You are on page 1of 18

Renewable Energy 126 (2018) 801e818

Contents lists available at ScienceDirect

Renewable Energy
journal homepage: www.elsevier.com/locate/renene

Investigation on aerodynamic performance of vertical axis wind


turbine with different series airfoil shapes
Ying Wang a, b, c, Sheng Shen a, b, Gaohui Li a, b, Diangui Huang a, b, *, Zhongquan Zheng c
a
School of Energy and Power Engineering, University of Shanghai for Science and Technology, Shanghai, 200093, China
b
Shanghai Key Laboratory of Multiphase Flow and Heat Transfer of Power Engineering, Shanghai, 200093, China
c
Centre for Bulk Solids and Particulate Technologies, Newcastle Institute for Energy and Resources, The University of Newcastle, Callaghan, Australia

a r t i c l e i n f o a b s t r a c t

Article history: Numerical research for vertical axis wind turbine with different thicknesses or different cambers of
Received 7 June 2017 airfoils was conducted in this paper. By comparing with the experiment results, the simulation results
Received in revised form were validated. Power coefficients under different tip speed ratios were calculated and compared for
15 January 2018
different types of airfoils: ① symmetrical airfoils with same maximum thickness position, but different
Accepted 18 February 2018
Available online 19 February 2018
maximum thicknesses; ② symmetrical airfoils with same maximum thickness, but different maximum
thickness positions; ③ non-symmetrical airfoils with same maximum thickness, same maximum
thickness position, same maximum camber position, but different maximum cambers; ④ non-
Keywords:
vertical axis wind turbine
symmetrical airfoils with same maximum thickness, same maximum thickness position, same
different airfoil shapes maximum camber, but different maximum camber positions. It was found that power coefficient of
wind energy VAWT showed tendency at first increase and then decrease with the increase of tip speed ratio for
aerodynamic performance symmetrical airfoils with same maximum thickness and same maximum thickness position. Moreover,
numerical simulation when the maximum thickness position moved from the leading edge to trailing edge, the power coef-
ficient firstly increased and then decreased for symmetrical airfoils with the same maximum thickness
but different maximum thickness positions. Besides, with the rise of areas enclosed by curves of surface
pressure coefficient when the azimuthal position increases from 0 to 90 , there showed much larger
work capability.
© 2018 Elsevier Ltd. All rights reserved.

1. Introduction HAWT, VAWTs are omni-directional, accepting wind from any di-
rection without any yawing mechanism [2]. Besides, VAWTs also
In recent years, due to pollution, global warming and shortage of have several well-known advantages: low noise, flexible and
traditional fossil energy, renewable energy resources have become convenient, safe and reliable, nice appearance and simple mainte-
increased significantly, such as biomass energy, solar energy, nance. VAWTs can be used for distributed power supply such as
geothermal energy, hydroelectric energy and wind energy, etc. public lighting in the urban central area, resident families, large
Among these resources, wind energy becomes a research hotspot outdoor advertising, telecommunication base station, oil field,
and hence extensive research efforts have been made to improve highway entire journey supervisory system, cruise ships and vil-
the technology of power generation through wind. lages, etc. VAWT is a good supplementary form of wind energy
Wind turbine mainly can be divided into two types based on utilization which shows broad application prospects [3,4]. How to
rotor structure and its position in airflow: horizontal axis wind improve the wind energy coefficient of VAWT has become a
turbines (HAWT) and vertical axis wind turbines (VAWT). HAWT is research hot point in academy [5,6].
widely utilized for the commercial generation of electricity due to VAWTs can be classified into Darrieus wind turbine, Savonius
many years of research and development [1]. Compared with rotor, combined Savonius and Darrieus rotor, two leaf semi rotary
VAWT, Sistan type wind mill and Zephyr turbine [2]. Darrieus wind
turbine was first designed in 1931 [7]. Among all these VAWTs, the
Darrieus wind turbines show the highest values of efficiency. Lift
* Corresponding author. School of Energy and Power Engineering, University of
Shanghai for Science and Technology, Shanghai, 200093, China.
type wind turbine is the main form of VAWT, where the lift forces
E-mail address: dghuang@usst.edu.cn (D. Huang). acting on the blades of turbine drive the rotor to rotate and hence

https://doi.org/10.1016/j.renene.2018.02.095
0960-1481/© 2018 Elsevier Ltd. All rights reserved.
802 Y. Wang et al. / Renewable Energy 126 (2018) 801e818

generate electricity, and the corresponding wind power coefficient while short chord and long-diameter rotors (i.e., lower solidity)
has already reached about 40%. According to longitudinal section performed better at high TSR. There also existed a maximum power
shape of the rotor, these wind turbines can be divided into H type output with particular pitch angle and helical angle. In order to
and F type. For F type Darrieus wind turbine, since there is quite increase the aerodynamic performance of a Darrieus wind turbine
small bending stress on the blade, it is inconvenient to achieve self- at fixed wind and rotational speeds, Bedon [17] created an opti-
starting and rotating speed control through variable pitch method. mization routine and developed a new airfoil shape. A new airfoil
H type Darrieus wind turbine has advantages including light shape showing a particular configuration of mean line and thick-
weight, simple structure, good balance, as well as relatively higher ness distribution was obtained. By estimating the aerodynamic
wind power coefficient and broader application prospects [8]. coefficients and pressure distributions at different operative con-
Some researchers experimentally investigated the Darrieus ditions, the airfoil aerodynamics was analyzed. Ying [18] numeri-
VAWT with different arrangement of blades. Li [9] studied the in- cally investigated the aerodynamic performance of a novel vertical
fluence of number of blades on aerodynamic forces for a straight- axis wind turbine with adaptive blades, whose shape deformed
bladed VAWT in wind tunnel test. The test airfoil blades were automatically into a desired geometry and thus achieved a better
NACA 0021 with two to five blades. It was found that the power aerodynamic performance. Compared to conventional VAWT, po-
coefficient decreased with the increase of number of blades. wer coefficient of this innovative design was improved by about
Moreover, Li [10] also experimentally studied the effects of solidity 15%. In order to maximize output torque coefficient and output
on aerodynamic forces around straight-bladed VAWT, where the power coefficient of Darrieus VAWT, Hameed [19] designed and
pressure difference substantially decreased with the increase of analyzed a straight bladed VAWT. In order to obtain the 1 kW po-
solidity. Elkhoury [11] conducted experimental investigation of a wer output, design parameters such as solidity, aspect ratio, pres-
three-dimensional micro VAWT with variable-pitch: for high so- sure coefficient, etc. were determined. Besides, at extreme wind
lidity VAWTs, thicker symmetric airfoils performed better; conditions, the blade design was analyzed and designed to attain
compared with airfoils with camber, symmetric airfoils performed the maximum values of deflection and bending stresses. Mohamed
better. However, only NACA 0021 and NACA 0018 were chosen for [20] carried out 20 different symmetric and non-symmetric airfoils
comparison of thickness effect on power coefficient, while only for 2D CFD simulation. Five different series including NACA 00XX,
NACA 0021 and NACA 634221 were chosen for comparison of NACA 63XXX, S-series, A-series and FX-series were considered for
camber effect on power coefficient. El-Samanoudy [12] designed comparison. However, all these airfoils from same series showed
and manufactured a Giromill VAWT, and the performance of this total different thicknesses and cambers and comparison cannot be
VAWT was tested with variation of design parameters including conducted directly. This is because both of virtual camber and
pitch angle, number of blades, airfoil type, turbine radius and chord thickness may have influence on performance of airfoil. Therefore, a
length. It was found that with the increase of number of blades systematic analysis of effects of airfoil shapes on aerodynamic
from two to four, there was noticeable increase in performance, performance is necessary to be conducted.
while there was only small increase in performance by increasing Accurate and efficient computational solvers are required for
the number of blades from three to four. Besides, symmetrical predicting unsteady and highly turbulent flows for VAWT. Un-
airfoil NACA 0024 resulted in higher performance compared with steady Reynolds-averaged Navier-Stokes (URANS) methods are
non-symmetrical airfoil NACA 4420 and NACA 4520. Furthermore, usually applied for wind turbine design. However, turbulence
increasing the chord length had a significant effect on increase of models are often unreliable in predicting complex flow phenomena
performance. such as flow separation, and validation with experiment is crucial
Some researchers conducted numerical simulation for Darrieus [21,22]. Daroczy [23] systematically analyzed and compared the
VAWT with different arrangements of blades. Vassberg [13] made experimentally measured characteristic curves with results of
the first attempt to improve the efficiency of a VAWT design different turbulence models applying CFD computations with
through application of CFD method. Parameters such as solidity, StarCCMþ and Fluent on H-Darrieus rotors. With 2D CFD simula-
TSR and variation of a NACA 0015 airfoil section were studied. The tions for an H-Darrieus rotor, four experimental measurements
research results indicated that items to be addressed can poten- were compared by applying different turbulence models. Finally, k-
tially improve the efficiency for VAWT. At various solidities and ε Realizable and k-u SST model were chosen as best candidates in
rotating speeds, Joo [14] numerically analyzed the aerodynamic 2D simulation. By testing Spalart-Allmaras (with modified strain-
characteristics of an H-Darrieus vertical axis wind turbine with two based formulation), k-u SST with low Reynolds corrections and
straight blades. By changing solidities and rotating speeds, there Transition SST, Gosselin [24] concluded that the k-u SST model was
was significant variation of blockage and interactions (blade-to- the most appropriate one. In all ranges of the TSR, a low Reynolds
flow and -blade), and the flow characteristics of H-Darrieus would regime was identified when the Reynolds number varied between
be then significantly influenced. Shahizare [15] investigated the 4.7  104 and 2.5  105 [16]. Since k-3 and Spallart-Allmaras model
influence of different Omni-direction-guide-vane (ODGV) guide cannot capture and predict the flow development, especially in
vanes angles on the aerodynamic performance of a vertical axis laminar separation bubble [25,26], the k-u SST model is then
wind turbine through 3D numerical simulation. The whole system applied for the laminar region as it can be used as a low Reynolds
including VAWT and shrouding ODGV for 52 angles of the ODGV turbulence model without any extra damping functions [27,28].
under four different tip speed ratios (TSR) was simulated. Accurate Moreover, formed by combining k-u and k-ε models, the shear
ODGV angles under different TSR were selected to achieve the stress transport (SST) formulation shows very good capability for
maximum performance of the VAWT. Lee [16] numerically studied simulating flow with great adverse pressure gradients and sepa-
the aerodynamic performance of a 500 W Darrieus-type vertical- rating flow [29]: k-ε model can be chosen in inlet free-stream
axis wind turbine. Parameters including chord length, helical angle, turbulence properties to avoids k-u problem, while k-u formula-
pitch angle, and rotor diameter were varied to examine the aero- tion can be used in the inner parts of the boundary layer [16].
dynamic characteristics and the separated flow occurring in the Many researchers have conducted numerical research for air-
vicinity of the blade, the interaction between the flow and the foils with different thicknesses or different cambers, as shown in
blade, and the torque and power characteristics. It was found that Table 1. However, systematic analysis for influence of maximum
longer chord length and smaller main diameter (i.e., higher solid- thickness, maximum camber, maximum thickness position and
ity) could increase the power performance of VAWT at low TSR, maximum camber position on aerodynamic performance of VAWT
Y. Wang et al. / Renewable Energy 126 (2018) 801e818 803

Table 1
Summaries of numerical researches for VAWT airfoils with different thicknesses or cambers.

Author TSR Airfoil types Reynolds number

(a) Research for airfoils with different thicknesses


Healy [30] 2.25 upwards NACA 0009, NACA 0012, NACA 0015, NACA 0018 Wide range of Re
Claessens [31] 2.5e5.0 1) NACA 0015, NACA 0018, and NACA 0020 300,000
2) NACA 0018, NACA 63-018, NACA 63-118, NACA 63-020, NACA 63-120 700,000

Kirke [32] 0e4.5 NACA 0012, NACA 0015, NACA 0018, NACA 0021 200,000
Mohamed [20] 2e10 1) NACA 0010, NACA 0015, NACA 0018 and NACA 0021 []
2) NACA 63418, NACA 63415 and NACA 6312
Bedon [33] 1e9 NACA 0015, NACA 0018, NACA 0021 and NACA 0025 100,000e3,000,000
Mei [34] [] 1) NACA 0012, NACA 0015, NACA 0018, NACA 0020 50,000e300,000
2) NACA 0015, NACA 2415, NACA 4415, NACA 6415
Liao [35] 1.04e5.2 NACA 0009, NACA 0012, NACA 0015, NACA 0018 Vin ¼ 10 m/s
Gao [36] NACA 0009, NACA 0012, NACA 0015, NACA 0018 Vin ¼ 10 m/s
Ye [37] 1.5e4.5 NACA 0012, NACA 0015, NACA 0018 Vin ¼ 10 m/s
Sun [38] 2.25 1) NACA 0012, NACA0013, NACA0014, NACA0015, NACA0016, NACA0017, NACA0018, NACA0019, []
NACA0020, NACA0021, NACA0022, NACA0023, NACA0024
2) NACA 0Y15, NACA 1Y15, NACA 2Y15, NACA 4Y15
3) NACA 63W15, NACA 64W15
(b) Research for airfoils with different cambers
Bianchini [39] [] 1) NACA 0018 Vin ¼ 8 m/s
2) 1.42% maximum camber at 50% of chord
3) 3.11% maximum camber at 50% of chord
Claessens [40] 2.5e5.0 1) NACA 0018, NACA 1418, NACA 2418, NACA 4418 300,000
2) NACA 0018, NLF-0018, NLF-0 (0.8)18, NLF-0 (0.8)20, NLF-0118, NLF-0416 700,000
Sabaeifard [41] 1e6 NACA0018, DUW200, S1210 Vin ¼ 10 m/s
Mohamed [20] 2e10 1) NACA 63418, NACA 63415 and NACA 6312 []
2) S-809, S-9000, S-1046, S1014
3) AG18, AH93W174, AH93W215, AH94W301
4) FX66S196, FX77W256, FX71L150, FXL142, FXLV152
Rainbird [42] 2.1, 3.1 NACA 0018 300,000
Danao [43] 2.5e4.0 1) NACA0012, NACA0022 135,000
2) NACA5522, LS0421

hasn't been developed yet. This paper numerically investigates the static mesh which contains less details of flow field, and the un-
aerodynamic performance of VAWT with different blade shapes. structured mesh is then applied; since the rotor rotating domain
Through numerical simulation, the interaction between vortex and includes the blades domain, it is necessary to generate mesh for
wind turbine blades is also analyzed. This paper discusses the blades domain separately. Particularly, the blades domain adopts O
relationship between aerodynamic performance and maximum type structured mesh, and the mesh around the blades is refined to
thickness, maximum camber, maximum thickness position and meet the wall boundary conditions required by turbulence model.
maximum camber position for VAWT. The research outcomes aim The whole computational domain which is 40 times of rotor
to provide promising guidance for the design of Darrieus VAWT. diameter is large enough to ensure that the boundary conditions
have no influence on the computational results. Boundary condi-
2. Numerical methodologies tions for this problem are showed in Fig. 2: velocity inlet for tunnel
inlet, pressure outlet for tunnel outlet, symmetry for tunnel side
2.1. Vertical axis wind turbine model walls and interface for the contact region between rotor and stator.
Fig. 3(a)e(e), shows the gird distribution around airfoils. This
NACA four-digit series are the low speed airfoils established by paper adopts a method by combining structured and unstructured
the United States, and NACA six-digit series are one of the laminar grid, and refined and structured gird is applied around airfoil.
flow aerofoil profiles. The meaning of digits for these two types of Moving mesh methodology and sliding mesh model (SMM) are
airfoils profiles are showed in Fig. 1. used for modeling the rotor rotation around the central axis. Fig. 3
This paper conducted model validation by using experimental (a) illustrates the whole computational domain. The area inside of
results of Darrieus wind turbine from Castelli's [44]. NACA 0021 is the rotating circle in Fig. 3 (b) shows the rotating domain, and it is
adopted and the parameters used in simulation are showed in set as rotor while set up simulation, while the rest of the area
Table 2. The velocity of the inlet and outlet, and non-slip wall are outside of the rotating circle is set as stator. The positions of inlet
adopted as boundary conditions. and outlet are both 40 m away from the rotating center, and it is
20 m away from the center along the vertical axis. The O type
structured grid is applied around the airfoil to ensure the orthog-
2.2. Computational grids
onality of grid cell.

As shown in Fig. 2, the whole computational domain is rectan-


gular and can be divided in to two parts: the external flow field 2.3. Governing equation
domain 1 and the rotor rotating domain 2. Moving mesh meth-
odology and sliding mesh model (SMM) are used for modeling the The numerical simulation is performed with a two dimensional
rotor rotation around the central axis. In order to ensure the mesh unsteady turbulent flow system. Governing equations are [45]:
quality, the whole computational domain is divided into several
parts for mesh generation since the external flow field domain has a Continuity equation:
804 Y. Wang et al. / Renewable Energy 126 (2018) 801e818

2.4. Key performance parameters


 !
V, r V ¼ 0 (1)
Power coefficient, Cp , is a measure of wind turbine efficiency
often used by the wind power industry. Cp equals to the ratio of
Momentum equations: actual electric power produced by a wind turbine divided by the
total wind power flowing into the turbine blades at specific wind
x - momentum: speed.

P
vðruÞ  ! vp vtxx vtyx Cp ¼ (6)
þ V, ru V ¼  þ þ (2) E
vt vx vx vy
where P is the actual electric power produced by a wind turbine, E
is the kinetic energy of the flowing air through the wind turbine.
y - momentum:
1
vðrvÞ  ! vp vtxy vtyy E¼ mv2 (7)
þ V, rv V ¼  þ þ (3) 2
vt vy vx vy
where m is the mass of air, v is the velocity of air. If the volume of air
flowing through cross-sectional area S in per unit time is V, then:

1 1 1 1
The RANS models are widly used for simulation of turbulence. E¼ mv2 ¼ rVv2 ¼ rSvv2 ¼ rSv3 (8)
As one of the RANS models, the SST model developed by Menter 2 2 2 2
[46] can effectively blend the k-u model in the close-wall region Thus, Cp can be represented as:
and the k-ε model in the far field to handle the complex flows with
adverse pressure gradients. The SST k-u model has been widely P P
Cp ¼ ¼ (9)
applied for turbulence modeling of VAWT and shows fair accuracy E 12 rSv3
and efficiency [47e49]. Two mathematical formulas, including k
and u equations, have been proposed in SST methods as below: In international system of units, the units for P, E, r, S and v are
W, W, kg=m3 , m2 and m/s, respectively. According to Betz's law,
! there is a limit value for wind power coefficient. No turbine can
v v v vk
ðrkÞ þ ðrkui Þ ¼ Gk þ Gk  Yk þ Sk (4) capture more than 16/27 (59.3%) of the kinetic energy in wind. The
vt vxi vxj vxj factor 16/27 (0.593) is known as Betz's coefficient, and it is
impossible for Cp to exceed Betz's coefficient:
!
v v v vu 16
ðruÞ þ ðruui Þ ¼ Gu þ Gu  Yu þ Su (5) Cpmax ¼ ¼ 0:593 (10)
vt vxi vxj vxj 27
For VAWT, tip speed ratio (TSR) is the ratio between the
where Gk and Gu express the active diffusivity of k and u, as well Sk tangential speed of the tip of a blade and the actual velocity of the
and Su that are user-defined source terms. In addition, Gk and Gu wind, and it is one of the most important parameters for wind
show the turbulence kinetic energy generation due to mean turbine design. Normally, TSR can be represented by l:
velocity gradients and u, Yk and Yu also mean the dissipation of k
and u due to turbulence [15]. Vt
l¼ (11)
V∞
m
Gk ¼ m þ t
sk where V∞ is the actual velocity of the wind, Vt is the tangential
speed of the tip of a blade.
mt
Gu ¼ m þ 2pn
su Vt ¼ u0 R ¼ R (12)
60
where sk and su are the turbulent Prandtl numbers for k and u, where R is the rotating radius of wind turbine, u0 is the angular
respectively. The turbulent viscosity, mt , is computed as follows: velocity of rotating wind turbine, n is the rotating speed.

rk 1
mt ¼ h i
u max 1* ; SF2 3. Model validation
a a1 u
3.1. Mesh independence study
where S is the strain rate magnitude and
Based on the numerical model set as shown above, this paper
1 conducted the mesh independence study. Firstly, the influence of
sk ¼ F1 ð1F Þ
sk;1 þ sk;2
1 nodes number along the airfoil is considered, and nodes number
including 80, 161 and 250 are selected for simulation, as shown in
Table 3. Fig. 4 shows the mesh near airfoil for all these three cases,
1 and Fig. 5 provides the comparison of torque coefficients under
su ¼ F1 ð1F1 Þ different surface mesh nodes. It can be observed that there is quite
su;1 þ su;2
small difference of torque coefficient between mesh number 161
and 250, while the simulation result with nodes number of 80 is not
Y. Wang et al. / Renewable Energy 126 (2018) 801e818 805

NACA XYZZ NACA 6AB-CZZ

Max
Thickness 6 Series
Camber
Designed Lift
Minimum Coefficient
Maximum
Pressure Location
Camber Position
Range of lift coefficient of B above Max
and below the design lift coefficient Thickness
where a low drag can be maintained
(a) NACA 4 digits series (b) NACA 6 digits series
Fig. 1. NACA 4 digits series and NACA 6 digits series.

Table 2 on k-u SST turbulence model matches better with experimental


Main geometrical features of the tested model. results. Besides, simulation with k-u SST turbulence model show
Items Symbol Value Unit better predicting results than those with S-A turbulence model. In
Castelli's simulation, the maximum power coefficient is about 56%
Rotor diameter Drotor 1030 [mm]
Rotor height Hrotor 1 (2D simulation) [mm] which is too close to the Betz Limit. In this paper, the maximum
Number of blades N 3 [] power coefficient with k-u SST turbulence model is about 37.45%.
Blade profile e NACA 0021 [] Moreover, the overall trend of simulation and value of optimum
Chord length c 85.8 [mm]
TSR are consistent with the experimental results. To sum up, based
Solidity s 0.5 []
on the validation of wind tunnel experimental model, the compu-
tation program and numerical calculation method applied on
simulation of vertical axis wind turbine is reliable.
very satisfactory. The relative error of power coefficient for nodes It is reasonable that the values of power coefficient Cp obtained
number of 161 and 250 is less than 0.28%. Therefore, nodes number with the 2D computer simulation are higher than the values
of 161 is chosen for the following numerical simulation of VAWT in measured in the experiment. In Castelli's experiment [44] whose
this paper. measurement data were used for comparison, a 3D wind turbine
prototype with blades having finite length was tested. However, for
3.2. Comparison between experiment and simulation results a blade of finite length, the pressure difference between the suction
and pressure side tends to compensate at the blade tip. This causes
For 2D numerical simulation of Darrieus vertical axis wind a cross flow over the side edge of the tip which creates a tip vortex
turbine, Fig. 6 conducts a comparison between experiment and that is part of the convected wake [50]. Thus, the blade tip loss
simulation results with S-A, k-ε realizable and k-u SST turbulence effects would reduce lift and power of wind turbine. The lower the
models, along with the simulation results from Castelli [44]. The aspect ratio (the ratio of blade height to chord length) is, the
horizontal axis represents the tip speed ratio, while the vertical axis stronger the blade tip effect shows.
represents the power coefficient. The simulation results from this In the present study, the simulations were conducted in 2D and
paper are closer to the experimental results compared with Cas- it assumed that the Darrieus rotor blades were made of infinite
telli's. Compared with curve tendency calculated from k-ε realizable airfoils joined together adjacently. As a result, there were no tip
turbulence model, curve tendency obtained from simulation based losses and it is reasonable that the 2D simulations showed an over

Symmetry Wall

1
Pressure outlet
Velocity inlet

2
Steady outer domain

Rotating domain
Interface
Symmetry

(a) 2D wind turbine computational domain (b) Mesh topological structure around rotor

Fig. 2. Computational domain and the overall view of the 3-bladed turbine model mesh.
806 Y. Wang et al. / Renewable Energy 126 (2018) 801e818

Interface

(a) Whole domain (b) Grid around three airfoils

(c) Grid around one airfoil (d) Leading edge (e) Trailing edge

Fig. 3. Entire computational domain and the grid near airfoil.

Table 3 won't change the fact that both 2D and 3D CFD simulations have
Mesh nodes and total mesh number adopted in simulation. now been widely used by researchers to assess the power perfor-
No. Nodes number along airfoil Total mesh number Power coefficient mance characteristics of various wind turbine types. We believe
Case 1 80 125,315 30.72%
that even though there exists an understandable degree of differ-
Case 2 161 175,376 37.45% ence between the values of Cp obtained by 3D experiments and the
Case 3 250 218,213 37.73% 2D model, the curve of Cp of the H-Darrieus VAWT obtained from
the numerical solution is in fair agreement with the curve obtained
from the experiment, and thus a 2D model is sufficient to charac-
estimation of the power coefficient Cp when compared to the terize the power output characteristics of a H-Darrieus VAWT.
experimental data obtained by Castelli.
Based on our experience, when the aspect ratio (AR) of the 4. Simulation results and analysis
straight Darrieus rotor blade equals to 1.5, the values for experi-
mental power coefficient Cp of a H-Darrieus turbine is about 70% of By taking NACA 4 digits series foils and NACA 6 digits series foils
those obtained in 2D CFD calculations; while AR equals to 1.0, the as prototypes for VAWT, symmetrical airfoils with different
experimental Cp is about 50% of the 2D predictions. The AR of the maximum thicknesses, symmetrical airfoils with different
Darrieus rotor tested in Castelli's experiments [44] is 1.4. Due to the maximum thickness positions, non-symmetrical airfoils with
effects of the tip vortices present on the real turbine, Cp gained by different maximum cambers and non-symmetrical airfoils with
Castelli should be about 65% of the numerical values predicted by different maximum camber positions are studied in this paper,
our 2D model. Consequently, the percent difference between the respectively. The mechanism of aerodynamic performance for
numerical predictions made using the 2D model and the experi- VAWT is analyzed accordingly.
mental data should fall between 60% and 70%.
As a matter of fact, it is very difficult to get an exact match be-
4.1. Symmetrical airfoils with different maximum thicknesses
tween experiments and numerical results, even in 3D, due to
limited accuracy of different numerical methodologies and due to
Fig. 7 shows the power coefficient of VAWT with different
inevitable simplifications introduced in the model. Nevertheless, it
thicknesses of symmetric NACA 4 digits airfoils. For each type of

(a) 80 (b) 161 (c) 250

Fig. 4. Mesh for airfoil with different nodes number.


Y. Wang et al. / Renewable Energy 126 (2018) 801e818 807

0.25
80 Nodes 161 Nodes 250 Nodes

0.20

0.15
Cm

0.10

0.05

0.00
0 0.25 0.5 0.75 1
t/T
Fig. 5. Comparison of torque coefficients for airfoil with different nodes number.

airfoil, with the increase of TSR, the power coefficient first increases maximum thickness, the maximum value of power coefficient for
to reach an optimum value and then decreases. Besides, when the each type of airfoil starts to decrease. Therefore, in order to obtain
maximum airfoil thickness changes from 6% to 15% for different high power coefficient under different TSRs, NACA 0018 can be the
symmetric NACA 4 digits airfoils, the optimum TSR decreases for best option for TSR which is lower than 2.33, while NACA 0015 can
each type of airfoil. When the maximum airfoil thickness changes be the best choice for TSR which is higher than 2.33.
from 15% to 27%, the optimum TSR maintains almost the same Fig. 8 gives the comparison of vortex structures in flow field
value. For all these types of airfoils, the maximum value of power around the wind turbine with NACA 0006 and NACA 0015 at
coefficient increases at first with the increase of airfoil thickness, different angles of rotation. It can be observed that for NACA 0006
and reaches the highest value of 39.47% when the maximum with thinner airfoil shape, there is relatively large vortex separation
thickness equals to 15%; When continuously increase the and shedding on blades, e.g. on blade A. Apart from thin trailing

Fig. 6. Comparison between experiment in wind tunnel measurement and 2D simulation results for a wind speed of 9 m/s at the test section entrance.
808 Y. Wang et al. / Renewable Energy 126 (2018) 801e818

0.50 0006 0008 rotor disc should be equal to unperturbed ambient static pressure.
0010 0012 This hypothesis is not verified in the present work.
0015 0018 Fig. 10 shows the instantaneous moment coefficient and power
0.40 0021 0023
0025 0027 coefficient of H-rotor Darrieus turbine constructed by NACA 0015
0.30 airfoil under different TSRs. Between the three, the blade with
TSR ¼ 2.64 which produces higher moment coefficient except for
0.20 q ¼ 42 to 69 , 162 to 189 and 282 to 309 . While power coeffi-
Cp (ave)

cient Cp always keeps the highest value when TSR ¼ 2.64 at all
0.10 azimuthal positions compared with other cases.
Fig. 11 shows the power coefficient of VAWT with different
0.00 thicknesses of symmetric NACA 6 digits airfoils. For each type of
airfoil, with the rise of TSR, the power coefficient also first increases
-0.10 to reach an optimum value and then decreases. Moreover, when the
maximum airfoil thickness changes from 6% to 15% for different
-0.20 symmetric NACA 6 digits airfoils, the optimum TSR decreases for
1 2 3 4 5 each type of airfoil; when maximum airfoil thickness changes from
TSR
15% to 27%, the optimum TSR shows almost the same value.
Fig. 7. Power coefficient of VAWT with different thicknesses NACA 4 digits airfoils Furthermore, for all these types of airfoils, the maximum value of
under different TSRs. power coefficient increases at first with the increase of airfoil
thickness, and reaches the highest value when thickness equals to
18%; When continuously increasing the thickness, the maximum
wake, no periodic vortex shedding phenomenon from VAWT with
value of power coefficient starts to decrease. Thus, in order to
NACA 0015 over the entire rotating cycle was observed.
obtain high power coefficient, NACA 63-018 can be chosen for
Fig. 9 displays the distribution of the instantaneous power co-
VAWT blades for TSR which is lower than 2.64, while NACA 0015
efficient as a function of azimuthal positions for the three rotor
can be the best choice for TSR which is higher than 2.64.
blades of NACA 0006 and NACA 0015 when TSR ¼ 2.64. The
Fig. 12 gives the comparison of vortex structures in flow field
contribution of each blade to total rotor performance and three
around the wind turbine with NACA 63-006 and NACA 63-015 at
peaks of the instantaneous power coefficient are also showed. The
different angle of rotation. It can be found that for NACA 63-015,
power coefficient for each blade as well as the total power coeffi-
there is no periodic vortex shedding phenomenon from wind tur-
cient for wind turbine with NACA 0006 is lower than wind turbine
bine with NACA 0015 over the entire cycle of its rotation. While
with NACA 0015. Moreover, the instantaneous power coefficient
relatively large vortex separation and shedding on blades was
locally (close to 96 , 216 and 336 azimuthal position) exceeds the
observed for wind turbine with NACA 63-006, which is similar to
Betz's limit for almost 60 of rotor revolution which shows similar
NACA 0006.
result as Castelli's research, with reaching a maximum value of
Fig. 13 displays the distribution of the instantaneous power
0.625 [44]. As a matter of fact, according to Betz's law, no turbine
coefficient as a function of azimuthal positions for the three rotor
could capture more than 59.3% of the kinetic energy flux associated
blades when TSR ¼ 2.64. The contribution of each blade to total
to wind [51]. Nevertheless, as pointed out by Jonkman [52], among
rotor performance and three peaks of the instantaneous power
the assumptions of Rankine-Froude actuator disc theory, the static
coefficient are also showed. The power coefficient for each blade as
pressure on the boundary of the streamtube portion enclosing the

1 1 3
1
A A 3
3 A 1 A
2
3
2
2
2

(a) NACA 0006

1 1 3
1 3
3 1
2
3
2 2 2

(b) NACA 0015


Fig. 8. Comparison diagram of vortex for VAWT with NACA 0006 and NACA 0015.
Y. Wang et al. / Renewable Energy 126 (2018) 801e818 809

0 Cp [-] total
0
4 356352348
4
12 80.70 356352348 Cp [-] blade1
12 80.30 20 16 344340
20 16 344340 24 336 Cp [-] blade2
24 336 28 332
28 332
32 0.25 328 36
32 0.60 328
324 Cp [-] blade3
36 324
40 0.20 320 44
40
0.50
320
6
316
Betz Limit
44 316
48 312
48 312
52 0.15 308 52 0.40 308
56 304 56 304
0.10 60 300
60 300 0.30
64 296 64 296
0.05
68 292 68 0.20 292
72 0.00 288 72 288
76 284 76 0.10 284
-0.05 80 280
80 280
84 0.00 276
84 -0.10 276
88 272 88 272
-0.15 -0.10
92 268 92 268
96 264 96 264
100 260 100 260
104 256 104 256
108 252 108 252
112 248 112 248
116 244 116 244
120
NACA 0006 240 120 240
124 236 124 NACA 0015 236
128 232 128 232
132 228 132 228
136 224 136 224
140 220 140 220
144 216 144 216
148 212 148 212
152 208 152 208
156 204 156 204
160164 196200 160164
168172176 192 168172176 196200
184188 184188192
180 180

Fig. 9. Evolution of instantaneous power coefficient as a function of blade azimuthal position for the three rotor blades of NACA 0006 and NACA 0015; TSR ¼ 2.64.

TSR=1.44 TSR=1.44
TSR=2.64 TSR=2.64
0.30 0.80
TSR=5.00 TSR=5.00
0.60
0.20

0.40
0.10
Cm

0.20
Cp

0.00
0.00

-0.10
-0.20

-0.20 -0.40
0 30 60 90 120 150 180 210 240 270 300 330 360 0 30 60 90 120 150 180 210 240 270 300 330 360
Azimuth angle Azimuth angle

Fig. 10. Instantaneous moment coefficient and power coefficient of H-rotor Darrieus turbine constructed by NACA 0015 airfoil under different TSRs.

well as the total power coefficient for wind turbine with NACA 63- and then decreases for each type of blades. For all these airfoils with
006 is lower than wind turbine with NACA 63-015. Furthermore, different maximum thickness positions, the power coefficients
the instantaneous power coefficient locally (close to 100 , 220 and reach the optimum values when TSR equals to 2.64. When
340 azimuthal position) exceeds the Betz's limit for almost 60 of maximum thickness position decreases from 60% to 45%, the power
rotor revolution with reaching a maximum value of 0.617. coefficient at different TSRs shows evident increase; when
Fig. 14 shows the instantaneous moment coefficient and power maximum thickness position changes from 45% to 30%, although
coefficient of H-rotor Darrieus turbine constructed by NACA 63-015 there is no obvious increase for power coefficient at different TSRs,
airfoil under different TSRs. Between the three, the blade with the entire curve of power coefficient reach the highest values when
TSR ¼ 2.64 which produces more moment coefficient except for maximum thickness position is 30%; when maximum thickness
q ¼ 52 to 69 , 172 to 189 and 292 to 309 . Compared with other position changes from 30% to 20%, there is only slight decrease for
cases, power coefficient Cp always keeps the highest value when power coefficient at different TSRs. According to Fig. 15, in order to
TSR ¼ 2.64 at all azimuthal positions. obtain high power coefficient, the maximum thickness position
should equal to 30% for VAWT with NACA 0018 airfoil.
4.2. Symmetrical airfoils with different maximum thickness Fig. 16 shows the surface pressure coefficient for NACA 0018
positions airfoil rotating from azimuthal position 0 to 90 with maximum
airfoil thicknesses of 30% and 60%. The pressure surface and suction
Fig. 15 shows the power coefficient of VAWT with different surface show almost no contribution to work at azimuthal position
thickness positions of NACA 0018 airfoils. With the rise of TSR, the 0 under different TSRs. When the azimuthal position increases
power coefficient also first increases to reach an optimum value from 0 to 90 , it can be observed that the blade surface pressure
810 Y. Wang et al. / Renewable Energy 126 (2018) 801e818

0.50 63-006 63-008


63-010 63-012
0.40 63-015 63-018
63-021 63-023
0.30 63-025 63-027

0.20
Cp (ave)

0.10

0.00

-0.10

-0.20
1 2 3 4 5
TSR
Fig. 11. Power coefficient of VAWT with different thicknesses NACA 6 digits airfoils under different TSRs.

1 1 3
1
3
3 1
2
3
2
2
2

(a) NACA 63-006

1 1 3
1 3
3 1
2
3
2 2 2

(b) NACA 63-015


Fig. 12. Comparison diagram of vortex for VAWT with NACA 63-006 and NACA 63-015.

coefficient at the suction side decreases quickly, the pressure dif- when maximum thickness position decreases from 25% to 15%, the
ference between the blade internal and external surfaces increases power coefficient starts to decrease at different TSRs. According to
continuously at the same value of TSR, and the work capability of Fig. 17, in order to obtain high power coefficient, the maximum
blade increases accordingly, especially for TSR ¼ 2.64. Thus, case thickness position should equal to 25% for TSR is lower than 3.09,
with TSR ¼ 2.64 and 30% maximum thickness position shows larger while maximum thickness position should equal to 35% for TSR is
work capability. higher than 3.09.
Fig. 17 shows the power coefficient of VAWT with different Fig. 18 shows the surface pressure coefficient for NACA 630018
thickness positions of NACA 63-018 airfoil. Except three airfoils airfoil rotating from azimuthal position 0 to 90 with maximum
whose maximum thickness positions are 15%, 55% and 60% and the airfoil thicknesses of 25% and 60%. Similar as case with NACA
corresponding optimum TSR is 3.09, all the other airfoils show a 0018 at azimuthal position 0 under different TSRs, the pressure
different optimum TSR which equals to 2.64. For airfoils whose coefficient on pressure surface and suction surface show almost no
maximum thickness position decreases from 60% to 25%, the power contribution to work. When the azimuthal position increases from
coefficient gradually increases at different values of TSRs. However, 0 to 90 , it can be seen that the blade surface pressure coefficient
Y. Wang et al. / Renewable Energy 126 (2018) 801e818 811

0
Cp [-] total
0
4
12 8 0.25 356352348 4
12 8 0.7 356352348 Cp [-] blade1
20 16 344340
20 16 344340
28
24 336
332 24 336 Cp [-] blade2
28 332
36
32 0.2 328
324 36
32 0.6 328
324
Cp [-] blade3
40 320 40 320
0 Betz Limit
44 0.15 316 44 0.5 316
48 312 48 312
52 0.1 308 52 308
0.4
56 304 56 304
60 0.05 300 60 300
0.3
64 296 64 296
68 0 292 68 292
0.2
72 288 72 288
76 -0.05 284 76 0.1 284
80 280 80 280
84 -0.1 276 0
84 276
88 272 88 272
-0.15 -0.1
92 268 92 268
96 264 96 264
100 260 100 260
104 256 104 256
108 252
108 252
112 248
116 NACA 0006 244
112 248
116 244
120 240
120 240
124 236
128 232
124 NACA 0015 236
128 232
132 228
136 224 132 228
140 220 136 224
144 216 140 220
148 212 144 216
152 208 148 212
156 204 152 208
160164 196200 156 204
168172176 160164 196200
184188192 168172176
180 184188192
180

Fig. 13. Evolution of instantaneous power coefficient as a function of blade azimuthal position for the three rotor blades of NACA 63-006 and NACA 63-015; TSR ¼ 2.64.

on the suction side decreases rapidly, the pressure difference be- 0.50
tween the blade internal and external surfaces rises continuously at
the same value of TSR, and the work capability of blade goes up 0.40
accordingly, particularly for TSR ¼ 2.64. Thus, case with TSR ¼ 2.64
and 25% maximum thickness position shows higher work 0.30
capability.
0.20
Cp (ave)

4.3. Non-symmetric airfoils with same thickness, same maximum 0.10


thickness position, same maximum camber position, but different 20% 25%
0.00
maximum cambers 30% 35%
-0.10 40% 45%
Fig. 19 shows the power coefficient of VAWT with different 50% 55%
cambers of NACA 63A612 airfoils. For airfoils whose maximum -0.20 60%
cambers changing from 0.81% to 4.36%, the corresponding opti-
mum TSR is 3.30; for airfoils whose maximum cambers are 4.95% -0.30
and 5.54%, the corresponding optimum TSR decreases to 3.09; for 1 2 3 4 5
TSR
airfoils whose maximum cambers changing from 6.13% to 7.71%, the
corresponding optimum TSR continuously decreases to 2.51. Thus, Fig. 15. Power coefficient of VAWT with different maximum thickness positions of
with the increase of camber, the optimum TSR goes down NACA 0018 airfoil.

TSR=1.44 TSR=1.44
TSR=2.64 TSR=2.64
0.30 0.80
TSR=5.00 TSR=5.00
0.60
0.20
0.40

0.10 0.20
Cp
Cm

0.00
0.00
-0.20
-0.10
-0.40

-0.20 -0.60
0 30 60 90 120 150 180 210 240 270 300 330 360 0 30 60 90 120 150 180 210 240 270 300 330 360
Azimuth angle Azimuth angle

Fig. 14. Instantaneous moment coefficient and power coefficient of H-rotor Darrieus turbine constructed by NACA 63-015 airfoil under different TSRs.
812 Y. Wang et al. / Renewable Energy 126 (2018) 801e818

10 10 10 10 0018-30%

pressure coefficient (Cp)


pressure coefficient (Cp)
pressure coefficient (Cp)

pressure coefficient (Cp)


5 5 5 5 0018-60%
0 0 0 0
-5 -5 -5 -5
-10 -10 -10 -10
-15 -15 -15 -15
-20 -20 -20 -20
-25 0018-30% -25 0018-30% -25 0018-30% -25
-30 0018-60% -30 0018-60% -30 0018-60% -30
-35 -35 -35 -35
-40 -40 -40 -40
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
x/c x/c x/c
x/c
(a) TSR=1.44
0018-30%
10 10 10 10

pressure coefficient (Cp)


pressure coefficient (Cp)

pressure coefficient (Cp)


pressure coefficient (Cp)

0018-60%
0 0 0 0
-10 -10 -10 -10
-20 -20 -20 -20
-30 -30 0018-30%
0018-30% -30 0018-30% -30
-40 0018-60% -40 0018-60%
-40 0018-30% -40
-50 -50
-50 0 0.2 0.4 0.6 0.8 1 -50
0 0.2 0.4 0.6 0.8 1
0 0.2 0.4 0.6 0.8 1 x/c 0 0.2 0.4 0.6 0.8 1
x/c
x/c x/c
(b) TSR=2.64
40 40 40 40
pressure coefficient (Cp)
pressure coefficient (Cp)

pressure coefficient (Cp)


pressure coefficient (Cp)

30 30 30 30
20 0018-30% 0018-30% 0018-30%
20 0018-30% 20 20
10 10 0018-60% 10 0018-60% 10 0018-60%
0018-60%
0 0 0 0
-10 -10 -10 -10
-20 -20 -20 -20
-30 -30 -30 -30
-40 -40 -40 -40
-50 -50 -50 -50
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
x/c x/c x/c
x/c
(c) TSR=5.00
Fig. 16. Pressure coefficient on NACA 0018 airfoil rotating from azimuthal position 0 to 90 with maximum thicknesses of 30% and 60%.

constantly for NACA 63A612 airfoils. Moreover, the power coeffi-


cient almost decreases constantly with the increase of camber at
0.50
different values of TSR. The outcome indicates that unsymmetrical
0.40 airfoils show lower power coefficient compared with symmetrical
airfoils.
0.30 Fig. 20 shows the surface pressure coefficient for NACA 63A612
airfoil rotating from azimuthal position 0 to 90 with maximum
0.20 airfoil cambers of 0.81% and 7.71% when TSR equals to 3.30 and
Cp (ave)

5.00. When the azimuthal position increases from 0 to 90 , it can


0.10
be found that the blade surface pressure coefficient at the suction
0.00 15% 20% side decreases dramatically, the pressure difference between the
25% 30% blade internal and external surfaces rises continuously at the same
-0.10 35% 40% value of TSR, and then the work capability of blade increases. At the
45% 50% same azimuthal position with different values of TSR, the areas
-0.20 enclosed by curves of surface pressure coefficients are much larger
55% 60%
for case with 0.81% maximum camber at azimuthal position of 90 ,
-0.30
and there shows much larger work capability.
1 2 3 4 5
TSR Fig. 21 shows the power coefficient of VAWT with different
cambers of NACA 4415bis airfoils. For airfoils whose maximum
Fig. 17. Power coefficient of VAWT with different maximum thickness positions of cambers varying from 0.61% to 4.16%, the corresponding optimum
NACA 630018 airfoil.
TSR is 2.51; for airfoils whose maximum cambers changing from
Y. Wang et al. / Renewable Energy 126 (2018) 801e818 813

10 10 10 10
pressure coefficient (Cp)

pressure coefficient (Cp)


5 5

pressure coefficient (Cp)

pressure coefficient (Cp)


5 5
0 0 0 0
-5 -5 -5 -5
-10 -10 -10 -10
-15 -15 -15 -15
-20 -20 -20 -20
-25 630018-25% -25 630018-25% -25 -25
-30 630018-25%
-30 630018-65% -30 -30 630018-25%
630018-25% -35 630018-65%
-35 -35 -35 630018-60%
-40 -40 -40 -40
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
x/c x/c x/c x/c
(a) TSR=1.44
20 20 20 20 630018-25%
pressure coefficient (Cp)

pressure coefficient (Cp)


630018-60%
pressure coefficient (Cp)

pressure coefficient (Cp)


10 10 10 10
0 0 0 0
-10 -10 -10 -10
-20 -20 -20 -20
-30 -30 -30 -30
-40 -40 -40 -40
-50 630018-25% -50 630018-25% -50 -50
630018-25% -60
-60 630018-60% -60 -60 630018-60%
630018-60% -70
-70 -70 -70
-80 -80 -80 -80
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
x/c x/c x/c x/c
(b) TSR=2.64
40 40 40 40 630018-25%
pressure coefficient (Cp)

630018-25%
pressure coefficient (Cp)

630018-25% 30 630018-60%
pressure coefficient (Cp)

pressure coefficient (Cp)


30 630018-25% 30 30
20 630018-60% 20 630018-60% 20 630018-60% 20
10 10 10 10
0 0 0 0
-10 -10 -10 -10
-20 -20 -20 -20
-30 -30 -30 -30
-40 -40 -40 -40
-50 -50 -50 -50
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
x/c x/c x/c x/c
(c) TSR=5.00
Fig. 18. Pressure coefficient on NACA 630018 airfoil rotating from azimuthal position 0 to 90 with maximum thicknesses of 30% and 60%.

4.75% to 5.94%, the corresponding optimum TSR increases to 3.09; corresponding optimum TSR continuously increases to 3.30. Thus,
for airfoils whose maximum cambers are from 6.53% to 7.71%, the with the increase of camber, the optimum TSR rises constantly for
NACA 63A612 airfoils. Additionally, the power coefficient decreases
almost constantly with the increase of camber at different values of
0.40 TSR. The results also mean unsymmetrical airfoils have lower po-
wer coefficient compared with symmetrical airfoils.
0.30 Fig. 22 presents the surface pressure coefficient for NACA 4415
airfoil rotating from azimuthal position 0 to 90 with maximum
0.20 airfoil cambers of 0.61% and 7.71% when TSR equals to 2.51 and
5.00. When the azimuthal position increases from 0 to 90 , it can
0.10
be observed that the blade surface pressure coefficient at the suc-
Cp (ave)

0.00 tion side decreases dramatically, and the pressure difference be-
tween the blade internal and external surfaces rises continuously at
-0.10 0.81% 1.40% the same value of TSR which shows increasing work capability.
1.99% 2.58%
-0.20 3.18% 3.81%
4.4. Non-symmetric airfoils with same thickness, same maximum
4.36% 4.95%
5.54% 6.13% thickness position, same maximum camber, but different maximum
-0.30
7.12% 7.71% camber positions
-0.40
1 2 3 4 5 Fig. 23 shows the power coefficient of VAWT with different
TSR camber positions of NACA 4X15. The camber of NACA 4015 actually
is 0%. Except for NACA 4015, all the other airfoils in Fig. 13 have 40%
Fig. 19. Power coefficient of VAWT with different cambers of NACA 63A612 airfoil.
camber at positions varying from 10% to 90%. Among all these
814 Y. Wang et al. / Renewable Energy 126 (2018) 801e818

20 20 20 20

pressure coefficient (Cp)


pressure coefficient (Cp)
pressure coefficient (Cp)

pressure coefficient (Cp)


10 10 10 10
0 0 0 0
-10 -10 -10 -10
-20 -20 -20 -20
-30 -30 -30 -30
-40 -40 -40 -40
-50 63A612-0.81% -50 63A612-0.81% -50 63A612-0.81% -50 63A612-0.81%
-60 63A612-7.71% -60 63A612-7.71% -60 63A612-7.71% -60 63A612-7.71%
-70 -70 -70 -70
-80 -80 -80 -80
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
x/c x/c x/c x/c

(a) TSR=3.30
40 40 40 40 63A612-0.81%
pressure coefficient (Cp)
pressure coefficient (Cp)

pressure coefficient (Cp)


pressure coefficient (Cp)
30 30 30 30 63A612-7.71%
20 20 20 20
10 10 10 10
0 0 0 0
-10 -10 -10 -10
-20 -20 -20 -20
-30 -30 -30 -30
-40 63A612-0.81% -40 63A612-0.81% -40 63A612-0.81% -40
-50 -50 -50 63A612-0.81% -50
-60 63A612-7.71% -60 63A612-7.71% -60 7.71% -60
-70 -70 -70 -70
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
x/c x/c x/c x/c
(b) TSR=5.00
Fig. 20. Pressure coefficient on NACA 63A612 airfoil rotating from azimuthal position 0 to 90 with maximum cambers of 0.81% and 7.71%.

0.45 leading edge to trailing edge. Based on the above analysis, 40%
to 70% camber positions would be recommended among this group
0.35 of NACA 4X15 airfoils.
Figs. 26 and 27 show the airfoil shape and power coefficient of
0.25 VAWT with different pitch angles of NACA X015 where all these
airfoils are symmetrical. Except for NACA 0015, with the increase of
0.15 pitch angle, the curve of power coefficient shows a decrease ten-
Cp (ave)

dency under all values of TSR, as shown in Fig. 27. Particularly, when
0.05 TSR is between 1.40 and 2.04, the curve of power coefficient for
0.61% 1.20% NACA 0015 without pitch angle shows lower value compared with
1.79% 2.38% airfoils with pitch angles including NACA 1015, NACA 2015, NACA
-0.05 2.98% 3.69%
4015, NACA 6015 and NACA 8015; when TSR is larger than 2.64, the
4.16% 4.75%
5.34% 5.94% symmetrical airfoil NACA 0015 still shows higher power coefficient.
-0.15
6.53% 7.12% Rezaeiha [53] also found similar result that a small negative pitch
7.71% angle could increase turbine CP compared to symmetrical airfoil.
-0.25
1 2 3 4 5 Figs. 28 and 29 show the power coefficient of VAWT with
TSR different camber positions of NACA 63215 and NACA 63415 who has
the same maximum thickness 15% but different maximum cambers
Fig. 21. Power coefficient of VAWT with different cambers of NACA 4415bis airfoil.
of 1.10% and 2.21% at the same position of 50%. For all these airfoils,
the whole curves of power coefficient rises generally with the in-
crease of camber positions and then decrease. Besides, with similar
airfoils, NACA 4015 shows the highest power coefficients at maximum camber position, the power coefficients for NACA 63215
different values of TSR as the airfoil is symmetrical. For all the other series airfoils are all higher than NACA 63415 series airfoils.
airfoils, with the increase of camber positions, the optimum TSR Therefore, it can be summarized that power coefficient of VAWT is
gradually goes down from 3.30 to 2.51. Moreover, the whole curve higher when the maximum camber is smaller and the foil is more
of power coefficient rises first and then decrease with the increase symmetrical.
of camber positions from 10% to 90%. Especially for NACA 4815 and Based on the above analysis, the best solution for wind turbine
NACA 4915, the power coefficient decreases dramatically when TSR blade selection under various values of TSRs has been summarized,
is higher than 2.64. As shown in Fig. 24, NACA 4115, NACA 4215, as shown in Table 4, which can provide guidance for selecting
NACA 4815, and NACA 4915 are not regular airfoils shapes and can proper airfoil shapes while designing vertical axis wind turbine.
be neglected, and the rest of the curves of power coefficient are
presented in Fig. 25. It can be observed that except for NACA 4015
5. Conclusions
which is symmetrical, the values of power coefficient for rest of the
airfoils increase when maximum camber positions moving from
This paper conducted numerical research on aerodynamic
Y. Wang et al. / Renewable Energy 126 (2018) 801e818 815

20 20 20 20 4415bis-0.61%
Pressure coefficient (Cp)

Pressure coefficient (Cp)


4415bis-7.71%

Pressure coefficient (Cp)


Pressure coefficient (Cp)
10 10 10 10
0 0 0 0
-10 -10 -10 -10
-20 -20 -20 -20
-30 -30 -30 -30
-40 4415bis-0.61% 4415bis-0.61% 4415bis-0.61%
4415bis-7.71% -40 -40 -40
-50 4415bis-7.71% 4415bis-7.71%
-50 -50 -50
-60 -60 -60 -60
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
x/c x/c x/c x/c
(a) TSR=2.51
4415bis-0.61%
30 30 30 4415bis-7.71%
Pressure coefficient (Cp)

Pressure coefficient (Cp)


Pressure coefficient (Cp)

30

Pressure coefficient (Cp)


10 10 10 10
-10 -10 -10 -10
-30 -30 -30 -30
-50 -50 4415bis-0.61% -50 -50
4415bis-0.61% 4415bis-0.61%
-70 4415bis-7.71% -70 -70
-70 4415bis-0.61% 4415bis-7.71%
-90 -90 -90 -90
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
x/c x/c x/c
x/c
(b) TSR=5.00
Fig. 22. Pressure coefficient on NACA 4415bis airfoil rotating from azimuthal position 0 to 90 with maximum thicknesses of 30% and 60%.

0.45
0.40
0.35
0.30
0.25
Cp (ave)

0.20
0.15
0.10
0.05 4015 4315 4415
0.00 4515 4615 4715

-0.05
1 2 3 4 5
TSR
Fig. 25. Power coefficient of VAWT with different camber positions of selecting NACA
4X15.

Fig. 23. Power coefficient of VAWT with different camber positions of NACA 4X15.

positions, the power coefficients of VAWT under different values of


performance for VAWT with different NACA 4 digits and 6 digits TSRs are compared, and the following conclusions are obtained:
series airfoils. By changing maximum thicknesses, maximum
thickness positions, maximum cambers and maximum camber 1) For NACA 4 digits and 6 digits symmetric series airfoils with
same thickness and same position of maximum thickness, the

(a) 4115 (b) 4215 (c) 4815 (d) 4915

Fig. 24. Airfoil shapes of NACA 4115, NACA 4215, NACA 4815 and NACA 4915.
816 Y. Wang et al. / Renewable Energy 126 (2018) 801e818

(a) NACA 1015 (b) NACA 2015 (c) NACA 4015

(d) NACA 6015 (e) NACA 8015 (f) NACA 9015

Fig. 26. Airfoil shape of NACA 1015, NACA 2015, NACA 4015, NACA 6015, NACA 8015 and NACA 9015.

0.50 0.40
0.35
0.40
0.30

0.30 0.25
0.20
Cp (ave)

Cp (ave)
0.20 0.15
19.8% 24.7%
0.10 29.6% 35.5%
0.10 0015 1015 40.4% 45.4%
2015 4015 0.05 55.5% 59.6%
6015 8015 64.8% 70.0%
0.00 0.00 79.9% 75.3%
9015 85.3%
-0.05
-0.10 1 2 3 4 5
1 2 3 4 5 TSR
TSR Fig. 29. Power coefficient of VAWT with different camber positions of NACA 63415.
Fig. 27. Power coefficient of VAWT with different pitch angles for NACA X015.

3) At the same azimuthal position with different values of TSRs, the


power coefficient of VAWT show tendency at first increases and areas enclosed by curves of surface pressure coefficients are
then decreases with the increase of TSR. much larger for case with lower maximum camber when blade
2) For symmetrical airfoils with same maximum thickness but rotating from azimuthal position of 0 to 90 , and there shows
different positions of maximum thickness, the power co- much larger work capability.
efficients firstly increases and then decreases when the 4) For symmetrical airfoils with the same maximum thickness,
maximum thickness position moves from the leading edge to except for NACA 0015, the power coefficients decrease at all
trailing edge. values of TSRs with the increase of pitch angle.

0.40

0.35

0.30

0.25
Cp (ave)

0.20
20.3% 25.1%
0.15 29.9% 35.7%
40.6% 45.5%
0.10 55.4% 59.7%
65.6% 70.0%
0.05 75.3% 79.9%
85.3%
0.00
1 2 3 4 5
TSR
Fig. 28. Power coefficient of VAWT with different camber positions of NACA 63215.
Y. Wang et al. / Renewable Energy 126 (2018) 801e818 817

Table 4
Recommendation of flapping hydrofoil profiles.

Prototype Suggested wind turbine blade types

Maximum Maximum thickness Maximum Maximum camber position Maximum pitch angle
thickness position camber

1 NACA 00XX 15% e 18% [] [] [] []


NACA 63 (3)0XX 15% e 18% [] [] [] []
2 NACA 0018 [] 30% [] [] []
NACA 63018 [] 25% to 35% [] [] []
3 NACA 63A612 [] [] 0.81% (the smaller [] []
the better)
NACA 4415bis [] [] 0.61% (the smaller [] []
the better)
4 NACA 4X15 [] [] [] 40%e70% (except for NACA 4015) []
NACA 63215 [] [] [] 35.7%e70% []
NACA 63415 [] [] [] 55.5%e70% []
5 NACA X015 [] [] [] [] (except for NACA 0015, the smaller the better)

Acknowledgements investigation on aerodynamic performance of a novel vertical axis wind tur-


bine with adaptive blades, Energy Convers. Manag. 108 (2016) 275e286.
[19] M.S. Hameed, S.K. Afaq, Design and analysis of a straight bladed vertical axis
This work is supported by National Natural Science Foundation wind turbine blade using analytical and numerical techniques, Ocean. Eng. 57
of China Grant Nos. 51406117, Nos. 51536006, USST Key Laboratory (2013) 248e255.
of Flow Control and Simulation (D15013). [20] M.H. Mohamed, Performance investigation of H-rotor Darrieus turbine with
new airfoil shapes, Energy 47 (2012) 522e530.
[21] D.C. Wilcox, Turbulence Modeling for CFD, second ed., DCW Industries, La
Can~ ada, California, 1998.
References [22] A.J. Buchner, M.W. Lohry, L. Martinelli, J. Soria, A.J. Smits, Dynamic stall in
vertical axis wind turbines: comparing experiments and computations, Dy-
[1] O. Eboibi, L.A.M. Danao, R.J. Howell, Experimental investigation of the influ- namic stall in vertical axis wind turbines: comparing experiments and com-
ence of solidity on the performance and flow field aerodynamics of vertical putations, J. Wind Eng. Ind. Aerodyn. 146 (2015) 163e171.
axis wind turbines at low Reynolds numbers, Renew. Energy 92 (2016) [23] L. Daroczy, G. Janiga, K. Petrasch, M. Webner, D. The venin, Comparative
474e483. analysis of turbulence models for the aerodynamic simulation of H-Darrieus
[2] M.M.A. Bhutta, N. Hayat, A.U. Farooq, Z. Ali, S.R. Jamil, Z. Hussain, Vertical axis rotors, Energy 90 (2015) 680e690.
wind turbine - a review of various configurations and design techniques, [24] R. Gosselin, G. Dumas, M. Boudreau, Parametric Study of H-darrieus Vertical-
Renew. Sustain. Energy Rev. 16 (2012) 1926e1939. axis Turbines Using URANS Simulations (Paper CFDSC-2013 178) 21st Annual
[3] C.X. Li, X.Y. Li, X.M. Ye, Reviews on improved structure and performance of Conference of the CFD Society of Canada, Sherbrooke, Canada, 2013.
vertical Axis wind turbines, Compress. Blower Fan Technol. 67e73 (2013) 1. [25] P. Marsh, D. Ranmuthugala, I. Penesis, G. Thomas, Three-dimensional nu-
[4] X.D. Mo, T. Li, Types and development advantages of vertical Axis wind tur- merical simulations of straight-bladed vertical axis tidal turbines investigating
bine, Energy Conserv. Technol. 28 (2010) 450e453. power output, torque ripple and mounting forces, Renew. Energy 83 (2015)
[5] L.A. Danao, J. Edwards, O. Eboibi, R. Howell, A numerical investigation into the 67e77.
influence of unsteady wind on the performance and aerodynamics of a ver- [26] J. McNaughton, F. Billard, A. Revell, Turbulence modelling of low Reynolds
tical axis wind turbine, Appl. Energy 116 (2014) 111e124. number flow effects around a vertical axis turbine at a range of tip-speed
[6] A. Shahzad, T. Asim, R. Mishra, A. Paris, Performance of a vertical axis wind ratios, J. Fluid. Struct. 47 (5) (2014) 124e138.
turbine under accelerating and decelerating flows, Proc. CIRP 11 (2013) [27] M. Jafaryar, R. Kamrani, M. Gorji-Bandpy, M. Hatami, D.D. Ganji, Numerical
311e316. optimization of the asymmetric blades mounted on a vertical axis cross-flow
[7] G. Darrieus. Turbine having its rotating Shaft Transverse to flow of Current, U. wind turbine, Int. Commun. Heat Mass Transf. 70 (2016) 93e104.
S. Pat. 1834018, 1931. [28] M.F. Ismail, K. Vijayaraghavan, The effects of aerofoil profile modification on a
[8] J.R. Baker, Features to aid or enable self starting of fixed pitch low solidity vertical axis wind turbine performance, Energy 80 (2015) 20e31.
vertical Axis wind turbines, J. Wind Eng. Ind. Aerodyn. 15 (1) (1983) 369e380. [29] F.R. Menter, M. Kuntz, R. Langtry, Ten years of industrial experience with the
[9] Q. Li, T. Maeda, Y. Kamada, J. Murata, K. Furukawa, M. Yamamoto, Effect of SST turbulence model, Turbul. Heat. Mass Transf. (2003) 4.
number of blades on aerodynamic forces on a straight-bladed vertical axis [30] J.V. Healy, The influence of blade thickness on the output of vertical axis wind
wind turbine, Energy 90 (1) (2015) 784e795. turbines, Wind Eng. 2 (1) (1978) 1e9.
[10] Q. Li, T. Maeda, Y. Kamada, J. Murata, K. Shimizu, T. Ogasawara, A. Nakai, [31] M.C. Claessens, The Design and Testing of Airfoils for Application in Small
T. Kasuya, Effect of solidity on aerodynamic forces around straight-bladed Vertical Axis Wind Turbines, Delft University of Technology, 2006 (MS thesis).
vertical axis wind turbine by wind tunnel experiments (depending on num- [32] B. Kirke, L. Lazauskas, Enhancing the performance if a vertical axis wind
ber of blades), Renew. Energy 96 (2016) 928e939. turbine using a simple variable pitch system, Wind Eng. 15 (4) (1991)
[11] M. Elkhoury, T. Kiwata, E. Aoun, Experimental and numerical investigation of 187e195.
a three-dimensional vertical-axis wind turbine with variable-pitch, J. Wind [33] G. Bedon, U.S. Paulsen, H.A. Madsen, F. Belloni, M.R. Castelli, E. Benini,
Eng. Ind. Aerodyn. 139 (2015) 111e123. Computational assessment of the Deep Wind aerodynamic performance with
[12] M. El-Samanoudy, A.A.E. Ghorab, S.Z. Youssef, Effect of some design param- different blade and airfoil configurations, Appl. Energy 185 (2015), 1100(9).
eters on the performance of a Giromill vertical axis wind turbine, Ain Shams [34] Y. Mei, Research on Blade Airfoil of H Type Vertical axis Wind Turbine in Low
Eng. J. 1 (1) (2010) 85e95. Wind Speed, Master Thesis, Harbin Institute of Technology, 2010 (In Chinese).
[13] J.C. Vassberg, A.K. Gopinath, A. Jameson, Revisiting the Vertical-axis Wind- [35] S.X. Liao, C. Li, J.B. Nie, W. Gao, The analysis of aerodynamic performance for
turbine Design Using Advanced Computational Fluid Dynamics, AIAA Paper, small H type VAWT based on different airfoils, Mach. Des. Res. 27 (3) (2011)
43rd AIAA ASM, Reno, NV, 2005. 108e116 (In Chinese).
[14] S. Joo, H. Choi, J. Lee, Aerodynamic characteristics of two-bladed H-Darrieus at [36] W. Gao, C. Li, Z. Ye, J.B. Nie, Dynamic characteristics and aerodynamic per-
various solidities and rotating speeds, Energy 90 (2015) 439e451. formance of vertical axis wind turbine, J. Drainage Irrigat. Mach. Eng. 32 (2)
[15] B. Shahizare, N. Nik-Ghazali, W.T. Chong, S. Tabatabaeikia, N. Izadyar, (2014) 146e150.
A. Esmaeilzadeh, Novel investigation of the different Omni-direction-guide- [37] T. Ye, M. Liu, Study on asphalt pavement high temperature deformation based
vane angles effects on the urban vertical axis wind turbine output power on creep model, J. Wuhan Univ. Technol. Transp. Sci. Eng. 37 (6) (2013)
via three-dimensional numerical simulation, Energy Convers. Manag. 117 1287e1295 (In Chinese).
(2016) 206e217. [38] X.J. Sun, Q.D. Lu, D.G. Huang, G.Q. Wu, Airfoil selection for a lift type vertical
[16] Y. Lee, H. Lim, Numerical study of the aerodynamic performance of a 500 W axis wind turbine, J. Eng. Thermophys. 33 (3) (2012) 408e410.
Darrieus-type vertical-axis wind turbine, Renew. Energy 83 (2015) 407e415. [39] A. Bianchini, F. Balduzzi, G. Ferrara, L. Ferrari, Virtual incidence effect on
[17] G. Bedon, S.D. Betta, E. Benini, Performance-optimized airfoil for Darrieus rotating airfoils in Darrieus wind turbines, Energy Convers. Manag. 111 (2016)
wind turbines, Renew. Energy 94 (2016) 328e340. 329e338.
[18] Y. Wang, X.J. Sun, X.H. Dong, B. Zhu, D.G. Huang, Z.Q. Zheng, Numerical
818 Y. Wang et al. / Renewable Energy 126 (2018) 801e818

[40] M.C. Claessens, The Design and Testing of Airfoils for Application in Small [47] K.W. McLaren, A Numerical and Experimental Study of Unsteady Loading of
Vertical Axis Wind Turbines, Delft University of Technology, 2006 (MS thesis). High-solidity Vertical Axis Wind Turbines, Ph.D. Thesis, McMaster University,
[41] P. Sabaeifard, H. Razzaghi, A. Forouzandeh, Determination of vertical axis Canada, 2011.
wind turbines optimal configuration through CFD simulations, in: 2012 In- [48] M.R. Castelli, A.D. Monte, M. Quaresimin, E. Benini, Numerical evaluation of
ternational conference on Future Environment and Energy, Singapore, 2012. aerodynamic and inertial contributions to Darrieus wind turbine blade
[42] A. Bianchini, F. Balduzzi, J.M. Rainbird, J. Peiro, J.M.R. Graham, G. Ferrara, et al., deformation, Renew. Energy 51 (2013) 101e112.
On the influence of virtual camber effect on airfoil polars for use in simula- [49] C. Li, S.Y. Zhu, Y.L. Xu, Y.Q. Xiao, 2.5D large eddy simulation of vertical axis
tions of darrieus wind turbines, Energy Convers. Manag. 106 (2015) 373e384. wind turbine in consideration of high angle of attack flow, Renew. Energy 51
[43] L.A. Danao, N. Qin, R. Howell, A numerical study of blade thickness and (2013) 317e330.
camber effects on vertical axis wind turbines, Proc. Inst. Mech. Eng. Part A J. [50] S. Wagner, R. Bareiss, G. Guidati, Wind Turbine Noise, Springer, Berlin, 1996,
Power & Energy 226 (7) (2012) 867e881. pp. 71e72.
[44] M.R. Castelli, A. Englaro, E. Benini, The Darrieus wind turbine: proposal for a [51] A. Betz, Das Maximum der Theoretisch Mo € glichen Ausnützung des Windes
new performance prediction model based on CFD, Energy 36 (2011) durch Windmotoren, Zeitschrift Für Das Gesamte Turbinenwesen, 1920.
4919e4934. [52] J.M. Jonkman, Modeling of the UAE wind turbine for refinement of Fast_AD,
[45] M. Jafaryar, R. Kamrani, M. Gorji-Bandpy, M. Hatami, D.D. Ganji, Numerical NREL/TP-500e34755, 2003. page 7.
optimization of the asymmetric blades mounted on a vertical axis cross-flow [53] A. Rezaeiha, I. Kalkman, B. Blocken, Effect of pitch angle on power perfor-
wind turbine, Int. Commun. Heat Mass Transf. 70 (2016) 93e104. mance and aerodynamics of a vertical axis wind turbine, Appl. Energy 197
[46] F.R. Menter, Two-equation eddy-viscosity turbulence models for engineering (2017) 132e150.
applications, AIAA J. 32 (1994) 1598e1605.

You might also like