underflow of the cyclone
d50 Cut size; diameter of particle with equal probability to m
report to the overflow and underflow of the
cyclone
d50(c) Corrected cut size
d75 Diameter of particle with 75% probability to report to m
the underflow of the cyclone
Eu Euler number, Eu ¼ 2DP rv2 e
FC Centrifugal force N
FD Drag force N
Fi Material parameter, i ¼ 1, 2, 3 e
h Distance between vortex finder and apex of cyclone m
I Inefficiency of separation e
Hydrocyclones 3

dcont'd
Symbol Meaning SI Units
k Constant
L Length of cyclone m
l Length of vortex finder m
P Pressure Pa
Qf Throughput m3/s
R Fraction of liquid in the feed recovered in the e
underflow stream
Re Reynolds number, Re ¼ rvd m e
r Radial position of particle inside hydrocyclone m
ðrsrÞvd50
2
Stk50 Stokes number, Stk50 ¼ 18mD c
e
ur Radial velocity of particle in hydrocyclone m/s
vr Radial velocity of fluid in hydrocyclone m/s
vq Tangential velocity of fluid in hydrocyclone m/s
Vp Volume of particle m3
v Velocity m/s
y Uncorrected efficiency e
y0 Reduced efficiency e

1. BACKGROUND
Hydrocyclones have been in use in industry since the 1940s, although
the first patent can be traced back to the nineteenth century (Bretnai, 1891).
Owing to their simple design, low cost, easy operation, and low maintenance,
they have assumed an important role in the separation of solids and liquids.
Although hydrocyclones are widely used at present, such as in closed circuit
grinding (Casali et al., 1998), desliming (Yalamanchili and Miller, 1995),
liquid clarification (Puprasert et al., 2004), degritting (Murray, 1980), and
thickening operations (Woodfield and Bickert, 2004; Yang et al., 2004),
the phenomena leading to separation are still not fully understood yet.
In a hydrocyclone, a slurry enters through a tangential inlet, giving rise to
a vortex in the stationary body. The particles and fluid are accelerated
centrifugally and separation occurs in the radial direction. Denser materials
migrate to the outer wall of the hydrocyclone chamber, while less dense ma-
terials move toward the inner axis. Flow in the cyclone cylinder is charac-
terized by two vortices that flow in opposite directions. Denser flow tends
to travel along the primary vortex to the underflow and less dense material
travels along the secondary vortex in the opposite direction to the overflow,
as indicated in Figure 1.
4 Chris Aldrich

Figure 1 Features of a hydrocyclone in operation.

According to classical theory, particles in the cyclone are subject to two

opposing forces, i.e., an outwardly acting centrifugal force and an inwardly
acting drag force, as indicated in Figure 2. Particles are separated by the
accelerating centrifugal force based on size, shape, and density, while the
drag force moves slower settling particles to the low pressure zone along
the axis of the cyclone, where they are carried upward through the vortex
finder to the overflow. Figure 2 also shows the equations representing the
classical force balance acting on a spherical particle with diameter d, at a po-
sition r in the cyclone, i.e., at equilibrium, the centrifugal force in the
cyclone is equal to the drag force and the buoyancy forces acting on the par-
ticle, or FC  FB  FD ¼ 0.

Figure 2 Force balance on a particle in a hydrocyclone.

Hydrocyclones 5

2. BASIC DESIGN
2.1 Geometry
Various designs of hydrocyclones have been proposed in order to
exploit the use of inertial and gravitational forces to separate particulate matter
from fluids. The main parameter of a hydrocyclone is its diameter, i.e., the
inside diameter of the cylindrical feed chamber. This diameter can range
from 10 mm to 2.5 m in commercial hydrocyclones, capable of separating
particles of sizes ranging from 1.5 to 300 mm (and densities of approximately
2700 kg/m3). This is followed by the area of the inlet nozzle at the point of
entry into the feed chamber. The inlet nozzle is usually a rectangular orifice,
with the larger dimension parallel to the axis of the cyclone. Typically, the
area of the inlet nozzle is approximately 5% of that of the square of the
diameter of the cyclone.
The primary function of the vortex finder is to control both the sep-
aration and the flow leaving the cyclone. The size of the vortex finder is
approximately 35% of the diameter of the cyclone. The upper cylindrical
section of the cyclone is located between the feed chamber and the
conical section, and it has the same diameter as the feed chamber. It
serves to extend the length of the cyclone in order to increase the reten-
tion time of particles. Its length is typically equal to the cyclone
diameter.
The lower conical section further adds to the retention time of particles
and has an included angle normally between 10 and 20 . The conical section
terminates in the apex orifice and the critical dimension of the orifice is its in-
side diameter at the discharge point. It has to be of sufficient size to permit
solids in the underflow to exit without plugging. The normal minimum
orifice size is typically from 10% to 35% of the cyclone diameter. Finally, a
splash skirt below the apex often helps to contain the underflow slurry.

2.2 Inlet Design

The feed inlet of the cyclone can be circular or rectangular, while the inlet
opening can also vary in size. Different inlet configurations are shown in
Figure 3. Modification of the inlet section of hydrocyclones is seen as a sim-
ple approach to control particle cut size and to improve the classification
performance of hydrocyclones (Nenu and Yoshida, 2009). Modifications
include the use of movable circular guide plates (Yoshida et al., 2006) or
rotational top blades (Norimoto et al., 2004).
6 Chris Aldrich

Figure 3 Feed inlet configurations in hydrocyclones.

Likewise, Nenu and Yoshida (2009) have compared one-inlet with two-
inlet cyclones and have concluded that the particle collection efficiency of
the two-inlet cyclone was better than that of the one-inlet cyclone with
the same total flow rate, owing to enhancement of the tangential velocity
profile of the particles in the former. The particle collection efficiency in
the two-inlet cyclone was also found to be marginally better than that in
the one-inlet cyclone under the same pressure drop. Similar results were
observed by other authors for gas cyclones (Lim et al., 2003; Zhao et al.,
2004) and hydrocyclones (Yoshida et al., 2006).
Wet size classification of particulate materials with a hydrocyclone with an
axial inlet was investigated by Yalcin et al. (2003) as an alternative to tangential
inlet cyclones traditionally used in the mineral processing industry. Experi-
mental work was done with copper–nickel mill tailings having a particle
size of 91% passing 300 mm at different inlet pressures, feed pulp densities,
and vortex finder lengths. Higher throughputs, coarser cut sizes in relatively
dilute pulps, and greater flexibility and control over the cyclone separation
process were observed in comparison with tangential inlet cyclones.

2.3 Materials of Construction

The construction of hydrocyclones varies widely, depending on the applica-
tion, but the majority of the designs include metal housings with replaceable
liners. The most common liner material is natural gum rubber, owing to its
comparatively low cost, high resistance to wear, and ease of handling. These
liners are not suitable in conditions where temperatures may exceed approx-
imately 60  C or in aggressive chemical environments, such as where the
slurry may contain oil or large amounts of other hydrocarbons. Under these
circumstances, other elastomers, such as neoprene may perform better. In
addition, urethane may be used in areas exposed to relatively fine solids.
In the presence of highly abrasive slurries, ceramic materials such as sili-
con carbide can be used for the apex orifice, as well as other areas which may
Hydrocyclones 7

require protection, such as the lower cone or vortex finder (Madge et al.,
2004). Nihard, a nickel-based steel alloy with a martensitic microstructure,
has also proven to be an acceptable wear material, especially for vortex
finders and other areas which require strength as well as abrasion resistance.

3. CHARACTERIZATION OF PERFORMANCE
The practical range of particle sizes that can be classified by hydrocy-
clones is from 40 mm to 400 mm, with some specialized applications sepa-
rating fines in the submicron range (Endres et al., 2012) or as coarse as
1000 mm. Operating pressures range from 50 kPa for large units to 1 MPa
for smaller ones (Cilliers, 2000).

3.1 Partition Curves

The commonest approach to describe the efficiency of a cyclone is through
its partition or performance curve in which the mass fraction of each particle
size in the feed that reports to the underflow is related to the particle size
itself, as indicated in Figure 3. The sharpness of the cut is indicated by the
slope of the central portion of the graph, and the larger this slope (the closer
to the vertical), the more efficient the cyclone.
Correction of the classification curve may be necessary to account for
solids of all sizes that are entrained in the underflow in direct proportion
to the fraction of feed water reporting to the underflow. The corrected
partition curve ðy0 Þ can be obtained from the uncorrected curve ðyÞ by using
Eqn (1). In this equation, y is the actual mass fraction of a given particle size
reporting to the underflow, y0 is the corrected mass fraction of a given par-
ticle size reporting to the underflow, and R is the fraction of liquid in the
feed recovered in the underflow stream. This is indicated by the blue (un-
corrected) curve and black (corrected) curve in Figure 4.
yR
y0 ¼ (1)
1R
This efficiency or imperfection ðIÞ can be expressed by using the points
at which 75% and 25% of the feed particles report to the underflow (d75 and
d25 , respectively) in addition to the cut size d50 , i.e., the particle size having
an equal probability of reporting to the overflow or underflow, as follows:
d75  d25
I¼ (2)
2d50
8 Chris Aldrich

Figure 4 Typical reduced efficiency curve for a hydrocyclone showing recovery as a

function of normalized particle size.

The so-called fish hook effect may occur in the partition curve when
progressively higher partition numbers are observed for particle sizes finer
than that at the minimum partition value (Kraipech et al., 2002; Neesse
et al., 2004a).
The effect is more prevalent in smaller hydrocyclones and may be attrib-
uted to turbulent dispersion of the particles. Moreover, under these circum-
stances water recovery may be considerably lower than the lowest observed
partition value and correction of such curves would be meaningless. Fish
hook partition curves can be modelled by summation of a corrected parti-
tion curve (such as represented by Eqn (3)) and the product of an inverted
partition curve and a bypass fraction (Cilliers, 2000).

eaX  1
y0 ¼ (3)
eaX þ ea  2

3.2 Factors Affecting the Performance of Hydrocyclones

In general, the following trends or heuristics have been accumulated from
experience. Classification efficiency depends on the cyclone size (diameter)
and geometry, generally decreases with an increase in feed solids concentra-
tion and increase in viscosity. Efficiency is increased by limiting water to the
underflow.
Hydrocyclones 9

Table 1 Effect of cyclone design and operating variables on performance

Cyclone Di/Dc Do/Dc L/Dc l/Dc q

Bradley 0.133e0.143 0.200 6.850 0.333 9

Krebs 0.267 0.159 5.874 - 12.7
Rietema 0.280 0.340 5 0.400 15e20

Table 2 Geometric proportions of hydrocyclones (Castilho and Medronho, 2000,

Silva et al. (2012))
Sharpness of
Factor Throughput (Q) Cut size (d50) classification

Cyclone diameter þ þ þ
Feed inlet þ e e
Vortex finder diameter þ þ þ
Spigot diameter þ e e
Cone angle N/A þ þ
Volumetric feed solids þ þ e
concentration
Pressure drop þ e þ Or -
Free vortex height þ þ þ

Asomah and Napier-Munn (1997) have found that angles of inclination

exceeding 45 with the vertical axis of the hydrocyclone can play a signifi-
cant role in its operation.
The effects of cyclone geometry (summarized in Table 2) and opera-
tional variables on the throughput, cut size, and sharpness of classification
in hydrocyclones are summarized in Table 1, with “þ” indicating an in-
crease and “–” indicating a decrease in the performance criterion with an in-
crease in the factor (Table 2).

4. HYDROCYCLONE MODELS
A large variety of hydrocyclone models have been proposed to
estimate separation efficiencies of solid particles and pressure drops in these
devices. These include empirical models encapsulating experimental data by
use of fitted formulae or equations, as well as semiempirical models based on
equilibrium orbit theory, residence time, and turbulent flow theory.

4.1 Fundamental and Empirical Models

These models are typically based on heuristics or some theoretical descrip-
tion of the phenomena in a hydrocyclone, as follows.
10 Chris Aldrich

4.1.1 Equilibrium Orbit Theory

According to the equilibrium orbit theory (Bradley and Pulling, 1959), first pro-
posed by Driessen (1951), a particle attains an equilibrium radial position in the
cyclone, where its terminal settling velocity is equal to the radial velocity of the
liquid (Coelho and Medronho, 2001). At this point, the centrifugal force on
the particle is equal to the drag force in the opposite direction. If the liquid
at that point flows inward towards the central axis of the hydrocyclone, the
particles will move with the liquid to the overflow, and conversely, if the liquid
flows outward, away from the central axis of the hydrocyclone, the particles
will move to the wall and become separated through the underflow.
Conditions defining an equilibrium orbit can be defined as follows:
1
FD ¼ CD ðvr  ur Þ2 rf A0 (4)
2
Here, FD is the drag force experienced by the particle, CD is the drag
coefficient, vr is the radial velocity of the fluid, ur is the radial velocity of
the particle, rf is the density of the fluid, and A0 is the cross-sectional
area of the particle.

vq2

FC ¼ Vp rs  rf (5)
r
The centrifugal force FC is represented by Eqn (5) as a function of the
tangential component of the particle velocity vector, vq , the volume of
the particle, Vp , and the difference between the fluid ðrf Þ and particle den-
sities ðrs Þ. When the drag force and the centrifugal forces are equal, the par-
ticle finds itself in an equilibrium orbit.

4.1.2 Crowding Theory

Crowding theory (Bloor et al., 1980; White, 1991) is based on the observation
that at higher feed concentrations, separation size is chiefly determined by the
discharge capacity of the spigot and the feed size distribution (e.g., Slechta and
Firth, 1984). Theoretically, any cut size within the feed size distribution can be
obtained by controlling the outlet dimensions of the hydrocyclone.

4.1.3 Residence Time Theory

Residence time theory models (Rietema, 1961; Dwari et al., 2004) are not
concerned with particular flow phenomena, other than that a particle can be
considered separated, if it can traverse the distance to the cyclone wall region
within the residence time of the particle in the hydrocyclone.
Hydrocyclones 11

4.1.4 Turbulent Two-Phase Flow

In turbulent two-phase models (Neesse et al., 2004a), separation arises
owing to the turbulent cross flow moving perpendicular to the direction
of the force field.

4.1.5 Empirical Modelling

Empirical models are not based on any specific theory of hydrocyclone
operation and are derived by fitting models to experimental data. A vari-
ety of these were constructed in the previous century to predict the
performance of industrial equipment, e.g., Shepherd and Lapple, (1939),
Lapple, (1951), Leith and Licht (1972). More recently, models have
been proposed by among other Nageswararo et al. (2004) and Kraipech
et al. (2006).
One of the most widely used empirical models for hydrocyclones is that
of Plitt (1976) and Lynch and Rao (1975). In the Plitt model, which has also
been incorporated in a number of commercial steady simulation packages
(Cilliers and Hinde, 1991), such as DYNAFRAG (Flament et al., 1993),
the corrected cut size in micron is expressed as


F1 39:7Dc0:46 Di0:6 Do1:21 m0:5 exp 0:063Cv
d50ðcÞ ¼  k (6)
rs 1
Du0:71 h0:38 Q0:45
f 1:6

In this model, Dc , Di , Do , and Du represent the inside diameters in cm of

the hydrocyclone, inlet, vortex finder, and apex, respectively, Cv is the
volumetric concentration of the solids in the feed, m is the liquid viscosity
(cP), rs is the density of the solids (g/cm3), F1 is a material specific constant
that needs to be determined from tests with the feed material, Qf is the feed
low rate (L/min), h is the distance between the vortex finder and the apex
(cm), and k is a hydrodynamic exponent that needs to be estimated exper-
imentally (default value of k ¼ 0:5 for laminar flow conditions).
Likewise, the volumetric flow rate of the slurry to the hydrocyclone can
be expressed by

0:49
F2 P 0:56 Dc0:21 Di0:53 h0:16 Du2 þ Do2
Qf ¼ (7)
expð0:0031Cv Þ
Here, like F1 , F2 is a material-specific constant that needs to be deter-
mined from tests with the feed material and P is the pressure drop across
the cyclone (kPa). Similar equations for the efficiency of the hydrocyclone,
12 Chris Aldrich

as well as the water split between the underflow and overflow are reported by
Plitt (1976).
Implicit models based on artificial neural networks (Stange, 1993; Van
der Walt et al., 1993) and genofuzzy systems (Karr et al., 2000) have been
proposed more recently with the growing use of machine-learning algo-
rithms in process engineering. These models are theoretically more powerful
than explicit equations, but are also more complex to use and have not
found general acceptance in industry yet.

4.2 Numerical Models

With advances in computer hardware and software, numerical modelling of
hydrocyclones is becoming more popular (Chen et al., 2000; Brennan,
2006; Delgadillo and Rajamani, 2007). Numerical methods are used to pre-
dict the multiphase flow fields in cyclones by solution of the basic equations
governing flow, including the effects of turbulence and liquid–solid interac-
tions. Numerical models can be categorized according to their characteris-
tics, i.e., as two-dimensional or three-dimensional models, or as Eulerian
(Dueck et al., 2000) or Lagrangian models. In the latter, the observer follows
an individual fluid element moving through space and time, while in the
former, the focus is one specific location in space through which fluid flows
as time passes. In the Eulerian approach, or mixture model, the particle
phases are treated as a pseudo-continuum and the transport equations for
phase concentrations and phase momentum are solved. State-of-the-art
approaches currently tend to focus on three-dimensional Lagrangian models
(Dyakowski et al., 1999; Nowakowski et al., 2004; Schuetz et al., 2004).
Although the particle and fluid equations can be treated independently
in dilute slurries, this is not the case as the particle size concentrations in-
crease, since this affects the density of the mixtures as well as their effective
viscosities (Pericleous and Rhodes, 1986).
The air core in the cyclone is treated similar to the particles. Density dif-
ferences between the air and the liquid give rise to centrifugal forces acting
on the air, which leads to a slip velocity of the particles, directed toward the
axis of the cyclone. Over time, air accumulates along the cyclone axis to
form the air core.

4.2.1 Modelling Turbulence

Fluid motion is usually represented by steady state, incompressible Navier–
Stokes flow and continuity equations. Turbulence is represented by the
effective pulp viscosity, which is variant under local flow conditions. These
Hydrocyclones 13

include simple mixing length models (Pericleous and Rhodes, 1986) or

higher order turbulence models, such as the k-ε model, requiring equations
to be solved for the kinetic energy (k), as well as the dissipation rate of the
energy (ε). The standard k–ε model may not be able to account for the high
degree of swirl found in hydrocyclones. As a result of this, more advanced
representations of turbulence phenomena have been developed over the
last couple of decades.
The three main approaches to the modelling of turbulence are RANS
model, LES model, and DNS model. RANS is the oldest approach to model
turbulence in cyclones, and in unsteady RANS, an ensemble averaged rep-
resentation of the governing equations, including transient terms, are solved
(Utikar et al., 2010). Turbulence closure can be accomplished by use of the
so-called Boussinesq hypothesis. This involves the use of an algebraic equa-
tion for the Reynolds stresses or by solving the transport equations for the
Reynolds stresses. Early examples of the use of RANS models are those
of Hsieh (1988) and Hsieh and Rajamani (1988) that solved the Reynolds
averaged Navier–Stokes (RANS) equations to simulate single phase flow
through a hydrocyclone. These results could be compared with the velocity
measurements obtained from a 75 mm diameter hydrocyclone.
In the LES approach, larger carrying eddies are simulated and smaller
eddies are filtered and modelled by means of a sub-grid scale model. DNS
models are very expensive, since fully resolved Navier–Stokes equations
are solved and all relevant scales of turbulent motion are simulated. As a
consequence, these models are not yet practical for solving large-scale indus-
trial problems. Although LES models are more costly than unsteady RANS
models, they can be used for modelling the performance of full-scale equip-
ment. For example, Delgadillo and Rajamani (2005) have compared the
renormalized group k–ε model, the Reynolds stress model and LES model
to predict the dimensions of the air core and velocity profiles in a cyclone.
The LES model was better able to predict the experimental data.
Several other examples of the use of RANS and LES models to study the
impact of various geometrical and operational variables on the performance
of hydrocyclones can be found in the literature (Cullivan et al., 2004; Now-
akowski et al., 2004; Schuetz et al., 2004; Mangadoddy et al., 2005).

4.2.2 Model Validation

Earlier studies were conducted with pitot tubes (Yoshioka and Hotta, 1955)
and kinematic photography (Knowles et al., 1973). Later on, Das and Miller
(1996) have used X-ray tomography to analyse the swirl flow characteristics
14 Chris Aldrich

in an air-sparged hydrocyclone flotation system. Apart from gaining crucial

insights into flow phenomena in hydrocyclones, experimental studies are
increasingly used in conjunction with numerical modelling for this purpose.
The data required to validate numerical models can be obtained through
use of among other, particle image velocimetry (PIV) and laser Doppler
velocimetry (LDV) (Lim et al., 2010; Marins et al., 2010). With PIV photo-
graphic frames containing a number of particles are cross-correlated. In phase
Doppler anemometry, the velocity distribution and size of particles are
acquired by detecting laser light scattered by the particle at two angles
(e.g., Monredon et al., 1992; Fisher and Flack, 2002; Marins et al., 2010).
Another efficient method of measuring flow processes in hydrocyclones is
by use of radioisotope tracers. Stegowski and Leclerc (2002) have used 64Cu
with a half-life of 12.7 h to investigate the classification of copper ore slurries.
Methods such as PIV and LDV track the average velocity distributions of
particles, as opposed to Lagrangian tracking, where individual particles of a
dispersed phase can be tracked in space and time. By examining the particle
trajectory, it can be seen how a particle interacts with the flow around it under
the specific conditions and any local instabilities can be observed in detail.
Similarly, Bamrungsri et al. (2008) have used a high-speed camera to
record the motion of a dyed oil drop in water in a hydrocyclone. Wang
et al. (2008) have also used a high-speed camera to record the motion of a
vegetable seed with density of 1140 kg/m3 in a water hydrocyclone. Two-
dimensional trajectories could be plotted by recording the positions in consec-
utive frames. The two-dimensional particle paths obtained by Wang et al.
(2008) show that local or instantaneous instabilities in the flow field can
have a major effect on the particle trajectories, and hence on the separation
performance of the hydrocyclone. These insights cannot be gained from
studies of average velocity distributions. Although an additional camera can
be set up in an orthogonal direction to obtain three-dimensional trajectories
of particles, the spatial resolution can become an issue.
Lagrangian tracking also includes the use of positron emission particle
tracking (PEPT), which is a more recent development in process engineering.
PEPT is closely related to the medical method referred to as positron emission
tomography (PET). The difference is that while PET is based on detection of
spatially distributed radioactive sources (positron emitters) and image recon-
struction, and therefore provides the time-dependent concentrations of radio-
actively labelled substances, PEPT locates a point-like positron emitter by
cross-triangulation. The spatiotemporal resolution of PEPT is considerably
higher than that for PET. For example, Chang et al. (2011) have studied
Hydrocyclones 15

the flow of a particle through a hydrocyclone by means of positron emission

particle tracking (PEPT) using 18F as radioactive tracer and could track the
particle in the cyclone with an accuracy of 0.2 mm/ms.

5. SCALE-UP AND DESIGN

5.1 Scale-up
Preliminary selection of hydrocyclones is usually done from manufac-
turers’ charts. These charts show regions of operation for each cyclone as a
function of the inlet pressure and volumetric flow rate. For more in-depth
analysis, the operation of hydrocyclones can be captured by three-
dimensionless groups, namely the Stokes, Euler, and Reynolds numbers, as
indicated in Eqns (8)–(10). In these equations, the velocity v is the mean axial
velocity in the body of the hydrocyclone, i.e., the ratio of the volumetric
flow to the cross-sectional area of the cylindrical section of the cyclone body.

2
rs  r vd50
Stk50 ¼ (8)
18mDc
DP
Eu ¼ (9)
rv 2 =2
rvDC
Re ¼ (10)
m
Svarovsky (1984) has shown that the product of the Stokes and Euler
numbers is approximately constant for geometrically similar hydrocyclones,
and as such these equations can serve as a basis for the scale-up and design of
hydrocyclones in process circuits.
Broadly speaking, there are essentially three major families of geometri-
cally similar hydrocyclones, viz. the so-called Rietema (1961) and Bradley
(1965) hydrocyclones. The geometric proportions of these families of cy-
clones are indicated in Table 1. The range of sizes of hydrocyclones in in-
dustry is comparatively limited, but each cyclone can be operated at
different opening sizes for the inlet, underflow, and overflow by means of
interchangeable parts (Castilho and Medronho, 2000).
Dc, Di, and Do are the diameters of the cyclone, the feed inlet, and the
overflow, respectively, L and l are the total length of the cyclone and the
length of the vortex finder, respectively, and q is the angle of the hydrocy-
clone cone.
16 Chris Aldrich

5.2 Hydrocyclone Networks

Satisfactory separation can be achieved in many conventional solid–solid
separation systems by means of a single pass through a hydrocyclone. How-
ever, under extreme conditions, such as dealing with low density differen-
tials, fine particles, friable particulates, etc., multiple stages of cyclones
may be required. The use of multiple stages of hydrocyclones is also neces-
sary where multiple products are required, such as when the dispersed phase
is separated into size or density fractions, or when cyclones are used for
dewatering in co- or countercurrent washing operations (Williams et al.,
1994). Likewise, multiple stages of hydrocyclones are used to improve the
efficiency of separation in closed circuit grinding, for example (Heiskanen,
1993). The same equations comprising dimensionless numbers that are
used for scale-up can also be used for basic analysis with regard to the design
of hydrocyclone networks or clusters.

6. MONITORING AND CONTROL OF HYDROCYCLONES

6.1 Control
To date, the control of hydrocyclones has not enjoyed the same atten-
tion as the control of other pieces of processing equipment, in part owing to
their simple design and robust operation. Nonetheless, this seems to be
changing with the realization that significant gains in plant efficiencies can
be realized by better operation of hydrocyclones in process circuits. For
example, an important part of control of hydrocyclones is stabilization of
the slurry density in the overflow or cut point for feeds with varying solids
concentrations and particle size distributions (Neesse et al., 2004b; Hodouin,
2011). Regulation of the volume split has been restricted to single large cy-
clones fitted with adjustable nozzles. This is not feasible with batteries of
smaller cyclones.

6.2 Soft Sensors

Soft sensors are inferential models using other variables correlated to the var-
iable of interest to estimate the value of the latter. In hydrocyclones,
nonlinear autoregressive models with exogenous variables (ARX) (Casali
et al., 1998), error projection (Sbarbaro et al., 2008), recursive least squares
(Du et al., 1997), and artificial neural networks and support vector machines
(Sun et al., 2008) have been used to estimate the particle size in the overflow
of hydrocyclones in closed grinding circuits.
Hydrocyclones 17

6.3 Online Monitoring of Hydrocyclones

Over the last 20 years, several approaches have been proposed for the online
monitoring of hydrocyclone behaviour or characteristics, as summarized in
Table 3. Although some of these approaches have been commercialized, on-
line monitoring of hydrocyclones has not yet been incorporated into process
control loops to a large extent and online monitoring technology can there-
fore not be considered as a mature technology yet. Table 3 gives a summary of
online monitoring methods proposed over the last two decades.

6.4 Example of Underflow Monitoring

An example of underflow monitoring of hydrocyclones as discussed by Janse
van Vuuren et al. (2011) can be considered briefly here. In this approach, a
video camera is used to monitor the spray profile of a hydrocyclone

Table 3 Online monitoring of hydrocyclones

Method Measured element References

Mechanical plate Underflow spray angle Hulbert (1993)

Load cell gravimetry Mass of solids in Neesse et al. (2004c)
hydrocyclone
X-ray tomography Internal distribution of Galvin and Smitham (1994)
solids
Electrical impedance Internal distribution of Williams et al. (1997),
tomography solids; air core size and Gutiérrez et al. (2000)
shape
Electrical resistance Internal concentration Williams et al. (1992),
tomography of solids; location, size Williams et al. (1995)
and shape of air core Williams et al. (1999),
West et al. (2000)
Acoustic monitoring Feed pressure, feed Hou et al. (1998, 2002),
solids concentration; Neesse et al. (2004c)
underflow discharge
oscillation
Ultrasound Underflow discharge Olson and Waterman
oscillation (2006)
Laser optical Underflow discharge Neesse et al. (2004c)
profile
Videography Underflow discharge Petersen et al. (1996), Van
spray angle Deventer et al. (2003),
Janse van Vuuren et al.
(2011)
Overflow Pressure Air core diameter Krishna et al. (2010)
18 Chris Aldrich

Figure 5 Preprocessing of an image of the underflow of a hydrocyclone. From Janse

van Vuuren et al. (2011); with permission from Elsevier.

dynamically. Images of the underflow are first enhanced to determine the

outline of the flow as best as possible, as shown in Figure 5. This entails
removal of the noise in the image on the left in the figure, to finally yield
the image on the right.
The width of this spray profile is subsequently measured at a fixed point
below the spigot, as a proxy for the angle of the spray profile as indicated in
Figure 6). Measurements at different points on the profile can also be ob-
tained for more robust analysis (Petersen et al., 1996).
This information is analysed dynamically from a sequence of images and
the underflow widths can then be mapped to a process chart, as shown in
Figure 7. On this chart, normal operating conditions are indicated by a re-
gion delineated by a 99% confidence limit (the confidence level is adjust-
able). Image data are continuously mapped to this chart and the operator

Figure 6 Measurement of the spray profile width from a denoised image of the under-
flow of a hydrocyclone. From Janse van Vuuren et al. (2011); with permission from Elsevier.
Hydrocyclones 19

Figure 7 Mapping of image data to a process chart form predictive monitoring of the
behaviour of the hydrocyclones.

can then detect problems as before arise (typically overloading of the cyclone
leading to roping), since the transition from the normal operating region to
abnormal behaviour may be preceded by instability.

7. FUTURE DEVELOPMENTS
The application of hydrocyclones as separation devices is exceedingly
diverse, and one of the recent trends is the development of hydrocyclones
for the classification of fine particles. However, classification inefficiencies,
such as their large bypass, represent a major hurdle that will have to be
surmounted before they could be competitive with other devices in this re-
gard. Even so, the potential of very small diameter hydrocyclones in the
20 Chris Aldrich

classification of submicron size particles would be large, if these problems can

be solved (Cilliers, 2000).
As indicated in this chapter, the online monitoring of hydrocyclones has
attracted considerable attention in recent years. These systems could for the
basis for advanced control of hydrocyclone operations, which could translate
into significant benefits to plant operations associated with a wide range of
products.
Finally, numerical modelling of hydrocyclones continues to advance
strongly on the back of concomitant improvements in computer technol-
ogy, software, and increasing insight into flow processes in hydrocyclones.
This will lead to improvement in hydrocyclone design and will extend their
range and flexibility even further.

REFERENCES
Asomah, A.K., Napier-Munn, T.J., 1997. An empirical model of hydrocyclones, incorpo-
rating angle of cyclone inclination. Miner. Eng. 10 (3), 339–347.
Bamrungsri, P., Puprasert, C., Guigui, C., Marteil, P., Bréant, P., Hébrard, G., 2008. Devel-
opment of a simple experimental method for the determination of the liquid field veloc-
ity in conical and cylindrical hydrocyclones. Chem. Eng. Res. Des. 8 (11), 1263–1270.
Bloor, M.I.G., Ingham, D.B., Laverack, S.D., 1980. An analysis of boundary layer effects in a
hydrocyclone. In: Proceedings of the International Conference on Hydrocyclones,
Cambridge, Paper 5, 49–62, BHRA Fluid Engineering, Cranfield, 1980.
Bradley, D., Pulling, D.J., 1959. Flow patterns in the hydraulic cyclone and their interpreta-
tion in terms of performance. Trans. Inst. Chem. Eng. 37, 34–45.
Bradley, D., 1965. The Hydrocyclone. Pergamon Press Limited, Oxford.
Brennan, M., 2006. CFD simulations of hydrocyclones with an air core comparison between
large eddy simulations and a second moment closure. Chem. Eng. Res. Des. 84 (A6),
495–505.
Bretnai, E., 1891. US Patent No. 453, 105.
Casali, A., Gonzalez, G.D., Torres, F., Vallebuona, G., Castelli, L., Gimnez, P., 1998. Particle
size distribution soft-sensor for a grinding circuit. Powder Technol. 99, 15–20.
Castilho, L.R., Medronho, R.A., 2000. A simple procedure for design and performance pre-
diction of Bradley and Rietema hydrocyclones. Miner. Eng. 13 (2), 183–191.
Chang, Y.-F., Ilea, C.G., Aasen, Ø.L., Hoffmann, A.C., 2011. Particle flow in a hydrocy-
clone investigated by positron emission particle tracking. Chem. Eng. Sci. 66 (18),
4203–4211.
Chen, W., Zydek, N., Parma, F., 2000. Evaluation of hydrocyclone models for practical
applications. Chem. Eng. J. 80, 295–303.
Cilliers, J.J., Hinde, A.L., 1991. An improved hydrocyclone model for backfill preparation.
Miner. Eng. 4 (7–11), 683–693.
Cilliers, J.J., 2000. Hydrocyclones for particle separation. In: II Particle Size Separation.
Academic Press.
Coelho, M.A.Z., Medronho, R.A., 2001. A model for performance prediction of
hydrocyclones. Chem. Eng. J. 84, 7–14.
Cullivan, J.C., Williams, R.A., Dyakowski, T., Cross, C.R., 2004. New understanding of a
hydrocyclone flow field and separation mechanism from computational fluid dynamics.
Miner. Eng. 17 (5), 651–660.
Hydrocyclones 21

Das, A., Miller, J.D., 1996. Swirl flow characteristics and froth phase features in air-sparged
hydrocyclone flotation as revealed by X-ray CT analysis. J. Miner. Process. 47 (3–4),
251–274.
Delgadillo, J.A., Rajamani, R.K., 2005. A comparative study of three turbulence closure
models for the hydrocyclone problem. Int. J. Miner. Process. 77 (4), 217–230.
Delgadillo, J.A., Rajamani, R.K., 2007. Exploration of hydrocyclone designs using compu-
tational fluid dynamics. Int. J. Miner. Process. 84 (1–4), 252–261.
Driessen, M.G., 1951. Theory of Flow in a Cyclone. Revue de L’Industrie Minerale,
Numero Special, Saint-Etienne, 449–461.
Du, Y.-G., del Villar, R., Thibault, J., 1997. Neural net-based softsensor for dynamic particle
size estimation in grinding circuits. Int. J. Miner. Process. 52 (2), 121–135.
Dueck, J., Matvienko, O.V., Neesse, T., 2000. In: Modelling of Hydrodynamics and Sepa-
ration in a Hydrocyclone. Theoretical Foundations of Chemical Engineering, vol. 34.
Kluwer Academic/Plenum Publishers, pp. 428–438.
Dwari, R.K., Biswas, M.N., Meikap, B.C., 2004. Performance characteristics of sand FCC
and fly ash in a novel hydrocyclone. Chem. Eng. Sci. 59, 671–684.
Dyakowski, T., Nowakowski, A.F., Kraipech, W., Williams, R.A., 1999. A three dimen-
sional simulation of hydrocyclone behaviour. In: Second International Conference on
CFD in the Minerals and Process Industries Melbourne, Australia.
Endres, E., Dueck, J., Neesse, T., 2012. Hydrocyclone classification of particles in the micron
range. Miner. Eng. 31, 42–45.
Flament, F., Thibault, J., Hodouin, D., 1993. Neural Networks Based Control of Mineral
Grinding Plants. Mineral Engineering 6 (3), 235–249.
Fisher, M.J., Flack, R.D., 2002. Velocity distributions in a hydrocyclone separator. Exp.
Fluids 32 (3), 302–312.
Galvin, K.P., Smitham, J.B., 1994. Use of X-rays to determine the distribution of particles in
an operating cyclone. Miner. Eng. 7, 1269–1280.
Gutiérrez, J.A., Dyakowski, T., Beck, M.S., Williams, R.A., 2000. Using electrical imped-
ance tomography for controlling hydrocyclone underflow discharge. Powder Technol.
108, 180–184.
Heiskanen, K., 1993. Particle Classification. Chapman and Hall, London, UK.
Hodouin, D., 2011. Methods for automatic control, observation, and optimization in mineral
processing plants. J. Process Control 21, 211–225.
Hou, R., Hunt, A., Williams, R.A., 1998. Acoustic monitoring of hydrocyclone
performance. Miner. Eng. 11 (11), 1047–1059.
Hou, R., Hunt, A., Williams, R.A., 2002. Acoustic monitoring of hydrocyclones. Powder
Technol. 124, 176–187.
Hsieh, K.T., 1988. A Phenomenological Model of the Hydrocyclone (Ph.D. thesis). Univer-
sity of Utah, UT, USA.
Hsieh, K.T., Rajamani, R.K., 1988. Phenomenological model of the hydrocyclone: model
development and verification for single-phase flow. Int. J. Miner. Process. 22 (1–4),
223–237.
Hulbert, D.G., 1993. Measurement methods and apparatus for hydrocyclones. In: Mintek
(Ed.), EP0522215A3.
Janse van Vuuren, M.J., Aldrich, C., Auret, L., 2011. Videographic monitoring of
hydrocyclones. Miner. Eng. 24, 1532–1544.
Karr, C.L., Stanley, D.A., McWhorter, B., 2000. Optimization of hydrocyclone operation
using a geno-fuzzy algorithm. Comput. Method. Appl. Mech. Eng. 186, 517–530.
Knowles, S.R., Woods, D.R., Feuerstein, I.A., 1973. The velocity distribution within a
hydrocyclone operating without an air-core. Can. J. Chem. Eng. 51, 263–271.
Kraipech, W., Chen, W., Parma, F.J., Dyakowski, T., 2002. Modelling the fish-hook effect
of the flow within hydrocyclones. Int. J. Miner. Process. 66 (1–4), 49–65.
22 Chris Aldrich

Kraipech, W., Chen, W., Dyakowski, T., Nowakowski, A., 2006. The performance of
the empirical models on industrial hydrocyclone design. Int. J. Miner. Process. 80,
100–115.
Krishna, V., Sripriya, R., Kumar, V., Chakraborty, S., Meikap, B.C., 2010. Identification
and prediction of air core diameter in a hydrocyclone by a novel online sensor based
on digital signal processing technique. Chem. Eng. Process. Process Intensif. 49,
165–176.
Lapple, C.E., 1951. Processes use many collector types. Chem. Eng. 58, 144–151.
Leith, D., Licht, W., 1972. Collection efficiency of cyclone type particle collector: a new
theoretical approach. AIChE Symp. Series (Air-1971) 68, 196–206.
Lim, E.W., Chen, Y.-R., Wang, C.-H., Wu, R.-M., 2010. Experimental and computational
studies of multiphase hydrodynamics in a hydrocyclone separator system. Chem. Eng.
Sci. 65, 6415–6424.
Lim, K.S., Kwon, S.B., Lee, K.W., 2003. Characteristics of the collection efficiency for a
double inlet cyclone with clean air. Aerosol Sci. 34, 1085–1095.
Lynch, A.J., Rao, T.C., 1975. Modelling and scale-up of hydrocyclone classifiers. In: Pro-
ceedings of the 11th International Mineral Processing Congress, Cagliari, pp. 245–269.
Madge, D.N., Romero, J., Strand, W.L., 2004. Hydrocarbon cyclones in hydrophilic oil sand
environments. Miner. Eng. 17 (5), 625–636.
Mangadoddy, N., Sripriya, R., Banerjee, P.K., 2005. CFD modelling of
hydrocyclonedprediction of cut size. Int. J. Miner. Process. 75, 53–68.
Marins, L.P.M., Duarte, D.G., Loureiro, J.B.R., Moraes, C.A.C., Silva Freire, A.P., 2010.
LDA and PIV characterization of the flow in a hydrocyclone without an air-core.
J. Pet. Sci. Eng. 70, 168–176.
Monredon, T.C., Hsieh, K.T., Rajamani, R.K., 1992. Fluid flow model of the hydrocy-
clone: an investigation of device dimensions. Int. J. Miner. Process. 35 (1–2), 65–83.
Murray, H.H., 1980. Major kaolin processing developments. Int. J. Miner. Process. 7 (3),
263–274.
Nageswararao, K., Wiseman, D.M., Napier-Munn, T.J., 2004. Two empirical hydrocyclone
models revisited. Miner. Eng. 17, 671–687.
Neesse, T., Dueck, J., Minkov, L., 2004a. Separation of finest particles in hydrocyclones.
Miner. Eng. 17, 689–696.
Neesse, T., Golyk, V., Kaniut, P., Reinsch, V., 2004b. Hydrocyclone control in grinding
circuits. Miner. Eng. 17, 1237–1240.
Neesse, T., Schneider, M., Golyk, V., Tiefel, H., 2004c. Measuring the operating state of the
hydrocyclone. Miner. Eng. 17, 697–703.
Nenu, R.K.T., Yoshida, H., 2009. Comparison of separation performance between single
and two inlets hydrocyclones. Adv. Powder Technol. 20, 195–202.
Norimoto, U., Fukui, K., Yoshida, H., 2004. Particle separation performance of hydrocy-
clone with forced vortex type. J. Soc. Powder Technol. Jpn. 43 (9), 666–675.
Nowakowski, A.F., Cullivan, J.C., Williams, R.A., Dyakowski, T., 2004. Application of
CFD to modelling of the flow in hydrocyclones. Is this a realizable option or still a
research challenge? Miner. Eng. 17 (5), 661–669.
Olson, T.J., Waterman, R.J., 2006. Hydrocyclone roping Detector and method. In: TUCSON
(Ed.), US6983850B2. Krebs Engineers Corporation, United States.
Pericleous, K.A., Rhodes, N., 1986. The hydrocyclone classifier – a numerical approach. Int.
J. Miner. Process. 17, 23–43.
Petersen, K.R.P., Aldrich, C., Van Deventer, J.S.J., McInnes, C., Stange, W.W., 1996.
Hydrocyclone underflow monitoring using image processing methods. Miner. Eng. 9
(3), 301–316.
Plitt, L.R., 1976. A mathematical model of the hydrocyclone classifier. Can. lnst. Min. Met-
all. Bull. 60 (776), 114–122.
Hydrocyclones 23

Puprasert, C., Hebrard, G., Lopez, L., Aurelle, Y., 2004. Potential of using hydrocyclone and
hydrocyclone equipped with Grit pot as a pre-treatment in run-off water treatment.
Chem. Eng. Process. Process Intensif. 43 (1), 67–83.
Rietema, K., 1961. Performance and design of hydrocyclones – parts I–IV. Chem. Eng. Sci.
15, 298–325.
Sbarbaro, D., Ascencio, P., Espinoza, P., Mujica, F., Cortes, G., 2008. Adaptive soft-sensors
for on-line particle size estimation in wet grinding circuits. Control Eng. Pract. 16 (2),
171–178.
Schuetz, S., Mayer, G., Bierdel, M., Piesche, M., 2004. Investigations on the flow and sep-
aration behaviour of hydrocyclones using computational fluid dynamics. Int. J. Miner.
Process. 73 (2–4), 229–237.
Shepherd, C.B., Lapple, C.E., 1939. Flow pattern and pressure drop in cyclone dust
collectors. Ind. Eng. Chem. 31, 972–984.
Silva, A.C., Silva, E.M.S., Matos, J.D.V. Comparison between Plitt’s model and its revisions
for hydrocyclones simulation. Accessed from: http://www.arber.com.tr/imps2012.org/
proceedingsebook/Abstract/absfilAbstractSubmissionFullContent369.pdf.
Slechta, J., Firth, B.A., 1984. Classification of fine coal with a hydrocyclone. Int. J. Miner.
Process. 12 (4), 213–237.
Stange, W.W., 1993. Using artificial neural networks for the control of grinding circuits.
Miner. Eng. 6 (5), 479–489.
Stegowski, Z., Leclerc, J.-P., 2002. Determination of the solid separation and residence time
distributions in an industrial hydrocyclone using radioisotope tracer experiments.
Int. J. Miner. Process. 66, 67–77.
Sun, Z., Wang, H., Zhang, Z., 2008. Soft sensing of overflow particle size distributions in
hydrocyclones using a combined method. Tsinghua Sci. Technol. 13 (1), 47–53.
Svarovsky, L., 1984. Hydrocyclones. Holt Rinehart and Winston.
Utikar, R., Darmawan, N., Tade, M., Li, Q., Evans, G., Glenny, M., Pareek, V., 2010.
Hydrodynamic simulation of cyclone separators. In: Oh, H.W. (Ed.), Computational
Fluid Dynamics. InTech. Available from. http://www.intechopen.com/books/
computational-fluiddynamics/hydrodynamic-simulation-of-cyclone-separators. ISBN:
978-953-7619-59-6.
Van der Walt, T.J., Van Deventer, J.S.J., Barnard, E., 1993. Neural nets for the simulation of
mineral processing operations: part II. Applications. Miner. Eng. 6 (11), 1135–1153.
Van Deventer, J.S.J., Feng, D., Petersen, K.R.P., Aldrich, C., 2003. Modelling of hydrocy-
clone performance based on spray profile analysis. Int. J. Miner. Process. 70, 183–203.
Wang, Z.-B., Chu, L.-Y., Chen, W.-M., Wang, S.-G., 2008. Experimental investigation of
the motion trajectory of solid particles inside the hydrocyclone by a Lagrange method.
Chem. Eng. J. 138 (1–3), 1–9.
West, R.M., Jia, X., Williams, R.A., 2000. Parametric modeling in industrial tomography.
Chem. Eng. J. 77 (1–2), 31–36.
White, D.A., 1991. Efficiency curve model for hydrocyclones based on crowding theory.
Trans. Inst. Min. Metall. Sect. C: Miner. Process. Extr. Metall. 100, 135–138.
Williams, R.A., Dickin, F.J., Dyakowski, T., Ilyas, O.M., Abdullah, Z., Beck, M.S., 1992.
Looking into mineral process plant? Miner. Eng. 5, 867–881.
Williams, R.A., Ilyas, O.M., Dyakowski, T., Dickin, F.J., Gutierrez, J.A., Wang, M.,
Beck, M.S., Shah, C., Rushton, A., 1995. Air core imaging in cyclonic separators:
implications for separator design and modelling. Chem. Eng. J. Biochem. Eng. J. 56,
135–141.
Williams, R.A., Jia, X., West, R.M., Wang, M., Cullivan, J.C., Bond, J., Faulks, I.,
Dyakowski, T., Wang, S.J., Climpson, N., Kostuch, J.A., Payton, D., 1999. Industrial
monitoring of hydrocyclone operation using electrical resistance tomography. Miner.
Eng. 12, 1245–1252.
24 Chris Aldrich

Williams, R.A., Albarran de Garcia Colon, I.L., Lee, M.S., Roldan Villasana, E.J., 1994.
Design targeting of hydrocyclone networks. Miner. Eng. 5, 253–256.
Williams, R.A., Dickin, F.J., Gutiérrez, J.A., Dyakowski, T., Beck, M.S., 1997. Using elec-
trical impedance tomography for controlling hydrocyclone underflow discharge.
Control Eng. Pract. 5, 253–256.
Woodfield, D., Bickert, G., 2004. Separation of flocs in hydrocyclones - significance of floc
breakage and floc hydrodynamics. Int. J. Miner. Process. 73 (2–4), 239–249.
Yalamanchili, M.R., Miller, J.D., 1995. Removal of insoluble slimes from potash ore by
air-sparged hydrocyclone flotation. Miner. Eng. 8 (1–2), 169–177.
Yalcin, T., Kaukolin, E., Byers, A., 2003. Axial inlet cyclone for mineral processing
applications. Miner. Eng. 16 (12), 1375–1381.
Yang, I.H., Shin, C.B., Kim, T.-H., Kim, S., 2004. A three-dimensional simulation of a
hydrocyclone for the sludge separation in water purifying plants and comparison with
experimental data. Miner. Eng. 17 (5), 637–641.
Yoshida, H., Norimoto, U., Fukui, K., 2006. Effect of blade rotation on particle classification
performance of hydrocyclone. Powder Technol. 164, 103–110.
Yoshioka, N., Hotta, Y., 1955. Liquid cyclones as a hydraulic classifier. Chem. Eng. Jpn. 19,
633–641.
Zhao, B., Shen, H., Kang, Y., 2004. Development of a symmetrical spiral inlet to improve
cyclone separator performance. Powder Technol. 145, 47–50.