Tunnelling and Underground Space Technology 71 (2018) 223–230

Contents lists available at ScienceDirect

Tunnelling and Underground Space Technology

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

Numerical investigation of pipeline transport characteristics of slurry shield MARK

under gravel stratum
⁎
Duan Yanga, Yimin Xiaa,b, , Dun Wuc, Peng Chenc, Guiying Zenga, Xianqiong Zhaoa
a
School of Mechanical and Electrical Engineering, Central South University, Changsha 410083, China
b
State Key Laboratory of High Performance Complex Manufacturing, Central South University, Changsha 410083, China
c
China Railway 14th Construction Bureau Co. Ltd., Jinan 250014, China

A R T I C L E I N F O A B S T R A C T

Keywords: In order to grasp the pipeline transport characteristics of slurry shield under gravel stratum, with a cross-river
Slurry shield tunnel construction of Lanzhou metro as the research background, the numerical simulation method is adopted,
Pipeline transport and a slurry-stone two-phase flow model is established based on multiphase flow theory by considering the
Two-phase flow rheological properties of slurry. Besides, corresponding simulation models validated by experiment are estab-
CFD-DEM
lished using CFD and DEM softwares, and the impacts of inlet slurry velocity, stone volumetric concentration and
pipe inclination angle on pressure drop and transport capacity of the pipeline are investigated. The results show
that stones are mainly distributed at the bottom of the pipeline for horizontal straight pipes. The pressure drop
and transport capacity increase exponentially with the increase of inlet slurry velocity and stone volumetric
concentration. For inclined straight pipes, the pressure drop of the pipeline increases slowly with the pipe in-
clination angle increasing from 0° to 60°, but increases rapidly when the pipe inclination angle increases from
60° to 90°; the transport capacity reduces slowly at first and then increases rapidly with the increase of pipe
inclination angle, and the lowest value appears at 45–60°.

1. Introduction horizontal pipe were found to be markedly higher than those in the
vertical pipe. Edelin et al. (2015) experimentally investigated the
Slurry shield machine is a kind of large-scale tunnel boring machine transport of fluids composed of water and polypropylene particles
which integrates functions such as digging and cutting soil, conveying (341–756 μm). The best operating conditions for energy optimum are
soil, lining and oriented correction. It injects pressurized slurry into the obtained close to the limit deposition velocity, when the fully sus-
cutter compartment to balance the water and earth pressure of the pended pattern is reached. Wu et al. (2015a, 2015b) investigated the
excavated surface, which can effectively maintain the pressure balance pipeline transportation of the cemented coal gangue-fly ash backfill
of the excavation surface and control the ground subsidence. For this (CGFB) mixtures using a test loop system and found the solid contents
reason, it is widely used in various kinds of cross-river and cross-sea and the mix proportions of the coal gangue, fly ash and cement have a
tunnel projects. When the slurry shield is excavating in gravel stratum, significant impact on the transportability and pressure drop of the CGFB
a large number of gravels generated by cutters cutting and crushers slurry in the pipe loop. Matousek (2011) established sand
crushing will need to be transported from the excavation chamber to (0.22–1.38 mm) transport formula based on experiments in the la-
the outside of the tunnel through the slurry pipeline. Therefore, boratory circuit with a pipe of inner diameter of 100 mm. Using PIV
grasping the pipeline transport characteristics of the slurry shield is the (particle image velocimetry) and EIT (electrical impedance tomo-
key to ensuring the efficiency and safety of tunnel constructions. graphy) technologies, Vlasak et al. (2014b), Pinto et al. (2014) and
At present, a number of scholars have used experimental method to Hashemi et al. (2014) observed the motion of particles (< 8 mm) in
studied the pipeline transport characteristics. With water as the car- pipes under different fluid velocities, and determined the critical ve-
rying fluid, Ravelet et al. (2013) and Vlasak et al. (2014a) experimen- locities for different flow regimes. Nevertheless, these experimental
tally investigated the pressure drop of horizontal and inclined pipes of studies mostly focus on small diameter pipes transport with water as the
inner diameter D = 100 mm. The hydraulic gradient under different carrying fluid, which is different from slurry shield pipeline transport.
mixing velocities was obtained and the frictional pressure drops in the With the development of computer, the numerical simulation

⁎
Corresponding author at: School of Mechanical and Electrical Engineering, Central South University, Changsha 410083, China.
E-mail address: [email protected] (Y. Xia).

http://dx.doi.org/10.1016/j.tust.2017.08.011
Received 20 January 2017; Received in revised form 30 May 2017; Accepted 8 August 2017
Available online 19 September 2017
0886-7798/ © 2017 Elsevier Ltd. All rights reserved.
D. Yang et al. Tunnelling and Underground Space Technology 71 (2018) 223–230

method is increasingly applied to the research of pipeline transporta- ∂ k

(αf ρf T ) + ∇ ·(αf ρf Uf T ) = ∇ ·⎛ T ∇T ⎞ + sT
⎜ ⎟

tion, which can significantly shorten research cycle, cut research cost ∂t ⎝ cP ⎠ (3)
and improve research quality. Kaushal et al. (2012) numerically si-
mulated the pipeline slurry flow of mono-dispersed fine particles at where αf , ρf , Uf and pf are the fluid volume fraction, density, velocity
high concentration using Mixture and Eulerian two-phase models, and vector and pressure respectively, τ is the viscous stress tensor, g is the
their results show that the Eulerian model gives more accurate pre- gravitational acceleration, Fpf is the force exerted by all particles on
dictions for both the pressure drop and concentration profiles at high fluid, T is the temperature, kT and cP are the heat transfer coefficient
concentration. Capecelatro and Desjardins (2013) and Arolla and and the fluid specific heat ratio, respectively, and sT is the viscous
Desjardins (2015) used Euler–Lagrange large eddy simulation metho- dissipation phase.
dology to predict the physics of turbulent liquid–solid slurry flow Due to the large diameter of pipe and high flow velocity, the slurry
through a horizontal periodic pipe. The simulation results suggest that flow in the slurry shield pipeline is a typical turbulent flow. In this
this computational strategy is capable of predicting critical deposition paper, the two-equation k−ε turbulence model (Launder and Spalding,
velocity. Based on the kinetic theory of granular flow, Chen et al. 1974), which has been widely used in studies of turbulent flow (Grewal
(2009) and Gopaliya and Kaushal (2016) simulated the flow of coal- et al., 2013; Zhao et al., 2014; Wadnerkar et al., 2012; Torano et al.,
water slurry (CWS) and sand-water slurry (SWS) in horizontal pipelines 2011; Wang et al., 2011), is adopted to simulate the turbulence.
using an Eulerian multiphase approach, and the pressure drop, velocity The slurry used in the cross-river tunnel construction of Lanzhou
distribution as well as concentration distribution of CWS and SWS were metro is a mixture of water, bentonite, red clay and pulping agent. The
obtained. Using the computational fluid dynamics (CFD) model, Eesa pulping agent, which is mainly made of polyacrylic acid salts, cellulose
and Barigou (2009) investigated the pipe transport of coarse solids salts and anti-calcium agent modified soil, can effectively increase the
(2–9 mm) in laminar power law fluids, and found the maximum particle adsorption between particles. Therefore, the shear viscosity of slurry is
velocity is always significantly less than twice the mean flow velocity improved and the amount of slurry leaked into stratum is reduced,
for shear thinning fluids, but it can exceed this value in shear thickening which makes a great contribution to keeping the excavation surface
fluids. Tian et al. (2014) established the ice slurry flow model and in- stable. With the slurry on the construction site as the sample, the
vestigated the influences of ice packing fractions, ice particle size, ad- rheological curve is obtained using a six-speed rotary rheometer, as
ditive concentrations and pipe size on transportation safety. Liu et al. shown in Fig. 1. It can be observed from Fig. 1 that the shear rate of
(2015) used a simplified CFD-based procedure consisting of flow shield slurry is linear with the shear stress (determination coefficient
modeling, particle tracking, and penetration calculation to calculate the R2 = 0.9914 ), which can be described by Bingham fluid rheological
penetration rates in elbows for annular flow. By comparing with ex- equation:
perimental data, the new method is proved to be reasonable in sim-
plifying the simulation of annular flow field and shows good accuracy τ = τd + μp γ̇ (4)
in erosion prediction.
Conceivably, numerical simulation has been widely used to study where τ is shear stress, τd is yield stress, μp is plastic viscosity, and γ̇ is
the characteristics of pipeline transportation and desirable results have shear rate.
been achieved. In this paper, with a cross-river tunnel construction of In addition, when τd = 0 , Bingham fluid rheological equation can be
Lanzhou metro as the research background, a three-dimensional fluid- used to describe Newtonian fluid, such as water.
particle two-phase flow model is established based on multiphase flow
theory by considering the rheological properties of the fluid.
2.2. Particle phase
Subsequently, pipeline transport characteristics of slurry shield under
gravel stratum are investigated through numerical simulation based on
In the pipeline transportation process, the stone is forced by other
CFD-DEM technology.
stones, slurry and pipe wall, which can be described by Newton’s
second law of motion using the discrete element method originally
2. Mathematical model proposed by Cundall and Strack (1979), as follows:
n
The stones broken by slurry shield cutters and crushers are carried dUp,i
mi = mi g + ∑ (Fc,ij + Fd,ij ) + (Fc,iw + Fd,iw ) + Ff ,i
into the slurry pipeline with the slurry flow, and finally pumped to the dt j=1 (5)
outside of the tunnel by multi-stage slurry pumps. The flow process is
complicated due to the strong interaction between slurry and stones. In
the present work, the rheological property model of slurry is obtained
by experiment, and the three-dimensional slurry flow model, the stone
motion model and other related models regarding flow characteristics
are established by employing multi-phase flow theory.

2.1. Fluid phase

The fluid flow and heat transfer processes follow mass conservation
law, momentum conservation law and energy conservation law, with
their corresponding mathematical models being the continuity equa-
tion, Navier-Stokes equation and energy conservation equation, as fol-
lows:

∂αf ρf
+ ∇ ·(αf ρf Uf ) = 0
∂t (1)

∂
(αf ρf Uf ) + ∇ ·(αf ρf Uf Uf ) = −∇pf + ∇ ·αf τ + αf ρf g + Fpf
∂t (2) Fig. 1. The rheological curve of slurry.

224
D. Yang et al. Tunnelling and Underground Space Technology 71 (2018) 223–230

n
dωi method of standard wall functions (Tao, 2001).
Ii
dt
= ∑ Tij + Tiw
Velocity of grid node at the wall is expressed as:
j=1 (6)
∗
where mi , Up,i , Ii and ωi are the mass, velocity, motion of inertia and ⎧y ,y∗ < ym∗
u∗ = 1
rotational velocity of particle i ; Fc,ij , Fd,ij and Tij are the inter-particle ⎨ κ ln(Ey ) ,y∗ > ym∗
∗
(14)
⎩
contact force, viscous damping force and torque between particle i and
other particles; Fc,iw , Fd,iw and Tiw are the contact force, viscous damping u∗
where = (u−u w )/ u τ , u τ = (τw /ρf )1/2 ,
u is the tangential velocity of the
force and torque between particle i and pipe wall. Ff ,i is the force ex- fluid, u w is the tangential velocity at the wall, τw is the shear stress on
erted on particle i by the fluid phase. the wall, y∗ ≡ ρf Cμ1/4 k1/2y / μe , E and κ are the empirical coefficient and
In this paper, the contact force, viscous damping force and torque the von Karman coefficient, respectively, y is the distance from the wall
1
are determined by the Hertz-Mindlin contact model (Mindlin and to the first grid node, and ym∗ solves ym∗ − κ ln(Eym∗ ) = 0 .
Deresiewicz, 1953; Cundall, 1988). The turbulence parameters meet:
∂k
⎧ ∂n = 0
2.3. Particle-fluid interactions ⎪
⎪Gk = τw τw
κρf Cμ1/4 k1/2y
The interaction force between stone and slurry mainly consists of ⎨
⎪ Cμ3/4 k 3/2
buoyancy FB, drag force FD and pressure gradient force FP. ⎪ε =
⎩ κy (15)
For spherical particles, these forces can be calculated as follows:
ρf where k , Gk and ε are the turbulence kinetic energy, the production of
FB = − g kinetic energy and the dissipation rate; n is the local coordinate normal
ρp (7) to the wall.
18μe CDRep In addition, particles are all simplified to spherical particles of a
FD = (Uf −Up) certain diameter and stacked at the bottom of the pipe with a certain
ρP dp2 24 (8)
volumetric concentration. In order to ensure that the volume fraction
1 has a certain value, particles escaping from the outlet will flow into the
FP = − ∇p
ρp f (9) pipeline from the inlet again.

where Up is particle velocity, μe is fluid effective viscosity, ρP is particle 3. Validation of model effectiveness
density, dp is particle diameter, the particle Reynolds number Rep and
the drag coefficient CD are calculated as follows:
The hydraulic lifting test platform, which mainly consists of a pump,
ρdp |Up−Uf | vertical steel pipes, a transparent glass pipe and a water tank, is used to
Rep =
μ (10) verify the effectiveness of the fluid-solid two-phase model. Fig. 2 il-
lustrates the arrangement of the transparent glass pipe, through which
a a
CD = a1 + 2 + 3 the movement of particles in the pipeline can be observed. The location
Re Re (11)
of the particles is recorded by a Phantom V310 high-speed video
where Re is fluid Reynolds number; a1, a2 and a3 are constants which are camera, and then the particles’ suspending velocity is obtained by
given by Morsi and Alexander (1972) as follows: image processing software.
The density of experimental particles is 1860 kg/m3, the diameter is
⎧ 0,24,0 0 < Re < 0.1
5, 10, 15 and 20 mm, respectively, and the total weight for each ex-
⎪3.690,22.73,0.0903 0.1 < Re < 1
⎪1.222,29.1667,−3.8889 periment is 10 kg. Fig. 3 illustrates the distribution of particles under
1 < Re < 10
⎪
⎪ 0.6167,46.50,−116.67 different inlet water velocities when the particle diameter is 20 mm. It
10 < Re < 100
a1,a2,a3 = can be seen from Fig. 3 that particles leave the bottom of the pipe
⎨ 0.3644,98.33,−2778 100 < Re < 1000
⎪ 0.357,148.62,−47500 gradually with the increase of inlet water velocity. When the inlet water
1000 < Re < 5000
⎪ velocity reaches 0.61 m/s, the particles are evenly distributed in the
⎪ 0.46,−490.546,578700 5000 < Re < 10000
pipe and reach the suspended state. The inlet water velocity at this
⎪ 0.5191,−1662.5,5416700 Re⩾10000
⎩
moment is recorded as the particle group suspension velocity.
With water as the carrying liquid, the corresponding rheological
2.4. Boundary conditions properties equation is τ = 0.001γ ̇, the hydraulic lifting process is

The slurry shield pipe is a circular pipe made of steel. One end of the
pipeline is the inlet, and the other end is the outlet. The inlet boundary
is assumed to be a uniform distribution of fluid velocity. The turbulence
kinetic energy and the dissipation rate at the inlet are calculated as
follows (Nakayama et al., 1996; Tao, 2001):
kin = 0.005vin2 (12)

εin = Cμ kin3 /0.015De (13)

where kin and εin are the turbulence kinetic energy and the dissipation
rate at the inlet, vin is the fluid velocity, Cμ = 0.085 is an experimental
constant, and D is the equivalent diameter of pipe.
The outlet boundary is assumed to be a uniform distribution of
pressure and the value of pressure is set at 0 (which equals to atmo-
spheric pressure). The pipe wall boundary is treated with the law of
solids, which assumes that all the nodes of the solid wall have a non-slip
Fig. 2. The arrangement of the transparent glass pipe.
condition. The first grid nodes around the wall can be solved using the

225
D. Yang et al. Tunnelling and Underground Space Technology 71 (2018) 223–230

Fig. 5. Physical model of pipe for slurry shield.

With the shield slurry pipeline of this cross-river tunnel construction

as the research object, a pipe physical model with the diameter of
300 mm and the length of 5 m is established and the stone is simplified
as a spherical particle with the diameter of 20 mm, as shown in Fig. 5.
The static friction coefficient and the rolling friction coefficient are set
as 0.3 and 0.01, respectively, and the sliding friction coefficient is as-
sumed to be equal to the static friction coefficient.
Through numerical simulation of the above model, the impacts of
Fig. 3. The distribution of particles under different inlet water velocities (m/s). inlet slurry velocity, stone volumetric concentration and pipe inclina-
tion angle on the pressure drop and transport capacity of the pipeline
are investigated

4.1. Horizontal straight pipe

In order to reduce the wear rate and extend the service life of the
pipeline, the inlet slurry velocity is always less than 5 m/s. In this work,
the inlet slurry velocity of 1, 2, 3, 4 and 5 m/s and the stone volumetric
concentration of 0.05, 0.1, 0.15 and 0.2 are taken to explore their im-
pacts on the pressure drop and transport capacity of the horizontal
straight pipe.
When the stone volumetric concentration is 0.1, the velocity dis-
tribution of stones under different inlet slurry velocities is shown in
Fig. 6. As can be seen from Fig. 6, it is difficult for the stones to over-
come the gravity and reach a state of complete suspension motion
under the actions of slurry buoyancy and turbulence since the stone
diameter is 20 mm (the corresponding mass is 11.3 g). Therefore, the
stones are mainly distributed at the bottom of the pipeline. The slurry
Fig. 4. The comparison of particle group suspension velocity between experiment data
flows into the pipeline and exchanges kinetic energy with stones, which
and simulation predicted. drives the stones to move forward. Due to the accumulation of stones,
the closer the stone is to the bottom of the pipe, the less affected by the
slurry flow it is. As a result, the velocity distribution of stones is also
simulated based on the above fluid–solid two-phase model. The com-
uneven: the velocity is high for stones at the top, but low for stones at
parison of particle group suspension velocity between experiment data
the bottom. When the inlet slurry velocity is 1 m/s, it is difficult for
and simulation predicted is illustrated in Fig. 4. Due to the uneven
stones at the bottom to overcome the friction between the wall and
surface of the experimental particles which increases the drag force
continues to remain in a static state. With the increase of inlet slurry
exerted on particles by water, experiment data is slightly smaller than
velocity, the stones at the bottom begin to move gradually and the
simulation predicted with the average relative error as 11.97%, which
overall velocity of stones increases, which means the transport capacity
validates the effectiveness of the model.
of the pipeline is improving. Fig. 7 illustrates the slurry velocity dis-
tribution on x = 0 profile under different inlet slurry velocities. It can
4. Results and discussions be observed from Fig. 7 that the velocity distribution of the slurry is
obviously stratified: the velocity is high where the stone volumetric
The shield slurry pipeline in the cross-river tunnel construction of concentration is low and the velocity is low where the stone volumetric
Lanzhou metro is mainly composed of horizontal straight pipes and concentration is high.
inclined straight pipes with the diameter of 300 mm. The diameter and Fig. 8 illustrates the total pressure distribution of slurry when inlet
density of the stones produced by tunneling are about 20 mm and slurry velocity is 3 m/s and stone volumetric concentration is 0.1. It can
2700 kg/m3, respectively. The density of shield slurry is 1100 kg/m3, be observed from Fig. 8 that: at the inlet (z = 0 m), the total pressure at
and the rheological properties equation is given as follows: the pipe bottom is higher than the other positions. With the increase of
τ = 3.98 + 0.00912γ ̇ (16) the distance from the inlet, the total pressure at the pipe bottom drops
more sharply than the other positions due to the collision and friction

226
D. Yang et al. Tunnelling and Underground Space Technology 71 (2018) 223–230

Fig. 6. The velocity distribution of stones under dif-

ferent inlet slurry velocities (m/s).

between stones and pipe wall. varying curves of the transport capacity of the pipeline with the inlet
With the average total pressure difference between inlet and outlet slurry velocity and the stone volumetric concentration are obtained, as
as the pressure drop, Fig. 9 illustrates the relationship between pressure shown in Figs. 11 and 12, respectively. It can be seen from
drop and inlet slurry velocity, and Fig. 10 illustrates the relationship Figs. 11 and 12 that the transport capacity of the pipeline increases
between pressure drop and stone volumetric concentration. It can be exponentially with the increase of inlet slurry velocity and stone vo-
seen from these two figures that the pressure drop increases ex- lumetric concentration. When the stone volumetric concentration is
ponentially with the increase of inlet slurry velocity and stone volu- 0.05, 0.1, 0.15 and 0.2, respectively, the corresponding stone mass flow
metric concentration. rate increases by an average of 4.54, 10.15, 16.67 and 30.86 kg/s for
The hydraulic gradient I is defined as follows: each additional 1 m/s of inlet slurry velocity.
ΔP
I (m/m) = 4.2. Inclined straight pipe
ρw gL (17)

where ΔP is the pressure drop, ρw is the density of water, and L is the In this work, the pipe inclination angle of 0°, 30°, 45°, 60° and 90° is
length of pipeline. taken to explore its impact on the pressure drop and transport capacity
When the stone volumetric concentration is 0.05, the pressure drop of the slurry pipeline by numerical simulation based on the above fluid-
increases slowly with the increase of inlet slurry velocity. It indicates solid two-phase model.
that the inlet slurry velocity makes a small impact on pressure drop When the stone volumetric concentration is 0.1 and the inlet slurry
when the stone volumetric concentration is low. As the stone volu- velocity is 1 m/s, the distribution of the stone velocity in the z-axis
metric concentration increases, the impact of inlet slurry velocity on direction (the pipe axial direction) under different inclination angles is
pressure drop is increasing: when the stone volumetric concentration is shown in Fig. 13. Unlike horizontal pipelines, the stones distributed at
0.2, the hydraulic gradient increases by an average of 1.32 m/m for the bottom of the pipe originally slip to the pipe inlet gradually under
each additional 1 m/s of inlet slurry velocity, while the corresponding the action of gravity with the increase of the inclination angle. Due to
value is only 0.15 m/m when the stone volumetric concentration is the accumulation of stones at the pipe inlet, the cross-sectional area of
0.05. slurry flow is reduced, and, as a result, the slurry flow at the inlet area is
Based on the statistics of the stone mass flow rate at the outlet, the accelerated of which the maximum velocity is even greater than 4 m/s,

Fig. 7. The slurry velocity distribution on x = 0 profile

under different inlet slurry velocities.

227
D. Yang et al. Tunnelling and Underground Space Technology 71 (2018) 223–230

Fig. 8. The total pressure distribution of slurry

(CV =0.1,v = 3 m/s ).

Fig. 11. The relationship between transport capacity and slurry velocity under different
Fig. 9. The relationship between pressure drop and inlet slurry velocity under different
stone volumetric concentrations.
stone volumetric concentrations.

Fig. 12. The relationship between transport capacity and stone volumetric concentration
under different inlet slurry velocities.
Fig. 10. The relationship between pressure drop and stone volumetric concentration
under different inlet slurry velocities.

228
D. Yang et al. Tunnelling and Underground Space Technology 71 (2018) 223–230

Fig. 15. The relationship between pressure drop and inclination angle under different
inlet slurry velocities.

inclination angle under different inlet slurry velocities. Because the

hydraulic gradient of inclined straight pipe changes in a relatively large
range over time due to the unstable movement of stones when inlet
slurry velocity v ⩽ 2 m/s , only the hydraulic gradient change curves
under inlet slurry velocity of 3, 4 and 5 m/s are shown in Fig. 15. It can
Fig. 13. The distribution of the stone velocity in the z-axis direction under different in- be seen from Fig. 15 that the hydraulic gradient of the inclined pipe
clination angles (CV = 0.1,v = 1 m/s ).
increases with the increase of inlet slurry velocity. When the inclination
angle is smaller than 60°, the hydraulic gradient of the pipeline in-
creases slowly with the increase of the inclination angle: when the
slurry velocity is 3, 4 and 5 m/s, the corresponding hydraulic gradient
only increases by an average of 0.04, 0.04 and 0.18 m/m, respectively,
for each additional 15° of inclination angle. When the inclination angle
is greater than 60°, the hydraulic gradient increases dramatically with
the increase of inclination angle: when the slurry velocity is 3, 4 and
5 m/s, the corresponding hydraulic gradient increases by an average of
0.58, 0.93 and 0.88 m/m for each additional 15° of inclination angle.
This is mainly because the gravity in the radial component of the pi-
peline becomes very small when the inclination angle is greater than
60°. Then the stones leave the pipe wall and are distributed more evenly
under the action of slurry turbulence, which greatly increases the
contact surface between slurry and stones, resulting in a significant
increase in the hydraulic gradient.
Fig. 16 illustrates the relationship between transport capacity and
inclination angle under different inlet slurry velocities. It can be seen

Fig. 14. The distribution of the slurry velocity under different inclination angles
(CV = 0.1,v = 1 m/s ).

as shown in Fig. 14. The stones slipping to the pipeline inlet then
change their movement to the outlet direction at the inlet area under
the action of accelerated slurry flow. However, the closer the stones are
to the pipe outlet, the smaller the stone volumetric concentration is and
the larger the cross-sectional area of the slurry flow is. With the slurry
velocity decreasing, the stones cannot get enough thrust to overcome
the gravity effect, and thus fall down to the pipe inlet again. Therefore,
no stone can be carried out when the inlet slurry velocity is 1 m/s. Fig. 16. The relationship between transport capacity and inclination angle under dif-
Fig. 15 illustrates the relationship between pressure drop and ferent slurry velocities.

229
D. Yang et al. Tunnelling and Underground Space Technology 71 (2018) 223–230

from Fig. 16 that the transport capacity increases with the increase of turbulent pipe flow using Euler-Lagrange large eddy simulation. Int. J. Multiph. Flow
75, 1–11.
slurry velocity, and decreases at first and then increases with the in- Capecelatro, J., Desjardins, O., 2013. Eulerian-Lagrangian modeling of turbulent li-
crease of inclination angle where the minimum value appears at quid–solid slurries in horizontal pipes. Int. J. Multiph. Flow 55, 64–79.
45–60°. This is because, when the inclination angle increases from 0° to Chen, L., Duan, Y., Pu, W., et al., 2009. CFD simulation of coal-water slurry flowing in
horizontal pipelines. Korean J. Chem. Eng. 26 (4), 1144–1154.
60°, on the one hand, the radial component of gravity is sufficient to Cundall, P.A., 1988. Formulation of a three-dimensional distinct element model—Part I.
make the stones distributed over the pipe wall against the action of A scheme to detect and represent contacts in a system composed of many polyhedral
slurry turbulence; on the other hand, the axial component of gravity blocks. Int. J. Rock Mech. Min. Sci. Geomech. Abstr. 25 (3), 107–116.
Cundall, P.A., Strack, O.D.L., 1979. A discrete numerical model for granular assemblies.
prevents the stones from moving forward, thus reducing the transport Géotechnique 29, 47–65.
capacity of the slurry pipeline. When the inclination angle is greater Edelin, D., Czujko, P., Castelain, C., et al., 2015. Experimental determination of the en-
than 60°, stones are distributed more evenly in the pipeline. Although ergy optimum for the transport of floating particles in pipes. Exp. Therm. Fluid Sci.
68, 634–643.
this increases the hydraulic gradient, the transport capacity of the
Eesa, M., Barigou, M., 2009. CFD investigation of the pipe transport of coarse solids in
slurry pipeline is also increased significantly. It is important to note that laminar power law fluids. Chem. Eng. Sci. 64, 322–333.
the minimum inlet slurry velocity must be greater than 2 m/s when the Gopaliya, M.K., Kaushal, D.R., 2016. Modeling of sand-water slurry flow through hor-
pipe inclination angle is between 30° and 60°. Otherwise, the stones izontal pipe using CFD. J. Hydrol. Hydromech. 64 (3), 261–272.
Grewal, H.S., Agrawal, A., Singh, H., 2013. Design and development of high-velocity
cannot be carried out. slurry erosion test rig using CFD. J. Mater. Eng. Perform. 22, 152–161.
Hashemi, S.A., Sadighian, A., Shah, S.I.A., et al., 2014. Solid velocity and concentration
6. Conclusions fluctuations in highly concentrated liquid–solid (slurry) pipe flows. Int. J. Multiph.
Flow 66, 46–61.
Kaushal, D.R., Thinglas, T., Tomita, Y., et al., 2012. CFD modeling for pipeline flow of fine
In this study, with a cross-river tunnel construction of Lanzhou particles at high concentration. Int. J. Multiph. Flow 43, 85–100.
metro as the research background, a three-dimensional fluid-particle Launder, B.E., Spalding, D.B., 1974. The numerical computation of turbulent flows.
Comput. Meth. Appl. Mech. Eng. 3 (2), 269–289.
two-phase flow model is established based on multiphase flow theory Liu, M., Liu, H., Zhang, R., 2015. Numerical analyses of the solid particle erosion in
by considering the rheological properties of the fluid. The corre- elbows for annular flow. Ocean Eng. 105, 186–195.
sponding simulation model is established with CFD and DEM softwares, Matousek, V., 2011. Solids transport formula in predictive model for pipe flow of slurry
above deposit. Part. Sci. Technol. 29, 89–106.
and its effectiveness is validated through the comparison between ex-
Mindlin, R.D., Deresiewicz, H., 1953. Elastic spheres in contact under varying oblique
periment data and simulation predicted. Then based on this model, the forces. J. Appl. Mech. 20 (3), 327–344.
impacts of inlet slurry velocity, stone volumetric concentration and Morsi, S.A., Alexander, A.J., 1972. An investigation of particle trajectories in two-phase
flow systems. J. Fluid Mech. 55 (2), 193–208.
pipe inclination angle on pressure drop and transport capacity of the
Nakayama, S., Uchino, K., Inoue, M., 1996. 3 Dimensional flow measurement at heading
pipeline are investigated. face and application of CFD dimensional flow measurement at heading face and
The results show: (a) for horizontal straight pipes, it is difficult for application of CFD. Shigen-To-Sozai 112 (9), 639–644.
the stones to overcome the gravity and reach a state of complete sus- Pinto, T.C.S., Moraes, J.D., Slatter, P.T., et al., 2014. Modelling the critical velocity for
heterogeneous flow of mineral slurries. Int. J. Multiph. Flow 65, 31–37.
pension motion under the actions of slurry buoyancy and turbulence. Ravelet, F., Bakir, F., Khelladi, S., et al., 2013. Experimental study of hydraulic transport
Therefore, the stones are mainly distributed at the bottom of the pi- of large particles in horizontal pipes. Exp. Therm. Fluid Sci. 45, 187–197.
peline. The pressure drop and transport capacity increase exponentially Tao, Wen-quan, 2001. Numerical Heat Transfer, second ed. Xi’an, Xi’an Jiaotong
University Press, pp. 283–430.
with the increase of inlet slurry velocity and stone volumetric con- Tian, Q., He, G., Wang, H., Cai, D., 2014. Simulation on transportation safety of ice slurry
centration, (b) For inclined straight pipes, the pressure drop of the pi- in ice cooling system of buildings. Energy Build. 72, 262–270.
peline increases slowly with the pipe inclination angle increasing from Torano, J., Torno, S., Menendez, M., et al., 2011. Auxiliary ventilation in mining road-
ways driven with roadheaders: validated CFD modelling of dust behavior. Tunn.
0° to 60°, but increases rapidly when the inclination angle increases Undergr. Space Technol. 26, 201.
from 60° to 90°. The transport capacity reduces slowly at first and then Vlasak, P., Chara, Z., Krupicka, J., et al., 2014a. Experimental investigation of coarse
increases rapidly with the increase of pipe inclination angle, and the particles-water mixture flow in horizontal and inclined pipes. J. Hydrol. Hydromech.
62 (3), 241–247.
minimum value appears at 45–60°. Besides, for vertical pipes, the Vlasak, P., Kysela, B., Chara, Z., 2014b. Fully stratified particle-laden flow in horizontal
minimum inlet slurry velocity must be greater than 1 m/s and greater circular pipe. Part. Sci. Technol. 32, 179–185.
than 2 m/s for inclined pipes with an inclination angle of 30–60°. Wadnerkar, D., Utikar, R.P., Tade, M.O., et al., 2012. CFD simulation of solid–liquid
stirred tanks. Adv. Powder Technol. 23, 445–453.
Wang, X., Zhang, Z., Zhou, Z., Chen, H., 2011. Numerical simulation of TBM construction
Acknowledgments ventilation in a long diversion tunnel. Tunn. Undergr. Space Technol. 26, 560–572.
Wu, D., Yang, B., Liu, Y., 2015a. Transportability and pressure drop of fresh cemented
The work was carried out by School of Mechanical and Electrical coal gangue-fly ash backfill (CGFB) slurry in pipe loop. Powder Technol. 284,
218–224.
Engineering and the department of Mechanical Engineering, Central Wu, D., Yang, B.G., Liu, Y.C., 2015b. Pressure drop in loop pipe flow of fresh cemented
South University, China. This research was supported by the National coal gangue-fly ash slurry: experiment and simulation. Adv. Powder Technol. 26,
High Technology Research and Development Program of China 920–927.
Zhao, H., Zhang, Z., Zhang, T., 2014. Experimental and CFD studies of solid−liquid slurry
(2012AA041803), the Science and Technology Major Project of Hunan tank stirred with an improved Intermig impeller. Trans. Nonferrous Met. Soc. China
Province, China (2014FJ1002), and the Fundamental Research Funds 24, 2650–2659.
For The Central Universities Of Central South University (2016zzts292).

References

Arolla, S.K., Desjardins, O., 2015. Transport modeling of sedimenting particles in a

230