Suction and pulsed blowing for control of local wing-engine-slat-cut out flow separation: The INAFLOWT Project

: An experimental study on the application of active flow control (AFC) to a 1:8.4 scale model of a swept wing in a landing configuration was conducted. The wing is fitted with an Ultra High Bypass Ratio (UHBR) engine nacelle. The highly efficient UHBR engines characterized by a large diameter that interferes with the flow around the wing, degrading its performance. An innovative active flow control device, creating steady suction and pulsed blowing (PB), was installed in the leading-edge region of the wing, above the nacelle, and its performance was experimentally evaluated. The effects of the suction and PB mechanisms were examined individually and simultaneously, using relevant normalized parameters to pave the way to a full-scale wind tunnel test. It was shown that the AFC devices increase the lift by up to 3%, redirected the flow to the desired down-stream direction and reduced the size of the separation zone created due to the implementation of the UHBR nacelle. The next step is validating the small-scale results of this study in full-scale wind tunnel tests that hopefully make the technology flight test ready.


INTRODUCTION
The number of commercial flights rises steadily, the impact on the environment becomes more substantial as aircrafts release a considerable amount of greenhouse gasses to the atmosphere, contributing to climate change. One way of reducing aircraft emissions is by installing a more fuel-efficient engine such as the Ultra High Bypass Ratio (UHBR) turbofan [1]. This engine includes a significantly larger diameter fan relative to the engine core, which maximizes the mass flow around the core with minimal speed increment, thus reducing the thrust specific fuel consumption while increasing the overall diameter of the nacelle. Incorporating this larger diameter engine in a typical airliner poses an engineering challenge [2]. The engine is to be mounted under the wing, to reduce cabin noise, thereby its size is limited by the minimal safe ground clearance below, and by the wing above. The solution for this challenge, with the current length of the landing gear, is to mount the engine closer to the wing, with the slats cut wider away from the pylon to avoid clashing with the nacelle. The close coupling of the nacelle, wing and slat complicates the flow and creates structures that significantly alter the direction and near-wall momentum of the flow, leading to local separations, reduced lift and increased drag [3][4][5]. As part of several European-Union-funded studies, the effects of this close coupling have been examined in several wind tunnel experiments [3,5,6] and Computational Fluid Dynamics (CFD) simulations [3,[7][8][9]. These studies show that the main harmful effects can be traced to a region downstream of the nacelle, between the inboard side of the pylon and the slats, where the flow is reversed (marked blue in Figure 1), and longitudinal vortices are prominent ( Figure 2). Thus, to incorporate the UHBR engines in new commercial aircraft, a method to redirect and reenergize the near-wall flow is needed. region is inboard of the nacelle [3].
To deal with the local flow separation described above, a passive flow control solution in the form of a strake was positioned on the nacelle [6,10]. However, this solution is insufficiently effective [6]. Therefore, the current solution is for Active Flow Control (AFC) devices [11], creating periodic excitation [12,13]. Periodic excitation has been shown in the past as a plausible flow control mechanism for increasing lift, stall postponement, controlling complex flow structures and even acting as virtual control surfaces [13][14][15]. Conveniently, the leading edge (LE) of the disturbed area can house these devices where slats do not exist directly downstream of the pylon. The implementation of AFC devices in the wing-pylon junction has been examined in past efforts as part of the EU AFLoNext project [16], where pulsed jet actuators (PJA) and Synthetic Jet Actuators (SJA) were simulated, installed [17] and tested. These tests showed promising results, but the high mass flow consumption (of PJA) and the insufficient momentum provided (by the SJA), call for a different device to be used.

Figure 2. CFD simulation of the longitudinal vortices
formed by the geometry of the slat edges, strake and nacelle [9].
Within the scope of the EU TAU lead INAFLOWT project [18,19], an innovative AFC device in the form of Suction and Oscillatory Blowing (SaOB) actuator is used to alleviate the local boundary layer separation caused by the UHBR installment. The SaOB actuator (Figure 3), developed at Tel-Aviv University, is a fluidic device that consists of two main elements; the ejector which is creating the steady suction and the oscillator (by-stable fluidic amplifier) which is responsible for the oscillatory blowing flow. Its principle of operation and performance has been extensively studied [20,21]. This paper presents the experimental study conducted on a small-scale model of a wing segment with a UHBR nacelle at Tel Aviv University's Knapp-Meadow low-speed wind tunnel. SaOB actuators were adapted to fit into the small-scale model, their performance was evaluated, and they were installed in the model for a series of wind tunnel tests. The conducted tests covered a range of free-stream velocities, angles of attack (AoA) and actuation magnitudes. Wind tunnel test data was collected to assess the change in pressure distributions, integrated lift and the development of streamwise vortices in the near wake. This study was conducted along with CFD simulations by the INAFLOWT project partners from VZLU [21], University of Stuttgart (USTUTT) [3] and Israel Aerospace Industry (IAI). The goal of this study is to verify that the SaOB actuator is capable to re-energize, redirect the flow and to recover the lift loss due to the geometry dictated by the UHBR nacelle, in preparation of a near-full-scale wind tunnel test to be conducted in the Central Aerohydrodynamic Institute (TsAGI) in Zhukovskiy, Russia. The structure of the paper is as follows: section 2 will present the experimental setup; first by describing the wind tunnel, wing test model and AFC devices. Section 2 will also describe the sensors and equipment used in the tests. Section 3 will cover the actuators' bench-top test results, a parametrized discussion of the wind tunnel test results and an analysis of the near wake of the model. Conclusions and future directions will be discussed in section 4. Figure 3. A sketch of a SaOB device, parts and operation principle [22]. In continuation to previous studies on the specific geometry of swept-wing with UHBR nacelle, the current study uses a scaled-down version of the large-scale model (LSM) (Figure 4) that was windtunnel tested at the closed-loop subsonic wind tunnel T-101 at TsAGI [23][24][25] [26], with flap and slat deflected at 35° and -28°, respectively, representing a landing configuration. The pylon and nacelle are mounted at the mid-span of the wing with the slats on both sides of the nacelle. The nacelle is a hollow outer shell of an UHBR engine with only the hollow core of the engine inside it and a strake outside, on the inboard side of the nacelle (shown in Figure 5).   Figure 5). The main element of the SSM was fabricated from a 5 [mm] thick shell of fiberglass-epoxy, the flap and slats were made of CNC-machined aluminum and the nacelle with its core and pylon were 3D printed from PLA material, assembled from a few parts, polished and painted. The lift generated by the wing was evaluated by integrating the pressures along two sections in the wing, flap and slats. These inboard (55 pressure taps) and outboard (69 pressure taps) sections are seen in red and green, respectively, in Figure 6, along with the forward (14 pressure taps) and aft (11 pressure taps) spanwise rows of pressure taps, marked by blue and yellow lines, respectively. Several additional pressure taps are located between the actuators and in other key locations in the separated flow region, counting a total of 159 pressure taps on the surface of the wing. The inner cavity of the wing could be accessed by removing a cover plate at the bottom of the wing.

The AFC array module
Incorporating the SaOB devices in the SSM posed several challenges, due to the small volume of the LE segment above and inboard of the nacelle, where AFC is required ( Figure 7). Ultimately, it was decided to fit 4 fluidic oscillators and 8 corresponding suction holes in this volume. The AFC array module is a single, 3D printed block, manufactured using an advanced SLA printer with a printing layer resolution of 25 [µm]. It was also decided that the suction and fluidic oscillator of each actuator will be two independent mechanisms, thus allowing flexibility to control the steady suction and pulsed blowing (PB) independently at different actuation magnitudes, unlike in the conventional SaOB actuator where they are coupled. This de-coupling allows us to study the optimal relation between suction and PB, and support the design of the full-scale SaOB device, which will be used in the next step of the INAFLOWT project, outside the scope of this paper.   As mentioned, the two AFC mechanisms are adjustable separately. The fluidic oscillators inletport can be supplied with compressed air using a mass-flow controller located outside the WT. The inlet air is routed into the wing through the side of the model into a manifold, spreading the flow evenly to the actuators' 4 inlet ports. The suction holes are open to the inner cavity of the wing, where low pressure can be applied using an external pump. The suction cavity pressure was measured during the bench-top tests, and therefore the suction velocity is known. The 4 fluidic oscillators are operated at identical phases relative to each other, meaning that when one slot of an oscillator is blowing, all slots on the same side will blow, creating synchronized oscillation. The fluidic oscillator periodic behavior is created by the feedback tube (FBT) (seen in Figure  3 and Figure 9), by connecting the feedback tubes of the oscillators, using a synchronization tube (termed sync-tube), the oscillators operate "inphase". The frequency of the oscillators is measured by an unsteady pressure sensor installed in the sync-tube. PB and suction velocity measurements were conducted using a hot-wire (HW) probe mounted on a motorized computer-controlled traverse system capable of moving in 3 dimensions. The probe was calibrated by positioning it in the potential core of a calibration jet with a diameter of 20 [mm]. The hot-wire's voltage is measured in a series of points, covering all the relevant velocity range. At each point, the measured voltage is correlated with the jet's velocity, eventually fitting a third-order Polynomial. An overall velocity uncertainty of less than 5% is achieved at low speeds and 2% at speeds above 10 [m/s]. Prior to its installment in the wing, the AFC array module was fixed to a test-bench setup to assess the PB mechanism performance. The air velocity at the exit slots was measured using a hot-wire probe ( Figure 10.b.). A Y-coordinate (spanwise) velocity scan (termed Y-scan) of the PB slots was conducted by moving the HW along the row of slots using a motorized, computer-controlled, traverse system. After this scan, the HW was positioned in the location where peak velocity was detected, at each slot, and the air-velocity for a range of inlet pressures was measured at this location (termed P-scan). These tests were repeated for 4, 3, and 2 of the most inboard actuators so the effectiveness of a different number of actuators can be assessed. To create and test arrays of 4, 3, and 2 actuators, the inlettube and the FBT were disconnected and the synctube was adapted to connect only between active actuators. Although the suction mechanism is separate from the PB mechanism, throughout the WT tests, whenever a fluidic oscillator was disconnected the corresponding pair of suction holes were taped over as well, to closely mimic SaOB operation. The AFC module was glued in place at the LE of the wing, using epoxy resin, and polished to create a seamless surface. The interior of the wing was then sealed completely, apart from the 8 (or 6 or 4) suction holes in the AFC module enabling suction.
To measure the suction air velocity, the entire wing model (nacelle removed) was fixed to a test-bench and a HW probe was carefully positioned at the suction holes center at the surface elevation ( Figure 10.a.). The suction air velocity was measured for a range of suction pressures and distances from the surface to assess the sensitivity of the measurement to the placement of the probe, found to be very high.

Near wake measurements
The air-velocity vectors and vortex flow-field in the wake of the model were calculated from the measurements of an array of 4, seven-holes probes (7HP). These probes calculate the wind velocity vector by correlating the pressures measured by the seven holes in its conical tip ( Figure 11) as described in-depth in [27]. The probes were calibrated using the rapid calibration method described in [28], where a calibration matrix was created to translate the pressures reading to an air-velocity vector (as described in depth in [29]). The probes are mounted on a traverse system connected to two stepper motors capable of moving the probes in two axes, scanning a 2D (Y-Z) surface normal to the freestream direction. The tip of the probes was located 1350 [mm] (approximately 3.5 [x/c]) from the model's AoA rotation axes ( Figure 12). With these two axes of motion, the velocity vectors and vorticity of almost the entire wind tunnel crosssection can be measured and velocity vectors calculated. Figure 11. Front view (a) and side view (b) of a Seven-Holes Probe with the ports numbered and principle axes marked [27]. The inner diameter of a single hole is 0. 6 [mm] and the diameter of the entire probe is 4 [mm].

AFC module array bench-top tests
In this section of the paper, we describe the results of the actuator module calibration in otherwise still air. First, a Y coordinate (visible in Figure 10.b.) hot-wire scan was performed at the PB exit slots using a single HW. The HW probe was carefully positioned so it will traverse along the Y-coordinate at a constant distance from the surface and the PB slots of the AFC array module. This is crucial to assure an accurate measurement verifying symmetrical behavior for each actuator. The results shown in Figure 13 display symmetrical behavior for the blowing velocity of all 4 actuators. The 8 peaks shown in Figure 13 correspond to the 8 blowing slots of the actuators, where the mean velocity (blue rhombus, left-side ordinate) is almost uniform across each slot, peaking close to the splitter. The standard deviation (STD) of the air velocity (red squares, left-side ordinate) is used as an indication of the switching quality (SQ) of the actuators (green triangles, right-scale ordinate). The SQ indicates the quality of the oscillations is simply defined as the ratio between the mean and STD velocities and it indicates if the flow fully switches sides from one slot to the other [22]. The SQ of all 4 actuators is above 0.5 and at places reaches 0.7, which is the SQ of a pure sine wave (between 0 and 2), indicating near-full switching between the two exit slots of each actuator as it oscillates. Inlet pressure scans (P-scan) were conducted by positioning the HW in the point where peak velocity was measured in one of the slots in the Y-scans, as shown in Figure 13. Note that this procedure requires careful positioning of the probe near the slot, as the velocity changes significantly at different distances from the slot. This means that relocating the probe in the exact peak-velocity point is inherently prone to uncertainty. Figure 14 shows the blowing velocity vs. the mass flux through one fluidic oscillator for an AFC array of 2, 3, and 4 of the inboard actuators. A maximum velocity of 100 [m/s] was reached. The P-scan measurements produce an important correlation between the inlet pressure, Pin, and the PB velocity, as the latter cannot be measured in realtime during the wind tunnel tests. The PB frequency is also important for calculating the nondimensional Strouhal number, F + , common between different scales and speeds, or identify any unsteady behaviors in the flow and correlate it to past tests and specific unsteady structures.
Analyzing the frequency measured with the HW probe (and the FBT pressure signal) during a Pscan provides the relation seen in Figure 15. The frequency measured with the unsteady pressure sensor inside the FBT was identical to the HW probe measurements as seen in Figure 15. The excitation frequencies were chosen to fit F + ≈1, which is very receptive in flow control applications. The data from the P-scan measurements also show a linear correlation between the PB frequency and velocity.  It is also important to quantify the phase in which each actuator slot operates when analyzing WT measurements. Since it is not possible to conduct intrusive HW measurements in the WT, we can use the unsteady pressure sensor in the FBT and add the phase lag between it and the PB slots at the surface of the wing. We obtain this phase lag by calculating the phase lag between the unsteady pressure sensor and the HW probe, measured during the bench-top tests. The phase lag calculated during a Y-scan of 4 actuators is shown in Figure 16, showing the approximately 180 [deg] phase between each slot pair of each actuator and a phase variation of 30-60° along a given slot. The results of the suction holes velocity scans with the probe positioned accurately at the surface of the holes are presented in Figure 17.a. The data shows a maximum air velocity of nearly 120 [m/s] at a suction pressure of -8. 6 [kPag], which is the maximum suction pressure available in the benchtop tests. It was shown during suction holes measurements that the measured velocity decreases by about 8% for every 0.1 [mm] the HW probe is pulled away from the surface. Moreover, in the WT tests, it is possible to reach a suction pressure of up to -10.3 [kPag]. To evaluate the suction-velocity at higher suction pressures, a linear fit was created for the suction-velocity dynamic-pressure vs. the suction pressure ( Figure  17.b.). This linear fit was used to predict the dynamic pressure at suction pressures higher than the maximum measured in the bench-top tests and to obtain the suction-velocity estimation at the maximum suction pressure of -10.3 [kPag], used in subsequent WT tests.

Pressure distribution and integrated Lift
The actuation magnitudes of the AFC array are presented in the parametric none-dimensional form of momentum coefficient, Cµ, and mass-flow coefficient, Cq. The suction magnitude is indicated in the form of momentum coefficient, Cµ, defined by Seifert [13]: Where ρs is the air density, Us is the suction air velocity, ∞ and ∞ are the free-stream density and flow, respectively, As is the suction holes cross-section area and Ar is the reference area of the model. It is assumed that for suction = ∞ .
Since we have direct measurements of the massflow to the fluidic oscillator, the momentum coefficient for PB will be calculated as such: Where ṁ is the mass flux through one actuator and is the mean blowing velocity as shown in Figure 14, measured by HW during the bench-top calibration stage. However, in this study we use the mass flux of air through the actuators to quantify the efficiency of the system, thus the actuation magnitude of the PB mechanism is presented by the mass-flow coefficient, defined in [11]: The relation between Cµ,PB and Cq is shown in Figure 18: The main interest of this study is to recover the lift lost due to the implementation of the new wingpylon-nacelle geometry, therefore measuring the change in the lift is a key topic of this study. By integrating the pressures along the inboard and outboard pressure tap sections ( Figure 6) it was possible to receive an estimate of the sectional lift coefficient. Note that the two sections are perpendicular to the LE of the wing and thus angled by 28° relative to the free-stream direction. The measured Cp distributions are shown in Figure  19. Data is presented for the inboard (left) and outboard (right) sections for the baseline case and an actuated case. The data of Figure 19 shows changes in Cp in the flap of the outboard section and slight changes in the trailing edge of the inboard slat and the main element of the outboard section. This data shows that the effects of the  Figure 20, displays a distinct linear lift curve up to 9° AoA, after which separation begins, achieving max-CL at 11° AoA and stalling at 12° AoA. Next, a suction scan was conducted with the PB slots still taped over. This set of measurements is presented along with the baseline results ( Figure  20, left) and with the change in lift, ΔCL, relative to the taped baseline measurement (Figure 20, right).
As expected, the no-actuation case with AFC cavities open (dashed blue line, Figure 20) is detrimental for lift, since this case simply adds holes near the LE, disrupting the flow and causing premature separation. However, activating the lowest magnitude of suction alleviates the disruptive effects of the suction holes. As the suction magnitude is increased it helps maintain the linear curve up to 10° AoA. Examining the change in the lift it is indicated that suction has a clear and consistent effect on the lift, increasing it by up to 2.5% (ΔCL≈0.05), whereas further increasing the suction pressure, it seems to saturate at Cµ = 9.2×10 -3 . The effects of steady suction and PB were examined by using each method separately and then simultaneously. The lift and lift increment shown for 35 [m/s] free-stream velocity in Figure , demonstrate the same behavior for baseline cases as seen for 25 [m/s] ( Figure 20). The baseline lift slightly degrades as the suction holes and PB slots are exposed. When no mass flux is introduced into the fluidic oscillators and using only a suction pressure yielding Cµ = 2.5×10 -3 (continuous blue line and triangles in Figure ), the sectional lift coefficient is increased substantially, peaking at 11° AoA. At AoAs higher than 11° the effect of suction diminishes until failing at the stall angle of 12° AoA. When introducing PB alone (green line and rectangles in Figure ) with an actuation   magnitude yielding Cq = 2.5×10 -4 it is apparent that it maintains a steady increase in the lift for AoAs between 10° and 12° (stall angle). Interestingly, when examining the PB with suction case relative to the suction alone case (red line relative to blue line in Figure ), it is apparent that adding PB to suction has a stronger contribution to CL, compared to each mechanism separately, between 11°-12° AoA rather than at AoAs lower than 11°. During the experiments at U∞ = 45 [m/s], the effects of a different number of actuators on the lift were measured at 10° and 11° AoA. In each case, the most outboard actuator(s) were disabled by disconnecting the fluidic oscillator from the inlet manifold and taping over its corresponding suction holes and PB slots. Figure 22. shows the change in the lift as affected by the suction momentum coefficient, calculated using Eq. (1). These results show that the 3 actuators case reaches a slightly higher ΔCL than the 2 actuators case, albeit using a higher Cµ. Whereas the 4 actuators case improves the lift slightly less than the 3 actuators case even when a much higher Cµ per actuator is used. This data allows us to conclude that for suction alone, the 3 most inboard actuators suffice.
To assess the effectiveness of a different number of fluidic oscillators, the PB magnitude was scanned for 2, 3 and 4 actuators while using a constant suction pressure of -7.9 [kPag] (Cµ = 1.2×10 -3 , 1.7×10 -3 , 2.4×10 -3 for 2, 3, and 4 actuators, respectively). The mean nozzle-airvelocity is derived using the data of Figure 14 and is referenced to the free-stream velocity by the Velocity-Ratio (VR) which is a quotient of the two. PB VR in Figure 23, left. It is demonstrated that up to VR = 1.5 the PB has no added effect on the lift, but for higher VR the lift increases significantly. As expected, the more actuators, the better the lift increment for the same relative PB velocity. However, adding more actuators means that a higher mass flux is needed. Presenting the same data versus the mass flow coefficient, Cq, of the actuators (Figure 23, right) shows that the efficiency of the 3 actuators case is higher in terms of the invested mass flux. It is evident that the 3 actuators case reaches almost the same lift as the 4 actuators case for a substantially lower mass flux. Clearly, the contribution of the 4 th actuator is much lower than the 3 rd actuator (in addition to the 2 most inboard actuators).

Drag estimation from wake velocity scans
Once the air-velocity 3D vectors and pressures were obtained for the 2D surface in the wake of the model, an assessment of the drag coefficient can be made. By using the equation to calculate drag from wake measurement derived by van Dam [30] and subtracting from it the thrust created by the AFC mechanisms, we arrive at eq. (4): Eq. (4) includes the Betz correction which considers the pressure gradient in the wake. An alternative form for this formula is suggested by Giles and Cummings [31] where the Betz correction is replaced by air-velocity components in the wake, normal to the free stream direction: Since wake measurements are time-consuming, only a handful of measurements were made. By using both equations (4) and (5) and subtracting the baseline case, we obtain the change in drag coefficient, ΔCD, presented in Figure 23. From this data it appears that at 10° AoA applying AFC is beneficial to drag reduction, reaching about 20% when PB actuation magnitude is increased to Cq = 4.0×10 -4 (or VR = 3). However, at 12° AoA (stall angle) AFC does not lead to drag reduction. The agreement between the two drag estimation methods is ±0.003 of the mean ΔCD.

Longitudinal vortices
To investigate the creation and development of longitudinal vortices, the spanwise pressure distribution on the wing is examined together with the velocity vectors and streamwise vorticity in the wake of the model. The two are linked by Helmholtz's 2 nd theorem and Prandtl's lifting line theory. Figure 25 presents the vortex field (color map) overlaid by the velocity vectors field (small blue arrows) of the Z-Y plane at the wake of the model, and below these, sharing the same horizontal axis, are the spanwise pressure distributions on the airfoil at two x/c stations. The wake velocity vectors field ( & ) is measured using the 7HP data and the longitudinal vortex field ( ̅ ) is calculated using the velocity vectors. The spanwise pressure distribution is measured with the two pressure-taps rows as seen in Figure 6.  Figure 25 show the aft spanwise pressure distributions are almost constant when compared to the forward spanwise pressure distributions. The forward spanwise pressure distribution shows a gradient on both sides of the wing that peaks and plateaus where the nacelle and slat-cut out are located. Closely examining the difference between the 3 test cases in Figure 25, it is observed that as the actuation magnitude is increased a counterclockwise vortex appears between the 4 baseline vortices causing them to shift slightly. The effect of the actuation on the spanwise pressure distributions is apparent in the immediate vicinity of the actuators. Namely, in the forward row of pressure taps, around Y≈-0.08 coordinate a sharp pressure gradient appears as the actuation magnitude is increased.

CONCLUSIONS
A small-scale experimental study of an innovative active flow control device installed on a wingengine-slat-cut out junction model is presented. This study is part of the EU Cleansky 2 INAFLOWT project where few CFD partners and large-scale WT tests are also conducted. The goal is to reduce detrimental flow effects of the coupling between the engine-wing-slat cut-out of this geometry, incorporating a large diameter engine in close proximity to the wing high-lift system. The study makes use of steady suction and pulsed blowing implemented at the leading-edge region of the wing above the nacelle, where a wide slat cut-out exists. The parameter study shows the positive effects of steady suction, applied at about 1.5% chord on the lift coefficient and the added contribution of the pulsed blowing mechanism, implemented at about 10%c, on maintaining the flow attached, where it separated. With the goal of lowering the AFC mass-flow requirements, the effect of a reduced number of actuators is examined, and it is determined that the most effective area for actuation is from the inboard slat edge, towards the pylon, and up to, but not including, the nacelle's pylon. This is supported by surface oil flow visualization showing the reattachment of the flow in the separated area due to actuation, contributing to the understanding of the development of complex flow structures on this geometry. The evolution of longitudinal vortices is also examined and a link between the effect of the actuation, spanwise pressure gradient and a longitudinal vortex is established. It is shown that as a result of the location of pressure sensors, the current method for evaluating the lift coefficient is under-estimating the changes in lift caused by actuation. Nonetheless, a 2.5-3% improvement in lift coefficient is achieved for most test cases of steady suction and pulsed blowing combined. The effect of actuation magnitudes and actuators' location on the flow structures is established parametrically for use in future near-full-scale experiments and industrial applications. Future studies involving large scale experiments planned for the INAFLOWT project will benefit from the current study in several aspects: • The future studies will use the results of this study to define the requirements of a full-scale SaOB device. This device will create suction using the ejector and using the sucked air to increase the PB mass flow. This will significantly reduce the required inlet massflow and yield a much higher energy efficiency compared to pulsed blowing alone used to date. • The full-scale tests will measure the lift coefficient using an integral force balance, creating a more accurate assessment of the AFC impact, alongside Cp measurements that can be directly compared to the current results. • The AFC array of the full-scale tests will be installed in the most effective areas, found in this study, thus lowering the mass-flow requirements even further and compacting the system and its mass flux. • Although drag reduction is not an official goal of this study, the drag can be measured more easily and precisely in the full-scale experiment. Potentially validating that the SaOB AFC also creates certain drag reduction benefits. Even in the current, locally separated flow case.