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

net/publication/258796371

Vortex-induced vibrations of pipes conveying fluid in the subcritical and

supercritical regimes

Article  in  Journal of Fluids and Structures · May 2013

DOI: 10.1016/j.jfluidstructs.2013.02.015

CITATIONS READS

58 452

4 authors, including:

Huliang Dai Lucien Wang

Huazhong University of Science and Technology Shanghai Jiao Tong University
97 PUBLICATIONS   2,104 CITATIONS    381 PUBLICATIONS   4,538 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Ang-2 promotes lung cancer metastasis by increasing epithelial-mesenchymal transition View project

MEMS & NEMS View project

All content following this page was uploaded by Huliang Dai on 28 February 2018.

The user has requested enhancement of the downloaded file.

This article appeared in a journal published by Elsevier. The attached
copy is furnished to the author for internal non-commercial research
and education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elsevier’s archiving and manuscript policies are
encouraged to visit:
http://www.elsevier.com/authorsrights
 Author's personal copy

Journal of Fluids and Structures 39 (2013) 322–334

Contents lists available at SciVerse ScienceDirect

Journal of Fluids and Structures

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

Vortex-induced vibrations of pipes conveying fluid

in the subcritical and supercritical regimes
H.L. Dai a,b, L. Wang a,b,n, Q. Qian a,b, Q. Ni a,b
a
Department of Mechanics, Huazhong University of Science and Technology, Wuhan 430074, China
b
Hubei Key Laboratory for Engineering Structural Analysis and Safety Assessment, Wuhan 430074, China

a r t i c l e in f o abstract

Article history: In this paper, the vortex-induced vibrations of a hinged–hinged pipe conveying fluid are
Received 3 May 2012 examined, by considering the internal fluid velocities ranging from the subcritical to the
Accepted 18 February 2013 supercritical regions. The nonlinear coupled equations of motion are discretized by
Available online 21 March 2013
employing a four-mode Galerkin method. Based on numerical simulations, diagrams of
Keywords: the displacement amplitude versus the external fluid reduced velocity are constructed
Vortex-induced vibration for pipes transporting subcritical and supercritical fluid flows. It is shown that when the
Pipe conveying fluid internal fluid velocity is in the subcritical region, the pipe is always vibrating periodically
Subcritical fluid velocity around the pre-buckling configuration and that with increasing external fluid reduced
Supercritical fluid velocity
velocity the peak amplitude of the pipe increases first and then decreases, with jumping
Chaotic motion
phenomenon between the upper and lower response branches. When the internal fluid
Lock-in
velocity is in the supercritical region, however, the pipe displays various dynamical
behaviors around the post-buckling configuration such as inverse period-doubling
bifurcations, periodic and chaotic motions. Moreover, the bifurcation diagrams for
vibration amplitude of the pipe with varying internal fluid velocities are constructed
for each of the lowest four modes of the pipe in the lock-in conditions. The results show
that there is a significant difference between the vibrations of the pipe around the pre-
buckling configuration and those around the post-buckling configuration.
& 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Vortex-induced vibration (VIV) is a phenomenon that could be observed in many potential areas such as chemical
industry, heat exchanger tubes, bridges, power transmission lines, marine structures such as cables and risers placed
within ocean current, etc. Due to vortex-induced forces, the cylindrical structures could be subjected to lateral vibrations.
Especially when the shedding frequency approaches to a natural frequency of the structure, lock-in effect takes place,
which can easily excite the body into transverse resonance, resulting in relatively large oscillations which contribute to
fatigue failure of the cylindrical structures. It is thus imperative for structural engineers to analyze and simulate the effects
of VIV on the dynamic responses of cylinders/pipes with or without internal fluid flow. Comprehensive reviews on VIV of
cylindrical structures can be found in many publications (see, e.g., Gabbai and Benaroya, 2005; Huang et al., 2009;
Paı̈doussis et al., 2011; Williamson and Govardhan, 2004, 2008; Zdravkovich, 1996).

n
Corresponding author at: Department of Mechanics, Huazhong University of Science and Technology, Wuhan 430074, China. Tel.: þ 86 02787543438.
E-mail address: wanglindds@mail.hust.edu.cn (L. Wang).

0889-9746/$ - see front matter & 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.jfluidstructs.2013.02.015
 Author's personal copy

H.L. Dai et al. / Journal of Fluids and Structures 39 (2013) 322–334 323

In recent decades, much attention has been devoted to the modeling and analysis of rigid cylinders in order to
investigate the fundamental mechanism of VIV phenomena, by both experimental and theoretical methods (Gabbai and
Benaroya, 2005; Wu et al., 2012). Khalak and Williamson (1997) and Williamson and Roshko (1988) reported that the
vortex shedding modes in the wake of the cylinder were 2s, 2p, pþs, 2p þ2s. And three amplitude response branches (the
initial branch, the upper branch and the lower branch) were detected with increasing external fluid reduced velocity in
their experiments. By coupling the structure and wake oscillators, Facchinetti et al. (2004) analyzed the VIV of the coupled
models and developed the most appropriate coupling terms and the values of model parameters. Some of the other
notable contributions in this area were made by Govardhan and Williamson (2004), Pan et al. (2007) and Skop and
Alasubramanian (1997).
However, cylindrical structures in ocean engineering are mostly flexible, especially pipes and risers. Over the past
decades, therefore, increasing attention has also been devoted to the VIV of elastic cylinders. Modal superposition was
firstly adopted by Skop and Griffin (1975) through extending the model of rigid cylinders to elastic cylinders. In another
work by Brika and Laneville (1993), the VIV of a long flexible circular cylinder with low mass-damping ratios was
investigated. It was observed that the hysteresis loop described by the steady-state displacement amplitudes was partially
different from those reported in earlier studies (Khalak and Williamson, 1997) and each branch of the loop was associated
with a vortex shedding mode. The wake oscillator model was employed to simulate the vortex-induced lift force by
Violette et al. (2007) for investigating the responses of flexible cylinders subjected to VIV. Quite good agreement with
experimental results was obtained. Many other reports on VIV of long flexible cylinders have been shown, e.g., by
Evangelinos and Karniadakis (1999), Wang et al. (2003), Yamamoto et al. (2004) and Zhou et al. (1999). However, it is
necessary to clarify that they were mostly restricted to cylinders without internal fluid flow.
VIV of slender elastic pipes containing internal fluid flow is often considered to be one of the important and more
complicated issues in ocean engineering. Indeed, even for a fluid-conveying pipe without external cross flow, the system is
capable of displaying very rich dynamical behavior (see, e.g., Ghayesh et al., 2012; Giacobbi et al., 2012; Guo et al., 2010;
Hellum et al., 2012; Rinaldi and Paı̈doussis, 2010; Sinir, 2013; Wang et al., 2012). Recently, vibration characteristics of
risers with axial internal fluid in external cross flow were theoretically and experimentally investigated by Guo and Lou
(2008). Their results indicated that the strain amplitudes of the pipe in lateral and in-line directions increase with
increasing internal flow speed. Nonlinear free vibrations of marine pipes transporting fluid were explored by
Chucheepsakul et al. (2003) and Kaewunruen et al. (2005), who found that the internal fluid velocity has significant
effect on the vibration behaviors of such systems. The effect of weak structural nonlinearity on the dynamical behavior of
pipes was studied by Keber and Wiercigroch (2008). In their work, the internal fluid flow can highlight the effect of
structural nonlinearity on the dynamics of the pipe. Based on the principle of virtual work, Meng and Chen (2012)
constructed a model formulation for analyzing the VIV of extensible steel catenary risers. The remarkable influence of
internal fluid velocities on nonlinear dynamics of the riser subjected to VIV was also found in their results. Very recently,
based on the forced system model (Paı̈doussis et al., 2011), Dai and Wang (2012) analyzed the steady-state responses of
supported pipes conveying fluid subjected to VIV by using the method of multiple scales.
It should be pointed out that, although some of the above studies have been focused on the effect of internal fluid
velocities on the VIV of pipes, the values of internal fluid velocities were restricted to the subcritical regions, i.e., the
internal fluid velocity being lower than the critical value. It is noted that the subcritical and supercritical regimes are
respectively defined as the values of internal fluid velocities being lower and higher than the critical value (denoted as ncr
in this paper). According to Paı̈doussis (1998), the critical internal fluid velocity means a situation that the pipe begins to
lose stability. Therefore, unstable behaviors of pipes would occur as the internal fluid velocity increases to the supercritical
regime, which may affect the safe operation of pipes. It is then noted that the investigation on VIV of pipes conveying fluid
in the case of internal fluid velocity beyond the critical value is vital. This motivates the current work.
The rest of this paper is organized as follows. The nonlinear coupled equations of motion for VIV of the pipe are first
formulated with consideration of the effects of structural nonlinearity and internal fluid flow in Section 2. The method of
high-order Galerkin discretization is utilized to derive the discrete equations in the cases of internal fluid velocities being
in the subcritical and supercritical regions under lock-in conditions in Section 3. The bifurcation diagrams and phase
portraits which describe responses of the pipe are shown in Section 4. It will be indicated that the nonlinear dynamics of
the pipe with supercritical internal fluid flow is capable of resulting in complex dynamics and chaos. Finally, the paper
gives conclusions in Section 5.

2. The analytical model

The system under consideration consists of a pinned–pinned flexible slender pipe conveying internal axial fluid with
velocity Ui in a uniform external cross-flow fluid flow with velocity Ue, as depicted in Fig. 1.
In many experiments, the cross-flow responses were often observed to be dominant over the in-line motions in the
lock-in conditions (see, e.g., Brika and Laneville, 1993; Khalak and Williamson, 1997; Marcollo and Hinwood, 2006). In this
paper, therefore, we assume that the motions of the pipe are dominant in the transverse direction (y direction). This
implies that the cross-flow responses of the pipe are much more intensive than those of the other two directions (x and z).
In the theoretical component of this study five principal idealizations were introduced: (i) the cross section of the slender
pipe is uniform and the effect of shear, torsion, rotational inertia, and Poisson’s ratio is neglected; (ii) the internal and
 Author's personal copy

324 H.L. Dai et al. / Journal of Fluids and Structures 39 (2013) 322–334

y
L

x D Ui

Ue

Fig. 1. Schematic of a pipe conveying fluid subjected to VIV.

external fluids are thought to be inviscid and irrotational; (iii) the motions of the flexible pipe are in a plane perpendicular
to the external flow; (iv) a four-mode Galerkin discretization of the equations of motion will be employed for analysis;
(v) the post-buckling configuration of the pipe is assumed to have a deformation in the direction of the cross-flow, which is
consistent with the y direction of the vortex-induced force.
If gravity, buoyancy, internal damping, externally imposed tension and pressurization effects are either absent or
neglected, the equation of motion for the fluid-conveying pipe subjected to the vortex-induced force takes the following
form:
" Z   #
@2 yðx,tÞ @4 yðx,tÞ @2 yðx,tÞ 2
2 @ yðx,tÞ EAp L @yðx,tÞ 2 @2 yðx,tÞ
m 2
þ EI 4
þ 2mf U i þ mf U i 2
 dx ¼ f ðx,tÞ, ð1Þ
@t @x @x@t @x 2L 0 @x @x2

where m¼mf þmp þ md with mf, mp and md respectively defined as the mass per unit length of the internal fluid, the pipe
and the additional fluid (the fluid displaced by the structure when in motion). EI is the flexural stiffness of the pipe, Ap is
the cross-section area of the pipe, L is the pipe length, y(x,t) is the lateral displacement of the pipe, and x and t are the axial
coordinate and time, respectively. f(x,t) Represents the effect of vortex-induced force acting on the pipe, which may be
given as (Keber and Wiercigroch, 2008; Vandiver, 1993)
f ðx,tÞ ¼ f D ðx,tÞ þ f L ðx,tÞ, ð2Þ

where fD(x,t) is the hydrodynamic damping force acting in the lateral direction, and fL(x,t) is the lift force. These can be
respectively expressed as
1 @yðx,tÞ
f D ðx,tÞ ¼  C D ro DU e , ð3Þ
2 @t

f L ðx,tÞ ¼ 14C L0 ro DU 2e qðx,tÞ, ð4Þ

where CD and C L0 are the damping and lift coefficients in the transverse direction, which can be respectively taken as 1.2
and 0.3 in the region of well-developed wakes (Facchinetti et al., 2004). In Eq. (4), q(x,t) is defined as the reduced lift
coefficient, which describes the behavior of the near wake and can be modeled by the following van der Pol equation
(Facchinetti et al., 2004):
@2 q @q P @2 y
2
þ los ðq2 1Þ þ os 2 q ¼ , ð5Þ
@t @t D @t 2
where os is the vortex-shedding frequency, and the values found for l and P are 0.3 and 12 respectively (Facchinetti et al.,
2004).
We introduce the following non-dimensional quantities in order to make Eqs. (1) and (5) dimensionless:
Z ¼ y=D, x ¼ x=L, t ¼ ðEI=mÞ0:5 t=L2 ,
Ap D2
v ¼ ðmf =EIÞ0:5 =U i L, b ¼ mf =m, g ¼ ,
2I
rffiffiffiffiffi
ro DU e L2 C L0 ro U 2e L4 m 2
c ¼ CD pffiffiffiffiffiffiffiffiffi , a¼ , Os ¼ os L :
2 mEI 4EI EI

then the nonlinear coupled equations for VIV of the fluid-conveying pipe can be written in dimensionless forms as follows:
" Z   #
@2 Z @Z pffiffiffi @2 Z @4 Z 2
2@ Z
1
@Z 2 @2 Z
þ c þ 2 b v þ þ v  g d x ¼ aq, ð6Þ
@t 2 @t @x@t @x 4
@x
2
0 @x @x
2

@2 q 2 @q 2 @2 Z
þ lOs ðq 1Þ þ O s q ¼ P : ð7Þ
@t2 @t @t2
 Author's personal copy

H.L. Dai et al. / Journal of Fluids and Structures 39 (2013) 322–334 325

3. Transformation of the PDEs into a set of ODEs

3.1. Pre-buckling problem

In this study, the influence of each of the lowest four modes on the VIV of structural vibration behaviors is taken into
account. Therefore, the lock-in dynamics for each of the first four natural frequencies will be explored. On the base of the
above assumptions and idealizations, we simplify the pipe to the model of Euler–Bernoulli beams. Thus, a fourth-order
modal truncation is adopted to recast the partial differential equations (PDEs) (6) and (7) as a set of ordinary differential
equations (ODEs) by choosing a suitable, complete set of functions. We introduce
X
4
Zðx, tÞ ¼ fi ðxÞZi ðtÞ, ð8Þ
i¼1

X
4
qðx, tÞ ¼ fi ðxÞqi ðtÞ, ð9Þ
i¼1
pffiffiffi
where fi ðxÞ ¼ 2 sinðwi xÞ are the eigenfuncitons of a simply supported beam, Zi ðtÞ and qi ðtÞ are the corresponding
generalized coordinates, and wi is the dimensionless eigenvalue of the ith mode, the value of which is wi ¼ip, i ¼1,2,3,4.
Substituting Eqs. (8) and (9) into (6) and (7), multiplying by fi(x) on both sides and integrating from 0 to 1, and
employing the orthogonality of modes, we can obtain eight coupled ordinary differential Eqs. (10)–(17) of Zi and qi
(we substitute Zi and qi for Zi and qi for writing simply)
pffiffiffi
Z€ 1 þ2 bvðB12 Z_ 2 þ B14 Z_ 4 Þ þ a1 Z1 ¼ gðH11 Z1 2 þ H22 Z2 2 þ H33 Z3 2 þ H44 Z4 2 ÞR11 Z1 þ aq1 cZ_ 1 , ð10Þ
pffiffiffi
Z€ 2 þ2 bvðB21 Z_ 1 þ B23 Z_ 3 Þ þ a2 Z2 ¼ gðH11 Z1 2 þ H22 Z2 2 þ H33 Z3 2 þ H44 Z4 2 ÞR22 Z2 þ aq2 cZ_ 2 , ð11Þ
pffiffiffi
Z€ 3 þ2 bvðB32 Z_ 2 þ B34 Z_ 4 Þ þ a3 Z3 ¼ gðH11 Z1 2 þ H22 Z2 2 þ H33 Z3 2 þ H44 Z4 2 ÞR33 Z3 þ aq3 cZ_ 3 , ð12Þ
pffiffiffi
Z€ 4 þ2 bvðB41 Z_ 1 þ B43 Z_ 3 Þ þ a4 Z4 ¼ gðH11 Z1 2 þ H22 Z2 2 þ H33 Z3 2 þ H44 Z4 2 ÞR44 Z4 þ aq4 cZ_ 4 , ð13Þ

q€ 1 þ O2s q1 ¼ lOs q_ 1 lOs ðq3 q4 q_ 2 þ q3 2 q_ 1 q1 q4 q_ 2 1=2q1 2 q_ 3 þ 2q1 q3 q_ 3 þ q4 2 q_ 1

q1 q2 q_ 4 þ q2 q3 q_ 2 q1 q3 q_ 1 þ 1=2q2 2 q_ 3 q2 q4 q_ 1 þ q2 2 q_ 1 þ3=2q1 2 q_ 1 þ q2 q3 q_ 4
þ q2 q4 q_ 3 þ 2q1 q4 q_ 4 þ 2q1 q2 q_ 2 Þ þP Z€ 1 , ð14Þ

q€ 2 þ Os 2 q2 ¼ lOs q_ 2 lOs ð3=2q2 2 q_ 2 þq1 q4 q_ 3 þ q2 q3 q_ 1 þq3 q4 q_ 1 þ q1 q3 q_ 2

1=2q1 2 q_ 4 þ q3 q4 q_ 3 þ 2q2 q4 q_ 4 þq3 2 q_ 2 þ 1=2q3 2 q_ 4 þ q1 q3 q_ 4
þ 2q1 q2 q_ 1 þ2q2 q3 q_ 3 þ q4 2 q_ 2 þq1 2 q_ 2 þ q1 q2 q_ 3 q1 q4 q_ 1 Þ þP Z€ 2 , ð15Þ

q€ 3 þ Os 2 q3 ¼ lOs q_ 3 lOs ðq1 2 q_ 3 þ q2 2 q_ 3 þ 1=2q2 2 q_ 1 1=2q1 2 q_ 1 þ 2q1 q3 q_ 1

þ q2 q4 q_ 3 þ q2 q3 q_ 4 þ q4 2 q_ 3 þ 3=2q3 2 q_ 3 þ q1 q2 q_ 2 þq1 q3 q_ 4 þ q1 q4 q_ 2
þ q2 q4 q_ 1 þ q3 q4 q_ 2 þ 2q2 q3 q_ 2 þ2q3 q4 q_ 4 Þ þ PZ€ 3 , ð16Þ

q€ 4 þ Os 2 q4 ¼ lOs q_ 4 lOs ðq1 q3 q_ 2 q1 q2 q_ 1 þq2 q3 q_ 3 þ q2 q3 q_ 1 þ q1 q2 q_ 3

þ 2q1 q4 q_ 1 þq1 q_ 4 þ2q3 q4 q_ 3 1=2q1 2 q_ 2 þ 1=2q3 2 q_ 2 þ2q2 q4 q_ 2 þ q3 2 q_ 4
þ q2 2 q_ 4 þ 3=2q4 2 q_ 4 Þ þP Z€ 4 , ð17Þ
where a1, a2, a3, a4, H11, H22, H33, H44, R11, R22, R33, R44, B12, B14, B21, B23, B32, B34, B41 and B43 are the coefficients computed
from the integrals of the eigenfunctions fi(x), analytically or numerically. These coefficients are given by
Z 1 Z 1 Z 1 Z 1 Z 1
00 0 0 00 0
an ¼ fn fn iv dx þ v2 fn fn dx, Hnn ¼ fn fn dx, Rnn ¼ fn fn dx, Bnm ¼ fn fm dx, ðn,m ¼ 1,2,3,4Þ
0 0 0 0 0

The first four natural frequencies of the fluid-conveying pipe can be derived from computing the eigenvalues of the
following four Eqs. (18)–(21) by dropping the damping and non-linear terms of Eqs. (10)–(13), and
pffiffiffi
Z€ 1 þ2 bvðB12 Z_ 2 þ B14 Z_ 4 Þ þ a1 Z1 ¼ 0, ð18Þ
pffiffiffi
Z€ 2 þ2 bvðB21 Z_ 1 þ B23 Z_ 3 Þ þ a2 Z2 ¼ 0, ð19Þ
pffiffiffi
Z€ 3 þ2 bvðB32 Z_ 2 þ B34 Z_ 4 Þ þ a3 Z3 ¼ 0, ð20Þ
pffiffiffi
Z€ 4 þ2 bvðB41 Z_ 1 þ B43 Z_ 3 Þ þ a4 Z4 ¼ 0: ð21Þ
 Author's personal copy

326 H.L. Dai et al. / Journal of Fluids and Structures 39 (2013) 322–334

Natural frequency, Ω
Natural frequency, Ω

v v
Fig. 2. Variation of the lowest four natural frequencies with increasing internal fluid velocity for the vibrations around (a) the pre-buckling configuration
and (b) the post-buckling configuration.

The lowest four dimensionless natural frequencies Oi (i¼1,2,3,4) around the pre-buckling configuration are shown in
Fig. 2(a). It should be emphasized here that the variation of the first frequency due to the internal flow velocity (v) is
relatively slow as it is away from the critical value (vcr1 ¼ p), but the drop is much more pronounced close to the critical
flow velocity. Higher natural frequencies have higher values of critical speed. Therefore, it is understandable that in the
range of [0, p] for the first natural frequency, higher modes will exhibit only a moderate change in the natural frequency.
These results show good agreement with those given by Paı̈doussis (1998).

3.2. Post-buckling problem

When the value of internal fluid velocity is higher than the critical one (vcr1 ¼ p), the pipe will display instability
behaviors, such as buckling for a simply supported system, or flutter for a cantilevered system (Paı̈doussis, 1998). For the
sake of convenience in researching the post-buckling problems, the first buckled instability of the pipe is taken into
account in this study. Therefore, the dynamics of the hinged–hinged pipe subjected to VIV under lock-in conditions in the
first buckled mode will be investigated. For that purpose, the equations of motion of the pipe around the first buckled
configuration will be deduced first.
The first buckled configuration is denoted by the buckled equilibrium solution (Nayfeh and Emam, 2008), which is
obtained as follows by dropping the time-dependent, damping and vortex-induced forcing terms in Eq. (6):
Z  
@4 Z @2 Z @2 Z 1 @Z 2
4
þ v2 2 g 2 dx ¼ 0, ð22Þ
@x @x @x 0 @x

and the hinged–hinged boundary conditions are given by

Zðx, tÞ ¼ 0 and Z00 ðx, tÞ ¼ 0 at x ¼ 0,1: ð23Þ
2 2 2 0:5
Based on Eqs. (22) and (23), we can get the first buckled equilibrium solution Z~ ðxÞ ¼ ½ðv p Þ=ðp gÞ f1 ðxÞ.
To investigate the dynamical responses of the pipe under the post-buckling condition, we let
Zðx, tÞ ¼ Z^ ðx, tÞ þ Z~ ðxÞ, ð24Þ

where Z^ ðx, tÞ is a small dynamic disturbance around the first buckled configuration Z~ ðxÞ. Substituting (24) into (6),
we obtain the first buckled equilibrium equation for the simply supported pipe conveying fluid subjected to VIV
(we substitute Z for Z^ in order to write simply). Thus
Z   Z
@2 Z pffiffiffi @2 Z @4 Z 2
2@ Z @2 Z 1 @Z 2 @2 Z~ 1 @Z @Z~
þ 2 b v þ þ v g d xg 2 dx
@t2 @x@t @x4 @x
2 2
@x 0 @x @x 0
2 @x @x
Z   Z   Z
@2 Z 1 @Z~ 2 @2 Z~ 1 @Z 2 @2 Z 1 @Z @Z~ @Z
g 2 dxg 2 dxg 2 2 dx ¼ aqðtÞc : ð25Þ
@x 0 @x @x 0 @ x @x 0 @x @ x @t

The shape functions of the buckled mode can also be obtained by dropping the nonlinear, damping, and vortex-induced
forcing terms from (25). According to the work by Nayfeh and Emam (2008), the post-buckling mode shape functions may
be chosen as the same as the pre-buckling mode functions for the hinged–hinged beam. A fourth-order modal truncation is
also employed to discretize the governing differential Eq. (25). Then we substitute (8) into (25) and utilize the method of
Galerkin discretization to derive the following four ODEs:
pffiffiffi
Z€ 1 þ 2 bvðB12 Z_ 2 þ B14 Z_ 4 Þ þ a1 Z1 ¼ 2gm1 H11 R11 Z1 2 þ 3gm1 2 H11 R11 Z1
 Author's personal copy

H.L. Dai et al. / Journal of Fluids and Structures 39 (2013) 322–334 327

þ gm1 ðH11 Z1 2 þ H22 Z2 2 þ H33 Z3 2 þ H44 Z4 2 ÞR11

þ gðH11 Z1 2 þ H22 Z2 2 þH33 Z3 2 þH44 Z4 2 ÞR11 Z1 þ aq1 cZ_ 1 , ð26Þ
pffiffiffi
Z€ 2 þ2 bvðB21 Z_ 1 þ B23 Z_ 3 Þ þ a2 Z2 ¼ 2gm1 H11 R22 Z1 Z2 þ gm1 2 H11 R22 Z2
þ gðH11 Z1 2 þ H22 Z2 2 þH33 Z3 2 þH44 Z4 2 ÞR22 Z1 þ aq2 cZ_ 2 , ð27Þ
pffiffiffi
Z€ 3 þ2 bvðB32 Z_ 2 þ B34 Z_ 4 Þ þ a3 Z3 ¼ 2gm1 H11 R33 Z1 Z3 þ gm1 2 H11 R33 Z3
þ gðH11 Z1 2 þ H22 Z2 2 þH33 Z3 2 þH44 Z4 2 ÞR33 Z3 þ aq3 cZ_ 3 , ð28Þ
pffiffiffi
Z€ 4 þ2 bvðB41 Z_ 1 þ B43 Z_ 3 Þ þ a4 Z4 ¼ 2gm1 H11 R44 Z1 Z4 þ gm1 2 H11 R44 Z4
þ gðH11 Z1 2 þ H22 Z2 2 þH33 Z3 2 þH44 Z4 2 ÞR44 Z4 þ aq4 cZ_ 4 , ð29Þ

where the various coefficients in Eqs. (26)–(29) are the same as those defined previously and m1 ¼ ½ðv2 p2 Þ=ðp2 gÞ0:5 .
As before, we can obtain the first four natural frequencies under the buckled condition. By dropping the damping and
non-linear terms from Eqs. (26)–(29), the following four equations can be derived:
pffiffiffi
Z€ 1 þ2 bvðB12 Z_ 2 þ B14 Z_ 4 Þ þ ða1 3gm1 2 H11 R11 ÞZ1 ¼ 0, ð30Þ
pffiffiffi
Z€ 2 þ2 bvðB21 Z_ 1 þ B23 Z_ 3 Þ þ ða2 gm1 2 H11 R22 ÞZ2 ¼ 0, ð31Þ
pffiffiffi
Z€ 3 þ2 bvðB32 Z_ 2 þ B34 Z_ 4 Þ þ ða3 gm1 2 H11 R33 ÞZ3 ¼ 0, ð32Þ
pffiffiffi
Z€ 4 þ2 bvðB41 Z_ 1 þ B43 Z_ 3 Þ þ ða4 gm1 2 H11 R44 ÞZ4 ¼ 0: ð33Þ

Then the lowest four dimensionless natural frequencies around the first buckled configuration are determined by
solving the eigenvalues of Eqs. (30)–(33). Fig. 2(b) shows the first four natural frequencies around the first buckled mode.
It can be seen that as the internal fluid velocity is increasing from the critical value, the first frequency O1 increases from
zero with increasing v. However, the other three higher frequencies hardly have changes because of their higher values of
critical fluid velocity. It is also observed from Fig. 2 that, when the fluid velocity v ¼vcr1 ¼ p, the natural frequency of each
mode predicted by the pre-buckling equations is equal to that predicted by the post-buckling equations.

4. Results and analysis

In this section, based on numerical simulations, the dynamical responses of the pipe are displayed for the coupled
system with different system parameters under lock-in and out of lock-in conditions. Typical results of complex dynamics
different from those reported in previous work will be given as follows.
In the following calculations, Young’s modulus E¼210 GPa, the length L¼150 m, the outer diameter D ¼0.25 m, the
inner diameter Di ¼0.125 m, the density of the pipe rp ¼ 7850 kg/m3, the density of the external fluid ro ¼ 1020 kg/m3 and
the density of the internal fluid ri ¼870 kg/m3.
Throughout, solutions of Eqs. (10)–(17) for the pre-buckling problem or those of Eqs. (26)–(29) and (14)–(17) for the
post-buckling problem were obtained by using a fourth-order Runge–Kutta integration algorithm. Unless otherwise stated,
the positive initial conditions for the pipe employed were Zi ð0Þ ¼ qi ð0Þ ¼ Z_ i ð0Þ ¼ q_ i ð0Þ ¼ 0:000025ði ¼ 1,2,3,4Þ.

4.1. Dynamics of the coupled system

4.1.1. Pre-buckling analysis

The steady-state responses of the pipe before buckling are shown in Fig. 3. The responses are described in terms of the
maximum vibration amplitude A/D (multiple of pipe’s diameter D) at x¼ 0.5L with respect to the external fluid reduced
velocity Ur. For a given internal fluid velocity v, the first natural fequency O1 of the pipe can be obtained (see Fig. 2(a)).
Then we can respectively define the external fluid reduced velocity and the shedding frequency as U r ¼ 2pU e =ðO1 DÞ and
Os ¼UrSO1, where S is the Strouhal number which is constant with a value of 0.2 for a large range of Reynolds numbers
(Chen, 1987). Now it is clear that if Ur equals to 5, the shedding frequency Os equals to O1, which is included in the lock-in
conditions for the first mode.
It is observed from Fig. 3 that the amplitude of the periodic motion decreases with increasing internal fluid velocity v
for a given Ur. For a given internal velocity v, however, the steady-state vibration amplitude increases first and then
decreases in the increment of external fluid velocity, with jumping pehomenons (see the arrows in this figure), which is
similar to that obtained by experiments for cylinders without internal fluid flow (Brika and Laneville, 1993). It is possible
that, in the parameter region of 4oUr o7.5, this is the upper response branch, which was refered to those of the previous
works (Brika and Laneville, 1993; Khalak and Williamson, 1997).
 Author's personal copy

328 H.L. Dai et al. / Journal of Fluids and Structures 39 (2013) 322–334

v
v
v

AD

Ur

Fig. 3. Relationship between the peak amplitude and the reduced velocity of external fluid with different internal fluid velocities below the critical value
for the first mode.

v = 3.5 v = 3.7
AD
AD

Ur Ur

v=4 v=5
AD

AD

Ur Ur

Fig. 4. The displacement amplitudes versus external fluid velocities for different internal fluid velocities beyond the critical value.

4.1.2. Post-buckling analysis

When the internal fluid velocity is beyond the first critical value (vcr1 ¼ p), however, the pipe will exhibit a variety of
dynamical behaviors as well as periodic motions. The bifurcation diagrams for amplitude responses of the pipe in terms of the
displacement amplitude A/D (recorded when vibration velocity Z_ ¼ 0 at x¼0.5L) with respect to reduced fluid velocity Ur are
shown in Fig. 4, where the shedding frequency is also defined as Os ¼UrSO1, with O1 being the first natural frequency around the
first buckled configuration, as shown in Fig. 2(b). Moreover, it is noted that Ur ¼5 implies the situation of lock-in.
It can be seen from Fig. 4 that great changes of the pipe motions take place in the increment of Ur for a given v. For
instance, in the case of v¼ 3.5, the first natural frequency O1 of the pipe system is 6.82 and the pipe vibrates periodically
around a non-zero equilibrium as the value of Ur is away from 5. When the external fluid reduced velocity Ur is gradually
close to 5, there is a sudden transition from periodic to chaotic motions of the pipe.
As for the internal fluid velocity v increasing to 3.7 (O1 ¼8.64) and 4 (O1 ¼10.93), the whole trends of the pipe
responses are similar with that of v ¼3.5. But the ranges of chaotic motions become narrower and the phenomenon of
 Author's personal copy

H.L. Dai et al. / Journal of Fluids and Structures 39 (2013) 322–334 329

AD

AD
v v
AD

AD

v v
Fig. 5. Bifurcation diagrams for the lowest four modes under lock-in conditions in the subcritical regime. (a) The dynamic response at x ¼0.5L when the
first mode is locked-in. (b) The dynamic response at x¼ 0.25L when the second mode is locked-in. (c) The dynamic response at x¼ 0.5L when the third
mode is locked-in. (d) The dynamic response at x ¼0.125L when the fourth mode is locked-in.
AD
AD

v v
AD
AD

v v
Fig. 6. Bifurcation diagrams for the lowest four modes under lock-in conditions in the supercritical regime. (a) The dynamic response at x ¼0.5L when the
first mode is locked-in. (b) The dynamic response at x¼ 0.25L when the second mode is locked-in. (c) The dynamic response at x¼ 0.5L when the third
mode is locked-in. (d) The dynamic response at x ¼0.125L when the fourth mode is locked-in.
 Author's personal copy

330 H.L. Dai et al. / Journal of Fluids and Structures 39 (2013) 322–334

periodic-doubling bifurcation is more evident with increasing v. When the internal fluid speed v further increases to 5
(O1 ¼17.07), the range of chaotic motions becomes much narrower and the pipe always undergoes periodic motions.

4.2. Vortex-induced responses in the lock-in condition

In this section, we investigate the vortex-induced vibration of the pipe under lock-in condition in the case of the value
of internal fluid velocity being in the subcritical and supercritical regions respectively.

4.2.1. Pre-buckling dynamics

The bifurcation diagrams for dynamical behaviors of the pipe with internal fluid speed in the subcritical regime for each
of the first four modes under lock-in condition are shown in Fig. 5. The responses of the pipe are displayed in terms of the
maximum vibration amplitude A/D versus the internal fluid velocity v.
It can be seen from Fig. 5 that each of the lowest four modes can be excited independently at its own natural frequency
during lock-in. In the case of lock-in, we equals the vortex shedding frequency Os to the natural frequency of the pipe
system Oi which is derived for a given internal flow velocity v. Then we can obtain the external fluid velocity Ue through
pffiffiffiffiffiffiffiffiffiffiffi
U e ¼ Os D=ð2pS m=EIL2 Þ.
When the internal fluid velocity is in the subcritical range (v o p), the pipe is always vibrating periodically around the
zero equilibrium position. In the lock-in condition for the first mode, the vibration amplitude of the pipe decreases sharply
as the internal fluid velocity approaches to the critical value (vcr1 ¼ p). However, in the lock-in condition for each higher
mode (the second, third or fourth mode), the vibration amplitude appears to change little in the increment of v, since the
natural frequency of each higher mode is slightly affected by increasing v (see Fig. 2(a)).

1.5
AD

0


-1.5

-3
0 0.3 0.6
v 

1.5
AD

0


-1.5

-3
-0.6 -0.3 0
v 
Fig. 7. Expanded versions of Fig. 6(a) for a smaller range of v and the phase portraits for v¼ 4.08: (a) for positive initial conditions and (b) for negative
initial conditions.
 Author's personal copy

H.L. Dai et al. / Journal of Fluids and Structures 39 (2013) 322–334 331

4.2.2. Post-buckling dynamics

The dynamical responses of the pipe under lock-in condition for the lowest four modes around the first buckled
configuration are explored in this subsection. Fig. 6 shows the bifurcation diagrams for the dynamical behaviors of the pipe
in the lock-in condition with the internal fluid velocity being in the supercritical range (v4 p). The displacement
amplitude A/D of the pipe is also recorded as the vibration velocity Z_ ¼ 0 at x ¼0.5. It is obviously seen in Fig. 6 that the
dynamics of the pipe around the post-buckling configuration under lock-in condition are different from that around the

0.8
0.1 v = 3.15 v = 3.2

0.4

0 0

-0.4
-0.1
-0.8
-0.1 0 0.1 -0.2 0 0.2

v = 3.5 v = 3.57
1 1

0 0


-1 -1

-0.5 -0.25 0 0.25 0.5 -0.5 0 0.5

v = 3.7 v = 3.798
1
1

0 0


-1
-1

0 0.25 0.5 0.1 0.2 0.3 0.4 0.5

v = 3.82 v=4

1 1.5


0 0

-1 -1.5

0.1 0.2 0.3 0.4 0.5 0 0.35 0.7

Fig. 8. Typical phase portraits for different internal fluid velocities when the first mode is locked-in.
 Author's personal copy

332 H.L. Dai et al. / Journal of Fluids and Structures 39 (2013) 322–334

pre-buckling configuration given in Fig. 5, especially for the pipe system in the first lock-in mode. From Fig. 6(a), it is noted that at
the beginning of increasing internal fluid velocity, the pipe oscillates periodically and the vibration amplitude increases with
increasing v. As the value of internal flow speed v increases to a larger region, the pipe is shown to be in the chaotic regime.
The more detailed version of Fig. 6(a) for the range 3.68 ovo4.38 can be found in Fig. 7(a). In this figure, a inverse
period-doubling bifurcation springs up, and the tip amplitude increases in the increment of v. Typical phase portraits for
different internal fluid velocities in the lock-in condition for the first buckled mode are plotted in Figs. 8 and 9.
In Fig. 7(a), it was also found that the symmetry of the solutions was broken for certain values of v. Hence, simulations were
also carried out with negative initial conditions were Zi ð0Þ ¼ qi ð0Þ ¼ Z_ i ð0Þ ¼ q_ i ð0Þ ¼ 0:000025 ði ¼ 1,2,3,4Þ, leading to a
different solution branch, as can be seen in Fig. 7(b). Therefore, there appear to be two competing local basins of attraction, which
are related to two distinct solution branches within the range v¼ 3.68–4.38. The outcome of the dynamical behavior of the pipe
will be around one of the non-zero equilibrium positions and is dependent on the initial conditions. It is also found that the pipe
would dispaly three kinds of chaotic motions, which are further represented by means of phase portraits in Figs. 8 and 9. For
v¼3.5 we see a wider-band chaotic motion around the two non-zero equilibrium positions. For v¼3.7 or 3.75, it is obvious that
only narrow-band chaotic motions around either of these two non-zero equilibrium positions could occur.

v = 3.75 v = 4.17

1 1.5

0 0

-1 -1.5

-0.54 -0.27 0 -0.7 -0.35 0

Fig. 9. Typical phase portraits for the pipe vibrating around the negative equilibrium position.

Ωs = Ω1 Ωs = Ω2
3 10

0 0

-3 -10

0.1 0.55 1 -0.4 0.2 0.8



Ωs = Ω3 Ωs = Ω4

20 40

0 0

-20 -40

-0.5 0 0.5 -0.5 0 0.5

Fig. 10. Phase portraits for each of the lowest four modes in the lock-in conditions and v¼5.
 Author's personal copy

H.L. Dai et al. / Journal of Fluids and Structures 39 (2013) 322–334 333

Compared with the various dynamics of the pipe in the first lock-in mode, however, the pipe only displays periodic
motion in the lock-in condition for each higher mode, which could be observed in Fig. 6(b–d). It is also found that the
vibration amplitudes are slightly increased with increasing v, which is different from the trend shown in Fig. 5 for the pipe
vibrating around the pre-buckling configuration. Moreover, at the critical fluid velocity (vcr1 ¼ p), for a certain mode in the
lock-in condition, the vibration amplitude predicted by the pre-buckling equations is equal to that predicted by the post-
buckling equations, which can be clearly seen by comparing Figs. 5 and 6.
Fig. 10 shows the phase portraits for the vibrations of the pipe at v ¼5 for the lowest four modes during lock-in. It is
noted that the higher mode in the lock-in condition means higher oscillation velocity of the pipe.

5. Conclusions

In this study, the dynamical behaviors of fluid-conveying pipes subjected to vortex-induced vibration are analyzed. The
effects of both subcritical and supercritical internal fluid flows on the nonlinear dynamics of the pipe have been
investigated. It is found that the internal fluid velocity has a strong effect on the nonlinear dynamics of the pipe, especially
for a pipe system conveying supercritical fluid flow. These results open the door for further investigations into the rich and
complicated dynamics of fluid-conveying pipes subjected to vortex-induced vibrations. From the results obtained, the
following conclusions can be drawn:

(i) When the internal fluid velocity is in the subcritical region, the pipe displays periodic motion and the vibration
amplitude is gradually decreasing with the increase of internal fluid velocity in the lock-in regions.
(ii) When the internal fluid velocity is in the supercritical region, the pipe under lock-in condition displays a variety of
dynamical behaviors, such as inverse period-doubling bifurcations, periodic and chaotic motions.
(iii) When the vortex shedding frequency is far away from the natural frequency of the pipe system, there is little effect of
the internal fluid velocity on the responses. Once the shedding frequency is close to the natural frequency, however,
the internal fluid flow has a prominent effect on the dynamics of the pipe.

Further work requires comprehensive experimental data for better understanding the dynamical behavior of the
system. In the in-flow direction, of course, there may be an external-flow-induced force exerted on the pipe which must be
included in some cases. For such a more complex problem, an analysis of three-dimensional (3-D) responses of the pipe is
inevitable. These and many more are some of the aspects of the problem should be considered in the further study.

Acknowledgments

The authors gratefully acknowledge the support provided by the National Natural Science Foundation of China
(Nos. 11172107 and 11172109) and the Program for New Century Excellent Talents in University (NCET-11-0183).

References

Brika, D., Laneville, A., 1993. Vortex-induced vibrations of a long flexible circular cylinder. Journal of Fluid Mechanics 250, 481–508.
Chen, S.S., 1987. Flow-induced Vibration of Circular Cylindrical Structures. Hemisphere Publishing Corp., Washington, DC.
Chucheepsakul, S., Monprapussorn, T., Huang, T., 2003. Large strain formulations of extensible flexible marine pipes transporting fluid. Journal of Fluids
and Structures 17, 185–224.
Dai, H.L., Wang, L., 2012. Vortex-induced vibration of pipes conveying fluid using the method of multiple scales. Theoretical and Applied Mechanics
Letters 2, 022006.
Evangelinos, C., Karniadakis, G.E., 1999. Dynamics and flow structures in the turbulent wake of rigid and flexible cylinders subject to vortex-induced
vibrations. Journal of Fluid Mechanics 400, 91–124.
Facchinetti, M.L., Langre, E.de., Biolley, F., 2004. Coupling of structure and wake oscillators in vortex-induced vibrations. Journal of Fluids and Structures
19, 123–140.
Giacobbi, D.B., Rinaldi, S., Semler, C., Paı̈doussis, M.P., 2012. The dynamics of a cantilevered pipe aspirating fluid studied by experimental, numerical and
analytical methods. Journal of Fluids and Structures 30, 73–96.
Gabbai, R.D., Benaroya, H., 2005. An overview of modeling and experiments of vortex-induced vibration of circular cylinders. Journal of Sound and
Vibration 282, 575–616.
Ghayesh, M.H., Amabili, M., Paı̈doussis, M.P., 2012. Thermo-mechanical phase-shift determination in Coriolis mass-flowmeters with added masses.
Journal of Fluids and Structures 34, 1–13.
Govardhan, R., Williamson, C.H.K., 2004. Critical mass in vortex-induced vibration of a cylinder. (European) Journal of Mechanics B/Fluids 23, 17–27.
Guo, H.Y., Lou, M., 2008. Effect of internal flow on vortex-induced vibration of risers. Journal of Fluids and Structures 14, 496–504.
Guo, C.Q., Zhang, C.H., Paı̈doussis, M.P., 2010. Modification of equation of motion of fluid-conveying pipe for laminar and turbulent flow profiles. Journal
of Fluids and Structures 26, 793–803.
Hellum, A., Mukherjee, R., Hull, A.J., 2012. Flutter instability of a fluid-conveying fluid-immersed pipe affixed to a rigid body. Journal of Fluids and
Structures 27, 1086–1096.
Huang, X.D., Zhang, H., Wang, X.S., 2009. An overview on the study of vortex-induced vibration of marine riser. Journal of Marine Sciences 27, 95–101
(in Chinese).
Kaewunruen, S., Chiravatchradej, J., Chucheepsakul, S., 2005. Nonlinear free vibrations of marine risers/pipes transporting fluid. Ocean Engineering 32,
417–440.
Keber, M., Wiercigroch, M., 2008. Dynamics of a vertical riser with weak structural nonlinearity excited by wakes. Journal of Sound and Vibration 315,
685–699.
 Author's personal copy

334 H.L. Dai et al. / Journal of Fluids and Structures 39 (2013) 322–334

Khalak, A., Williamson, C.H.K., 1997. Dynamics of a hydroelastic cylinder with very low mass and damping. Journal of Fluids and Structures 10, 455–472.
Marcollo, H., Hinwood, J.B., 2006. On shear flow single mode lock-in with both cross-flow and in-line lock-in mechanisms. Journal of Fluids and Structures
22, 197–211.
Meng, D., Chen, L., 2012. Nonlinear free vibrations and vortex-induced vibrations of fluid-conveying steel catenary riser. Applied Ocean Research 34,
52–67.
Nayfeh, A.H., Emam, S.A., 2008. Exact solution and stability of postbuckling configurations of beams. Nonlinear Dynamics 54, 395–408.
Paı̈doussis, M.P., 1998. Fluid–Structure Interactions: Slender Structures and Axial Flow, vol. 1. Academic Press Limited, London.
Paı̈doussis, M.P., Price, S.J., Langre, E.de., 2011. Fluid Structure Interactions, Cross-Flow-Induced Instabilities. Cambridge University Press.
Pan, Z.Y., Cui, W.C., Miao, Q.M., 2007. Numerical simulation of vortex-induced vibration of a circular cylinder at low mass-damping using RANS code.
Journal of Fluids and Structures 23, 23–37.
Rinaldi, S., Paı̈doussis, M.P., 2010. Dynamics of a cantilevered pipe discharging fluid, fitted with a stabilizing end-piece. Journal of Fluids and Structures
26, 517–525.
Sinir, B.G., 2013. Pseudo-nonlinear dynamic analysis of buckled pipes. Journal of Fluids and Structures 37, 151–170.
Skop, R.A., Alasubramanian, S.B., 1997. A new twist on an old model for vortex-induced vibrations. Journal of Fluids and Structures 11, 395–412.
Skop, R.A., Griffin, O.M., 1975. On a theory for the vortex-excited oscillations of flexible cylindrical structures. Journal of Sound and Vibration 41, 263–274.
Vandiver, J.K., 1993. Dimensionless parameters important to the prediction of vortex-induced vibration of long, flexible cylinders in ocean currents.
Journal of Fluids and Structures 7, 423–455.
Violette, R., Langre, E.de., Szydlowski, J., 2007. Computation of vortex-induced vibrations of long structures using a wake oscillator model: comparison
with DNS and experiments. Computers and Structures 85, 1134–1141.
Wang, L., Dai, H.L., Qian, Q., 2012. Dynamics of simply supported fluid-conveying pipes with geometric imperfections. Journal of Fluids and Structures 29,
97–106.
Wang, X.Q., So, R.M.C., Chan, K.T., 2003. A non-linear fluid force model for vortex-induced vibration of an elastic cylinder. Journal of Sound and Vibration
260, 287–305.
Williamson, C.H.K., Govardhan, R., 2008. A brief review of recent results in vortex-induced vibrations. Journal of Wind Engineering and Industrial
Aerodynamics 96, 713–735.
Williamson, C.H.K., Roshko, A., 1988. Vortex formation in the wake of an oscillating cylinder. Journal of Fluids and Structures 2, 355–381.
Wu, X.D., Ge, F., Hong, Y.S., 2012. A review of recent studies on vortex-induced vibrations of long slender cylinders. Journal of Fluids and Structures 28,
292–308.
Yamamoto, C.T., Meneghini, J.R., Saltara, F., Fregonesi, R.A., Ferrari Jr., J.A., 2004. Numerical simulations of vortex-induced vibration on flexible cylinders.
Journal of Fluids and Structures 19, 467–489.
Zdravkovich, M.M., 1996. Different modes of vortex shedding: an overview. Journal of Fluids and Structures 10, 427–437.
Zhou, C.Y., So, R.M.C., Lam, K., 1999. Vortex-induced vibrations of an elastic circular cylinder. Journal of Fluids and Structures 13, 165–189.

View publication stats