Journal of the Brazilian Society of Mechanical Sciences and Engineering (2020) 42:86

https://doi.org/10.1007/s40430-020-2170-4

TECHNICAL PAPER

Effects of the hinge position and suction on flow separation

and aerodynamic performance of the NACA 0012 airfoil
Esmaeel Fatahian1 · Ali Lohrasbi Nichkoohi2 · Hesamoddin Salarian1 · Jahanfar Khaleghinia1

Received: 6 February 2019 / Accepted: 4 January 2020 / Published online: 13 January 2020
© The Brazilian Society of Mechanical Sciences and Engineering 2020

Abstract
In the present study, the effect of hinge position (H) has been numerically investigated to find the appropriate position for
improving the aerodynamic performance of the NACA 0012 flapped airfoil. In addition, perpendicular and tangential suctions
have been applied to control the flow separation and enhance the aerodynamic performance over the NACA 0012 flapped
airfoil at each different hinge positions. The simulations were carried out at a Reynolds number of 5 × 105 (Ma = 0.021) based
on two-dimensional incompressible unsteady Reynolds-averaged Navier–Stokes calculations to determine the adequate hinge
position. The turbulence was modeled using the shear stress transport k–ω turbulence model. The effect of perpendicular
suction (θjet = − 90°) and tangential suction (θjet = − 30°) was computationally studied over NACA 0012 flapped airfoil for
five different hinge positions (H = 0.7c, 0.75c, 0.8c, 0.85c and 0.9c) and a flap deflection (δf) of 15°. Based on the results,
the hinge position significantly affects the aerodynamic performance of the airfoil. The lift coefficient increased clearly as
the hinge position moved to the trailing edge of the airfoil. Using perpendicular suction caused to increase the lift coefficient
and decrease the drag coefficient. Consequently, the maximum value of the lift-to-drag ratio (CL/CD) for perpendicular and
tangential suctions was achieved about 35.8% and 25.1% higher than that of the case without suction at an angle of attack
of 12° and H = 0.9c. Also, the effect of perpendicular suction was more considerable compared to the tangential suction.
This caused a reduction in the size of the recirculation zone from 0.5 to 0.09 of the airfoil chord length and also transferred
it from 1.13 to 1.18 of the airfoil chord length.

Keywords Flow separation · Suction · Lift coefficient · Drag coefficient · NACA 0012 · Hinge position

List of symbols μ Viscosity

α Angle of attack ρ Density
β The angle between the free-stream velocity direc- τ Shear stress
tion and the local jet surface δf Flap deflection
c Airfoil chord length P Pressure
CP Pressure coefficient H Hinge position
CL Lift coefficient 𝜌∞ Free-stream density
CD Drag coefficient h Hinge moment
Cf Skin friction coefficient Sf Reference area of the control surface
ω (ε/k) Specific dissipation rate Cs Control surface reference chord
k Turbulent kinetic energy q∞ Free-stream dynamic pressure
Ch Hinge moment coefficient
Re Reynolds number
Technical Editor: André Cavalieri.
Ma Mach number
* Ali Lohrasbi Nichkoohi U∞ Free-stream velocity
lohrasbi@iauns.ac.ir y+ Dimensionless wall distance
θjet Jet angle
1
Department of Mechanical Engineering, Nour Branch, Ljet Jet location
Islamic Azad University, Nour, Iran
Ujet Jet velocity
2
Department of Mechanical Engineering, Nowshahr Branch, Rjet Jet velocity ratio
Islamic Azad University, Nowshahr, Iran

13
Vol.:(0123456789)
86 Page 2 of 14 Journal of the Brazilian Society of Mechanical Sciences and Engineering (2020) 42:86

1 Introduction conditions. Their study showed that the best results can
be achieved with suction only. Lei et al. [26] investigated
Flaps are high-lift devices consisting of a hinged panel a suction control of the laminar separation bubble over
or panels mounted on the wing trailing edge. They help the NACA 2415 airfoil at low Reynolds number. Their
to increase the value of both lift and drag and a reduction results demonstrated that suction control can decrease the
in the stall speed. These factors cause an improvement in velocity gradient of the boundary layer, delay the transi-
takeoff and landing performance. Flaps help to increase tion and prevent the generation of the separation bubble.
the value of the lift coefficient (CL) and the stall angle Zhou et al. [27] numerically analyzed the effects of Mach
during the takeoff phase of an aircraft. The aircraft wings number on the flow separation control of an airfoil with a
are mainly intended to provide the maximum value of the small plate near the leading edge. They created a mutual
lift-to-drag ratio (CL/CD) [1]. This ratio is considered to be interference between the trailing vortex generated by the
a measure of the efficiency of an aircraft [2] and is related small plate and the airfoil boundary layer to control the
to the required amount of power for thrust generation. The flow separation. They concluded that better aerodynamic
performance of the aircraft wing is seriously damaged if performance of the airfoil achieved in this way. Ma et al.
flow separation occurs [3]. [28] numerically and experimentally studied the effects of
Nowadays, modern airplanes use complex multi-ele- suction flow control on the aerodynamic performance of
ment high-lift devices consisting of the slat and single or the wing at low Reynolds numbers. In their study, when
multiple flaps in order to generate the high lift required the suction holes are close to the separation point, the suc-
during takeoff and landing for reducing runway lengths tion has the best control effect. While the controllability in
and ground speeds [4]. Multi-element wing designs, how- an active flow control method is the main advantage for the
ever, are found unfavorable from a weight and complexity system, the implementation and maintenance of mechani-
point of view, and they increase the wing efficiency. This cal mechanisms are still the most important barriers for
is the reason for replacing the multi-element flap with a its industrial applications. Passive flow control methods
single-hinged flap in the current designs for decreasing the are much cheaper and easier to be implemented but they
complexity [1, 5]. Instead of incorporating complex, heavy are usually less effective than active flow control. Pas-
and expensive multi-element high-lift devices, single flaps sive flow control systems have to be implemented in their
without slats are desirable [6]. optimum positions because their final configuration cannot
In the domain of aerospace and aviation, the exist- be changed during their application. Therefore, finding an
ence of flow separation leads to the instability of flying optimum position is a critical point for their performance.
vehicles, the generation of noise, increasing the drag and From the literature review, any numerical study has not
decreasing the lift. The flow separation control technol- been yet conducted on using an active flow control method
ogy has been gradually developed as an important branch (suction) by considering and finding the appropriate hinge
of mechanics [6–8]. Traditionally, passive [7–9] or active position to enhance the aerodynamic performance and con-
[10–13] devices are used for flow separation control. Pas- trol the flow separation over the flapped airfoil. Moreover,
sive control devices are those which are not energy con- for the first time, a special attempt is done to utilize perpen-
sumptive such as trailing-edge flaps [14–16] and leading- dicular and tangential suctions in order to control the bound-
edge slats [17, 18]. In contrast, active control devices are ary layer separation and improve the aerodynamic perfor-
energy consumptives such as a small energy input by suc- mance at each different hinge positions. The outcome of the
tion and blowing jets [19–23]. present study could be important for the aviation community
Many experimental and numerical studies have been for designing flapped airfoils with better controllability in
conducted to control the flow separation and both active takeoff and landing conditions.
and passive methods have been suggested. Obeid et al. [1]
investigated a 2D subsonic flow over the NACA 0015 air-
foil with a 30% trailing-edge flap. Their results revealed 2 Computational methodology
that increasing the flap deflections enhanced the maxi-
mum lift coefficient and decrease the stall angle. Ahmed 2.1 Numerical method and governing equations
et al. [24] numerically studied the behavior of the NACA
0012 airfoil at different flap angles and Mach numbers. In the present study, the pressure-based finite volume solver
Their results indicated that the increment in the flap angle of commercial software ANSYS Fluent 18.2 [29] was used
caused to increase the thickness of the boundary layer. for numerical simulation and modeling the turbulent proper-
Genc et al. [25] numerically considered blowing and ties of the flow by utilizing the unsteady Reynolds-averaged
suction methods on the NACA 2415 airfoil in transient Navier–Stokes (URANS) equations. The calculations were
carried out over the NACA 0012 flapped airfoil with 1 m chord

13
Journal of the Brazilian Society of Mechanical Sciences and Engineering (2020) 42:86 Page 3 of 14 86

length and a chord Reynolds number of 5 × 105 (Ma = 0.021) where μt and k represent the turbulent viscosity and turbulent
using the CFD flow solver. kinetic energy. 𝛿ij is the Kronecker delta.
A low free-stream turbulence level was applied to match the Direct numerical simulation (DNS) and large eddy simu-
wind tunnel characteristics, so the stream turbulence intensity lation (LES) are useful tools for fundamental research on
was selected as less than 0.1% [19, 30]. Also, the SIMPLE turbulent boundary flows [35–38]; however, they are pro-
algorithm [19, 22, 30–34] for pressure–velocity coupling and hibitively costly for realistic engineering problems. Com-
upwind second-order method was employed to discretize the pared with DNS and LES, RANS is rather cheap in the cost
pressure, momentum and turbulence transport equations. A of computations and is widely used for airfoils [7, 22, 25].
sufficiently time step of 1 × 10−4 was used to achieve Cou- Therefore, the URANS approach using the SST k–ω model
rant–Friedrichs–Lewy (CFL) number lower than 1, and the is chosen to achieve a balance between accuracy and com-
results were converged when the scaled residual was less than putation cost. The turbulence model adopted in this study
1 × 10−6. was the shear stress transport (SST) k–ω [39] in order to
The fluid flow was modeled as a two-dimensional, unsteady, simulate the boundary layer turbulence, due to its success in
turbulent, viscous incompressible flow. The incompressible accurately predicting flow separation and reattachment phe-
URANS equations consisting of the continuity equation, which nomena [40, 41] as mentioned in previous studies of similar
represent the conservation of mass, and the momentum equa- flows [14, 19, 22, 30].
tion, representing the conservation of momentum which can The transport equations for the SST k–ω model are given
be written as follows [34]: by [39]:
𝜕 ( )
[ ]
u =0 (1) 𝜕 𝜕 𝜕 𝜕k
𝜕x i i (𝜌k) + (𝜌kui ) = 𝛤 + Gk − Yk (4)
𝜕t 𝜕xi 𝜕xj k 𝜕xj
( )
𝜕 𝜕P 𝜕 𝜕ui 𝜕
( ⟨ ⟩)
𝜌 𝜕t𝜕 ui −𝜌 u�j u�i
( )
+ 𝜌 (ui uj ) = − +𝜇 + [ ]
𝜕xj 𝜕xi 𝜕xj 𝜕xj 𝜕xj 𝜕 𝜕 𝜕 𝜕𝜔
(𝜌𝜔) + (𝜌𝜔ui ) = 𝛤 + G𝜔 − Y𝜔 + D𝜔 (5)
(2) 𝜕t 𝜕xi 𝜕xj 𝜔 𝜕xj
where ui is the ensemble-averaged velocity in the i-direction,
where Гk and Гω represent the effective diffusivity of k and
and symbols ρ, µ and P correspond to the density, molecular
ω, respectively. Gk and Gω indicate the generation of k and
viscosity and ensemble-averaged pressures, respectively.
ω due to mean velocity gradients, respectively. Yk and Yω
The single term in the continuity equation represents the
are the dissipation of k and ω, respectively. Dω denotes the
mass advection, while the first and second terms on the left-
cross-diffusion term.
hand side in the momentum equation are the transient and
momentum advection terms. The three terms on the right-
hand side of the momentum equation are the pressure force,
2.2 Parameters of suction flow control
viscous ⟨
term ⟩ and the Reynolds stress. The Reynolds stress
term −𝜌 u�j u�i in Eq. (2) contains the mean of the product
In this study, the effect of perpendicular suction (θjet = − 90°)
of turbulent fluctuating velocities modeled by the Boussin- and tangential suction (θ jet = − 30°) was computation-
esq approximation. This term requires a turbulence model ally studied over the NACA 0012 flapped airfoil at angles
to close the system of equations. of attack of 8° and 12° for five different hinge positions
(H = 0.7c, 0.75c, 0.8c, 0.85c and 0.9c) and a flap deflec-
𝜕ui 𝜕uj
( )
2
⟨ ⟩
−𝜌 u�j u�i = 𝜇t + − 𝜌k𝛿ij (3) tion (δf) of 15°. Three control parameters including jet loca-
𝜕xj 𝜕xi 3 tion (Ljet = 0.1), jet velocity ratio (Rjet = 0.15) and jet angles

Fig. 1  Schematic of the NACA

0012 flapped airfoil with flow
control parameters

13
86 Page 4 of 14 Journal of the Brazilian Society of Mechanical Sciences and Engineering (2020) 42:86

(θjet = − 30°, − 90°) were used, which are schematically length is placed on the airfoil upper surface to simulate the
shown in Fig. 1. suction jet as stated by Dannenberg and Weiberg [43]. They
In order to apply suction jet, inlet velocity was considered indicated that an increment in the suction area beyond 2.5%
as a boundary condition. The jet entrance velocity is defined of the chord length will not significantly increase the lift. In
as [25]: Eqs. 5–7, Rjet was set equal to 0.15, θjet had values of − 30°
(tangential) and − 90° (perpendicular), Ujet was set equal to
Ujet
Rjet = (6) 1.095 m/s and U∞ had a value of 7.3 m/s.
U∞

(
u = Ujet cos 𝜃jet + 𝛽
)
(7) 3 Computational domain

The computational domain was extended from 6 chords

(8)
( )
v = Ujet sin 𝜃jet + 𝛽 upstream to 12 chords downstream, and the upper and
lower boundaries were extended 6 chords from the airfoil
Here, θjet is the angle between the local jet surface and jet
with 1 m chord length, large enough to prevent the outer
entrance velocity direction [19, 20, 25, 30]. It should be
boundary from affecting the near flow field around the air-
noted that a negative θ indicates the suction condition. The
foil [20]. The computational grid was a C-type structured
suction slot was located at 10% of the chord length from
grid which is shown in Fig. 2a. On the airfoil surface,
the leading edge, which is highly recommended by previ-
no-slip conditions were applied. Inlet, upper and lower
ous studies [3, 20, 42]. U∞ is the free-stream velocity and β
boundaries were considered as velocity inlet, and outlet
is the angle between the free-stream velocity direction and
was set as pressure outlet (Fig. 2b). Three different compu-
the local jet surface. A jet with a width of 2.5% of the chord
tational domains with 4, 6 and 8 chords upstream to 9, 12

Fig. 2  Schematic view of a a

C-type structured grid of NACA
0012 and b boundary conditions

13
Journal of the Brazilian Society of Mechanical Sciences and Engineering (2020) 42:86 Page 5 of 14 86

and 15 chords downstream from the airfoil were generated 4.5

to consider the pressure coefficient (CP) for domain extent 4
α=10°
independence test as shown in Fig. 3. The domain extent α=16°
independence test was done at Re = 5 × 105 and α = 18°. 3.5

It is clear that the domain with 6 chords upstream and 12 3

chords downstream is sufficient for the present study.
2.5

y+
2

1.5
4 Grid independence study
1

To verify the adequacy of the final grids, a thorough grid 0.5

independence study was conducted by increasing the 0
number of grid cells, namely 19,000, 34,000, 52,000 and 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
74,000 to compare the lift and drag coefficients. Table 1 x/c
indicates the lift and drag coefficients with a Reynolds
number of 5 × 10 5 and angles of attack of 10° and 16°. Fig. 4  y+ distribution over the NACA 0012 airfoil at angles of attack
of 10° and 16°
There was a negligible difference between the results
of the smallest grid and the grid with 52,000 cells. The
grid size giving a grid independent result with reasonable 5 Validation
accuracy was selected to be 52,000 nodes. Figure 4 illus-
trates the value of y+ which was ensured to be less than 1 In the present study, the pressure coefficient ­(Cp) was com-
reported by Versteeg and Malalasekera [44]. Table 2 shows pared with the numerical results of Huang et al. [20], Farhadi
further details of four different grids including the number et al. [45] and experimental results of Gregory and O’reilly
of cells and ­y+ distribution. [46] at Reynolds number of 5 × 105 (Ma = 0.021), and angle
of attack of 18°, as shown in Fig. 5. As it can be seen, our
numerical results had a good agreement with the mentioned
numerical and experimental results.
-14
4c-9c Moreover, computational results of lift and drag coeffi-
-12 6c-12c cients were compared with the available experimental results
8c-15c of Critzos et al. [47], Jacobs et al. [48], as well as with
-10
numerical results of Huang et al. [20] and Yousefi et al. [19]
Pressure coefficient

-8 which are presented in Fig. 6a, b. A good agreement was

achieved between computational results and mentioned stud-
-6
ies. Therefore, it can be stated that the numerical procedure
-4 and the applied turbulent model had acceptable accuracy.
Figure 7 shows the comparison between the present CFD
-2
results of pressure coefficient (Cp) with DNS results of Shan
0 et al. [37] and LES results of Zhang et al. [38] at Re = 105 and
α = 4° for the airfoil without flap. It is clear that the present
2
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 CFD results completely simulated the separation point in a
x/c good agreement with DNS and LES results due to the adverse
pressure gradient on the airfoil surface. With these proper
Fig. 3  Domain extent independence test

Table 1  Grid independence Number of cells α = 10° α = 16°

study for the NACA 0012 airfoil
at angles of attack of 10° and Lift coefficient Drag coefficient Lift coefficient Drag coefficient
16°
19,000 1.091 0.0362 1.016 0.1252
34,000 1.022 0.0279 0.982 0.1106
52,000 0.981 0.0232 0.974 0.0928
74,000 0.979 0.0231 0.972 0.0926

13
86 Page 6 of 14 Journal of the Brazilian Society of Mechanical Sciences and Engineering (2020) 42:86

Table 2  Details of the grid cells Grid Number of cells Growth factor The height of the Max y+ Min y+ Average y+
and y+ distribution at α = 10° first cell (m)
and Re = 5 × 105
1 19,000 1.1 3 × 10−3 22.13 7.72 10.14
2 34,000 1.1 1 × 10−4 9.11 3.32 4.89
3 52,000 1.1 1 × 10−5 3.93 0.04 0.85
4 74,000 1.1 5 × 10−6 3.86 0.03 0.81

(a) 1.8
CL- Experimental, Jacobs et al.
1.6
CL-Numerical, Huang et al.
-14 1.4 CL-Numerical, Yousefi et al.
Experimental, Gregory and O'reilly

Lift coefficient (CL)

1.2 CL-Numerical, Present study
-12 Farhadi et al.
Huang et al.
Present study 1
-10
Pressure coefficient

0.8
-8
0.6
-6
0.4
-4
0.2
-2 0
0 2 4 6 8 10 12 14 16 18
0
Angle of attack (α)
2 (b) 0.35
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 CD- Experimental, Critzos et al.
x/c 0.3 CD-Numerical, Huang et al.
CD-Numerical, Yousefi et al.
Fig. 5  Comparison between the pressure coefficient (Cp) with 0.25
Drag coefficient (CD)

CD-Numerical, Present study

the numerical results of Huang et al. [20], Farhadi et al. [45] and
experimental results of Gregory and O’reilly [46] for Re = 5 × 105 0.2
(Ma = 0.021) and α = 18°
0.15

0.1

0.05
validations (Experimental, DNS and LES), it was guaranteed
0
that the URANS approach captured right physics in the region 0 2 4 6 8 10 12 14 16 18
where the flow is separated. Angle of attack (α)
In addition, a comparison between the effect of flap deflec-
tion on lift coefficient for α = 10° of the present study with Fig. 6  Comparison between a lift coefficient and b drag coefficients
numerical results of Obeid et al. [1] and experimental results with experimental results of Critzos et al. [47], Jacobs et al. [48] and
numerical results of Huang et al. [20] and Yousefi et al. [19]
of Williamson et al. [49] over the NACA 0015 is presented
in Fig. 8. A good agreement was achieved between compu-
tational results and mentioned numerical and experimental 6 Results and discussion
studies.
In Table 3, due to the lack of experimental results on suc- Figures 9 and 10 demonstrate the effect of hinge position
tion flow control over the NACA 0012 airfoil at the present (H) on the lift and drag coefficients at α = 8° and 12° with
flow conditions, the computational results of CL/CL-Base ratio flap deflection of 15°. In the case without suction, five
were compared with the numerical results of Huang et al. [20] hinge positions from 0.7 to 0.9 of the airfoil chord length
under a Reynolds number of 5 × 105, the angle of attack of 18° compared to the case with perpendicular and tangential
and a suction jet width of 2.5% of the airfoil chord length. As it suctions. It is obvious that the lift coefficient increased
can be seen, the numerical results had an excellent agreement dramatically as the hinge position moved to the trailing
with the numerical results of Huang et al. [20].

13
Journal of the Brazilian Society of Mechanical Sciences and Engineering (2020) 42:86 Page 7 of 14 86

-1.5 1.9 0.06

DNS-Shan et al.
LES-Zhang et al. 1.8
-1 Present study 0.05
1.7
Pressure coefficient

0.04
-0.5 1.6

CD
CL
1.5
0.03
0
1.4 CL-without suction
CL-tangential suction
CL-perpendicular suction 0.02
0.5 1.3 CD-without suction
CD-tangential suction
CD-perpendicular suction
1.2 0.01
1 0.65 0.7 0.75 0.8 0.85 0.9 0.95
-0.1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Hinge positions (H)
x/c
Fig. 9  Effect of hinge positions (H) on lift and drag coefficients at
Fig. 7  Comparison between the pressure coefficient of the present α = 8° with flap deflection of 15° and perpendicular and tangential
study with DNS results of Shan et al. [37], LES results of Zhang et al. suctions
[38] at Re = 105 and α = 4° for the airfoil without flap

2.4 0.08
2
0.07
1.8 2.2
1.6 0.06

1.4 2
0.05

CD
1.2
CL

1.8 0.04
1
CL

0.03
0.8 1.6
CL-without suction
0.6 CL-tangential suction 0.02
Experimental, Williamson CL-perpendicular suction
1.4 CD-without suction
0.4 0.01
Present study CD-tangential suction
0.2 CD-perpendicular suction
Obeid et al. 1.2 0
0 0.65 0.7 0.75 0.8 0.85 0.9 0.95
0 5 10 15 20 25 30
Hinge positions (H)
Flap deflection (δf)°

Fig. 10  Effect of hinge positions (H) on lift and drag coefficients at

Fig. 8  Comparison between the effect of flap deflection on lift coeffi- α = 12° with flap deflection of 15° and perpendicular and tangential
cient for α = 10° with numerical results of Obeid et al. [1] and experi- suctions
mental results of Williamson et al. [49]

coefficient was increased only 2.5% and 2.8% for the cases
without and with perpendicular suction, respectively.
Table 3  Comparison between the computational results of the
CL/CL-Base ratio with the numerical results of Huang et al. [20] On the other hand, tangential suction had less effect than
perpendicular suction. It caused to increase in the lift coef-
Case study Suction (θjet = − 30°) Suction (θjet = − 90°) ficient only 0.4% and decreased the drag coefficient about
CL/CL-base CL/CL-base 2.8% at H = 0.9c. By comparing the results at angels of
Huang et al. 1.342 1.502 attack of 8° and 12°, it was found that both lift and drag coef-
Present study 1.338 1.497 ficients raised by increasing the angle of attack. Although
the value of the lift coefficient in both cases increased at
H = 0.9c, the decrease in the drag coefficient was more sig-
nificant. This is more evident in the comparison of the lift-
edge of the airfoil. The maximum increment in the lift to-drag ratio (CL/CD).
coefficient was obtained at H = 0.9c which was about 25% Figures 11 and 12 depict the value of CL/CD at α = 8°
and 26% more than H = 0.7c for the cases without and with and α = 12°. As it can be seen from this figure, the value
perpendicular suction, respectively. Furthermore, the drag of C L/C D increased as the hinge position moved to the

13
86 Page 8 of 14 Journal of the Brazilian Society of Mechanical Sciences and Engineering (2020) 42:86

Without suction
suctions was obtained about 35.8% and 25.1% higher than
36
Perpendicular suction that of the case without suction at H = 0.9c.
Tangential suction
Figure 13 shows the distribution of the skin friction
34
coefficient ­(Cf) for the upper surface of the NACA 0012
airfoil without and with perpendicular suction at an angle
32
of attack of 8° for two different hinge positions (H = 0.7c
and H = 0.9c), respectively. After applying suction flow
CL/CD

30
control, the skin friction coefficient increased sharply near
28 the suction control area because the shear stress increased
within the boundary layer. The recirculation zone was
26 located after the hinge positions where the reverse flow
was generated. Consequently, the skin friction coefficient
24 of this zone was less than zero. Applying suction posi-
0.65 0.7 0.75 0.8 0.85 0.9 0.95 tively increased the value of the skin friction coefficient at
Hinge positions (H) the recirculation zone. This positive effect is more evident
by applying suction when the hinge position was located at
Fig. 11  Effect of hinge positions (H) on the lift-to-drag ratio (CL/CD) H = 0.9c where the skin friction coefficient is greater than
at α = 8° with flap deflection of 15° and perpendicular and tangential
zero with suction flow control, which is another evidence
suctions
of the reduction in the recirculation zone using suction
flow control.
38 A control surface deflection alters both the mean cam-
Without suction
Perpendicular suction ber line and the effective angle of attack of the airfoil sec-
36
Tangential suction tion, and it causes a change in the pressure distribution on
34 the geometry. Hinge moment (h) is non-dimensionalized
using the hinge moment coefficient which is defined as
32
[50, 51]:
CL/CD

30
h
Ch =
q∞ sf cs (9)
28

26 where Sf is the reference area of the control surface, typi-

cally taken to be a planform area of the control surface of
24
the hinge line. Cs is the control surface reference chord and
22 is typically the mean distance from the hinge line to the
0.65 0.7 0.75 0.8 0.85 0.9 0.95
Hinge positions (H)

0.08
Fig. 12  Effect of hinge positions (H) on the lift-to-drag ratio (CL/CD)
Without suction, H=0.7
at α = 12° with flap deflection of 15° and perpendicular and tangential 0.07 Suction, H=0.7
suctions Without suction, H=0.9
0.06 Suction, H=0.9

0.05
trailing edge of the airfoil. The maximum value of CL/CD
was obtained when the hinge position was located at 0.9 0.04
of airfoil chord length (H = 0.9c). In addition, applying
Cf

0.03
perpendicular and tangential suctions increased the value
of CL/CD for all cases. The maximum value of CL/CD for 0.02

perpendicular and tangential suctions were obtained about 0.01

15% and 6.9% higher than that of the case without suction
0
at H = 0.9c and α = 8°. The effect of using perpendicu-
lar and tangential suctions was more effective at α = 12°. -0.01
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
This is due to the higher increase of lift coefficient and x/c
decrease of the drag coefficient. At this angle of attack, the
maximum value of CL/CD for perpendicular and tangential Fig. 13  Distribution of the skin friction coefficient (Cf) for without
and with perpendicular suction at α = 8° and H = 0.7c and H = 0.9c

13
Journal of the Brazilian Society of Mechanical Sciences and Engineering (2020) 42:86 Page 9 of 14 86

trailing edge of the control surface measured normal to the positions. The maximum reduction in hinge moment was
hinge line. q∞ is the free-stream dynamic pressure. Dynamic obtained about 3% when the hinge position was located at
pressure can be written as: H = 0.9c.
Figure 15a, b presents the flow pattern over the flapped
1
q∞ = 𝜌 U2 (10) airfoil with hinge positions (H) equal to 0.7 and 0.9 at α = 8°
2 ∞ ∞
in terms of flow velocity (m/s) and streamlines. A compari-
where 𝜌∞ is the free-stream density, and U∞ is the free- son between these figures indicates as the hinge position
stream velocity. moved toward the trailing edge of the airfoil, the recircula-
The effect of hinge positions on the hinge moment ­(Ch) tion zone became smaller. It is clear that when the hinge
at α = 8° is illustrated in Fig. 14. It is obvious that the hinge position was located at 0.9 of airfoil chord length (H = 0.9c),
moment increased gradually with negative value as the hinge the flow exhibited the maximum acceleration over the upper
position moved to the trailing edge of the airfoil. Although surface of the airfoil and caused a reduction in the size of
the maximum increase in the lift coefficient was obtained recirculation zone. Figure 16a, b demonstrates the effect of
at H = 0.9c, it produced a higher hinge moment compared perpendicular suction on the flow pattern over the flapped
to the other hinge positions. On the other hand, applying airfoil for the hinge position equal to 0.9 at α = 8°. As it
suction positively decreased the hinge moment for all hinge was expected, perpendicular suction had a positive effect on
controlling the flow separation over the flapped airfoil. This
-0.45
was more significant when the hinge position was located at
0.9 of the airfoil chord length (H = 0.9c). The perpendicular
-0.5 Without suction
suction was more capable to reduce the size of the recircula-
-0.55 Suction
tion zone and transfer it downstream. Figure 17a, b shows
-0.6 the comparison of using perpendicular and tangential suc-
-0.65 tions on the flow pattern over the flapped airfoil for the hinge
-0.7
position equal to 0.9 at α = 8°. It is clear that perpendicular
Ch

suction is more effective in the reduction in the recirculation

-0.75
zone than tangential suction.
-0.8 The size (Sr) and length (Lr) of the recirculation zone at
-0.85 α = 8° are presented in Table 4. Lr represents the distance
-0.9 between the endpoint of the airfoil to the point S and Sr
-0.95
denotes the size of the recirculation zone in terms of the
0.65 0.7 0.75 0.8 0.85 0.9 0.95 airfoil chord length (x/c) which is illustrated in Figs. 15, 16
Hinge position (H) and 17. In Table 4, the effect of perpendicular and tangential
suctions on the NACA 0012 flapped airfoil is compared to
Fig. 14  Effect of hinge positions (H) on the hinge moment the case without suction. It was obvious that as the hinge

Fig. 15  Flow pattern over the flapped airfoil with a H = 0.7c and b H = 0.9c at α = 8° colored by velocity magnitude (m/s) contours (color figure
online)

13
86 Page 10 of 14 Journal of the Brazilian Society of Mechanical Sciences and Engineering (2020) 42:86

Fig. 16  Effect of perpendicular suction on the flow pattern over the flapped airfoil a without suction and b with suction for H = 0.9c at α = 8°
colored by velocity magnitude (m/s) contours (color figure online)

Fig. 17  Effect of suction on the flow pattern over the flapped airfoil with a perpendicular suction and b tangential suction for H = 0.9c at α = 8°
colored by velocity magnitude (m/s) contours (color figure online)

Table 4  Size (Sr) and length Case Hinge positions Length (Lr) of the Size (Sr) of
(Lr) of the recirculation zone in (H) recirculation zone the recircula-
terms of the airfoil chord length tion zone
(x/c) at α = 8°
The flapped airfoil 0.7c 1.12 0.5
0.9c 1.13 0.2
Perpendicular suction on flapped airfoil 0.9c 1.18 0.09
Tangential suction on flapped airfoil 0.9c Two recirculation Two recir-
zones: (1) 1.07, (2) culation
1.13 zones: (1)
0.15, (2)
0.11

13
Journal of the Brazilian Society of Mechanical Sciences and Engineering (2020) 42:86 Page 11 of 14 86

position moved forward to the trailing edge of the airfoil, The contours of mean vorticity for the flapped airfoil
the size of the recirculation zone decreased and its location in three different cases including without suction, perpen-
on the upper surface of the airfoil moved to the trailing edge dicular suction and tangential suction at H = 0.9c at α = 8°
of the airfoil. On the other hand, the effect of perpendicular are shown in Fig. 19a–c. It was obvious that perpendicular
suction was more considerable. This caused to minimize suction had a more significant effect on suppressing vortex
the size of the recirculation zone from 0.5 to 0.09 of airfoil shedding and made the flow structures became steady but
chord length and also transferred it from 1.13 to 1.18 of the tangential suction had less effect on flow structures and
airfoil chord length. The reduction in the size of the recir- the shape of vortex shedding. As the hinge position moved
culation zone was more significant using perpendicular suc- toward the leading edge of the airfoil, computational time
tion at H = 0.9c compared to using tangential suction. Using and cost were remarkably increased because of vortex shed-
tangential suction divided the recirculation zone into two ding and intensive fluctuation appeared in the flow structures
separate zones which have opposite circulation directions. around the flapped airfoil.
Figure 18a–c illustrates the contours of turbulent kinetic Figure 20 indicates the boundary layer velocity profiles
energy (TKE) ­(m2/s2) for the flapped airfoil in three differ- at different chordwise positions on the upper surface of
ent cases including without suction, perpendicular suction the airfoil at α = 8° for both cases including H = 0.7c and
and tangential suction at H = 0.9c and α = 8°. It is clear that H = 0.9c. As it can be seen, the magnitude of velocity was
a high amount of TKE was generated around the flapped decreased as the x/c increased for both hinge positions.
airfoil. As it can be seen, the wake regions had a sinusoidal Near the trailing edge, the deceleration of the bound-
pattern that got weakened by applying perpendicular and ary layer can be obviously seen at H = 0.7c compared to
tangential suctions. Applying perpendicular suction had a H = 0.9c, and there was a reverse flow in the boundary
more remarkable effect on suppressing the TKE region. layer for H = 0.7c. In addition, at 3/4 chord size, turbulent

Fig. 18  Contours of turbulent kinetic energy (TKE) ­(m2/s2) for the flapped airfoil a without suction, b perpendicular suction and c tangential
suction at H = 0.9c and α = 8°

Fig. 19  Contours of mean vorticity for the flapped airfoil a without suction, b perpendicular suction and c tangential suction at H = 0.9c and
α = 8°

13
86 Page 12 of 14 Journal of the Brazilian Society of Mechanical Sciences and Engineering (2020) 42:86

Fig. 20  Boundary layer velocity 0.08 0.08

H=0.7 H=0.7
profiles at 1/4 and 3/4 chord
0.07 H=0.9 0.07 H=0.9
sizes on the upper surface of the
1/4 chord length 3/4 chord length
airfoil 0.06 0.06
0.05 0.05
0.04 0.04

y/c

y/c
0.03 0.03
0.02 0.02
0.01 0.01
0 0
-0.2 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 -0.2 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6
u/U∞ u/U∞

0.08 25
Without suction
0.07
With suction 20
0.06
3/4 chord length
0.05
15
0.04
y/c

u+
0.03 Log-law
10
0.02
u+=y+
0.01
5
Without suction
0
With suction
-0.2 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6
0
u/U∞ 0.1 1 10 100 1000
y+
Fig. 21  Effect of perpendicular suction on the boundary layer veloc-
ity profiles at 3/4 chord size on the upper surface of the airfoil Fig. 22  Effect of perpendicular suction on the boundary layer veloc-
ity profiles as a function of the viscous scale at 3/4 chord size

profiles produced a thicker boundary layer than did the

√
𝜏w
laminar profiles at 1/4 chord size. Also, Fig. 21 illustrates u𝜏 = (11)
𝜌
the effect of applying perpendicular suction on the bound-
ary layer velocity profiles at 3/4 chord size on the upper u
surface of the airfoil at α = 8° and H = 0.9c. It is clear that u+ = (12)
u𝜏
perpendicular suction had a positive effect on the accel-
eration of the boundary layer for all positions. Accelera- yu𝜏
tion of the boundary layer increased the lift coefficient and y+ = (13)
𝜈
lift-to-drag ratio (CL/CD), decreased the drag coefficient
and boundary layer thickness and also delayed the flow where u𝜏 is called the friction velocity or shear velocity. 𝜌
separation. is the fluid density. 𝜏w represents the wall shear stress. 𝜈 is
Figure 22 depicts the effect of perpendicular suction on kinematic viscosity. u+ is the dimensionless velocity: the
the boundary layer velocity profiles as a function of vis- velocity u parallel to the wall as a function of y (distance
cous scale profiles at 3/4 chord size on the upper surface of from the wall), divided by the friction velocity u𝜏 . y+ is the
the airfoil at α = 8° and H = 0.9c. Turbulent flows are sig- wall coordinate: the distance y to the wall, made dimension-
nificantly affected by the presence of walls [52]. Successful less with the friction velocity u𝜏 and kinematic viscosity 𝜈 .
predictions of turbulence models used for wall-bounded tur- It can be seen that u+ = y+ for y+ < 5 (viscous sublayer) in
bulent flows depend on an accurate description of the flow in both cases as shown in Fig. 22. Furthermore, the logarithmic
the near-wall region. In the viscous sublayer (for a distance region agreed well with the log-law. It is clear that suction
from the wall y+ < 5), u+ can be approximated by u+ = y+ had no significant effect on the boundary layer in the vis-
[53, 54]. The u+ profile will be obtained from the resolution cous sublayer but it positively affected the acceleration of
of the momentum equation [53]. the boundary layer in the log-law region and outer layer.

13
Journal of the Brazilian Society of Mechanical Sciences and Engineering (2020) 42:86 Page 13 of 14 86

7 Conclusions 4. Günther B, Carnarius A, Thiele F (2010) Numerical investiga-

tion of active flow control applied to an airfoil with a camber
flap. In: King R (ed) Active flow control II. Springer, Berlin, pp
The present study mainly focused on the effect of hinge 45–61
position on the aerodynamic performance of the NACA 5. Günther B, Thiele F, Petz R, Nitsche W, Sahner J, Weinkauf T,
0012 flapped airfoil and analyzing the appropriate hinge Hege HC (2007) Control of separation on the flap of a three-
element high-lift configuration. In: 45th AIAA aerospace sci-
position. In addition, perpendicular and tangential suctions ences meeting and exhibit, p 265
were employed to control the boundary layer separation 6. Prabhakar A, Ohri A (2013) CFD analysis on MAV NACA 2412
and improve the aerodynamic performance at each different wing in high lift take-off configuration for enhanced lift genera-
hinge positions. The simulations were carried out at a chord tion. J Aeronaut Aerosp Eng 2:125
7. Amini Y, Liravi M, Izadpanah E (2018) The effects of Gurney
Reynolds number of 5 × 105 (Ma = 0.021). Two-dimensional flap on the aerodynamic performance of NACA 0012 airfoil in
incompressible URANS calculation was performed using the rarefied gas flow. Comput Fluids 170:93–105
the CFD flow solver. For turbulence simulation, the SST k–ω 8. Xie YH, Jiang W, Lu K, Zhang D (2016) Numerical investiga-
turbulence model was employed. It was concluded that per- tion into energy extraction of flapping airfoil with Gurney flaps.
Energy 109:694–702
pendicular suction can effectively improve the aerodynamic 9. Fatahian E, Nichkoohi AL, Salarian H, Khaleghinia J (2019)
performance of the flapped airfoil and the hinge position Comparative study of flow separation control using suction and
can play a notable role in changing the separation zone. The blowing over an airfoil with/without flap. Sādhanā 44(11):220
lift coefficient increased dramatically as the hinge position 10. Mohammadi M, Taleghani AS (2014) Active flow control
by dielectric barrier discharge to increase stall angle of a
moved to the trailing edge of the airfoil. Using perpendicular NACA0012 airfoil. Arab J Sci Eng 39(3):2363–2370
suction caused to increase the lift coefficient and decrease 11. Cattafesta LN III, Sheplak M (2011) Actuators for active flow
the drag coefficient. The maximum increment in the lift control. Annu Rev Fluid Mech 43:247–272
coefficient was obtained at H = 0.9c which was about 25% 12. Yang L, Li J, Cai J, Wang G, Zhang Z (2016) Lift augmenta-
tion based on flap deflection with dielectric barrier discharge
and 26% more than H = 0.7c for the cases without and with plasma flow control over multi-element airfoils. J Fluids Eng
perpendicular suction, respectively. The maximum value 138(3):031401
of CL/CD for perpendicular and tangential suctions were 13. Svorcan JM, Fotev VG, Petrović NB, Stupar SN (2016) Two-
obtained about 35.8% and 25.1% higher than that of the case dimensional numerical analysis of active flow control by steady
blowing along foil suction side by different URANS turbulence
without suction at H = 0.9c and α = 12°. It was obvious that models. Therm Sci 20(6):1–14
as the hinge position moved forward to the trailing edge of 14. Hao W, Ding Q, Li C (2019) Optimal performance of adaptive
the airfoil, the size of the recirculation zone decreased and flap on flow separation control. Comput Fluids 179:437–448
its location on the upper surface of the airfoil moved to the 15. Shehata H, Zakaria M, Hussein A, Hajj MR (2018) Aerody-
namic analysis of flapped airfoil at high angles of attack. In:
trailing edge of the airfoil. At a certain hinge position, the 2018 AIAA aerospace sciences meeting, p 0037
effect of applying perpendicular suction was considerable 16. Hafien C, Mbarek TB (2019) Reduced order model for the lift
and caused to minimize the size of the recirculation zone coefficient of an airfoil equipped with extrados and/or trailing
and also transferred it to the trailing edge of the airfoil. The edge flexible flaps. Comput Fluids 180:82–95
17. Genç MS, Kaynak Ü, Lock GD (2009) Flow over an aerofoil
reduction in the size of the recirculation zone was more sig- without and with a leading-edge slat at a transitional Reynolds
nificant using perpendicular suction at H = 0.9c compared to number. Proc Inst Mech Eng G J Aerosp Eng 223(3):217–231
using tangential suction. It was concluded that perpendicu- 18. Tung C, McAlister KW, Wang CM (1993) Unsteady aerody-
lar suction had a positive effect on the acceleration of the namic behavior of an airfoil with and without a slat. Comput
Fluids 22(4–5):529–547
boundary layer for all positions. Acceleration of the bound- 19. Yousefi K, Saleh R, Zahedi P (2014) Numerical study of blow-
ary layer increased the lift coefficient and lift-to-drag ratio ing and suction slot geometry optimization on NACA 0012 air-
(CL/CD), decreased the drag coefficient and boundary layer foil. J Mech Sci Technol 28(4):1297–1310
thickness and also delayed the flow separation. 20. Huang L, Huang PG, LeBeau RP, Hauser T (2004) Numerical
study of blowing and suction control mechanism on the NACA
0012 airfoil. J Aircr 41(5):1005–1013
21. Fatahian H, Salarian H, Nimvari ME, Fatahian E (2018) Numer-
ical study of suction and blowing approaches to control flow
References over a compressor cascade in turbulent flow regime. Int J Auto-
mot Mech Eng 15(2):1–21
22. Tadjfar M, Asgari E (2018) Active flow control of dynamic stall
1. Obeid S, Jha R, Ahmadi G (2017) RANS simulations of aero-
by means of continuous jet flow at Reynolds number of 1 × 106.
dynamic performance of NACA 0015 flapped airfoil. Fluids
J Fluids Eng 140(1):011107–011110
2(1):1–27
23. Fatahian E, Nichkoohi AL, Fatahian H (2019) Numerical study
2. Anderson John D (2001) Fundamentals of aerodynamics.
of the effect of suction at a compressible and high Reynolds
McGraw-Hill series in aeronautical and aerospace engineering,
number flow to control the flow separation over Naca 2415 air-
3rd edn. McGraw-Hill, New York, pp 283–335
foil. Prog Comput Fluid Dyn Int J 19(3):170–179
3. You D, Moin P (2008) Active control of flow separation over an
airfoil using synthetic jets. J Fluids Struct 24(8):1349–1357

13
86 Page 14 of 14 Journal of the Brazilian Society of Mechanical Sciences and Engineering (2020) 42:86

24. Ahmed T, Amin MT, Islam SR, Ahmed S (2014) Computational 42. Resendiz Rosas C (2005) Numerical simulation of flow separation
study of flow around a NACA 0012 wing flapped at different flap control by oscillatory fluid injection. Doctoral dissertation, Texas
angles with varying Mach numbers. Glob J Res Eng 13:4–16 A&M University
25. Genç MS, Kaynak Ü, Yapici H (2011) Performance of transition 43. Dannenberg RE, Weiberg JA (1952) Section characteristics of a
model for predicting low Re aerofoil flows without/with single 10.5-percent-thick airfoil with area suction as affected by chord-
and simultaneous blowing and suction. Eur J Mech B Fluids wise distribution of permeability. NASA TN 2847
30(2):218–235 44. Versteeg HK, Malalasekera W (2007) An introduction to compu-
26. Lei J, Liu Q, Li T (2017) Suction control of laminar separation tational fluid dynamics: the finite volume method. Pearson Educa-
bubble over an airfoil at low Reynolds number. Proc Inst Mech tion, London
Eng G J Aerosp Eng. https​://doi.org/10.1177/09544​10017​72702​5 45. Farhadi A, Rad EG, Emdad H (2017) Aerodynamic multi-param-
27. Zhou Y, Hou L, Huang D (2017) The effects of Mach number on eter optimization of NACA0012 airfoil using suction/blowing jet
the flow separation control of an airfoil with a small plate near the technique. Arab J Sci Eng 42(5):1727–1735
leading edge. Comput Fluids 156:274–282 46. Gregory N, O’reilly CL (1973) Low-speed aerodynamic charac-
28. Ma D, Li G, Yang M, Wang S (2018) Research of the suction flow teristics of NACA 0012 aerofoil section, including the effects of
control on wings at low Reynolds numbers. Proc Inst Mech Eng upper-surface roughness simulating hoar frost. HM Stationery
G J Aerosp Eng 232(8):1515–1528 Office, London
29. Fluent A (2009) 12.0 theory guide. Ansys Inc., 5(5) 47. Critzos CC, Heyson HH, Boswinkle Jr RW (1955) Aerodynamic
30. Yousefi K, Saleh R (2015) Three-dimensional suction flow control characteristics of NACA 0012 airfoil section at angles of attack
and suction jet length optimization of NACA 0012 wing. Mec- from 0 deg to 180 deg (No. NACA-TN-3361). National Aeronaut-
canica 50(6):1481–1494 ics and Space Administration, Washington
31. Ge M, Zhang H, Wu Y, Li Y (2019) Effects of leading edge 48. Jacobs EN, Sherman A (1937) Airfoil section characteristics as
defects on aerodynamic performance of the S809 airfoil. Energy affected by variations of the Reynolds number. NACA report no.
Convers Manag 195:466–479 586-231
32. Monir HE, Tadjfar M, Bakhtian A (2014) Tangential synthetic jets 49. Williamson G (2012) Experimental wind tunnel study of airfoils
for separation control. J Fluids Struct 45:50–65 with large flap deflections at low Reynolds numbers. Master’s the-
33. Liu D, Nishino T (2019) Unsteady RANS simulations of strong sis, aerospace engineering, The University of Illinois at Urbana-
and weak 3D stall cells on a 2D pitching aerofoil. Fluids 4(1):40 Champaign, Champaign, pp 14–60
34. Ockfen AE, Matveev KI (2009) Aerodynamic characteristics of 50. Simpson CD (2016) Control surface hinge moment prediction
NACA 4412 airfoil section with flap. Int J Nav Archit Ocean Eng using computational fluid dynamics. Master dissertation, The
1(1):1–12 University of Alabama
35. Yang XI, Sadique J, Mittal R, Meneveau C (2016) Exponential 51. Perry III B (1978) Control-surface hinge-moment calculations for
roughness layer and analytical model for turbulent boundary layer a high-aspect-ratio supercritical wing. Technical memorandum
flow over rectangular-prism roughness elements. J Fluid Mech 78664, NASA
789:127–165 52. Hinze JO (1975) Turbulence. McGraw-Hill, New York
36. Nichkoohi AL, Tousi AM (2014) Numerical investigation of high 53. Absi R (2009) A simple eddy viscosity formulation for turbu-
pressure and high Reynolds diffusion flame using Large Eddy lent boundary layers near smooth walls. Comptes Rendus Mec
Simulation. J Therm Sci 23(5):412–421 337(3):158–165
37. Shan H, Jiang L, Liu C (2005) Direct numerical simulation of 54. Hanjalić K, Launder BE (1976) Contribution towards a Reynolds-
flow separation around a NACA 0012 airfoil. Comput Fluids stress closure for low-Reynolds-number turbulence. J Fluid Mech
34(9):1096–1114 74(4):593–610
38. Zhang W, Zhang Z, Chen Z, Tang Q (2017) Main characteristics
of suction control of flow separation of an airfoil at low Reynolds Publisher’s Note Springer Nature remains neutral with regard to
numbers. Eur J Mech B Fluids 65:88–97 jurisdictional claims in published maps and institutional affiliations.
39. Menter FR (1994) Two-equation eddy-viscosity turbulence mod-
els for engineering applications. AIAA J 32(8):1598–1605
40. Vuddagiri A, Halder P, Samad A, Chaudhuri A (2016) Flow analy-
sis of airfoil having different cavities on its suction surface. Prog
Comput Fluid Dyn Int J 16(2):67–77
41. Catalano P, Amato M (2003) An evaluation of RANS turbulence
modelling for aerodynamic applications. Aerosp Sci Technol
7(7):493–509

13