Ocean Engineering 162 (2018) 341–352

Contents lists available at ScienceDirect

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

Study on the influence of cut-opening ratio, cut-opening shape, and cut- T

opening number on the flow field of a cubic artificial reef
Gang Wanga, Rong Wana,b,c,d,∗, XinXin Wanga, FenFang Zhaoa, XiaoZheng Lana, Hui Chenga,
WeiYao Tanga, QingLong Guana
a
Fishery College, Ocean University of China, Qingdao, 266003, China
b
National Engineering Research Center for Oceanic Fisheries, Shanghai, 201306, China
c
College of Marine Sciences, Shanghai Ocean University, Shanghai, 201306, China
d
Key Laboratory of Sustainable Exploitation of Oceanic Fisheries Resources, Shanghai Ocean University, Ministry of Education, Shanghai, 201306, China

A R T I C LE I N FO A B S T R A C T

Keywords: Upwelling and back eddy are usually selected as obvious indices to study the characteristics of the flow field of
Artificial reef artificial reefs. The influences of cut-opening ratio (OR), cut-opening shape (OS), and cut-opening number (ON)
Cut-opening ratio on the flow field of a cubic artificial reef were investigated. A Reynolds-averaged Navier–Stokes equation was
Cut-opening shape established to solve the flow field of an artificial reef. Based on the finite volume method, the standard k-ε model
Cut-opening number
was used to discretize the Reynolds-averaged Navier–Stokes equation. Flume model experiments were carried
Flow field
out to validate the accuracy of the numerical simulation. A multivariate analysis of variance was used to analyze
the diversities of the OR, OS, and ON. The results showed that the upwelling velocities, including the maximum
upwelling velocity and the volume-weighted average velocity, changed less with the above three factors. The ON
and OR had prominent effects on the volume of upwelling and back eddy, while the OS had the least effect. A
reasonable arrangement of ON and OR can enhance the field of upwelling and back eddy while an excessive OR
or ON can weaken them. This study provided a scientific reference for the optimal design of cubic artificial reefs.

1. Introduction
Globally, coastal ecosystems are suffering from global climate
change, marine environmental pollution, natural habitat degradation,
and overfishing. It has been reported that some coastal ecosystems are
no longer suitable for fishing activities, and many fish populations have
been overexploited or fully exploited in recent decades (FAO, 2014;
García and Leiva Moreno, 2003; Garcia and Newton, 1994; Lundin and
Lindén, 1993; Myers and Worm, 2003; Pauly et al., 2002, 2005). To
prevent this situation, marine ranching and the development of artifi-
cial reefs are important measures to alleviate the decline of fishery
resources, improve the ecological environment of coastal waters, and
provide a natural shelter for marine organisms (Collins et al., 1990;
Jensen et al., 1994; Steimle et al., 2002; Seaman and Sprague, 2013;
Fang et al., 2013).
An artificial reef is a submerged structure deployed on the seabed to
emulate some features of a natural reef such as protecting, con-
centrating and enhancing populations of living marine organisms
(London Convention and Protocol/UNEP, 2009; Kim et al., 2016). After
putting reefs on the seabed, an optical shadow is created around the
reefs, forming a dark area in the water and changing the original light
and shade distribution in the area (He and Shi, 1995). The presence of
artificial reefs affects the spatial distribution of seawater. When the
reefs are put in combination on the seabed, they have a great influence
on the seafloor space and will directly change the seafloor topographic
conditions (Baine, 2001). In addition, the main function of an artificial
reef is to alter the water flow field and circulate the water column much
better. The flow is separated by the reef blockage, and the separated
part of water flows upwards to form an upwelling, which helps to
promote the exchange layer of sea water, increase the bottom and
surface water nutrient transport and enhance the effect of bait diffusion
around reef area (Haro et al., 2004; Lan et al., 2004). Accordingly,
upwelling region has obvious effects in attracting fish. On the other
hand, part of the separated flow will cross the reef to bring in a series of
flow field changes such as vortex and sediment flow, which accelerates
the frequency of seawater exchange in the whole seabed area. Fur-
thermore, behind the reef, there is a small velocity swirl zone that
vortex flow speed is slow and stable. Many studies have indicated that
fish tend to group around reefs because the back-eddy flow provides the
fish with favorable conditions (Bohnsack and Sutherland, 1985; Collins

∗
Corresponding author. College of Marine Sciences, Shanghai Ocean University, Shanghai, 201306, China.
E-mail addresses: wg@stu.ouc.edu.cn (G. Wang), rongwan@ouc.edu.cn (R. Wan).

https://doi.org/10.1016/j.oceaneng.2018.05.007
Received 24 September 2017; Received in revised form 16 March 2018; Accepted 5 May 2018
0029-8018/ © 2018 Elsevier Ltd. All rights reserved.
G. Wang et al.
Fig. 1. Schematic diagram of flow-field velocity effect around artificial reef (m/
s).
et al., 1990; Godoy et al., 2002), the characteristics of back eddy around
the reef is stable and tranquil compared with turbulent flow. From the
structure perspective, porous reef structures are natural breeding ha-
bitats to repair the marine ecosystem as well as to conserve the fisheries
resources. As shown in Fig. 1, complex back eddies and upwelling are
formed when currents encounter an artificial reef.
Much research has focused on the effect of reef shape and structure
on the hydrodynamics and flow field effects of reefs because of the
extreme complications owing to different reef structures and shapes
(Pickering and Whitmarsh, 1997). Because cubic artificial reefs have
simple mechanical structures and low manufacturing costs, many in-
vestigations have been carried out on the flow field characteristics of
cubic reefs by wind tunnel experiments, flume experiments, and nu-
merical simulation technology based on computational fluid dynamics
(CFD). Tang et al. (2007) studied the influence of the velocity of in-
coming currents at different angles on the flow field and the resistance
of cubic reefs by the flume test and found that the resistance of reefs to
incoming current at an angle of 45° was larger than that at an angle of
0°. Liu et al. (2012) concluded that the height and area of the upwelling
regions changed less with various incoming velocities using the cubic
artificial reef model. Liu and Su (2013) and Cui et al. (2011) analyzed
the influence of different pitch arrangement on flow field around cubic
reefs using CFD and a wind tunnel experiment and found that the op-
timum spacing adjustment between cubic reefs was about 1–1.5 times
the length or width of a reef. Shao et al. (2014) and Huang et al. (2014)
conducted numerical simulations on the influences of the cut-opening
ratio (OR) and incoming velocity on the flow field of cubic reefs and
found that the OR had a significant impact on the flow field and the
incoming velocity seemed to have no effect on the flow field of reefs.
Based on the results of a two-way analysis of variance, Fu et al. (2012)
Fig. 2. Artificial reef models (OR 0.2, circular OS).
342
Ocean Engineering 162 (2018) 341–352
studied the influence of the OR and the shape of a reef on the flow field
and showed that the effect of reef shape on the flow field was less
significant than that of the OR. All of the aforementioned works pro-
vided important information for the optimal design and layout of cubic
reefs.
It has been shown that variations of structure, including outline
shape, cut-opening shape (OS), cut-opening ratio (OR), and cut-opening
number (ON) of artificial reefs might have significant effects on the
flow field around reefs, but related research is still lacking. OR is the
ratio of the projected area of the cut-opening hole to the whole pro-
jected area of the reef surface in the direction perpendicular to the flow.
In this study, a statistical method was employed to investigate the
combined influences of cut-opening factors on the flow field of cubic
artificial reefs based on numerical analysis results of CFD by ANSYS
Fluent 15.0. Illuminating the influence of multi-structure factors on
flow fields is instrumental in improving organizational efficiency of cut-
openings with less expenditure by combining appropriate parameters of
OR, OS, and ON. The aim of this study was to provide a scientific re-
ference for the optimal design of cubic reefs.
In order to simplify this work, several hypotheses should be dis-
cussed before carrying out investigations on cut-openings and flow
field. Firstly, the seafloor was assumed to be horizontal and infinity,
without considering the existence of other marine organism on the
seabed. Reefs fell on the seabed without being affected by other objects
during the process of deploying. Secondly, we assumed that the flow
velocity and direction of current in the reef area remained constant, so
the effects of change of the current on the flow field around reefs are
ignored. In reality, after the reefs deployed on the seabed, occasional
stacking or burial around the reefs might occur in the reef area, forming
reef sets or groups and enhancing the flow field effects around the reefs
to some degree (Grove and Sonu, 1985; Yoon et al., 2016). Further-
more, stacking or burial around the reef might lead to the intervals or
spaces around reefs, causing the change of effective usable volume or
facilitate volume (Yoon et al., 2016). But study subject of this paper
focused on impacts of cut-openings of single reef on the flow field, reef
sets or groups were not included for the time being. Finally, considering
the scope of this study, it was assumed that the reef kept stable after
layout in current, although cut-openings of the reef caused the weight
loss of the reef and consequently affected the stability of the reef for
field engineers, such as rolling and slipping on the seabed. For this
reason, it is suggested that the arrangement of cut-openings of reef
should be symmetric as far as possible.
2. Materials and methods
2.1. Materials
In this study, a cubic artificial reef (3.0 m × 3.0 m × 3.0 m) without
a top or bottom cover was used as the prototype design. The height,
width, and length of the cubic reef model were 150 mm with a
G. Wang et al. Ocean Engineering 162 (2018) 341–352

Table 1
Details of model cases.
Case Factor Volume Case Factor Volume
[cm3] [cm3]
ON OS OR ON OS OR

1 1 Square 0.1 573.75 31 4 Octagon 0.1 573.75

2 1 Square 0.2 506.25 32 4 Octagon 0.2 506.25
3 1 Square 0.3 438.75 33 4 Octagon 0.3 438.75
4 1 Square 0.4 371.25 34 4 Octagon 0.4 371.25
5 1 Square 0.5 303.75 35 4 Octagon 0.5 303.75
6 1 Hexagon 0.1 573.75 36 4 Circle 0.1 573.75
7 1 Hexagon 0.2 506.25 37 4 Circle 0.2 506.25
8 1 Hexagon 0.3 438.75 38 4 Circle 0.3 438.75
9 1 Hexagon 0.4 371.25 39 4 Circle 0.4 371.25
10 1 Hexagon 0.5 303.75 40 4 Circle 0.5 303.75
11 1 Octagon 0.1 573.75 41 9 Square 0.1 573.75
12 1 Octagon 0.2 506.25 42 9 Square 0.2 506.25
13 1 Octagon 0.3 438.75 43 9 Square 0.3 438.75
14 1 Octagon 0.4 371.25 44 9 Square 0.4 371.25
15 1 Octagon 0.5 303.75 45 9 Square 0.5 303.75
16 1 Circle 0.1 573.75 46 9 Hexagon 0.1 573.75
17 1 Circle 0.2 506.25 47 9 Hexagon 0.2 506.25
18 1 Circle 0.3 438.75 48 9 Hexagon 0.3 438.75
19 1 Circle 0.4 371.25 49 9 Hexagon 0.4 371.25
20 1 Circle 0.5 303.75 50 9 Hexagon 0.5 303.75
21 4 Square 0.1 573.75 51 9 Octagon 0.1 573.75
22 4 Square 0.2 506.25 52 9 Octagon 0.2 506.25
23 4 Square 0.3 438.75 53 9 Octagon 0.3 438.75
24 4 Square 0.4 371.25 54 9 Octagon 0.4 371.25
25 4 Square 0.5 303.75 55 9 Octagon 0.5 303.75 Fig. 5. Grid generation in computational domain.
26 4 Hexagon 0.1 573.75 56 9 Circle 0.1 573.75
27 4 Hexagon 0.2 506.25 57 9 Circle 0.2 506.25
28 4 Hexagon 0.3 438.75 58 9 Circle 0.3 438.75
29 4 Hexagon 0.4 371.25 59 9 Circle 0.4 371.25
30 4 Hexagon 0.5 303.75 60 9 Circle 0.5 303.75

Fig. 6. Control volume used to illustrate discretization of a scalar transport

equation: a two-dimensional, triangular cell.
Fig. 3. Computational domain of cubic artificial reef.
types of ONs (1, 4, and 9), four types of OSs (square, hexagonal, oc-
tagonal, and circular), and five types of ORs (0.1, 0.2, 0.3, 0.4, and 0.5)
were employed. The details of all model cases are shown in Table 1.

2.2. Numerical simulations

2.2.1. Governing equations and standard k-ε turbulence model

To simplify the simulation, the water in the flow field was assumed
to be an incompressible, steady, and viscous Newtonian fluid. Heat
exchange in the fluid was neglected and the water temperature was
isothermal at 25 °C. A continuity equation (Eq. (1)) and a three-di-
mensional Reynolds-averaged Navier–Stokes equation (Eq. (2)) were
used to calculate the velocity of the water particles.

∂ρ ∂
+ (ρui ) = 0
Fig. 4. Mesh optimization test: Circular mark represents the maximum up- ∂t ∂x i (1)
welling velocity, square mark represents volume of upwelling.
∂ ∂ ∂p ∂ ⎛ ∂ui ⎞ ∂
(ui ) + (ui uj ) = − + ⎜μ ⎟ + (−ui′ u′j )
geometric scale of 1:20, and the thickness of reef wall was 7.5 mm. As ∂t ∂x j ∂x i ∂x j ⎝ ∂x j ⎠ ∂x j (2)
shown in Fig. 2, cut-openings on the side surfaces of the reef were
where ui and uj represent the average velocity (i = 1, 2, 3 represent
symmetrically centered on the centroid, and each side surface had 1-,
components x, y, and z in the Cartesian coordinate system, respec-
4-, or 9-cut openings. In this study, 60 cross-combinations with three
tively), p is the static pressure, ρ is the density of the fluid, μ is the

343
G. Wang et al. Ocean Engineering 162 (2018) 341–352

Fig. 7. Layout of reef model in flume experiment.

Fig. 8. Distribution of measurement points in flume experiment.

dynamic viscosity of the fluid, ui′ and u′j represent the fluctuating ve- above methods. The details of the k and ε equations are shown as fol-
locity, and −ui′ u′j are the Reynolds stresses. lows (incompressible flow, user-defined source terms are not included):
Owing to the different methods dealing with Reynolds stresses,
∂ (ρk ) ∂ (ρkui ) ∂ ⎡⎛ μ ∂k ⎤
turbulence models consisted of the Reynolds Stress Model (RSM) and an + = ⎜ μ + t⎞⎟ + Gk − ρε
eddy viscosity model (EVM). To simplify the calculation under the ∂t ∂x i ∂x j ⎢
⎣⎝ σ ⎥
k ⎠ ∂x j ⎦ (4)
premise of ensuring accuracy, the EVM was adopted to solve the
Reynolds stresses. In the EVM, Reynolds stresses were not directly ∂ (ρε ) ∂ (ρεui ) ∂ ⎡⎛ μ ∂ε ⎤ ε ε2
+ = ⎜ μ + t⎞
⎟ + C1ε Gk − C2ε ρ
solved but were expressed by a function of eddy viscosity. ∂t ∂x i ∂x j ⎢
⎣⎝ σε⎠ ∂x ⎥
j⎦ k k (5)
The viscosity of a turbulent flow was derived from the eddy visc-
where Gk represents the generation of turbulence kinetic energy owing
osity hypothesis of Boussinesq. This assumption established the re-
to the mean velocity gradients; Cμ, C1ε, and C2ε are constants; and σk
lationship between the Reynolds stress and the mean velocity gradient
and σε are the turbulent Prandtl numbers for k and ε, respectively.
(Eq. (3)).
These parameters were calculated as follows (Wang, 2004).

∂u ∂uj ⎞ 2⎛ ∂u ∂u ∂uj ⎞ ∂ui k2

− ρui′ u′j = μt ⎜⎛ i + ⎟ − ⎜ ρk + μt i ⎞ δij
⎟
⎝ ∂x j ∂x i ⎠ 3⎝ ∂x i ⎠
where μt is the turbulent viscosity, μi is the time-averaged velocity, δij is
the Kronecker delta, and k is the turbulent kinetic energy.
Based on the Boussinesq hypothesis, the key to the calculation of
turbulent motion is the determination of μt. In this study, the standard
k-ε model was used to derive μt. The finite volume method (FVM) was
used to discretize the Reynolds-averaged Navier–Stokes equation. Thus,
the specific form of the turbulent flow field was calculated by all of the
Gk = μt ⎜⎛ i + ⎟ , μt = ρCμ , Cμ = 0.09, C1ε = 1.44,
(3) ⎝ ∂x j ∂x i ⎠ ∂x j ε
C2ε = 1.92, σk = 1.0, σε = 1.3
2.2.2. Computational domain and boundary conditions
As shown in Fig. 3, the computational domain of
4.0 m × 1.2 m × 1.0 m (length × width × height) was set in the Car-
tesian coordinate system of which the origin was set at the bottom
center of the reef 1.0 m downstream from the inflow entrance.

344
G. Wang et al.
Fig. 9. Upwelling and back eddy of cubic reef: (a)–(c) front view, side view, and top view of upwelling, respectively; (d)–(f): front view, side view, and top view of
back eddy, respectively.
Boundary conditions were classified as inlet, outflow, and wall as fol-
lows:
(1) Inlet was applied at the front side of the domain, which is the inflow
entrance of the computational domain. Furthermore, k and ε, which
represent turbulent intensity of the entire domain, were initialized
at the inlet.
(2) Outflow was applied at the back side of the domain and used to
model flow exits when the details of the flow velocity and pressure
were not known prior to solving the flow problem. This boundary
condition was applied to the case where the flow was fully devel-
oped at the outlet.
(3) A\ stationary no-slip wall boundary condition was employed at the
bottom of the domain. In addition, the surface roughness coefficient
of reef had impacts on its hydrodynamics and flow filed (Colebrook,
1939; Shockling et al., 2006; Zhang et al., 2017). So the reef sur-
faces were defined as no-slip wall with a roughness coefficient of
0.014, the reinforced concrete's roughness coefficient. In the
345
Ocean Engineering 162 (2018) 341–352
numerical scheme, the roughness coefficient was associated with
roughness height. For a uniform sand-grain roughness, the height of
the sand-grain can simply be taken for roughness height, and the
sand-grain height of reinforced concrete was set to 0.003 m.
(4) A moving no-slip wall boundary condition was applied at the top of
the domain, the free surface of fluid, with a zero shear stress and the
same speed as the inflow.
2.2.3. Grid generation
In this study, tetrahedral unstructured grids were adopted suitable
for the complex geometric features of a reef. As the surfaces of a reef
and the surrounding regions were the focus of our calculation, the grids
on the surfaces of the reef were specially treated. The standard k-ε
model was a high-Reynolds-number turbulent model suitable for a fully
developed turbulent flow field. Nevertheless, because of the low
Reynolds number and slow velocity in the boundary layer near the reef,
the standard k-ε model could not be used to discretize the
Navier–Stokes equation directly. Therefore, in this study, standard wall
G. Wang et al. Ocean Engineering 162 (2018) 341–352

the whole range of Reynolds numbers up to 109. Δy is the height of the

first cell near the reef surface. The y+ of the turbulent flow is in the
logarithm layer when y+ ≥ 11.63. The characteristics of the logarithm
layer are identical to those at the core of the turbulence. Consequently,
the near-wall treatment was carried out if the height of the first layer,
Δy, was arranged in the logarithmic layer and y+ ≥ 11.63 as shown in
Fig. 4. The accuracy and efficiency of the calculation can be improved
by an appropriate mesh size; however, excessively large or small sizes
may lead to a decrease in the accuracy of the calculation or an increase
in the workloads of the workstation. As a consequence, mesh size of
grids was obtained through a mesh optimization test to ensure stable
and convergent calculations (Fig. 4, case 17 for example), volume of
upwelling and the maximum upwelling velocity were selected as the
target indices to carry out the test. It's obvious that two indices got
stable after the gird number reached 5.25 × 106, so the maximum size
of an element in the computational domain and reef surfaces was
0.05 m and 0.001 m, respectively, as shown in Fig. 5. The total number
of nodes and elements were approximately 1.07 × 106 and 5.25 × 106
in each case after meshing, respectively.

2.2.4. Numerical solution

The entire computational domain was composed of cells and ele-
ments discretized by the FVM. A control-volume-based technique was
used to convert a general scalar transport equation to an algebraic
equation that can be solved numerically (ANSYS-Inc, 2013). Dis-
cretization of the governing equations can be illustrated most easily by
considering the unsteady conservation equation for transport of a scalar
quantity φ . To specify the scalar transport equation, a two-dimensional,
triangular cell (Fig. 6) example was used to φ how discretization of this
equation between cells.
Nfaces Nfaces
∂ (ρφ) ⎯→
⎯ ⎯→
⎯
+ ∑ ρf →
vf φf ⋅ Af = ∑ Γφ ∇φf ⋅ Af + Sφ V
∂t f f (12)

where Nfaces is number of faces enclosing cell, φf is the value of φ

Fig. 10. Comparison of Vx and Vy between experimental and numerical simu-
lation results. convected through face f, ρ →
v φ represents mass flux through the face,
f f f
⎯→
⎯
Af is the area of face f, ∇φf is the gradient of φ at face f. V is the cell
functions were used for the near-wall treatment to simulate a low volume. This equation took the same general form as the one given
Reynolds flow in the boundary layer using a set of formulas (Eqs. above and apply readily to multi-dimensional, unstructured meshed
(6)–(10)) to associate parameters in the boundary layer with a Reynolds composed of arbitrary polyhedra. It contained the unknown scalar
number and a reference length. variable φ at the cell center as well as the unknown values in sur-
rounding neighbor cells, such as velocity or pressure. However, face
u
u+ = values φf were required for the convection terms in Eq. (12) and must
uτ (6)
be interpolated from the cell center values. In this paper or similar
u⋅Lbl terrain, this was accomplished using the second-order upwind scheme
Re =
ν (7) (Fang et al., 2011), and the gradient ∇φf was computed by least squares
cell-based method.
Cf = [2lg (Rex ) − 0.65]−2.3 (8) During the CFD simulations, a 3-D pressure-based solver and
SIMPLEC (semi-implicit method for pressure-linked equations-con-
1
τw = Cf ⋅ ρu2 sistent) algorithm were used to iterate the pressure-linked equations
2 (9)
(Yao et al., 2011). Accordingly, the scalar φ that we studied at the cell
Δyρu τ Δy τw centers was obtained. The convergence standard was 10−4 and the
y+ = =
μ ν ρ (10) maximum iteration step was 5000 in this study, i.e., the calculation was
stopped if the residual of continuity, velocity, pressure, turbulent ki-
u τ = (τw / ρ)1/2 (11) netic energy, and energy-dissipation rate fell below 10−4.
where u is the incoming velocity and uτ is the wall friction velocity. τw is
the shear stress of the wall and also a function of the Reynolds number 2.3. Flume model experiments
and the length of the boundary layer (Eqs. (7)–(9)). Lbl is the length of
the boundary layer, and Cf is the skin friction coefficient related to the To verify the accuracy of the numerical calculation results, reef
Reynolds number. For a turbulent boundary layer, Prandtl-Schlichting model experiments were carried out in a circulating tank at the Fishery
skin-friction formulas for a flat plate can be used to estimate Cf with Engineering Hydrodynamics Laboratory at the Ocean University of
varying Reynolds numbers. Schlichting (1979) fitted the relations be- China. The circulating tank was 4.0 m × 1.2 m × 1.0 m
tween the coefficient of local skin coefficient and Reynolds number (length × width × water depth) as shown in Fig. 7. Vectrino current
forming the piecewise and empirical equations by the classification for meters (range: 0–2 m/s, accuracy: ± 0.5%) were equipped to measure
different parts of Reynolds number, according to this, eq. (8) is valid in the flow velocity. The Cartesian coordinate system was set in the flume

346
G. Wang et al. Ocean Engineering 162 (2018) 341–352

Fig. 11. Maximum upwelling velocity changes with three factors: (a) ON = 1, (b) ON = 4, and (c) ON = 9.

Table 2
Numerical simulation data.
Case Data Case Data

Vmax(m/s) v (m/s) Iu Iw Vmax(m/s) v (m/s) Iu Iw

1 0.251 0.213 124.56 5.44 31 0.248 0.214 137.88 6.40

2 0.248 0.213 96.26 4.30 32 0.236 0.213 184.61 6.89
3 0.245 0.213 80.05 3.59 33 0.239 0.212 178.44 4.33
4 0.241 0.213 66.16 2.91 34 0.239 0.212 118.96 2.51
5 0.241 0.213 50.07 2.14 35 0.244 0.212 58.43 1.79
6 0.249 0.213 127.51 5.55 36 0.250 0.214 136.15 6.41
7 0.253 0.213 97.00 4.30 37 0.240 0.213 183.06 6.92
8 0.246 0.213 78.24 3.64 38 0.237 0.212 181.51 4.18
9 0.243 0.213 62.02 2.96 39 0.247 0.212 119.48 2.51
10 0.239 0.213 48.76 2.40 40 0.244 0.212 58.51 1.76
11 0.260 0.213 123.64 5.37 41 0.246 0.214 145.81 6.78
12 0.251 0.213 94.36 4.24 42 0.238 0.213 200.53 7.26
13 0.246 0.213 76.57 3.61 43 0.236 0.212 218.01 4.99
14 0.242 0.213 63.31 2.98 44 0.240 0.212 215.02 1.83
15 0.239 0.213 49.22 2.38 45 0.244 0.211 126.92 0.88
16 0.253 0.213 128.77 5.50 46 0.247 0.214 145.96 6.84
17 0.248 0.213 91.99 4.28 47 0.237 0.213 194.75 6.90
18 0.246 0.213 76.27 3.66 48 0.237 0.212 220.63 4.86
19 0.248 0.213 62.08 2.97 49 0.240 0.212 205.42 1.70
20 0.240 0.213 49.48 2.40 50 0.248 0.211 111.76 0.85
21 0.243 0.213 134.48 6.40 51 0.246 0.214 148.40 6.63
22 0.235 0.213 184.29 7.04 52 0.237 0.213 199.41 7.26
23 0.237 0.212 184.66 4.39 53 0.237 0.212 218.55 4.87
24 0.239 0.212 122.23 2.52 54 0.241 0.212 206.65 1.68
25 0.246 0.212 63.21 1.79 55 0.247 0.211 111.87 0.84
26 0.250 0.214 138.25 6.43 56 0.247 0.214 148.16 6.81
27 0.241 0.213 187.76 6.83 57 0.238 0.213 198.27 7.14
28 0.237 0.212 180.75 4.48 58 0.237 0.212 217.25 4.89
29 0.241 0.212 118.85 2.50 59 0.238 0.212 208.09 1.59
30 0.248 0.212 58.53 1.77 60 0.246 0.211 114.43 0.84

347
G. Wang et al. Ocean Engineering 162 (2018) 341–352

Table 3 2.4. Data statistics

Multivariate analysis of variance.
Source Dependent Type III Sum of Df Mean Square F Value Sig. For a cubic reef with cut-openings or guide plates, 5% relative di-
Variable Squares vergence level was adopted to distinguish velocity of upwelling from
incoming velocity, so a region exceeding 1.05 times the incoming ve-
ON Iu 91717.241 2 45858.621 74.745 0.000
locity was defined as an upwelling (Jiang et al., 2013, 2016) as pre-
Iw 4.976 2 2.488 4.117 0.022
OS Iu 76.951 3 25.650 0.042 0.988
sented in Fig. 9a–c. The back-eddy region is defined as the recirculating
Iw 0.009 3 0.003 0.005 1.000 flow around reefs (Kim et al., 2014) (Fig. 9d–f). To study the properties
OR Iu 57017.981 4 14254.495 23.233 0.000 of a flow field around a cubic reef, typical indicators such as the
Iw 209.445 4 52.361 86.652 0.000 maximum velocity in the upwelling region and the height of the up-
Error Iu 30676.813 50 613.536
welling region were chosen in previous works (Huang et al., 2014; Kim
Iw 30.213 50 0.604
Total Iu 1228161.712 60 et al., 2014). The same maximum velocity in the upwelling region was
Iw 1269.149 60 adopted in this study to describe the characteristics of upwelling. In
addition, the volume-weighted averaged velocity, v was defined to in-
dicate the stability of a turbulent flow and the average disturbance
with the origin located at the bottom center of the cubic reef. The di- caused by ON, OS and OR. In the numerical scheme, v was computed by
rection of the positive axis x was the incoming flow direction, axis y was dividing the summation of the product of velocity in each cell center
the vertical direction, and axis z was perpendicular to the plane com- and cell volume by the total volume of the cell zone based on FVM
posed of axis x and axis y. (ANSYS-Inc, 2013).
As a representative example, a model test of the reef (Case 36 in
n
Table 1) was carried out in the flume tank. As shown in Fig. 8, the ∫V vdV 1
velocity at 15 points was measured; all measurement points were dis-
v =
V
=
V
∑ vi Vi
i=1 (13)
tributed on the mid-vertical plane of the computational domain (z = 0).
Dimensions of the experimental model and the simulation model re- where V is the total volume of upwelling or back eddy, dV is the dif-
mained the same. The cubic reef without a top or bottom cover was ferential element, and v is the upwelling or back eddy velocity of the
constructed of plexiglass with a roughness coefficient of 0.014. differential element, vi is the cell-center velocity in each cell of up-
(roughness height: 0.003 m). welling or back eddy region, |Vi| is the volume of each cell in the up-
To prevent reef model from slippage and rolling at the bottom of the welling or back eddy.
flume, the model was fixed with a smooth thin copper plate at the Influence volume is considered a better indicator for the assessment
bottom of the flume. According to Zheng et al. (2014), the use of a of upwelling and back eddies than influence length and area because
smooth copper plate in a smooth horizontal plane allows the impact of the 1-D reef length and 2-D area parameters may be inaccurate com-
the boundary layer on the flow field to be neglected. The velocity of pared to the influence of volume (3-D) (Huang et al., 2017). For ex-
each measurement point in the horizontal and vertical directions were ample, the height of upwelling or length of a back eddy are usually
measured by Vectrino current meters when the incoming velocity was studied with a mid-vertical plane in the computational domain as a
0.2 m/s. Comparisons between experimental and numerical simulation reference. However, because of the different cut-openings on the side
results are presented in Section 3.1. surfaces in this study, the height of the upwelling or the length of the
back eddy based on the mid-vertical plane may not equal the actual
length. Therefore, 3-D parameters and volume were used to assess the

Fig. 12. Relative divergences between Square OS and others: (a) Iu, (b) Iw.

348
G. Wang et al.
Fig. 13. Impact of OS on upwelling and back eddy when OR = 0.1 and OR = 0.2: (a) (b) Cases when OR = 0.1, and (c) (d) Cases when OR = 0.2.
hydrodynamic characteristics of the upwelling and back eddies. The
efficiency indice of upwelling and the efficiency indice of back eddy
were defined to indicate the volume of upwelling and back eddies
produced by a unit volume of reef, respectively (Kim et al., 2016).
Vup
Iu =
VAR (14)
Vw
Iw =
VAR (15)
where Vup represents the volume of the upwelling, Vw is the volume of
the back eddy, and VAR is the actual volume of the reef. According to
FVM and numerical solution methods described in the above parts,
values of pressure or velocities at the cell centers were acquired.
Benefited from its highly efficient and easy-to-use, CFD-Post was em-
ployed to conduct the postprocessing based on the cell-vertex scheme
and element-based finite volume method (EbFVM) (Filippini et al.,
2014; Joodaki and Ashrafizadeh, 2014; Profito et al., 2015). For the
tetrahedral cell study, the certain cell is regarded as an upwelling or
back eddy finite volume only if the velocities at all nodes of tetrahedral
cell satisfy the value conditions of upwelling or back eddy. Conse-
quently, the volume of back eddy or upwelling was obtained by sum-
ming all these cell volumes meet the criteria (Kim et al., 2016). To
ensure that these indices could be used in aided design of actual reefs in
the future work, efficiency indices of upwelling or back eddy volumes
should be non-dimensionalized. Iu and Iw associate the influence vo-
lumes with the actual reef volumes and express the relation between
flow field effects and cost effects. In other words, larger the indices are,
larger the cost-benefit of reefs are.
Based on the results of the numerical simulation, the influences of
ON, OS, and OR of cubic reefs on Iu and Iw were analyzed through a
multivariate analysis of variance carried out by the SPSS 19.0 statistical
package, and diversities were taken to be significant when P < 0.05.
349
Ocean Engineering 162 (2018) 341–352
3. Results and discussions
3.1. Validation results of flume experiment
Fig. 10 presents comparisons between the numerical results and
experimental results. As shown in Fig. 10, at the front of the reef, the
velocity of each point was about 0.2 m/s, which is consistent with the
incoming velocity. The velocity Vy at points 3 and 9 presented an in-
creased trend while Vx of which were close to 0, it was normal due to
the existence of upwelling in front of reef model. In the back-eddy re-
gion, Vy at points 6, 10, 11, 14, and 15 were negative compared with
others considering the existence of the back eddy. It can be seen that the
simulated results had good agreement with the experimental ones ex-
cept for a few points. Therefore, it can be considered that the numerical
approach by CFD can accurately reflect any change of flow field around
a cubic reef and can be used to simulate the flow field of a cubic reef. In
this study, it is hoped that the influences of ON, OS, and OR on the flow
field of cubic reefs can be analyzed by divergence analysis based on the
results of numerical simulation.
3.2. Results and analysis of numerical simulation
3.2.1. Influence of cut-opening factors on upwelling velocity
The influences of three factors (ON, OS, and OR) on the maximum
upwelling velocity are shown in Fig. 11. When the ON was 1, the
maximum upwelling velocity decreased when the OR increased. This is
consistent with the results obtained by Huang et al. (2014). When the
ON was 4 or 9, the maximum upwelling velocity first decreased and
then increased with an increasing OR. However, it is clear that the OS
had the least impact on the maximum upwelling velocity compared
with the ON and OR. As shown in Table 2, the maximum upwelling
velocity in all cases shown in Table 2 ranged from 0.24 m/s to 0.26 m/s
G. Wang et al.
Fig. 14. Comparison of flow-field velocities (m/s) around reefs: (a) Case 21,
square OS, (b) Case 36, circular OS.
when the incoming velocity was 0.2 m/s.
When the incoming velocity remained unchanged, v in all cases
changed little, ranging from 0.211 m/s to 0.214 m/s as shown in
Table 2. The results indicated that the influences of the three factors on
the average disturbance of upwelling were the same within a margin of
error, the turbulent flow was stable and steady when varying cut-
openings.
Fig. 15. Influence on upwelling and back eddy caused by ON: (a) Efficiency indice of upwelling Iu, (b) Efficiency indice of back eddy Iw.
350
Ocean Engineering 162 (2018) 341–352
3.2.2. Influence of cut-opening factors on volume of upwelling and back
eddy
According to the statistical results in Table 3, at a given confidence
level 95%, it was shown that the ON and OR had a significant impact on
Iu and Iw (P < 0.05), while the OS had the least impact (P > 0.05).
3.2.2.1. Effects of cut-opening shape on volume of upwelling and back
eddy. As previously mentioned, the OS had no significant effect on Iu
and Iw in comparison with the OR and ON, in order to specify the
irrelevance with OR and ON, relative divergences of Iu or Iw with
different OS and same OR and ON were descripted in detail in Fig. 12.
Besides, the cases whose OS were Square were regarded as standard
group (control group). As can be noted from Fig. 12, except for few
points, relative divergences were shown clearly that the maximum
divergence was within 10% between any two group with different OS.
Fig. 13 illustrates the effects of the OS and ON on the efficiency indices
of upwelling and back eddy when the OR equaled 0.1 (a, b) and the OR
equaled 0.2 (c, d). Fig. 14 illustrates the flow field around reefs with a
square OS (Case 21) and circular OS (Case 36) on the mid-vertical plane
(z = 0). It is obvious that Fig. 14a and (b) are identical and provide an
easy-to-understand example of the minimal influence of the OS on the
upwelling and back eddy.
3.2.2.2. Effects of cut-opening number on volume of upwelling and back
eddy. Because the OS had the least impact on the upwelling and back
eddy, Fig. 15 is used to explain how the ON effected Iu and Iw in the
example of a circular cut-opening. Iu obviously increased with an
increase in the ON except when the OR equaled 0.1 (Fig. 15a), and Iw
decreased remarkably when the ON increased or when the OR was 0.4
and 0.5 (Fig. 15b). However, when the OR was less than 0.3, Iw
increased slightly as the ON increased (Fig. 15b). On the basis of
material expense, a proper increase in the ON of a cubic reef is
beneficial to expand the upwelling and back eddy range of the flow
field, especially when the OR is less than or equal to 0.3.
3.2.2.3. Effect of cut-opening ratio on volume of upwelling and back
eddy. Fig. 16 shows the effects of the OR on Iu and Iw. Iu decreased
when the OR increased and the ON was 1, while Iu first increased then
decreased when the OR increased and the ON was 4 or 9. Furthermore,
the maximum Iu occurred when the OR was 0.2 or 0.3, and Iu was
relatively small when the OR attained 0.5 among all cases as shown in
Fig. 16a. In conclusion, it is better when the cases of ON and OR
attained 4–9 and 0.2–0.4, respectively, indicating that an excessive OR
G. Wang et al.
Fig. 16. Influence on upwelling and back eddy caused by OR: (a) Efficiency indice of upwelling Iu, (b) Efficiency indice of back eddy Iw.
may result in the enhanced permeability of a cubic reef. Therefore, the
decrease of Iu was obvious when the OR was relatively large (0.4 and
0.5) compared with the cases in which the OR was small (0.1 and 0.2).
Iw decreased with an increase in the OR when the ON was 1, while Iw
first increased then decreased when the OR increased when the ON was
4 or 9, and the maximum Iw occurred when the OR achieved 0.2. The Iw
when OR was 0.1 or 0.2 was much greater than that when the OR was
0.4 or 0.5, indicating that an excessive OR might result in a large
permeability of a cubic reef for the same reason as described above.
When the OR was relatively large (0.4 and 0.5), the Iw when the ON was
1 was larger than that when the ON was 4 or 9, proving that increasing
the ON is unfavorable for forming a back eddy when the OR is relatively
large. The wake effect is related to the fact that the flow passes directly
through with little reef shading.
4. Conclusion
This study was carried out by numerical simulation and flume ex-
perimental verification. The results of the flume experiments agreed
well with the numerical simulation within a margin of error. A multi-
variate analysis of variance was used to study the influences on up-
welling and back eddies caused by the OR, OS, and ON in the design of
a cubic artificial reef. The conclusions are summarized as follows:
(1) The three factors (OR, OS, and ON) had no significant influence on
the maximum upwelling velocity. When the ON was 1, the max-
imum upwelling velocity decreased as the OR increased, while the
maximum upwelling velocity first decreased, then increased when
the OR increased and the ON was 4 or 9. None of the three factors
had an influence on the volume-weighted average velocity.
(2) The ON and OR had a significant influence on the efficiency indice
of upwelling and the efficiency indice of back eddy, respectively,
while the OS had no significant influence. Considering the reduc-
tion in material expenses and the collocations of reefs, it is re-
commended that an ON of 4 or 9 and an OR of 0.2 or 0.3 is a good
balance for the maximum influence of upwelling and back eddies in
flow fields around cubic artificial reefs.
(3) In terms of the OR and expenses, a proper increase in the ON of a
cubic reef is beneficial to expand the range of upwelling and back
eddies in the flow field. In addition, excessive OR may cause en-
hanced permeability of a cubic artificial reef, resulting in an ob-
vious reduction of Iu and Iw at a high-level of OR.
Previous studies took just a single cut-opening factor into account,
which might cause conclusions about cut-openings partial or incorrect.
351
Ocean Engineering 162 (2018) 341–352
Nevertheless, the purpose of this study laid foundation for future de-
signs and practical application of cubic artificial reef by studying in-
fluences of comprehensive cut-opening factors on flow field effects,
especially variations of velocities indices and efficiency indices of up-
welling and back eddy.
Only three cut-opening factors (e.g., ON, OS and OR) have been
considered in this paper. More cut-opening factors should be considered
in future study, such as symmetry properties of cut-openings layout.
The hydrodynamics and structural mechanics, as well as flow field
characteristics of reefs, are both pivotal parameters in CFD issues,
which can reflect the ecological effects and stability properties of reefs.
Cut-openings of reef might lose its weights, consequently causing reefs
rolling or slipping on the seabed and structural damage under the actual
engineering conditions. Therefore, further research should focus on the
effects of cut-opening arrangements of reefs on stability properties,
hydrodynamics and structural mechanics. In addition, both seabed ef-
fects and flow velocities effects on flow field and hydrodynamics should
be also consummated in the future work.
Acknowledgements
This study was supported by the Specialized Research Fund for the
Doctoral Program of Higher Education (20120132130001) and the
Shandong scientific and technologic development programme
(2011GHY11524). And the numerical simulations in this paper was
finished in the Linux server system in the Lab of Marine Fishery
Techniques, Ocean University of China, Qingdao.
Appendix A. Supplementary data
Supplementary data related to this article can be found at http://dx.
doi.org/10.1016/j.oceaneng.2018.05.007.
References
ANSYS-Inc, 2013. ANSYS Fluent Theory Guide, Release 15.0.
Baine, M., 2001. Artificial reefs: a review of their design, application, management and
performance. Ocean Coast Manag. 44, 241–259.
Bohnsack, J.A., Sutherland, D.L., 1985. Artificial reef research: a review with re-
commendations for future priorities. Bull. Mar. Sci. 37, 11–39.
Colebrook, C.F., 1939. Turbulent flow in pipes with particular reference to the transition
region between the smooth and rough pipe laws. J. Inst. Civ. Eng. 11, 133–156.
Collins, K.J., Jensen, A.C., Lockwood, A.P.M., 1990. Fishery enhancement reef building
exercise. Chem. Ecol. 4, 179–187.
Cui, Y., Guan, C.T., Wan, R., Li, J., Huang, B., 2011. Numerical simulation of influence of
diposal space on effects of flow field around artificial reefs. Trans. Oceanol. Limnol. 2,
59–65 (in Chinese with English abstract).
Fang, L.C., Chen, P.M., Chen, G.B., Tang, Y., Yuan, H.R., Feng, X., 2013. Preliminary
G. Wang et al.
evaluation on resources enhancement of artificial reef in the east corner of Zhelang
Shanwei. Asian Agric. Res. 5, 111–115.
Fang, J., Ling, X., Sang, Z.F., 2011. Experimental and numerical studies of the flow field in
a stirred tank equipped with multiple side-entering agitators. Chem. Eng. Technol.
34, 1619–1629.
FAO, 2014. The State of World Fisheries and Aquaculture. Food and Agriculture
Organization of the United Nations.
Filippini, G., Maliska, C.R., Vaz, M., 2014. A physical perspective of the element-based
finite volume method and FEM-Galerkin methods within the framework of the space
of finite elements. Int. J. Numer. Meth. Eng. 98, 24–43.
Fu, D.W., Luan, S.G., Zhang, R.J., Chen, Y., 2012. Two-way analysis of variance of effects
of cut-opening ratio and surface shape facing flowing in artificial fish reefs on the
flowing field. J. Dalian Ocean Univ. 27, 274–278 (in Chinese with English abstract).
García, S., Leiva Moreno, I., 2003. Global overview of marine fisheries. Conf. Responsible
Fish. Mar. Ecosys. 103–123.
Garcia, S.M., Newton, C.H., 1994. Responsible fisheries: an overview of FAO policy de-
velopments (1945-1994). Mar. Pollut. Bull. 29, 528–536.
Godoy, E., Almeida, T., Zalmon, I., 2002. Fish assemblages and environmental variables
on an artificial reef north of Rio de Janeiro, Brazil. ICES J. Mar. Sci. 59, S138–S143.
Grove, R.S., Sonu, C.J., 1985. Fishing reef planning in Japan. In: D'Itri, F.M. (Ed.),
Artificial Reefs Marine and Freshwater Applications. Lewis Publishers, Chelsea, MI,
USA, pp. 187–251.
Haro, A., Castro-Santos, T., Noreika, J., Odeh, M., 2004. Swimming performance of up-
stream migrant fishes in open-channel flow: a new approach to predicting passage
through velocity barriers. Can. J. Fish. Aquat. Sci. 61, 1590–1601.
He, D.R., Shi, Y.M., 1995. Attractive effect of fish reef model on black porgy (sparus
macrocephalus). J. Xiamen Univ. (Natural Sci) 34, 653–658.
Huang, L.Y., Cheng, H., Tang, Y.L., Yang, Q., Wang, X.X., 2017. Comparison of three ways
to assess the influence range of different artificial reefs. Int. J. Eng. Technol. 9,
105–110.
Huang, Y.D., Fu, D.F., He, W.R., 2014. Three-dimensional numerical simulation on in-
fluence of cut-opening ratio of artificial reef on flow effect. J. Water Resour. Water
Eng. 25, 39–43 (in Chinese with English abstract).
Jensen, A.C., Collins, K.J., Free, E.K., Bannister, R.C.A., 1994. Lobster (Homarus
Gammarus) movement on an artificial reef-The potential use of artificial reefs for
stock enhancement. Crustaceana 67, 198–211.
Jiang, Z., Liang, Z., Liu, Y., Tang, Y., Huang, L., 2013. Particle image velocimetry and
numerical simulations of the hydrodynamic characteristics of an artificial reef. Chin.
J. Oceanol. Limnol. 31, 949–956.
Jiang, Z., Liang, Z., Zhu, L., Liu, Y., 2016. Numerical simulation of effect of guide plate on
flow field of artificial reef. Ocean Eng. 116, 236–241.
Joodaki, A., Ashrafizadeh, A., 2014. Surface shape design in fluid flow problems via
hybrid optimization algorithms. Aero. Sci. Technol. 39, 639–651.
Kim, D., Woo, J., Yoon, H.S., Na, W.B., 2016. Efficiency, tranquillity and stability indices
to evaluate performance in the artificial reef wake region. Ocean Eng. 122, 253–261.
Kim, D., Woo, J., Yoon, H.S., Na, W.B., 2014. Wake lengths and structural responses of
Korean general artificial reefs. Ocean Eng. 92, 83–91.
Lan, C.H., Chen, C.C., Hsui, C.Y., 2004. An approach to design spatial configuration of
artificial reef ecosystem. Ecol. Eng. 22, 217–226.
Liu, T.L., Su, D.T., 2013. Numerical analysis of the influence of reef arrangements on
352
Ocean Engineering 162 (2018) 341–352
artificial reef flow fields. Ocean Eng. 74, 81–89.
Liu, Y., Zhao, Y.P., Cui, Y., Dong, G.H., 2012. Experimental study of the flow field around
cube artificial reef. Ocean Eng. 30, 103–108 (in Chinese with English abstract).
London Convention and Protocol/UNEP, 2009. London Convention and Protocol/UNEP
Guidelines for the Placement of Artificial Reefs, London, UK. pp. 2–9.
Lundin, C.G., Lindén, O., 1993. Coastal manage ecosystems: attempts to manage a
threatened resource. Ambio 22, 468–473.
Myers, R.A., Worm, B., 2003. Rapid worldwide depletion of predatory fish communities.
Nature 423, 280–283.
Pauly, D., Christensen, V., Guénette, S., Pitcher, T.J., Sumaila, U.R., Walters, C.J., Watson,
R., Zeller, D., 2002. Towards sustainability in world fisheries. Nature 418, 689–695.
Pauly, D., Watson, R., Alder, J., 2005. Global trends in world fisheries: impacts on marine
ecosystems and food security. Philos. Trans. R. Soc. Lond. B Biol. Sci. 360, 5–12.
Pickering, H., Whitmarsh, D., 1997. Artificial reefs and fisheries exploitation: a review of
the “attraction versus production” debate, the influence of design and its significance
for policy. Fish. Res. 31, 39–59.
Profito, F.J., Giacopini, M., Zachariadis, D.C., Dini, D., 2015. A general finite volume
method for the solution of the Reynolds lubrication equation with a mass-conserving
cavitation model. Tribol. Lett. 60.
Schlichting, H., 1979. Boundary Layer Theory. MeGraw-Hill, New York.
Seaman, W., Sprague, L.M., 2013. Artificial Habitats for Marine and Freshwater Fisheries.
(Artificial Habitats for Marine and Freshwater Fisheries).
Shao, W.J., Liu, C.G., Nie, H.T., Zheng, D.B., 2014. Analysis of hydrodynamic char-
acteristics and flow field around artificial reefs. Chin. J. Hydrodyn. 29, 580–585 (in
Chinese with English abstract).
Shockling, M.A., Allen, J.J., Smits, A.J., 2006. Roughness effects in turbulent pipe flow. J.
Fluid Mech. 564, 267–285.
Steimle, F., Foster, K., Kropp, R., Conlin, B., 2002. Benthic macrofauna productivity en-
hancement by an artificial reef in Delaware Bay, USA. ICES J. Mar. Sci. [ICES J. Mar.
Sci.] 59, 100–105.
Tang, Y.L., Wang, L., Liang, Z.L., Jiang, Z.Y., Shi, H.W., 2007. Test of the Hydrodynamic
Performance of Square Artificial Reefs. Period, vol. 37. Ocean University China, pp.
713–716 (in Chinese with English abstract).
Wang, F., 2004. Computational Fluid Dynamics Analysis. Tsinghua University Press,
Beijing (in Chinese).
Yao, Y., Wang, H.J., Gong, R.Z., et al., 2011. Numerical study of Francis turbine flow field
under small opening condition. In: Proceedings of 2011 International Conference on
Electronic and Mechanical Engineering and Information Technology. China, pp.
1675–1678.
Yoon, H.S., Kim, D., Na, W.B., 2016. Estimation of effective usable and burial volumes of
artificial reefs and the prediction of cost-effective management. Ocean Coast Manag.
120, 135–147.
Zhang, X., Wang, G., Zhang, M., Liu, H., Li, W., 2017. Numerical study of the aerodynamic
performance of blunt trailing-edge airfoil considering the sensitive roughness height.
Int. J. Hydrogen Energy 42, 18252–18262.
Zheng, Y.X., Liang, Z.L., Guan, C.T., Song, X.F., Li, J., Cui, Y., Li, Q., Shan, X.L., Xu, W.W.,
2014. Numerical simulation and experimental study on flow field of artificial reefs in
three tube-stacking layouts. Oceanol. Limnol. Sinica 45, 11–19 (in Chinese with
English abstract).