American Institute of Aeronautics and Astronautics

1

Efficient Control of Separation Using Microjets

Vikas Kumar* and Farrukh S. Alvi.† Department of Mechanical Engineering

Florida A & M University and Florida State University 2525 Pottsdamer Street

Tallahassee, FL-32310, USA

Flow separation in engine inlets and ducts can significantly compromise the performance of aircraft propulsion systems. The study presented herein describes an experimental investigation carried out to study the feasibility of using microjets to control this boundary layer separation in an adverse pressure gradient. The geometry used for this study is a diverging “Stratford ramp” equipped with arrays of 400µm microjets. Detailed PIV investigations and Unsteady Pressure Measurements have been carried out to study the flow control over a wide parametric space. The results indicate that by activating these microjets, the flow becomes attached to the surface with minimal mass flux, less than 0.2% of the primary flow based on 30% Boundary Layer Ingesting duct.

1. Introduction low separation and its control are of considerable interest both from fundamental fluid dynamics and practical perspectives. Flow separation can lead to significant reductions in performance for internal and external flows,

such as lift loss, increase in drag, buffeting, and pressure recovery losses (in engine inlet/transmission ducts), among others. In particular, in the proposed Blended Wing Body (BWB)1 configuration, shown in Fig. 1, flow distortion and separation is a major cause of concern. Serpentine inlet ducts2(Fig. 2) used in these BWB are located on the aft end to reduce the size of aircraft and to diminish radar signatures from the engines. This placement of serpentine inlet ducts requires ingestion of a thick boundary layer developed over the aircraft surface. This thick degraded boundary layer is much more susceptible to separation when it encounters adverse pressure gradients in the inlet/diffuser ducts. The pressure loss due to this separation reduces the overall system efficiency. Moreover, flow distortion and unsteadiness created due to this separation can result in aerodynamic stall and surge of the compressor and the fan blades3,4. Henceforth, it is highly desirable to avoid boundary layer separation as it can significantly compromise the performance of aircraft propulsion systems.

Numerous techniques5-11 have been explored to control this flow separation in adverse pressure gradients. These range from the use of passive devices in form of vanes11, bumps11, Vortex Generators10 (v.g’s) etc. to the use of synthetic jets6, acoustic excitation7,8 as active control devices. However, to date, the performance of these techniques has been somewhat limited. For example, passive devices such as v.g.’s have been found to be effective in controlling separation. However, they need to be optimized for their location, size and other parameters for specific operating conditions and induce parasitic drag when not in use. As a result, active flow control devices have been suggested as an attractive control technique. Some of these active control devices, such as synthetic jets have also been explored by Amitay6. Their flow attachment was, however, limited to a certain region of the flow field and complete attachment was limited to few cases. Similar control devices (piezoelectric synthetic jets) were also employed by Jenkins et al.12 over “Stratford ramp”. Based on their results, Jenkins et al. concluded that synthetic jets did not have sufficient velocity/momentum to provide effective control. Acoustic excitation devices employed by Ahuja8 and Zaman7 showed some benefits. However, these studies were facility dependent9 and as such are limited from a practical perspective.

A detailed investigation, examining the nature of separation and the means to control it, was initiated over the last few years at the Fluid Mechanics Research Laboratory (FMRL) located at the Florida State University (FSU). In this study, strategetically-located microjets were used to control the flow separation generated by adverse pressure gradients. Based on their success in other applications13, their ability to produce relatively high-momentum streams

* Graduate Research Assistant, Mechanical Engineering, FSU, Student Member AIAA. † Associate Professor, Mechanical Engineering, FSU, Senior Member AIAA.

F

35th AIAA Fluid Dynamics Conference and Exhibit6 - 9 June 2005, Toronto, Ontario Canada

AIAA 2005-4879

Copyright © 2005 by the author(s). Published by the American Institute of Aeronautics and Astronautics, Inc., with permission.

American Institute of Aeronautics and Astronautics

2

and because of their relative simplicity and ease of implementation, we anticipated these microjet-based actuators to be effective in separation control. Initial results from this research have been discussed in Kumar et. al.14 where we explored the effectiveness of microjets as an active flow control device for the control of the boundary layer separation in adverse pressure gradients and in the presence of secondary flows. As the results discussed in our earlier paper indicate, microjets were very successful in controlling the flow separation. This paper presents additional results from this ongoing study where we have greatly extended the parametric range over which microjet flow control has been examined. As a result, we have been able to significantly optimize the use of microjets, achieving separation control with mass flow rates that are significantly lower, by almost an order of magnitude in some cases, than those reported in Kumar et al14. More detailed measurements, including that of the velocity field, have also been obtained through which we hope to obtain a better fundamental understanding of the effect of flow control on the flow dynamics. These results including our approach to optimize the use of microjet control is described in this paper.

2. Experimental Set-up

2.1 Test Model The experiments were performed in a subsonic, closed return, wind tunnel with a maximum freestream velocity

of 65m/s (at its present configuration) in the 24”x24” test section. A Pitot static probe was used for measuring the inlet flow speed of the wind tunnel.

The geometry used for the test model is a simple diverging ‘Stratford’ ramp, where a picture of this model and its schematic are shown in Fig. 3a and Fig. 3b respectively. In theory, the ramp profile produces a Stratford like pressure gradient15 in the test section. It is similar to a model used at NASA Langley Research Center where other flow control devices are being evaluated12 for comparison. The ramp is instrumented with more than 60 pressure taps along the centerline and laterally across the ramp at selected locations. The model is highly flexible, allowing us to change the base flow as well as the control parameters. For example, the ramp can be rotated about a pivot point giving us the flexibility (see Fig. 3b) to change the adverse pressure gradient experienced by the flow, thereby controlling the nature – size and location – of the separated flow. Fig. 3a and Fig. 3b also indicates the region where PIV measurements were conducted, discussed later in this paper.

The model also has 7 arrays of supersonic microjet incorporated on a modular insert. The locations of these microjets and the modular insert are also shown in Fig. 3a and Fig. 3b. This actuator insert contains seven slots for microjet modules where different microjet modules are capable of supplying momentum at various angles with respect to the local flow and are interchangeable. The microjets are 400 µm in diameter with spacing of approximately 5mm between microjets. In addition, the model is also instrumented with 4 high frequency response pressure transducers for unsteady pressure measurement at selected locations. The location of these pressure transducers was chosen to study the regions both upstream and downstream of separation; the location of these microjets and transducers is presented in Table 1 and is also shown in Fig. 3a.

These microjets were powered by supplying ‘air’ from compressed Nitrogen tanks. Nitrogen was used because of its easy availability in pure form and as it has essentially the same gas dynamic properties as air.

Fig.1- Blended Wing Body (BWB) Configuration Fig.2- Serpentine Inlet

American Institute of Aeronautics and Astronautics

3

Table 1: Ramp Configuration

Ramp Height (H) 2.25˝ Approximate Locations (X/H)

TR1 1.1 TR2 1.6 TR3 2.2

Transducers

TR4 2.7 MJ2 0.7 MJ3 0.9 MJ4 1.0 MJ5 1.3 MJ6 1.5

Microjets

MJ7 1.9 Separation begins 40 m/s 1.7 Reattachment 40 m/s 3.1

2.2 Data Acquisition Detailed PIV measurements and unsteady pressure measurements were conducted to examine the flow-field and

to explore the possibility of using unsteady surface pressure as means of detecting separation. To obtain whole field velocity data in this flow, quantitative measurements were obtained using 2D Particle

Image Velocimetry (PIV) technique. A New Wave Nd-YAG pulsed laser with a repetition rate of 15Hz was used to illuminate the particles introduced into the flow field. Each PIV image pair was then acquired using a Kodak ES1.0 high-resolution CCD camera capable of recording 10-bit digital image pairs in separate frames at a rate of 15-image pairs/second. Further details of this PIV technique can be found in Lourenco16 et al. One of the main advantages of this PIV technique is a novel processing scheme with high spatial resolution that uses image matching to extract the particle displacements, hence the velocities, from particle image pairs17. An average of 1000 such images was used to obtain the velocity field.

Unsteady pressure measurements were made using high frequency pressure Endevco® transducers with a range of 0-1psi. These pressure transducers were calibrated before each data acquisition. The signals through these transducers were amplified and then PC based Data acquisition system, using National Instruments software and hardware, was used to obtain the data.

MJ2 MJ1 MJ3 MJ4

MJ5 MJ6

MJ7

TR1 TR2

TR3 TR4

b) a)

Y

Z X

Ramp L.E

α Microjets

Microjet Insert

Centerline

Base

H

Pivot Angle of Attack

PIV Region

PIV Region

Fig.3- a) Test model mounted in the test section b) Schematic of the test model

American Institute of Aeronautics and Astronautics

4

2.3 Test Conditions We conducted the experiments up to a free-stream velocity of 65m/s. However, due to restrictions imposed by

the cooling capacity of the heat exchanger in the wind tunnel, the tests at 65m/s could only be conducted over a short time, limiting the parametric range of the experiments at this velocity (we are in the process of making modifications to improve the maximum run times at the higher velocities). As such, extensive studies were conducted at a freestream velocity of 40m/s as it illustrates the principal flow features and its response to microjet control. Most of the conclusions drawn based on the 40 m/s are also applicable for the high velocity, 65 m/s, case.

The Reynolds number at the leading edge of the ramp for freestream velocity of 40 m/s is 1.3 x 106. At this velocity, the boundary layer thickness, δ was measured at the leading edge of the ramp at the centerline (using a boundary layer pitot probe) and was found to be 0.75″. The boundary-layer profile was in close agreement with seventh power law profile indicating that the incoming boundary layer is nominally turbulent. Under these conditions, the corresponding displacement thickness, δ* and the momentum thickness, θ at the ramp leading edge were estimated to be 0.12″ and 0.08″, respectively.

3. Results and Discussions

3.1 VELCOITY FIELD MEASUREMENTS 3.1.1 Baseline Case: No Control

Quantitative velocity field data was obtained using 2D-PIV along the centerline plane. A representative of the ensemble averaged velocity field for freestream velocity of 40m/s has been shown in Fig. 4 where the ramp is at an angle of 5° (see Fig. 3b) and flow is from left to right. In these plots, length scales are non-dimensionalized with respect to ramp height, H. (Note that increasing the ramp angle corresponds to increasing the adverse pressure gradient on the ramp surface). A closer look at Fig. 4a reveals that as one proceeds downstream in the vicinity of the surface, there is a rapid deceleration in the fluid velocity, which eventually leads to a region of reverse flow. This reverse flow zone which corresponds to dark blue velocity contours, starting at around X/H=1.7 and extending till X/H=3.1, where H = 2.25″, indicates that flow has separated locally and a separation bubble with recirculating flow is visibly present. In physical dimensions, the separation region for this case is ~3.2″, which is considerably larger to the case where the ramp is at lower angles, e.g. for 0° case, the recirculation bubble is only ~1.9″. This of course is expected, since higher ramp angles results in more adverse pressure gradients hence larger separation regions. Although not shown here, similar results can be observed from the vertical velocity component also where low magnitude velocity moving away from the boundary can be observed in the same region where reverse flow field is seen in the streamwise velocity plots.

To summarize, the adverse pressure gradient generated due to ramp geometry leads to local separation of the incoming boundary layer, where the size of the bubble increases with the freestream velocity and ramp angle (adverse pressure gradient). In the following section, we examine the effect of microjet control on this separated flow.

3.1.2 Effect of Microjet Control

When microjets were turned on at suitable locations, it was observed that, the reverse flow or the separated flow region is eliminated, with very low mass flux. This result can be verified by the velocity contour plot shown in Fig. 4b where the microjet array at mj4, at an angle of 90° relative to the local surface, has been activated at a stagnation pressure of 10psig. This effect of microjets was observed for all the conditions where the separated flow was present for the baseline case. A comparison with Fig. 4a (No control) indicates that with the activation of microjets not only was the reverse flow eliminated but the momentum near the surface is increased significantly. This effect of microjet control and the optimization procedure, i.e. identifying parameters to formulate a control strategy which has a maximal impact with minimal effort is discussed later on in this paper. Although not shown here, similar effect of control was observed at the highest velocity examined (65 m/s), where the extent and the magnitude of the separation region were much greater.

In addition to simply characterizing the effect of separation and its control on the unsteady surface flow, another main goal of this measurement was to identify properties that are relatively sensitive to the state of the flow above the surface. These can then be used as input ‘sensors’ for the control strategy. Since separation is characterized by an increase in the unsteadiness and as such pressure fluctuations on the surface, we decided to examine unsteady surface pressure as means for identifying flow separation.

American Institute of Aeronautics and Astronautics

5

3.2 UNSTEADY PRESSURE MEASUREMENTS High frequency response Endevco® (Model 8510B-1) pressure transducers with a range of 0-1psi were used to

obtain the unsteady pressure data at selected locations. These sensors were placed such that they were located both upstream and downstream of separation location. The location of these transducers has been shown in Fig. 3a. A representative dynamic pressure spectra for one of the transducers (TR4), located inside the separation bubble, is presented in Fig. 5. Measurements shown in the Fig. 5 includes the spectra for flow with and without control at 0° and 5° angle of attack case at X/H=2.7. 3.2.1 Baseline Flow (No Control)

A comparison of the spectra shown as dashed lines – red and black - in Fig. 5 shows that for the baseline flow, the pressure fluctuations increase dramatically as we increase the ramp angle from 0° to 5°. The shift in spectra between the 0° and 5° cases represents an increase by almost a factor of two in the overall PRMS levels. This corresponds to a significant increase in the pressure fluctuations as we increase the ramp angle of attack. This behavior is also reflected in the PIV data; where changing the angle of attack increased the magnitude of mean reverse flow velocity and the size of the separation bubble. This is also accompanied by substantial increase in the velocity fluctuations, VRMS, in the separated region with increasingly large separations. This behavior suggests that the unsteady pressure distribution, either alone or coupled with the mean pressures, may be used as a measure of the separation location and its size. This issue is addressed in more detail in the subsequent sections of the paper.

3.2.2 Effect of Microjet Control

The dynamic pressures with the microjets activated are also presented in Fig. 5 as solid lines. It is clearly evident that the activation of microjets results in a significant reduction in the pressure fluctuations for both the 0° and 5° case. For the cases shown here, the microjets are operating at a stagnation pressure of 25psig at an angle of 90° relative to the local surface. For both 0° and 5° cases the pressures fluctuations have been reduced by more than a factor of 2. This clearly illustrates the beneficial effects of the present flow control strategy, while again pointing to the use of the unsteady surface pressures as a possible input sensor for adaptive control strategies.

To summarize, microjets were very effective in controlling separation and attaching the flow field globally. This phenomenon is also reflected in unsteady pressure distribution over the surface. Would this pressure distribution respond to the microjet control parameters or not is the next question. The answer to this question is explored through a systematic, parametric study of microjet control parameters on properties such as the velocity and the pressure field, presented next.

Fig.4- a) Streamwise velocity data @40m/s: No Control b) Streamwise velocity data @40m/s: MJ4; 90°; 10psig

b) a)

American Institute of Aeronautics and Astronautics

6

4. A Closer look at the Effect of Microjets

4.1 Effect of Microjet Location The location of the actuation has been found to be critical by numerous investigators12. It stands to reason and is

generally believed that actuation closer to the separation location would produce the most beneficial results. From our velocity field measurements, the separation location for the given flow conditions was known. Accordingly, the flowfield was examined for a number of cases where the microjet actuators were placed at varying distances, both upstream and downstream, relative to the separation location. As discussed subsequently, in some detail, the velocity field data reveals that in general the microjets are most effective when activated upstream of the separation location. There appears to be a range of locations upstream of separation, where their effect is optimal; once outside this range, either very close or too far upstream, the effect of control diminishes. Similarly, locating the actuators downstream of the separation is also non-optimal. However, for all the actuator locations examined to date, separation control could be achieved when the microjets were operated at the appropriate pressure. This required pressure increases as one moves outside the optimal actuator placement zone.

To facilitate our discussion, Table 1 provides the approximate locations for the separation, actuators, and the pressure transducers. A representative example of the velocity profiles for the 40 m/s case at a ramp angle of 5° is presented in Fig. 6, where U/Uinf is shown as a function of the vertical distance from the surface (Z/H) at a streamwise location of X/H=2.3. The location at which this data was extracted from the velocity field is shown as a dashed vertical white line on Fig. 4. For the controlled cases, the microjet angle relative to the surface is 68° and the microjet stagnation pressure is 10psig. As evident in Fig. 6a, different locations have a different effect on the velocity profile. Keeping in mind that for this case, the flow separates roughly at X/H~1.7, the activation of mj2, which is significantly far upstream (at X/H ~ 0.7), at 10 psig does not lead to complete attachment and very small reverse velocities can still be seen in the plot (mj2: green curve). In contrast, the flow is completely attached and the velocity profiles become much ‘full’ with the activation of mj5 (orange curve), which is located at X/H~ 1.3, i.e. much close but upstream of separation. Along the same lines, when mj7 (@ X/H ~ 1.9, located downstream of the separation location, is activated the velocity profile deteriorates slightly relative to mj5, although the difference is minimal.

The corresponding unsteady pressure distribution at X/H=2.7, shown in Fig 7a, reveals similar results with higher overall PRMS values for mj2. As was the case for the mean velocity profiles of Fig. 6a, the PRMS distributions of mj5 and mj7 are very similar, primarily, because the flow is attached for both cases. Partial flow attachment for mj2 at these conditions, however, does not imply that the flow can not be attached. For example, using mj2, separation could be completely eliminated using either a different microjet angle or a higher pressure. A sample result at the same location, obtained from the PIV data with different microjet angle is shown in Fig. 6b, where the microjet angle is 90° relative to the ramp surface. It can be noticed that using the same microjet pressure (10 psig), the flow is now attached using mj2. Although not shown here, the results show that when 90° microjets are used, flow attachment is achieved at all actuator locations examined- mj2 through mj7. Similar effects can be seen from the corresponding unsteady pressure measurements shown in Fig. 7b where the spectra for mj2 at two different injection angles are compared. As seen in the plot, there is a significant reduction in the overall PRMS values at 90°

Fig.5- Unsteady pressure spectra for U∞= 40 m/s

American Institute of Aeronautics and Astronautics

7

relative to 68°. Given that the flow was still separated for 68° and attached for 90°, these results further substantiate the fact that a significant reduction in PRMS is an indication of the flow attachment. What PRMS values or range of values, indicate complete attachment has yet to be explored and is discussed later in this paper. What this does indicate is that, with proper scaling, it may be used to identify the state of the flow.

The same effect, although not as dramatic, can also be observed by increasing the microjet pressure. This effect can be seen in Fig. 6c where for the same microjet angle of 68° at the same location, mj2, the microjet pressure is varied. The flow is attached at a pressure of 15psig and the velocity profile continues to become ‘fuller’ as the microjet pressure is increased. The unsteady pressure distribution for this case is shown in Fig. 7c. Similar observations can be made that as the microjet stagnation pressure is increased, the pressure spectra shows a decrease in PRMS values. However, Figs. 6c and 7c also indicate that successive increments in the microjet pressures yield increasingly lower returns in terms of additional momentum in the boundary layer, i.e. fuller profiles, and lower PRMS values. This issue is further addressed in § 4.3.

Having established that the worst control case scenario (farthest upstream to the separation location) also attaches the flow field completely – if sufficient supply pressures are used, and that the unsteady pressure reflects these characteristics, we now move on to other control parameters. The results discussed above leads naturally suggest two additional parameters of importance, namely, microjet angle and microjet pressure which appear to have a significant impact on the flow properties and the control efficacy.

4.2 Effect of Microjet Angle The microjet angle here refers to the angle relative to the local surface through which the flow is injected into the

main flowstream. Four different angles were studied within the constraints of our present set-up. The effect of microjet angle control was primarily examined at two locations, mj4 and mj5, primarily to

determine if the effect of the microjet angle and location are coupled. Representative velocity profiles at X/H=2.3, with microjets activated at mj4 and mj5 are shown in Fig. 8. Similar to those shown in Fig. 6, these profiles were extracted from whole field velocity data and the microjets are operating at a constant pressure of 10psig for all the cases shown in Fig. 8

In Fig. 8a, the microjet array mj4 has been activated; using the fullness of the profile, i.e. the presence of high-momentum fluid near the surface, these velocity profiles suggest that the best control is achieved with an injection angle of 75°. It can also be seen that the 68° injection angle produces a profile that is not as full as the 90° case. This suggests that there is an optimum angle to achieve the best results. It should be mentioned here that 68° doesn’t mean that the microjets are blowing opposite to the freestream. On the contrary, after taking ramp geometry and ramp angle into account, it is blowing at an angle of 3° with respect to and in the direction of freestream.

To check whether this optimum angle is coupled with location, similar experiments were performed with the array mj5. These results are shown for X/H=2.3 in Fig. 8b. It can be seen that the optimal angle is no longer 75°. In fact, there is very little difference in the profiles in the near –wall region, between the 68°, 75° and 90° injection angles. This is indicative of the complex flow behavior and the strong interdependence between the various control parameters. This also suggests the need for other metric(s) that captures the effect of more than one parameter. Nevertheless, even though we are still struggling to understand the details of the various parametric effects in order to further fine-tune and optimize the use of microjets for separation control, what is encouraging is the fact that this very simple approach is relatively robust, in that it yields significant benefits over a wide range of conditions with very little cost.

The unsteady pressure distributions shown in Fig. 9 further confirm the trends observed in the velocity profiles shown in Fig. 8. In the case of mj4 shown in Fig. 9a, 75° and 105° seems to be performing better. The difference in the spectra and the overall values for the cases are not so significant though. This is probably because the flow is attached for all the control cases. Similarly, as seen in Fig. 9b, 68° and 105° are seen to be performing better than 75°. This indicates that the optimal control angle depends on the control location.

American Institute of Aeronautics and Astronautics

8

Fig.6- Effect of Parametric variation (Location, Angle, Pressure) on mean velocity profilesX/H=2.3, U = 40 m/s, Ramp angle = 5° a) For MJ2, MJ5, MJ7; 68°; 10psig b) For MJ2; 68°, 90°; 10psig c) For MJ2; 68°; 10, 15, 20, 25psig

Fig.7- Effect of Parametric variation (Location, Angle, Pressure) on Unsteady pressure spectra at X/H=2.7, U = 40 m/s, Ramp angle = 5° a) For MJ2, MJ5, MJ7; 68°; 10psig b) For MJ2; 68°, 90°; 10psig c) For MJ2; 68°; 10, 15, 20, 25psig

b) b)

c) c)

a) a)

American Institute of Aeronautics and Astronautics

9

The question that naturally arises is “why is the behavior different for different angles?” A closer study of Fig. 8

suggests that for a given pressure, increasing the injection angle initially makes the velocity profile fuller, however subsequent increments in the injection angle, yield lower gains in terms of added momentum in the near wall region. This can be explained as follows: As we are increasing the angle, we are distributing a part of the vertical component of momentum to the horizontal component. Normally, one would expect that as the horizontal component is increased, the profile should grow stronger and stronger. On the contrary, observations from the above results reveal that is not happening. Or, alternatively that is also happening, but on lower injection angles, some other strong effect is making the profile even more resistant and which is probably weakened at higher injection angles. If one would think about it, the vertical component of momentum has reduced significantly, because of higher injection angles. Hence, direct momentum injection is not the mechanism behind separation control in the

Fig.9- Effect of Parametric variation (Angle) on Unsteady surface pressure X/H=2.7, U = 40 m/s, Ramp angle = 5° a) For MJ4; 10psig b) For MJ5; 10psig

Fig.8- Effect of Parametric variation (Angle) on mean velocity profilesX/H=2.3, U = 40 m/s, Ramp angle = 5° a) For MJ4; 10psig b) For MJ5; 10psig

a) b)

a) b)

American Institute of Aeronautics and Astronautics

10

present approach; in fact the momentum added once the flow is attached is at least an order of magnitude greater than the total momentum injected via the microjets (see Kumar et al14., for more details). Hence, one must look elsewhere for an explanation. It is well-known that jets in cross-flow, which is in essence the flow generated by the actuation of microjets, can generate longitudinal or streamwise vortices18-20 in a boundary layer. These vortices in turn appear to increase cross-stream mixing thus increasing the streamwise momentum of the near-wall fluid. It was our expectation, that in a manner similar to solid vortex generators (but perhaps more efficiently), high-speed microjets would energize the boundary layer fluid by creating strong streamwise vorticity19, thereby enhancing mixing with the more energetic flow. With the higher injection angle and as such, weakened vertical momentum component, the effect created by these streamwise vortices is reduced. This reduction due to secondary effects generated by the microjet is probably not compensated by the redirection of momentum parallel to the surface. This effect of streamwise vorticity needs to further studied and 3D-PIV experiments have been planned to investigate this effect. But, what this does indicates is that it is not just the direct addition of momentum, rather a more powerful secondary effect, created by the streamwise vortices, which is dominant in the flow control.

4.3 Effect of Microjet Pressure As mentioned earlier, and as evident from the results presented so far, microjet pressure is also a significant

parameter. As seen in Fig. 6, at certain locations higher microjet pressure is required to control the flow. With this in mind, we investigated the pressure variation at the ‘optimized control location’, i.e. the location which required the least amount of momentum influx for control, mj5. The microjet supply pressure was varied from 2 to 20 psig in increments of 2 psig and the microjet angle was fixed at 90°. Representative vertical velocity profiles from this experiment are shown in Fig. 10a for X/H =2.3. As seen from Fig. 10a, the separation was well controlled at 6 psig and saturation was observed beyond this pressure. This variation in the control effect is also seen in the unsteady pressure measurements shown in Fig. 10b. The spectra shown in Fig. 10b reveal that overall PRMS values decreases as the microjet pressure is increased, but after 10psig, the microjet effect is saturated.

Another question that is raised is “What should be the criterion for control?” OR, “When have we achieved

separation control where further increases in the microjet pressure are not beneficial?” Should the criterion just be that the flow is incipiently attached everywhere or should the increase in the momentum in the near-wall region of the boundary layer, i.e. the fullness of the boundary layer profile, be also considered? The answer to this question will likely come from a detailed cost function analysis, which is beyond the scope of this paper

However, as an initial step, in order to gain some insight into the cost and benefit of this control, we plot the length of the separation bubble, LB, and the shape factor, H, as a function of microjet pressure in Fig. 11. Variation in the shape factor with microjet pressure has been shown in Fig. 11a, where the shape factor is calculated from the velocity profiles at X/H = 2.3 for the microjet array mj5. The turbulent boundary layer separates near and above a shape factor of H = 2.4. Fig. 11a then shows that at a pressure of 6 psig the flow should be attached. Also, as we go for higher pressures, the shape factor decreases. Once the flow is attached, however, with the additional pressure

Fig.10- Effect of Pressure Variation (for U=40m/s, Ramp angle=5°) a) on mean Velocity profile at X/H=2.3;MJ5;90°;0,1,2,4,6,10,15,25psig b) on Unsteady surface pressure at X/H=2.7; MJ5; 90°;0,2,10,20psig

a) b)

American Institute of Aeronautics and Astronautics

11

applied gain in the shape factor, or alternatively, the stability is not significant and appears to become saturated. This can also be seen from the unsteady pressure spectra in Fig. 10b. The variation in bubble length with microjet pressure shown in Fig. 11b, when extrapolated also shows the same results. Also, both the criteria show that the microjet pressure needed for the flow to be incipiently attached is ~5psig. When extrapolated to a theoretical inlet duct, e.g. a 30% boundary layer ingesting duct – a nominal value used for BWB inlet, this results in a mass flow requirement that is well below the maximum permissible bleed flow values as quoted by engine manufacturers. This suggests that engine bleed flow may be used to supply the gas for the microjet actuators.

4.4 Unsteady Pressure as means for detecting Separation From our earlier discussions, it can be observed that the unsteady pressure distribution over the surface

somewhat captures the state of the flow and the effect of control on the flow. To further examine this, we plot the unsteady pressure at various locations for baseline flow to see if the pressure distribution can be used to detect separation. This plot of the unsteady pressure distribution at various stations (Transducer Location 1, 2, 3, 4) is shown in Fig. 12. PIV measurements reveal that the flow separates at X/H ~ 1.7. The unsteady pressure measurements show that the fluctuations closer to wall increases as we get closer to the separation location and there is a sudden jump in the pressure spectra at a frequency of around 200 Hz after separation. The overall PRMS values upstream of separation for Transducer 1 (TR1) at X/H=1.1 and Transducer 2 (TR2) at X/H= 1.6 are almost the same at ~ 20 Pa, whereas for TR3 and TR4, it has a value of ~27 Pa and 32 Pa respectively.

These corresponding locations and size of the separation region can also be seen in Fig. 4. The TR3 location is just inside the separation region whereas TR 4 is near the middle of the separation zone, which likely explains the higher PRMS values at TR4. It confirms our notion that PRMS, scaled with certain parameters or even without scaling, can be an indicator for separated flows. Also as discussed earlier, unsteady pressure measurements also reflect the control effort to a great extent and the PIV measurements can be correlated very well to the unsteady pressure measurements. As such, unsteady pressure measurements can be used as “input sensors” and PRMS can be used as an efficiency parameter for separation control using microjets, as in other applications.

5.Concluding Remarks To summarize, our results based on PIV data and unsteady pressure measurements show the dramatic effect of

microjets in controlling flow separation, at least in the present test geometry. Our results indicate that the effect of microjets varies with the amount of momentum supplied via the microjets (which is proportional to microjet pressure) and also with their location and angle at which momentum is injected. These controlling parameters are coupled together, but yet by changing any variable, the separated flow can be attached.

Fig.11- a) Shape factor at X/H=2.3; MJ5; 90° Vs. Microjet Pressure b) Bubble Length (MJ5; 90°) Vs. Microjet Pressure

Bubble Length Vs. Microjet Pressure

0

0.2

0.4

0.6

0.8

1

1.2

1.4

0 1 2 3 4 5 6

Microjet Pressure (in psig)

Non

-Dim

ensi

onal

ized

Bub

ble

Leng

th (L

B/H

)

Experimental Results

Linear (ExperimentalResults)

1

1.5

2

2.5

3

3.5

4

4.5

4 6 8 10 12 14 16 18 20

Microjet Pressure (in psig)

Shap

eFa

ctor

Separation at this Shape Factor for Turbulent Flows

Shape Factor for Flat Plate Turbulent Flow

a) b)

American Institute of Aeronautics and Astronautics

12

The mechanism, as anticipated, is that these microjets serve the dual purpose of energizing the low momentum boundary layer flow and also act as “Fluidic Vortex Generators13” to prevent or delay separation. These vortices promote the mixing of lower momentum boundary layer fluid with outer higher momentum fluid and this enhanced mixing helps to control or delay the separation. The most significant advantage of using these microjets is that they can be turned on or off, as and when desired, thereby eliminating the associated penalty of reduced drag as with the conventional vortex generators. In addition, they are “very low mass-high momentum” devices, which make them all the more attractive from a practical point of view.

The effect of separation and the control efforts on the unsteady pressure is significant. As such they can be used as sensing devices for the separated flow (indicating location and the magnitude of the separation) and correspondingly control the flow through a closed-loop strategy using microjets. If better understood and properly utilized, this behavior can be of significant practical importance for developing and implementing online control.

6. Acknowledgements This study has been supported by grants from the NASA Langley Research Center – grants which initiated this

research; the authors are grateful for this support. This support is also gratefully acknowledged. We also thank Mr. B. DePriest for his considerable assistance in designing and fabricating the hardware and models. We also appreciate the contributions of Dr. B. Alkislar in conducting the PIV experiments.

7. References 1Anabtawi A.J., Blackwelder R.F., Lissaman P.B.S., Leibeck R.H., “An Experimental Investigation of Boundary Layer

Ingestion in a diffusing S-duct with and without Passive Flow Control”, AIAA Paper 99-0739. 2Mayer, D.W., Anderson, B.H., Johnson T.A., “3D subsonic diffuser design and Analysis”, AIAA Paper 98-3418. 3Rabe, D., Boles, A. and Russler, P. “Influence of Inlet Distortion on Transonic Compressor Blade Loading”, AIAA Paper

95-2461. 4Wygnanski, Seifert A., “Delay of Airfoil Stall by Periodic Excitation” AIAA J Aircraft 1996; Vol. 33(4):691–8. 5Gad-el-Hak M, Bushnell D.M., “Separation Control: A Review”, Journal of Fluids Engg., 1991; 113:5-30. 6Amitay M., Pitt D., Glezer A. “Separation control in duct flows”, Journal of Aircraft Vol. 39, No. 4, July – August 2002. 7Zaman K.B.M.Q., Bar Sever A, Mangalam S.M., “Effect of Acoustic Excitation on the flow over a Low-Re Airfoil”,

Journal of Fluid Mechanics, 1987,182:127-48. 8Ahuja K.K., Whipkey R.R., Jones G.S., “Control of Turbulent Boundary Layer by Sound”, AIAA Paper 83-0726, 1983. 9Greenblatt David, Wygnanski, I.J., “The Control of Flow Separation by Periodic Excitation”, Progress in Aerospace

Sciences Vol. 36(2000), p. 487-545. 10Lin, John C. “Review of Research on Low Profile Vortex Generators to control Boundary Layer Separation”, Progress in

Aerospace Sciences Vol. 38(2002), p. 389-420. 11Storms B. L, Ross JC. “Experimental Study of Lift-Enhancing Tabs on a Two-Element Airfoil.” AIAA J Aircraft 1995; Vol.

32: 1072-8 12Jenkins, L.N., Gorton, S.A., Anders S.G., “Flow Control Device Evaluation for an Internal Flow with an Adverse Pressure

Gradient” AIAA Paper 2002-0266. 13Alvi, F. S., Elavarsan R., Shih, C., Garg G., and Krothapalli, A., “Control of Supersonic Impinging Jet Flows using

Microjets,” AIAA Journal, 2003 Vol. 41(7), p. 1347-1355.

Fig.12- Unsteady pressure Spectra; U=40 m/s; No control

American Institute of Aeronautics and Astronautics

13

14Kumar, V., Alvi, F.S. “Use of Supersonic Microjets for active separation control in Diffusers”, AIAA 2003-4160, 33rd AIAA conference, Orlando, June 23-26, 2003.

15Stratford, B.S., “The Prediction of Separation of Turbulent Boundary Layer”, J. Fluid Mech. 5, 1-16, 1959b. 16Lourenco, L.M., “True Resolution PIV: A Mesh Free Second Order accurate Algorithm” International Conference in

application of lasers in Fluid mechanics, Lisbon, Portugal, July 2000 17Lourenco, L.M., Krothapalli A., “Mesh Free Second Order Algorithm for PIV Processing”, Proceedings of the International

conference on Optical technology and Image Processing in Fluid, Thermal and Combustion Flows, Yokohoma, Japan, Dec. 1998, pp. 24.

18Johnston J.P., Nishi M., “Vortex Generator Jets- Means for Flow Separation Control”, AIAA Journal, 1990 Vol. 28, No. 6, pp.989-994.

19Alvi, F.S., “The role of Streamwise Vorticity in the Control of Impinging Jets”, FEDSM2003-45062. 20Marzouk, Y. M., Ghoniem, A.F. “Vorticity formulation for an actuated jet in crossflow”, AIAA Paper 2004-0096. 21Gorton S.A., Owens L.R., Jenkins L.N., Allan B.G, Schuster E.P., “Active Flow Control on a Boundary-Layer-Ingesting

Inlet”, 42nd AIAA Aerospace Sciences Meeting and Exhibit, Reno, Nevada, AIAA Paper 2004-1203.