Você está na página 1de 9

E-proceedings of the 36th IAHR World Congress

28 June 3 July, 2015, The Hague, the Netherlands

AN INTEGRATED FINITE ELEMENT METHOD MODEL FOR WAVE-SOIL-PIPELINE INTERACTION


ZAIBIN LIN(1), YAKUN GUO(2), DONG-SHENG JENG(3), NICK REY(4), CHENGCONG LIAO(5)
(1)

School of Engineering, University of Aberdeen, Aberdeen, United Kingdom


e-mail: zaibin.lin@abdn.ac.uk

(2)

School of Engineering, University of Bradford, Bradford, United Kingdom


e-mail:y.guo16@bradford.ac.uk

(3)

Griffith School of Engineering, Griffith University Gold Coast Campus, Queensland, Australia.
e-mail: d.jeng@griffith.edu.au
(4)

Wood Group Kenny, Aberdeen, United Kingdom


e-mail: nick.rey@woodgroupkenny.com

(5)

Department of Civil Engineering, State Key Laboratory of Ocean Engineering, Shanghai Jiao Tong University, China
e-mail: billaday@gmail.com

ABSTRACT
In this study, a monolithically integrated Finite Element Method (FEM) model has been developed for wave-soil-structure
interaction (WSSI). The Reynolds-Averaged Navier-Stokes (RANS) equations are applied to describe wave motion in a fluid
domain, while the seabed domain is described by quasi-state Biot equation. The interface of water and air is tracked by
conservative Level Set method (LSM). The Finite Element Method (FEM) and backward differentiation formula (BDF) are
respectively selected for space discretization and time discretization in this integrated FEM model. In the present model,
only one-way coupling has been considered to integrate fluid model and seabed model. Several available laboratory
experiments are adopted to validate present model. The objective of this paper is to present an integrated FEM model for
the prediction on stability of pipeline subjected to water waves loading and to discuss the dynamic seabed response for
several specific pipeline layouts. The result of validation demonstrates that numerical results have a good agreement with
experimental results and this integrated FEM model can be further extended to practical engineering application. Some
cases on the dynamic seabed response under the pipeline mounted on a flat seabed or inside a trench are investigated for
pipeline stability. The conclusion drawn from this study is that the present WSSI model can accurately capture the
mechanism of wave-induced transient seabed liquefaction around a half or fully buried pipeline.
Keywords: Wave-soil-pipeline interaction, Finite element method, Seabed liquefaction, Integrated model
1.

INTRODUCTION

In recent years, the investigation on wave-induced soil response has been of particular importance in the design of pipeline
installation. The motivation of this increasing study is primarily due to a report on the damage of several marine structures
caused by wave-induced instability (Sumer and Fredse, 2002) instead of construction or material breakdown. Uncoupled
or one-way coupling approaches (Jeng, 2013; Jeng et al., 2013) may not fully capture the real mechanism behind this marine
process. Thus, it is required to develop an integrated model to properly understand and accurately predict wave-induced
soil response by considering the effect of installation of a partially or fully buried pipeline in a porous seabed.

Numerical simulation on wave-seabed-structure interaction has been obtained more attention than ever due to the amazing
development on the computing resource, which allows researchers to simulate large-scale and realistic model and to couple
soil model together with a complicated fluid model. The study on wave-induced seabed response and flow field in the vicinity
of pipeline has been drawn ever-increasing attention concerning the stability and operation of submarine pipeline. Cheng
and Liu (1986) solved the Biot model numerically with a Boundary Integral Equation Method regarding the impact of uplift
seepage force on a buried pipeline. Jeng and Cheng (2000) demonstrated a Finite Difference Method based numerical
model to investigate wave-induced soil response around a buried pipeline. Up to date, Zhou et al. (2013) and Zhou et al.
(2014) analysed wave-induced seabed dynamic response in the neighbourhood of a buried pipeline in a multi-layered
seabed or in an anisotropic seabed, where the wave pressure acting on the seabed is however produced by analytical wave
solution. In consequence, these preceding models coupling with analytical wave may not be able to predict within a certain
1

E-proceedings of the 36th IAHR World Congress


28 June 3 July, 2015, The Hague, the Netherlands

seabed response around pipeline when a wave is propagating over a pipeline which is partially buried or mounted on a
seabed. Thus, it is required to develop a fully integrated WSSI model to predict the stability of submarine pipeline.

For coupling soil model and wave model, Mizutani and Mostafa (1998) and Mostafa et al. (1999) combined Boundary
Element Method (BEM) and FEM to develop an integrated model for the investigation on WSSI. This BEM-FEM model
employs the Poissons equation and the poro-elastic Biot consolidation equations to the irrotational wave field of an
incompressible, inviscous fluid and the porous seabed, respectively. Jeng et al. (2001) examined the wave-induced seabed
response in the proximity of a composite breakwater exposed to a linear wave. Additionally, Jengs research group proposed
an integrated WSSI model based on FVM-FEM model, in which the wave field is described by Reynolds-Averaged NavierStokes (RANS) equations and Volume of Fluid (VOF), and the dynamic Biot equations are adopted for the description on
the mechanical behaviour of a porous seabed (Jeng et al., 2013; Ye and Jeng, 2011).

Most of aforementioned investigations have only examined the wave-induced response in the neighbourhood of pipeline
with analytical wave or FVM-FEM model. Therefore, the objective of this paper is, for the first time, to introduce a
monolithically integrated FEM model for the investigation on the transient stability of pipeline located on or inside an isotropic
seabed in the context of water wave propagating overhead. This integrated FEM model is proposed within the environment
of COMOSL Multiphysics. To examine the applicability of capturing the mechanism of WSSI, the RANS equations are applied
to describe the flow motion, whereas the interface between water and air is traced by conservative LSM (Olsson and Kreiss,
2005; Olsson et al., 2007) . The mechanical response of a porous seabed will be described by the quasi-state (QS) Biot
equations. In this present model, the integration process is one-way coupling, which indicates that only excess wave
pressure is applied on seabed surface to obtain corresponding dynamic seabed response in the vicinity of a partially buried
or fully buried pipeline. Several available experimental results are then used to validate the present model. Subsequently, a
series of parametric investigations on wave-induced seabed response and the corresponding potential liquefaction zone
around pipeline have been performed. At last, the conclusion of the present model will be drawn.
2.

NUMERICAL MODEL

The present integrated FEM WSSI model consists of two component models, which are two phase wave model and QS Biot
model. One-way coupling approach will be adopted in this proposed model, which denotes that only dynamic wave pressure
is acted on the seabed surface. The wave-induced soil response, including the pressure, soil displacement and effective
stress, is described by QS Biot equations.
2.1 Wave model
The two phase incompressible fluid motion can be expressed by the RANS equations, which are denoted as mass
conservation and momentum conservation equations demonstrated in Ignat et al. (2000)
0

(1)
2
3

(2)

in which,
is the averaged velocity field; is the fluid density; is the time;
is the fluid pressure;
is the kinematic
fluid viscosity;
/ is the eddy viscosity as well as known as turbulent viscosity; is the identity matrix; the
coefficient
is 0.09;
is the turbulent kinetic energy;
is the gravity acceleration;
is the source term which can be
demonstrated as either wave generation source term indicated in Choi and Yoon (2009) and Ha et al. (2013) and shown as
Eq.(3) and Eq. (4) or wave damping source function proposed in Wei and Kirby (1995) and shown as Eq. (7) and Eq.(8).
g 2
g

(4)

cos

and

where
and
are the coordinates;
can be illustrated as

(3)

sin

are the source term in x and y direction respectively. The other variables

where

g
1

0.530
cos

(5)
(6)

in which angular frequency , wave number , wave direction , and wave amplitude are required to generate a target
wave train.
/
/4 where 80/ /
, where
is the wavelength and is a parameter that denotes
the width of internal wave generation domain.
2

E-proceedings of the 36th IAHR World Congress


28 June 3 July, 2015, The Hague, the Netherlands

(7)
1
1

(8)

for

is the coordinate of the start location in sponge layer; is the length of sponge layer;
and
are the
where
empirical coefficients determined by numerical test, respectively. The study of Lin and Liu (2004) suggested that can be
taken as 200 and
=10. The width of wave damping region is usually adjusted to double or triple wave length ( ).

The turbulence model adopted in this study is the standard k-epsilon turbulence model (Wilcox, 1998).

(9)
(10)

where
=1.0,

is turbulent kinematic energy;


=1.3,
=1.44,
=1.92.

(11)

is the dissipation rate; the constant coefficients are determined from Wilcox (1998)

The interface of two incompressible fluid is captured by conservative LSM recommended by Olsson and Kreiss (2005) and
Olsson et al. (2007), which is presented as Eqn.(12).

| |

(12)

in which is the level set function;


is the reinitialization parameter and its default value is 1 m/s;
is the parameter
controlling interface thickness. Although Olsson et al. (2007) demonstrated that the conservative LSM conserves ideally the
mass of each fluid, the computational time generally costs more than normal LSM.

2.2 Seabed model


For the soil model, the mechanical behaviour of a hydraulically isotropic porous seabed is demonstrated by the quasi-static
Biot equations (Biot, 1956). Jeng (2013) indicated that this Biot equation is in general acceptable to govern behaviour of the
compressible pore fluid flowing within a compressible porous medium.
(13)
is wave-induced oscillatory pore pressure;
is the permeability in different coordinate directions and in this
where
is the soil porosity;
is the unit of pore water;
and
are respectively the
study we treat it as the same value;
soil displacements in x and y direction; the compressibility of pore fluid
can be illustrated through the relation of the true
bulk modulus of elasticity of water
(generally taken as 2x109N/m2), the absolute water pressure
, and the degree of
soil saturation , which is defined as Eq. (14).
1

(14)

The governing equations applied to describe the force equilibrium in a porous poro-elastic soil can be illustrated as

in which

(15)
(16)

is the shear modulus of soil and is related to Youngs modulus (E) as well as Poissons ratio

The relationship between soil displacement and effective stress in terms of the generalized Hookies law can be denoted as

(17)
3

E-proceedings of the 36th IAHR World Congress


28 June 3 July, 2015, The Hague, the Netherlands

(18)

where
stress.

(19)

and are the effective normal stresses in the horizontal and vertical direction respectively;

is the shear

and soil displacement ( and


), several appropriate boundary conditions are
Prior to obtain the pore pressure
recommended to apply to the seabed domain. At the surface of seabed, where
, the vertical effective normal stress
and shear stress dissipate and correspondingly are set to zero, while the pore pressure is equivalent to the dynamic wave
pressure computed from the wave model.
,
where

is the hydrostatic water pressure,

0,

0 at

(20)

is water depth.

The bottom boundary condition of seabed domain (


, is soil depth) is limited to an impermeable rigid bottom,
which indicates zero displacement and the typical condition where no vertical flow occurs as shown in Eq. (21).
0,

0 at

(21)

For the lateral boundary conditions, both sides of seabed domain are considered to be impermeable as suggested by Zhou
et al. (2013) and shown in Eq.(22) . Based on the study of Ye and Jeng (2012) which proposed that the influence of lateral
boundary conditions vanishes when the simulated soil domain is longer than double wave length, a triple wave length soil
domain is adopted to the soil domain concerning to eliminate the impact of lateral boundary conditions.
0 ,
where

and

0 at

and

(22)

are the horizontal coordinate of the beginning and the end in the soil domain.

The boundary condition of the pipeline is assumed to be an elastic impermeable material and no fluid flows through the
pipeline surface, which can be expressed as
0
where

(23)

is normal to the surface of pipeline.

2.3 WSSI integration process


The coupling process of the present integrated WSSI model can be illustrated by several steps in COMSOL Multiphysics.
When integrating the two sub-models, the segregated solver 1 couples the all aforementioned equations of both sub-models
except the standard k-epsilon turbulence model, which are included in the segregated solver 2, into a huge stiffness matrix
and can be treated as a monolithic solver. The matching mesh on the jointly used boundary between wave model and
seabed model is adopted, which implies that the surface of seabed model shares the same mesh node as that on the bottom
boundary of wave model and no additional interpolation algorithm is required to exchange data for both models.

The present model can be further developed to a two-way coupling WSSI model considering the motion of seabed surface,
while the displacement is limited to a relatively small range and does have a limited influence on the flow field around the
pipeline (Ye et al., 2013). Thus, concerning the computational time and efficiency, the proposed integrated WSSI model in
this study offers a much better choice compared to two-way coupling WSSI model.
3.

VALIDATION

To validate the accuracy of current integrated WSSI model, the proposed model is verified against the available experimental
data in the literatures. The numerical set-up is presented in Figure. 1 in detail. All the characteristic parameters of available
laboratory experiments are tabulated in Table 1.

E-proceedings of the 36th IAHR World Congress


28 June 3 July, 2015, The Hague, the Netherlands

Figure. 1 Sketch of wave-seabed-pipeline interaction

Table 1 Wave characteristics and soil properties for validation case


Experiment
Jacobsen et al. (2012)

(m)

(s)

(m)

0.12

1.63

0.505

0.0524

Turcotte et al. (1984)

0.9

0.533

(N/m2)
0

0.95

0.42

0.95

0.42

0.95

0.33

1.1x10-3

2.3

0.533

6.4x105

0.168

0.42

-3

1.1x10

0.0302

0.167

1.1x10

0.33

6.4x10

0.826

0.533

(m)

0
-3

6.4x10

1.75

e (m)

0.143

(m)

(m/s)

0.33

3.1. Wave model


The wave model without coupling soil model in COMSOL is firstly verified against analytical solution and another wave
generation model using Finite Volume Method (FVM) in OpenFOAM. The target wave is regular progressive waves with
wave height 0.12m, wave period 1.63s, and water depth 0.505m as listed in Table 1. The results indicated in Figure. 2 are
wave amplitude against time history at x=1.0m from the right end of wave generation zone. It can be observed in Figure. 2
that the wave train computed in COMSOL is identical with analytical solution and the traditional wave generation in
OpenFOAM. Consequently, this wave model can be further extended to an integrated WSSI model and is enough to
accurately produce the wave train over a porous seabed.

Figure. 2 Comparisons of wave model in COMSOL and wave model in


OpenFOAM and analytical solution
3.2. Wave-seabed-pipeline interaction
An experimental investigation on wave-seabed-pipeline interaction has been carried out by Turcotte et al. (1984), which
mainly focused on the pipeline buried in a porous seabed. The comparison on wave-seabed-pipeline interaction is presented
in Figure. 3, where P is the excess water pressure imposed on seabed surface. All the details of the experiment are
5

E-proceedings of the 36th IAHR World Congress


28 June 3 July, 2015, The Hague, the Netherlands

tabulated in Table 1. As shown as Figure. 3, numerical results totally have a good agreement with the experimental results
except the case where
ranges from 45to 180when
= 0.9s. This same deviation is also indicated in Cheng and Liu
(1986). The overall validation determines that the present model is able to capture the mechanism of wave-soil-pipeline
interaction.

(a)

= 0.0524m and

= 0.9s

(c)

(b)

= 0.0302m and

= 0.143m and

= 1.75s

= 2.3s

Figure. 3 Comparison of numerical and experimetnal (Turcotte et al., 1984) pore water pressure around pipeline

4.

APPLICATIONS

The advantage of present integrated WSSI model is that it can be applied to evaluate the flow field around pipeline and the
relevant wave-induced dynamic soil response around pipeline in a FEM model, which allows us to investigate the mechanical
behaviour of wave-seabed-pipeline interaction further in various engineering practice. The research on flow field and waveinduced dynamic soil response in the vicinity of the pipeline laid on or inside a flat seabed has been investigated in the
following section. All related parameters for the study of wave-seabed-pipeline interaction are listed in Table 2.
Liquefaction in general occurring around the pipeline may accelerate the scour and drive pipeline to sink into seabed under
the impact of a propagating wave. Sumer et al. (2001) demonstrated that the main influences on producing onset of scour
underneath pipeline are the excessive seepage flow and corresponding piping, and that the liquefaction, which results in
the self-buried pipeline, may lead the soil to lose its carrying capability and to fail to support pipeline. Thus, the investigation
on the mechanism of liquefaction around a pipeline laid on a flat seabed exposed to a progressive wave is presented in this
section. The embedded depth of pipeline are selected to be 0 m, 1/20 of pipeline diameter, half pipeline diameter, and fully
buried as indicated in Figure. 4. The criterion for identifying the liquefaction zone under a 2D progressive wave is suggested
by Zen and Yamazaki (1990) and used in this study, which is depicted as Eq. (24).
| |

(24)

is the unit weight of seabed,


is the unit weight of water,
is the vertical coordinate indicated in Figure. 1,
where
is the wave-induced pore pressure within a seabed,
is the pore pressure on the mudline
is water depth over seabed,
where | |
= 0. It can be noted that when the excess pore pressure exceeds the overburden soil pressure liquefaction
may take place in a porous seabed.

The liquefaction zones of four different embedded depths below pipeline are displayed in Figure. 4 when wave trough is
/ = 0 indicates soil displacement field and blue vector field over
over pipeline, where green vector field below | |
zero demonstrates the flow velocity field. The liquefaction zone varying with the different embedment depth together with
6

E-proceedings of the 36th IAHR World Congress


28 June 3 July, 2015, The Hague, the Netherlands

the detailed vortex structure and soil displacement in the vicinity of pipeline under wave trough are also presented in Figure.
4. It is clearly seen that the transient liquefied depth under pipeline decreases evidently with the growth of embedded depth.
Obviously, the vortex structure induced by wave propagation diminishes when embedment depth advances. The
disappearance of vortex may lower the possibility of onset of scour around pipeline, in which the soil particles can survive
from being moved away massively. It can be noted from Figure. 4 that under certain condition the maximum transient
liquefied depth compared to the bottom of pipeline decreases with the increment of embedded depth and even vanishes
below the bottom of pipeline when pipeline is fully buried.

(a) Pipeline positioned on a flat seabed

(b) Embedment depth up to 1/20 diameter of pipeline

(c) Embedment depth up to half diameter of pipeline

(d) Embedment depth e=0.525D

Figure. 4 Liquefaction zone and wave-induced vortex for different embedment depth under wave trough

Table 2 Parameters for the study of wave-seabed-pipeline interaction


Wave characteristics
Wave height(m)

Wave period (s)

Water depth (m)

20

Wave length (m)

88.73

Seabed thickness(m)

40

Poisson ratio

0.4

Shear modulus (Pa)

1.5x107

Permeability (m/s)

1x10-5

Degree of saturation

0.98

Seabed characteristics

Pipeline characteristics
Diameter (m)

5.

CONCLUSION

A new fully integrated WSSI model for evaluating wave-induced excess pore pressure around pipeline has been developed
in this study. It can be concluded from the validation that the present model agrees well with experimental measurements
and can be further applied to predict the mechanism of wave-seabed-pipeline interaction. According to the preceding
investigations, some following conclusions can be drawn:

The presented WSSI model is capable of predicting wave-seabed-pipeline interaction and the corresponding results
are reliable for engineering application since its good agreement with available experimental results in the literatures.

Under a certain oceanic environment, the enough thickness of backfilled depth can fully prevent the porous seabed
being liquefied and the soil particle being transferred by flow vortex in the vicinity of pipeline. The above study suggests that
the pipeline being fully buried will ensure the soil to maintain its carrying capacity and efficiently support the pipeline to be
away from liquefaction zone.
7

E-proceedings of the 36th IAHR World Congress


28 June 3 July, 2015, The Hague, the Netherlands

ACKNOWLEDGEMENTS
The authors would like to acknowledge the financial support from Energy Technology Partnership (ETP).
REFERENCES
Biot, M.A., 1956. Theory of propagation of elastic waves in a fluidsaturated porous solid. I. Lowfrequency range. J.
Acoust. Soc. Am. 28, 168-178.
Cheng, A.H., Liu, P.L., 1986. Seepage force on a pipeline buried in a poroelastic seabed under wave loadings. Appl. Ocean
Res. 8, 22-32.
Choi, J., Yoon, S.B., 2009. Numerical simulations using momentum source wave-maker applied to RANS equation model.
Coast. Eng. 56, 1043-1060.
Ha, T., Lin, P., Cho, Y.-S., 2013. Generation of 3D regular and irregular waves using Navier-Stokes equations model with
an internal wave maker. Coast. Eng. 76, 55-67.
Ignat, L., Pelletier, D., Ilinca, F., 2000. A universal formulation of two-equation models for adaptive computation of turbulent
flows. Comput. Methods Appl. Mech. Eng. 189, 1119-1139.
Jacobsen, N.G., Fuhrman, D.R., Fredse, J., 2012. A wave generation toolbox for the open source CFD library:
OpenFoam. Int. J. Numer. Methods Fluids. 70, 1073-1088.
Jeng, D.-S., Cha, D., Lin, Y., Hu, P., 2001. Wave-induced pore pressure around a composite breakwater. Ocean Eng. 28,
1413-1435.
Jeng, D.-S., Cheng, L., 2000. Wave-induced seabed instability around a buried pipeline in a poro-elastic seabed. Ocean
Eng. 27, 127-146.
Jeng, D.-S., Ye, J.-H., Zhang, J.-S., Liu, P.L., 2013. An integrated model for the wave-induced seabed response around
marine structures: Model verifications and applications. Coast. Eng. 72, 1-19.
Jeng, D.-S., 2013. Porous Models for Wave-Seabed Interactions. Springer Berlin Heidelberg.
Lin, P., Liu, P.L., 2004. Discussion of Vertical variation of the flow across the surf zone [Coast. Eng. 45 (2002) 169198].
Coast. Eng. 50, 161-164.
Mizutani , N., Mostafa, A.M., 1998. Nonlinear wave-induced seabed instability around coastal structures. Coast. Eng. J. 40,
131-160.
Mostafa, A.M., Mizutani, N., Iwata, K., 1999. Nonlinear wave, composite breakwater, and seabed dynamic interaction.
Journal of Waterway, Port, Coastal, and Ocean Engineering. 125, 88-97.
Olsson, E., Kreiss, G., 2005. A conservative level set method for two phase flow. Journal of computational physics. 210,
225-246.
Olsson, E., Kreiss, G., Zahedi, S., 2007. A conservative level set method for two phase flow II. Journal of Computational
Physics. 225, 785-807.
Sumer, B.M., Fredse, J., 2002. The Mechanics of Scour in the Marine Environment. World Scientific Publishing Co Pte Ltd.
Sumer, B.M., Truelsen, C., Sichmann, T., Fredse, J., 2001. Onset of scour below pipelines and self-burial. Coast. Eng. 42,
313-335.
Turcotte, B.R., Kulhawy, F.H., Liu, P.L., 1984. Laboratory Evaluation of Wave Tank Parameters for Wave-Sediment
Interaction. School of Civil and Environmental Engineering, Cornell University.
8

E-proceedings of the 36th IAHR World Congress


28 June 3 July, 2015, The Hague, the Netherlands

Wei, G., Kirby, J.T., 1995. Time-dependent numerical code for extended Boussinesq equations. Journal of Waterway, Port,
Coastal, and Ocean Engineering. 121, 251-261.
Wilcox, D.C., 1998. Turbulence Modeling for CFD. DCW industries La Canada, CA.
Ye, J.-H., Jeng, D.-S., Wang, R., Zhu, C., 2013. Validation of a 2-D semi-coupled numerical model for fluidstructureseabed
interaction. J. Fluids Struct. 42, 333-357.
Ye, J.-H., Jeng, D.-S., 2011. Effects of bottom shear stresses on the wave-induced dynamic response in a porous seabed:
PORO-WSSI (shear) model. Acta Mech Sin. 27, 898-910.
Ye, J.-H., Jeng, D.-S., 2012. Response of Porous Seabed to Nature Loadings: Waves and Currents. J. Eng. Mech. 138,
601-613.
Zen, K., Yamazaki, H., 1990. Mechanism of wave-induced liquefaction and densification in seabed. Soil Found. 30, 90-104.
Zhou, X., Jeng, D.-S., Yan, Y., Wang, J., 2013. Wave-induced multi-layered seabed response around a buried pipeline.
Ocean Eng. 72, 195-208.
Zhou, X., Wang, J., Zhang, J., Jeng, D.-S., 2014. Wave and current induced seabed response around a submarine pipeline
in an anisotropic seabed. Ocean Eng. 75, 112-127.

Você também pode gostar