See

discussions, stats, and author profiles for this publication at: http://www.researchgate.net/publication/259092608

Hydrodynamic characteristics of an oscillating

circular disk under steady in-plane current
conditions
ARTICLE in OCEAN ENGINEERING JANUARY 2014
Impact Factor: 1.35 DOI: 10.1016/j.oceaneng.2013.11.011

READS

87

3 AUTHORS, INCLUDING:
Xinliang Tian
Shanghai Jiao Tong University
8 PUBLICATIONS 7 CITATIONS
SEE PROFILE

Available from: Xinliang Tian

Retrieved on: 29 September 2015

Ocean Engineering 75 (2014) 5363

Contents lists available at ScienceDirect

Ocean Engineering
journal homepage: www.elsevier.com/locate/oceaneng

Hydrodynamic characteristics of an oscillating circular disk under

steady in-plane current conditions
Jianmin Yang, Xinliang Tian n, Xin Li
State Key Laboratory of Ocean Engineering, Shanghai Jiao Tong University, Shanghai 200240, China

art ic l e i nf o

a b s t r a c t

Article history:
Received 6 November 2012
Accepted 6 November 2013
Available online 26 November 2013

The heave motion performance of a Spar platform could be improved signicantly by installing heave
plates. The heave plates not only increase the vertical added mass of the platform but also provide extra
damping in heave. This paper presents an investigation of the hydrodynamic characteristics of a heave
plate under steady current conditions using three-dimensional direct numerical simulations. A circular
disk of diameter D is forced to oscillate along its axis sinusoidally. Three aspect ratios of the disk, i.e.,
t d =D 0:1; 0:05 and 0.02, are considered, respectively. The current is introduced in the plane of the disk.
The inuences of three parameters, the KeuleganCarpenter number (KC), the reduced current velocity
(Vr) and the aspect ratio of the disk, on the ow structures, the current drag force, the added mass and
the damping of the heave plate are examined both qualitatively and quantitatively. The ranges of the KC
and Vr numbers considered in this study are 0 r KC r 1 and 0 r V r r 0:25, respectively. The frequency
number for the oscillating disk is xed at 200.
& 2013 Elsevier Ltd. All rights reserved.

Keywords:
Heave plate
Added mass
Damping
Vortex shedding
DNS

1. Introduction
Deepwater offshore structures at sea, e.g., a Spar platform, may
experience motion responses due to incident wave forces (Tao and
Dray, 2008). The motion in the vertical direction, i.e., the heave
motion, is a particular concern of engineers because large heave
motion may cause damage in risers, drilling pipes and mooring
lines. Heave plates are often installed in Spar platforms, and they
have been demonstrated to be effective components to minimise
the heave response (Downie et al., 2000). The heave plates
signicantly increase the effective heave added mass of the platform. Therefore, the natural frequency of the heave motion is
decreased and is displaced further out of the wave-frequency
region of signicant wave energy to avoid resonance. On the other
hand, heave plates can provide extra damping forces due to vortex
shedding at the edges of the plates.
In the past two decades, many experimental (see, e.g., Prislin
et al., 1998; Thiagarajan and Troesch, 1998; Lake et al., 2000;
Thiagarajan et al., 2002; He, 2003; Tao and Dray, 2008) and
numerical (see, e.g., Magee et al., 2000; Holmes et al., 2001;
Molin, 2001; Tao and Thiagarajan, 2003a,b; Tao and Cai, 2004; Tao
et al., 2007; Shen et al., 2012) studies on the hydrodynamic
characteristics of heave plates have been conducted. The added
mass and the damping of the heave plates have been found to be
n

Corresponding author. Tel.: 86 13681957161; fax: 86 34207058.

E-mail address: tianxinliang@sjtu.edu.cn (X. Tian).

0029-8018/$ - see front matter & 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.oceaneng.2013.11.011

dependent on both the amplitude and the frequency of the heave

motion. Moreover, the hydrodynamic characteristics of the heave
plates are also inuenced by the geometric congurations of the
heave plates, e.g., the size of the plates (Thiagarajan et al., 2002),
the distances between multiple plates (Magee et al., 2000; Tao
et al., 2007), the thickness (Thiagarajan et al., 2002; Tao and
Thiagarajan, 2003a,b; He, 2003), the porosity (Molin, 2001; Tao
and Dray, 2008) and the edge taper angles (Shen et al., 2012).
It is known that the current is a phenomenon that is widely
existing in the majority of ocean areas and may have a prominent
inuence on the hydrodynamic performance of offshore structures, e.g., the VIV (vortex induced vibration) of risers and
pipelines. However, investigations on heave plates under current
conditions are quite scarce. To the best of our knowledge, the
exceptions are the studies conducted by Thiagarajan and Troesch
(1998) and Holmes et al. (2001). Thiagarajan and Troesch (1998)
experimentally investigated the inuence of current on the damping of a circular cylinder with a circular disk attached to its bottom
in a towing tank, and the existence of a small current was reported
to increase the heave damping by 87%. Holmes et al. (2001)
performed numerical simulations for an isolated square heave
plate with a constant pitch angle (71) under current conditions.
The load distributions on the plate were not symmetric about the
plate centre, in contrast to the cases at a zero pitch angle and
without current. It is apparent that the hydrodynamic characteristics of heave plates under current conditions have not been
well addressed, and the conclusions reported by Thiagarajan and

54

J. Yang et al. / Ocean Engineering 75 (2014) 5363

2. Mathematical formulations and numerical methods

The Cartesian coordinate system (x; y; z) is used in this study.
If these coordinates are written as x1 ; x2 ; x3 and the velocity
component in the xi-direction is denoted as ui, where i13, the
NavierStokes (N-S) equations for an incompressible viscous uid
are written as follows:
ui
0
xi

ui
u
1 p
2 ui
uj i 

t
xj
xi xj xj

Fig. 1. Schematic diagram of the oscillating circular disk under in-plane current.

Troesch (1998) and Holmes et al. (2001) are limited to the specic
ow congurations in their studies.
In this study, the hydrodynamic characteristics of a heave plate
under steady in-plane current conditions are investigated with the
technique of Computational Fluid Dynamics (CFD) using direct
numerical simulation (DNS). We consider a thin circular disk that
is forced to oscillate sinusoidally along its axis (z) in a uniform
steady in-plane current, see Fig. 1. The displacement of the disk
follows a function of time (t): zdisk t a sin t, where a is the
single amplitude of the oscillation, 2 =T (T is the period of the
oscillation) is the angular frequency of the oscillation. Thus, the
axis velocity of the disk could be written as uz t U max cos t,
where Umax is the maximum speed of the oscillation motion, i.e.,
U max 2 a=T. Two dimensionless parameters characterising the
oscillation motion are the KeuleganCarpenter number (KC) and
the frequency number (), and they are dened, respectively as
KC

2 a U max T

D
D

D2
T

where D is the diameter of the disk and is the kinematic viscosity

of the uid. The Reynolds number for the oscillation motion is
dened as
Reo

U max D

KC

The current velocity is nondimensionlised in a form similar to

the reduced velocity, which is widely used in the area of VIV, and
is given as
Vr

Uc

UcT
UcD

2 D 2

where Uc is the magnitude of the current velocity. The Reynolds

number for the steady current ow is dened as
Rec

Uc D

For a prototype Spar under extreme sea conditions, assume

D30 m, a3 m, T25 s and Uc 2 m/s, which results in KC  0:63
and V r  0:25. The ranges of KC and Vr considered in this study are
0 r KC r1 and 0 rV r r0:25, respectively. This study focusses on
the inuences of KC and Vr on the hydrodynamic characteristics of
the heave plates while the number is xed at 200. In order to
investigate the inuence of aspect ratios (thickness/diameter) of the
disk, three disks with t d =D 0:1; 0:05 and 0.02 are considered,
respectively. This paper provides detailed discussions on the inuences of KC and Vr on the ow structures, the current drag force, the
added mass and the damping of the disk. The inuences of aspect
ratio of the disk on the hydrodynamic characteristics of the disk are
also discussed in this paper.

where p is the pressure and is the density of the uid.

The N-S equations are discretised using the nite volume
method (FVM) based on the open source CFD code OpenFOAM.
OpenFOAM is mainly applied to solve problems in continuum
mechanics. It is based on the tensorial approach and object oriented
techniques (Weller et al., 1998). The PIMPLE algorithm, which
merges the PISO (Pressure Implicit with Splitting of Operators)
and SIMPLE (Semi-Implicit Method for Pressure-Linked Equations)
algorithms, is used in this study. The spatial schemes for the
interpolation, the gradient, the Laplacian and the divergence are
linear, Gaussian linear, Gaussian linear corrected and Gaussian
linear schemes, respectively. All of these schemes are in second
order. The second order CrankNicolson scheme is used for the
time integration. Further details of these schemes are given in
OpenFOAM (2009).
In this study, the axis of the disk is in the z-direction, and the
current is in the x-direction. Therefore, the drag force coefcients
in x- and z-directions are calculated, respectively, as
Cx

Cz

Fx
1
U 2c Dt d
2

Fz
1
U 2ref D2
8

where Fx and Fz are the force components in the x- and

z-directions acting on the disk, respectively, and they are directly
calculated by integrating the pressure and viscous shear stress
over the disk surfaces. In the case of a steady ow normal to a
stationary disk for code validation (see Section 3), the reference
speed (Uref) is taken as the free stream speed (Us), and, in the
oscillating disk case, Uref is taken as the maximum speed of the
disk (Umax). The time-averaged value of Cx and the r.m.s. (root
mean square) value of Cx are denoted as C x and Cxrms, respectively.
Czrms denotes the r.m.s. value of Cz.
Following the formulations used by Tao and Dray (2008), the
axis viscous drag force Fzd acting on the disk can be written in
Morison's equation-like form as


1 D2
uz t
uz tj
F zd t C d
2 4

10

where Cd is the drag coefcient. By using Fourier analysis, the drag

and added mass coefcients could be calculated, respectively, as
(Sarpkaya et al., 1981)
Z T
3
Cd 
F z t cos t dt
11
2 2
D U max 0
Ca

1
8 U max

T
0

F z t sin t dt

12

where 8 is the immersed volume of the disk, which is taken as

8 1=4 D2 t d .

J. Yang et al. / Ocean Engineering 75 (2014) 5363

55

The axis viscous drag force can also be expressed using an

equivalent linear damping coefcient B as
F zd t Buz t

13

Free-slip

Outlet

Then, the linear damping coefcient is obtained as

B 13 DKCC d

Disk

14

The added mass and the damping coefcients are further

nondimensionalised against the theoretical ideal uid added mass,
m 1=3D3 . Thus, the added mass and damping coefcients are
rewritten, respectively, as
Cam
m

15

B
KCC d

2m
4

16

where m is the displacement mass of the disk, m 8 .

It should be noted that the Reynolds numbers investigated in this
study are relatively low, i.e., Reo r 200 and Rec o 315; therefore, the
ow is laminar and highly periodic (for oscillating disk cases).
Therefore, for the stationary disk cases (KC0), the results presented
in the following part are obtained after the ow reaches a steady
stage. For the oscillating disk cases (KC 40), the results are calculated
from the data in one oscillation cycle after the ow has repeated
itself.
The vorticity components in the y- and z-directions are
denoted as y and z , respectively. They are dened as

u1 u3

x3 x1

17

u2 u1

x1 x2

18

The so-called Q-criterion proposed by Hunt et al. (1988) is used

to identify the three-dimensional vortical structures. The denition of Q is
Q  12 J S J 2  J J 2 ;

19

where S and denote the strain and the rotation tensor, respectively.
3. Code validation
To validate this numerical approach, benchmark calculations of
the steady ow normal to a stationary circular disk were performed. Following the setup in Shenoy and Kleinstreuer (2008),
we considered a circular disk with the thickness of 0.1D and used a
cylindrical computational domain with a cross-section of diameter
12D; see Fig. 2(a). The origin of the coordinates (x; y; z) was located
at the centre of the disk and the z-axis coincided with the axis of
the disk. The inlet and outlet boundaries were located at 2:5D
upstream and 15D downstream of the disk, respectively. This
computational domain had been demonstrated to be large enough
to eliminate the boundary effects (Shenoy and Kleinstreuer, 2008).
On the disk surface, no-slip and zero normal pressure gradient
boundary conditions were employed. At the inlet boundary, a
uniform velocity (0,0,Us) and zero normal pressure gradient
boundary conditions were prescribed. At the outlet boundary,
the velocity was set to a zero normal gradient, and the pressure
was xed to zero. Free-slip and zero normal gradient boundary
conditions were applied for the velocity and the pressure, respectively, on the side boundaries of the computational domain.
The mesh structures also followed the topology used by Shenoy
and Kleinstreuer (2008), and the computational domain is discretised with hexahedral elements; see Fig. 2(b). The grids were
rened near the surface of the disk. To further assess the effects of
the spatial and temporal resolutions on the results, three cases

12D

Inlet

2.5D

15D

Disk

Fig. 2. Stationary disk (t d =D 0:1): (a) computational domain, boundary conditions and (b) grid structures near the disk surface.

with different spatial and temporal resolutions were investigated;

see the three cases in Table 1, where n1 denotes the element size
in the normal-wall direction next to the disk surface, t is the
time step, Res denotes the Reynolds number based on the free
stream velocity (Us) and the diameter of the disk (D), and Lw
denotes the recirculation length, i.e., the streamwise distance from
the centre of the disk to the position where the streamwise
velocity changes its sign from negative to positive. In the Reynolds
number region considered (i.e., Res r 100), the ow is steady and
axisymmetric (Shenoy and Kleinstreuer, 2008). As shown in
Table 1, only very small variations exist in the results of Cz and
Lw calculated from the cases with different spatial and temporal
resolutions. Moreover, the axis drag coefcient (Cz) and the
recirculation length(Lw) obtained from Mesh 3 are in good agreement with the DNS results by Shenoy and Kleinstreuer (2008) and
the experimental measurements by Roos and Willmarth (1971),
see Fig. 3. Therefore, the numerical approach presented here is
able to give reliable results for the ow over a circular disk.

4. Computational overview
4.1. Computational domain and boundary conditions
A spherical computational domain with a diameter of 40D was
used for the oscillating disk; see Fig. 4(a). The origin of the
coordinates was located at the centre of the computational domain
and the z-axis coincided with the axis of the disk. The mean
position of the disk was at the centre of the computational
domain. On the disk surface, a no-slip boundary condition was
specied for the velocity, and the pressure was set to a zero
normal gradient. At the inlet boundary, a uniform velocity U c ; 0; 0
and a zero normal gradient pressure were specied. At the outlet

56

J. Yang et al. / Ocean Engineering 75 (2014) 5363

Table 1
Results of the axis drag coefcient (Cz) and the recirculation length (Lw) obtained from the cases with different spatial and temporal resolutions for the ow normal to a
stationary circular disk.
Mesh

Elements
n1 =D
tU s =D
Res 10
Res 30
Res 50
Res 75
Res 100

Lw =D

Cz
Mesh 1

Mesh 2

Mesh 3

Mesh 1

Mesh 2

Mesh 3

74 880
0.02
0.01
4.19
2.24
1.77
1.50
1.34

253 440
0.0133
0.0075
4.21
2.25
1.78
1.50
1.34

599 040
0.01
0.005
4.21
2.25
1.78
1.50
1.34

74 880
0.02
0.01
0.41
0.89
1.25
1.62
1.89

253 440
0.0133
0.0075
0.41
0.90
1.26
1.63
1.92

599 040
0.01
0.005
0.41
0.90
1.27
1.64
1.93

Fig. 3. Results of Cz and Lw for the steady ow normal to a stationary circular disk
(t d =D 0:1).

boundary, the pressure was zero, and the velocity was set to a zero
normal gradient. The whole computational domain was discretised with hexahedral elements, and the grids near the surface of
the disk were rened in order to resolve the steeper gradient
there. Fig. 4(b) shows the grids near the disk surface of the disk of
t d =D 0:1, and the meshes for the disks with t d =D 0:05 and 0.02
are created in a similar manner.
The oscillation motion of the disk was realised using the
deforming mesh method, which means that the topology of the
mesh does not change throughout the entire simulation. The grid
points on the disk surface are relocated at each time step, and the
grid points at the far boundary do not move. The grids in the uid
domain are updated in a spring-like way at each time step.
Because the amplitude of the oscillation was small in this study
(a o 0:16D), a good mesh quality was preserved for all the cases.

4.2. Convergence studies

To assess the effects of the spatial and temporal resolutions on
the calculated results, for each aspect ratio (t d =D), three cases with
different grid elements and time steps were investigated using the
parameter choices KC 1 and Vr 0.15. The details of the cases and
the hydrodynamic characteristics are shown in Table 2. As shown,
for each aspect ratio, the results of C x , Cxrms, Czrms, Ca and Cd
obtained from the three cases agree well with each other, indicating a reasonable convergence for the grid resolution and the time
step. Therefore, the medium mesh of 397 440 elements and a time
step of T/5000 were selected for the remainder of the simulations
presented in this study. The maximum CFL (CourantFriedrichs
Lewy) number in this study is approximately 0.3.

Fig. 4. Oscillating disk (t d =D 0:1): (a) computational domain, boundary conditions and (b) grid structures near the disk surface.

5. Results and discussion

In this study, both the KC and Vr numbers were considered at
6 levels with equal steps, i.e., KC 0, 0.2, 0.4, 0.6, 0.8, 1 and Vr 0,
0.05, 0.1, 0.15, 0.2, 0.25. KC 0 represents the stationary disk case,
and Vr 0 represents the case without current. Therefore, 35 sets
of ow parameters were considered for each aspect ratio, except

J. Yang et al. / Ocean Engineering 75 (2014) 5363

the case of KC 0; V r 0. In the following part, the inuences of KC,

Vr and the aspect ratio of disk on the hydrodynamic characteristics of
the disk are presented both qualitatively and quantitatively.

5.1. Flow visualisations

This section presents the visualisations of the ow around a
stationary/oscillating disk of t d =D 0:1 with/without an in-plane
current. The cases under four sets of typical conditions, i.e.,
(a) current only, (b) oscillation only, (c) a constant current velocity
and various oscillation amplitudes and (d) a constant oscillation
amplitude and various current velocities, are presented below.
First, let us consider the cases of a stationary circular disk
(KC 0) under a steady in-plane current. The ow is still steady
and laminar at low Reynolds numbers, i.e., Rec o315. As an
example, visualisations of the steady current around a circular
Table 2
Results of the hydrodynamic coefcients obtained from the cases with different
spatial and temporal resolutions at KC 1 and Vr 0.15.
Mesh
Mesh
Mesh
Mesh
Mesh
Mesh
Mesh
Mesh
Mesh
Mesh

4
5
6
7
8
9
10
11
12

t d =D

Elements

n1 =D

Cx

Cxrms

Czrms

Ca

Cd

0.10
0.10
0.10
0.05
0.05
0.05
0.02
0.02
0.02

117 760
397 440
942 080
117 760
397 440
942 080
117 760
397 440
942 080

0.0055
0.0036
0.0027
0.0055
0.0036
0.0027
0.0055
0.0036
0.0027

T/2500
T/5000
T/7500
T/2500
T/5000
T/7500
T/2500
T/5000
T/7500

4.35
4.38
4.40
5.73
5.73
5.73
9.46
9.49
9.49

0.75
0.78
0.79
1.76
1.76
1.76
3.86
3.89
3.90

5.17
5.18
5.18
5.11
5.11
5.11
5.23
5.23
5.23

5.50
5.50
5.50
10.47
10.49
10.50
25.91
25.94
25.96

2.84
2.87
2.89
3.50
3.52
3.53
4.09
4.11
4.12

57

disk of t d =D 0:1 at Vr 0.25 are shown in Fig. 5. Fig. 5(a) shows

some streamlines near the upper half of the disk, and it appears
that the ow goes around the disk in a laminar way. Fig. 5(b)
shows zvorticity contours in the z 0 plane, and the ow is
observed to be symmetric about the centre plane of the disk
(y0). Similarly, Fig. 5(c) shows the y vorticity contours in the
y0 plane, and the ow is also symmetric about the centre plane
of the disk (z0). Furthermore, there is no evident ow separation
on the surface of the disk.
Fig. 6 shows the instantaneous visualisations of the ow
around an oscillating disk of t d =D 0:1 at KC 0.2, 0.4, 0.6,
0.8 and 1 in the absence of a current (Vr 0), when the disk has
a maximum axis velocity. The gures on the left-hand side of Fig. 6
represent the y vorticity contours in the y0 plane and the
gures on the right-hand side of Fig. 6 represent the vortical
structures around the disk identied by the iso-surface of Q0.1.
Tao and Thiagarajan (2003a) investigated the characteristics of the
ow around an oscillating circular disk attached to the bottom of a
circular cylinder using two-dimensional DNS based on the axisymmetric assumption. The authors reported three distinct vortex
shedding patterns depending on KC number and the aspect ratio
(thickness/diameter) of the disk: the independent, interactive and
uni-directional regimes. Although the present geometric conguration differs signicantly with that used by Tao and Thiagarajan
(2003a), the cases shown in Fig. 6 appear to be in the interactive
ow regime, where the dominant vortex pattern is still symmetric
over one oscillation cycle and both diffusion and convection play
an important role. As shown in Fig. 6, the vortices generated by the
oscillating disk do not leave the disk, and the newly generated
vortex interacts with the previous one. Moreover, the size and
spread of the vortex rings excited by the oscillating disk increase

Fig. 5. Flow visualisations of the in-plane steady current around a stationary circular disk (t d =D 0:1, KC 0; V r 0:25): (a) streamlines, (b) z vorticity contours in the z 0
plane and (c) y vorticity contours in the y 0 plane. The vorticity contours are plotted at 12 levels of z (or y) ( 72, 7 4, 7 8, 7 16, 732, 7 64)U c =D. The solid and dashed
contour lines indicate the positive and negative vorticity contours, respectively.

58

J. Yang et al. / Ocean Engineering 75 (2014) 5363

Fig. 6. Flow visualisations of the oscillating disk (t d =D 0:1) at various KC numbers in the absence of a current (Vr 0) when the disk has a maximum axis velocity: (a, b)
KC 0.2, (c, d) KC 0.4, (e, f) KC0.6, (g, h) KC 0.8, (i, j) KC 1. The gures on the left-hand side represent the vorticity contours in the y 0 plane, which are plotted at 14
levels of y 7 0:25; 7 0:5; 7 1; 7 2; 7 4; 7 8; 7 16U max =2 a. The solid and dashed contour lines indicate the positive and negative vorticity contours, respectively.
The gures on the right-hand side represent the vortical structures around the disk identied by the iso-surface of Q 0.1.

as KC increases. As shown in the gures on the right-hand side of

Fig. 6, the ow is actually axisymmetric about the axis of the disk
for all of the KC numbers considered.
Fig. 7 shows the instantaneous visualisations of the ow
around an oscillating disk of t d =D 0:1 at various KC numbers,
i.e., KC 0, 0.2, 0.4, 0.6, 0.8 and 1, and in a constant in-plane current
(Vr 0.15). The gures on the left-hand side of Fig. 7 represent the
vorticity contours in the y0 plane, and the gures on the righthand side of Fig. 7 represent the vortical structures around the
disk identied by the iso-surface of Q 0.1. As shown in Figs. 5 and
7(a, b), there is no evident vortex generated by the in-plane

current around a stationary disk at the present Reynolds numbers.

For the cases at KC 4 0, vortices are generated at the edges of the
oscillating disk. Compared with the cases in the absence of a
current (see Fig. 6), the ow is no longer symmetric about the axis
of the disk in the steady current condition. Due to the existence of
a current, the vortex shedding at the front edge of the disk is
suppressed, and the vortex generated at the rear edges of the disk
is forced to move downstream (the x-direction in the gures).
Therefore, the ring-like vortex shown in the gures on the righthand side of Fig. 6 is transformed to the worm-like vortices at low
KC numbers (see Fig. 7(d, f)) or the hairpin-like vortices at high KC

J. Yang et al. / Ocean Engineering 75 (2014) 5363

59

Fig. 7. Flow visualisations of the steady constant current (Vr 0.15) around an oscillating circular disk of t d =D 0:1 at various KC numbers when the disk has a maximum axis
velocity: (a, b) KC 0, (c, d) KC 0.2, (e, f) KC 0.4, (g, h) KC 0.6, (i, j) KC 0.8, (k, l) KC 1. The gures on the left-hand side represent the vorticity contours in the y0 plane
which are plotted at 14 levels of y 7 0:5; 7 1; 7 2; 7 4; 7 8; 7 16; 7 32U c =V r D. The solid and dashed contour lines indicate the positive and the negative vorticity
contours, respectively. The gures on the right-hand side represent the vortical structures around the disk identied by the iso-surface of Q 0.1.

numbers (see Fig. 7(h, j, l)). Similar hairpin-like vortices have been
observed in various ow congurations (see, e.g., Johnson and
Patel, 1999; Shenoy and Kleinstreuer, 2008). As shown in the
gures of the vorticity contours on the left-hand side of Fig. 7, the
vortex shedding and convection at the rear side of the disk are
indeed enhanced as KC increases.
Fig. 8 shows the instantaneous visualisations of the ow
around the oscillating disk of t d =D 0:1 with a constant oscillation
amplitude (KC 0.6) and various in-plane currents, i.e., Vr 0, 0.05,
0.1, 0.15, 0.2 and 0.25, when the disk has a maximum axis velocity.
As observed in Fig. 8, the axisymmetric ow around the circular
disk is destroyed by introducing the current. The ow pattern
appears to be signicantly inuenced by the magnitude of current.
The diffusion of the vortices generated by the oscillating disk is
enhanced as Vr increases. Furthermore, the streamwise distance of
the neighboring vortices increases as Vr increases.
5.2. Current drag force coefcient
Fig. 9 shows the results of the time-averaged current drag force
coefcient (C x ) for the disk with t d =D 0:1 as a function of the

reduced current velocity (Vr) at different KC numbers, i.e., KC 0,

0.2, 0.4, 0.6, 0.8 and 1. As shown, C x decreases nonlinearly as Vr
increases for all of the KC numbers considered, which is similar to
the results for low Reynolds number ows around other structures: see, e.g., Roos and Willmarth (1971) for a circular disk or a
sphere and Sumer and Fredse (1997) for a circular cylinder. For a
xed Vr number, C x increases as KC increases in the range of KC r1
because as KC increases, the vortex shedding patterns are
enhanced (see Fig. 7), and the base pressure at the rear side of
the disk is reduced. Then, the drag force acting on the disk is
increased. Furthermore, for the cases where KC o 0:4, the difference between the value of C x for an oscillating disk and that for a
stationary disk (KC 0) at the same Vr number does not vary
signicantly as Vr increases; however, for the cases where
KC 4 0:4, this difference decreases as Vr increases. In other words,
the inuence of KC on C x is dependent on both KC and Vr.
As mentioned in Section 5.1, the current around a stationary
disk remains steady at these Reynolds numbers (i.e., Rec o 315):
therefore, the drag force is steady without uctuation, i.e.,
C xrms 0. However, for an oscillating disk, the drag force uctuates
due to the periodic vortex shedding at the rear side of the disk.

60

J. Yang et al. / Ocean Engineering 75 (2014) 5363

Fig. 8. Flow visualisations of the oscillating disk of t d =D 0:1 at KC 0.6 in the currents with various velocities when the disk has a maximum axis velocity: (a, b) Vr 0, (c, d)
Vr 0.05, (e, f) Vr 0.1, (g, h) Vr 0.15, (i, j) Vr 0.2, (k, l) Vr 0.25. The gures on the left-hand side represent vorticity contours in the y 0 plane, which are plotted at 14
levels of y 7 0:25; 7 0:5; 7 1; 7 2; 7 4; 7 8; 7 16U max =2 a. The solid and dashed contour lines indicate the positive and negative vorticity contours, respectively.
The gures on the right-hand side represent the vortical structures around the disk identied by the iso-surface of Q 0.1.

Fig. 10 shows the variations of the r.m.s. values of Cx versus Vr at

different KC numbers for the disk with t d =D 0:1. It should be
noted that Cxrms is plotted in the form of a relative percentage
with respect to C x . For the cases with small oscillation amplitudes (KC r0:2), the values of C xrms =C x are relatively small
(C xrms =C x o 0:5%) in the range of Vr numbers up to 0.25 and
increase slightly as Vr increases. For the cases at KC Z 0:4, it is
evident that the relative uctuations of Cx increase as KC increases.
It appears that, for a xed KC number in the range of KC Z 0:4, the
C xrms =C x curves have a decreasing trend at low Vr numbers and
have a slightly increasing trend at high Vr numbers in the range of
V r r 0:25. The critical number of Vr appears to increase as KC
increases.

5.3. Added mass coefcient

The results of the added mass coefcient (A) are plotted
against KC in Fig. 11. The results of A for the disks with
t d =D 0:1; 0:05 and 0.02 are shown in Fig. 11(ac), respectively.
As shown in Fig. 11(a), for the disk with t d =D 0:1, the results
of A at the same KC number decrease as Vr increases. Similar
features are also observed for the disks with t d =D 0:05 and 0.02,
see Fig. 11(b, c). For the disks of the aspect ratios of t d =D 0:1; 0:05
and 0.02, the values of A at Vr 0.25 are approximately 1520%
lower than the values of A at Vr 0. For a Spar platform at sea, the
added mass of the heave plates plays an important role in
determining the natural heave frequency of the platform, and this

J. Yang et al. / Ocean Engineering 75 (2014) 5363

61

Fig. 9. Variations of the time-averaged current drag force coefcient (C x ) for the
disk with t d =D 0:1 versus Vr at different KC numbers.

Fig. 10. Variations of relative uctuation of the current drag force coefcient
(C xrms =C x ) versus Vr for different KC numbers for the disk with t d =D 0:1.

frequency increases as the added mass decreases. Therefore, we

expect a high added mass of the heave plates and a low natural
heave frequency of the platform to avoid heave resonance with
incident waves. However, the decrease in the added mass due to
the existence of a current would result in an increased natural
heave frequency, which is harmful to the heave motion suppression of a Spar platform.
For the cases for small currents, e.g., Vr 0.05, the inuence of
the existence of current on A is relatively weak and becomes
stronger as KC increases, see Fig. 11(ac). Moreover, this inuence
becomes stronger as the aspect ratio of the disk decreases.
Fig. 12 shows the variations of A versus KC at Vr 0 and 0.25 for
the disks with three different aspect ratios. For the cases in the
absence of current (i.e., Vr 0), A decreases as t d =D decreases in
the small KC region and A increases as t d =D decreases in the large
KC region. That means the aspect ratio of the disk inuences the
slope of A curve with respect to KC. He (2003) conducted
experimental investigations on the inuence of the aspect ratio

Fig. 11. Results of the added mass coefcient (A) versus KC for different Vr
numbers: (a) t d =D 0:1, (b) t d =D 0:05 and (c) t d =D 0:02.

of the disk on A, and the slope of the A curve was found to

increase slightly as the aspect ratio of the disk decreases. Tao and
Dray (2008) conducted experimental measurements for a solid
oscillating disk with an aspect ratio of 0.02 and found that A
increases as KC increases. For the cases at Vr 0.25, A decreases as
the aspect ratio of the disk decreases within the region of KC from
0.2 to 1.0.

62

J. Yang et al. / Ocean Engineering 75 (2014) 5363

Fig. 12. Inuences of the aspect ratio of the disk on the added mass coefcient (A).

5.4. Damping coefcient

Fig. 13 shows the variations of the damping coefcient (B)
versus KC at different Vr numbers. The results of B for the disks
with t d =D 0:1; 0:05 and 0.02 are shown in Fig. 13(ac),
respectively.
As shown in Fig. 13, it appears that B increases as KC increases
for the cases within the range of Vr from 0 to 0.25. Similar
phenomena have been observed in the experimental measurements by Tao and Dray (2008) for a forced oscillating solid disk in
still water. The authors (Tao and Dray) reported that B increases as
KC increases and that B is linearly dependent on KC in the range of
KC from 0.2 to 1.2. It is reported that there are two critical values
existing in the damping curve for a solid disk oscillating at KC
values of approximately 0.075 and 0.13 (Tao and Thiagarajan,
2003a; He, 2003). The minimum non-zero KC number considered
in this study is 0.2: therefore, this feature is not indicated in Fig. 13.
As shown in Fig. 13(a, b), for the disks with the aspect ratio of
0.1 and 0.05, the inuence of Vr on B appears to be weak at low Vr
values, i.e., V r r 0:15. And the values of B at Vr 0.25 are
signicantly larger than the values of B for the lower Vr numbers
at the same KC number. Therefore, for the disks with the aspect
ratio of 0.1 and 0.05, the inuence of the in-plane current on the
damping of the heave plate appears to take effect with large
current velocities (i.e., V r Z 0:2). However, the features of the B
curves for the disk with the aspect ratio of 0.02 are more
complicated than those for the disks with the other two aspect
ratios and it appears that the small currents would also have a
signicant inuence on the damping coefcient of the disk.
Fig. 14 shows the variations of B versus KC at Vr 0 and 0.25 for
the disks with three different aspect ratios. For the cases in the
absence of current (i.e., Vr 0), B increases as the aspect ratio of
the disk decreases. For the cases at Vr 0.25, a similar feature is
indicated in Fig. 14. It appears that the damping coefcient of a
thin disk is higher than that of a thick disk at the same KC and Vr.
According to the experimental measurement reported by He
(2003), it is reported that there is a signicant reduction in B
with an increase in the aspect ratio of the disk.

6. Conclusions
Direct numerical simulations were performed for a steady
in-plane current around a stationary/oscillating circular disk. The
inuences of three parameters (KC, Vr and the aspect ratio of the
disk) on the ow structures and the hydrodynamic characteristics

Fig. 13. Results of the damping coefcient (B) versus KC for different Vr numbers:
(a) t d =D 0:1, (b) t d =D 0:05 and (c) t d =D 0:02.

of the disk are investigated. The main conclusions are summarised

as follows:
(1) The ow of the steady current around a stationary disk
remains steady and symmetric about the centre-plane (y0 and
z0) of the disk with these ow parameters. For an oscillating
disk in the absence of a current, the ow is symmetric about the
axis of the disk at KC numbers up to 1. However, the existence of
an in-plane current signicantly affects the vortical structures

J. Yang et al. / Ocean Engineering 75 (2014) 5363

63

Acknowledgements
This work was been supported by the National Natural Science
Foundation of China (Grant no. 51239007). The simulations were
performed on TianHe-1(A) at National Supercomputer Center in
Tianjin, China. These supports are gratefully acknowledged. The
authors would also like to thank the editor and anonymous
referees for their comments which have led to a much
improved paper.
References

Fig. 14. Inuences of the aspect ratio of the disk on the damping coefcient (B).

around the disk. Periodic vortex shedding along the rear edges of
the disk was observed.
(2) The characteristics of the current drag force coefcients (Cx)
are inuenced by the oscillation amplitude of the disk. The timeaveraged current drag force coefcient (C x ) decreases as Vr
increases and increases as KC increases. The relative uctuation
of the current drag force coefcient (C xrms =C x ) increases as KC
increases.
(3) The added mass coefcient (A) and the damping coefcient
(B) of the oscillating disk are dependent on KC, Vr and the aspect
ratio of the disk. For the considered cases, A decreases as Vr
increases, and B increases as KC increases. The inuence of Vr on
A is relatively weak at low Vr numbers, i.e., V r o0:05, and this
inuence becomes more signicant as the aspect ratio of the disk
decreases.
It should be noted that this study was initially motivated by the
hydrodynamic characteristics of the heave plates in a Spar platform under current conditions. However, there is still a large
difference between the present ow conguration and a real
heave plate at sea. The hydrodynamic characteristics of heave
plates may also be inuenced by the geometric congurations
(e.g., the plate shape, the plate porosity, and the stiffener and truss
structures), platform motions (e.g., roll and pitch) and some other
ow parameters (e.g., the Reynolds number, the turbulence
intensity and the current proles). Nevertheless, the present
numerical simulations provide some preliminary understandings
of the ow features and the hydrodynamic characteristics of a
heave plate under current conditions. The ow structures, the
current drag forces, the added mass and the damping coefcients
were found to be dependent on KC, Vr and the aspect ratio of the
plate. The results presented in this paper are of signicance
for engineering applications and future investigations on similar
issues.

Downie, M., Graham, J., Hall, C., Incecik, A., Nygaard, I., 2000. An experimental
investigation of motion control devices for truss spars. Mar. Struct. 13, 7590.
He, H., 2003. Hydrodynamics of Thin Plates (Ph.D. thesis). University of Michigan.
Holmes, S., Bhat, S., Beynet, P., Sablok, A., Prislin, I., 2001. Heave plate design with
computational uid dynamics. J. Offshore Mech. Arct. Eng. 123, 2228.
Hunt, J.C.R., Wray, A.A., Moin, P., 1988. Eddies, streams, and convergence zones in
turbulent ows. In: Center for Turbulence Research Report CTR-S88, pp. 193208.
Johnson, T., Patel, V., 1999. Flow past a sphere up to a Reynolds number of 300. J.
Fluid Mech. 378, 1970.
Lake, M., He, H., Troesch, A., Perlin, M., Thiagarajan, K., 2000. Hydrodynamic
coefcient estimation for TLP and Spar structures. J. Offshore Mech. Arct. Eng.
122, 118124.
Magee, A., Sablok, A., Maher, J., Halkyard, J., Finn, L., Datta, I., 2000. Heave plate
effectiveness in the performance of truss Spars. In: Proceedings of the ETCE/
OMAE2000 Joint Conference, New Orleans, LA, pp. 111.
Molin, B., 2001. On the added mass and damping of periodic arrays of fully or
partially porous disks. J. Fluids Struct. 15, 275290.
OpenFOAM, 2009. The Open Source CFD Toolbox, Programmer's Guide, Version 1.6.
OpenCFD Limited.
Prislin, I., Blevins, R., Halkyard, J., 1998. Viscous damping and added mass of solid
square plates. In: 17th International Conference on Offshore Mechanics and
Arctic Engineering. ASME, Lisbon.
Roos, F.W., Willmarth, W.W., 1971. Some experimental results on sphere and disk
drag. AIAA J. 9, 285291.
Sarpkaya, T., Isaacson, M., Isaacson, M., Isaacson, M., 1981. Mechanics of wave forces
on offshore structures. Van Nostrand Reinhold, New York.
Shen, W., Tang, Y., Liu, L., 2012. Research on the hydrodynamic characteristics of
heave plate structure with different form edges of a spar platform. China Ocean
Eng. 26, 177184.
Shenoy, A.R., Kleinstreuer, C., 2008. Flow over a thin circular disk at low to
moderate Reynolds numbers. J. Fluid Mech. 605, 253262.
Sumer, B., Fredse, J., 1997. Hydrodynamics around cylindrical structures, vol. 12.
World Scientic Publishing Company Incorporated.
Tao, L., Cai, S., 2004. Heave motion suppression of a spar with a heave plate. Ocean
Eng. 31, 669692.
Tao, L., Dray, D., 2008. Hydrodynamic performance of solid and porous heave plates.
Ocean Eng. 35, 10061014.
Tao, L., Molin, B., Scolan, Y., Thiagarajan, K., 2007. Spacing effects on hydrodynamics
of heave plates on offshore structures. J. Fluids Struct. 23, 11191136.
Tao, L., Thiagarajan, K., 2003a. Low KC ow regimes of oscillating sharp edges I.
Vortex shedding observation. Appl. Ocean Res. 25, 2135.
Tao, L., Thiagarajan, K., 2003b. Low KC ow regimes of oscillating sharp edges. II.
Hydrodynamic forces. Appl. Ocean Res. 25, 5362.
Thiagarajan, K., Datta, I., Ran, A., Tao, L., Halkyard, J., 2002. Inuence of heave plate
geometry on the heave response of classic spars. In: 21st International
Conference on Offshore Mechanics and Arctic Engineering. ASME.
Thiagarajan, K., Troesch, A., 1998. Effects of appendages and small currents on the
hydrodynamic heave damping of TLP columns. J. Offshore Mech. Arct. Eng. 120,
3742.
Weller, H.G., Tabor, G., Jasak, H., Fureby, C., 1998. A tensorial approach to
computational continuum mechanics using object-oriented techniques. Comput. Phys. 12, 620631.