Journal of Applied Fluid Mechanics, Vol. 9, No. 2, pp. 711-718, 2016.

Available online at www.jafmonline.net, ISSN 1735-3572, EISSN 1735-3645.

Numerical Simulation of Local Scour around Underwater

Pipeline based on FLUENT Software
Z. Zhang1,2,3 and B. Shi2
1.
Zhejiang Institute of hydraulics and Estuary, Hangzhou,310020,China
2.
College of Engineering, Ocean University of China,Qingdao,266100,China
3.
Zhejiang Provincial Key Laboratory of Estuary and Coast

Corresponding Author Email:82457114zzy@163.com

(Received March 6, 2014; accepted April 15, 2015)

ABSTRACT

This paper proposed a new methodology which was based on computational fluid dynamics for
predicting the scouring process of underwater pipeline. By redeveloping a commercial CFD
computer code, the governing equations for the flow model was solved by finite volume method
and wall shear stress which acted as the key parameter to judge the incipient motion of sediment
was firstly calculated. Then the morphological change of the sandy bed was simulated by dynamic
mesh technology. Based on the comparisons between experimental results and numerical results, it
was confirmed that present numerical modeling method can simulate the flow field and scouring
process around underwater pipeline accurately. Besides, the influence of gap ratios on the scour
behaviors was investigated by the present simulation.

Keywords: Underwater pipeline; Local scour; Numerical simulation; FLUENT; Yser defined
function.

1. INTRODUCTION al.2004) established a 2D numerical model to

Offshore pipeline, which transfers oil or gas from
offshore to onshore, is one of the most important
marine engineering equipment. As for the offshore
pipelines laid directly on sandy seabed, the
complex marine hydrodynamic conditions often
induce the local scour and cause the suspension of
pipeline, which makes the offshore pipeline
unstable and have potential risk of fracture. At
present, many researchers have studied the local
scour around pipelines under currents by physical
model tests and proposed different empirical
equations describing the equilibrium scour depth
of submarine pipeline respectively, such as (Mao
Y.1986), (Sumer and Fredse1990), (Chiew1991),
(Moncada and Aguirre 1999), (Yang et
al.2012).Besides, numerical solution is also
another efficient method to study the local scour
around underwater pipeline. Brrs (1999)
developed a scour model which considered the
suspended sediment transport and bed-load
sediment transport. (Li and Cheng1999, 2001)
proposed a mathematical model to simulate the
scour hole around offshore pipelines. This model
assumed that the shear force at the bed surface
was not larger than the critical shear force when
scour reached to equilibrium state. (Liang et
simulate the flow field and scour around offshore
pipeline using finite difference method. (Lu et
al.2007) simulated local scour around submarine
pipelines based on the finite element method and
unstructured triangle grid system. The equilibrium
scour depth was obtained by this model. However,
the scouring process could not be simulated using
this model. (Zhao and Fernando 2007) employed
an Eulerian two-phase model embedded in the
FLUENT software to simulate the scour around
pipeline. The simulation results certified the
capability of simulating the temporal evolution of
the scour profile and also found that the model
over-predicted the sediment phase velocities under
the water and the sediment interface within the
bed. (Abbas et al.2011) used the Euler-Euler two
phase model to simulate the tunnel erosion
beneath underwater pipeline, however, the
scouring process after the tunnel erosion was not
simulated. (Zhang et al.2013) established a new
numerical model to simulate the seepage flow by
assuming the seabed as porous media and adding
source terms in momentum equations. Using this
model, they discussed the influence of
impermeable plate below pipeline on the seepage
flow under steady flow and gave the critical length
of plate for preventing scour. The above numerical
Z. Zhang and B. Shi /JAFM, Vol. 9, No. 2, pp. 711-718, 2016.
studies show that the sediment transport model is
more suitable for modeling the sediment scour
than the two phase model.
FLUENT software is an advanced commercial CFD
code. It can simulate the flow filed of any case and
shows good reliability. However, there is no
sediment scour model in it and the scour process
can not directly be simulated using this software.
FLUENT software supplies a secondary
development technology which is called user
defined function (UDF). Using this function any
complex models, such as moving items (Belkheir et
al.2012) can be simulated. (Xiong et al. 2014)
redeveloped the software and simulated the scour
around pier, which shows that this software can be
redeveloped to simulate the scour process. In this
study a bed-load sediment transport equation
dependent on relative wall shear force is derived
and the sediment transport model is embedded into
FLUENT. Deformation of the scour hole is
captured using dynamic mesh technology. Based on
this model, the influence of initial gap distance on
flow field and scour behaviors around underwater
pipeline is studied.
2. FLOW MODEL
2.1 Governing Equations
The governing equations for the flow around
underwater pipeline are the continuity equation and
the momentum equation which can be referred by
(Zhang et al. 2013).
2.2Turbulence Model
The k- turbulence model is one of the most
common model used to predict the effects of
turbulence.(Liang et al.2004) and (Xiong et al.
2014) used k- turbulence model to simulate the
scouring process around pipeline and pier
respectively and all got more accurate results
than other turbulence model, such as
Smagorinsky subgrid scale (SGS) model.
Consequently, standard k- turbulence model was
also used in this paper. The transport equations
governing the turbulent kinetic energy k and its
dissipation rate can be described as follows.
k equation
k kui
k q bi q 0
t Gk
t x i x j k x j
equation
ui t 2
C1 Gk C2 (2)
t xi x j x j k k
Where =fluid density; ui=velocity component
along xi direction; xi=axis direction. (i=1,2);=fluid
viscosity and t= turbulent eddy viscosity;C1,
C2,k,k are the constants, and they are
1.44,1.92,1.0,1.03 respectively; G= generation of
turbulence kinetic energy. Besides, the standard
wall function is implemented near the wall.
3. SEDIMENTTRANSPORTMODEL
3.1 Bed Load Transport
In this study the bed load sediment is considered.
The model used here is similar to the model
described by (Xiong et al.2014).With the increase
of flow velocity, the shear force on sediment also
increases and when the force exceeds the critical
incipient shear force of sediment, the removal or
transportation of sediment will occur.
For a flat sandy bed, the critical shear force of
sediment incipient motion can be calculated as
cr 0 g (s 1)d 50cr 0 (3)
where cr0=the critical shear force over flat sandy
bed;g= the gravity acceleration; s= sediment
specific weight; d50=the mean sediment diameter;
cr0=critical shields number (dimensionless shear
force) and can be obtained by the formula.
0.3
cr 0 0.055[1 exp(0.02D*)] (4)
11.2D*
where D*=d50[g(s-1)/2]1/3.
As for the sediment incipient on the slope, the
critical shear force is adjusted to a higher threshold
condition for sediment motion upslope and lower
threshold condition down slope compared with the
flat bed thresholdcr0, using the formula.
s in
c r c r 0 (c o s ) (5)
ta n
where is slope angle of sandy bed, is repose
angle of the sediment.
The sediment transport rate formula described by
Souls by is chosen for this model.
3
q 0 12 g ( s 1) d 5 0 ( c r ) (6)
where q0 is the bed load sediment transport rate per
unit width over a flat bed.
The bed load sediment transport rate at the slope
will be adjusted according to the formula
i h
Cq0 (7)
x i
(1) where iis the shear force at xi direction. The
second term is a slope correction term. Laboratory
measurements indicate that C is in the range of 1.5-
2.3, and C =1.5 is used in the present study. h is the
sandy bed level.
According to the theory of mass balance of
sediment transportation, the change of bed level h
can be expressed as
h 1 q (8)
( bi )
t (1 n ) x i
Where n is porosity of sediment.
712
Z Zhang and B. Shi /JAFM, Vool. 9, No. 2, pp. 7711-718, 2016.
Z.

4. DYNA
AMIC MESH
H TECHNOLG
GY structuured 2D cells.

The sandy bed boundarys

T b motion is decided byy the 6.
6 PHYSCIA
AL EXPERIM
MENT
i
interaction withh the water arouund it. Sandy bbeds
s
shear forces aree obtained fromm the CFD solvvers 6.1 Experiment Settup
s
solution of last time step and used
u to calculatee the
b
boundary nodess new positionns by the sedim ment The experiments
e werre carried out in an annular
t
transport modell. In order to move
m the sandy bed flume which is 25.0 m long, 0.5 m wide w and 0.6 m
b
boundary, the equation
e of mottion is written as a deep (Fig.2).Part
( of fllume was filled with
w sediment,
D
DEFINE_GRID _MOTION U
UDF which is whichh represented thhe sandy bed. The T other side
e
embedded into FLUENT. Affter the boundaarys was a fixed bed, whicch was at the sam me level as the
r
refreshing, the finite
f volume meesh also needs tto be sandy bed. One testted pipe was laaid above the
u
updated. The automatic mesh deformationn is sandy bed to study thhe scour processs and the other
i
implemented byy a Laplacian sm
mooth operator (Liu pipe was
w laid on thee fixed bed to study s the flow
a Marcelo 200
and 08). characcteristics. The ddiameters of thee tested pipes
investigated in experriments were D=0.1 D m. The
5. NUM
MERICAL SO
OLUTION AAND initial gap distance beetween pipe and sandy bed can
be adjjusted by vertical sliding suppoort. The sandy
BOUN
NDARY CON
NDITIONS bed was
w of 0.15m heeight in the fluume. The sand
used here
h is noncohessive with a median size of d50=
As for the flow model, the goveerning equationss are
A 0.3 mm
m and a porosiity n=0.5, the reepose angle of
d
discretized by finite
fi volume meethod. For presssure- the sediment is = 332, th critical Shields
S number
v
velocity couplinng, the SIMPLEE scheme is useed in
for a horizontal bedd is 0.038.A high h accuracy
t model. The sediment transpport model is wrritten
this instrumment ADV (Accoustic Dopplerr velocimetry)
a a user defined function in FLUENT. The bed
as
was selected
s to measure the velocitty across four
c
change equatio
on (Eq.8) is solved
s using finite
f
cross sections
s at varioous gap ratios (F
Fig.3). Besides,
d
difference methood.
each case
c in this expeeriment lasted 150
1 minutes to
T computation
The nal domain is a vertical 2D dommain get thee equilibrium scoour depth.
(
(Fig.1). The velocity inlet boun
ndary is used forr the
u
upstream bounddary. The boundary conditionss are
d
defined as

u z
u ( z ) m in * ln ( ), u 0 (9)
z 0
z
2
k z max C 1/ 2 1 u 2 , 0.0005u 02

(10)
C 3 / 4 k z
3/2
z (11
1)
l z Fig.2. Layou
ut of physical model.
m

z
l z m in ,,C (12)
1 1 .5 z /
The friction veloocity is evaluateed as u*=u0/lnn (
T
/ 0), is the watter depth in thiss model. For thee bed
/z
r
roughness lengthh, z0can be defin
ned as 2.5d50/30.The
v Karman con
von nstant is 0.4.

F
Fig.3. Positions oof velocity crosss sections.

6.2 Experiment Paarameters

In the experiment, the experimennt cases with
differeent gap ratios (G
G/D=0.0, 0.1, 0.3 and 0.5) on
Fig.1. Computationaal domain. both fixed
f bed and scoured bed werre investigated
(Tablee1). The inflow velocity was 0.25m/s and the
Outflow boundaary is applied at the outlet bounddary.
O water depth was keppt as constant of 0.35m.The
T wall bounddary condition ass well as a stanndard
The experiimental parameteers were listed inn Tab.1.
w
wall function is used at the t water-sedimment
i
interface. The water
w surfaces change
c is negleccted.
T computation
The nal domain is discretized
d with nnon-

713
Z. Zhang and B. Shi /JAFM, Vol. 9, No. 2, pp. 711-718, 2016.

Table1 Experimental parameters decreasing from 1.4D to 0.8D.There is a backflow

Pipe below x=0.5Dand the velocity is small. At x=3.5D,
Case Velocity Gap ratio
Bed type diameter the backflow velocity is larger than the velocity at
No. (m/s) (G/D)
(m) x=1.0D while the shear flow region becomes larger.
Scoured bed 1 0.1 0.25 0.0 At x=7.5D cross section which locates at the end of
Scoured bed 2 0.1 0.25 0.1 vortex B, there is nearly no backflow below the
Scoured bed 3 0.1 0.25 0.3 region of z=1D. The flow velocity increases with
Scoured bed 4 0.1 0.25 0.5 the increasing height. With the increasing distance
Fixed bed 5 0.1 0.25 0.0 away from the pipe, the influence of pipe becomes
Fixed bed 6 0.1 0.25 0.1 insignificant and the velocity gradually returns to
Fixed bed 7 0.1 0.25 0.3 the incoming velocity distribution, which has been
Fixed bed 8 0.1 0.25 0.5 verified by the velocity distribution at x=12.5D.

7. RESULTS AND DISCUSSIONS

7.1 Flow Structure

As for the flow around pipe, the flow structure at
G/D=0 is different from other cases.(Mao Y. 1986)
found that there are three vortexes in case of
G/D=0.Present numerical and experimental results
also certified his viewpoint.Fig.4 shows the flow
streamlines around pipe with no gap between pipe (a) G/D=0.1
and wall. It can be seen that there are a larger
clockwise vortex B behind the pipe and two smaller
vortexes A and C both in front of and behind the
pipe near the wall. The vortexes A and C may
induce the sediment suspension and hence cause the
occurrence of scour tunnel. Fig.5 shows the typical
flow structures when G/D=0.1,0.3 and 0.5. Three
cases all present vortex shedding phenomenon
which is not seen in case of G/D=0. However,
because of the walls influence, the interaction
between the underneath shear layer and external (b) G/D=0.3
layer in wake flow region is blocked, which
prevents the increase of vorticity. Consequently, the
vortex below is not fully developed and its size is
smaller than above one in case of G/D=0.1 and 0.3.
In case ofG/D=0.5, the walls influence decreases
and both vortexes are nearly kept as the same size.

(c) G/D=0.5
Fig.5. Flow patterns at various gap ratios.

As for the flow around pipe with a gap, flow

Fig. 4. Flow pattern around the pipeline when
G/D=0.
7.2Velocity Distribution
Due to the different flow structures with or without
gap between pipe and wall, the velocity
distributions are also different.Fig.6 and Fig.7 listed
experimental and numerical results of the horizontal
velocity at four cross sections when gap ratio equals
0.0 and 0.3 respectively and the velocity
distribution at G/D=0.1 and 0.5 are similar with
G/D=0.3, which are not shown here. Seen from
Fig.6,it can conclude that due to the blockage of
pipe, at x=1.0D,thevelocity above the top of pipe is
larger and there is a sharp decrease with the height
separates from the pipe surface at two sides, leading
to the generation of wake flow. An S-shape velocity
profile is found at the cross sections of x/D=1 and
x/D=3.5. Section of x/D=1 is in the lee-wake region
and the negative velocity indicates the presence of
the vortex. No backflow is found at the section of
x/D=3.5, indicating that the vortex disappears at that
distance away from the pipeline. At x/D=7.5,the
flow distribution nearly returns to its initial state.
Fig.6 and Fig.7 also show that the numerical results
are in agreement with experimental results,
especially the velocity near the bed, which lays the
foundation for simulating the scouring process
accurately.

714
Z Zhang and B. Shi /JAFM, Vool. 9, No. 2, pp. 7711-718, 2016.
Z.

7 Scour Characteristics
7.3

(a)x =1.00D (bb)x =3.5D

(c)x =7.5D
D (d))x =12.5D
Fig
g. 6. Velocity diistributions in case
c of G/D=0.00.

(a)x =1.0D
D (bb)x =3.5D

(c)x =7.5D
D (d))x =12.5D
Fig
g. 7. Velocity diistributions in case
c of G/D=0.33.

IIn this study, thee evolution of thhe scour hole in case processs.Fig.8 shows the pressure fields around
o G/D=0.5 wass selected to illustrate the scouuring
of scouriing bed at differrent times. With the change of

715
Z. Zhangg and B. Shi /JA
AFM, Vol. 9, No
o. 2, pp. 711-7188, 2016.

bed level, the pressuree changes correespondingly.
When t==150 min, the prressures along th
he sandy bed
underneath the pipelinee are larger thhan those in
previouss time, and the pressures on thhe bottom of
pipeline also increase more clearly than other
positionss of the pipeline.
The san ndy beds scouuring process can c also be
observedd in Fig.8. From t=0 min to t=5min, the
scouringg process is simiilar with the scouur caused by
horizonttal jet. The scourr hole changes from
fr shallow
to deepp upstream witth the extensioon of dune
downstrream increases. And at this earlly stage, the
scouringg is vigorous. After the initiial scouring
process,, the dune is mov ved towards dow wnstream by
flow, annd the scour hole
h changes slowly. From
t=30minn to 150 min. thhe scour hole haas a smaller
change, and the dunee nearly disapppears. When
t=150 min,
m the profile of o sandy bed is inn agreement
with thee scour hole which
w is shown in Fig.9 in
experim
ment.
The scouring processes at otther gap ratios except
G/D=0.0 are similar
s to the casse of G/D=0.5. In
I case
of no gap, thhe scouring is caused
c by the seeepage
piping which h cannot be ssimulated by present p
numerical moodel at early stage. And aftter the
appearance of flow tunnel betweenb pipelinne and
sandy bed, thee scouring will ggo on due to flow w. The
equilibrium sccour holes in aall cases are shoown in
Fig.10. The scour profiles are similar in n four
different casess in the experimment, as well as for the
two cases (G/D/D=0.3 and 0.5) iin the simulationn using
the same sand d and velocity. TThe agreement beetween
numerical reesults and expperimental resuults at
various gap ratios validatees the mathem matical
models accurracy. It can alsoo be observed thhat the
maximum sco our depth decreasses with the incrreasing
of initial gap ratio.
r

(a) t=0 min

m

Fig.9. Scour hole in experim

ment when t=150 min
(G/D=0.55).

(b) t=1 min

m

(c) t=5 min

m
Fig.10. Equillibrium scour profiles at variou
us gap
ratios.

The equilibriuum scour depthss in different caases in

this study are shown with preevious ones in Fig.11.
F
Note that there are some diffferences betweeen the
present condiitions and prevvious ones. It can c be
observed thatt the experimenntal results are a little
(d) t=300 min deeper than nuumerical results,, and the variatioon rule
of equilibrium
m scour depth veersus initial gap ratio
r is
linear. The eqquilibrium scourr depth at variouus gap
ratios can be obtained
o as folloows:
S S 0 m (G / D ) (13)

WhereS0 is thhe equilibrium sccour depth whenn there

is no gap betw
ween pipeline annd sandy bed, an nd m is
correlation co
oefficient. The correlation coeffficient
(e) t=150 min m is 0.29 by data fitting, inddicating that thee fitted
Fig..8. Pressure (Un
nit: Pa) distribu
utions at curves (Fig.11) agree well with the scatters and
different times. Eq.13 can be used as a semii-empirical form mula to

716
Z Zhang and B. Shi /JAFM, Vool. 9, No. 2, pp. 7711-718, 2016.
Z.

ddescribe the eqquilibrium scou

ur depth at varrious Brrs,, B. (1999).Num
merical Modelingg of Flow and
i
initial gap ratios. Scour
S at Pipeelines.Journal of Hydraulic
Engineering
E 125((5), 511-523.
Chiew
w,Y.M. (1991). P Prediction of maximum
m scour
deepth at submaarine pipelines.. Journal of
Hyydraulic Engineeering 117(4), 4552-466.
Gao, F.P., B.Yang,Y. X..Wuand S.
M.Yan(2006).Ste
M eady current innduced seabed
scour around a vibrating pipeeline. Applied
Ocean
O Research228(5), 291-298.
Hanseen, E.A., J.Fredsse and Y. Maoo (1986).Two-
dimensional
d scour below
pipelines.Proceed
p dings of the Fifth
International
In Symposium o
on Offshore
Fig.11. Equiliibrium scour deepth versus initiial Mechanics
M andd Arctic Enginneering, New
gap ratio. York670-678.
Y
Huai, W.X., Z. W.WWang,Z. D.Qian nand Y.Q.Han
8. CON
NCLUSIONS 2011).Numericaal simulation of
(2 o sandy bed
erosion
e by 22D vertical jet. Science
The clear-water scour below unnderwater pipelinne is
T China:Technolog
C gical Science 54(12),
5 3265-
s
simulated using finite volume method
m and dynaamic 3274.
3
m
mesh. A sedimeent model that only considers bed
l
load sediment is established using user deffined Li, F.and L. Cheng(11999).A numeriical model for
f
function in FLU UENT, in which the Shields num mber loocal scour underr offshore pipeliines.Journal of
i the major parameter. The new
is w sediment moddel is Hydraulic
H Engineeering125(4), 4000-406.
t
then used to sim mulate the bed deformation aroound Li, F. and L. Cheng (2001).Predictioon of lee-wake
u
underwater pipeeline. The main conclusions cann be scouring of piipelines in cu urrents.Journal
s
summarized as follows.
f ofWaterway,Port
of t,Coastal annd Ocean
The comparisonns between the experimental
T e results Engineering
E 127((2), 106-112.
a numerical results
and r show thaat the mathemaatical Liang,, D.F., L.Chengg and F. Li (20004). Numerical
m
model can simullate the process of local scour beelow modelling
m of sscour below a pipeline in
u
underwater pipeline well and can obtain equilibrrium currents.part:Sco
c our simulatio
on. Coastal
p
profileswhicharee similar to the experimental
e resuults. Engineering52(1
E ), 43-62.
On condition of
O o the same saandy bed and flow Liu, X.F.and H. G. Marcelo (2008).Three-
v
velocity, the scoour holes show geometric similarity Dimensional
D Nuumerical Modeel with Free
a various initiaal gap ratios. Th
at he clear-water sscour Water
W Surface and Mesh Deeformation for
e
equilibrium deptth decreases witth the increasingg gap Local
L Sedim
ment Scourr. Journal
r
ratio, and the seemi-experience formula
f is proposed ofWaterway,Port
of t,Coastal annd Ocean
a
according to computational and experimeental Engineering134(
E 4), 203-217.
r
results.
Lu, L., Y. C. Li and J. M. Qin (20007). Numerical
ACKNOWL
LEDGEMENTS simulation of thee equilibrium profile
p of local
scour around suubmarine pipeliines based on
reenormalized grooup turbulence model.Ocean
The authors would
T w like to acknowledge the
Engineering
E 32(117), 2007-2019.
f
financial supporrts from the Natiional Nature Science
F
Fund of China (Grant No. 413376099, 512791189), Mao, Y. (1986).The iinteraction betw
ween a pipeline
t
the Zhejiang Province
P Sciencce and Technoology and
a an erodiible bed. Deenmark: The
C
Conditions of Coonstruction Project(2016C330866). Technology
T Univversity of Denmaark.
Moncaada, M.A.andJJ. Aguirre-Pe (1999).Scour
REFERENCES below
b pipeline in
i river crossinggs. Journal of
Hydraulic
H Engineeering 125(9), 953-958.
9
Abbas,Y.
A B.,H.K.Mohammad,E.S.A Amir,
H.B.Jacoannd L. Cheng (2011).
( EulerEEuler Qi E
E.R., G. Y.L Li,W.Li,J.Wu and a X.Zhang
two-phase flow simulationn of tunnel eroosion (2
2006).Study of vortex characteristics of the
beneath marinepipelines.
m . Applied Occean flow
fl around a horizontal
h circullar cylinder at
Research333(2), 137-146. various
v gap-ratioos in the cross-flo
ow. Journal of
Hydrodynamics,
H Ser. B, 18(3), 3334-340.
Belkheir, N., R.Dizene
B R and S.Khelladi
S (2012).A
Numerical Simulation off Turbulence F Flow Sumerr, B.M. and J. Fredse (1990). Scour below
around a BladeProfile off HAWT Rotoor in pipelines
p iin waves. Journal
Moving Pulse.Journal
P of Applied F Fluid ofWaterway,Port
of t,Coastal annd Ocean
Mechanics55(1),1-9. Engineering
E 116((3), 307-323.

717
Z. Zhang and B. Shi /JAFM, Vol. 9, No. 2, pp. 711-718, 2016.

Xiong, W., C. Cai andB. Kong (2014).CFD Underwater Pipeline Under Steady Current.
Simulations and Analyses for Bridge-Scour Science China: Technolical Sciences
Development Using a Dynamic-Mesh 56(5),1232-1240
Updating Technique. Journal of Computing in
Civil Engineering, 10.1061/(ASCE)CP.1943- Zhao, M.and L. Cheng (2008).Numerical modeling
5487.0000458. of local scour below a piggyback pipeline in
currents. Journal of Hydraulic Engineering
Yang, Z.D. (1996). Sediment transport Theory and 134(10), 1452-1463.
Practice. New York: The McGraw-Hill
Companies. Zhao, Z.H.and H. J. S.Fernando (2007).Numerical
simulation of scour around pipelines using an
Yang,L.P., B.Shi, Y.K.Guo and X.Y.Wen(2012). EulerEuler coupled two-phase model.
Calculation and experiment on scour depth for Environment Fluid Mechanics 7(2), 121-142.
submarine pipeline with a spoiler.Ocean
Engineering55(1),191-198. Zhu, Z.W. and Z. Liu (2012). CFD prediction of
local scour hole around bridge piers. Journal
Zhang, Z.Y., B. Shi,Y. K. Guo andL. P. Yang of Central South University of
(2013). Numerical Investigation On Critical Technology19(1),273-281.
Length of Impermeable Plate Below

718