R1 V2 NoChangeMarked

See discussions, stats, and author profiles for this publication at: https://www.researchgate.
net/publication/365882257
Recommendation for strut designs of vertical axis wind turbines: Effects of

strut proﬁles and connecting conﬁgurations on the aerodynamic performance
Article in Energy Conversion and Management · January 2023

DOI: 10.1016/j.enconman.2022.116436
CITATIONS READS
0 75
8 authors, including:
Weipao Miao Qingsong Liu

University of Shanghai for Science and Technology University of Shanghai for Science and Technology
34 PUBLICATIONS 231 CITATIONS 15 PUBLICATIONS 132 CITATIONS
SEE PROFILE SEE PROFILE
Zifei Xu Chun Li
Liverpool John Moores University University of Shanghai for Science and Technology
16 PUBLICATIONS 179 CITATIONS 91 PUBLICATIONS 1,415 CITATIONS
SEE PROFILE SEE PROFILE
Some of the authors of this publication are also working on these related projects:
Wind Farm Analysis and Optimization View project
fluid-induced vibration due to seals View project
All content following this page was uploaded by Weipao Miao on 06 December 2022.
The user has requested enhancement of the downloaded file.

Recommendation for strut designs of vertical axis wind
turbines: Effects of strut profiles and connecting
configurations on the aerodynamic performance
Weipao Miaoa, Qingsong Liua, Qiang Zhanga, Zifei Xua,b, Chun Lia,c,*, Minnan Yuea, Wanfu Zhanga, Zhou Yea
a School of Energy and Power Engineering, University of Shanghai for Science and Technology, Shanghai, 200093. China
b Department of Maritime and Mechanical Engineering, Liverpool John Moores University, Liverpool, Byrom Street, L3 3AF, UK
c Shanghai Key Laboratory of Multiphase Flow and Heat Transfer in Power Engineering, Shanghai 200093, China
Abstract: The strut is a necessary support structure that connects the vertical axis wind turbine blade to the
rotating shaft and transfers the torque, and is one of the key factors affecting the aerodynamic performance of the
wind turbine. However, research on struts is limited and the appropriate strut design remain uncertain. Therefore,
in this paper, the struts of a straight-bladed vertical axis wind turbine are investigated by a three-dimensional
computational fluid dynamics approach, to find the appropriate strut design parameters that minimize the impact
on the aerodynamic performance. Thirty strut profiles with different chord lengths, thicknesses, areas, and
streamlined shapes were parametrically created and analyzed, including a novel strut-specific profile proposed in
this paper. Furthermore, all strut connecting parameters, such as spanwise and chordwise positions, connecting
angle and fairing were investigated. The results show that the power loss of a vertical axis wind turbine has a
relatively linear relationship with the dimensionless absolute thickness of the strut within a certain range. The strut
chord length is another important influencing factor. Equal chord lengths of the strut and the blade should be
avoided as much as possible. The proposed novel olivary strut profile has lower drag characteristics and more
favorable geometrical structural properties compared to the most used NACA airfoil. The influence of the spanwise
connecting position (within the range around the blade quarter span) is minor. But the chordwise position has a
significant effect, and special attention should be paid to avoid the strut maximum thickness position connecting
with the blade aerodynamic center, which will result in the greatest performance degradation. Analysis of the
connecting angles shows that inclined struts are suitable for small wind turbines. The fairing is effective when the
strut chord length is equal to or longer than the blade chord length. With these results, a set of recommended design
parameters for vertical axis wind turbine struts is presented.
Key words: Strut Effects; H-type VAWT; Aerodynamic Drags; Strut Parameters; CFD;
Nomenclature
AB Cross-sectional Area of Blade [m2] LE Leading Edge [-]
AC Aerodynamic Center [-] MTP Maximum Thickness Position [-]
ARC Asymptotic Range of Convergence [-] My Bending Moment [N·m]
AS Cross-sectional Area of Strut [m2] N Number of Blades [-]
c Chord Length [m] Nk Number of Grids [-]
CAD Computer Aided Design [-] P Power Output [W]
cB Chord Length of Blade [m] R Rotor Radius [m]
CFD Computational Fluid Dynamic [-] RANS Reynolds-Averaged Navier-Stokes [-]
COE Cost of Electricity [-] Re Reynolds Number [-]
CP Power Coefficient [-] S Strut Position [-]
CPre Pressure Coefficient [-] SIMPLE Semi-Implicit Method for Pressure Linked Equations [-]
CQ Torque Coefficient [-] SST Shear Stress Transport [-]
cS Chord Length of Strut [m] t Time [s]
D Rotation Diameter [m] Tabs Strut Absolute Thickness [-]
dLE Distance Between Strut LE and Blade LE [m] TCR Thickness-Chord Ratio [-]
dMTP Distance Between Strut MTP and Blade LE [m] TSR λ Tip Speed Ratio [-]
DT Tower Diamater [m] V∞ Incoming Velocity [m/s]
FD Drag Force [N] VAWTs Vertical Axis Wind Turbines [-]
FL Lift Force [N] W Synthetic Velocity [m/s]
FN Normal Force [N] α Angle of Attack [-]
FT Tangential Force [N] β Fixed Pitch Angle [°]
GCI Grid Convergence Index [-] ζ Strut Connecting Angle [°]
H Blade Height [m] θ Azimuth Angle [°]
AC Aerodynamic Center [-] θS Tangent Angle of Strut Profile Edge [°]
ARC Asymptotic Range of Convergence [-] μ Dynamic Viscosity [Pa·s]
h Half Height of Blade [m] ρ Air Density [kg/m3]
HAWTs Horizontal Axis Wind Turbines [-] σ Solidity of Rotor [-]
IDDES Improved Detached Eddy Simulation [-] σS Stress [Pa]
Ix Moment of Inertia [m4] Ω Rotational Speed [rad/s]
* Corresponding author. Tel, +021-55271729

E-mail address: linchun_usst@163.com
1
1. Introduction
Wind energy is clean and renewable, with considerable potential to alleviate the future fossil fuel scarcity and
price increase[1]. Through years of development, horizontal axis wind turbines (HAWTs) are the dominant wind
energy utilization devices on the market today[2]. However, recent studies reveal that it is difficult to further reduce
the cost of electricity (COE) by increasing the size of HAWTs[3-5]. In addition, for the current global emphasis on
offshore wind energy, HAWTs have to invest more in the construction and subsequent operation and maintenance
of offshore floating platforms due to their high center of gravity[5]. As these issues of HAWTs became more
prominent, attention is once again being focused on vertical axis wind turbines (VAWTs), especially the lift-type
VAWT with efficiency comparable to those of HAWTs[6, 7]. VAWTs have no yaw system, low center of gravity,
segmental blades, less deformation and edgewise loading, which is more favorable to design as floating type, and
their mooring system cost can be reduced by more than 20% compared with HAWTs[8, 9]. These advantages make
VAWTs potentially more suitable than HAWTs as the future offshore wind energy. Many countries have made
VAWTs a research priority, leading to their commercialization[7].
However, as the development of VAWTs has lagged behind that of HAWTs in the last few decades, resulting
in the technological maturity of VAWTs still facing challenges[10]. In particular, despite the evident advantages
of VAWTs, optimal aerodynamic design is still a prominent issue[11]. As an example, Fig. 1 shows the design
parameters that affect the aerodynamic performance of a straight-bladed VAWT, including blade tip design[12, 13],
pitch angle[14], coning angle[15], solidity[16], blade aspect ratio[17], airfoil[18, 19], blade orientation[20], wind turbine
height-to-radius ratio[21], tower[22] and strut[23], etc. Many of these factors are also interrelated and thus affect the
overall power efficiency. A detailed review of these parameters is presented by Hand et al.[24]. Fortunately, with
the rapid development of wind turbine technology, considerable studies on VAWTs are being carried out by
academics around the world in an endeavor to provide optimum design solutions for VAWTs. In addition, many
studies have attempted to improve the aerodynamic performance of the VAWT in terms of flow control methods,
such as the active flow control using suction[25, 26], blowing[27], dynamics flap[28] and variable pitch[29], or the
passive flow control by means of adaptive flap[30], Gurney flap[31], deflector[32], leading edge serrations[33], slotted
airfoil blades[34], or dual blades[35, 36].
Bla
de
As
pec
tR
Coning Angle atio
Tip Design Pitch Angle
ity
f Bla
des Solid
ber o atio
Num s R
diu atio
o-Ra ius R
Chord-t -Rad
ht-to
Heig
Reynolds Number
n
Desig
Strut
ut
e-Str
Blad on Point
o n n ecti
Tower Design
AERODYNAMIC
Airfoil PERFORMANCE
Efficiency
Blade Orientation Power Output
Operational Range
Optimum Tip Speed Ratio
Fig. 1. Design parameters affecting the aerodynamic performance of the VAWT[24].

It is important to noted that most of the past VAWT performance studies are investigated through two-
dimensional (2D) models[28, 37, 38] due to the limitations in computer technology and resources. When three-
2
dimensional (3D) models are used, the aerodynamic performance of most optimally designed VAWTs will be
substantially lower than that of the 2D models. In addition to the blade tip loss effect[12] of 3D blades, the strut
effect of VAWTs is another important factor. However, the current limited strut studies are not sufficient to provide
suitable design parameters, and further investigation is needed. This is the primary research objective of this paper.
1.1. Strut Effect of VAWTs

Currently, lift-type VAWTs that have been validated as efficient and feasible by extensive research include:
the straight-blade “Musgrove H-rotor”, the helical-blade “Gorlov” and the curved-blade “Darrieus Φ-rotor”[39].
The strut (also called support structures or arms) is an essential generic structure for a VAWT. For the helical
and straight blade VAWTs, the struts not only provide structural support to resist aerodynamic, gravitational, and
inertial loads on the blade, but are also responsible for transferring torque to the shaft for power generation and
strongly influence the natural frequency of the rotor[40, 41]. For the Φ-type wind turbine, although the 34 m VAWT
first developed by Sandia Labs stabilized the curved blade by its tension without the need for struts, they
emphasized the necessity of using struts for reinforcement due to the increasing weight of the large-scale VAWT
blades[11]. Consequently, despite their different types, all three VAWTs invariably require struts to provide the
necessary structural connection and support, especially for large-scale VAWTs.
However, the presence of struts generates two types of drag during the turbine rotation: the direct profile drag
caused by the struts cross-sectional shape and the induced drag due to the interference at the strut-blade connecting
interface[11],[42]. Many studies have demonstrated that both types of drag significantly reduce the VAWT power
output. Tests by SANDIA showed that the struts reduced the maximum power output of the VAWT by 26%[43]. An
experimental study by Li et al.[44] showed that the output power of the VAWT with struts measured with a torque
meter and a six-component balance was significantly lower than the power measured with the pressure taps on the
blade surface. Recent numerical simulations by Aya et al.[23, 45] also showed that at high tip-speed-ratio (TSR), the
struts reduced the power of their 12 kW H-type VAWT by 43%. This significant degradation in aerodynamic
performance is due to the faster rotor speed at high TSR, resulting in an exponential increase in the strut effect[14].
Therefore, due to the pronounced influence of the strut, any attempt to model the VAWT without modelling struts
will lead to inaccurate solutions.
1.2. Influences of Strut Parameters

It is clear that a reasonable strut design is essential to reduce the power loss of VAWT. However, the design
parameters of the strut are numerous, and the related research is still limited. The following four sections involve
the overview of different strut parameters.
1.2.1. Strut Profiles
According to fundamental aerodynamic principles, it is known that the cross-sectional profile shape plays a
dominant role in the drag characteristics of the strut. Early VAWTs use cylindrical strut profiles, considering only
the support function[14]. The drag effect of cylinder struts proved to be extremely significant. Peter et al. [46]
compared the drag of a NACA 0021 airfoil and a cylinder strut using a hydraulic tow tank. They rotated the struts
individually to measure their drag characteristics and found that the high drag produced by the cylinder strut
prevented the turbine from producing power output, while the power losses of the NACA airfoil strut were much
lower. Apparently, the streamlined cross section reduces the profile drag of the strut itself. For this reason,
Mazharul et al.[41],[47] designed a special profile for the strut named MI-Struct1, and they confirmed through their
in house code that the profile drag of MI-Struct1 is lower than that of the E862 airfoil.
However, it is not enough to only focus on the strut profile drag. The strut-to-blade connection type also
affects the aerodynamic performance of the blade itself, known as the induced drag. Philip et al. [48] experimentally
evaluated the effect of two types of strut-to-blade connections on the power efficiency of a vertical axis tidal
turbine bolted flat strut and a fused NACA0012 profiles strut. Their results showed a 50% decrease in efficiency
3
for the turbine with flat struts. Unfortunately, it is difficult to isolate in the experiment whether profile or induced
drag has a stronger effect. But at least their combined effect on the turbine performance is significant. Yutaka et al.
[49]
compared the effect of three strut profiles, NACA 0018 airfoil, rectangular and circular, of the same thickness
on the resistance torque (the direct drag) and the tangential torque (the induced drag). They decomposed these two
types of torque to the differential pressure and frictional drags, and found that the effect of struts on the blade
frictional drag was much smaller. Another interesting finding was that the profile drag was higher than the induced
drag for the rectangular and circular struts, while the opposite was true for NACA airfoil. Unfortunately, they did
not further investigate the flow mechanism responsible for this phenomenon.
These previous studies analyzed the effect of different strut profiles with the same chord length or the same
thickness, but another important factor was not sufficiently considered. That is, the structural strength of these
struts is different, which leads to unfair comparisons in many studies. The design of struts involves a compromise
between aerodynamic performance and structural strength[50]. Mojtabadeng et al.[40] considered this factor and
compared different strut profiles such as airfoil, cylinder, circular, rectangular and diamond with the same cross-
sectional area. The total power loss and structural stresses of different struts were analyzed by their in-house
program. The results show that the drag of rectangular and diamond profiles is greater than that of circular profile,
and the streamlined airfoil profile is obviously the best choice. Unfortunately, however, their analysis is unable to
identify the effect of strut profile on the blade aerodynamic performance (the induced drag).
1.2.2. Strut Connecting Position
The connecting position of the strut could be discussed in two aspects, the spanwise and the chordwise.
(i) The Spanwise Position
To balance the rotor, all types of struts generally installed symmetrically to the central height position of the
blade. Depending on the spanwise connecting position, the VAWT struts are classified into the three types
illustrated in Fig. 2, the middle span, the quarter span, and the end span. Noted that the quarter span in Fig. 2 (b)
is not fixed at the 1/4 position of the blade exactly, but at the position between the middle span and the end span.
Intuitively, the middle span should have the least drag because it uses one strut on each blade[51]. Nevertheless,
the results of Philip et al.[48] and Thierry et al.[51] showed that the end span installed struts have higher VAWT
aerodynamic efficiency. This is because the middle or quarter span struts interrupt the high-performance region of
the blade surface. On the contrary, the end span installed struts do not interfere with the blade, but also contribute
to reducing the blade tip losses because it acts like an endplate[51].
However, the installation of struts should consider not only the aerodynamic performance but also the
necessary structural support, which is even more important for large-scale VAWTs. If only one strut is used for
each blade, the aerodynamic forces and the gravity of the blade will generate significant bending stresses at the
joints. Mojtaba et al.[40] analyzed and optimized the blade bending stresses for these three strut spanwise positions,
and showed that the maximum stresses for the blade with end span struts were nearly five times higher than those
for the blade with quarter span struts. This was also confirmed by Hameed et al.[52] through finite element analysis.
Therefore, despite the relatively better aerodynamic performance of the end span struts, the structural risk is
significantly higher. On the contrary, the quarter span strut is more balanced in terms of aerodynamic and structural
performance and is a relatively better choice[24]. In addition, blades with quarter span struts can also reduce the tip
loss through a blade tip design[12], thus achieving similar results as end span struts, but with better structural
properties.
4
(a) (b) (c)
[50]
Fig. 2. VAWTs with different strut spanwise connecting positions . Ω is the rotating speed. (a) Middle Span. (b) Quarter Span. (c)
End Span.
(ii) The Chordwise Position
In VAWT fabrication, the struts are connected at the mid-chord position (i.e., 0.5c) of the blade in most
cases[53]. However, Bianchini et al. [54] showed that since the aerodynamic center (AC) of most subsonic airfoils is
located at 0.25c, the actual force on the blade will change when the connecting position is located at 0.5c as shown
in Fig. 3 (a). This is equivalent to a slight increase in the blade radius and an additional pitch angle, which also
causes a pitching moment.to be generated by the normal component force of the blade. This phenomenon is more
pronounced in smaller wind turbines with a higher solidity. Therefore, they recommend a chordwise connecting
position of 0.25c for the struts.
Aerodynamic Center
Strut Connecting Position
R
R
c B
0.25c B
0.5c B
R
(a) (b)
Fig. 3. Different strut chordwise connecting positions for various strut chords. (a) ACs on different radius. (b) ACs on the same
radius. cB represents the blade chord. R is the rotor radius.
However, since the 2D VAWT numerical model[53, 54] used in previous studies on the blade-strut chordwise
connection could not practically include struts, what they actually analyzed was the variation of the rotor radius
and the blade pitch angle due to different connection positions. For the 3D VAWT model, another important factor,
the strut width, must be considered when analyzing the spanwise connecting position, as shown in Fig. 3 (b). In
the case where the strut chord length is smaller than the blade chord, for blades with AC on the same radius, the
variation of the strut chordwise position does not affect the AC location and the pitch angle, but affects the flow
on the blade surface resulting in different induced drags. Research on this subject has not been carried out so far.
1.2.3. Strut Connecting Angles
In addition to the common horizontal installation, the strut can be inclined at a certain angle, as shown in Fig.
4. For a straight-bladed VAWT, the advantage of using inclined struts is that the tower height and the center of
gravity of the rotor can be reduced[11], even to the bottom, as shown in Fig. 4 (c). Many concept designs of large-
scale offshore VAWTs[7, 55] have adopted this V-rotor because its low center of gravity characteristic is beneficial
for the stability of floating platforms. Agostino et al. [56] proposed a combined horizontal and inclined VAWT strut,
and their numerical simulations and experimental results showed that the inclined part of the strut could captures
additional wind energy and compensate for its own drag. Thierry et al. [51] also found that the inclined 45° strut
provide positive moment when analyzing the inclined angle of the strut mounted on the end span, but it
significantly interfered with the blade aerodynamic efficiency. Since their struts extended from the blade tip, it is
actually more like the effect of different blade tip designs.
It should be noted that the non-vertical connection of the inclined strut to the blade will result in bending
moment, rather than pure compression and tension as in the case of horizontal struts[57]. Therefore, for structural
reasons, Hand et al. [24] suggested a horizontal form of strut for large-scale VAWTs. Nevertheless, a comparison of
5
the effects of horizontal and inclined struts on the VAWT aerodynamic performance is still lacking.
(a) (b) (c)

Fig. 4. VAWTs with inclined struts. ξ is the inclined angle. (a) Φ-rotor. (b) H-rotor. (c) V-rotor
1.2.4. Strut-Blade Connecting Fairing
Another parameter of the strut that should be considered is the geometry at the connection. When a blade is
connected to a strut, a horseshoe vortex is generated at the “T” shape joint, resulting in greater wake extension[58].
This is one of the reasons for the degradation of the blade aerodynamic performance caused by the strut.
Fig. 5. The horseshoe vortex at the blade-strut connecting joint[24].

This performance loss could be reduced by using a fairing, which is a streamlined geometry that smoothly
merges the blade and the strut at the connecting joint [24]. Hoerner [59] suggested using fairings with 4%~8% airfoil
chord length circular radius as transition joints in aircraft design. The SANDIA labs added manual fairings to the
blade joints of their 34 m VAWT, resulting in a small increase in performance in both low and high winds [11]. But
their report did not provide details on the design of the fairing. Denis et al. [60] developed a 600 kW onshore VAWT
with fairings, but also no details were provided. Thierry et al. [51] analyzed the effect of a circular transition between
the end span strut and the blade, which was also considered as a fairing. They found that a radius of 0.5c works
most effectively and could significantly decrease the kinetic energy of the vortex wake, which thus reduces the
induced drag. Unfortunately, the effect of fairings on the performance of VAWT blades has not been sufficiently
studied so far, especially the issue of matching different strut sizes to the fairing needs further investigation.
1.3. Present Work

1.3.1. Motivations and Objective
The above overview shows that the strut is an essential part of the VAWT, but it significantly affects the
aerodynamic efficiency. Although some studies have been carried out, several issues still deserve future
investigation.
To this end, the VAWT struts are investigated in this paper. The most aerodynamically efficient straight-bladed
VAWT[6] is the subject of this paper. To balance the aerodynamic and structural performance, struts installed near
the quarter span are used. Based on this, the strut parameters are numerically investigated to minimize their
influence on the VAWT aerodynamic performance. The study focuses on the following issues:
⚫ It is known that using streamlined profile reduces the strut drag. The drag characteristics of a strut profile
is highly dependent on its chord length and relative thickness. However, discussions on these two parameters
have not been conducted yet, and studies regarding their influence on the aerodynamic performance are still
lacking in research. In addition, there is a periodic inversion of the pressure and suction sides of the VAWT
blade. The influence of the struts on the different sides needs to be studied.
⚫ At present, most of the strut use the same airfoil cross-sections as the VAWT blade, or a symmetrical NACA
airfoil. However, it is still uncertain how to choose the strut profile size when the strut satisfies the structural
strength. Are there other more suitable strut cross-sectional profiles?
6
⚫ The effects of the strut connecting positions require confirmation. In particular, the influence of different
chordwise positions has not been discussed when the strut chord length is smaller than the blade chord
length.
⚫ Inclined struts are frequently used in many large-scale VAWTs in recent years, but few analyses of have
been published regarding their effects on the aerodynamic performance.

⚫ How the connecting fairing between the strut and the blade affects the VAWT needs further study, especially
the matching with different strut chord lengths.

In response to the above issues, this paper investigates VAWT struts, focusing on their influence on the
aerodynamic performance. Using the parametric technology, nearly thirty different strut profiles including a novel
low drag olivary shape were created by CAD method, and about eighty VAWT cases were calculated by 3D CFD
simulation. The effects of chord length, thickness and area of struts were analyzed. Then the influences of the strut
connecting configurations on VAWT aerodynamic performance were analyzed. This paper tried to provide some
suggestions for the future VAWT strut design.
1.3.2. Organization of Present Study
The rest of this paper is organized as follows. Sections 2.1-2.2 present the geometrical and operational
parameters of the baseline VAWT and the basic aerodynamic characteristics. Section 2.3 describes different strut
parameters investigated in this paper, including the strut profiles (2.3.1), the spanwise and chordwise positions
(2.3.2), the connecting angle (2.3.3) and the fairings (2.3.4). The models and settings required for the CFD
approach, and the validations about mesh number, time step, revolution, turbulence model, and comparison of
experimental data are presented in Sections 2.4-2.5.
The results are discussed in Section 3 in terms of the effect on both the aerodynamic performance and the
fluid field characteristic. Section 3.1 analyzes the influence of the existence of the strut. The discussion in Sections
3.2-3.5 correspond to the different strut configurations in Section 2.3. The relatively optimal strut parameters are
summarized in Section3.6. Finally, Sections 4-5 provide the conclusions and limitations of this paper, and further
work that can be considered in the future.
2. Physical and Numerical Models

2.1. The Baseline VAWT Model
In this paper, a two-straight-bladed VAWT model of Li et al. [61, 62] is used, which provides elaborate
experimental data to facilitate validation. Fig. 6 shows the VAWT model and the wind tunnel experimental facility
[62]
. The specifications of this VAWT are illustrated in Table 1.
Li et al. [40] did not provide the detailed parameters of the strut. Based on their image data[40], a thin and round-
headed strut displayed in Fig. 6-(a) is created as a baseline wind turbine model. The strut has a chord length of
0.5cB and a thickness of 0.01H. A semicircular transition is used at both ends. The blade AC (0.25cB) is located on
the rotation radius. The center of the strut is connected to the 0.5cB position of the blade. The influence of this
chordwise connecting position will be discussed in subsequent sections. The tower diameter is the same as the
blade chord length. The tower height is slightly higher than the strut position.
7
NACA 0015
(a)
β=6° S=0.25cB
Fixed Pitch Angle Strut Connected Position
in CAD and CFD model

D=
Split into 20 segments

c S =0.5c B 1.7
0m
z=0.3H
TAbs =0.01H
DT =1.0c
Split into 10 segments
Ω in CAD and CFD model
25m
Symmetrically Model in CFD cB=0.2
m/s
V∞ =7 Pressure Taps
z=0.80h
h=H/2=0.51m
z=0.70h
z=0.55h
H=1.02m
z=0.05h
e
wis
l
Downwind
a
θ=0°
enti
ord
g
Ch
Tan
Upwind
FD
FN θ
W FW FT
α β
FL
V∞ Ω·
R
Z
Y X θ=180°
Fig. 6. The VAWT model and the tunnel experiment facility. The VAWT is simulated symmetrically. So the gray part of the blade
are not meshed in the CFD model. (a) VAWT model with key specifications. Detailed parameters are seen in Table 1. (b)
Experimental facilities from Ref [61, 62].
To facilitate comparison with the experiment, the baseline struts are fixed at 0.3H from the blade tip. Note
that only half of the blade height is modeled, i.e., h=H/2=0.51m, while the other half is mapped in the CFD model
using the symmetry plane boundary condition. This helps to speed up the simulation. Although the symmetric
model results in a slight discrepancy at the bottom part of the tower, Aihara et al.[23] have shown that influence of
the tower is relatively insignificant in the presence of struts.
The experimental blade were added with pressure measuring taps at four different heights of z=0h, 0.55h,
0.7h and 0.8h[39], as shown in Fig. 6-(a). In the numerical model of this paper, the blade and strut surfaces are split
into 20 and 10 segments respectively, to monitor the forces varying along the span direction. This is also convenient
to extract the data for four heights to compare experimental and simulation results.
Table 1 Geometrical and operational parameters of the VAWT.
Parameter Value
Airfoil NACA 0015
Blade Chord cB 0.225 m
Strut position S 0.25c
Fixed pitch angle β 6°
Rotation diameter D 1.7 m
Blade Height H 1.02 m
Number of blades N 2
Solidity σ=Nc/D 0.265
Tower diameter DT c
Strut chord cS 0.5c
Strut absolute thickness TAbs 0.01H
Reynolds number Re ~2.30×105
Incoming velocity V∞ 7 m/s
Rotational speed Ω 15.23-20.75 rad/s
Tip Speed Ratio (TSR)  1.85-2.52
2.2. Aerodynamic Characteristics of Straight-Bladed VAWT

The aerodynamic characteristics of the straight-bladed VAWT cross-section is presented in Fig. 6-(a). We
define θ as the azimuth angle, so 0°＜θ＜180° is called the upwind region and 180°＜θ＜360° is called the
downwind region. V∞ is the incoming wind speed, and Ω is defined as the rotational angular speed of the rotor. R
is the rotor radius. The relationship between the tangential velocity (Ω ·R) and incoming speed (V∞) can be
expressed by the tip speed ratio (TSR) λ:
ΩR
= (1)
V
The normal force FN and the tangential force FT of the VAWT blade are calculated as following:
8
 FT = FL sin( +  ) − FD cos( +  ) (2)

 FN = FL cos( +  ) + FD sin( +  ) (3)
where FL and FD are the lift and drag forces generated by the synthetic velocity of the incoming speed and
tangential velocity, α and β are the angle of attack and the fixed pitch angle respectively.
The tangential force generates a torque Q, which is the power source of the VAWT. A VAWT with N blades
can produce the power output P:
P = ΩQ (4)
Q = NFTa R (5)
1 2
2 0
FTa = FT ( )d (6)
where FTa is the averaged tangential force of a single blade over one revolution.
Typically, we use the torque coefficient CQ (Eq.(9)) and the power coefficient CP (Eq.(10)) to evaluate the
aerodynamic efficiency of VAWTs, which provides a convenient metric for different VAWTs [63].
Q
CQ = (7)
1 / 2  DHRV2
P
CP = (8)
1 / 2 DHV3
where ρ is the air density (assumed 1.18415 kg/m3 in this paper), and other parameters as described in Table 1.
2.3. Strut Models with Different Configurations

2.3.1. Strut Profiles
The drag of a strut consists of profile drag and induced drag. The profile drag is caused by the frictional force
and pressure difference generated by the strut profile. The induced drag is caused by the strut interfering with the
flow on the blade surface. Apparently, the strut drags highly depends on its profile shape and size.
However, another key factor needs to be considered before investigating the strut drag characteristics, namely
the structural properties of the strut. It should be strength enough to enable the strut to withstand the bending
stresses generated by the blade weight and the axial stresses generated by the normal component of aerodynamic
forces. For a horizontal strut, the internal stress σS can be approximated as[40]:
 S = FN / AS + M y / I x (9)
where AS is the cross-sectional area of the strut, My is the bending moment generated by the blade weight and the
vertical component of the aerodynamic force, and Ix is the moment of inertia of the strut cross-section. Assuming
the variation of strut profile does not affect FN and My, the stresses on the strut are similar when the cross-sectional
area is constant. Although the moment of inertia varies with the profile shape, it is still assumed to be the same
here because the moment of inertia is strongly related to the cross-sectional area.
This equation suggests that the basic principle in analyzing the drag characteristics of different struts is to
ensure that they have the same cross-sectional area, so that a fair comparison can be made. Therefore, a smaller
chord length requires a larger relative thickness to ensure the same area, and vice versa. Based on this principle,
the NACA XYZZ 4-digit series airfoils (ZZ is the thickness-chord ratio, TCR) which is most adopted in VAWT
(both blades and struts) are used in this paper, to create various struts.
The blade cross-sectional area (AB) is used as a reference. The strut cross-sectional areas (AS) of different
struts are classified as 0.5AB, 0.67AB, 0.75AB, 1.00AB and 1.50AB. The strut chord lengths (cS) are classified as
0.5cB, 0.75cB, 1.00cB, 1.25cB and 1.50cB based on the blade chord cB. Once the strut area and chord length are
defined, the relative thickness of the NACA 4-digit airfoil is determined. However, in this case, not all airfoil
thicknesses are integers. Here a parametric description of the NACA airfoil profile is adopted in the form of a
polynomial[64] and modeled in CAD software using a user-defined program. The airfoil polynomial is as follows:
9
2 3 4
y x  x  x  x  x
 = a0 + a1   + a2   + a3   + a4   (10)
c c c c c c
where x, y are the airfoil horizontal and vertical coordinates, c is the chord length, a0~a4 are the coefficients. Taking
NACA 0020 as an example, the coordinates are also subject to the following constraints:
(i) Maximum ordinate
x y dy
= 0.3 = 0.1 =0 (11)
c c dx
(ii) Ordinate at trailing edge：

x y
= 1.0 = 0.002 (12)
c c
(iii) Trailing-edge angle：

x dy
= 1.0 = 0.234 (13)
c dx
(iv) Nose shape：

x y
= 0.1 = 0.078 (14)
c c
The coefficients a0= 0.2969, a1= -0.1260, a2= -0.3516, a3 = 0.2843 and a4= -0.1015 are obtained by above
equations. Airfoils with different TCRs are created by adjusting the value of y/c. Twenty-two NACA airfoils are
shown in Table 2. For comparison purposes, the strut thickness is defined as a dimensionless parameter of the ratio
of the absolute thickness to the blade height Tabs/H, where Tabs=cS·TCR.
Table 2 The cross sections of VAWT struts with different profiles, areas, chord lengths, and thicknesses analyzed in this paper.
Taking the NACA airfoil in rows 1-6 as an example, the first row indicates the ratio of the strut chord (cS) to the blade chord (cB),
while the first column indicates the strut cross-sectional area (AS) to the blade cross-sectional area (AB). Thus, the cross section of
strut in the 2nd row and 2nd column has an area AS=0.5AB and a chord length cS=0.5cB. The absolute thickness (Tabs) of the strut is
represented as a ratio relative to the blade height (H). Several profiles with pentagrams have the same absolute thickness and are
used for comparison in later sections. The olivary struts in rows 8-11 are defined by the same rules, where DS-LE in the 5th column is
the leading/trailing edge diameter. The last two columns of rows 8-11 are used for the strut of the baseline wind turbine.
NACA Airfoil c =0.50c
S B c =0.75c
S B c =1.00c
S B c =1.25c
S B c =1.50c
S B
2.2%H
1.3%H
3.3%H
1.7%H
A =0.50A
S B
2.9%H
1.8%H
4.4%H
2.2%H
A =0.67A
S B
5.0%H
2.0%H
3.3%H
2.5%H
1.7%H
A =0.75A
S B
2.8%H
3.3%H
2.2%H
6.6%H
4.4%H
A =1.00A
S B
4.0%H
6.6%H
5.0%H
3.3%H
A =1.50A
S B
Olivary c =0.50c
S B c =0.75c
S B c =1.00c
S B c =1.25c
S B
3.3%H
2.2%H
A =0.50A
S B Baseline A =0.21A
S B
s
1.0%H
2.5%H
3.3%H
A =0.75A
S B
Flat 0.50c
02 c
s
=0.0
1.4%H
3.3%H
4.4%H
D s-LE
A =1.00A
S B Airfoil 0.50c
3.7%H
6.6%H
3.3%H
A =1.50A
S B Cylinder
In addition to the NACA series airfoil, a symmetrical olivary-shaped profile (see the left-bottom part of Table
2.) is proposed specifically for the VAWT struts, which has lower drag characteristics and can be fabricated simply.
Two reasons motivate the design of this olivary profile:
(i) Lift performance is not the primary function of a VAWT strut. Therefore, the NACA 4-digital series airfoil
with a high maximum lift coefficient is not necessary and not suitable for a strut. Since the relatively large
diameter of the NACA airfoil leading edge is not beneficial for drag reduction, the leading edge section is
10
modified to a similar size as the trailing edge.
(ii) The struts of a large-scale VAWT prefer to be manufactured as a hollow structure using composite
materials similar to the HAWT blade. This composite structure is supported by the internal spars and shear
webs, where the spars are usually located at the maximum thickness position (MTP). However, the
maximum thickness of the NACA airfoil is at 0.3cB, resulting in a forward center of gravity and a large
panel area in the trailing edge region, which is susceptible to buckling. Therefore, a centrosymmetric
profile looks like an olive-shape core is proposed.
The maximum thickness of the olivary profile is the same as that of the NACA airfoil, following Equation
(11). Both leading and trailing edges follow Equation (12), but with a circlar transition between upper and lower
surface. To make a fair comparison, the values of the tangent angle θs for the olivary leading and trailing edges are
varied to change the profile curve curvature, so that the olivary profile has the same area as the NACA airfoil when
their TCRs are the same. Besides, the profiles in the right-bottom of Table 2 in dark red are the NACA airfoil and
cylindrical struts with the same area as the baseline flat experimental VAWT model, which are also used to analyze
the effects caused by the profile geometry.
The three-dimensional model of the struts with different chord lengths are illustrated in Fig. 7. When
cs≤1.00cB, the ACs of the blade and the strut are overlapped. When cs>1.00cB, the strut will extend a distance so
that it wraps around both sides of the blade as seen in Fig. 7 (d), which is common in some small or experimental
wind turbines[65].
Tabs =3.3%H
Tabs =3.3%H
.5 0 cB 5 cB
S =0 cS=0.7
0A B c 5 AB
A S =0.5 AS =0.7
(a) (b)
Top View
Top View
Tabs =3.3%H
Tabs =3.3%H
0 cB
1.00 c
B
cS=1.5
0 AB cS= .50 A
B
A =1.0
S AS=1
(c) (d)
Fig. 7. Three-dimensional geometric model of struts with different chord lengths. The spanwise connecting position is the same as
the baseline model. (a) cS=0.50cB; (b) cS=0.75cB; (c) cS=1.00cB; (d) cS=1.50cB.
It should be noted that those extremely thin struts (with small TCR) in Table 2 are feasible for small VAWTs
because of their low aerodynamic forces and blade weight. Such thin struts have been used in many experimental
turbines[65, 66]. However, as the size of VAWT increases, the blade weight and aerodynamic forces will increase by
cubic and quadratic orders of magnitude of the size, respectively. Apparently, such thin struts are not practical for
large-scale VAWTs because it is difficult to fabricate and provide support. Nevertheless, for this paper, designing
these extremely thin strut profiles is still valuable, as this allows analyzing the effect of strut geometry on the blade
aerodynamic performance.
Moreover, in practice, large VAWT struts prefer variable cross-sections like HAWT blades, i.e., thicker
airfoils at the root to provide more strength, and progressively decreasing thickness along the spanwise direction
to reduce drag. However, this would lead to too many parameters and increase the complexity of the analysis, so
the constant cross-sectional area is used in this paper to facilitate comparison.
11
2.3.2. Strut Connecting Positions
(i) The Spanwise Position
As mentioned, this paper is aimed at the strut installed near the quarter span position. Mojtaba et al.[40] showed
that the structural stresses in the blade are minimize when the strut are installed at 0.2H position from the blade
tip. On the other hand, to facilitate comparison with the reference experimental model, the baseline VAWT struts
are installed at the 0.3H position in this paper.
Therefore, the effects of these two spanwise positions are compared here, as shown in Fig. 8. Eight struts are
modeled at the 0.2H position: the NACA airfoils and the olivary profiles (marked with pentagrams with absolute
thickness of 3.3%H in Table 2). Other connecting parameters are identical to 0.3H cases if the profile is the same.
z=0.2H
z=0.3H
Profiles
Profiles
(a) (b)
Fig. 8. Different spanwise connecting position of struts for various NACA and olivary profiles. All struts keep the same
absolute thickness of 3.3%H but have different chords. (a) z=0.3H. (b) z=0.2H.
(ii)The Chordwise Positions
Previous studies on the connecting position have focused on the relative position of the blade AC to the rotor
center. However, the effect of different chordwise connecting position in the 3D case is neglected when the strut
chord length is smaller than the blade chord length.
Therefore, the different strut chordwise connecting positions are modeled as illustrated in Fig. 9. Three
profiles of the baseline flat, NACA airfoil (two chord lengths of cS=0.50cB and cS=0.75cB, both AS=0.5AB) and the
olivary (cS=0.50cB, AS=0.5AB) are analyzed so as to exclude the influence of profile shape. The ACs of all blades
are located on the same radius. Here the dLE and dMTP are used to define the chordwise positions of struts. These
two parameters are helpful to investigate the flow mechanism for the different profiles, which will be discussed in
detail in later sections.
Aerodynamic Center c = 0.50c
S tru t Bla d e Profiles
3.3%H 3.3%H 1.0%H
Strut Max-Thickness Position 0.50c
d MTP =-0.125cB d MTP =0cB
d LE =0.125c B
dLE =0cB
d MTP =0.125cB d MTP =0.375cB
d LE =0.25cB d LE =0.5cB
cS tru t= 0.75cBla d e Profile

2.2%H
d MTP =-0.0625cB
d MTP =0cB
dLE =0cB d LE =0.0675cB

d MTP =0.0625cB d MTP =0.1875cB
d LE =0.25cB
d LE =0.125cB
Fig. 9. Different chordwise connecting position of struts. The dLE represents the distance between the blade leading edge and
the strut leading edge. The dMTP represents the distance between the strut MTP and the blade AC. For the situation of cS=0.50cB,
three profiles as shown in right-top are used. For the situation of cS=0.75cB, only the NACA profile is used.
2.3.3. Strut Connecting Angle
The effect of strut angle on the aerodynamic performance of the VAWT is analyzed according to the inclined
strut, as shown in Fig. 10. One principle of the inclined strut is to make the tower height in the middle of the blade.
This allows the load on the strut to be balanced. This principle leads to the strut angle being decided when the rotor
radius, the blade height and the spanwise connecting position are determined. In this paper, the inclined strut angle
12
is 15°. As in the case of the spanwise position investigation, a total of eight profiles of the VAWT model with
inclined strut are designed.
les
ofi
Pr
ζ=15°
Z
Y X
Fig. 10. The VAWT model with inclined struts. It is also simulated using symmetric boundary condition. Various NACA and olivary
profiles are used. All struts keep the same absolute thickness of 3.3%H but have different chords.
2.3.4. Strut-Blade Connecting Fairings
To analyze the effect of fairing, a circular transition surface at the connecting conner between the strut and
the blade is used. Hand et al. [24] suggested that the fillet radius r=0.06cB. Here two radiuses, r=0.06cB and r=0.12cB
are analyzed. The fairings for struts of different chord lengths are shown in Fig. 11. Among these the case of
cs=1.00cB is special. Since the circular transition surface results in a bunt trailing edge with a large thickness, a
distance must be extended to achieve a streamlined shape. This design is similar to the fairing in Ref. [60].
r=0.12cB r=0.12cB r=0.12cB r=0.12cB
Front View Front View Front View Front View
r=0.06cB r=0.06cB r=0.06cB
r=0.06cB
Front View Front View Front View
Tabs =3.3% H
Back View
Tabs =3.3%H
Tabs =3.3% H
Tabs =3.3%H
.50 c
B
5 cB 0 cB
S =0 S =0
.7 1.00 c cS=1.5
B
0A c 5A c
B
A cS= .50A
A =0.5 A =0.7
B B
0
A =1
B
S
(a)
S
(b) A =1.0
S
(c)
S
(d)
Fig. 11. The strut-blade connecting fairings for different strut chords and fairing radiuses. Both the NACA and olivary profiles are
used, but here only show the former. All struts also use the profiles with pentagrams as described in Table 2. (a) cS=0.50cB; (b)
cS=0.75cB; (c) cS=1.00cB; (d) cS=1.50cB.
2.4. CFD Model

2.4.1. Boundary Conditions and Mesh Topology
In this paper, CFD simulations are performed using Star-CCM+. Fig. 12 displays the computational domain
and boundary conditions. In the top left figure, 5D or 8D represent distance of five or eight times the rotor diameter,
which is far enough to minimize the boundary reflection effect [67]. The bottom surface uses the symmetry plane
condition to map the half blade as mentioned above. The left, right and top surfaces use the slip wall condition to
avoid the influence of wall boundary as Zhang et al. [13] did. The velocity inlet condition of 7 m/s inflow speed and
the pressure outlet condition for standard atmospheric pressure are employed on the front and back surfaces
respectively. The inner interface condition is used to connect the rotating region to the stationary region, while the
rotating region use a sliding mesh method to simulate rotor rotation. Other sub-figures show the mesh topology.
There are 24 layers of prism meshes around the blade surface with the first layer having a height of 0.0015 mm,
which makes the y+ ~1 to deal with the sub-layer of wall boundary layer. Three levels of local mesh refinements
13
are used to better capture the flow characteristics in the wake.
Fig. 12. An illustration of the computational domain, the boundary conditions and the mesh topology. The white dashed lines in
sub-figure show the local mesh refinement. Three levels of refinement in total, which are the blade near wake refined region, the
blade rotating wake refined region and the rotor wake refined region.
2.4.2. Solver Settings
Considering small velocity and the absence of heat diffusion in VAWT simulations, the implicit segregated
flow model of an incompressible constant density gas is used to solve the flow equations without energy. The
pressure-velocity coupling is combined with a SIMPLE-type algorithm and the second-order upwind schemes is
applied for velocity, pressure, and turbulence. The second-order upwind scheme is set to the convection term. In
the unsteady simulation, a second-order central difference scheme is used for temporal time discretization. All
simulations were adopted 10 inner-iterations at each time step and were carried out on in-house parallel computing
server with 10 nodes. Node 1st is used for system management and the remaining nine nodes can be used for CFD
calculations. Each node is equipped with two Intel (R) Xeon (R) Silver 4116 CPU @ 2.10 GHz processors with
eighteen cores each. Each node has 128GB of memory. For a single VAWT model, a complete converged CFD
simulation with a single node of 36 cores takes approximately 80 hours. If three nodes of 108 cores are used, the
computing time for a single case can be reduced to 24 hours.
2.5. Validation of Simulations

Verifying numerical models to improve their reliability is essential for simulation studies. Our previous
study[12] provided comprehensive numerical simulation validations, including grid convergence, time step, rotor
revolution, turbulence model and comparison with experimental values. The numerical model for this paper uses
the same strategy as the previous ones. Therefore, a brief description of the numerical model setup is presented
here. The complete validation details can be found in Ref. [12].
2.5.1. Grid Convergence Study
[68]
Methods for examining the grid convergence of CFD simulations are presented in the book by Roache ,
which is based on the Richardson's extrapolation method. It should perform simulations on several successively
[69]
finer grids. Zadeh et al. provided a detailed process for VAWT meshes. Table 3 shows the grid convergence
study of the VAWT without struts at two TSRs in the validation study[12].
The ARC of 1.0048 for λ=1.85 and 1.0329 for λ=2.29, are close to 1, indicating that the results are in the range
14
fine
of asymptotic convergence. The maximum error GCI i +1,i are 0.014% for λ=1.85 and 0.497% for λ=2.29, both
simulations are below than 3%. This indicates that further refinement of the mesh size does not markedly improve
the accuracy of the calculations, and the medium mesh is fine enough. Therefore, all VAWTs with struts use the
medium mesh strategy in this paper.
Table 3. Discretization error for λ=1.85 and λ=2.29. The critical variable fi is used to verify the influence of mesh refinement
based on the simulation result. The power coefficient (Cp) of VAWT is used here. All parameters on the left are coefficients required
for GCI calculation.
Value
Parameters
λ=1.85 λ=2.29
Fine Mesh Number (N1) 5,319,717
Medium Mesh Number (N2) 3,589,852
Coarse Mesh Number (N3) 2,916,044
r21 1.5 1.5
r32 1.5 1.5
f1 0.1416 0.3085
f2 0.1410 0.2987
f3 0.1110 0.2102
ε32 -0.0300 -0.0885
ε21 -0.0007 -0.0098
p 9.3636 5.4200
R 0.0224 0.1111
ea21 0.475% 3.185%
ea32 21.252% 29.619%
GCI 21Fine 0.014% 0.497%
GCI 32Medium 0.610% 4.626%
ARC 1.0048 1.0329
2.5.2. Time Step and Revolution Verification

According to the study of Rezaeiha et al. [70], the azimuthal increments corresponding to the time step and
rotor revolutions will affect the accuracy of the VAWT simulations.
For wind turbine simulations, the azimuthal increment associated with the rotating speed Ω is used to verify
the time step. Four different azimuthal increments include Ωdt=4°, 2°, 1° and 0.5° were tested[12]. Considering the
computational efficiency and simulation accuracy, the azimuthal increment Ωdt=1° was used finally.
The revolution verification of wind turbine simulations is essential because the wake from the upwind moves
downstream and thereby affecting the aerodynamic performance of the VAWT blades. Therefore, the number of
rotor revolutions should be tested to ensure the simulation obtains a statistically steady state. The revolution was
verified by monitoring the variation of CP. Most cases get a statistically steady state after the 6th revolutions[12],
which is the same as Ref. [13]. But for some cases with fluctuations, a few additional revolutions were calculated.
2.5.3. Turbulence Model Verification
The turbulence model has a remarkable influence on CFD simulations. It is recognized that the k-ω shear
stress transport (SST) model in the unsteady Reynolds-averaged Navier-Stokes (URANS) equations is suitable for
VAWT simulations [71, 72]. However, it is mostly studied in 2D simulations. Rezaeiha et al.[73] showed that the
accuracy of URANS is not sufficient for 3D VAWT simulations, while the hybrid RANS/LES model can better
capture the dynamic stall and the vortex characteristics in the wake. Therefore, the Improved Detached Eddy
Simulation (IDDES) [74] model based on the k-ω SST model and RANS turbulence model were verified. Previous
turbulence model verification showed that IDDES does have better agreement with the experimental data[12, 75].
Therefore, the IDDES model is adopted in this paper.
2.5.4. Validation Using Experimental Data
Finally, the simulation results are verified with experimental data. Fig. 13 shows the pressure coefficients CPre
at the blade center position z=0h for different azimuthal angles when λ=2.29. The inflow velocity V∞=7 m/s was
15
used to calculate the pressure coefficient. The present CFD result uses the medium mesh with 1°azimuthal
increment, IDDES model and the sampled data of the 6th revolution. As illustrated in Fig. 13, the pressure
coefficients show a good arrangement with the experiment at most azimuthal angles.
-30 Exp -30 Exp -30 Exp -30 Exp

-25 θ=0° CFD [Li et al.]
-25 θ=30° CFD [Li et al.] -25 θ=60° CFD [Li et al.]
-25 θ=90° CFD [Li et al.]
Present CFD Present CFD
Pressure Coefficient-CPre
Present CFD Present CFD
-20 -20 -20 -20
-15 -15 -15 -15
-10 -10 -10 -10
-5 -5 -5 -5
0 0 0 0
5 5 5 5
10 10 10 10
15 15 15 15
0 0.2 0.4 0.6 0.8 1 0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
x/c x/c x/c x/c
-30 Exp -30 Exp -30 Exp -30 Exp
-25 θ=120° CFD [Li et al.]
-25 θ=150° CFD [Li et al.]
-25 θ=180° CFD [Li et al.]
-25 θ=210° CFD [Li et al.]
Present CFD Present CFD Present CFD Present CFD
-20 -20 -20 -20
-15 -15 -15 -15
-10 -10 -10 -10
-5 -5 -5 -5
0 0 0 0
5 5 5 5
10 10 10 10
15 15 15 15
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
x/c x/c x/c x/c
-30 Exp -30 Exp -30 Exp -30 Exp
-25 θ=240° CFD [Li et al.]
-25 θ=270° -25 θ=300° CFD [Li et al.]
-25 θ=330° CFD [Li et al.]
Present CFD Present CFD Present CFD
-20 -20 -20 -20

-15 -15 -15 -15
-10 -10 -10 -10
-5 -5 -5 -5
0 0 0 0
5 5 5 5
10 10 10 10
15 15 15 15
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
x/c x/c x/c x/c
Fig. 13. Pressure coefficient distribution on the blade center position z=0h, λ=2.29. The black dot and red line are the experimental
data and CFD simulation result of Li et al. [61]. The blue lattice proves the accuracy between present CFD simulation and the
experimental value.
The torque coefficients of the experiment and the baseline model with flat struts at different heights are shown
in Fig. 14. At z=0.05h, the baseline model is in good agreement with the experiment data. As the height close to
the blade tip, the simulated CQ are slightly higher. This variation may be caused by the different positions of the
strut connection. However, the phenomenon that CQ significantly decrease near the strut is the same in both the
experiment and the simulation.
0.35 Exp_0.05h
Exp_0.55h
0.3 Exp_0.70h
Exp_0.80h
Torque Coefficient CQs (-)
0.25 BaselineFlatStrut_0.05h
BaselineFlatStrut_0.55h
0.2 BaselineFlatStrut_0.7h
BaselineFlatStrut_0.8h
0.15
0.1
TSR=2.29
0.05
0
-0.05
0 30 60 90 120 150 180 210 240 270 300 330 360
θ (◦)
Fig. 14. Torque coefficients (CQ) of the experiment and the baseline model with flat struts at different blade pressure tap heights.
The above validation studies prove that the CFD simulation in this paper has a good accuracy compared with
the experimental data, which guarantees the effectiveness of the subsequent simulations.
3. Results and Discussion

3.1. The Strut Effects
The variation of the aerodynamic performances of VAWTs with and without struts are analyzed firstly. Three
baseline struts in the right-down of Table 2 are compared here. Their cross-sectional profiles are flat bar, NACA
16
airfoil and cylinder respectively.
(i) Aerodynamic Performance
The power coefficients of the baseline wind turbine at different TSRs are shown in Fig. 15. Two important
results can be obtained:
⚫ The presence of struts substantially reduces the VAWT efficiency. In sub-figure (a), it is observed that the
CP is overestimated when the strut (green curve) is neglected in the numerical simulation, which is one of
the reasons why many 2D simulations are much higher than the experimental values. When the simulation
considers struts, the total CP of the whole baseline VAWT (red curve, the total torque including struts and
tower) and the blade CP (blue curve, calculating torque only from blades) show good agreement with the
trendy of the experimental value (black triangles, integrating the value from the blade pressure taps). This
also proves the accuracy of the numerical simulations in this paper. Besides, the experiment also tested the
CP of the whole VAWT (yellow diamond) by means of a six-component balance. However, these values are
even lower since there is mechanical damping between experimental parts, which leads to a further power
loss. This loss is difficult to be calculated in CFD simulations.
⚫ For these three struts with the same cross-sectional area, it indicates that the strut cross-section profile
shapes have a significant influence on the aerodynamic efficiency, as shown in sub-figure (b). The ∆CP-Profile
and ∆CP-Induced values are close for the non-streamlined struts (the flat and cylinder), while both the profile
and induced drags of the cylinder strut are significant, resulting in a negative total efficiency. For the
streamlined airfoil strut with similar chord and thickness to the flat strut, both direct and induced drag can
be reduced, especially for the former.
0.32
0.35 Exp_6-Component Blance
Exp_Interal for Spanwise (a) (b)
Without Struts 0.28 - 0 .0 2 4 6
- 0 .0 3 1 6
0.3 BS_Flat-OnlyBlades
- 0 .0 1 4 7
BS_Flat-Total 0.24
∆CP-Induced=-11.21%
Power Coefficient - Cp (-)
- 0 .0 4 0 2
∆CP-Total=-25.47% - 0 .1 6 1 8
0.25 0.2
Power Coefficient - Cp (-)
∆CP-Profile=-16.07%
0.16
0 .2 8 1 6
0.2
0.12
0 .2 8 1 6 0 .2 8 1 6
0 .2 5 0 0 0 .2 5 7 0
0.15 0.08 0 .2 0 9 8
0 .2 4 2 2
0 .11 9 9 7- 0 .1 7 5 3
0.04
0.1
0
Flat Airfoil Cylinder
WithoutStruts WithStruts_Onl yBla de s WithStruts_Tot al - 0 .0 5 5 6
0.05 -0.04
C P - I n ter f er en ce C P - Pr o f ile
1.8 1.9 2.0 2.1 2.2 2.3 2.4 2.5 2.6
TSR -0.08 TSR=2.29
Fig. 15. Power coefficients of the baseline VAWTs. (a) CP of the flat strut VAWT at different TSR. The black triangles are the
integral of the spanwise pressure taps in experiment. The yellow diamonds are experimental data obtained from 6-component
balance. Other colors are simulated results. (b) CP of the simulated baseline VAWTs with different strut profiles at TSR=2.29. ∆CP-
Induced indicates the efficiency reduction caused by the variation of the blade torque in the presence of the strut, which is the induced
drag. ∆CP-Profile indicates the efficiency reduction due to the drag of the strut itself, which is the direct drag.
To investigate the influence of strut at different azimuth angles, the CQ of a single VAWT blade without and
with struts is displayed. Fig. 16 shows the torque coefficients of the whole blade, while Fig. 17 shows the torque
coefficients for each segment at different blade heights (20 segments in total). It can be observed that:
⚫ The distributions of the two drags are completely different at various azimuth angles, as shown in the light
red and light blue areas in Fig. 16, respectively. The induced drag is concentrated at 60○<θ<180○ (when the
strut is on the suction side of the VAWT blade), while the profile drag is more significant at 0○<θ<30○ and
270○<θ<360○ (when the relative wind speed of the strut is higher).
⚫ The strut affects a wide range in the spanwise direction of the blade. Comparing Fig. 17 (a) and (b), it is
observed that the blade CQ has a significant reduction at the strut connecting position. Fig. 17 (c) shows the
variation of CQ for VAWT blade with and without struts. This variation represents the induced drag caused
17
by the flat strut. It is observed that in the upwind region, the influence of the strut spreads from the blade
middle (0.0h) to the tip (1.0h), despite its installation at 0.45h. In the downwind region, the influence of the
strut is relatively small, and there is even a slight increase in CQ near the blade tip part at 240○<θ<300○.
This increase is most likely due to the local velocity variation caused by the blade tip vortex.
0.30
TSR=2.29 WithoutStruts
0.25 BaselineFlatStruts-OnlyFromBlade
BaselineFlatStruts-Total
0.20
CQ-Interal for Spanwise (-)

Loss caused by the induced drag
0.15
0.10
Loss caused by the profile drag
0.05
0.00
-0.05
0 30 60 90 120 150 180 210 240 270 300 330 360
θ (°)
Fig. 16. Torque coefficients of the VAWT single blade without and with strut at TSR=2.29. The light red and light blue areas
indicate the torque losses caused by the induced drag and the profile drag, respectively
(c) Difference of CQ at Various Hegihts △ CQ-Induced

1.0 0.09
0.9 0.07
0.8 0.05
0.7 0.03
0.6 0.01
z/h
0.5
-0.01
0.4
-0.03
0.3 Strut Height
-0.05
0.2
-0.07
0.1
-0.09
0 30 60 90 120 150 180 210 240 270 300 330 360
TSR=2.29
θ (°）
Fig. 17. Torque coefficients of single blade varying along the spanwise direction at TSR=2.29. (a) the blade without struts; (b) the
blade with flat struts, but the CQ values are only from blade; (c) Difference of torque coefficients (∆CQ-Induced) between the blades
without and with flat struts. The blue area shows the spatial location where the blade torque coefficient decreases, and the
boundaries of this area are identified using contours.
(ii) Fluid Field Characteristics
The effects of the struts on the fluid field characteristics and the blade surface streamlines are displayed in
Fig. 18. The surface streamlines can be used to determine whether the fluid attaches to the blade surface to generate
aerodynamic forces. The cylinder strut is also shown here. From the perspective of the pressure and suction
surfaces of the VAWT blade:
⚫ For the pressure side, the influence of the strut is not significant as seen by the blade surface limited
streamlines in Fig. 18 (a), (c) and (e). In the downwind region at θ=270°, the flat and cylinder struts only
interfere with the blade pressure surface flow within their thickness range. In the upwind region at θ=90°,
both struts barely affect the blade surface flow.
⚫ For the suction side, the impact of the struts is extremely pronounced. In the upwind region at θ=90° , the
flow separation occurs at the strut position, resulting in a reduction of the blade aerodynamic force. This
separation is much larger than the thickness range of the strut, which is clearly related to the shedding vortex
scale induced by the strut cross section shape. In the downwind region at θ = 270°, the strut still affects the
blade suction side, although it is located on the opposite (pressure) side at this point. Importantly, it is
observed in the lower left corner of Fig. 18 (b) and (d) that the streamlines near the flat strut blade tip are
instead better attached than the blade without strut, as confirmed by the CQ curves in Fig. 16. For the cylinder
strut, the flow on the suction side is almost completely separated at θ = 270°. Combined with the vorticity
field, the presence of the struts significantly aggravates the vortex scale and intensity of the VAWT wake.
As seen above, for the VAWT with struts, the reduction in aerodynamic performance is primarily caused by
18
the struts interfering with the blade suction surface in the upwind region, while in the downwind region it is caused
by the changes in the wake induced by the struts.
Fig. 18. Flow characteristics of the VAWTs without struts, with flat struts and with cylinder flats at TSR=2.29. The surfaces of
blades and struts are colored in static pressure and covered with the limited streamlines. The z-vorticity at z=0h, z=0.55h, z=0.7h
and z=0.8h in the fluid field are also displayed to show the wake.
3.2. Influences of Strut Profiles

The influence of the strut profiles is discussed here, focusing on the variations of chord, thickness, and cross-
sectional area. In addition, the performance of the specialized strut profile proposed in this paper is analyzed.
3.2.1. Effects of Chords, Thicknesses and Areas
Fig.19 illustrates the CP of VAWTs for NACA airfoil struts with different chords, thicknesses and areas.
Several results are observed:
⚫ The VAWT efficiency decreases obviously with the increase of the strut cross-sectional area. For the struts
with same area, their profile drags increase as the relative thickness increases (with a corresponding
decrease in chord), which essentially follows an exponential power low, as shown in Fig. 19 (b). This is
determined by the drag characteristics of the NACA airfoil itself[76]. However, an interesting phenomenon
is that, when AStrut/ABlade≤1.0, the total CP suddenly decreases at cStrut/cBlade=1.0, as shown by the red circle
in Fig. 19 (a). In addition, this phenomenon is also observed for struts with the same absolute thickness
(TStrut/HBlade=3.3%) but different chord lengths
⚫ The drag magnitude of the VAWT strut is approximately linearly related to the strut absolute thickness. As
illustrated in Fig. 19 (c), all CP values representing the NACA airfoil struts are located within the narrow
red region. This linear relationship is not obvious if the horizontal coordinates are replaced by relative
thicknesses, as seen in Fig. 19 (d).
19
0.26 1.4% 0
(a) (b) 1.4% 2.2% 1.7% 1.3% 1.7%
1.0% 1.3% 1.7%
0.22 -0.02 3.3% 2.2% 1.8%
2.2% 1.7% 1.0% 2.9% 2.2%
3.3% 1.8% 2.0%
0.18 2.9% 2.2% -0.04 4.1%
4.4% 2.0% 2.2% 2.5%
3.3%
0.14 4.1%
-0.06 2.6%
CP_Total（-）
2.5%
3.3% 3.3%
2.6%
∆CP_Profile (-)
4.4% 3.3% 3.3%
0.1 -0.08 4.4%
4.4% 3.3% 5.0%
5.0% 4.0%
0.06 4.0% -0.1
5.0%
0.02 5.3% BS_Flat -0.12
5.0% BS_Airfoil BS_Flat
BS_Airfoil
-0.02 AStrut/ABlade=0.5 -0.14 5.3% AStrut/ABlade=0.5
AStrut/ABlade=0.67 6.6% AStrut/ABlade=0.67
AStrut/ABlade=0.75 AStrut/ABlade=0.75
-0.06 6.6% 6.6% AStrut/ABlade=1.0 -0.16 AStrut/ABlade=1.0
AStrut/ABlade=1.5 6.6% AStrut/ABlade=1.5
-0.1 -0.18
0.25 0.5 0.75 1 1.25 1.5 0.25 0.5 0.75 1 1.25 1.5
cStrut/cBlade (-) cStrut/cBlade (-)
0.26 0.5 1 0.75 0.26 0.5
1.25 (c)
0.5 0.75 0.5 (d)
0.22 1.5 0.5 1.5
0.5
0.22 1.25 0.75 0.5
1 0.75
0.75 0.5 BS_Flat
0.18 0.4 0.18
1.5 1.5 BS_Airfoil
0.75 AStrut/ABlade=0.5
0.75
0.14 0.14 1.5 AStrut/ABlade=0.67
CP_Total（-）
0.75 0.4
CP_Total（-）
1.25 1.25 AStrut/ABlade=0.75
0.1 1 AStrut/ABlade=1.0
1.25
0.5 0.1 AStrut/ABlade=1.5
1.25
1.25 1
0.06 1 1 0.06 1 0.5
1
1 1.25
1.5
0.02 BS_Flat 0.02 0.4
0.75 1 1.25
BS_Airfoil
0.75
-0.02 AStrut/ABlade=0.5 -0.02
AStrut/ABlade=0.67 1.25
AStrut/ABlade=0.75 0.5 0.5
-0.06 AStrut/ABlade=1.0
0.4 -0.06
AStrut/ABlade=1.5
-0.1 -0.1
0.5% 1.5% 2.5% 3.5% 4.5% 5.5% 6.5% 7.5% 0.0% 10.0% 20.0% 30.0% 40.0% 50.0% 60.0% 70.0%
TStrut/HBlade (-) TStrut/cStrut (-)
Fig. 19. CP of VAWTs with different chord lengths and thicknesses NACA airfoils struts. The bubble diameter represents
AStrut/ABlade, (a) The total CP varies with cStrut/cBlade, the label value represents the absolute thickness ratio TStrut/HBlade. (b) The power
losses caused by the strut profiles drags. (c) The total CP varies with TStrut/HBlade, the label value represents the relative thickness
cStrut/cBlade. (d) The total CP varies with the relative thickness TStrut/cStrut.
The blade CQ curves for the struts with same areas AStrut=0.75ABlade but different chords are shown in Fig. 20
(a). When cS=1.00cB, the blade torque in the upwind region is significantly reduced, even lower than the case of
cS=0.50cB with the maximum relative thickness. Fig. 20 (b) shows the difference in CQ at various heights for
cS=1.00cB and cS=0.75cB, and it indicates that the torque reduction in the former occurs mostly on the blade surface
above the strut when 60○<θ<180○.
Difference of CQ at Various Hegihts △ CQ
0.25 1.0 0.13
(a) cS=0.50cB (b)
cS=0.75cB 0.9
0.10
0.2 cS=1.00cB 0.8
0.07
cS=1.50cB
CQ-Only From Blade
0.7
0.04
0.15
0.6
0.01
λ=2.29
z/h
0.5
0.1 -0.01
AStrut=0.75ABlade
0.4
Only Form Blades -0.04
0.05 0.3 Strut Height
-0.07
0.2
-0.10
0
0.1 △ CQ=CQ-C =1.00C - CQ-C =0.75C
0 30 60 90 120 150 180 210 240 270 300 330 360 S B S B
-0.13
0 30 60 90 120 150 180 210 240 270 300 330 360
-0.05
θ (◦) θ (°）
Fig. 20. Torque coefficients (CQ) when AStrut=0.75ABlade. (a) CQ of the single blades for different strut chords. (b) ∆CQ is the
difference between the blade struts cS=0.75cB and cS=1.00cB along the blade spanwise direction.
To investigate the effect of strut size on the flow characteristics, the surface limited streamlines and Q-
criterion iso-surfaces for blades with different chords (and thickness) are extracted at θ = 120°, as shown in Fig.
21. This azimuth angle is selected because the CQ curves indicate a large difference in the performance of each
blade at this upwind position. Several results are found:
⚫ The strut chord is an important factor affecting the flow characteristics of the blade surface. By vertically
comparing any two cases, such as (a) with (e), (b) with (f), etc., it is observed that the limited streamlines
and the shedding vortices around the struts are similar as long as the struts have the same chord, regardless
20
of whether the thickness (or cross-sectional area) is the same.
⚫ When the strut chord is equal to the blade chord, i.e., cS=1.00cB, the flow separation on the blade surface is
remarkable. As seen in Fig. 21 (c) and (g), the vortex scales and their influence range on the blade surface
along the spanwise direction are the largest among these cases. Compared to the streamlines, the separation
occurs earlier in the region of blade leading edge above the strut when cS=1.00cB, as illustrated in the red
dashed lines.
⚫ When the strut chord is longer than the blade chord, the vortex formed at the connection is significantly
reduced. Comparing Fig. 21 (e)~(h), the increase in chord increases the cross-sectional area for the same
absolute thickness, but the streamlines at cS=1.5cB is similar to that at cS=0.50cB, with the former having
less separation at the blade leading edge. This phenomenon implies that the strut chord length exceeding
the blade chord length may also be a good choice.
Fig. 21 Flow characteristics for the VAWTs with different struts. The iso-surface of Q-Criterion (Q=5 s-2) colored by the velocity
magnitude are used to display the vortex around the blade. (a)~(d) Struts have same areas of AStrut=0.75ABlade. (e)~(h) Struts have
same absolutely thickness of TStrut=3.3% HBlade.
3.2.2. The Olivary Profiles
Based on the characteristics of the VAWT strut, an olivary-shaped specialized profile is proposed in this paper.
Several olivary struts with different areas and chords in Table 2 are calculated. The power coefficients of these
VAWTs are shown in Table 4. The diagonal lines in the table show several cases with same absolute thicknesses
of 3.3%H. Fig. 22 shows the variations of the CP of VAWTs with two type struts relative to that of VAWTs without
struts. It is observed that:
⚫ The power coefficients of the VAWTs with olivary struts are better than those of the NACA airfoil in all
cases. Comparing the CP-Total and the CP-Blade for both profiles, the olivary struts not only have less profile
drag than the NACA airfoil, but also less induced drag. Compared to the blades without struts, for the eight
cases in Table 4, the CP-Total of the turbines with NACA struts are reduced by an average of 39.4%, while
the turbines with olivary struts are reduced by a averaged of 29.2%. Thus, the drag of the olivary strut is
reduced by 25.8% relative to the NACA strut.
⚫ The olivary struts with longer chord than the blade chord has less drag. As seen in Fig. 22, the power
reductions of NACA airfoil and olivary struts are close when cS<1.00cB. When cS≥1.00cB, the drag of
olivary strut can be reduced significantly. When cS=1.50cB, the total CP of the olivary strut VAWT is reduced
by 22.4% compared to the blades without struts. Although the CP still has significant reduction, it is superior
to most struts and even the baseline VAWT with the thin flat strut as seen Fig. 15. What’s more, the cross-
sectional area of the olivary strut is 7 times larger than that of the flat strut. This indicates that the olivary
21
struts can provide higher strength while keeping small drag characteristics.
Table 4. Comparison of power coefficients of NACA airfoils and the olivary profiles for different chords and thicknesses
corresponding to the left-bottom part of Table 2. The upper half of this table is the total power coefficient of the VAWT (including
the struts and the tower). The bottom half is the power coefficient only produced by the blades. The numbers colored by red and
blue are the NACA and olivary profiles with the same absolute thickness, respectively.
CP_Total cS=0.50cB cS=0.75cB cS=1.00cB cS=1.50cB
0.1819 0.1774 NACA
AS=0.50AB
0.1907 0.1998 Olivary
0.1473 0.1291 NACA
AS=0.75AB
0.1700 0.2147 Olivary
0.0973 0.0848 NACA
AS=1.00AB
0.1192 0.1690 Olivary
-0.0350 0.1391 NACA
AS=1.50AB
-0.0014 0.2185 Olivary
CP_Blades cS=0.50cB cS=0.75cB cS=1.00cB cS=1.50cB
0.2145 0.2066 NACA
AS=0.50AB
0.2252 0.2226 Olivary
0.1909 0.1734 NACA
AS=0.75AB
0.2028 0.2267 Olivary
0.1584 0.1446 NACA
AS=1.00AB
0.1690 0.2013 Olivary
0.0893 0.1860 NACA
AS=1.50AB
0.1114 0.2378 Olivary
cStrut/cBlade (-) cStrut/cBlade (-)

0.25 0.5 0.75 1 1.25 1.5 0.25 0.5 0.75 1 1.25 1.5
0.25 0.25
-35.4% -23.8%
-37.0% ∆C P-Total_NACA -26.6% ∆C P-Blades_NACA
∆C P-Total_Olivary ∆C P-Blades_Olivary
0.5 0.5
-29.0% -20.9%
-32.3% -20.0%
AStrut/ABlade（-）
AStrut/ABlade（-）
-47.7% -54.1% -32.2% -38.4%

0.75 0.75
-65.4% -39.6% -28.0% -19.5%
-43.7% -28.0%
-69.9% -48.6%
1 1
-59.6% -40.0% -41.1% -28.5%
1.25 -22.4% 1.25 -15.5%
-112.4% -100.5% -68.3% -60.4%
(a) -49.6% (b) -33.4%
1.5 1.5
Fig. 22. Relative variations in CP for VAWTs with and without struts, where ∆CP = (CP-WithStruts - CP-WithoutStruts)/CP-WithoutStruts. The
hollow circles represent CP of VAWTs with NACA airfoil struts. The solid circles represent CP of VAWTs with olivary profiles. The
labels represent relative variation values. (a) The total power coefficients of VAWTs. (b) The power coefficients only from blades.
The CQ curves of blades installed these two strut profiles and their variation along the spanwise direction are
illustrated in Fig. 23. It shows that the induced drag of the olivary strut is smaller (higher CQ than the NACA airfoil
strut). The torque improvement mostly occurs in the upwind region at 60○<θ<180○ and in the downwind region at
240○<θ<300○. When cS<1.00cB, the torque improvement of the blade with olivary strut blade only occurs in a
small part of the blade and is not significant. When cS≥1.00cB, the CQ of the olivary strut blade is significantly
increased up to 0.23.
22
0.25 0.25 0.25 0.25
(b) Olivary (c) Olivary Olivary
(a) Olivary
NACA NACA (d)
NACA 0.2 NACA
0.2 0.2 0.2
CQ-Only From Blade
CQ-Only From Blade
CQ-Only From Blade

CQ-Only From Blade
0.15 0.15 0.15 0.15
λ=2.29 λ=2.29 λ=2.29 λ=2.29

0.1 0.1 0.1 0.1
AS=0.50AB _cS=0.50cB AS=0.75AB _cS=0.75cB AS=1.00AB _cS=1.00cB AS=1.50AB _cS=1.50cB
0.05 0.05 0.05 0.05
0 0 0 0
0 60 120 180 240 300 360 0 60 120 180 240 300 360
0 60 120 180
θ (◦)
240 300 360
-0.05
θ (◦) θ (◦) 0 60 120 180
θ (◦)
240 300 360
-0.05 -0.05 -0.05
TSR=2.29, AS=0.50AB ,CS=0.50CB △ CQ TSR=2.29, AS=0.75AB ,CS=0.75CB △ CQ TSR=2.29, AS=1.00AB ,CS=1.00CB △ CQ TSR=2.29, AS=1.50AB ,CS=1.50CB △ CQ
1.0 0.23 1.0 0.23 1.0 0.23 1.0 0.23
0.9 (e) 0.18 0.9 (f) 0.18 0.9 (g) 0.18 0.9 (h) 0.18
0.8 0.13 0.8 0.13 0.8 0.13 0.8 0.13
0.7 0.08 0.7 0.7 0.08 0.7
0.08 0.08
0.6 0.03 0.6 0.6 0.6
0.03 0.03 0.03
z/h
z/h
z/h
z/h
0.5 -0.03 0.5 0.5 0.5
-0.03 -0.03 -0.03
0.4 0.4 0.4 0.4
-0.08 -0.08 -0.08 -0.08
Strut Height Strut Height Strut Height
0.3 0.3 0.3 0.3 Strut Height
-0.13 -0.13 -0.13 -0.13
0.2 0.2 0.2 0.2
-0.18 -0.18 -0.18 -0.18
0.1 △ CQ=CQ-Olivary - CQ-NACA 0.1 △ CQ=CQ-Olivary - CQ-NACA 0.1 △ CQ=CQ-Olivary - CQ-NACA △ CQ=CQ-Olivary - CQ-NACA
-0.23 -0.23 0.1
-0.23 -0.23
0 60 120 180 240 300 360 0 60 120 180 240 300 360 0 60 120 180 240 300 360 0 60 120 180 240 300 360
θ (°） θ (°） θ (°） θ (°）
Fig. 23. Comparison of CQ (only from blades) between the NACA and olivary struts with same absolutely thicknesses of
TStrut=3.3% HBlade. (a)~(d) CQ of the single blade. Different color of these figures corresponding to Table 2. (e)~(h) Difference of
blade torque coefficients (∆CQ) between the NACA and olivary strut along the blade spanwise direction.
The effects of two types of strut profiles, the NACA airfoil and the olivary, with different chords but the same
absolute thickness on the flow characteristics is illustrated in Fig. 24. It shows that in sub-figure (a), the limited
streamlines are stagnant at the blade leading edge, which is caused by the large leading edge radius of the NACA
airfoil. For the olivary profile in sub-figure (e), the smaller leading edge radius reduces the separation and
contributes to mitigate the interference between the strut and the blade. This mitigation effect of the olivary strut
is more pronounced when cS=0.75cB and cS=1.00cB. There is better attached flow in the upper part of the olivary
strut blade and the scale of vortex is significantly reduced.
Fig. 24 Comparison of flow characteristics between the VAWTs with different strut profiles. All profiles have the same absolutely
thicknesses of TStrut=3.3% HBlade (a)~(d) The NACA airfoil struts. (e)~(h) The olivary struts.
3.3. Influences of Strut Connecting Positions

This section shows the effects of the strut connecting positions, discussed in two aspects: the spanwise
position and the chordwise position.
3.3.1. Effects of Spanwise Positions
The structural analysis by Mojtaba et al.[40] suggested that the strut should preferably be installed at 0.2H
from the blade tip, while the experimental VAWT strut referenced in this paper are at 0.3H. Therefore, the effect
of these two spanwise connecting positions is investigated. The variations of the VAWT power coefficients are
shown in Fig. 25. Both the NACA airfoil and the olivary profiles are analyzed.
It is observed that the spanwise connecting position closed to the blade tip is beneficial to increase the CP of
23
the NACA airfoil strut VAWT, but the effect on the olivary strut VAWT is negligible. In addition, as seen from the
difference between CP_Blade and CP_Total in Fig. 25, the variation of the strut spanwise connecting position primarily
affects the induced drag, while it has less effect on the profile drag.
A Strut/A Blade=0.50 _C P-Total A Strut/A Blade=0.50 _C P-Blades A Strut/A Blade=0.50 _C P-Total A Strut/A Blade=0.50 _C P-Blades
(a) NACA A Strut/A Blade=1.50 _C P-Total A Strut/A Blade=1.50 _C P-Blades (b) Olivary A Strut/A Blade=1.50 _C P-Total A Strut/A Blade=1.50 _C P-Blades
0.25 0.25
0.225 0.225
0.2 0.2
0.175 0.175
0.15 0.15
CP (-)
CP (-)
0.125 0.125
0.1 0.1
0.075 0.075
0.05 0.05
0.025 0.025
0 0
0.3H 0.2H 0.3H 0.2H
Distance form blade tip (-) Distance form blade tip (-)
Fig. 25 The power coefficients of VAWTs with different strut spanwise connecting positions. The solid histograms represent the
total CP. The hollow histograms represent the CP only form blades. (a) The struts use NACA airfoil profiles. (b) The struts use
olivary profiles.
The flow characteristics of the different chord struts installed at 0.2H are shown in Fig. 26. Comparing with
the cases of 0.3H struts in Fig. 24, the spanwise connecting position at 0.2H facilitates the attachment of air in the
region below the struts, especially for the NACA airfoil struts with cS<1.00cB. This is because the large-scale
interference generated at the NACA airfoil strut connection overlaps with the spanwise flow induced by the tip
loss effects, thus indirectly minimizing the interference below the strut. For the olivary struts, they are insensitive
to the variations in spanwise position because their interference with the blades is inherently insignificant.
Fig. 26 Flow characteristics of the VAWTs with different profiles of struts installed at z=0.2H. All profiles have the same absolutely
thicknesses of TStrut=3.3% HBlade (a)~(d) The NACA airfoil struts. (e)~(h) The olivary struts.
3.3.2. Effects of Chordwise Positions
The chordwise connecting positions has not been discussed in past studies when the strut chord is smaller
than the blade chord. Therefore, several VAWTs with different strut chordwise connecting positions were
established. Fig. 27 shows the power coefficients of these VAWTs.
The CP are classified by the parameter dLE firstly, as shown in Fig. 27 (a). However, it is difficult to find a
general rule for the influence of the strut chordwise position on the aerodynamic performance. In particular, when
the strut is changed to olivary profile, there is a remarkable difference with the NACA airfoil strut in the same
positions. This raises an issue that why the same area, chord, maximum thickness (either relative or absolute) and
24
connecting position of the olivary and NACA airfoil struts result in a significant reduction of induced drag for the
former.
Observing these two profiles, it is noted that in addition to the difference in leading edge radius, another
notable difference is the location of the maximum thickness. Therefore, another parameter dMTP, representing the
relative distance between the strut MTP and the blade AC is used to identify the different chordwise positions, as
displayed in Fig. 27 (b). Two rules are observed here:
⚫ The aerodynamic efficiency of the VAWT is minimized when the strut MTP coincides with the blade AC
(dMTP=0c), as shown by the red dashed line in Fig. 27 (b). This phenomenon is remarkable for the blade
with NACA and olivary struts, but not for the baseline flat strut (black histogram). This is explainable
because the flat strut does not have a maximum thickness. The CP of the baseline flat strut VAWT improves
with increasing distance from the blade leading edge.
⚫ The aerodynamic performance of all VAWTs is significantly improved when the trailing edge of the strut
coincide with the trailing edge of the blade, which is the farthest position of each strut from the blade AC,
as illustrated by the marked pentagrams in Fig. 27 (b).
0.27 0.27 Baseline_Flat_Total Baseline_Flat_OnlyFromBlades
Baseline_Flat_Total Baseline_Flat_OnlyFromBlades (a) (b)
0.26 0.26 NACA_0.50c_Total NACA_0.50c _OnlyFromBlades
NACA_0.50c_Total NACA_0.50c _OnlyFromBlades
NACA_0.75c _Total NACA_0.75c _OnlyFromBlades
0.25 NACA_0.75c _Total NACA_0.75c _OnlyFromBlades
0.25
Olivary _0.50c_Total Olivary _0.50c _OnlyFromBlades
Olivary _0.50c_Total Olivary _0.50c _OnlyFromBlades
0.24 0.24
0.23 0.23
0.22 0.22
CP (-)
CP (-)
0.21 0.21
0.2 0.2
0.19 0.19
0.18 0.18
0.17 0.17
0.16 0.16
0c 0.0675c 0.125c 0.25c 0.5c -0.125c -0.0675c 0c 0.0675c 0.125c 0.1875c 0.375c 0.5c
d LE d MTP
Fig. 27. The power coefficients of VAWTs with different strut chordwise connecting positions. The solid histograms represent the
total CP and the hollow histograms represent CP only form blades. (a) Claffied by dLE. (b) Claffied by dMTP. The cases with red
dotted line and pentagrams are the worst and the best performance respectively.

The flow characteristics of each strut at different chordwise positions are displayed in Fig. 28. Observed
together with the aerodynamic performance in Fig. 27:
⚫ For the streamlined strut profiles with thickness variations, either NACA airfoil or olivary shape, when
dMTP=0c, the flow state on the blade surfaces is worse compared to the other strut positions, as shown in
Fig. 28 (f) and (m). The reason for this phenomenon may be due to the airflow passing over the blade
aerodynamic center (also near the blade MTP) is in a more easily separated state. When the blade AC is
connected to the strut MTP, the blade surface flow in the separation critical state is interfered by the strut to
the greatest extent. But this speculation needs more evidence to confirm.
⚫ For all struts with different profiles, the closer to the trailing edge, the less influence on the blade surface
flow. Taking the baseline flat strut as an example, Fig. 28 (d) shows that the interference of the strut on the
blade is slight, even similar to the case without the strut. Comparing with Fig. 18 (b) at θ=90○, it is observed
that the surface flow near the trailing edge of the blade without strut will be separated by the spanwise flow
due to the tip loss effect. Therefore, installing the strut at the trailing edge position will minimize its
influence on the blade. However, this connecting position will cause the aerodynamic forces to generate a
pitching moment and induce local stress concentrations at the connection, which is unacceptable for large-
scale VAWTs.
In summary, considering the influence of the strut connecting position on the aerodynamic and structural
25
performance, dMTP=0.125c is a relative better choice.
Fig. 28. Flow characteristics of VAWTs for different strut spanwise connecting positions. The red dashed lines are used to indicate
the blade AC and the strut MTP. The baseline flat strut is identified with the 0.25cS since it has no maximum thickness. (a)~(d) The
baseline flat struts. (e)~(h) The NACA airfoil struts with cS=0.50cB. (i)~(l) The NACA airfoil struts with cS=0.75cB. (m)~(p) The
olivary struts with cS=0.50cB.
3.4. Influences of Strut Connecting Angle

Inclining struts at an angle to lower the center of gravity of the tower is commonly used for large-scale VAWTs.
Although Hand et al.[24] suggested a horizontal installed strut to reduce its bending moment, the influence of the
inclined strut on the aerodynamic performance is still lacking and worth discussed. Therefore, based on the
different strut sizes and profiles, the effect of the inclined strut is investigated in this section.
The effect of the inclined strut on the VAWT power coefficients is displayed in Fig. 29. It shows that the drag
characteristics of the inclined strut are related to the strut cross-sectional profiles. For the NACA airfoil struts, the
inclination is beneficial to improve the efficiency. However, for the olivary profiles, the inclined strut slightly
reduces the aerodynamic performance. In addition, compared to the horizontal strut, the inclined NACA airfoil
strut reduces both the induced drag and the profile drag, but the inclined olivary strut has insignificant variations
in the profile drag. This difference may be due to the fact that the aerodynamic component force of the inclined
NACA airfoil strut are able to provide some positive torque[56], while the olivary strut fails to provide lift force due
to its special shape.
26
0.25 0.25
0.225 0.225
0.2 0.2
0.175 0.175
0.15 0.15
CP (-)
CP (-)
0.125 0.125
0.1 0.1
0.075 0.075
0.05 0.05
0.025 0.025
0 0
Horizontal Inclined Horizontal Inclined
Strut connecting angle (-) Strut connecting angle (-)
Fig. 29. The power coefficients of VAWTs with different strut connecting angles. The solid histograms represent the total CP. The
hollow histograms represent the CP only form blades. (a) The struts use NACA airfoil profiles. (a) The struts use olivary profiles.
The flow characteristics of the VAWTs with horizontal and inclined struts are illustrated in Fig. 30. For the
NACA airfoil struts, the inclined installation significantly reduces the vortex scale above the strut, thus improving
the flow separation in this region, as shown by the dashed red circles in Fig. 30 (e)~(h). However, the inclined
NACA airfoil strut compresses the space below the strut, leading to an increase in separation. These flow
characteristics correspond to the variation in aerodynamic performance in Fig. 29. For the olivary struts, the
inclined struts have little effect on the blade surface flow, except for the case cS=1.00cB. This indicates that the
olivary profile is not sensitive to the connecting angle in most cases.
Therefore, if only high aerodynamic efficiency is desired, the inclined VAWT struts may be a good choice,
especially for small-scale turbines. However, considering the issue of bending stresses exacerbated by the inclined
strut as stated by Hand et al.[24], horizontally installed olivary struts may be more suitable for large-scale VAWTs.
27
Fig. 30. Flow characteristics of VAWTs for horizontal and inclined struts. (a)~(d) The horizontal NACA airfoil struts. (e)~(h) The
inclined NACA airfoil struts. (i)~(l) The horizontal olivary struts. (m)~(p) The inclined olivary struts.
3.5. Influences of Strut-Blade Connecting Fairing

The effect of VAWT strut-blade connecting fairing on aerodynamic performance has not been discussed yet.
Consequently, the fairings for struts with different chords and profiles are established.
The aerodynamic performance of the VAWT with the NACA airfoil struts adding fairings is shown in Fig. 31
(a). It is observed that the effect of fairing is highly dependent on the strut chord. For cases of cS<1.00cB (red and
blue histograms), the fairing has a negative effect. The larger the fairing radius, the more significant the blade
performance decrease. However, for cases of cS≥1.00cB (green and yellow histograms), the addition of the fairing
improves the VAWT efficiency, and the improvement is more obvious for larger radius, especially for cS=1.00cB.
In addition, to verify whether the fairing effect is affected by the cross-sectional profile, the olivary struts
with 0.12c radius are calculated, as shown in Fig. 31 (b). It is observed that the olivary struts also follow the above
rule, i.e., the fairing has a positive effect if cS≥1.00cB. But the improvement of the olivary strut fairing is not as
pronounced as that of the NACA airfoil strut.
28
0.25 0.25
0.225 0.225
0.2 0.2
0.175 0.175
0.15 0.15
CP (-)
CP (-)
0.125 0.125
0.1 0.1
0.075 0.075
0.05 0.05
0.025 0.025
0 0
No Fairing 0.06c 0.12c No Fairng 0.12c
Strut-Blade Connecting Fairing Radius (-) Strut-Blade Connecting Fairing Radius (-)
Fig. 31. The power coefficients of VAWTs with different strut-blade connecting fairings. The solid histograms represent the total
CP. The hollow histograms represent the CP only form blades. (a) The struts use NACA airfoil profiles. (a) The struts use olivary
profiles. Only the 0.12c fairing radius are calculated for the olivary struts here.
The flow characteristics of NACA and olivary struts with 0.12c fairings for different chords are illustrated in
Fig. 32. Comparing with the VAWT struts without fairings in Fig. 24, it shows that the fairing exacerbates the
separation scale at the leading edge of the connection when cS<1.00cB, as illustrated by the red dashed lines in
(a)~(b) and (e)~(f), which in turn affects the blade surface flow and leads to a reduction in blade torque. For cases
of cS≥1.00cB, the fairing significantly improves the interference of the strut on the blade surface. In particular,
when cS=1.00cB, the large-scale separation in the blade leading edge region above the NACA airfoil strut is
mitigated and replaced by relatively better attached flow.
Fig. 32. Flow characteristics of VAWTs with strut-blade connecting fairings radius 0.12c. (a)~(d) The NACA airfoil struts of
different chords. (e)~(h) The olivary struts of different chords.
3.6. Summary of Strut Parameters

In summary of the above analysis, Table 5 gives the recommended values for the straight-blade VAWT strut
design parameters. These parameters are beneficial for minimizing the influence of the strut on the aerodynamic
performance, as well as for the structural strength design requirements of the VAWT struts.
Table 5 Recommended parameter values of straight-bladed VAWT strut.
Parameter Recommended Value
Cross section Olivary
Strut chord length 1.5c
Spanwise position 0.2H (From blade tip)
Chordwise posintion dMTP=0.125c (If cS<cB)
Connecting angle ξ ξ=0°(For large tuebine)
Fairing radius 0.12c (When cS≥cB)
29
4. Conclusions
In this paper, the struts of the straight-bladed VAWT are investigated to explore the suitable design parameters
that minimize the influence on the VAWT aerodynamic performance. Thirty types of strut profiles including a
novel olivary profile with different chord lengths, thicknesses and areas are parametrical modeled based on the
equal cross-sectional area principle for the basic supporting strength requirements. VAWTs with these struts are
simulated using the accuracy validated CFD model. Then the strut connecting parameters, including the spanwise
and chordwise positions, connecting angle, and connecting fairing are analyzed. Some primary conclusions are as
following:
(1) The profile and induce drags of the strut substantially reduce the VAWT efficiency. For the strut profile
drag related to the characteristic of the cross-sectional shape itself, it is more significant at the azimuth
angles where the relative wind speed is stronger (270○<θ<390○). For the induced drag, which is caused
by the strut interfering with the blade surface, the interference spreads over the entire blade surface in the
spanwise direction, even for the thinnest baseline flat strut. In addition, the induced drag is concentrated
on the suction side of the VAWT blade and is most significant in the upwind region (60○<θ<180○). In the
downwind region, the induced drag is primarily caused by the wake generated by the struts.
(2) The analysis of VAWT performance for different strut chord lengths, thicknesses and areas shows a general
linear relationship between the struts drag characteristics and their absolute thicknesses (a dimensionless
parameter based on the blade height). The higher the absolute thickness, the higher the strut drag, and the
lower the VAWT performance. On the other hand, for the struts with same cross-sectional area, the larger
the chord length (corresponding to a smaller relative thickness) the lower the drag. However, attention
should be paid to avoid equal chord lengths for struts and blades, as this will generate the largest vortex
scale at the strut-blade connection, resulting in the most severe degradation of blade performance. The
strut chord longer than the blade chord is a good choice, which can minimize the vortex scale generated
at the connection and reduce the spanwise flow by acting as an endplate.
(3) The olivary profile proposed specifically for VAWT strut has less profile drag and induced drag than the
NACA airfoil strut of the same size. Considering the variation of CP-Total of VAWT, the drag of olivary
strut is reduced by 25.8% compared to the NACA strut. Besides, this olivary strut can keep similar drag
characteristics with a cross-sectional area 7 times larger than that of the baseline flat strut, which is
beneficial for designing a strut with higher structural strength. The flow characteristics analysis shows
that the smaller leading edge radius of the olivary profile and the more backward position of the maximum
thickness contributes to reduce the interference of the strut with the blade surface flow.
(4) Variations in both the spanwise and chordwise connecting positions affect the VAWT aerodynamic
performance. For the spanwise position, the NACA airfoil strut will slightly improve the VAWT efficiency
due to the overlap of interfered vortex and blade tip loss effects when the strut is close to the blade tip
(0.2H). However, the olivary strut is not sensitive to the variation of spanwise position. For the chordwise
positions (when cS<1.00cB), the maximum thickness position of the strut should be avoided to coincide
with the blade aerodynamic center. This is because such coincidence leads to the most significant
degradation of aerodynamic performance. This holds true for both NACA and olivary struts.
(5) The inclined strut slightly improves the aerodynamic performance of the VAWT with NACA airfoil struts
but has a negligible effect on the VAWT with olivary struts. This is because the inclined NACA airfoil
strut reduces the vortex scale in the region above the strut and generate some positive torque by itself at
the same time. Based on the analysis, it is concluded that an inclined strut is a good choice for small
VAWTs with low structural demands. But for large-scale VAWTs, the horizontal installed olivary strut is
more recommended to reduce the bending stress inside the strut.
(6) The strut-blade connecting fairing is more suitable for the case of cS≥1.00cB, which is valid for both NACA
30
and olivary struts. Flow analysis shows that the fairing expands the vortex scale at the connection leading
edge, resulting in performance degradation when cS<1.00cB. The opposite is true for the case of cS≥1.00cB.
5. Future Work
This paper focuses on the influence of strut design parameters on the aerodynamic performance of the
straight-bladed VAWT. Some further works may be studied in the future:
(1) A small VAWT with low aspect ratio is use in this paper because it is convenient to verify with the
experimental wind turbine. Although the dimensionless parametric design of the strut scale has been
carried out, the effects of some strut parameters related to the spanwise direction, such as the thickness
and the spanwise connecting position, still require validation on a high aspect ratio VAWT.
(2) In this paper, struts with constant cross section along the radius direction of the VAWT are employed.
However, struts with variable chord length and profile are commonly used in practical applications,
especially for large-scale VAWTs. The drag characteristics of variable cross-section struts with different
parameters needs to be further confirmed.
(3) Experiments should be conducted in the future to confirm the characteristics of the olivary profile and the
strut connecting parameters discussed in this paper.
Acknowledgments
The authors would like to acknowledge the support of National Natural Science Foundation of China (grand
No. 52106262, No. 52006148, and No. 51976131). Besides, special thanks to researcher Hand et al[24]. for the
work of VAWT that inspired this study.
Reference
[1] Simon-Philippe Breton, Geir Moe. Status, plans and technologies for offshore wind turbines in Europe and North
America [J]. Renewable Energy, 2009, 34(3): 646-654.
[2] P. Veers, K. Dykes, E. Lantz, S. Barth, C. L. Bottasso, O. Carlson, A. Clifton, J. Green, P. Green, H. Holttinen, D.
Laird, V. Lehtomaki, J. K. Lundquist, J. Manwell, M. Marquis, C. Meneveau, P. Moriarty, X. Munduate, M.
Muskulus, J. Naughton, L. Pao, J. Paquette, J. Peinke, A. Robertson, J. Sanz Rodrigo, A. M. Sempreviva, J. C.
Smith, A. Tuohy, R. Wiser. Grand challenges in the science of wind energy [J]. Science, 2019, 366(6464): 1-9.
[3] T Composites, Warren, Cost study for large wind turbine blades: WindPACT blade system design studies.[R].
Sandia National Laboratories, 2003.
[4] H. Arabian-Hoseynabadi, H. Oraee, P. J. Tavner. Failure Modes and Effects Analysis (FMEA) for wind turbines [J].
International Journal of Electrical Power & Energy Systems, 2010, 32(7): 817-824.
[5] G. Sieros, P. Chaviaropoulos, J. D. Sørensen, B. H. Bulder, P. Jamieson. Upscaling wind turbines: theoretical and
practical aspects and their impact on the cost of energy [J]. Wind Energy, 2012, 15(1): 3-17.
[6] Willy Tjiu, Tjukup Marnoto, Sohif Mat, Mohd Hafidz Ruslan, Kamaruzzaman Sopian. Darrieus vertical axis wind
turbine for power generation I: Assessment of Darrieus VAWT configurations [J]. Renewable Energy, 2015, 75:
50-67.
[7] Willy Tjiu, Tjukup Marnoto, Sohif Mat, Mohd Hafidz Ruslan, Kamaruzzaman Sopian. Darrieus vertical axis wind
turbine for power generation II: Challenges in HAWT and the opportunity of multi-megawatt Darrieus VAWT
development [J]. Renewable Energy, 2015, 75: 560-571.
[8] Michael Borg, Maurizio Collu, Athanasios Kolios. Offshore floating vertical axis wind turbines, dynamics
modelling state of the art. Part II: Mooring line and structural dynamics [J]. Renewable and Sustainable Energy
Reviews, 2014, 39: 1226-1234.
[9] Michael Borg, Maurizio Collu. Offshore floating vertical axis wind turbines, dynamics modelling state of the art.
Part III: Hydrodynamics and coupled modelling approaches [J]. Renewable and Sustainable Energy Reviews, 2015,
46: 296-310.
[10] Michael Borg, Andrew Shires, Maurizio Collu. Offshore floating vertical axis wind turbines, dynamics modelling
state of the art. part I: Aerodynamics [J]. Renewable and Sustainable Energy Reviews, 2014, 39: 1214-1225.
[11] Sutherland H J Ashwill T D, Berg D E, A retrospective of VAWT technology[R]. Sandia National Laboratories,
2012.
[12] Weipao Miao, Qingsong Liu, Zifei Xu, Minnan Yue, Chun Li, Wanfu Zhang. A comprehensive analysis of blade
31
tip for vertical axis wind turbine: Aerodynamics and the tip loss effect [J]. Energy Conversion and Management,
2022, 253: 115140.
[13] Tian-tian Zhang, Mohamed Elsakka, Wei Huang, Zhen-guo Wang, Derek B. Ingham, Lin Ma, Mohamed
Pourkashanian. Winglet design for vertical axis wind turbines based on a design of experiment and CFD approach
[J]. Energy Conversion and Management, 2019, 195: 712-726.
[14] M. Elkhoury, T. Kiwata, E. Aoun. Experimental and numerical investigation of a three-dimensional vertical-axis
wind turbine with variable-pitch [J]. Journal of Wind Engineering and Industrial Aerodynamics, 2015, 139: 111-
123.
[15] Willmer A. Wind tunnel tests on a 3m diameter musgrove windmill [J]. Int J Ambient Energy, 1980, 1(1): 21-27.
[16] Longhuan Du, Grant Ingram, Robert G. Dominy. Experimental study of the effects of turbine solidity, blade profile,
pitch angle, surface roughness, and aspect ratio on the H‐Darrieus wind turbine self‐starting and overall
performance [J]. Energy Science & Engineering, 2019, 7(6): 2421-2436.
[17] Stefania Zanforlin, Stefano Deluca. Effects of the Reynolds number and the tip losses on the optimal aspect ratio
of straight-bladed Vertical Axis Wind Turbines [J]. Energy, 2018, 148: 179-195.
[18] Claessens MC. The design and testing of airfoils for application in small vertical axis wind turbines: aerospace
engineering[D]. Delft University of Technology, 2006
[19] Mojtaba Tahani, Narek Babayan, Seyedmajid Mehrnia, Mehran Shadmehri. A novel heuristic method for
optimization of straight blade vertical axis wind turbine [J]. Energy Conversion and Management, 2016, 127: 461-
476.
[20] Islam M, Ting D, Fartaj A. Desirable airfoil features for smaller-capacity straight-bladed VAWT [J]. Wind Energy,
2007, 31(3): 165-196.
[21] S. Brusca, R. Lanzafame, M. Messina. Design of a vertical-axis wind turbine: how the aspect ratio affects the
turbine’s performance [J]. International Journal of Energy and Environmental Engineering, 2014, 5(4): 333-340.
[22] Abdolrahim Rezaeiha, Ivo Kalkman, Hamid Montazeri, Bert Blocken. Effect of the shaft on the aerodynamic
performance of urban vertical axis wind turbines [J]. Energy Conversion and Management, 2017, 149: 616-630.
[23] Aya Aihara, Victor Mendoza, Anders Goude, Hans Bernhoff. A numerical study of strut and tower influence on the
performance of vertical axis wind turbines using computational fluid dynamics simulation [J]. Wind Energy, 2022, :
1-17.
[24] Brian Hand, Ger Kelly, Andrew Cashman. Aerodynamic design and performance parameters of a lift-type vertical
axis wind turbine: A comprehensive review [J]. Renewable and Sustainable Energy Reviews, 2021, 139: 110699.
[25] Abdolrahim Rezaeiha, Hamid Montazeri, Bert Blocken. Active flow control for power enhancement of vertical
axis wind turbines: Leading-edge slot suction [J]. Energy, 2019, 189: 116131.
[26] Jinjing Sun, Xiaojing Sun, Diangui Huang. Aerodynamics of vertical-axis wind turbine with boundary layer suction
– Effects of suction momentum [J]. Energy, 2020, 209: 118446.
[27] Haitian Zhu, Wenxing Hao, Chun Li, Qinwei Ding, Baihui Wu. Application of flow control strategy of blowing,
synthetic and plasma jet actuators in vertical axis wind turbines [J]. Aerospace Science and Technology, 2019, 88:
468-480.
[28] Qingsong Liu, Weipao Miao, Qi Ye, Chun Li. Performance assessment of an innovative Gurney flap for straight-
bladed vertical axis wind turbine [J]. Renewable Energy, 2022, 185: 1124-1138.
[29] You-Lin Xu, Yi-Xin Peng, Sheng Zhan. Optimal blade pitch function and control device for high-solidity straight-
bladed vertical axis wind turbines [J]. Applied Energy, 2019, 242: 1613-1625.
[30] Wenxing Hao, Musa Bashir, Chun Li, Chengda Sun. Flow control for high-solidity vertical axis wind turbine based
on adaptive flap [J]. Energy Conversion and Management, 2021, 249: 114845.
[31] Alessandro Bianchini, Francesco Balduzzi, Daniele Di Rosa, Giovanni Ferrara. On the use of Gurney Flaps for the
aerodynamic performance augmentation of Darrieus wind turbines [J]. Energy Conversion and Management, 2019,
184: 402-415.
[32] Kok Hoe Wong, Wen Tong Chong, Nazatul Liana Sukiman, Yui-Chuin Shiah, Sin Chew Poh, Kamaruzzaman
Sopian, Wei-Cheng Wang. Experimental and simulation investigation into the effects of a flat plate deflector on
vertical axis wind turbine [J]. Energy Conversion and Management, 2018, 160: 1019-1125.
[33] Zhenyu Wang, Yuchen Wang, Mei Zhuang. Improvement of the aerodynamic performance of vertical axis wind
turbines with leading-edge serrations and helical blades using CFD and Taguchi method [J]. Energy Conversion
and Management, 2018, 177: 107-121.
[34] Omar S. Mohamed, Ahmed A. Ibrahim, Ahmed K. Etman, Amr A. Abdelfatah, Ahmed M. R. Elbaz. Numerical
investigation of Darrieus wind turbine with slotted airfoil blades [J]. Energy Conversion and Management: X, 2020,
5: 100026.
[35] F. Arpino, M. Scungio, G. Cortellessa. Numerical performance assessment of an innovative Darrieus-style vertical
axis wind turbine with auxiliary straight blades [J]. Energy Conversion and Management, 2018, 171: 769-777.
[36] M. Scungio, F. Arpino, V. Focanti, M. Profili, M. Rotondi. Wind tunnel testing of scaled models of a newly
32
developed Darrieus-style vertical axis wind turbine with auxiliary straight blades [J]. Energy Conversion and
Management, 2016, 130: 60-70.
[37] Qingsong Liu, Weipao Miao, Chun Li, Winxing Hao, Haitian Zhu, Yunhe Deng. Effects of trailing-edge movable
flap on aerodynamic performance and noise characteristics of VAWT [J]. Energy, 2019, 189: 116271.
[38] Lulu Ni, Weipao Miao, Chun Li, Qingsong Liu. Impacts of Gurney flap and solidity on the aerodynamic
performance of vertical axis wind turbines in array configurations [J]. Energy, 2021, 215: 118915.
[39] L. Battisti, A. Brighenti, E. Benini, M. Raciti Castelli. Analysis of Different Blade Architectures on small VAWT
Performance[C]//Journal of Physics: Conference Series. 2016,753:062009.
[40] Mojtaba Ahmadi-Baloutaki, Rupp Carriveau, David S. K. Ting. Straight-bladed vertical axis wind turbine rotor
design guide based on aerodynamic performance and loading analysis [J]. Proceedings of the Institution of
Mechanical Engineers, Part A: Journal of Power and Energy, 2014, 228(7): 742-759.
[41] M. Islam, M.R. Amin, D.S-K Ting, A Fartaj. Performance Analysis of a Smaller-capacity Straight-bladed VAWT
with Prospective Airfoils[C]//46th AIAA Aerospace Sciences Meeting and Exhibit, Reno, Nevada, USA. 2008.
[42] Brian Hand, Andrew Cashman. Aerodynamic modeling methods for a large-scale vertical axis wind turbine: A
comparative study [J]. Renewable Energy, 2018, 129: 12-31.
[43] Mark H. Worstell. Aerodynamic Performance of the DOE/Sandia 17-m-Diameter Vertical-Axis Wind Turbine [J].
Journal of Energy, 1981, 5(1): 39-42.
[44] Qing’an Li, Takao Maeda, Yasunari Kamada, Junsuke Murata, Kento Shimizu, Tatsuhiko Ogasawara, Alisa Nakai,
Takuji Kasuya. Effect of solidity on aerodynamic forces around straight-bladed vertical axis wind turbine by wind
tunnel experiments (depending on number of blades) [J]. Renewable Energy, 2016, 96: 928-939.
[45] Aya Aihara, Victor Mendoza, Anders Goude, Hans Bernhoff. Comparison of Three-Dimensional Numerical
Methods for Modeling of Strut Effect on the Performance of a Vertical Axis Wind Turbine [J]. Energies, 2022,
15(7): 2361.
[46] Peter Bachant, Martin Wosnik, Budi Gunawan, Vincent S. Neary. Experimental Study of a Reference Model
Vertical-Axis Cross-FlowTurbine [J]. PLoS One, 2016, 11(9): 1-20.
[47] Mazharul Islam, M. Ruhul Amin, David S-K. Ting, Amir Fartaj. A NewAirfoil for the Supporting Struts of Smaller-
capacity Straight-Bladed VAWT[C]//12th AIAA/ISSMO Multidisciplinary Analysis and Optimization Conference,
Victoria, British Columbia Canada. 2008.
[48] Philip Marsh, Dev Ranmuthugala, Irene Penesis, Giles Thomas. Three-dimensional numerical simulations of
straight-bladed vertical axis tidal turbines investigating power output, torque ripple and mounting forces [J].
Renewable Energy, 2015, 83: 67-77.
[49] Yutaka Hara, Naoki Horita, Shigeo Yoshida, Hiromichi Akimoto, Takahiro Sumi. Numerical Analysis of Effects of
Arms with Different Cross-Sections on Straight-Bladed Vertical Axis Wind Turbine [J]. Energies, 2019, 12(11):
2106.
[50] Mazharul Islam, Amir Fartaj, Rupp Carriveau. Analysis of the Design Parameters Related to a Fixed-Pitch Straight-
Bladed Vertical Axis Wind Turbine [J]. Wind Engineering, 2008, 32(5): 491-507.
[51] Thierry Villeneuve, Grégoire Winckelmans, Guy Dumas. Increasing the efficiency of vertical-axis turbines through
improved blade support structures [J]. Renewable Energy, 2021, 169: 1386-1401.
[52] M. Saqib Hameed, S. Kamran Afaq. Design and analysis of a straight bladed vertical axis wind turbine blade using
analytical and numerical techniques [J]. Ocean Engineering, 2013, 57: 248-255.
[53] John Keithley Difuntorum, Louis Angelo M. Danao. Improving VAWT performance through parametric studies of
rotor design configurations using computational fluid dynamics [J]. Proceedings of the Institution of Mechanical
Engineers, Part A: Journal of Power and Energy, 2018, 233(4): 489-509.
[54] Alessandro Bianchini, Francesco Balduzzi, Giovanni Ferrara, Lorenzo Ferrari. Influence of the blade-spoke
connection point on the aerodynamic perfomance of Darrieus wind turbines[C]//In: Proceedings of the ASME turbo
expo 2016, Seoul, South Korea. 2016:1-9.
[55] Pierre Blusseau, Minoo H. Patel. Gyroscopic effects on a large vertical axis wind turbine mounted on a floating
structure [J]. Renewable Energy, 2012, 46: 31-42.
[56] Agostino De Marco, Domenico P. Coiro, Domenico Cucco, Fabrizio Nicolosi. A Numerical Study on a Vertical-
Axis Wind Turbine with Inclined Arms [J]. International Journal of Aerospace Engineering, 2014, 2014: 1-14.
[57] Brian K. Kirke. Evaluation of Self-Starting Vertical Axis Wind Turbines for Stand-Alone Applications[D]. Griffith
University, 1998
[58] H. F. Lam, H. Y. Peng. Study of wake characteristics of a vertical axis wind turbine by two- and three-dimensional
computational fluid dynamics simulations [J]. Renewable Energy, 2016, 90: 386-398.
[59] Sighard F. Hoerner, Practical information on aerodynamic drag and hydrodynamic resistance[R]. Fluid
dynamic drag. Horner fluid dynamics, CA, 1965.
[60] Denis Pitance, Simon Horb, oanna Kluczewska-Bordier, Alexandre Immas, Frédéric Silver, Experimental
validation of pharwen code using data from vertical-axis wind turbines[R]. WindEurope PO, 2016.
33
[61] Qing'an Li, Takao Maeda, Yasunari Kamada, Junsuke Murata, Toshiaki Kawabata, Kento Shimizu, Tatsuhiko
Ogasawara, Alisa Nakai, Takuji Kasuya. Wind tunnel and numerical study of a straight-bladed vertical axis wind
turbine in three-dimensional analysis (Part I: For predicting aerodynamic loads and performance) [J]. Energy, 2016,
106: 443-452.
[62] Qing'an Li, Takao Maeda, Yasunari Kamada, Junsuke Murata, Toshiaki Kawabata, Kento Shimizu, Tatsuhiko
Ogasawara, Alisa Nakai, Takuji Kasuya. Wind tunnel and numerical study of a straight-bladed Vertical Axis Wind
Turbine in three-dimensional analysis (Part II: For predicting flow field and performance) [J]. Energy, 2016, 104:
295-307.
[63] Manwell JF, McGowan JG, Rogers AL. Wind energy explained: Theory, design and application[M]. John Wiley
& Sons, 2002
[64] C. L. Ladson, Cuyler W.Jr. Brooks, Development of a computer program to obtain ordinates for NACA 4-digit, 4-
digit modified, 5-digit, and 16 series airfoils[R]. No. L-10375, 1975.
[65] Marco Raciti Castelli, Guido Ardizzon, Lorenzo Battisti, Ernesto Benini, Giorgio Pavesi. Modeling Strategy and
Numerical Validation For A Darrieus Vertical Axis Micro-Wind Turbine[C]//Proceedings of the ASME 2010
International Mechanical Engineering Congress & Exposition, Canada，British Columbia. 2010.
[66] L. Battisti, G. Persico, V. Dossena, B. Paradiso, M. Raciti Castelli, A. Brighenti, E. Benini. Experimental
benchmark data for H-shaped and troposkien VAWT architectures [J]. Renewable Energy, 2018, 125: 425-444.
[67] Ning Ma, Hang Lei, Zhaolong Han, Dai Zhou, Yan Bao, Kai Zhang, Lei Zhou, Caiyong Chen. Airfoil optimization
to improve power performance of a high-solidity vertical axis wind turbine at a moderate tip speed ratio [J]. Energy,
2018, 150: 236-252.
[68] P. J. Roache. Quantification of Uncertainty in Computational Fluid Dynamics [J]. Annu Rev Fluid Mech, 1997, 29:
123–160.
[69] S. Zadeh, M. Komeili, M. Paraschivoiu. Mesh convergence study for 2-d straight-blade vertical axis wind turbine
simulations and estimation for 3-d simulations [J]. Transactions of the Canadian Society for Mechanical
Engineering, 2014, 38(4): 487-503.
[70] Abdolrahim Rezaeiha, Hamid Montazeri, Bert Blocken. Towards accurate CFD simulations of vertical axis wind
turbines at different tip speed ratios and solidities: Guidelines for azimuthal increment, domain size and
convergence [J]. Energy Conversion and Management, 2018, 156: 301-316.
[71] László Daróczy, Gábor Janiga, Klaus Petrasch, Michael Webner, Dominique Thévenin. Comparative analysis of
turbulence models for the aerodynamic simulation of H-Darrieus rotors [J]. Energy, 2015, 90: 680-690.
[72] Abdolrahim Rezaeiha, Hamid Montazeri, Bert Blocken. On the accuracy of turbulence models for CFD simulations
of vertical axis wind turbines [J]. Energy, 2019, 180: 838-857.
[73] Abdolrahim Rezaeiha, Hamid Montazeri, Bert Blocken. CFD analysis of dynamic stall on vertical axis wind
turbines using Scale-Adaptive Simulation (SAS): Comparison against URANS and hybrid RANS/LES [J]. Energy
Conversion and Management, 2019, 196: 1282-1298.
[74] Mikhail L. Shur, Philippe R. Spalart, Mikhail Kh Strelets, Andrey K. Travin. A hybrid RANS-LES approach with
delayed-DES and wall-modelled LES capabilities [J]. International Journal of Heat and Fluid Flow, 2008, 29(6):
1638-1649.
[75] Hang Lei, Dai Zhou, Yan Bao, Ye Li, Zhaolong Han. Three-dimensional Improved Delayed Detached Eddy
Simulation of a two-bladed vertical axis wind turbine [J]. Energy Conversion and Management, 2017, 133: 235-
248.
[76] Jacobs EN, Sherman A, Airfoil section characteristics as affected by variations of the Reynolds number. Tech.
rep.[R]. Washington. D.C National Advisory Committee for Aeronautics, 1939.
34
View publication stats

R1 V2 NoChangeMarked

Uploaded by

Copyright:

Available Formats

R1 V2 NoChangeMarked

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

R1 V2 NoChangeMarked

Uploaded by

Copyright:

Available Formats

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

Recommendation for strut designs of vertical axis wind turbines: Effects of

Article in Energy Conversion and Management · January 2023

Weipao Miao Qingsong Liu

SEE PROFILE SEE PROFILE

SEE PROFILE SEE PROFILE

Wind Farm Analysis and Optimization View project

fluid-induced vibration due to seals View project

The user has requested enhancement of the downloaded file.

* Corresponding author. Tel, +021-55271729

Tip Design Pitch Angle

Fig. 1. Design parameters affecting the aerodynamic performance of the VAWT[24].

1.1. Strut Effect of VAWTs

1.2. Influences of Strut Parameters

(a) (b) (c)

Fig. 5. The horseshoe vortex at the blade-strut connecting joint[24].

1.3. Present Work

been published regarding their effects on the aerodynamic performance.

the matching with different strut chord lengths.

2. Physical and Numerical Models

in CAD and CFD model

Split into 20 segments

2.2. Aerodynamic Characteristics of Straight-Bladed VAWT

2.3. Strut Models with Different Configurations

(ii) Ordinate at trailing edge：

(iii) Trailing-edge angle：

(iv) Nose shape：

Strut Max-Thickness Position 0.50c

d MTP =-0.125cB d MTP =0cB

cS tru t= 0.75cBla d e Profile

dLE =0cB d LE =0.0675cB

r=0.12cB r=0.12cB r=0.12cB r=0.12cB

Front View Front View Front View Front View

r=0.06cB r=0.06cB r=0.06cB

2.4. CFD Model

2.5. Validation of Simulations

2.5.2. Time Step and Revolution Verification

-30 Exp -30 Exp -30 Exp -30 Exp

Present CFD Present CFD

-20 -20 -20 -20

3. Results and Discussion

CQ-Interal for Spanwise (-)

(c) Difference of CQ at Various Hegihts △ CQ-Induced

3.2. Influences of Strut Profiles

cStrut/cBlade (-) cStrut/cBlade (-)

-47.7% -54.1% -32.2% -38.4%

CQ-Only From Blade

CQ-Only From Blade

CQ-Only From Blade

0.15 0.15 0.15 0.15

λ=2.29 λ=2.29 λ=2.29 λ=2.29

0.05 0.05 0.05 0.05

3.3. Influences of Strut Connecting Positions

(ii) Fluid Field Characteristics

3.4. Influences of Strut Connecting Angle

3.5. Influences of Strut-Blade Connecting Fairing

3.6. Summary of Strut Parameters

View publication stats

You might also like