1 Introduction

Laminar boundary layer separation is a common phenomenon for turbines operating in the low-Reynolds-number regime (Re < 5 × 105) and not only has a negative effect on aerodynamic performance, but is also responsible for aerodynamic noises and flow-induced vibrations [1, 2]. The leading aircraft operating at low Re numbers are unmanned aerial vehicles that come to the fore in military and civil applications. In addition to being widely used in recent years, they have the potential to be deployed in the future [3]. These devices, which perform military surveillance and reconnaissance, have constraints such as quickly locating due to flow-induced noise. On the other hand, hydro/wind turbines used in ocean/urban applications have to comply with strict noise limitations. Consequently, boundary layer separation is usually an unfavorable phenomenon as it adversely affects the performance of aerodynamic vehicles.

The boundary layer at the onset of the pressure recovery region for an airfoil at low Reynolds number flows may continue to be a susceptible laminar layer (Fig. 1). Besides, in the laminar region, waves start to form. Due to the nature of the geometry of the airfoil, the pressure increases, thus adverse pressure gradient forms as it is close to the trailing edge of the airfoil. In the laminar region, this laminar boundary layer, incapable of withstanding the developing adverse pressure gradients, often results in flow separation and forming a separated shear layer. In the separated shear layer, the transition process continues, and spanwise vortices and 3D breakdown causing exponential and nonlinear growth in the flow occur [4,5,6,7]. Depending on the flow characteristics, the flow can remain separated, or it can be followed by the transition in the separated flow and flow reattachment as a result of gaining adequate energy. As seen in Fig. 1 sketched using our oil-flow visualization experimental result, this dead region trapped between separation and reattachment points is called a LSB.

Fig. 1
figure 1

Laminar separation bubble over an airfoil

Full size image

Although numerous studies have been carried out in recent years to understand the physics underlying this phenomenon, it is still current and worth investigating the topic [6,7,8,9]. For this reason, scientists studied flow control to attenuate or even relocate the formation of LSB [7,8,9,10,11,12,13,14,15,16,17,18,19]. With respect to their definitions [6,7,8,9], flow control methods are categorized into two groups, namely active and passive. In the active flow control methods, the flow is forced, whereas the passive flow controllers induce the flow without auxiliary energy. Vortex generators (VGs) [10] have been utilized as widespread passive control methods until now. Besides VGs, other passive control techniques have been offered over decades including riblet, slat-flap, roughness element, the control plate, dimple, groove, bio-inspired devices, slits, control rods and to name but a few [11,12,13,14,15,16,17,18,19,20,21,22,23]. An efficient hydrofoil design was numerically studied to reduce flow separation over an S1210 hydrofoil using a dimple with vortex generators and a dimple with tubes by Kundu [11]. The aerodynamic performance increased at a higher angle of attack, and the stall angle was delayed. Experimental studies were carried out for the drag reduction on an airfoil with riblets by Chamorro and Sotiropoulos [12]. Aerodynamic force measurements were performed, and using the riblet provided drag reduction. Especially, the most efficient result was for a completely covered airfoil with the riblet. Beyhaghi and Amano [13] used a slot through the leading edge of a NACA 4412 airfoil to study its impact on aerodynamic performance. The optimum values of the position and width of the slot and inlet angle were determined. The lift coefficient improvement was observed, while the drag did not decrease. Genç et al. [14] investigated numerically and experimentally the flow over NACA 4412 airfoil with roughness. Using the roughness, the transition at a low Reynolds number was forced to become the bypass transition with flow momentum increasing in the boundary layer with resulting lift increment and drag reduction. The study of Rinoie et al. [15] studied flow control with a rectangular plate at the leading edge of the NACA 0012 airfoil using an actuator that automatically changed the plate height based on the angle of attack. The upper sides of a wind turbine blade were changed with spherical dimples by Sedighi et al. [16] to observe the effect of dimples on the aerodynamic performance of the turbine. It was concluded that dimples could be effective in energy output. An airfoil with a groove was considered numerically [17, 18]. In these studies, the positions and shapes of grooves were analyzed at different low Reynolds numbers, and the lift-to-drag ratio improvements by groove were shown. Moreover, aerodynamic noise was investigated, and the noise reduction effect at the trailing edge direction was observed. A bio-inspired leading edge was performed to prevent ice formation to delay the ice accretion time [19]. And, the spanwise continuous ice was converted to discontinuous ice. It was [20, 21] presented the effects of wavy or tubercules on leading edge on the flow. These systems raised the maximum aerodynamic efficiency and decreased the induced drag. Larabi et al. [22] investigated numerically the influence of a cylindrical rod on the flow over NACA0012. The rod eliminated the boundary layer separation with a decrease in drag and an increase in lift. Karasu [23] evaluated the effects of a slit on the flow of a diamond cylinder using the particle image velocimetry (PIV) system. The slits decreased the drag and fluctuations of velocity.

Except for the conventional passive techniques, aerodynamics researchers focused on adapting flexibility to airfoil surfaces to control the separation bubble, reduce vibration produced by its presence and improve aerodynamic performance as inspired by flying animals. This inspiration in relation to using flexibility on wings is due to their highly stable and agile flight abilities. Researchers have done a good deal of investigations in this field which focused on reducing detrimental load and noise, ensuring a more advantageous lift-to-drag ratio, thus improving the aerodynamic performance. Lian et al. [24, 25] considered the aerodynamics of flexible and rigid wings of similar geometry under micro-air vehicle (MAV) flight conditions. It resulted that the stall delayed on the membrane wing. In addition, they have mentioned that aerodynamic gusts as well as high turbulence levels may possibly be accomplished by using flexibility. Besides, a series of comprehensive studies were carried out on unsteady flow on membrane wings at low Reynolds number flows both experimentally [26,27,28,29,30,31] and computationally [32, 33]. Rojratsirikul et al. [26,27,28] conducted PIV and flow visualization studies on flexible wings with particular emphasis on fluid–structure interaction at low Reynolds numbers. The findings showed a strong coupling between membrane oscillations and the unsteady flow, especially with vortex shedding in the wake. It was concluded that flexibility postpones the stall, considering the rigid and flexible membrane airfoil. On the other hand, two-dimensional rectangular membrane wings with excess length represented higher vibration modes, an earlier roll-up of vortices and smaller separated flow regions, while the membranes with pre-strain generally exhibited more similar behavior to the rigid wing. Genç [29] also experimentally evaluated how excessive length affects the unsteady aerodynamics behavior over a rectangular membrane blade with a low aspect ratio. The results revealed that the separated shear layer was stimulated by an excess length cambered membrane wing. The shear layer was also substantially closer to the surface, which caused wings with excessive length to exhibit separated small regions. The other experimental studies [30, 31] of Genç’s research group were performed on rectangular membrane wings with different aspect ratios at low Reynolds numbers. The experimental results indicated that membrane vibrations occurred at the natural frequencies close to the harmonics of the wake instabilities, and also tip vortices were effective on flexible wings with low aspect ratio. Gordnier [32] carried out a numerical analysis for a 2D flexible wing with a sixth-order Navier–Stokes solver at Re = 2.5 × 103–1.0 × 104. The unsteady flow on the wing was demonstrated to be considerably altered by the dynamic motion of the membrane surface. It was revealed that this dynamic motion caused a delay in the stall and increased lift. Gordnier [33] also performed a simulation on the geometry of the membrane airfoil compatible with the experimental setup used by Rojratsirikul et al. [26]. For angles of attack of α = 8° and α = 14°, a definition of the unsteady fluid/structure interaction was provided, indicating a strong correlation between the unsteady flow behavior and the structural response. This computational finding showed good agreement in conjunction with the experimental study.

The previous experimental studies on flexible membrane airfoils/wings were conducted on fully flexible wings, and very little in the literature can be found related to the partially flexible applications in the last decade. Naderi and Mojtahedpoor [34] studied numerically on local flexible material (LFM) for different airfoils at low Reynolds numbers. It was reported that the stall was delayed and the LSB was shrunk. Kang et al. [35] investigated the effects of flexibility on the lift and the associated flow development to better understand the nature of the interaction between fluid and structure with a numerical model using a method coupling finite elements with a laminar Navier–Stokes solver. The findings demonstrate that the lift of the airfoil with various elastic stiffnesses is significantly influenced by the interaction between the fluid and structure. It was also revealed that the mean deviation raised as the elastic stiffness decreases. Sun et al. [36] experimentally investigated the influences of the locally flexible polydimethylsiloxane membrane characteristics implemented on the upper surface of the NACA 0012 airfoil on aerodynamic performance. After a detailed glance at the study, they performed commonly used experimental techniques which are often encountered in the literature but do not explain the flow physically as much as possible. Their results belonging to the low-definition smoke wire experiment pointed out that the flow separation was hindered with the controlled airfoil compared to its uncontrolled case. Additionally, they just presented the results of aerodynamic force measurement without highlighting the flow-induced fluctuations. Their simple-based aerodynamic force measurement results showed that the stall angle was postponed when utilizing a partially flexible membrane.

Apart from those literature studies mentioned above, the authors of current study investigated the different types of airfoils with LFM mounted on suction surface [37,38,39,40] and suction-pressure surfaces [41, 42] by implementing experimental tests. Their objective was to learn how LFM affected the airfoils in terms on aerodynamic features. Experimental findings at different Reynolds number and angles of attack clearly noticed that LSB formation was enormously diminished, and wake region size was minimized, implying that higher lift coefficient as well as less aerodynamic vibration. Additionally, as indicated and observed in the studies [43, 44], large eddies occurred because of the presence of flow separation on suction surface were suppressed via LFM.

This research aims to gain a comprehensive understanding of flow physics on the partially flexible SD7062 airfoil and its rigid counterpart at Reynolds number of 1.05 × 105 the first in the literature. A time-dependent force measurement, which provides an explanation of flow-induced vibrations, was carried out. A hot-wire experiment using a boundary layer and glue-on probes was performed to give detailed information on the laminar-turbulent transition. The oil-flow visualization technique was utilized to scrutinize the formation of LSB and its effect on surface flow. In addition, all experimental methods were used to show the impact of partial flexibility on the flow of the SD7062 airfoil. At the end of the paper, a comprehensive intermittency analysis using hot-wire results was carried out to analyze the flow physics of the partial flexibility the first in the literature.

2 Experimental Methodology

In this section, the experimental setup including the test specimen as well as experimental rigs were discussed as follows:

2.1 Test Specimen

As the experimental model, the SD7062 airfoil was selected in this study. Two reasons were paid attention for choosing this airfoil: (i) it was proper for flight vehicles such as the UAVs and MAVs which had thin highly cambered airfoils [45], and (ii) there were no detailed studies including flow characteristics formed over the suction surface in the literature. As the technical identification, its maximum thickness was 14% of the chord at a distance of 25.5% of the chord, while its maximum camber was 3.5% of the chord at a distance of 38.8% of the chord. The chord length was 200 mm, and its span was 300 mm. A grooved channel with 5-mm depth along the spanwise was composed between x/c = 0.2 and x/c = 0.4 (at a distance of 40% of the chord from the leading edge) for the location of LFM. As demonstrated in Fig. 2, the 3D printer was used for the manufacturing of the airfoil. During the manufacturing, the rate of material occupancy was selected as ~ 18% in order to not have a tough and bulky airfoil. To obtain a polished surface, the manufactured model was rubbered with an automatic sandpaper vehicle, and then an acrylic spray was employed to paint the surfaces. The double-sided tape with a thickness of 0.05 mm was utilized at each tip of the airfoil to cover all surfaces of the airfoil with a flexible membrane. Concerning the technical description of the flexible membrane, its material was the latex rubber sheet with Young’s modulus (E) of 2.2 MPa and a density of 1 g/cm3. Additionally, end plates were employed at each tip of the airfoil to protect it from the influence of tip vortices. It was foreseen that utilizing LFM on surface of SD7062 airfoil positively affected aerodynamic characteristics by enhancing aerodynamic performance and providing more stable situations.

Fig. 2
figure 2

a Utilized materials, b the closed view of the controlled SD7062 airfoil with LFM

Full size image

2.2 The Research Facility and Experimental Rigs

As demonstrated in Fig. 3, a wind tunnel belonging to the Research Group Laboratory with a square test section of 0.5 m × 0.5 m was employed for all experiments. The length of test section was nearly 2 m, surrounded by transparent plexiglass to visualize the flow phenomena. The wind speed range of the wind tunnel was 1–45 m/s. The free-stream velocity in the tunnel can be switched from 3 m/s (TI = 0.91%) to 40 m/s (TI = 0.35%). Comprehensive information related to the wind tunnel was tabulated as denoted in Table 1.

Fig. 3
figure 3

View of the wind tunnel and experimental rigs

Full size image
Table 1 Technical specification of WEAR group’s wind tunnel
Full size table

In this suction type wind tunnel, the velocity was measured via DANTEC hot-wire system (multi-channel constant temperature anemometer). Figure 4 shows the position of hot-wire probe in the test chamber. In this study, the boundary layer probe (55P15) was utilized to measure the flow in the boundary layer. The measurement was performed chordwise in conjunction with the intervals of x/c = 0.1 except for the location of the flexible membrane. The area between x/c = 0.2 and x/c = 0.4 was measured with an interval of x/c = 0.05 to understand clearly how the LFM affect the flow characteristics. As seen in Fig. 3, the probe was acted by means of a traverse system which could move at both the x-axis and y-axis. Utilizing the traverse system allowed the probe to approach the suction surface as much as possible, leading to achieving more accurate data in the boundary layer. During the experiment, the probe was run at 10 s with a sample rate of 1 kHz for each x/c. Totally, 10,000 data were gathered for each x/c.

Fig. 4
figure 4

Measurement image of the hot-wire experiment

Full size image

In addition to using a boundary layer probe in the hot-wire system, the hot-film sensor (glue-on type-55R47) as shown in Fig. 5 was employed using the hot-wire system at the same x/c points of the suction surface to obtain more information about boundary layer, flow physics and formation of the LSB, etc. The size of this probe was as similar as 1 cent, allowing the measurement without destroying the flow. Regarding the calibration process of the glue-on probe, it was too sophisticated due to the existence of unwanted issues such as dust particles and material properties [14]. Therefore, the calibration process was not conducted in this study. Instead of it, the raw data (the voltage variance) which ensured semi-measurable information was gathered to peruse the boundary layer as employed by Hodson and Howell [46] and Akhlaghi et al. [47]. To compare hot-film sensor data with the boundary layer probe data accurately, the measurements were at the same x/c points and durations with sample rates.

Fig. 5
figure 5

Place of the hot-film sensor along the chordwise

Full size image

To evaluate both mean and time-dependent forces, the aerodynamic force measurement was performed. As illustrated in Fig. 6, the system consisted of a data converter and load cells which measured the lift and drag forces. Before the experiment, the calibration process was conducted by joining the load cells in conjunction with certain weights. Also, it was repeated before each experiment so as to provide more consistency and accuracy. Concerning the data recording, the sample rate was 1 kHz for 10 s which was similar to those just as performed at the hot-wire experiments. The measured forces were converted to the lift (CL) and drag (CD) coefficients:

$$ C_{{\text{L}}} = \frac{{2F_{{\text{L}}} }}{{\rho U_{\infty }^{2} S}} $$
(1)
$$ C_{{\text{D}}} = \frac{{2F_{{\text{D}}} }}{{\rho U_{\infty }^{2} S}} $$
(2)

where U∞ was flow velocity, ρ was air density and S was the area.

Fig. 6
figure 6

Sketch of the force measurement experiment in the wind tunnel

Full size image

In terms of better understanding flow events in the boundary layer, the surface oil-flow visualization technique was fulfilled as seen in Fig. 7. For the oil mixture to successfully signal the formation of the boundary layer, the consistency must be just correct in this technique. In order to avoid changing the conditions at the surface, the inertia forces of the moving oil should also be larger than the forces of viscosity and surface tension. The most popular oils are kerosene, light diesel, and transformer oil, while the most popular pigments are titanium dioxide, china clay, and fluorescent chrysene [48]. In order to observe the pigment deposit on the oiled surface clearly, oleic acid can also be added to the mixture. In this study, the utilized materials for the oil mixture and their amounts were provided. For the oil mixture, titanium dioxide, kerosene, oil and oleic acid were used. The amount of titanium dioxide was relatively less than other materials. They were put in a beaker, and they were mixed for 5 min. The prepared oil mixture was then applied to the airfoil surface with the help of a thin brush.

Fig. 7
figure 7

Oil-flow visualization method, a ingredients for mixed oil, b mixing procedure

Full size image

2.3 Intermittency Analysis

Laminar to turbulent transition starts as turbulent spots and streamwise fluctuations originate in the flow, and then, as the number of spots increases downstream, it merges to form a fully turbulent regime. Turbulent spots are intermittent, and such intermittent (γ) nature of the flow can be measured as a fraction of the time the flow remains turbulent, with values ranging from 0 to 1. γ points out the transition formation. The intermittency having values more than 0.5 indicates that the transition occurs and the energy of the free-stream increases. This topic has been researched for several decades but determining intermittency (especially experimentally) remains challenging. A brief review of the available intermittency models can be noticed in Veerasamy and Atkin [49]. Interestingly, most of the models in the literature have followed a standard procedure including three sequential steps (i) Detector function—sensitize the signal to amplify the turbulent spots; (ii) Criterion function—smooth the sensitized signal to avoid laminar spikes; and (iii) Indicator function (I(t))—select an appropriate threshold value and indicate the laminar and turbulent regimes (turbulent spot) in the signal. Finally, the indicator function can be averaged over a specified amount of time to produce the intermittency distribution. Of these three procedures, choosing the threshold value has a significant impact on the intermittency distribution’s correctness. In this research, we have adopted a dual-slope technique for finding the threshold value, proposed by Kuan and Wang [50] for the experimental results of hot-wire and glue-on probes. In this method, the intersection of two separate slopes will be taken into account as the threshold value. The probability distribution function (PDF) of a sensitized signal is shown versus the values of the sensitized signal. In order to determine the intermittency distribution, a threshold value is used.

$$ \gamma = \frac{1}{T}\mathop \int \limits_{0}^{T} I\left( t \right)dt $$

3 Results and Discussions

Figure 8 shows the aerodynamic force results with fluctuation density for the uncontrolled and the controlled cases at Re = 1.05 × 105. Using the partial flexibility, the stall angle raised from 10° to 12° and the drag force improved, although the lift force decreased. In the lift forces curves versus the angle of attack, force fluctuations were the same in the uncontrolled and the controlled cases, but the increase in the force fluctuations formed after the stall occurred near the stall angle in the controlled case due to the fluid–structure interaction.

Fig. 8
figure 8

Aerodynamic force measurement results with fluctuation density for the uncontrolled and controlled SD7062 airfoil, gray straight line: controlled case, red dotted line: uncontrolled case, Re = 1.05 × 105

Full size image

To understand the flow physics of the partial flexibility case, both glue-on and boundary layer probes operated in the hot-wire system for the uncontrolled and the controlled cases in this study. Since the membrane surface was not suitable for the glue-on probe measurement, this experiment could not carry out in the controlled case. In the control case, instead of that experiment, the hot-wire experiment with the boundary layer probe was used for measurements in near-surface flow. In Fig. 9, the voltage measurement results using the glue-on probe were presented for the uncontrolled airfoil case at Re = 1.05 × 105, α = 8°. In the graphs in Fig. 9, time-dependent voltage variance for different locations over the suction surface of the airfoil was sketched. While the voltage variance in the laminar region fluctuated at a low level, with the onset of the transition process, these fluctuations increased according to the development of flow phenomena. For example, when T-S waves began to occur, the voltage value increases as seen at the x/c = 0.2 point, but when spanwise vortices and their 3d breakdown occurred, exponential and nonlinear changes caused larger fluctuations in the voltage variance as seen at the x/c = 0.3 and x/c = 0.4 points. After completing the transition from laminar to turbulence, the variance fluctuations were raised fully.

Fig. 9
figure 9figure 9

Voltage variance of glue-on probe experiment for the uncontrolled case, Re = 1.05 × 105, α = 8°

Full size image

In Fig. 10, the voltage variance of the boundary layer probe was given graphically for both the uncontrolled and the controlled cases of the airfoil at Re = 1.05 × 105, α = 8°. When it was first looked at the uncontrolled case, it was seen that consistent results with the glue-on probe results were obtained. At the x/c = 0.5 location and beyond, due to the transition to turbulence formation, the fluctuations were fully manifested. The important point in such controlled case graphs was the flow at the flexible membrane region and beyond. With the use of flexible membrane material between x/c = 0.2 and x/c = 0.4 over the airfoil, the oscillation of this flexible material triggered the transition to turbulence and a bypass transition occurred, which caused the flow to gain momentum and to be the reattached flow.

Fig. 10
figure 10figure 10figure 10figure 10

Voltage variance of hot-wire experiments for both the uncontrolled and the controlled cases, Re = 1.05 × 105, α = 8°

Full size image

Figure 11 shows that transition onset location (where γ start deviating from 0) measured using both the hot-wire probe and glue-on probe are not the same for Re = 1.05 × 105, where the transition onset at x/c = 0.35 for hot-wire measurement as compared to x/c = 0.3 for the glue-one probe. This discrepancy can be reasoned by the position of the hot-wire. In free-stream turbulence-induced boundary layer transition, intermittency decrease in the wall-normal direction, therefore in the present experiment hot-wire may not be close enough to capture the near wall intermittency, which in turn decreases in intermittent value compared to the glue-on probe.

Fig. 11
figure 11

Intermittency values graph at Re = 1.05 × 105, α = 8° from hot-wire data for both the uncontrolled and the controlled cases

Full size image

Laminar, transitional and turbulent flow regions via glue-on and boundary layer probes voltage variance mentioned in Fig. 10 were analyzed using by oil-flow visualization experimental technique in Fig. 12 at Re = 1.05 × 105 for different angles of attack in the uncontrolled case. In Fig. 12, the airfoils whose surfaces were marked by each 10% chord were given as a top view. The separation, the reattachment and the trailing edge separation lines were pointed out using green, red and blue lines, respectively. As seen in Figs. 9 and 10, the transition processes such as spanwise vortices and 3D breakdowns began to occur before the separation point. After the separation point, the laminar separation bubble formed until the reattachment point. With the realization of the reattachment phenomena, the flow completely passed into the turbulence and the fully turbulent attached flow was observed. The trailing edge flow separation was seen near the trailing edge of the airfoil at α = 14. Moreover, the laminar separation bubble length at lower angles of attack was smaller than that at higher angles of attack, and the bubble both shrank and moved toward the front of the airfoil as the angle of attack raised.

Fig. 12
figure 12

Oil-flow visualization results at Re = 1.05 × 105 for different angles of attack in the uncontrolled case

Full size image

In order to illustrate flow physics over the airfoil with the LFM, the oil-flow visualization experiment was carried out at α = 8° and Re = 1.05 × 105 for both the uncontrolled and the controlled case as seen in Fig. 13. In this experiment, the oil mixture with titanium dioxide, kerosene, oil and oleic acid damaged the flexible membrane material in a short time. Therefore, this experiment was able to do for observing the flow at the flexible membrane region and beyond at only α = 8° for the controlled case. In the uncontrolled case, the LSB formed between x/c = 0.16 and x/c = 0.42 over the airfoil. The important point in such a controlled case photograph was the first observing the flow at the flexible membrane region and beyond. It was shown via this experiment that with the use of flexible membrane material between x/c = 0.2 and x/c = 0.4 over the airfoil the oscillation of this flexible material triggered the transition to turbulence and a bypass transition occurred, which caused the flow to gain momentum and be the reattached flow.

Fig. 13
figure 13

Oil-flow visualization results at α = 8° and Re = 1.05 × 105 in a the uncontrolled and b the controlled case

Full size image

4 Conclusion

This paper pointed out the effect of the partially flexible membrane material located on the suction surface on the flow over SD7062 wind turbine airfoil at low Reynolds number flow. For this investigation, the time-dependent force measurement, the hot-wire experiment and the oil-flow visualization technique were utilized to scrutinize the formation of LSB, the laminar-turbulent transition and the effect of unique flow control technique using the LFM. In addition, a comprehensive intermittency analysis using hot-wire results of a boundary layer and a glue-on probe was performed to analyze the flow physics of the partial flexibility the first in the literature. Important and new findings from this study are as follows:

  • Using the partial flexibility, the stall delayed from 10° to 12° and the drag force improved, although the lift force decreased. In the lift forces curves, force fluctuations in the controlled case increased near the stall angle due to the fluid–structure interaction.

  • The hot-wire results pointed out over the airfoil laminar, transitional and turbulent regions in the voltage variance graphs, and the transition onset and fully turbulent regions in the intermittency analysis graph. In the time-dependent voltage variance graphs for different locations over the suction surface of the airfoil, the voltage fluctuations in the laminar region were at a low level, after the onset of the transition process these fluctuations raised. Furthermore, as transition flow phenomena such as T-S waves, spanwise vortices and their 3d breakdown occurred, exponential and nonlinear changes caused larger fluctuations in the voltage variance. After completing the transition from laminar to turbulence, the variance fluctuations increased fully.

  • With the use of flexible membrane material between x/c = 0.2 and x/c = 0.4 over the airfoil, the oscillation of this flexible material triggered the transition to turbulence and a bypass transition occurred, which caused the flow to gain momentum and to be reattached flow.

  • The oil-flow visualization experiment results illustrated that the LSB formed between x/c = 0.16 and x/c = 0.42 over the airfoil at α = 8° in the uncontrolled case; the use of flexible membrane material over the airfoil supplied that the oscillation of this flexible material triggered the transition to turbulence and a bypass transition occurred, and to be the reattached flow.

For the future works, the present study can be extended to other technique such as digital image correlation (DIC) for deformation measurement so as to deeply understand fluid–structure interaction and vibration analysis. This research study will be an alternative study in terms of enhancing aerodynamic performance of wind turbines by employing novel control technique.