See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/346966531

Analysis of Particle Size Distribution of Multisize Particulate Slurry Flow in

Horizontal Pipeline Using CFD

Conference Paper · March 2016

CITATIONS READS

0 173

3 authors:

Rahul Tarodiya Subodh Khullar

Visvesvaraya National Institute of Technology Indian Institute of Technology Roorkee
31 PUBLICATIONS 340 CITATIONS 19 PUBLICATIONS 67 CITATIONS

SEE PROFILE SEE PROFILE

Bhupendra K. Gandhi
Indian Institute of Technology Roorkee
112 PUBLICATIONS 3,742 CITATIONS

SEE PROFILE

All content following this page was uploaded by Rahul Tarodiya on 12 December 2020.

The user has requested enhancement of the downloaded file.

 International Conference on Research Innovations in Science and Engineering (RISE’2015),
March 20-21, 2016, Mauritius

ANALYSIS OF PARTICLE SIZE DISTRIBUTION

OF MULTISIZE PARTICULATE SLURRY FLOW
IN HORIZONTAL PIPELINE USING CFD
Rahul Tarodiya1, Subodh Khullar2, and Bhupendra K. Gandhi3

 reported that most of the equations developed are based on

Abstract— For commercial slurry transportation, particle sizes the limited data on materials having equisized and narrow
are varied over wide ranges. It is important to analyze their size-range particles. Keeping this, they further modified
distribution across the pipe cross section in order to enhance the Karabelas model by improving some of the restrictive
design of slurry pipeline. This paper shows distribution of six
particle sizes in a 105 mm diameter horizontal pipeline transporting
assumptions and discussed the concentration distribution of
multisized particulate slurry of zinc tailing and water at 3.5 m/s flow six particle size ranges from 38 to739 µm of zinc tailing
velocity and 26% efflux concentration using three dimensional CFD slurry flowing in 105 mm diameter pipe.
modelling. The computational model was developed using CFD With the advent of increased computational capabilities, it
code Fluent 14.0. The validation of simulation results was becomes endeavor of the researchers to develop the
performed with experimental results reported by Kaushal and
computational fluid dynamics (CFD) model to predict the
Tomita [6]. The predicted distribution of vertical solid concentration
profile for different particle sizes showed excellent agreement. flow characteristic during solid liquid transportation. Since
Keywords— Multisized particulate slurry, CFD, Concentration slurry flow is very complex and therefore applicability of CFD
profiles. code has certain limitations. In spite that CFD studies of solid
liquid multiphase flow has been increasingly used now a
I. INTRODUCTION days. The advantage to perform CFD study is that it saves the
time and cost of performing experiments and also wide range

S olid-liquid (slurry) transportation through long

pipeline has been used for decades for transporting
different solid materials such as coal, mineral, ore,
concentrates, tailings and many more. It has certain
of parameters affecting flow conditions can be evaluated
which is not possible with experimental study. Ling et al [7]
investigated the distribution of silica sand particle in a
advantages, like eco-friendliness, low maintenance and horizontal pipe line of diameter 22.1 mm using a simplified
operating cost. In any commercial application, the multisized three dimensional algebraic mixture model along with RNG
solids are being transported with maximum to minimum ratio k-ε turbulent model in CFD code Fluent 5.0. They reported
of particle sizes up to 103. The knowledge of particle size and that the distribution of solid concentration is symmetric in the
concentration distribution during conveying improve the horizontal direction but have large gradients in vertical
design of the slurry pipeline. In spite of significant research, direction. Lahiri and Ghanta [8] predicted the solid particle
the design of slurry pipeline is based on empirical (glass beads) distribution in a vertical plane by using
correlations obtained through experiments. Karabelas [1] Eulerian–Eulerian multiphase model along with standard k-ε
developed a model for the prediction of concentration turbulence model of CFD code Fluent 6.2 for slurry flow in
distribution of multisized slurry. During further research in horizontal pipe of 105 mm diameter up to 50% solid loading.
this field the inconsistency of Karabelas model compared to They found reasonable agreement between the predicted
experimental results has been highlighted due to the results and the experimental results of Kaushal et al. [9]. The
assumption of constant particle diffusivity, which limits this limitation of mixture model of Fluent for multiphase used by
model to lower concentration levels only [2]. Many efforts Ling et al. [7] was highlighted by Kaushal et al. [10] for its
have been done for the development of a reasonable model for incapability at high solid concentration. They further
the prediction of concentration distribution of multisized simulated the slurry flow consisting of water and glass bead
slurries, which includes the work of Seshadri et al. [3], Roco particles in a horizontal pipeline of 54.9 mm internal
and Shook [4], Gillies et al. [5]. Kaushal and Tomita [6] diameter for up to 50% concentration by using Mixture and
Eulerian multiphase models. They observed that the
Rahul Tarodiya1,Research Scholar, Department of Mechanical and Industrial concentration distribution obtained from Eulerian model
Engineering IIT Roorkee, India (e-mail: [email protected]) showed good agreement with their experimental results.
Subodh Khullar 2, Research Associate, Department of Mechanical and
Industrial Engineering IIT Roorkee, India (e-mail: The study of literatures reveals that numerical study to
[email protected]). predict the concentration distribution was restricted only for
Bhupendra K. Gandhi 3, Professor, Department of Mechanical and Industrial
Engineering IIT Roorkee, India (corresponding author’s phone: +91-1332-
mono sized particulate slurry, which is not the case in actual
285544, e-mail: [email protected]). slurry transportation application. In this work distribution of
1/4
multisized particulate slurry in a fully developed pipe flow Cvm is the coefficient of virtual mass force and
has been performed using Eulerian multiphase model in 105 CL s f (vf  vs )  (vf ) is lift force, where C L is the lift
mm diameter horizontal pipe on zinc tailings with specific
coefficient. Due to minor contribution of lift and virtual mass
gravity of 2.82, for six particle sizes ranging from 38-739 µm,
forces, these are not considered in the simulation.
keeping the flow velocity 3.5 m/s and solid concentration
26% by volume. The predicted concentration profiles are III. NUMERICAL METHOD
compared with the experimental results of Kaushal and
Tomita [6]. 3.1 Geometry and Meshing
O-type structured mesh was generated using Ansys ICEM
II. MATHEMATICAL MODEL 14.0 for the 5 m long, 105 mm internal diameter horizontal
Slurry flow of zinc tailings and water through horizontal pipe consists of 1,658,972 nodes. The pipe length was much
pipeline has been modeled using Eulerian-Eulerian greater than the required entrance length for fully developed
multiphase model of CFD code Fluent 14.0. This approach flow. The minimum orthogonal quality of the mesh is 0.788
allowed the determination of separate flow field solution after and the maximum aspect ratio is 4.39.
solving continuity and momentum equation for each phase.
For pressure-velocity coupling, the Phase Coupled SIMPLE
(PC-SIMPLE) algorithm [11] has been used. This algorithm
is an extension of the SIMPLE algorithm [12]. The velocities
are solved coupled by phases, but in a segregated fashion.
Pressure and velocities are then corrected so as to satisfy the
continuity equation. To model the turbulence, the standard k-
є model has been used.

In the Eulerian-Eulerian approach, the slurry flow consists

of solid(s) and fluid(f) phases, are considered to be Fig.1 3D O-grid meshing of slurry pipeline
interpenetrating continua, so that s   f  1 ,
where  f and s are the volumetric concentrations of fluid and 3.2 Boundary Conditions
The calculation domain is bounded by three faces, the inlet
solid phase, respectively. The laws for the conservation of
mass and momentum are satisfied for each phase boundary, the wall boundary and the outlet boundary. The
individually. inlet and outlet boundaries are specified with a single primary
phase (water) and six numbers of secondary phases (zinc
2.1 Continuity equation tailings),  s  C vjf and  f  1  C vjf . Secondary phases have
.( t t vt )  0 (1)
different mean particle diameter and initial solid volume
where, t is either s or f.
fraction (Table 1). The initial volume fraction of the solids of
2.2 Momentum equations
For fluid phase: different sizes was selected based on the data of efflux
concentration of multisized zinc tailings slurry reported by
.( f f vf vf )   f P  .f   f f g  K sf (vs  v f )
Kaushal and Tomita [6]. Inlet velocity and volume fraction at
 Cvm s f (vs . vs  vf . vf ) (2) inlet boundary condition and pressure outlet and back flow
 CL s f (vf  vs )  (  vf ) volume fraction at outlet boundary was set. The initial
For solid phase: velocity of all the phases at inlet is set to be equal. The outlet
.(s s vs vs )  s P  Ps  .s   s s g  K fs (v f  vs ) pressure was set to zero gauge pressure. No slip condition was
used to model liquid and solid velocity at the wall. The drag
 Cvm s f (vf . vf  vs . vs ) (3)
force was modelled using the Syamlal-O’Brien model [13].
 CL s f (vs  vf )  (  vf )
where, P , Static pressure gradient, Ps , Inertial force due to 3.3 Solution strategy and convergence
particle interaction Ksf (vs  vf ) , Drag force caused by the A first order upwind discretization scheme was used for
velocity difference between two phases, where K sf is the the momentum equation, volume fraction, turbulent kinetic
and turbulent dissipation energy. These schemes ensured, in
momentum exchange coefficient between fluid and solid
general, satisfactory accuracy, stability and convergence. The
phase, v s and v f are velocity of solid and fluid phase, convergence criterion is based on the residual value of the
respectively. .f is the stress tensor for fluid. g is the body calculated variables, i.e., mass, velocity components for each
phases, turbulent kinetic energies, turbulent energy
force, Cvm s f (vf .vf  vs .vs ) is Virtual mass force, where
dissipation rate and volume fraction. In the present

2/4
calculations, the threshold values were set to a 10 -4 times the
initial residual value of each variable. The under relaxation
factors of momentum, volume fraction, turbulence kinetic
energy and turbulence energy dissipation is taken as 0.7. The
initial volume fraction of solids for each solid phase was set
by using the patch option in software.

TABLE 1
DETAILS OF PARTICLE SIZE DISTRIBUTION
Solid phase Mean particle diameter, Initial solid volume
dj (µm) fraction, Cvjf (%) (e) d5= 91 µm (f) d6= 38 µm
Fig.2. Comparison of predicted and experimental (Kaushal and
1 739 0.935 Tomita [6]) solids concentration profiles of different particle sizes in
2 255 2.543 a multisize slurry flow.
3 180 1.765
4 128 5.19
5 91 3.504
6 38 12.02

IV. RESULTS AND DISCUSSION:

Figure 2 (a-f) shows the comparison of experimental results
of Kaushal and Tomita [6] and the predicted results from
modelling, for the distribution of concentration of different
particle sizes along the vertical profile of the pipeline at Fig.3. Predicted concentration profiles for different particle sizes in
outlet. These curves show an excellent agreement of the multisize slurry flow.
predicted and experimental results. This indicates that the
CFD model has capability to capture the flow of multisize Figure 4(a-f) shows the contours of solid concentration
particulate slurry in a pipeline and based on that it can be distribution for different particle sizes. It is observed that the
possible to evaluate the effect of each particle size during the distribution of smallest size particles (38µm) is nearly
flow of multisize slurry, which helps in further prediction of uniform at this velocity although their fraction in multisized
minimum velocity required to keep particles in suspension slurry is the highest. On increasing the size of particles,
with multisize slurry flow. The relative concentration settling characteristic has been observed at this velocity and
distribution of different particles sizes is presented in Figure the particles are settled at lower half portion of the pipe and
3. It is observed that for all the sizes, higher concentration is at top most of the flow is only water. Distribution of 739 µm
observed at bottom of the pipe due to the effect of gravity. particles which is the largest size is also nearly uniform due
to their low volume fraction in the slurry. Based on this
behavior it seems that on keeping the volume fraction of
bigger size particles small the flow behavior of multisized
slurry improves.

(a) d1= 739 µm (b) d2= 255 µm

(a) d1= 739 µm (b) d2= 255 µm

(c) d3= 180 µm (d) d4= 128 µm

(c) d3= 180 µm (d) d4= 128 µm

3/4
 [13] M. Syamlal, W. Rogers, and T.J. O'Brien, MFIX Documentation, U.S.
Department of Energy, Federal Energy Technology Center, Morgantown,
1993.

(e) d5= 91 µm (f) d6= 38 µm

Fig.4. Contours of the predicted solid concentration distributions of
different particle sizes in multisize slurry flow.

V. CONCLUSION
The modelling through CFD for the transportation of
multisized particulate slurry from horizontal pipeline has
been carried out. This was done to explore the capability of
CFD modelling for multisized particulate slurry flow. The
following conclusions are made through study:
1. The proposed model is capable to capture the distribution
of different solid particle sizes across the pipe cross section
for multisize slurry flow.
2. Concentration gradient is present along the vertical plane
in the pipe for all particle sizes. The higher concentration
zone is at lower half portion of the pipeline.
3. Keeping small volumetric concentration of large size
particles in multisize slurry improves the flow behavior of
large size particles.

REFERENCES
[1] A. J. Karabelas, “Vertical distribution of dilute suspensions in turbulent
pipe flow,” AIChE J., vol. 23, pp. 426–434, 1977 .
[2] V. Seshadri, R. C. Malhotra and C. Adiseshu, “Prediction of distribution of
solids across the cross section of a slurry pipeline,” in 1984 Proc. 13th
Nat. Conf. on FMFP, R.E.C., Trichurapalli, India.
[3] V. Seshadri, R. C. Malhotra and K. S. Sunder, “Concentration and size
distribution of solids in a slurry pipe line,” in 1982 Proc. 11th Nat. Conf.
on FMFP, BHEL Hyderabad, India.
[4] M. C. Roco and C. A. Shook, “Computational method for coal slurry pipe
line with heterogeneous size distribution,” J. of Powder Technology, vol.
2, pp. 159-176, 1984.
[5] R. G. Gillies, C. A. Shook and K. C. Wilson, “An improved two layer
model for horizontal slurry pipeline flow,” The Canadian J. of Chemical
Engineering, vol. 69, pp. 173–178, 1991.
[6] D. R. Kaushal and Y. Tomita, “Solids concentration profiles and pressure
drop in pipeline flow of multisized particulate slurries,” Int. J. Multiphase
Flow, vol. 28, pp. 1697-1717, 2002.
[7] J. Ling, P. V. Skudarnov, C. X. Lin and M. A. Ebadian, “Numerical
investigations of liquid–solid slurry flow in a fully developed turbulent
flow region,” Int. J. Heat Fluid Flow, vol. 24, pp. 389–398, 2003.
[8] S. K. Lahiri and K. C. Ghanta, “Slurry flow modelling by CFD,” Int. J.
Chemical Industry & Chemical Engineering Quarterly, vol. 16, no. 4, pp.
295−308, 2010.
[9] D. R. Kaushal, K. Sato, T. Toyota, K. Funatsu and Y. Tomita, “Effect of
particle size distribution on pressure drop and concentration profile in
pipeline flow of highly concentrated slurry,” Int. J. Multiphase Flow, vol.
31, pp. 809–823, 2005.
[10] D. Kaushal, T. Thinglas, Y. Tomita, S. Kuchii and H. Tsukamoto, “CFD
modeling for pipeline flow of fine particles at high concentration,” Int. J.
Multiphase Flow, vol. 43, pp. 85-100, 2012.
[11] S. A. Vasquez and V. A. Ivanov, “A phase coupled method for solving
multiphase problem on unstructured meshes,” Fluids Engineering
Division Summer Meeting, ASME, Boston, USA, 2000.
[12] S. V. Patankar, Numerical Heat Transfer and Fluid Flow, McGraw-Hil,
New York, 1980.

4/4

View publication stats