Hindawi Publishing Corporation

Advances in Mechanical Engineering

Volume 2014, Article ID 815945, 10 pages
http://dx.doi.org/10.1155/2014/815945

Research Article
Coupling Analysis of Fluid-Structure Interaction and Flow
Erosion of Gas-Solid Flow in Elbow Pipe

Hongjun Zhu,1,2 Hongnan Zhao,1 Qian Pan,1 and Xue Li1

1
State Key Laboratory of Oil and Gas Reservoir Geology and Exploitation, Southwest Petroleum University, Chengdu,
Sichuan 610500, China
2
State Key Laboratory of Hydraulics and Mountain River Engineering, Sichuan University, Chengdu, Sichuan 610065, China

Correspondence should be addressed to Hongjun Zhu; ticky863@126.com

Received 8 June 2013; Revised 27 November 2013; Accepted 11 December 2013; Published 12 February 2014

Academic Editor: Mohammad Reza Salimpour

Copyright © 2014 Hongjun Zhu et al. This is an open access article distributed under the Creative Commons Attribution License,
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

A numerical simulation has been conducted to investigate flow erosion and pipe deformation of elbow in gas-solid two-phase flow.
The motion of the continuous fluid phase is captured based on calculating three-dimensional Reynolds-averaged-Navier-Stokes
(RANS) equations, while the kinematics and trajectory of the discrete particles are evaluated by discrete phase model (DPM), and
a fluid-structure interaction (FSI) computational model is adopted to calculate the pipe deformation. The effects of inlet velocity,
pipe diameter, and the ratio of curvature and diameter on flow feature, erosion rate, and deformation of elbow are analyzed based
on a series of numerical simulations. The numerical results show that flow field, erosion rate, and deformation of elbow are all
sensitive to the structural changes and inlet condition changes. Higher inlet rate, smaller curvature diameter ratio, or smaller pipe
diameter leads to greater deformation, while slower inlet rate, larger curvature diameter ratio, and larger pipe diameter can weaken
flow erosion.

1. Introduction Figure 1 shows two actual damaged elbows in gas-solid

Erosion wear, resulting in mass loss from the inner wall
of pipe, wall thinning, and even pipe break, is an essential
degradation mechanism for industrial piping [1]. Especially,
high-speed gas-particle flow usually causes very serious
erosion failure of bend pipes in many industrial practices,
such as pneumatic conveying of powders and sand discharge
pipe in gas drilling, sand blasting, and so forth. Bends
are also known to be responsible for dramatic change in
flow field, high pressure loss, and secondary flow, which
inevitably induce vibration and deformation [2]. In addition,
particles continuously impinging on pipe wall can also cause
pipe vibration and deformation. Therefore, flow erosion
and pipe deformation both are present in elbows in gas-
solid flow. Under the combined action of flow erosion and
pipe deformation, elbows are undoubtedly vulnerable and
fail easily, which threaten the piping system reliability and
personal safety.
flow. The left one, 114.30 mm in outer diameter, is a failure
of bend in chemical plant after repairing, which had pre-
sented perforation failure just 3 months after putting into
production. The right one is a failure of bend in gas pipeline
with outer diameter of 88.90 mm. It is clearly seen that gas is
ejecting from the crack in elbow.
Erosion in elbows has been experimentally or numeri-
cally analyzed by different investigators [3–6]. The velocity
of erosive particles, angle of impact, and type of material
are regarded as the three most important factors governing
erosion [7]. Several empirical or semiempirical correlations
have been built to predict the erosion rate of elbows in gas-
solid flow. But they are limited to special working conditions
and cannot be used for all erosion calculations, because that
erosion strongly depends on piping structure, piping layout,
flow localized distribution, and so forth. Pipe deformation
has also been investigated by several researchers [8, 9]. But
they mainly focused on structural strength analysis. Flow
2 Advances in Mechanical Engineering

(a) (b)

Figure 1: Actual damaged elbows: (a) a failure of bend in chemical plant after repairing; (b) a failure of bend in gas pipeline.

induced bend deformation is rarely studied. Therefore, not where

only flow erosion but also flow induced deformation needs 2
further studies. 𝜏𝑓 = [−𝑝 + (−𝑔 − 𝜇) ∇ ⋅ V] 𝐼 + 2𝜇𝐸,
3
Furthermore, few literatures have taken flow erosion and (2)
pipe deformation into account together. There is a mutual 1
𝐸 = (∇V + ∇V𝑇 ) .
influence between the two. On the one hand, high-speed 2
gas-particle flow can induce erosion wear and result in wall
Simultaneously, realizable k-𝜀 turbulence model [11] is
thinning, which brings about local change in the flow channel
employed to close the flow governing equations and describe
and results in change of the flow field. On the other hand,
the turbulent properties:
gas-particle flow can also cause bend deformation, leading
to change of the flow field. Then, the modified flow field will 𝜕 (𝜌𝑓 𝑘) 𝜕 (𝜌𝑓 𝑘V𝑖 ) 𝜕 𝜇 𝜕𝑘
have a new different effect on erosion and pipe deformation. + = [(𝜇 + 𝑡 ) ]
Thus, it is urgently needed to conduct coupling analysis of 𝜕𝑡 𝜕𝑥𝑖 𝜕𝑥𝑗 𝜎𝑘 𝜕𝑥𝑗
fluid-structure interaction and flow erosion of gas-solid flow
in elbow pipe and identify the location and magnitude of the + 𝐺𝑘 + 𝐺𝑏 − 𝜌𝑓 𝜀 − 𝑌𝑀,
maximum erosion and deformation.
Since gas-solid flow feature in elbow as well as erosion 𝜕 (𝜌𝑓 𝜀) 𝜕 (𝜌𝑓 𝜀V𝑖 ) 𝜕 𝜇 𝜕𝜀
+ = [(𝜇 + 𝑡 ) ] + 𝜌𝑓 𝐶1 𝑆𝜀
rate and deformation displacement is difficult to quantify 𝜕𝑡 𝜕𝑥𝑖 𝜕𝑥𝑗 𝜎𝜀 𝜕𝑥𝑗
experimentally, in this work, they are obtained by use of a
three-dimensional fluid-structure interaction computational 𝜀2 𝜀
− 𝜌𝑓 𝐶2 + 𝐶1𝜀 𝐶3𝜀 𝐺𝑏 ,
model coupling with a combined continuum and discrete 𝑘 + √𝜐𝜀 𝑘
model. The flow field distribution of gas-solid flow and the (3)
erosion rate and deformation of elbow are captured under
different inlet conditions or different structural conditions. where
And the effects of inlet velocity, pipe diameter, and the ratio 𝜂
of curvature and diameter are discussed finally. 𝐶1 = max (0.43, ),
𝜂+5
1/2 𝑘
2. Simulation Method 𝜂 = (2𝑆𝑖𝑗 ⋅ 𝑆𝑖𝑗 ) , (4)
𝜀
2.1. Mathematical Model. In this study, gas seemed as
continuous fluid phase evaluated by a Navier-Stokes solver, 1 𝜕V𝑖 𝜕V𝑗
𝑆𝑖𝑗 = ( + ).
and solid particles are treated as spherical particles added 2 𝜕𝑥𝑗 𝜕𝑥𝑖
into continuous phase flow field as discrete phase, which
are captured by discrete phase model (DPM). The motion After continuous phase flow field such as velocity and
of the continuous fluid phase is governed by the three- pressure distribution obtained by solving the above equa-
dimensional Reynolds-averaged-Navier-Stokes (RANS) tions, turbulent properties of discrete phase can be described
equations, including continuity and momentum equations, by a particle motion equation, including the trajectory of
presented as follows [10]: particles, the attack angle, and velocity perturbation [12, 13].
This particle motion equation is called as DPM model and is
𝜕𝜌𝑓 written as
+ ∇ ⋅ (𝜌𝑓 V) = 0,
𝜕𝑡 𝑑V𝑠 𝐶𝐷Re𝑑𝑠 𝑔 (𝜌𝑠 − 𝜌𝑓 ) 𝜌𝑓 𝑑 (V − V𝑠 )
(1) = (V − V𝑠 ) + + 0.5 ,
𝜕𝜌𝑓 V 𝑑𝑡 24𝜏𝑡 𝜌𝑠 𝜌𝑠 𝑑𝑡
+ ∇ ⋅ (𝜌𝑓 VV − 𝜏𝑓 ) = 𝑓𝑓 , (5)
𝜕𝑡
Advances in Mechanical Engineering 3

O
ut
Re

le
ar

t
s tra
ig h
tp
L ip
e se
c ti
on
d
Fixed

R
D Inner wall Bend
y section
x
o
Outer wall Cross section of Longitudinal section
fluid region of pipe
L Y
Fixed Y
t X
Inle Front straight pipe section Z Z X

(a) (b)

Figure 2: Schematic and numerical grid representation of computational domain: (a) overall view of the computational domain and boundary
conditions; (b) grid distribution of fluid and solid computational domain.

where In calculations, SIMPLE algorithm is employed to solve

the pressure-velocity coupling, and second-order upwind
𝜌𝑠 𝑑𝑠2 scheme and second-order central-differencing scheme are
𝜏𝑡 = ,
18𝜇 used for convective terms and diffusion terms, respectively.
󵄨 󵄨 The convergent criteria for all calculations are set such that
𝜌𝑓 𝑑𝑠 󵄨󵄨󵄨V𝑠 − V󵄨󵄨󵄨 the residual in the control volume for each equation is smaller
Re𝑑𝑠 = , (6)
𝜇 than 10−5 .
24 𝑏 𝑏 Re
𝐶𝐷 = (1 + 𝑏1 Re𝑑𝑠2 ) + 3 𝑑𝑠 .
Re𝑑𝑠 𝑏4 + Re𝑑𝑠 2.2. Simulation Conditions. In this work, gas-solid flow in a
90∘ horizontal bend is studied. As shown in Figure 2, the pipe
Then, the erosion rate can be determined by the mass is divided into three sections: front straight pipe section, bend
transfer rate of magnetite on the metal surface, which takes section, and rear straight pipe section. The lengths of front
the form: straight pipe and rear straight pipe are both 5 m. In order
𝑁𝑠𝑑 to analyze the effect of pipe diameter, the inner diameter of
1.8 × 10−9 𝑚𝑠
𝑒= ∑ . (7) pipe is defined as 74.22 mm, 84.84 mm, or 95.00 mm, with
𝑠𝑑=1
𝐴𝑓 outer diameter (D) of 88.90 mm, 101.60 mm, and 114.30 mm,
respectively. The curvature and diameter ratio of elbow (𝑅/𝐷)
The flow induced deformation of bend pipe is governed is set to 3, 4, and 5, respectively, in order to observe the effect
by the following equation [14]: of curvature change.
ICEM CFD mesh-generator and ANSYS mesh-generator
𝑑2 𝑟 𝑑𝑟
𝑀𝑝 + 𝐶𝑝 + 𝐾𝑝 𝑟 + 𝜏𝑝 = 0. (8) are employed to perform the fluid and pipe geometry gener-
𝑑𝑡2 𝑑𝑡 ation and meshing, respectively. Figure 2(b) shows the grid
And the fluid-structure interaction meets the dynamic con- distribution of fluid and solid (pipe) computational domain.
dition on the interface as follows: It is clearly seen that fluid computational domain is divided
into five blocks and progressive mesh is used to capture
𝑛 ⋅ 𝜏𝑓 = 𝑛 ⋅ 𝜏𝑝 . (9) the near-wall flow properties (seen from the cross section).
Each block is discretized with hexahedral cells in order to
Finite volume method (FVM) and finite element method control the grid distribution and computational stability.
(FEM) are used to discretize the fluid and pipe motion Solid computational domain is discretized with uniform
equations, respectively. All the simulations are carried out hexahedral cells, but the grid density in radial direction is less
using a commercial software package ANSYS Workbench than that of fluid computational domain.
14.0, a fluid-structure interaction computing platform, in In order to ensure that the numerical results are indepen-
which FLUENT is used to calculate gas-solid flow field dent of grid size, five different mesh densities are selected to
including DPM being adopted to capture discrete phase based conduct grid resolution tests for elbow with outer diameter of
on Lagrangian approach, and ANSYS mechanical analysis 101.60 mm and 𝑅/𝐷 of 4 with inlet velocity Vin = 20 m/s. The
module is used to calculate the pipe deformation. mesh generation parameters and the maximum calculated
4 Advances in Mechanical Engineering

0.5 0.5 0.5

in = 20 m/s 340 in = 30 m/s 720 in = 40 m/s 1200
332 700 1167
0.4 323 0.4 680 0.4 1133
315 660 1100
0.3 307 0.3 640 0.3 1067

Y (m)
Y (m)
P (Pa)

Y (m)
298 620 1033
290 600 1000
0.2 282 0.2 580 0.2 967
273 560 933
0.1 265 0.1 540 0.1 900
257 520 867
248 500 833
0 240 0 480 0 800
0 0.1 0.2 0.3 0.4 0.5 0 0.1 0.2 0.3 0.4 0.5 0 0.1 0.2 0.3 0.4 0.5
X (m) X (m) X (m)

0.5 22 0.5 0.5 44

34
19 30 39
0.4 0.4 0.4
17 26 33
0.3 14 0.3 22 0.3 28
v (m/s)

Y (m)
Y (m)

Y (m)

12 18 23
0.2 9 0.2 0.2 18
14
7 10 12
0.1 0.1 0.1
4 6 7
0 2 0 2 0 2
0 0.1 0.2 0.3 0.4 0.5 0 0.1 0.2 0.3 0.4 0.5 0 0.1 0.2 0.3 0.4 0.5
X (m) X (m) X (m)
3.7377e − 9 8.2483e − 9 1.448e − 8
3.3224e − 9 7.3319e − 9 1.287e − 8
2.9071e − 9 6.4154e − 9 1.126e − 8
5.4989e − 9 9.653e − 9
de (m)

2.4918e − 9
2.0765e − 9 4.5824e − 9 8.044e − 9
1.6612e − 9 3.6659e − 9 6.435e − 9
1.2459e − 9 2.7494e − 9 4.827e − 9
8.306e − 10 1.833e − 9 3.218e − 9
4.153e − 10 9.1648e − 10 1.609e − 9
0 0 0

3148.8 6693.9 11455

2855.4 6075.1 10402
2561.9 5456.3 9349.2
VMS (Pa)

2268.5 4837.4 8296.2

1975 4218.6 7243.3
1681.6 3599.7 6190.4
1388.1 2980.9 5137.4
1094.6 2362 4084.5
801.19 1743.2 3031.5
507.74 1124.3 1978.6

X X
X
Y Y
Y
Z Z
Z

3.4E
4 − 07 6.0E − 07 8.0E − 07
0
e (kg/(m2 .s))

3.1E − 07 5.5E − 07 7.4E

4 − 07
0
2.9E − 07 5.0E − 07 6.7E
7 − 07
0
2.6E − 07 4.5E − 07 6.1E − 07
0
2.3E − 07 4.0E − 07 5.5E − 07
0
2.0E − 07 3.5E − 07 4.8E − 07
0
1.8E − 07 3.0E − 07 4.2E − 07
0
1.5E − 07 2.5E − 07 3.5E − 07
0
1.2E − 07 2.0E − 07 2.9E − 07
0
9.5E − 08 1.5E − 07 2.3E − 07
0
6.7E
7 − 08 1.0E − 07 1.6E − 07
0
4.0E − 08 5.0E − 08 1.0E − 07
0

Figure 3: Effects of inlet velocity (Vin ) on flow field (pressure and velocity distribution), deformation, von-Mises stress, and erosion in elbow.
Advances in Mechanical Engineering 5

Table 1: Mesh dependence check.

Mesh Elements in fluid domain Elements in solid domain 𝑒max (kg/(m2 ⋅s))
M1 56931 19678 5.21𝐸 − 07
M2 82436 35872 5.56𝐸 − 07 (6.72%)
M3 101780 43260 5.95𝐸 − 07 (6.12%)
M4 115579 51000 6.00𝐸 − 07 (1.69%)
M5 134853 60142 6.01𝐸 − 07 (0.17%)

erosion rate (𝑒max ) are given in Table 1, in which the percent- 4.8
age changes are indicated inside the brackets. The percentage 4.3
difference in 𝑒max between the two grids of M2 and M3 is
3.8

Wall shear stress (Pa)

6.12% and this amount of difference reduces with the increase
in the number of grids as of percentage of difference between
two grids M3 and M4 is just 1.69%. The values undergo a
variation of 0.17% when we pass from M4 to M5 grid systems.
The simulations are executed in the Intel(R) Core(TM) 2 with
specifications of CPU 4 GHz and 2 GB RAM with windows
XP platform. Simulations consumed typically 26 and 45 CPU
hours for M4 and M5 grids, respectively. The CPU time
undergoes an increase of 173.08% while numerical results
undergo a maximum variation of only 0.17%, when we pass
from M4 to M5 grid systems. Therefore, we conclude that the
grid M4 with the minimum size of the grid 0.005 m, shown in
Figure 2, can give a good compromise between precision and
calculation time and is sufficient for carrying out numerical
simulations in this study.
Uniform gas and solid velocity is taken as the inlet
condition in this work. After passing through the front
straight pipe, they will form a fully developed flow. The
same rate of gas and solid in inlet is set to 20 m/s, 30 m/s,
or 40 m/s, in order to observe the effect of inlet condition.
Pressure outlet boundary condition is used for the outlet of
fluid computational domain. And the value is defined as 0 Pa
in order to facilitate comparative analysis. No slip boundary
condition is imposed on the pipe inner wall. Density and
viscosity of gas seemed constant as 1.225 kg/m3 and 1.8 ×
10−5 Pa ⋅ s, respectively. And the density of solid particle is
1123 kg/m3 .
As shown in Figure 2, the elbow is fixed at both ends.
The corresponding outer wall is set to be fixed boundary.
Steel is the pipe material with density, Young’s modulus, and
damping ratio of 7850 kg/m3 , 196 GPa, and 0.05, respectively.
3. Numerical Results and Discussion
3.1. Effect of Inlet Velocity. Gas-solid flow in elbow is affected
by a range of variables related to operational conditions, pipe
structure, and particle properties. Inlet condition, being as
the most important one of operational conditions, is selected
to be analyzed in the present work firstly. This is done by
performing a number of simulations by varying flow inlet
rate while other variables are fixed at their base values (𝐷 =
101.60 mm and 𝑅/𝐷 = 4).
Figure 3 shows the distribution of pressure and gas veloc-
ity in axial section, and the spatial distribution of deformation
(de), von-Mises stress (VMS), and erosion rate (𝑒). It can
3.3
2.8
2.3
1.8
1.3
0.8
0 0.09 0.18 0.27 0.36 0.45
Position (m)
in = 20 m/s in = 40 m/s
in = 30 m/s
Figure 4: Wall shear stress distribution along the outer wall of
elbows with different inlet velocities.
be seen that the radial pressure gradient increases with the
increase of inlet rate. For inlet rate of 40 m/s, the value
of radial pressure gradient is 4714.76 Pa/m, which is four
times as great as that with 20 m/s inlet rate. Gas velocity
distributions in axial section are similar to different inlet
rates, which show a tongue-like shape. High-speed region is
near to the extrados while low speed region is close to the
intrados of elbow.
The maximum pipe deformation locations are observed
at 30–50∘ of the elbow (the inlet of bend is set to 0∘ and the
outlet of bend corresponds to 90∘ ). The higher the inlet rate,
the larger the maximum deformation. The maximum value
is 1.448 × 10−2 𝜇m presenting at the elbow with 40 m/s inlet
rate, which is about 3.87 times than that with 20 m/s inlet rate.
The von-Mises stress of elbow presents the same tendency
as pipe deformation; the maximum value, 11455 Pa, occurs
when the inlet rate is 40 m/s. While the von-Mises stress is
just 3148.8 Pa with the inlet rate 20 m/s, which is about 27.49%
of the former one.
The erosion rates along the elbow wall for different inlet
speeds are also predicted and presented in Figure 3. It is
clearly seen that erosion rate of outer wall is much larger
than that of inner wall, and the maximum erosion rate is
located at 30–45∘ of the outer wall. The larger the inlet rate, the
greater the kinetic energy of particles, resulting in the greater
impinging on elbow wall. The maximum value of erosion rate
increases from 3.4 × 10−7 kg/(m2 ⋅ s) to 8.0 × 10−7 kg/(m2 ⋅ s) as
the inlet flow rate increases from 20 m/s to 40 m/s. Therefore,
6 Advances in Mechanical Engineering

0.4 740 0.5 0.6

717 R/D = 4 720 R/D = 5 700
R/D = 3 693 700 0.5 680
0.4 680
0.3 670 660
647 660
640 0.4 640
623 0.3
Y (m)
P (Pa)

Y (m)
620

Y (m)
0.2 600 620
577 600 0.3 600
553 0.2 580 580
0.1 530 560 0.2 560
507 0.1 540
520 0.1 540
483
0 460 500 520
0 480 0 500
0 0.1 0.2 0.3 0.4
X (m) 0 0.1 0.2 0.3 0.4 0.5 0 0.1 0.2 0.3 0.4 0.5 0.6
X (m) X (m)
0.5 0.6
0.4 32 34 34
28 0.4 30 0.5 30
0.3 24 26 26
0.4
21 0.3
v (m/s)

22
Y (m)

22

Y (m)
0.2 17 0.3
Y (m)

18 18
0.2
13 14 14
0.2
0.1 9 10
0.1 10
6 6 0.1 6
0 2 0 2 0 2
0 0.1 0.2 0.3 0.4 0 0.1 0.2 0.3 0.4 0.5
X (m) 0 0.1 0.2 0.3 0.4 0.5 0.6
X (m) X (m)
5.5359e − 9 8.2483e − 9
4.9208e − 9 7.3319e − 9 1.1234e − 8
4.3057e − 9 9.9856e − 9
de (m)

6.4154e − 9
3.6906e − 9 5.4989e − 9 8.7374e − 9
3.0755e − 9 7.4892e − 9
4.5824e − 9
2.4604e − 9 6.241e − 9
3.6659e − 9
1.8453e − 9 4.9928e − 9
2.7494e − 9
1.2302e − 9 3.7446e − 9
1.833e − 9
6.151e − 10 2.4964e − 9
9.1648e − 10
0 1.2482e − 9
0
0
6087 6693.9
6844.3
5545.3 6075.1
6194
5003.7 5456.3
5543.7
VMS (Pa)

4462.1 4837.4
4893.4
3920.4 4218.6
4243.1
3378.8 3599.7
3592.9
2837.2 2980.9
2942.6
2295.5 2362
2292.3
1753.9 1743.2
1642
1212.3 1124.3
991.7

X
X X
Y
Y Y
Z
Z Z
6.0E − 07
5.5E − 07 6.0E − 07
5.0E − 07 5.5E − 07 6.0E
0 − 07
e (kg/(m2 .s))

4.5E − 07 5.0E − 07 5.5E

5 − 07
4.5E − 07 5.0E
0 − 07
4.0E − 07 4.5E
5 − 07
3.5E − 07 4.0E − 07
3.5E − 07 4.0E
0 − 07
3.0E − 07 3.0E − 07 3.5E
5 − 07
2.5E − 07 2.5E − 07 3.0E
0 − 07
2.0E − 07 2.5E
5 − 07
2.0E − 07 2.0E
0 − 07
1.5E − 07 1.5E − 07 1.5E
5 − 07
1.0E − 07 1.0E − 07 1.0E
0 − 07
5.0E − 08 5.0E − 08 5.0E
0 − 08

Figure 5: Effects of ratio of curvature and diameter (𝑅/𝐷) on flow field (pressure and velocity distribution), deformation, von-Mises stress,
and erosion in elbow.
Advances in Mechanical Engineering
2.5
2.4
2.3
Wall shear stress (Pa)
2.2
2.1
2.0
1.9
1.8
1.7
0 0.1 0.2 0.3
Position (m)
R/D = 3
R/D = 4
Figure 6: Wall shear stress distribution along the outer wall of
elbows with different ratios of curvature and diameter.
the severity of erosion is enhanced as inlet rate is increased,
which is also drawn from the distribution of wall shear stress.
The wall shear stress distributions along the outer wall
under different inlet velocities are exhibited in Figure 4. As
the fluid flows into the elbow, the wall shear stress has a slight
decline but soon increases rapidly. The larger the inlet rate, the
greater the wall shear stress and the faster the increase rate of
wall shear stress. For inlet rate of 40 m/s, the wall shear stress
increases about 1.3 times from the elbow inlet to the outlet.
3.2. Effect of the Ratio of Curvature and Diameter. Pipe
structure is another important parameter affecting the flow
feature in elbow. The details of the erosion and deformation
effects are obtained for different curvature diameter ratios
and different pipe diameters.
Flow field, deformation, von-Mises stress, and erosion
rate distributions for different curvature diameter ratios are
shown in Figure 5. It is seen that flow field is greatly affected
by curvature diameter ratio. Pressure drop is greater in elbow
with smaller ratio, which is 280 Pa for 𝑅/𝐷 of 3, being 1.4
times than that for 𝑅/𝐷 of 5. And the tongue-like-shape
velocity contours for smaller ratio present a greater degree of
curvature and a larger length. The main reason is that flow
channel curves sharply in small ratio elbow, bringing about
more power centrifugal force acting on fluid flow.
The erosion rate is affected by this flow field. The smaller
the ratio, the more serious the flow erosion. We can clearly see
that a bigger maximum erosion rate region presents in outer
wall of elbow with 𝑅/𝐷 of 3, although the maximum value is
the same for different ratios. And as the ratio increases, the
location of the maximum erosion rate region moves to the
upstream. For 𝑅/𝐷 of 3, the location sites are at around 45∘ ,
while it moves to about 35∘ for 𝑅/𝐷 of 5.
However, the deformation of elbow shows an opposite
trend. The larger the ratio, the greater the pipe deformation.
Due to the fact that both ends are fixed, the arc length plays
7
an important role in bend deformation. As shown by von-
Mises stress in Figure 5, the maximum value increases from
6087 Pa to 6844.3 Pa with the increase of curvature diameter
ratio from 3 to 5.
Figure 6 shows the wall shear stress distribution along the
outer wall of elbow with different ratios. There is a decline of
wall shear stress after fluid flow entering the bend. And the
minimum value appears to be about 0.09 m downstream from
the elbow inlet. However, the wall shear stress has a rapid
growth after the trough. The fastest increase rate is 2.85 Pa/m,
appearing in elbow with R/D of 3.
3.3. Effect of Pipe Diameter. Figure 7 shows the effects of
0.4 0.5
pipe diameter on flow field, deformation, von-Mises stress,
and erosion in elbow. The differences of pressure drop and
R/D = 5 velocity contours are not obvious for different pipe diameters.
However, the deformation and erosion rate of elbow change
significantly as diameter changes. The maximum deforma-
tion decreases from 8.7607 × 10−9 m to 7.4619 × 10−9 m
as diameter increases from 88.9 mm to 114.3 mm. And the
maximum von-Mises stress for the three different diameter
elbows is 7074.8 Pa, 6693.9 Pa, and 5929.3 Pa, respectively.
The severity of flow erosion is enhanced as the pipe diam-
eter is decreased. The erosion rate has a sharp decrease as 𝐷
increases from 88.9 mm to 114.3 mm. However, the declining
magnitude of erosion rate from 101.6 mm to 114.3 mm is larger
than that from 88.9 mm to 101.6 mm.
The wall shear stress distribution also indicates that small
diameter pipe has more power flow force acting on it, as
shown in Figure 8. The increase rate of wall shear stress along
outer wall is 2.33 Pa/m for diameter of 88.9 mm, while it is
1.63 Pa/m for diameter of 114.3 mm.
The dimensionless deformation and erosion rate of bends
at different cases are shown in Figures 9 and 10. We can see
that the dimensionless deformation increases obviously with
the increase in Re or 𝑅/𝑑𝑜 , while it reduces with the increase
of ratio of wall thickness and diameter. For 𝑅/𝑑𝑜 , there
are no obvious changes in dimensionless erosion rate. With
the increase in Re, the dimensionless erosion rate increases
firstly and then has no significant change. The dimensionless
erosion rate decreases rapidly with the increase in ratio of wall
thickness and diameter.
4. Conclusions
Numerical simulations based on FSI-CFD and DPM have
been conducted to study the flow induced deformation and
erosion in elbow pipe. Effects of the inlet rate, curvature
diameter ratio, and bend diameter on deformation and
erosion rate are discussed. According to the numerical results
and the analysis above, the following conclusions can be
drawn.
(1) Flow field, erosion rate, and deformation of elbow
are all sensitive to the structural changes and inlet
condition changes. But the influence of inlet rate is
greater than that of structural changes including 𝑅/𝐷
and 𝐷 on elbow deformation and flow erosion.
8 Advances in Mechanical Engineering

0.4 800 0.5 6640

782 0.5 720 D = 114.3mm
D = 88.9 mm D = 101.6mm 700 6622
763 6603
745 0.4 680 0.4 5585
0.3 660
727 5567
708 640 5548
0.3

Y(m)
0.3 620

Y(m)
P (Pa)

690
Y(m)

0.2 5530
672 600 5512
653 0.2 580 0.2 4493
0.1 635 560 4475
617 0.1 540 0.1 4457
598 520 4438
0 580 500 4420
0 480 0
0 0.1 0.2 0.3 0.4 0 0.1 0.2 0.3 0.4 0.5 0 0.1 0.2 0.3 0.4 0.5
X(m) X(m) X(m)

0.5
0.4 34 34 0.5 34
30 0.4 30 30
26 0.4 26
0.3 26
0.3

Y(m)
22 22 22
Y(m)

0.3
 (m/s)

18
Y(m)

0.2 18 18
0.2 0.2
14 14 14
0.1 10 0.1 10 10
0.1
6 6 6
0 2 0 2 0 2
0 0.1 0.2 0.3 0.4 0 0.1 0.2 0.3 0.4 0.5 0 0.1 0.2 0.3 0.4 0.5
X(m) X(m) X(m)
8.7607e − 9 8.2483e − 9 7.4619e − 9
7.7873e − 9 7.3319e − 9 6.6328e − 9
de (m)

6.8139e − 9 6.4154e − 9 5.8037e − 9

5.85139e − 9 5.4989e − 9 4.9746e − 9
5.8405e − 9
4.5824e − 9 4.1455e − 9
4.867e − 9
3.8936e − 9 3.6659e − 9 3.3164e − 9
2.9202e − 9 2.7494e − 9 2.4873e − 9
1.9468e − 9 1.833e − 9 1.6582e − 9
9.7341e − 10 9.1648e − 10 8.291e − 10
0 0 0
7074.8 6693.9 5929.3
6435.7 6075.1 5389.9
5796.6 5456.3 4850.6
VMS (Pa)

5157.5 4837.4 4311.2

4518.4 4218.6 3771.9
3879.3 3599.7 3232.5
3240.2 2980.9 2693.1
2601.1 2362 2153.8
1962 1743.2 1614.4
1322.9 1124.3 1075.1

X X
X
Y Y
Y
Z Z
Z
6.2E − 07
5.7E − 07 6.0E − 07
e (kg/(m2 .s))

5.2E − 07 5.5E − 07 5 −
5.5E 007
4.6E − 07 5.0E − 07 0 −
5.0E 007
4.1E − 07 4.5E − 07 6 −
4.6E 007
3.6E − 07 4.0E − 07 4.1E − 007
3.5E − 07 7 −
3.7E 007
3.1E − 07 2 −
3.2E 007
2.6E − 07 3.0E − 07
2.5E − 07 8 −
2.8E 007
2.1E − 07 3 −
2.3E 007
2.0E − 07 9 −
1.9E 007
1.5E − 07 1.5E − 07
1.0E − 07 4 −
1.4E 007
1.0E − 07 5 −
9.5E 008
5.0E − 08 5.0E − 08 0 −
5.0E 008

Figure 7: Effects of pipe diameter (𝐷) on flow field (pressure and velocity distribution), deformation, von-Mises stress, and erosion in elbow.
Advances in Mechanical Engineering 9

2.6 2.7
Wall shear stress (Pa) 2.5 2.6
2.4
2.5

emax × 10 / 𝜌v
2.3

12
2.4
2.2
2.3
2.1
2.0 2.2
1.9 2.1
1.8 0.8 1.0 1.2 1.4 1.6 1.8 2.0
0 0.1 0.2 0.3 0.4 0.5 Re × 10−5, R/3do ,10(do − di)/do
Position (m)
Re × 10−5 10(do − d i)/do
D = 88.9 mm D = 114.3 mm R /3do
D = 101.6 mm

Figure 8: Wall shear stress distribution along the outer wall of Figure 10: Dimensionless erosion rate of bends at different cases.
elbows with different diameters.

1.8
1.6 𝐶3𝜀 : Empirical constants taken as 0.09
𝐶𝐷: Drag coefficient
1.4
𝐶𝑝 : Damping of pipe
de × 10 /d i
7

1.2 𝑑𝑠 : Diameter of solid particle

1.0 𝑓𝑓 : Volume force of fluid
0.8 𝐺𝑘 : Production term of turbulent kinetic energy
due to the average velocity gradient
0.6
𝐺𝑏 : Production term of turbulent kinetic energy
0.4 due to lift
0.8 1.0 1.2 1.4 1.6 1.8 2.0
g: Gravitational acceleration
Re × 10−5, R/3do ,10(do − di)/do
I: Second-order unit tensor
Re × 10−5 10(do − d i)/do k: Turbulent kinetic energy per unit mass
R /3do 𝐾𝑝 : Stiffness of pipe
𝑀𝑝 : Mass of pipe
Figure 9: Dimensionless deformation of bends at different cases.
𝑚𝑠 : Mass rate of particles
n: Unit vector in the normal direction of the
fluid-structure interface
(2) Higher inlet rate, smaller curvature diameter ratio, or 𝑁𝑠𝑑 : Number of particles
smaller pipe diameter leads to greater deformation, p: Pressure
while slower inlet rate, larger curvature diameter r: Displacement of pipe
ratio, and larger pipe diameter can weaken flow Re𝑑𝑠 : Particle equivalent Reynolds number
erosion. 𝜌𝑓 : Density of gas
𝜌𝑠 : Density of solid particle
(3) The maximum pipe deformation locates at 30–50∘ of v: Velocity of gas
the elbow, while the maximum erosion rate locates V𝑠 : Velocity of solid particle
at 30–45∘ of the outer wall. And the location of the 𝑌𝑀: Impact of compressible turbulence inflation
maximum value region moves to upstream when on the total dissipation rate
curvature diameter ratio increases. 𝜀: Turbulent kinetic energy dissipation rate per
unit mass
𝜏𝑓 : Stress of fluid
Nomenclature
𝜏𝑝 : Stress of structure
𝐴 𝑓: Projected area of particles in the wall 𝜏𝑡 : Particle relaxation time
𝑏1 : Constant taken as 0.186 𝜇: Dynamic viscosity
𝑏2 : Constant taken as 0.653 𝜐: Molecule kinetic viscosity
𝑏3 : Constant taken as 0.437 𝜎𝑘 : Prandtl numbers corresponding to turbulent
𝑏4 : Constant taken as 7178.741 kinetic energy
𝐶1𝜀 : Empirical constants taken as 1.44 𝜎𝜀 : Prandtl numbers corresponding to turbulent
𝐶2 : Empirical constants taken as 1.9 kinetic energy dissipation rate.
10 Advances in Mechanical Engineering

Conflict of Interests separation by hydrodynamic unit operations,” Environmental

The authors declare that there is no conflict of interests
regarding the publication of this paper.
Acknowledgments
Science and Technology, vol. 43, no. 21, pp. 8220–8226, 2009.
[14] W. W. Xu, D. Z. Wu, and L. Q. Wang, “Coupling analysis of fluid-
structure interaction in fluid-filled elbow pipe,” IOP Conference
Series: Earth and Environmental Science, vol. 15, part 6, Article
ID 062001, 2012.

Research work was supported by Open Fund (no. PLN1210)

of State Key Laboratory of Oil and Gas Reservoir Geology
and Exploitation (Southwest Petroleum University) and Key
Project of Sichuan Provincial Education Department (no.
12ZA189). Without the support, this work would not have
been possible.

References
[1] Y. M. Ferng, “Predicting local distributions of erosion-
corrosion wear sites for the piping in the nuclear power plant
using CFD models,” Annals of Nuclear Energy, vol. 35, no. 2, pp.
304–313, 2008.
[2] K. W. Chu and A. B. Yu, “Numerical simulation of the gas-
solid flow in three-dimensional pneumatic conveying bends,”
Industrial and Engineering Chemistry Research, vol. 47, no. 18,
pp. 7058–7071, 2008.
[3] J. Postlethwaite and U. Lotz, “Mass transfer at erosion–corrosion
roughened surfaces,” Canadian Journal of Chemical Engineer-
ing, vol. 66, no. 1, pp. 75–78, 1988.
[4] Y. Tan, H. Zhang, D. Yang, S. Jiang, J. Song, and Y. Sheng,
“Numerical simulation of concrete pumping process and inves-
tigation of wear mechanism of the piping wall,” Tribology
International, vol. 46, pp. 137–144, 2012.
[5] J. R. Fan, K. Luo, X. Y. Zhang, and K. C. Cen, “Large eddy
simulation of the anti-erosion characteristics of the ribbed-
bend in gas-solid flows,” Journal of Engineering for Gas Turbines
and Power, vol. 126, no. 3, pp. 672–679, 2004.
[6] R. Li, A. Yamaguchi, and H. Ninokata, “Computational fluid
dynamics study of liquid droplet impingement erosion in the
inner wall of a bent pipe,” Journal of Power and Energy Systems,
vol. 4, no. 2, pp. 327–336, 2010.
[7] P. Tang, J. Yang, J. Zheng et al., “Failure analysis and prediction
of pipes due to the interaction between multiphase flow and
structure,” Engineering Failure Analysis, vol. 16, no. 5, pp. 1749–
1756, 2009.
[8] Y. Jiang and A. Arabyan, “A new pipe element for mod-
eling three-dimensional large deformation problems,” Finite
Elements in Analysis and Design, vol. 22, no. 1, pp. 59–68, 1996.
[9] M. S. Won, H. I. Ling, and Y. S. Kim, “A study of the deformation
of flexible pipes buried under model reinforced sand,” KSCE
Journal of Civil Engineering, vol. 8, no. 4, pp. 377–385, 2004.
[10] Y. He and K. Li, “Two-level stabilized finite element methods
for the steady Navier-Stokes problem,” Computing, vol. 74, no.
4, pp. 337–351, 2005.
[11] I. Kimura and T. Hosoda, “A non-linear k-𝜀 model with realiz-
ability for prediction of flows around bluff bodies,” International
Journal for Numerical Methods in Fluids, vol. 42, no. 8, pp. 813–
837, 2003.
[12] L. Sun, J.-Z. Lin, F.-L. Wu, and Y.-M. Chen, “Effect of non-
spherical particles on the fluid turbulence in a particulate pipe
flow,” Journal of Hydrodynamics, vol. 16, no. 6, pp. 721–729, 2004.
[13] J. A. Dickenson and J. J. Sansalone, “Discrete phase model
representation of Particulate Matter (PM) for simulating PM