High Bandwidth Micro-Actuators for Active Flow Control

John T Solomon1, Rajan Kumar2 and Farrukh S. Alvi3

Florida A & M University and Florida State University, Tallahassee, FL 32310 Abstract

This paper describes an experimental study conducted at the Advanced Aero propulsion Lab (AAPL) for the design and development of actuator systems capable of producing high bandwidth, high momentum microjet arrays for active flow control applications. A systematic approach for designing micro-actuators with high unsteady and mean momentum efflux is followed. Beginning with a simple configuration, i.e., supersonic impinging microjets, we added more geometric complexity to the actuator design to finally arrive at an actuator configuration that provides the desired flow properties. Our first generation actuator design consists of a primary source jet, incident upon a cylindrical cavity. The lower surface of this cavity contains micronozzles through which the unsteady microjets (400µm) issue. Results clearly show that microjets produced by this actuator contain very high mean momentum (300-400 m/s) as well as a very significant unsteady component (70-100 m/s). Experiments were conducted over a large range of parameters, in terms of cavity length, source jet NPR and source jet impingement distance. The results unequivocally demonstrate the ability to vary the frequency as well as the amplitude of the mean and unsteady momentum of the microjets issuing from this actuator. By varying the dimensions of the actuator by only few hundred microns, we were able to tune the frequency of the unsteady component over intervals of 10-15 kHz. The ability to produce, unsteady flow with significant mean and unsteady components, where the dynamic range can be easily varied makes these actuators promising for a number of flow control applications.

I Introduction

ctive Control of flows for a wide array of applications has seen a surge of activity in recent years due to the potentially substantial gains in performance offered by flow control schemes. For example, control or delay of flow separation over airfoils and lifting bodies can significantly extend the operating

envelope of aircraft by improving their aerodynamic performance. The control of aeroacoustically induced flow oscillations in cavity flow is another area where various active (and passive) control methods are being explored1. Another ideal candidate for the use of flow control is the flowfield generated by supersonic impinging jets2-4. Such flows are ubiquitous in many applications, such as in STOVL aircraft during hover, and result in a highly unsteady flowfield with very high acoustic levels and dynamic pressure loads in the near-field. These examples demonstrate that flows where control can be applied are wide and varied with more applications are likely to appear as the technology matures.

Efficient control of flows requires the use of effective actuators, which can be adapted for specific applications. Among the most common actuators used are piezoelectric material based actuators, which can be fabricated in various shapes, such as flaps and wedges. Such actuators have been used for the control of cavity flows (Cattafesta et al.5) and shear flows (Wiltse & Glezer6). Actuators based on synthetic jets have also been used for separation control over airfoils and cylinders (Amitay et al.7). Although relatively successful at low speeds, most actuators are not very efficient when the primary flow velocities are high. This is primarily because the magnitude of the control input - be it tip deflection of the piezoelectric flap actuators or the momentum of the synthetic jet - becomes small relative to the inertia of the mean flow. The potential of Hartmann tube or modified Hartmann resonators for active flow control has also been explored by a few researchers8-10 with limited success in part because their response is in a limited frequency range. Consequently, there is a need for actuators, which produce high-amplitude disturbances over a large frequency range that can be used for the control of high-speed flows.

1 Graduate Research Assistant, Department of Mechanical Engineering, Student Member AIAA 2 Research Scientist, Department of Mechanical Engineering, Member AIAA 3 Professor, Department of Mechanical Engineering, AIAA Associate Fellow

A

14th AIAA/CEAS Aeroacoustics Conference (29th AIAA Aeroacoustics Conference)5 - 7 May 2008, Vancouver, British Columbia Canada

AIAA 2008-3042

Copyright © 2008 by Authors. Published by the American Institute of Aeronautics and Astronautics, Inc., with permission.

2

Supersonic impinging jet flows have been the subject of extensive study at our laboratory over the past few years. In order to reduce the feedback-driven unsteadiness in the impinging jet flowfield, an array of 400 µm supersonic steady microjets, were mounted circumferentially around the main jet. This control technique proved to be very effective, where the overall noise levels were reduced by up to 12 dB and 15dB for cold and hot impinging jets, respectively. The reason for the success of this technique is likely due to the fact that the high momentum associated with the supersonic microjets allows them to penetrate the primary jet shear layer sufficiently to disrupt the feedback loop (Alvi et al.2, Lou et al.3, and Kumar et al.4). Furthermore, the small diameter of the supersonic microjets drastically reduces the mass flow requirement while producing a very high momentum jets. Motivated by the success of steady supersonic microjets for flow control and the lessons learned from literature, we initiated a program to develop pulsed microjet actuators that produce unsteady microjets with high mean and unsteady momentum and whose properties can be varied over a large range of frequencies– i.e. a large bandwidth. As an example of the potential application of such unsteady microjets, we show the pressure spectra from a supersonic impinging jet in Fig. 1a (Kumar et al.4), and supersonic cavity flow in Fig. 1b (Zhuang et al.1). Both these spectra show the presence of strong, discrete tones approximately in the range of 4 -10 kHz. Hence, it is anticipated that actuators with a strong unsteady component in this frequency range could potentially be very useful in further enhancing the control efficacy of microjet-based control of such flows. Here we describe an experimental program, where high-speed microjets are combined with a cavity and other geometric features to generate highly unsteady, pulsed flow. Such a flow could be utilized to produce high momentum, unsteady jets, which can be used as fluidic actuators. Furthermore, the unsteady component (frequencies) of such actuators could be tuned for specific applications.

II Experimental Apparatus and Procedures

The experiments were conducted on an optical table at the Advanced Aero Propulsion Laboratory (AAPL) at the Florida State University. Measurements were made for the jet operating over a range of Nozzle Pressure Ratios (NPR) corresponding to moderately to strongly under-expanded jets and over a range of geometrical parameters. Measurements include flow visualization using a micro-schlieren system and unsteady pressure measurements using a miniature Kulite pressure transducer, the details of which are given below.

A. Micro-Schlieren System The length scales associated with the micro-actuator flowfield being examined in the present study require a spatial resolution of the order of tens of microns (~10µm), which is too small to be resolved by conventional optics. Hence, a specialized high magnification, high sensitivity schlieren system, capable of visualizing flows at micro-scales designed at AAPL was used to visualize the flow field associated with micro-actuators. The main parameters, which were controlled for optimal visualization, were: magnification, resolution, field of view and sensitivity. The micro-schlieren system utilizes a high-magnification, in-line achromatic lens-based optics, coupled with a graded filter (to minimize diffraction effects) and a Kodak Megaplus 1.4 camera with

a) Supersonic impinging jet4 b) Supersonic cavity flow1

Figure 1. Potential applications of unsteady microjets (the lower amplitude spectra in each plot reflects the effect of steady microjet control).

3

a CCD array resolution of 1008 x 1018 pixels. A white light stroboscopic lamp with adjustable frequency (up to 1 kHz) and intensity was used as the light source. The resultant magnification obtained from the lens-based system was as high as 4 – 5. More details of the optics used in the micro-schlieren system are available in Phalnikar et al.11.

B. Unsteady Pressure Measurements The unsteady flow coming out of these microjets was measured using a probe (diameter = 1.5 mm) with a 1.3 mm diameter, 0-100psia range Kulite pressure transducer. The unsteady pressure signals were acquired through high speed National Instruments digital data acquisition cards using LabviewTM. The transducer output was conditioned using a low-pass StanfordTM filter (cut-off frequency = 60 kHz) and sampled at 200 kHz. Standard FFT analysis was used to obtain narrowband pressure spectra. A total of 100 FFT’s of 4096 samples each were averaged in order to obtain statistically reliable narrowband spectra.

III. Building-block experiments for micro-actuator development In this section, we briefly discuss some initial experiments that were conducted using simple, building-block or canonical configurations. The aim of these studies was to gain some insight and guidance towards the design of the ‘final’ unsteady microjet actuators, which is discussed in section IV. A. Supersonic impinging microjets Large scale supersonic impinging jets are known to generate a highly unsteady flow field associated with sharp discrete tones of high intensity2-4. As a first step towards our primary goal - the micro actuator development, we have examined the characteristics of supersonic impinging microjets. Figure 2 shows the pressure spectra of impinging microjets at a nozzle pressure ratio, NPR = 5.8, over a range of nozzle to plate (h/d) distances. The spectra shows strong impinging tones in the frequency range of 25-55 kHz over the range of h/d tested. These results confirm that the resonance loop observed for larger impinging jets12-13 is present here and can thus be leveraged for our actuator development. However, due to the small physical scales the frequency range of these tones is much higher than those required for the development of micro-actuator for our applications (see Fig. 1). Hence, a modification to the impinging jet configuration is needed to produce unsteady microjets in the desired frequency range. B. Impinging microjets with hole tones The unsteady flowfield characteristics of edge/hole tones and the effect of various geometric parameters have been studied in detail by Powell12. To leverage the understanding gained from prior research by a number of investigators and our studies on impinging microjets, we carried out experiments on impinging microjets on a flat plate with a central hole. The experiments were designed such that the shear layer of microjets grazes the edges of the central hole and generates large amplitude tones, often referred to as hole tones (Note that the feedback mechanisms governing impinging jets, screeching jets and hole tones are very similar, Krothapalli, et al.13). Using this configuration, hole tones were produced that were lower in amplitude than those in Fig. 2, however equally significantly, the tones were also in a lower frequency range, around 3 kHz.

Impingement plate

Kulite Probe

h

d = 1mm

Figure 2. Pressure spectra of impinging microjets.

4

C. Microjets in a cavity Motivated by the well known Hartmann tube14 we next explored the unsteady flowfield of microjets impinging in a cylindrical cavity. The results corresponding to a 1mm diameter jet normally incident upon a cylindrical cavity of length L = 4.7mm and diameter D= 1.4mm are presented here. Figure 3 shows a schematic of this arrangement as well as the pressure spectra produced for microjets impinging into this cavity over a range of NPRs (1.9 – 6.3). As seen, here the spectra show distinct tones in the frequency range of 6-11 kHz for this configuration. These results were very encouraging both in terms of the tonal frequency range and the associated high amplitudes since in many applications actuators with a large unsteady momentum flux and the ability to tune their frequency is desired for different flow control applications.

IV Actuator Design: Results and Discussion A. Design details The results described in the previous section were very encouraging. They provided ample evidence of the potential to produce unsteady, high momentum microjets by expanding upon the simple building block configurations shown in Figs. 2 and 3. Based on these results a ‘first generation’ microactuator was designed, a schematic of which is shown in the Fig. 4. As seen here, the micro-actuator consists of three main components: a) A larger, primary source jet, which supplies the air to a cylindrical cavity, b) a cylindrical cavity upon which the source jet impinges, and c) multiple micronozzles (i.e microjet orifices) at the bottom of the cylindrical cavity, from which the high-momentum, unsteady microjets issue. In the present design, the source jet was issued from a 1.0mm diameter converging nozzle and the micronozzles array at the bottom of the cavity consists of four 400 µm holes in the pattern shown in Fig. 4. The unsteady microjet flow was visualized using the high magnification micro-schlieren system discussed previously (see section IIA) The flow properties of the microjets were measured using a Kulite total pressure probe, fabricated using a Kulite XCE-062-100A (0-100psia range) transducer. Admittedly, given the very small size of the microjets (~0.016”) some spatial averaging will occur. However, as discussed later, measurements made with the probe tangential to, i.e. ‘grazing’ the microjet plume (see Fig. 11), confirm that the unsteady properties discussed in the following represent the unsteady microjet behavior. The location of the Kulite probe relative to the actuator is shown in Fig. 4. B. Microjets Array Flow Properties The main parameters that govern the properties of the microjet array issuing from the actuator assembly are: a) the distance from the source jet h, b) the length of the cylindrical cavity, L and c) the source jet pressure ratio,

D

L

d

Kulite Pressure Probe (dia. = 1.5 mm)

h

Four 400 µm holes

0.5 mm

Figure 4. Schematic of Micro-actuator

D

L

d = 1.0 mm

h

Figure 3. Jet incident on a cavity (left); pressure spectra (right) for this configuration

5

NPR. The two geometric parameters are indicated in Fig. 4. Experiments were conducted over a wide range in terms of these geometric and flow parameters, where h/d (where d=1 mm) was varied from 1 to 2, L/d from 1 to 5 and NPR from 1.9 to 6.5. The aim was to examine and understand the effect of these parameters on the flow issuing from the microjet actuator array and to identify the optimal range and combination of these parameters that produce the desired micractuator flow. As a result, we hope to develop a preliminary design approach and scaling laws for such actuators. In the following we discuss the micro-actuator flow properties as these parameters are varied. a) L/d= 1 Figure 5 shows representative schlieren images of the flowfield associated with micro-actuator at NPR = 4.8 for two values of h/d. At h/d = 1.3, large oscillations of the Mach disc (see Fig. 5) and in the cavity flow were observed, which led to a strong tone in the pressure spectra presented next. There were no visually observable oscillations in the flow at h/d = 1.6, a property confirmed by the corresponding pressure spectra which was devoid of any discrete tones. The presence of shock cells in the microjets issuing from the bottom of the actuator clearly confirms that the flow is supersonic. In Fig. 6, we see the pressure spectra of the flow issuing from the microjet actuators for cavity length L/d=1. Fig. 6a, shows the effect of varying h/d while NPR is kept constant (NPR=4.8) and Fig. 6b shows the effect of varying the NPR for a fixed h/d = 1.4. The spectra in both figures clearly show the presence of high amplitude peaks indicating the presence of highly unsteady flow issuing from the actuators. Equally noteworthy is the trend of peak frequency variation seen in Fig.6a, where a modest variation of h/d (from 1.3 to 2), leads to a significant shift in the peak frequency. For example, at h/d =1.3, a spectral peak with an amplitude of ~ 157dB occurs at a frequency of approx. 58 kHz, whereas at h/d =1.8, the peak has shifted to a lower frequency of ~42 kHz and with an amplitude of roughly 141dB. Furthermore, the spectral peaks systematically become broader with increasing h/d and beyond h/d = 1.8, there is no measurable discrete peak. The presence of narrow spectral tones is expected as this actuator design leverages a number of flow-acoustic resonance phenomena to enhance the flow unsteadiness; some of these have already been discussed in section II. Similarly, the shift to lower frequencies with increasing spatial lengths or distances is also expected. Likewise, if one is utilizing impinging jet resonance, then the feedback loop governing such phenomenon is expected to become weaker beyond some critical distance due to the decay in the strength of the flow instabilities. For the present case, this critical distance appears to be h/d=2.

The results in Fig.6a are very encouraging as they clearly show the capability of this actuator to produce highly unsteady flow. Equally importantly, we see the presence of ‘control knobs’, namely NPR and h/d, that allows one to vary the nature – frequency and amplitude – of the unsteady flow from these actuators.

a) Variation in h/d b) Variation in NPR Figure 6. Pressure spectra of micro-actuator for L/d = 1.

Figure 5. Schlieren images of the

flowfield; NPR = 4.8

6

Keeping in mind that the two applications where we first plan to implement and test such actuators, i.e. the supersonic impinging jet and cavity flowfield - require frequencies in a range of 6-10 kHz (see Fig. 1), we next examine ways of reducing the unsteady frequencies, to more practical ranges. This is accomplished by increasing the cavity length, L/d, discussed next. b) L/d= 2 and L/d=3 Figures 7a and 7b show pressure spectra corresponding to those seen in Fig. 6a and 6b for a longer cavity, L/d=2. The trends observed in Fig. 6a are also seen here for this actuator with the longer cylindrical cavity. As seen in Fig. 7a, a variation in h/d from 1.0 to 1.5, shifts the peak frequency from 24 kHz to 36 kHz. Similarly, increasing the NPR from 4.4 to 5.5 (Fig. 7b), shifts the frequency from 26 kHz to 34 kHz while increasing the peak amplitude by roughly 15 dB. By using a longer cavity in the actuator (from L/d = 1 to L/d = 2), we have shifted the actuator output frequency range by 20 kHz. The trend continues for the actuator with L/d =3, the results of which are shown in Fig. 8. The same dependence on h/d and NPR is observed as in the previous two cases and the usable frequency range is now between 14-24 kHz.

a) Variation in h/d b) Variation in NPR

Figure 7. Pressure spectra of micro-actuator for L/d = 2.

a) Variation in h/d b) Variation in NPR

Figure 8. Pressure spectra of micro-actuator for L/d = 3.

7

c) Designing an actuator for our applications The effect of varying the three parameters on the actuator output is summarized in Fig. 9, which shows the variation in the peak frequency as a function of cavity length, L/d, for a range of h/d and NPR tested. As seen here, for a given actuator design, i.e. fixed L/d, very small changes in the source jet distance and operating pressure allows one to sweep the output frequencies over a rather large range of ∆factuator = 10-20 kHz. An extrapolation of the trend observed in this plot also provides guidance regarding the size of cavity needed to produce unsteady microjet flow in the desired range of 6-10 kHz. As seen in Fig. 9, this corresponds to a cavity with L/D ~ 5. Based on these results we designed an actuator with L/d = 5, the results of which are shown in Fig. 10. As predicted by the trend in Fig. 9, the peak frequencies for this actuator are in the range of 6-11 kHz, as needed for our control applications.

In summary, the results presented thus far show that by optimally leveraging multiple resonance phenomena, one can design a very compact actuator. In addition, as seen here, very small changes in the actuator dimensions allow us to control the frequency output as well as amplitude over a wide range. Fig. 11 compares frequency response from the actuator measured with a probe placed parallel to the actuated jet to that where the jet normally impinges on the probe. This test was conducted in part to verify that the fluctuations measured by the Kulite probe were not in fact due to resonance induced by impingement on the probe itself. As seen here, peaks are observed at the same frequency components in both cases. However, as expected, the amplitudes are lower for the tangential probe. These results substantiate the fact that the unsteady microjet flow measured by the Kulite is due to the aero-acoustic coupling within the actuator which

Figure 9. Summary of Micro-actuator output ( design criteria)

a) Variation in h/d b) Variation in NPR

Figure 10. Pressure spectra of micro-actuator for L/d = 5.

Figure 11. Effect of measurement direction on

pressure spectra

Desired frequencies

8

significantly enhances the unsteady component of the supersonic flow issuing from these actuators. Microjet actuator arrays designed using the above approach can potentially be used for a number of supersonic (and subsonic) flow control applications which require high unsteady and mean components., such as those produced herein.

C. Flow field visualization In addition to the spectra discussed earlier, micro-schlieren visualizations of the actuator flowfield provide ample evidence of the highly unsteady nature of the flow, both into (i.e primary jet into the cavity) and out of the actuator. Fig. 12 shows a series of images corresponding to various phases of this unsteady flowfield. An examination of these images clearly illustrates the coupling between the unsteadiness in the primary/source jet flowfield – seen by the change in the Mach-disk structure, and that in the microjet flow issuing from the bottom of the actuator. A change in the primary jet flow leads to a change in the strength of the microjet arrays. This can be seen by comparing the shock cells seen in these images, which is seen to vary considerably. A comparison of the shock cell structure to more detailed visualizations of microjets of the same scale (Phalnikar et al.11) allows one to roughly estimate the fully expanded (micro)jet Mach number. Based on such a comparison, the microjet flow varies from high subsonic (no shock cells, case ‘e’) to a Mach number ~1.5 (for the cases ‘a’ and ‘g’) as shown in Fig. 12. This corresponds to a very large and periodic variation in the strength of the actuator output, a property that is highly desirable. As mentioned earlier, the aeroacoustic properties of this actuator appear to be a combination of a number of resonance phenomena. In some ways, the flow behavior is similar to the Hartmann tube in which the resonator, the cavity, is alternately filled with high pressure air and then discharged. However, there are some differences, in part due to the fact that the Hartmann cavity is closed at one end whereas, by design, we allow flow to exit at the lower end, producing the unsteady microjets. A comparison of the frequencies produced by the present design to the quarter-wave correlation proposed by Hartmann14 shows poor agreement between these two. Finally, in Fig. 13, we show the effect of varying the primary jet distance on the amplitude and the frequency of the unsteady micro-actuator flow. As seen here, and as alluded to in our earlier discussion, there is a narrow range of h/d over which the frequency as well as the amplitude can be systematically changed. For the L/d=3 case shown in Fig. 13, this ‘region of instability’ approximately lies between h/d =1 to 2, where the amplitude first increases, peaks and then systematically decreases.

a b e g

Figure 12. Various phases of fluid dynamic oscillations

Figure 13. Frequency and intensity variation for L/d=3

9

Conclusions

In this paper, we use a systematic approach for designing micro-actuators with high unsteady and mean momentum efflux. Beginning with a simple configuration, i.e., supersonic impinging microjets, we added more geometric complexity to the actuator design to finally arrive at an actuator configuration that provides the desired flow properties. Our first generation actuator design consists of a primary source jet, incident upon a cylindrical cavity. The lower surface of this cavity contains micronozzles through which the unsteady microjets issue. Our results clearly show that microjets produced by this actuator posses very high mean momentum (they are supersonic for most cases, hence >300 m/s) as well as a very significant unsteady component (70-100 m/s). Using this design, experiments were conducted over a large range of parameters, in terms of cavity length, source jet NPR and source jet impingement distance. The results unequivocally demonstrate our ability to vary the frequency as well as the amplitude of the mean and unsteady momentum of the microjets issuing from this actuator. By varying the dimensions of the actuator by only few hundred microns, we were able to tune the frequency of the unsteady component over intervals of 10-15 kHz. The ability to produce, unsteady flow with significant mean and unsteady components, where the dynamic range can be easily varied makes these actuators promising for a number of flow control applications. In the near future, our goal is to use these actuators for controlling supersonic cavity flows, a problem that is ideally suited for such a control scheme.

References

1Zhuang N, Alvi FS., Alkislar MB, and Shih C. “Supersonic cavity flows and their control” AIAA Journal, Vol. 44, No. 9, 2006, pp. 2118-2128.

2Alvi, F. S., Shih, C., Elavarasan, R., Garg, G. and Krothapalli, A., “Control of supersonic impinging jet flows using supersonic microjets,” AIAA Journal, Vol. 41, No. 7, 2003, pp.1347-1355.

3Lou, H., Alvi, F. S. and Shih, C., “Active and adaptive control of supersonic impinging jets,” AIAA Journal, Vol. 44, No. 1, 2006, pp.58-66.

4Kumar, R., Lazic, S., and Alvi, F. S., “Active control of high temperature supersonic impinging jets,” AIAA-2008-360

5Cattafesta III, L. N., Garg, S., Choudhari, M. and Li, F. “Active Control of Flow-Induced Cavity Resonance,” AIAA Paper 97-1804.

6Wiltse, J. M. and Glezer, A., “Manipulation of Free Shear Flows Using Piezoelectric Actuators,” J. Fluid Mech., Vol. 249, 1993, pp. 261-285.

7Amitay, M., Kibens, V., Parekh, D.E., and Glezer, A., “Flow reattachment dynamics over a thick airfoil controlled by synthetic jet actuators,” AIAA Paper 99-1001.

8Raman, G. and Kibens, V., “Active flow control using integrated powered resonance tube actuators”. AIAA Paper 2001-31330.

9Dziuba, M. and Rossmann, T., “Active control of a sonic transverse jet in a supersonic cross flow using a powered resonance tube”, AIAA Paper 2005-897.

10Kastner, J and Samimy, M. “Development and characterization of Hartmann tube fluidic actuators for high-speed flow control”, AIAA Journal, Vol. 40, No.10, 2002, pp.1926-1934.

11Phalnikar, K. A., Kumar, R., and Alvi, F. S., “Experiments on free and impinging supersonic microjets”, Experiments in Fluids, 2007.

12Powell, A. “The sound producing oscillations of round underexpanded jets impinging on normal plates,” Journal of Acoustic Society of America, Vol. 83, No.2, 1988.

13Krothapalli, A., Rajkuperan, E., Alvi, F. S. and Lourenco, L., “Flow field and noise characteristics of a supersonic impinging jet,” Journal of Fluid Mechanics, Vol. 392, 1999, pp. 155-181.

14Hartmann, J and Trolle, B. “ A new acoustic generator”, Journal of Scientific Instruments, Vol. 4, No. 4, 1927, pp. 101-111.