Advanced Powder Technology 24 (2013) 473–481

Contents lists available at SciVerse ScienceDirect

Advanced Powder Technology

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

Original Research Paper

CFD–DEM investigation of particle separations using a sinusoidal jigging

profile
Stephen Viduka a,b, Yuqing Feng a,⇑, Karen Hapgood b, Phil Schwarz a
a
CSIRO Mathematics, Informatics, and Statistics, Box 312, Clayton South, 3169 Victoria, Australia
b
Monash Advanced Particle Engineering Laboratory, Department of Chemical Engineering, Monash University, PO Box 36, Clayton, 3800 Victoria, Australia

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

Article history: This paper presents a numerical investigation of solid separation in jigging device. Jigging is a gravity sep-
Received 15 June 2012 aration method commonly used by the minerals industry to separate coal, iron ore, diamonds and other
Received in revised form 29 October 2012 minerals on the basis of particle size and/or density. Separation is recognised as being heavily dependent
Accepted 8 November 2012
on fluid motion in the jig. This study explores the effects of the inlet time dependent velocity profile in
Available online 11 February 2013
relation to a wide criterion on jigging performance. Modelling of the liquid–solid system is performed
through a combination of computational fluid dynamics (CFD) to simulate liquid flow and discrete ele-
Keywords:
ment method (DEM) to resolve particle motion. The initial packing conditions consist of a binary-density
Discrete element method
Computational fluid dynamics
particle system of 1130 particles each 1 cm in diameter. A range of jigging profiles have been imple-
Liquid–solid flow mented in mineral processing. In this study the sinusoidal pulsation profile is selected adopting varia-
Jigging tions in both amplitude and frequency. The performance of profile variants are compared in terms of
Stratification solid flow patterns, separation kinetics, energy, and mean particle position. These quantitative compar-
Gravity concentration isons demonstrate significant differences in the segregation rate, energy, and solid phenomena, helping
find an alternative optimum operating setting for the system. In addition, boundaries of operation are
found in terms of frequency and amplitude limits and the concentration mechanics are investigated in
these regions.
Ó 2013 The Society of Powder Technology Japan. Published by Elsevier B.V. and The Society of Powder
Technology Japan. All rights reserved.

1. Introduction
Jigging is a gravity separation process adopted by the minerals
industry to separate ore minerals on the basis of particle size and/
or density [1]. Jigging units apply pulsed liquid flow dilating the
particle bed where particle stratification ensues due to influences
of hydrodynamic and gravity forces.
Much of the published research performed in jigging has been
experimental [2–11]. Further, commercial jigs date back as far as
the Neil Jig (1914) [12]. It is reasonable to assume many jigs histor-
ically were designed principally with the aid of experiments and
also analytical expressions in the absence of computational capa-
bilities. The past empirical studies develop understanding of how
the feed material macroscopically responds to various operating
conditions, they do not elucidate on the intricate transient behav-
iour of the fluid and particles, along with separation kinetics which
are important to develop a full understanding of the process.
Modern investigations of jigging phenomena using numerical
simulation techniques has shown to be a fast growing area. Solnor-
dal et al. [13], applied a single phase computational fluid dynamic
⇑ Corresponding author. Tel.: +61 3 9545 8982; fax: +61 3 9545 8669.
E-mail address: Yuqing.Feng@csiro.au (Y. Feng).
(CFD) technique but this was limited as it treated the slurry as a
single phase. Various studies applied discrete element method
(DEM) to simulate the motion of individual particles discretely
coupled with simplified fluid models giving some insights into mi-
cro-mechanical processes at the particulate level [3–5,14–17].
These modelling techniques assume a uniform fluid field and do
not account for the effect of non-uniform fluid velocity on the par-
ticle drag forces. The Euler–Lagrange (DEM–CFD) model, first pro-
posed by Tsuji et al. [18], remains the most attractive technique
because of its superior computational convenience as compared
to Direct Numerical Simulation-DEM, or Lattice Boltzmann-DEM
models, and the capability to capture the particle physics as com-
pared to DEM-simplified fluid models. The model has been increas-
ingly used to study a wide range of particle fluid systems [19]. The
liquid phase flow is solved using the Navier–Stokes and continuity
equations, while the motion of individual particles is obtained by
solving Newton’s second law of motion, with the liquid-particle
coupling treated using Newton’s third law of motion. This ap-
proach can generate detailed information about the trajectories
of particles and the transient forces between two particles and be-
tween particles and fluid. Other notable models used to investigate
jigging include Potential energy [20–22], Potential energy-Monte
Carlo [21], Artificial Neural Network (ANN) [23,24], Statistical
[25], and Unsteady-fluidisation [26].

0921-8831/$ - see front matter Ó 2013 The Society of Powder Technology Japan. Published by Elsevier B.V. and The Society of Powder Technology Japan. All rights reserved.
http://dx.doi.org/10.1016/j.apt.2012.11.012
474 S. Viduka et al. / Advanced Powder Technology 24 (2013) 473–481
Only a few jigging studies have adopted the DEM–CFD approach
[27–30]. Both the studies by Asakura et al. [27] and Xia and Peng
[28] are two-way coupled and consider drag on each particle indi-
vidually, but do not consider porosity. Xia and Peng [28,30] used a
2D column model and implemented forces including virtual mass
force, Magnus force [31], and Saffman force [32,33]. One study ana-
lysed the importance of different forces acting on a particle in jig-
ging and was performed for multi-sized and binary-sized particles
in a sinusoidal pulsion. Additionally, the authors studied the hin-
dered settling velocity as a function of particle densities and sizes,
and the effect of sinusoidal pulsation, amplitude and frequency on
the particle separation and fluid flows [28]. A separate study used
the same model and highlighted that the fluid is highly dynamic
and influenced by the presence of particles confirming the simpli-
fied idealised flow behaviour as assumed in DEM-simplified fluid
models does not exist [30]. Asakura et al. [27] went a step further
including the Basset force [34] and a 3D column model which stud-
ied the trajectory and response time of a single particle in a jig.
Dong et al. [29] applied a one-way coupled 3D model to a close-
to realistic geometry Inline Pressure jig. The study considered that
fluid flow is the dominant factor in the jig, and implemented a saw-
tooth-forward leaning jigging profile investigating vibration fre-
quency and amplitude, and the size and density of ragging
particles on the flow separation. However, one-way coupling does
not account for the influence of the local particles on the fluid.
Previous studies using the DEM–CFD model have used a sinu-
soidal pulsation profile with the exception of Dong et al. [29],
who used a forward leaning saw tooth cycle. No numerical inves-
tigations (including all various modelling techniques) have studied
what effect the sinusoidal profile has on concentration mechanics
by using two-way coupling in conjunction with a porous drag force
model. Further, these studies have not investigated jigging aspects
such as separation time, energy, and profile optimisation. The aim
of this study is to elucidate how the profile induces segregation,
and how variations of frequency and amplitudes affect perfor-
mance based on a range of criteria.
2. Simulation method
2.1. Governing equations
The DEM–CFD model has been well documented in the litera-
ture. For brevity, only the outline of the model structure is de-
scribed below. The solid phase is treated as a discrete phase and
solved using DEM. The translational and rotational motions of a
particle at any time, t, in the bed are determined by Newton’s sec-
ond law of motion. These can be written as:
X k
dv i i to produce three dimensional (3D) results. The liquid flow was
mi ¼ f f ;i þ ðf c;ij þ f d;ij Þ þ f g;i
dt j¼1
and
k
dxi X i
Ii ¼ T ij
dt
j¼1
where mi, Ii, ki, vi and xi are, respectively, the mass, moment of iner-
tia, number of contacting particles, translational and rotational
velocities of particle i, and ff,i, fg,i are fluid drag force, gravitational
force respectively. fc,ij, fd,ij and Ti,j are the contact force, viscous con-
tact damping force and torque between particles i and j. These in-
ter-particle forces and torques are summed over the ki particles in
contact with particle i.
The particle–particle and particle–wall contact force is based on
the soft-sphere method. The particle fluid interaction force is cal-
culated using the Di Felice drag force correlation [35], and Model
B formulation is adopted [36]. The liquid phase is treated as a con-
tinuous phase moving through a porous medium created by the
particles, and is modelled similarly to conventional two fluid mod-
els in which porosity (or liquid volume fraction) modifies the stan-
dard single phase Navier–Stokes equations. The governing
equations are then the conservations of mass and momentum in
terms of the local mean variables over a computational cell, given
by:
@e
þ r  ðeuÞ ¼ 0 ð3Þ
@t
and
Pk c
@ðqf euÞ i¼1 f f ;i
þ r  ðqf euuÞ ¼ rP  þ rðesÞ þ qf eg ð4Þ
@t DV
where qf, u and P are, respectively, the fluid density, velocity and
pressure; s, e and DV are the fluid viscous stress tensor, porosity
and volume of a computational cell.
The particle flow is solved numerically using an in-house DEM
code [36] with an explicit time integration method and established
geometrical and flow boundary conditions. The continuous liquid
phase is readily solved using a commercial CFD software package
(ANSYS CFX 10.0). The coupling between DEM and CFD is achieved
as follows. At each time step DEM will give information of posi-
tions and velocities of individual particles for the evaluation of
porosity and volumetric fluid drag force in a computational cell.
CFD will then use this data to determine the fluid flow field, which
in turn is used to determine the fluid drag forces acting on individ-
ual particles. Incorporating the resulting forces into DEM will pro-
duce information about the motion of individual particles for the
next time step. The fluid drag force acting on an individual particle
will react on the fluid phase from the particles, so that Newton’s
third law of motion is satisfied.
2.2. Simulation conditions
The model consists of a rectangular domain filled with a binary-
density spherical particle system and liquid. The particles were di-
vided into 565 light particles and 565 heavy particles with respec-
tive densities of 2540 kg/m3 (glass) and 4630 kg/m3 (ceramic), the
liquid used was water 1000 kg/m3 in density. Detailed model set-
tings are shown in Table 1. The side walls were treated with no-slip
boundary conditions. The bottom was considered as a wall for the
particle phase, so they cannot fall through, but as an inlet for liquid.
The top exit was treated with a zero normal gradient opening con-
dition. Periodic boundary conditions were applied to the front and
rear surfaces of the flow domain effectively creating infinite thick-
ness and economically reducing the number of particles required
ð1Þ considered in two dimensions (2D) using only one cell in the thick-
ness direction and hence does not resolve detailed flow fields in
this direction, while DEM modelling of the particles was in 3D,
with a bed thickness equal to five particle diameters. As all pulsa-
tion profiles are studied with a two dimensional model, the results
ð2Þ
remain acceptable for comparison.
Uniform liquid flow was injected through the inlet and the
flowrate varied with time according to the pulsation profile simu-
lated. The inlet flow for the sinusoidal pulsation profile was estab-
lished using a sinusoidal function. The pulsation profiles are
compared by holding the shape of the profile constant and using
three variations of period (T) and volumetric water input/exhaust
(A). These are 1, 2, and 3 s periods (or 60, 30, and 20 cycles/min),
and, 1.5, 2.25 and 3 L water amplitudes. The amplitudes are repre-
sented in litres not distance as the water/air free surface is not re-
solved (i.e. the domain at anytime is completely filled with water),
otherwise the amplitudes would be equivalent to 0.2, 0.3 and
 S. Viduka et al. / Advanced Powder Technology 24 (2013) 473–481 475

Table 1
Jig model specifications.

Particle phase Liquid phase

3
Density (kg m ) Light 2540 Viscosity (kg m1 s1) 1  103
Heavy 4630
Young’s modulus (N m2) 1.0  108 Density (kg m3) 1000
Poisson ratio (N m2) 0.3 CFD cell Width (m) 0.025
Sliding friction coefficient (–) 0.3 Height (m) 0.025
Damping coefficient (–) 0.2 Bed geometry Width (m) 0.15
Particle diameter (m) Light 0.01 Height (m) 0.9
Heavy 0.01 Thickness (m) 0.05
Number of particles (–) Light 565 Bed distributor Uniform
Heavy 565
Time step (s) 1  105 Time step (s) 1  103

better mixed packing before the start of the jigging. After settling,
0.9 0.9
liquid is injected through the bottom following the appropriate
0.6 0.6 pulsation profile and jigging begins.
Water (L) x 10 -1

0.3 0.3 Velocity (m/s)

0 0
-0.3
-0.6
-0.9
0 0.5 1 1.5 2 2.5
Time (s)
Water input/exhaust
Fig. 1. An example of the sinusoidal pulsation profile applied at inlet boundary
condition. T = 2 s, A = 3 L.
0.4 cm. A further study is performed with broader amplitude and
frequency values to establish and investigate operational limits
of these parameters. The particles were processed using 60 s of jig-
ging concluding with 1 s of settling after the last jigging cycle. An
example of the sinusoidal profile is displayed graphically in
Fig. 1. Each profile includes a pulsion period with an upward liquid
motion (positive value in velocity) and a suction period with a
downward liquid motion (negative value in velocity). The simula-
tion begins with the random generation of particles without over-
laps, followed by a period of gravitational settling to form an initial
mixed fixed bed. During the settling process, the buoyancy force to
particles is switched off to help prevent segregation and achieve a
3. Results and discussion
-0.3
3.1. Solid flow patterns
-0.6
Solid flow patterns are plotted first to obtain a visual under-
-0.9
standing of the stratification process. Fig. 2 shows the particle posi-
3 3.5 4
tions for the sinusoidal profile variant of T = 2 s and A = 3 L over six
jigging cycles. This variant is moderate profile in terms of fre-
Velocity quency and volumetric water input and therefore elucidates on
more general profile phenomena which exist in all the profile vari-
ants. The maximum and minimum particle displacements can be
visualised. The light particles are coloured black and heavy parti-
cles grey. Three snapshots are taken at each cycle: before pulsion,
during the cycle when particles are at their peak heights, and at the
end of suction, respectively.
In the initial cycle the particle bed begins well mixed. As pulsion
proceeds the bed starts to loosen from the bottom and total bed
expansion follows via a loosening wave rapidly spreading upwards
through the bed. The bed is left almost uniformly expanded (see
Fig. 2). Segregation proceeds both in pulsion and suction and heavy
particles begin to populate the bottom of the column. As segrega-
tion progresses particles towards the top of the bed increasingly
begin to cluster and lift to greater heights. In the final rested state
at the end of the sixth cycle particles are shown to be completely
separated.

Fig. 2. Solid flow patterns shown by particle position for variant T = 2 s, A = 3 L, at the: 1st, 3rd, 4th, and 6th cycle. Heavy and light particle are coloured grey and black
respectively.
476 S. Viduka et al. / Advanced Powder Technology 24 (2013) 473–481

Bed height (cm) 80
70
60
50
40
30
20
10
0 cle bed. As the volumetric input increases the bed maximum par-
0 2 4 6 8 10
Time (s)
Heavy Light
Fig. 3. Mean particle position for variant T = 2 s, A = 3 L.
The mean particle position can be used to quantify particle
behaviours by illustrating the average height of both particle types
separately. Fig. 3 shows the mean particle position for the sinusoi-
dal profile variant of T = 2 s and A = 3 L over 12 s of jigging. Starting
from a well mixed state where each type of particle has a similar
mean position, the light particles travel upwards faster than the
heavier particles during pulsion and reach greater heights passing
heavy particles. During the suction period heavier particles
increasingly settle first owing to a higher downward velocity and
being closer to the bottom of the column. Following jigging, more
and more light particles aggregate on top of heavier particles. As
this occurs lighter particles receive less constraint from the heavier
particles and are able to travel higher in proceeding jigging cycles
until a dynamically stable state is reached. All profiles display
gradual segregation shown by the growing differences in mean
particle positions.
The maximum particle displacements of all variants in the first
rested state when segregation is achieved. The results show vast
differences in maximum heights the particles reach in pulsion
and also bed expansion. Profiles with volumetric water inputs of
A = 1.5 L, have insufficient velocity and therefore little drag to lift
the bed to great heights. Consequently, there is little opportunity
for bed expansion which facilitates particle rearrangement. These
variants comparatively require a lot of time to segregate the parti-
12 ticle height and expansion increases. This is advantageous to
segregation and these profiles segregate much faster. Decreasing
the cycle period is shown to increase expansion and maximum
particle height. It is found a high frequency profile must lift the
bed to a large height which therefore increases settling time allow-
ing time for rearrangement before being quickly sucked back
down. A moderate or low frequency profile does not have to lift
a bed to such great heights as the duration of the cycle is longer
allowing more time for rearrangement.
Therefore two important criteria must be met for segregation: a
profile must sufficiently lift the bed to a height where expansion
and loosening can proceed, but also allow enough time for particles
to rearrange.
3.2. Particle separation time
The coordination number represents the average sum of con-
tacts of a particle type, with either similar or different particle
types, e.g. the heavy particles average contact number with light
particles. The heavy-light coordination number is a good indication
of particle segregation and can be written as:
k
1X l

C hl ¼ xl ð5Þ
cycle can be visualised in Fig. 4, together with the particles in a kl l¼1

Fig. 4. Solid flow patterns during the first cycle when particles are at their peak heights for all variants of the sinsuoidal profile (top) and snapshots at the end of suction once
separation is complete (bottom). Heavy and light particle are coloured grey and black respectively: (a) T = 1 s, (b) T = 2 s, and (c) T = 3 s.
 S. Viduka et al. / Advanced Powder Technology 24 (2013) 473–481
where for each light particle l the sum of contacts with heavy par-
ticles xl gives the coordination number for each light particle, which
is averaged over the number of light particles kl. The coordination
number fluctuates as the bed expands under pulsion and compacts
under suction. Only values when the bed is at rest are considered
here which coincides with the state of the final product. The value
gradually reduces from when the bed is in a mixed state through to
complete segregation. The lower the value the higher the segrega-
tion i.e. the less light particles are in contact with heavy particles.
Of the nine profiles, eight completely segregated the particle
bed within 60 s of jigging time. Complete segregation is indicated
by coordination numbers beginning to plateau after approximately
a value of 0.5, slight differences under this value are insignificant
(see Fig. 5). Steady separation is critical for a reliable jigging pro-
cess. The separation process can be unsteady and is able to reverse
and mix in subsequent cycles, in this situation the coordination
number of a rested bed will fluctuate [37]. In this study the sinu-
soidal profile is found to provide steady segregation for the eight
variants which induce segregation (see Fig. 5).
A consistently low coordination number indicates stable sepa-
ration and a segregated bed. However, it does not characterise in
which way particles have settled. Desirable settling occurs when
particles are stratified vertically with one particle type directly
on top of another, this ensures separate delivery to launders. In
addition to segregation the settling also needs to be stable. If set-
tling is inconsistent it is undesirable and unreliable in jig process-
ing even if the bed is perfectly segregated. The solid flow patterns,
as in Fig. 2, help visualise the stratification behaviour. Similar to
the coordination number the solid flow pattern is tracked through
all the jigging cycles. If a tendency of undesirable settling is pres-
ent, the variant is deemed impractical for particle processing of
this system. All variants of the sinusoidal profile were found to be-
have steadily in both coordination number and settling
configuration.
The effects of volumetric water input and cycle period on the
separation time is shown in Fig. 6. Here a broader range of cycle
profile parameters are used to illustrate the complete particle sep-
aration time phenomenon in terms of amplitude and frequency
selection. A strong relationship between volumetric input, A, and
separation time is seen in all but one variant of frequency
(Fig. 6a). As the volumetric input increases the separation time re-
duces. The profile which is the exception has a profile setting of
T = 1 s and A = 3 L. Unlike other frequency variants, segregation is
faster at an intermediate amplitude. This is found to be caused
by particles being far from settled at the end of suction before pul-
sion begins in the following cycle (see variant A = 3 L on the bottom
of Fig. 4a where particles are shown not to be fully settled at the
end of suction). In this variant, a portion of the light particles are
still settling during pulsion in the subsequent cycle. These light
particles under pulsion have a downwards momentum and low
drag force upwards (due to high porosity) resulting in a decelera-
tion of the particle downwards. Alternatively, the heavy particles
settled in the previous cycle. These heavy particles which have
no momentum while in a packed bed state and higher drag force
(due to low porosity) move upwards with an opportunity to pene-
trate back through the bed slowing down segregation. It is ex-
pected that further increases in amplitude for variants T = 2 s,
and T = 3 s, would eventually result in the same situation.
The relationship between period, T, and separation time is
shown in Fig. 6b. Reducing the cycle period has a similar effect of
increasing amplitude. As the cycle period is reduced, the resulting
higher inlet velocity substantially increases particle expansion
reducing separation time. The reduction in particle separation time
plateaus and reaches a point where further reduction in cycle per-
iod slows separation. The separation slows for the same reason as
explained previously, for variant of T = 1 s and A = 3 L, separation is
477
faster until the bed can no longer expand and settle within one cy-
cle. Therefore, for each variation in amplitude an intermediate fre-
quency provides the fastest separation.
It can be concluded from the results that an optimal amplitude
and frequency exists in terms of separation time.
The number of jigging cycles used to achieve segregation can be
important in terms of operating wear and fatigue. High cycle num-
bers may cause certain mechanical damages earlier. The two equal
fastest profile variants of frequency, T = 1 s and T = 2 s, separate
after 10 s of jigging. However, the high frequency variant of
T = 1 s requires five additional cycles, that is 100% more cycles.
3.3. Jigging profile optimisation
With exception of the high frequency variant of T = 1 s, the
mean particle positions for all other profiles remain at a constant
value for a significant period during suction. These moments corre-
spond to a fixed bed which is a waste of time and energy and show
an opportunity to improve jigging performance by modifying the
profile settings. Fig. 7a displays the mean particle position for a
high frequency variant of T = 1 s where the mean particle position
is constant for an instant. Fig. 7b displays values for a low fre-
quency profile of T = 3 s where a significant duration of the bed
remaining in a static state is present.
Using mean particle position data to identify at what point in
the cycle the bed is at rest, this point is made the beginning of
the following cycle. To illustrate how the profile setting can be im-
proved the settings where changed for variant of T = 2 s and A = 3 L.
The original variant period, T, of 2 s was changed to 1.7 s and as a
consequence the water input is not completely exhausted at the
conclusion of the cycle. This has no influence on results, but can
be a consideration for jigging design, the inlet and exit water veloc-
ity remains the same. Fig. 8a shows the modified profile setting.
The resulting mean particle position is shown in Fig. 8b which dis-
plays a reduction in time the bed is at rest when compared to Fig. 3.
After modification the profile separated after 8.5 s, which is 1.5 s
(15%) faster. With the exception of the high frequency variant of
T = 1 s all other profiles have the ability to increase performance
in terms of segregation speed, while the cycle numbers remain
the same. Applying the same treatment to the fastest profile of
T = 2 s and A = 3.5 L (which one would expect to have a minimal
time that the bed is at rest) the separation time is reduced from
8 s to 6.8 s, again 15% faster.
3.4. Power
There are various parameters to judge the performance of a jig-
ging device. In addition to the separation speed, number of cycles
to achieve segregation, and the final degree of separation as al-
ready discussed, the energy input is an important concern in indus-
trial processes.
The following formula is used to calculate the input power,
where power is a product of total pressure drop DPt, and volumet-
ric flow rate Q:
Power ¼ DPt Q ð6Þ
and
DPt ¼ DPfa þ DP pa þ DPfw þ DPsw þ DPshs þ DP shf ð7Þ
where the total pressure drop DPt is a summation of various pres-
sure drops due to, DPfa, fluid acceleration, DPpa, particle accelera-
tion, DPfw, fluid-to-wall friction, DPsw, solid-to-wall friction, DPshs,
static head of solids, and, DPshf, static head of fluid. The contribu-
tions of wall effects are not resolved in high resolution using the
current model due to the computational effort and complexity.
Although these effects do contribute they are relatively small. The
478 S. Viduka et al. / Advanced Powder Technology 24 (2013) 473–481

(a) 4.5
4

Coordination no. (-)

3.5
3
2.5
2
1.5
1
0.5
0
0 5 10 15 20 25 30 35 40 45 50 55 60
Time (s)
T=1 s, A=1.5 L T=1 s, A=2.25 L T=1 s, A=3 L

(b) 4.5
4
Coordination no. (-)

3.5
3
2.5
2
1.5
1
0.5
0
0 5 10 15 20 25 30 35 40 45 50 55 60
Time (s)
T=2 s, A=1.5 L T=2 s, A=2.25 L T=2 s, A=3 L

(c) 4.5
4
Coordination no. (-)

3.5
3
2.5
2
1.5
1
0.5
0
0 5 10 15 20 25 30 35 40 45 50 55 60
Time (s)
T=3 s, A=1.5 L T=3 s, A=2.25 L T=3 s, A=3 L

Fig. 5. Packed bed coordination number values for all profile variants: (a) T = 1 s, (b) T = 2 s, and (c) T = 3 s.

following power values are calculated using ANSYS CFX commercial
software. The values calculated are not absolute power values yield-
ing only qualitative results. The model does not consider the fluid
pushing through a distributer plate at the inlet which would cause
substantial drag on the fluid. By integrating power over the jigging
time the total energy can be calculated shown in Fig. 9.
The results show energy per cycle is proportional to water vol-
umetric input. Further, the energy required is shown not to be
dependent on the time taken to achieve segregation. A profile
can use little or large amount of energy and segregate either
quickly or slowly. Each profile has individual characteristics and
reasons which describe the energy required to achieve segregation.
There is no correlation between the profiles which can be used as a
quick aid for power evaluation. For example, neither: litres input,
segregation speed, pulsion or suction velocities or duration, alone
correlate to energy. It is a combination of all these variables which
decide the energy outcome.
Further, in six of the nine profiles considerable power is found
to be used in suction. This occurs in situations when the bed falls
into a packed or partially-packed bed while a high suction velocity
and pressure drop is present. Optimising the profile as shown pre-
viously by immediately beginning the following pulsion cycle once
the bed comes to rest can reduce or remove this redundant power
usage. This reduction depends on how high the suction velocity
and low the average porosity is in the remainder of the jigging cy-
cle. Another method is to use hutch water addition. This involves
 S. Viduka et al. / Advanced Powder Technology 24 (2013) 473–481 479

(a) 60 (b) 60

Separation time (s)

Separation time (s)
50 50
40 40
30 30
20 20
10 10
0 0
1 1.5 2 2.5 3 3.5 4 4.5 5 0 1 2 3 4
Volumetric input (L) Period (s)
T=1 s T=2 s T=3 s A=1.5 L A=2.25 L A=3 L

Fig. 6. The effect of separation time against pulsation profile input parameters: (a) volumetric input and (b) cycle period.

(a) 80 (b) 80

Bed height (cm)

Bed height (cm)

70 70
60 60
50 50
40 40
30 30
20 20
10 10
0 0
0 1 2 3 4 5 6 7 8 9 10 11 12 0 3 6 9 12
Time (s) Time (s)
Heavy Light Heavy Light

Fig. 7. Mean particle position showing where profile improvements can be made: (a) T = 1 s, A = 3 L and (b) T = 3 s, A = 3 L.

(a) (b)
0.9 0.9 80
Bed height (cm)

0.6 0.6 70
-1

Velocity (m/s)

60
Water (L) x 10

0.3 0.3 50
0 0 40
-0.3 -0.3 30
20
-0.6 -0.6 10
-0.9 -0.9 0
0 0.85 1.7 2.55 3.4 0 1.7 3.4 5.1 6.8 8.5 10.2 11.9
Time (s) Time (s)
Water input/exhaust Velocity Heavy Light

Fig. 8. Alternative profile: (a) improved profile setting and (b) mean particle position of improved profile.

introducing water from the bottom of the column with a positive the separation time up to a point where the bed ceases to settle be-
upwards velocity nullifying the high negative suction velocity. fore the proceeding cycle. At this point the bed undergoes pulsion
Hutch water addition is not a new concept in relation to controlling in an unsettled state which slows separation. Alternatively,
the dilation of the bed however no references have been made to increasing frequency has a similar effect of increasing amplitude.
controlling energy usage. Separation time reduces until the bed can no longer settle before
the beginning of pulsion in the proceeding cycle resulting in slower
4. Conclusion separation.
The number of cycles used to complete separation is found to be
A DEM–CFD model was used to investigate the particle separa- vastly different and could be a consideration in operation. For
tion process in a jigging device using a sinusoidal profile. The study example, two profiles where found to separate in an equal time
selected three variations in amplitude and applied those over three of 10 s, but with a large five cycle difference, which is 100% more
frequencies. Only one variant did not induce particle separation, cycles.
those that displayed separation did so at different rates and energy The mean particle position indicates the particle bed remains at
usage. Of the nine variants eight were shown to completely sepa- rest for some time during suction which varies depending on var-
rate the particles within 60 s, which was illustrated by a plateau iant. This static moment is a waste of processing time, by eliminat-
in the coordination number. ing this period the separation time is found to reduce significantly.
Broadening the study to more variations in amplitude and fre- Two variants investigated separated 15% faster.
quency it was found an optimal amplitude and frequency exists Total energy required to complete separation is shown not to be
in terms of separation time. Increasing the amplitude will decrease dependent on the time taken to achieve segregation. It is a combi-
480 S. Viduka et al. / Advanced Powder Technology 24 (2013) 473–481

(a) 450
434.4
400
373.6
350
272.1
Total energy (J) 300
250
200
150
100
50
0
0 5 10 15 20 25 30 35 40 45 50 55 60
Time (s)
T=1 s, A=1.5 L T=1 s, A=2.25 L T=1 s, A=3 L
(b) 450
425.9
400
350
Total energy (J)

300
243.8
250
200
184.5
150
100
50
0
0 5 10 15 20 25 30 35 40 45 50 55 60
Time (s)
T=2 s, A=1.5 L T=2 s, A=2.25 L T=2 s, A=3 L
(c) 450
400
350
306.2
Total energy (J)

300
250
200
162.6
150
100
50
0
0 5 10 15 20 25 30 35 40 45 50 55 60
Time (s)
T=3 s, A=1.5 L T=3 s, A=2.25 L T=3 s, A=3 L

Fig. 9. Total energy for all variants: (a) T = 1 s, (b) T = 2 s, and (c) T = 3 s.

nation of many factors which contribute to the final energy out-
come e.g. water velocity, pulsion duration, and separation time.
Further, in six of the nine profiles considerable power is found to
be used in suction. This arises from the fluid being forced at high
velocity through an already packed or partially-packed bed. This
redundant power usage can be reduced or removed by immedi-
ately beginning the following pulsion cycle once the bed comes
to rest. Another possible method is to use hutch water addition
to eliminate the high negative suction velocity.
Finally, this numerical study has proven helpful in the examina-
tion of jigging, however, a future experiment will help quantify the
accuracy of results.
Acknowledgement
This work was supported by a CSIRO OCE PhD scholarship.
References
[1] C.K. Gupta, Chemical Metallurgy: Principles and Practice: Mineral Processing,
Wiley-VCH, Weinheim, 2003 (Chapter 2).
[2] A.K. Mukherjee, V.K. Dwivedi, B.K. Mishra, Analysis of a laboratory jigging
system for improved performance, Minerals Engineering 18 (2005) 1037–
1044.
[3] A.K. Mukherjee, B.K. Mishra, An integral assessment of the role of critical
process parameters on jigging, International Journal of Mineral Processing 81
(2006) 187–200.
 S. Viduka et al. / Advanced Powder Technology 24 (2013) 473–481
[4] A.K. Mukherjee, D. Bhattacharjee, B.K. Mishra, Role of water velocity
for efficient jigging of iron ore, Minerals Engineering 19 (2006) 952–
959.
[5] A.K. Mukherjee, B.K. Mishra, Experimental and simulation studies on the role
of fluid velocity during particle separation in a liquid–solid fluidized bed,
International Journal of Mineral Processing 82 (2007) 211–221.
[6] H. Kellerwessel, Concentration by jigging—current investigations, concepts and
models, Aufbereitungs-Technik 39 (1998) 9–15.
[7] A.J. Clarke, X. Jia, R.A. Williams, D.J. Parker, Verification of distinct element
modelling of particle segregation in laboratory jigs using positron emission
tomography. In: Proceedings Frontiers in Industrial Process Tomography II,
Engineering Foundation/Technical University of Delft, New York, 1997, pp. 91–
96.
[8] H. Schubert, Review of the fundamentals of wet jigging, Aufbereitungs-Technik
35 (1994) 337–349.
[9] d.T.P.R.d. Jong, H.J. Witteveen, W.L. Dalmijn, Penetration velocities in a
homogeneous jig bed, International Journal of Mineral Processing 46 (1996)
277–291.
[10] W. Hentzschel, Die Bewegungsvorgänge monodisperser homogener
Kugelschüttungen unter dem Einfluß vertikaler harmonischer Schwingungen
des Mediums, Freiberger Forsch, Akademie-Verl, 1958.
[11] Y. Jinnouchi, S. Kita, Y. Sawada, M. Tanaka, New trends in theory and
technology of the air-pulsated jigs in Japan, Minerals and Metallurgical
Processing (1984) 76–81.
[12] R.O. Burt, Gravity Concentration Technology, vol. 5, Elsevier, Amsterdam,
Netherlands, 1984.
[13] C.B. Solnordal, T. Hughes, A.H. Gray, P.M. Schwarz, CFD Modelling of a Novel
Gravity Separation Device, in: Seventh International Conference on CFD in the
Minerals and Process Industries, CSIRO, Melbourne, Australia, 2009.
[14] A.J.G. Beck, P.N. Holtham, Computer simulation of particle stratification in a
two-dimensional batch jig, Minerals Engineering 6 (1993) 523–532.
[15] B.K. Mishra, S.P. Mehrotra, Modelling of particle stratification in jigs by the
discrete element method, Minerals Engineering 11 (1998) 511–522.
[16] B.K. Mishra, S.P. Mehrotra, A jig model based on the discrete element method
and its experimental validation, International Journal of Mineral Processing 63
(2001) 177–189.
[17] R. Srinivasan, B.K. Mishra, S.P. Mehrotra, Simulation of particle stratification in
jigs, Coal Preparation 20 (1999) 50–70.
[18] Y. Tsuji, T. Kawaguchi, T. Tanaka, Discrete particle simulation of two-
dimensional fluidized bed, Powder Technology 77 (1993) 79–87.
[19] H.P. Zhu, Z.Y. Zhou, R.Y. Yang, A.B. Yu, Discrete particle simulation of
particulate systems: a review of major applications and findings, Chemical
Engineering Science 63 (2008) 5728–5770.
481
[20] F.W. Mayer, Fundamentals of a Potential Theory of the Jigging Process, in: Proc.
7th Int. Miner. Proc. Cong. New York, 1964, pp. 75–97.
[21] L.M. Tavares, Monte carlo simulations on the potential energy theory of
jigging, Coal Preparation 20 (1999) 71–83.
[22] L.M. Tavares, R.P. King, A useful model for the calculation of the performance
of batch and continuous jigs, Coal Preparation 15 (1995) 99–128.
[23] L. Panda, A.K. Sahoo, A. Tripathy, S.K. Biswal, A.K. Sahu, Application of artificial
neural network to study the performance of jig for beneficiation of non-coking
coal, Fuel 97 (2012) 151–156.
[24] W. Sun, J. Liu, D. Yang, Real time prediction of ash content of clean coal using
Neural Network, Journal of China University of Mining & Technology 34
(2005).
[25] M.M. Ahmed, Optimization of a jigging process using statistical technique,
International Journal of Coal Preparation and Utilization 31 (2011) 112–123.
[26] Y. Jinnouchi, S. Kawashima, How to predict and optimise the pulsation in the
air-pulsated jigs, in: 8th Int. Coal Prep. Cong. (Donetz Doue), Paper B5, 1979.
[27] K. Asakura, M. Mizuno, M. Nagao, S. Harada, Numerical Simulation of Particle
Motion in a Jig Separator, in: 5th Joint ASME/JSME Fluids Engineering
Conference, San Diego, California USA, 2007.
[28] Y.K. Xia, F.F. Peng, Numerical simulation of behavior of fine coal in oscillating
flows, Minerals Engineering 20 (2007) 113–123.
[29] K.J. Dong, S.B. Kuang, A. Vince, T. Hughes, A.B. Yu, Numerical simulation of the
in-line pressure jig unit in coal preparation, Minerals Engineering 23 (2009)
301–312.
[30] Y. Xia, F.F. Peng, E. Wolfe, CFD simulation of fine coal segregation and
stratification in jigs, International Journal of Mineral Processing 82 (2007)
164–176.
[31] S.I. Rubinow, J.B. Keller, The transverse force on a spinning sphere moving in a
viscous fluid, Journal of Fluid Mechanics 11 (1961) (1961) 447–459.
[32] P.G. Saffman, Lift on a small sphere in a slow shear flow, Journal of Fluid
Mechanics 22 (1965) 385–400.
[33] P.G. Saffman, Corrigendum to ‘‘The lift on a small sphere in a slow shear flow’’,
Journal of Fluid Mechanics 31 (1968) 624.
[34] A.B. Basset, Treatise on Hydrodynamics, Deighton, Bell and Co., Cambridge,
1961.
[35] R. Di Felice, The voidage function for fluid-particle interaction systems,
International Journal of Multiphase Flow 20 (1994) 153–159.
[36] Y.Q. Feng, A.B. Yu, Assessment of model formulations in the discrete particle
simulation of gas-solid flow, Industrial and Engineering Chemistry Research
43 (2004) 8378–8390.
[37] S. Viduka, Y. Feng, K. Hapgood, M.P. Schwarz, Discrete particle simulation of
particle separation in a Jigging device, in: The 4th Conference on Industrial
Fluidization, Gauteng, South Africa, 2011.