viv of flexible riser conveying internal fluid subjected
TRANSCRIPT
VIV of Flexible Riser Conveying Internal Fluid Subjected to Uniform Current
Min Li, Di Deng, Decheng Wan*
Computational Marine Hydrodynamics Lab (CMHL), State Key Laboratory of Ocean Engineering,
School of Naval Architecture, Ocean and Civil Engineering, Shanghai Jiao Tong University, Shanghai, China *Corresponding author
ABSTRACT
Vortex-induced vibration (VIV) of a riser in consideration of both
internal and external flows is numerically studied based on the in-house
CFD solver viv-FOAM-SJTU. The main objective of the study is to
investigate effects of the internal flow on VIV responses of the riser,
when the velocity ratio of internal flow against external current varies.
The results from dynamic analysis indicate that the internal flow does
play a greatly important part in determining the typical VIV
characteristics of the riser including vibration amplitude, vibration
frequency and the dominant mode in both CF and IL directions. In
addition, multi-modal vibration and mode transition phenomenon may
occur when the internal flow velocity is relatively high.
KEY WORDS: Vortex-induced vibration; internal flow; viv-FOAM-
SJTU; velocity ratio.
INTRODUCTION
The flexible riser is an important part of the structure of deep-sea oil
exploitation system. When a circular cylinder is exposed to a flowing
fluid, it is easy to excite the vortex-induced vibration (VIV). When the
vortex shedding frequency is close to the natural frequency of the riser,
the phenomenon of "lock-in" (Yang and Li, 2009) will occur. Under
this circumstance, the vibration of the riser is significantly amplified
and the vortex shedding frequency becomes very close to the vibration
frequency, which may cause structural fatigue damage of the riser.
Therefore, it is of great significance to accurately predict the VIV
response of the marine risers for the practical design of deep-sea
structures.
Since the past decades, researchers have extensively studied the vortex-
induced vibration of flexible risers. Systematic experimental studies
were conducted, such as Chaplin (2005), Trim (2004) and Huera-
Huarte (2009). These experimental data provided reliable references for
later studies. In addition, massive numerical studies on VIV problems
of both rigid and flexible cylinders can be found such as Yamamoto et
al. (2004), Kaja et al. (2016) and Willden et al. (2001).
In practical engineering, the flexible riser works in the complex
environment, where external environment loads such as wind, wave
and current interact with the structure. Meanwhile, marine risers are
usually utilized for oil gas transportation. Though there exist massive
works carried out for VIV of the riser subjected to external current, the
VIV dynamic behavior of the flexible riser with internal flow has rarely
been studied relatively. When the internal fluid travels along the curved
pipe, it will generate Centrifugal acceleration and Coriolis acceleration,
so that the fluid dynamic pressure will periodically affect the dynamic
response of the riser and excite the additional vibration. It has been
proved that complicated phenomenon can emerge due to both internal
and external flows (Modarres and Paidoussis, 2013). Housner (1952)
concluded that the fluid inside the riser could cause the reduction of
natural frequency of the riser. The results indicate that when the
internal flow velocity increases to a certain critical velocity, the
dynamic dissipation and instability will occur in the riser. Paidoussis et
al. (1974) studied the dynamic characteristics of vertical pipes
transporting fluid through physical experiments as well as mathematical
models, and then analyzed the influence of Coriolis force and
Centrifugal force on the whole vibration system.
Empirical methods such as Van Der Pol wake oscillator model have
been used to calculate VIV of the riser with internal flow. Duan et al.
(2018) investigated VIV of a riser numerically considering both
internal and external flows based on the semi-empirical method
proposed by Thorsen et al. (2014). Typical VIV characteristics such as
the dominate mode and frequency in both IL and CF directions are
analyzed, as well as the Root Mean Square (RMS) of the amplitudes,
standing and traveling waves for the IL and CF responses. By using a
CAE technology which combines structural software with the CFD
technology, Chen et al. (2012) performed the numerical study about
VIV of a flexible riser model in consideration of internal flow
progressing inside. According to the result from dynamic analysis, it
has been found that the existence of upward-progressing internal flow
plays an important part in determining the dynamic behaviors of VIV.
Dai et al. (2013) focused on the VIV dynamic response of a hinged–
hinged pipe with internal fluid velocities ranging from the subcritical to
the supercritical regions.
2381
Proceedings of the Thirtieth (2020) International Ocean and Polar Engineering ConferenceShanghai, China, October 11-16, 2020Copyright © 2020 by the International Society of Offshore and Polar Engineers (ISOPE)ISBN 978-1-880653-84-5; ISSN 1098-6189
www.isope.org
In this paper, in order to investigate the dynamic response of the
flexible riser with internal flow, a series of numerical simulations are
conducted considering both internal and external flow with various
velocity combinations based on the in-house fluid-structure coupling
solver viv-FOAM-SJTU, and the reliability of the solver has been
effectively verified (Duan and Fu, 2016). In the following section, the
numerical method is introduced firstly. Then the structural modal is set
up based on the classical model test and numerically verified by
comparing with the experimental results in section 3. In section 4, the
numerical results of VIV response with different internal velocities of
the riser is analyzed and discussed. Finally, conclusions are given in
section 5.
NUMERICAL METHOD
Hydrodynamic governing equations
In this paper, the fluid is supposed to be incompressible. The governing
equations of the fluid domain are the Reynolds-averaged Navier-Stokes
(RANS) equations. In cartesian coordinate system, the governing
equations are described as follows:
0i
i
u
x
(1)
2i i j ij j i
j i j
pu u u S u u
t x x x
(2)
where 1
2
jiij
j i
uuS
x x
is the time-averaged rate of strain tensor,
ij= j iu u is regarded as the Reynolds stress resulted from the
fluctuating velocity field, representing the turbulence effect. For the
sake of turbulence closure, the SST k-ω turbulence model is used to
compute the Reynolds stresses in this paper.
Structural governing equations
The riser is regarded as a Euler-Bernoulli beam pinned at both ends.
The dynamical behavior of the riser caused by both external and
internal flows is taken into account. According to the design standards
of ocean risers and relevant studies (DNV, 2004), the effect of internal
flow including Centrifugal and Coriolis forces on the dynamic of the
riser is reflected in the following dynamic equation of the structure:
4 2 2
4 2
, , , , ,2 +m ) ,e i s i x
x z t x z t x z t x z t x z tEI T z mV m c f z t
z z z z t t t
( (3)
4 2 2
4 2
, , y , , ,2 +m ) ,e i s i y
y z t y z t z t y z t y z tEI T z mV m c f z t
z z z z t t t
( (4)
where L is the length of the riser, sm is the structural mass per unit
length of the riser, im is the internal fluid mass inside per unit length of
the riser, c is structural damping coefficient, eT z is the effective
tension of the riser. The axial tension of the riser varies along the riser
because of the influence of gravity, regardless of the variation of
tension with time. Referring to the analysis of internal flow by the
Norwegian design standard DNV-OS-F201 (2004), the effective
tension of the riser is influenced by both internal flow velocity and
internal fluid mass. The equations are expressed as follows:
t sT z T L z (5)
2
0 0e i i i iT z T z PA P A AV (6)
where T z is the axial tension, tT is the invariable applied top tension,
s is the weight per unit length of the riser submerged in water in
consideration of the effect of buoyancy, iP and 0P are the internal and
external pressures of the riser respectively, iA and 0A are the internal
and external cross-sectional area of the riser, i is the density of the
internal fluid, V is the internal flow velocity.
Based on the Finite Element Method (FEM), the differential equations
of motions in Eqs.3~4 for the riser can be expressed as follows:
xx x x F M C K (7)
yy y y F M C K (8)
where M is the mass matrix, C is the damping matrix, K is the
stiffness matrix, xF and yF represent the hydrodynamic forces
acting on the structure in the in-line and cross-flow direction
respectively. Moreover, the Newmark-beta method (Clough and
Penzien, 2001) is utilized to solve the structural dynamics equation.
Fluid-structure interaction
Because of the large scale of the flexible riser in deep water, the three-
dimensional numerical simulation will consume a huge amount of
computational resources. Therefore, the strip method which is widely
used to simplify the flow field is adopted to simulate the VIV dynamics
of the flexible riser, and the coupling iterative program developed by
our group can realize the coupling solution of structural dynamics and
fluid dynamics. The main idea of this method is to calculate the
hydrodynamic forces at each strip, and then interpolate the forces to
corresponding elements of the structural modal. Fig. 1 shows the
schematic diagram of the coupling process of the strip method.
Fig. 1 The schematic diagram of the strip method
NUMERICAL MODEL VALIDATION
Before applying the modified solver to investigate the dynamic
response of the riser with internal flow, two series of numerical
simulations are carried out to verify its effectiveness of the solution
method in this section.
Firstly, the internal flow is not considered. The experiment carried out
by Lehn (2003) is selected. The test was carried out in MARINTEK,
2382
Norway. The riser model is placed vertically in the pool and subjected
to the incoming flow, as shown in Fig. 2. The riser is 0.02m in diameter
and 9.63m long, the aspect ratio of which L/D is 481.5. Other key
characteristics of the riser modal are presented in Table 1. Based on the
model, Lehn conducted a series of experiments under uniform and
shear flows.
Fig. 2 Sketch of test rig (Lehn, 2003)
Table 1. Main parameters of the model riser
Parameter Symbol Value
Diameter D 0.02 m
Length L 9.63 m
Aspect ratio L/D 135.4
Top tension Tt 817 N
Bending stiffness EI 135.4 N m2
Mass ratio 2.23
1th eigenfrequency fn1 1.79 Hz
2th eigenfrequency fn2 3.67 Hz
In the validation condition, the riser is subjected to the external uniform
flow U=0.2m/s. In this simulation, 20 strips are set equidistantly along
the span of the cylinder, and the computational domain is set to be -20D
≤x≤40D, -20D ≤y≤20D. The distribution of strips along the span of the
riser is shown in Fig. 3. The boundary conditions and mesh division of
all strips are consistent. Fig. 4 shows the geometry domain and entire
mesh of each strip. It can be found that grids are gradually becoming
sparse from the structure to the outer region.
Fig. 3 Layout of the computational model
Fig. 4 Entire mesh on each strip
Because of the great significance of the grid quality, three different
simulations are carried out by varying strip numbers and mesh density.
Table 2 presents the information of the three meshes and dimensionless
Root Mean Square (RMS) values of the IL and CF displacement
respectively. The present results are compared with the experimental
data and numerical results of Wang (2016) and Wu (2019). It can be
found that the RMS values exported from the three meshes obtain
relatively good agreement with the experimental results. The calculated
result of Mesh 1 is slightly lower than the experimental value while the
RMS values with Mesh 2 and Mesh 3 are extremely close to the
experimental data in the CF direction. Meanwhile, the results all agree
well in the IL direction. The error of RMS values among three
simulations is acceptable. Fig. 5 shows the RMS of vibration
displacement along the span of the whole riser in CF and IL directions.
It can be found that the dominant vibration modes are first order and
second order in the CF and IL direction respectively. The vibration
trajectories along the cylinder calculated by Mesh 3 are shown in Fig.
6(c) and (d). The corresponding experimental results are also given in
Fig. 6 (a) and (b). It can be observed apparently that the vibration
shapes are in good agreement with the experimental results. Slight
disagreement (in magnitude and position) is probably related to some
simplification methods adopted in the process of the convergence
calculation and differences in the selection of sampling time.
Table 2. the details of three-meshes and simulation results
Case Strip
number
Number
of Cells
max
xrmsA / D max
yrmsA / D
Experiment - - 0.103 0.401
Wang (2016) - - 0.125 0.395
Wu (2019) 20 811000 0.114 0.404
Mesh 1 10 366720 0.102 0.368
Mesh 2 20 733440 0.110 0.398
Mesh 3 20 1026000 0.123 0.410
(a) (b)
Fig. 5 Comparation of RMS of the vibration displacement along the
cylinder span with three meshes: (a) CF direction and (b) IL direction
2383
(a) (b)
(c) (d)
Fig. 6 The comparison of the vibration trajectories along the cylinder:
(a) the CF direction (Lehn, 2003) (b) the IL direction (Lehn, 2003) (c)
the CF direction (Mesh 3) (d) IL direction (Mesh 3)
Moreover, a series of numerical simulations have been carried out in
consideration of internal flow, the results of which are compared with
the empirical results by Li (2010). The numerical simulation model
follows the experiment by Chaplin et al.(2005a) in which the riser is
subjected to a stepped flow. Fig. 7 shows the comparison between the
empirical result of vibration displacement concerning three different
internal flow velocities in the CF direction and the present
corresponding numerical results, shown in Fig.7 (a) and (b) respectively.
Although there exists a certain error in the vibration amplitude, which
is due to the considerable difference in two methods. It is known that
CFD method is more accurate, especially at the nodes along the span of
riser. And the trend of the vibration response affected by internal flow
calculated by the CFD method agrees well with the empirical results.
Therefore, the effectiveness of the method is verified.
(a) (b)
Fig. 7 The comparison of the RSM displacement with internal flow
velocities in the CF direction: (a) empirical result (Li, 2010) (b)
numerical result
RESULT AND DISSCUSION
Set up of simulation conditions
Aiming at investigating effects of internal flow on the dynamic
response of the riser under external current, the VIV dynamics with
different internal flow velocities V are investigated based on the above
test model by Lehn in this section. Ten simulation cases of various
combination of internal flow velocities (V=0 m/s, 1 m/s, 2 m/s, 5 m/s,
10 m/s) and external flow velocities (U=0.2 m/s, 0.36 m/s) are
performed respectively, as shown in Table 3. Here, the expression V =0
m/s means the water inside the riser being in still state.
According to the "lock-in" region of the first-order vibration mode and
the second-order vibration mode of the riser model, the above two
corresponding external flow velocities are chosen. It is generally
believed that 0 9 1 4s nf . . f is the " lock-in " resonance region
(Carberry and Sheridan, 2001).
ts
S Uf
D (9)
where sf is the vortex shedding frequency, tS is the Strouhal number.
Table 3. Set up of simulation conditions
Case External flow velocity Internal flow velocity
1 0.2 m/s 0 m/s
2 0.2 m/s 1 m/s
3 0.2 m/s 2 m/s
4 0.2 m/s 5 m/s
5 0.2 m/s 10 m/s
6 0.36 m/s 0 m/s
7 0.36 m/s 1 m/s
8 0.36 m/s 2 m/s
9 0.36 m/s 5 m/s
10 0.36 m/s 10 m/s
Dynamic analysis of flexible riser
Fig. 8 shows the dimensionless RMS value of the IL and CF
displacement concerning different velocities of internal flow (V) under
the external flow velocity U=0.2 m/s. The results show that when V=0
m/s, the vibration amplitude of the riser in both directions is close to
that of the riser without internal flow by comparison of the results in
the preceding section. The modal power spectral densities of the riser
subjected to the external flow U=0.2 m/s in the CF and IL directions are
presented in Fig. 9 and Fig. 10, respectively. The VIV dynamic
response of the riser influenced by the internal flow including vibration
displacement, vibration frequency and dominant mode will be analyzed
as follows:
As shown in Fig. 8, with lower increasing internal velocity (case 2 and
3), it can be found that the vibration amplitude of the riser decreases
slightly. And in cases 1~3, the riser vibrates at a single frequency, and
the main vibration frequency is relatively close in the CF direction as
shown in Fig. 9 (a), (b) and (c). The coverage range of the dominant
frequency in case 2 and 3 is slightly wider than that in case 1,
indicating that the vortex shedding frequency in case 1 is closer to the
first mode "lock-in" frequency, so that the vibration energy of the VIV
processes is highly concentrated, regarding as “narrow-band”. It can be
known that it is in the very range of locking vibration region and "lock-
in" phenomenon occurs. When the internal flow velocity increases, the
reduction of the natural frequency of the riser causes that the whole
riser gradually leaves the first order "lock-in" region. That is why the
vibration amplitude of the riser decreases. However, the vibration mode
remains the same and the vibration intensity is comparative in case 1~3
2384
as shown in subplot (a)~ (c) of Fig. 9 and Fig. 10. With the increase of
internal flow velocity (case 4 and 5), it is apparent that higher vibration
mode dominates the whole riser system as shown in Fig. 8. Because
when the internal flow velocity is high, the natural vibration frequency
of the riser decreases greatly, and the vortex shedding frequency may
fall into a higher adjacent mode under the same external flow velocity.
Thus, the riser may vibrate with higher vibration mode. In addition, the
maximum RMS value of the displacement of the riser decreases
significantly, with the reduction of 50% in the CF direction and 40% in
the IL direction approximately. From the perspective of the energy
transfer between the outflow and the riser, the high mode with a large
amplitude is difficult to be excited. Because of the higher vibration
mode, the vibration frequency increases correspondingly as shown in
(d)~(e) of Fig. 9 and 10. The vibration frequency is approximately
equal concerning two cases with V=5 m/s and V=10 m/s.
(a) (b)
Fig. 8 The RMS value with different V, when external flow velocity
U=0.2 m/s: (a) CF RMS value and (b) IL RMS value
In comparison of vibration response of two directions, under the
influence of internal flow, the frequency variation trend in the IL
direction is the same as that in the CF direction. In general, the in-line
vibration frequency is exactly twice as that in cross-flow direction, so
does the dominant mode.
In addition, some subordinate frequency components of vibration
modes with considerable vibration intensity emerges in the IL direction
as shown in Fig. 10 (a)~(c). The fourth modes have been excited
around the original second vibration mode and the total vibration
intensity of others is much smaller than the former two. By comparison
of various internal flow velocities, it can be seen that the existence of
internal flow will change the dominant vibration frequency to some
degree by influencing the natural frequency of the riser. It also can be
found that the lower internal flow mainly influences the values of
vibration displacement. However, a higher internal flow velocity may
excite a higher dominant vibration mode.
(a) (b) (c)
(d) (e)
Fig. 9 CF modal power spectral densities with different internal flow
velocities: (a) V=0 m/s, (b) V=1 m/s, (c) V=2 m/s, (d) V=5 m/s and (e)
V=10 m/s, when external flow velocity U=0.2 m/s
(a) (b) (c)
(d) (e)
Fig. 10 IL modal power spectral densities with different internal flow
velocities: (a) V=0 m/s, (b) V=1 m/s, (c) V=2 m/s, (d) V=5 m/s and (e)
V=10 m/s, when external flow velocity U=0.2 m/s
Fig. 11 shows the RMS value of displacement in the IL and CF
direction concerning different velocities of internal flow under the
external flow velocity U=0.36 m/s. The calculation results show that
the vibration mode of the riser increases compared with that of
U=0.2m/s with the same velocity of internal flow. In case 6, the
vibration response is mainly dominated by the 2nd mode in the CF
direction and 5st mode in the IL direction. Therefore, it can be found
that the external flow velocity has a great influence on the modal
response of the riser. The modal power spectral densities of the riser
subjected to the external current U=0.36 m/s in the CF and IL
directions are presented in Fig. 12 and Fig. 13, respectively. Contrary to
2385
the results of the preceding group (U=0.2m/s), when the internal flow
velocity is lower (case 7 and 8), it can be found that the vibration
amplitude of the riser increases in the CF direction, which is consistent
with the conclusion of Guo and Lou (2008). When the internal fluid is
still, it can be seen that the dominant vibration mode is second-order
with slight first mode component as shown in Fig. 12 (a). Therefore, it
is under the transition state from first mode to second mode at the
current velocity of U=0.36m/s which is relatively far from the second
mode "lock-in" region. With the increasing internal flow velocity and
then the decrease of the natural frequency of the riser, the number of
elements with "lock-in" phenomenon increases in this condition so that
the vibration becomes more violent.
As the internal flow velocity continues to rise (case 9 and 10), the
vibration amplitude in the CF direction decreases by about 33% and
20% respectively compared with that in case 6, which is distinctly
lower than that in the preceding group. Moreover, the vibration
displacement and dominant vibration mode in the IL direction become
relatively stable with the increase of the internal flow velocity. On the
whole, from the perspective of the fluctuation of displacement
amplitude and the dominant vibration mode, existence of internal flow
seems to have a less effect on riser’s vibration displacement response
compared with the condition of U=0.2 m/s. It is decided by the velocity
ratio of V/U. The larger the velocity ratio is, the more of its influence
contributes to the dynamic response of the riser in consideration of both
external and internal flow.
In general, the dominant vibration mode of the riser is determined by
the coefficient of the natural frequency and the Strouhal frequency of
flexible riser. The Strouhal frequency reflects the external energy,
while the natural frequency expresses the structural characteristics of
the riser itself. It is known that the corresponding vibration mode with
natural frequency closest to the Strouhal frequency, tends to be an
apparent or dominant vibration mode. When the natural frequency of
the riser changes, there is no doubt that the dominant mode of vibration
changes accordingly. Therefore, the existence of internal flow with
different velocities contributes to the vibration mode of the riser greatly
which should not be negligible. The foresaid analysis of the mode
response is consistent with this conclusion, which verifies the reliability
of the calculated results.
(a) (b)
Fig. 11 The RMS value with different V, when external flow velocity
U=0.36 m/s: (a) CF RMS value and (b) IL RMS value
Single-modal vibration state can still be widely observed in case 6~8,
shown in Fig. 12 (a), (b) and (c) subsequently. Although it seems that
the internal flow has little effect on riser’s frequency response, the extra
effect along the main driving portion of the flexible riser model implies
that it is doomed to be a potential fact to eliminate or excite some
candidate vibration mode in some way (Chen, 2010). Moreover, it can
be seen the transformation process of dominate vibration mode from
the second one to the third one with the increase of the internal flow
velocity, as shown in Fig. 12 (d) and (e). When V=5m/s, the third mode
has been excited and becomes the subordinate vibration mode with
considerable energy. Fig. 12 (e) shows that the vibration intensity of the
subordinate vibration mode has exceeded that of the original dominant
vibration mode. In the IL direction, the phenomenon of multi-modal
vibration response can be widely observed. However, the fifth vibration
mode remains the dominant one and some subordinate modes with
small vibration intensity could be negligible.
On the whole, the dynamic vibration response of the riser is more
complicated when external current U rises from 2 m/s to 3.6 m/s. Even
in the case of V=0 m/s when U=3.6 m/s, subordinate vibration modes
emerge in both directions. The internal flow velocity makes less effect
on the frequency response of the riser when the current velocity is
relatively higher, which is mainly reflected in the dominant vibration
frequency.
(a) (b) (c)
(d) (e)
Fig. 12 CF modal power spectral densities with different internal flow
velocities: (a) V=0 m/s, (b) V=1 m/s, (c) V=2 m/s, (d) V=5 m/s and (e)
V=10 m/s, when external flow velocity U=0.36 m/s
Fig. 14 shows the power spectra densities in case 2 and 9 respectively,
which is made by the CF and IL displacements at nine equidistant
arranged nodes of the riser via FFT. The displacement power spectra of
multiple positions along the span of the riser are easily used to analyze
the distribution characteristics of vibration frequency domain along the
span of the riser. It can be found that the vibration energy distribution
in case 2 is gradually weakened from the middle to both ends of the
riser, shown in Fig.14 (a). From another point of view, it indicates that
the dominant vibration mode is the first order. It is apparent that at a
higher outflow velocity, there exist the multi-frequency components
and multi-mode resonance of the riser as shown in Fig. 14 (c). The VIV
dynamic response is more complicated. Fig. 15 shows the power
2386
spectrum densities of the cross-flow displacements along the span in
case 6 and 10 subsequently. Multi-mode and mode transition
phenomenon can be more apparently presented.
(a) (b) (c)
(d) (e)
Fig. 13 IL modal power spectral densities with different internal flow
velocities: (a) V=0 m/s, (b) V=1 m/s, (c) V=2 m/s, (d) V=5 m/s and (e)
V=10 m/s, when external flow velocity U=0.36 m/s
(a) (b)
(c) (d)
Fig. 14 The power spectrum densities of the cross-flow displacements
along the span: (a) case 2 in CF direction, (b) case 2 in IL direction, (c)
case 9 in CF direction and (d) case 9 in IL direction
(a) (b)
Fig. 15 Non-dimensional CF vibration modal weight at the mid-span of
the cylinder: (a) case 6 and (b) case 10
Table 4 concludes the vibration frequency and vibration modes of the
riser according to the magnitude of the modal weight in both CF and IL
directions of ten calculation conditions. In general, it has been found
that the vibration frequencies are comparative concerning different
internal flow velocities when the internal flow velocity is relatively low
and the mode transition phenomenon does not occur. With the increase
of V, the multi-modal vibration and the mode transition phenomenon
will occur.
Table 4. The vibration frequency and vibration modes of the riser in
both CF and IL directions
Case CF IL
frequency modes frequency modes
1 1.7 Hz 1 2.29 Hz 2, 4
2 1.68 Hz 1 2.29 Hz 2, 4
3 1.68 Hz 1 2.28 Hz 2, 4
4 2.67 Hz 2 5.14 Hz 4
5 2.54 Hz 2 4.98 Hz 4
6 3.16 Hz 2, 1 8.41 Hz 5, 4, 1
7 3.14 Hz 2 8.40 Hz 5, 1
8 3.17 Hz 2 8.45 Hz 5, 4, 1
9 2.82 Hz 2, 3, 1 7.37 Hz 5, 1
10 4.1 Hz 3 ,2, 1 7.29 Hz 5, 1, 6
2387
CONCLUSIONS
Aiming at investigating the effect of internal flow on the VIVs dynamic
of the riser, numerical simulations in consideration of both the internal
flow and the external uniform flow have been carried out in this paper.
Centrifugal and Coriolis forces excited by the internal flow are both
included in the riser’s vibration. By analyzing the simulation results of
a series of cases, the following conclusions are drawn:
With the increase of the outflow velocity, the vibration of the cylinder
in both directions intensifies obviously. In addition, the internal flow
also has a great effect on VIV dynamic response of the flexible riser.
When the internal flow velocity is relatively low, it mainly affects the
vibration displacement as the natural frequency of the riser decreases
due to the increase of the internal flow velocity. However, when the
change of the natural frequency causes that the riser is closer to the
locking vibration region, the increase of the internal flow will
strengthen the vibration. And the comparatively low internal flow
velocity contributes little to the dominant vibration frequency.
When the internal flow velocity continues to increase to some extent,
multi-modal vibration and mode transition of the riser are widely
observed. What’s more, the vibration intensity of the higher dominant
mode may be attenuated and the dominant frequency increases
correspondingly. The complicated VIV dynamic response makes the
riser more prone to fatigue damage. Therefore, the influence of internal
flow should be paid much attention in the practical engineering design
of the pipe.
Compared with the cross-flow vibration response, the in-line
displacement amplitude is much lower, and the influence of internal
flow is relatively insensitive, but its vibration frequency is about twice
than that of in cross-flow. Moreover, the multi-modal vibration
phenomenon is more likely to occur, making the riser in a dangerous
state. Thus, it is necessary to make accurate prediction of vortex-
induced vibration responses in both directions at the same time.
ACKNOWLEDGMENTS
This work is supported by the National Natural Science Foundation of
China (51879159), The National Key Research and Development
Program of China (2019YFB1704200, 2019YFC0312400), Chang
Jiang Scholars Program (T2014099), Shanghai Excellent Academic
Leaders Program (17XD1402300), and Innovative Special Project of
Numerical Tank of Ministry of Industry and Information Technology of
China (2016-23/09), to which the authors are most grateful.
REFERENCES
Carberry J., Sheridan J. (2001). “Force and wake modes of an oscillating
cylinder,” J Fluids and Structures, 15, 523~532.
Chaplin J.R., Bearman P.W. and Huera Huarte F.J. and Pattenden R
(2005a). “Laboratory measurements of vortex-induced vibrations of a
vertical tension riser in a stepped current,” J Fluids and Structures, 21,
3-24.
Chen Z S , Kim W J ,and Xiong C B (2012). “Effect of upward internal
flow on dynamics of riser model subject to shear current,” J China
Ocean Engineering, 26(1), 95-108.
Clough R.W. and Penzien J. (2001). “Dynamics of Structures,” J Applied
Mechanics, 44(2), 366.
Dai H., Wang L., Qian Q., Ni Q. (2013). “Vortex-induced vibrations of
pipes conveying fluid in the subcritical and supercritical regimes,” J.
Fluids Struct. 39, 322–334.
Det Norske Veritas (2004). Recommended practice: DNV-OS-F205-
Global Performance Analysis of Deepwater Floating Structures,
Norway: Det Norske Veritas.
Duan J.L., Chen K., You Y.X. and Li J.L. (2018). “Numerical
investigation of vortex-induced vibration of a riser with internal flow,”
J Applied Ocean Research, 72, 110-121.
Duan M.Y. and Fu B.W. (2016). “Prediction of response for vortex-
induced vibrations of a flexible riser pipe by using multi-strip method,”
The Twenty-sixth International Ocean and Polar Engineering
Conference. Rhodes, Greece.
Guo H Y and Lou M (2008). “Effect of internal flow on vortex-induced
vibration of risers,” J Fluids and Structures, 24(4), 496-504.
Housner G (1952). “Bending vibrations of a pipe line containing flowing
fluid,” J Applied Mechanics-Transactions of the ASME, 19 (2), 205–
208.
Huera-Huarte F and Bearman P (2009). “Wake structures and vortex-
induced vibrations of a long flexible cylinder—part 1: dynamic
response,” J Fluids and Structures, 25 (6), 969–990.
Kaja K., Zhao M., Xiang Y. and Cheng L. (2016). “Three-dimensional
numerical simulations of vortex-induced vibrations of tapered circular
cylinders,” J Appl. Ocean Res. 601–11.
Lehn, E. (2003). “VIV Suppression Tests on High L/D Flexible
Cylinders,” Norwegian Marine Technology Research Institute,
Trondheim, Norway.
Li X M, Guo H Y, Meng F S (2010). “Effect of internal flow on the
dynamic behavior of top tensioned riser,” J Ship Mechanics, 14(9),
1021-1030.
Modarres-Sadeghi Y. and Paidoussis M.P. (2013). “Chaotic oscillations
of long pipe conveying fluid in the presence of a large end-mass,” J
Comput. Struct, 122,192–201.
Paidoussis M P and Issid N (1974). “Dynamic stability of pipes
conveying fluid,” J J sound and vibration, 33 (3), 267–294.
Thorsen M.J., Sævik S. and Larsen C.M. (2014). “A simplified method
for time domain simulation of cross-flow vortex induced vibrations,” J
Fluids Struct. 49, 135–148.
Trim A.D., Braaten H., Lie H., Herfjord K., King R., Makrygiannis C.M.
and Meling T.S (2004). “Experimental investigation of vortex-induced
vibration of long marine risers,” FIV2004 (Flow Induced Vibration)
Conference, Ecole Polytechnique, Paris, France.
Wang E.H. and Xiao Q. (2016). “Numerical simulation of vortex-induced
vibration of a vertical riser in uniform and linearly sheared currents,” J
Ocean Engineering, 121.
Willden, R.H.J. and Graham, J.M.R. (2001). “Numerical prediction of
VIV on long flexible circular cylinders,” J Fluids Struct. 15, 659-669.
Wu L., Deng D., Wan D.C. and Chen G. (2019), “Numerical Simulations
of Vortex-Induced Vibrations of Flexible Risers in Tandem
Arrangement,” The Twenty-ninth (2019) International Ocean and
Polar Engineering Conference, Hawaii, USA, 3458-3464.
Yamamoto C., Meneghini J., Saltara F., Fregonesi R. and Ferrari J.
(2004). “Numerical simulations of vortex-induced vibration on flexible
cylinders,” J. Fluids Struct, 19, 467–489.
Yang H Z and Li H J (2009). “Instability Assessment of Deep-Sea Risers
Under Parametric Excitation,” J China Ocean Engineering, 23(4), 603-
612.
2388