2 Experimental setup and measurement techniques

Size: px

Start display at page:

Download "2 Experimental setup and measurement techniques"

Vivian Kelly
5 years ago
Views:

1 TOPICAL PROBLEMS OF FLUID MECHANICS 5 DOI: GENERATION OF INTERMITTENT JET BY PULSE-MODULATED SYNTHETIC JETS Z. Broučková Department of Thermodynamics, Institute of Thermomechanics AS CR, v.v.i., Dolejskova 4/5, Prague 8, Czech Republic, 8 Abstract An axisymmetric air jet was excited using a pair of counter synthetic jets running in opposed direction. First, the control synthetic jets were measured alone. The driving signal was pulse modulated sinusoidal signal. After an adjustment, the primary axisymmetric jet was excited to the bifurcating mode, and its behavior was studied experimentally. For comparison purposes, a reference steady (unforced) jet from the same nozzle was also measured (Reynolds number was approximately 3). The flow visualization and hot-wire anemometry techniques were used. Keywords: pulse modulation, synthetic jet, flow control Introduction An active flow control is an important tool in many fluid mechanics studies and applications. Many sources of an additional control energy can be applied, e.g. acoustic excitation [], pulsating secondary jet [], fluidic flip-flop control [3], excitation using plasma actuators [4] to name but a few. Some of the mentioned methods is suitable rather for the high-speed applications. On the other hand, when the low speed jets are considered, it has been shown that synthetic jets can also be a useful variant of a control jet due to their relative simplicity, relatively low energetic demands, possibility of miniaturization and (in general) wide range of working parameters (frequencies, amplitudes) within one actuator. Synthetic jet (SJ) is a fluid flow generated by an periodical oscillation diaphragm (or piston etc.) in a cavity which is connected with its surroundings by some type of exit orifice. The time-mean mass flux of the oscillatory flow in the orifice is zero but out of the orifice the train of the vortices creates a non-zero flow (see e.g. [5]). SJs are usually used in application with heat transfer (e.g. [6,7,8]), control of flows, wakes and boundary layers (e.g. [9,,]). In order to improve the SJ properties, the modifications of the basic SJ input signals have also been proposed. Zhang and Wang [] used modified sinusoidal signal where the positive (extrusion) part was shortened but its amplitude was increased and on a contrary, the negative (suction) part was prolonged but with smaller amplitude. This approached resulted in stronger extruded vortex pair. Qayoum et al. [3] used amplitude and frequency modulation to introduce a new additional frequency into the main (controlled) jet. Amitay and Glezer [4] used a pulse modulated signal as an input signal of a SJ actuator. The resulting jet was used as a control jet for in a study of an influence of a extremely brief actuation on separated flow over a D airfoil. The present study deals with an round air jet which is actively controlled by a pair of SJs running in opposed direction. The input signal of the SJ actuators is a pulse modulated sinusoidal signal. The present work is motivated by an earlier research [5] which dealt with the axisymmetric jet excited by a synthetic jet system. The control synthetic jets were driven by a standard sinusoidal signal. The main jet was excited to the helical or bifurcating mode depending on the chosen combination of the synthetic jets and their electrical connection. Both the original experimental device and the present modified device aim to be the models of possible future miniature device intended for cooling of e.g. highly loaded electronic components in microchannels. The present device uses loudspeakers as a part of SJ actuators, in case of miniature device the usage of micro piezoelectric diaphragms seems to be more promising alternative. Experimental setup and measurement techniques The scheme of the present experimental setup including the coordinate system is shown in Fig.. The working fluid was air. The air flow (position in Fig. ) is supplied by a compressor. The air passes through the grid (), settling chamber (3a), honeycomb (4) and the second settling chamber (5) to the quadrant-contoured nozzle generating the main jet. The exit diameter and the area contraction ratio are

2 6 Prague, February -3, 8 y r 6a L 5 3b 4 R 6b 3a Figure : Scheme of the tested nozzle; : primary air flow supply, : grid, 3a, 3b: settling chambers, 4: honeycomb, 5: contoured contraction, 6: SJ actuators. D = mm and ε =., respectively. The exit channel of the nozzle is a short cylindrical pipe (of 3 mm length) with a constant cross-section. It is equipped with two counter SJ actuators (6a, 6b) denoted as L (Left) and R (Right) in Fig.. Both SJ actuators are driven elecrodynamically by a diaphragm (originated from loudspeaker Monacor SP-7/4S). Both actuators generates SJ issuing from rectangular slot. 6 mm in perpendicular direction to the primary jet axis. The cavities of the actuators are connected with their exit slots by a pipe of inner diameter and length 9. mm and l = 63 mm, respectively. The actuators are electrically series-connected. They are running in opposed direction. The driving signal is pulse modulated sinusoidal signal with frequency f = Hz. Only one pulse is generated at the beginning of every working period with the modulation frequency f MOD = 6 Hz. The left SJ actuator begins its working cycle with extrusion, the right SJ actuator works in opposed direction. The SJ actuators are fed from the sweep/function generator (AGILENT 33A) which is amplified by the audio amplifier (PIONEER A-9R). The actuator input peak-to-peak voltage is kept at 9.6 V. Two experimental methods were applied in this study: qualitative flow visualization and quantitative measurement using hot-wire anemometry technique. The flow visualization was performed using water fog which was added into the main jet. The visualization was done in phase-locked regime under the stroboscopic light. The light was synchronized with the driving signal and it was shifted to the chosen phase angle. In this particular case the two-channel function generator RIGOL DG46 was used. The pictures were taken using digital camera connected with PC via an USB. The exposure time was s, so the resulting pictures are the multi-exposures of working cycles. A hot-wire anemometer (MiniCTA 55T3, Dantec Dynamics) with the single-wire probe (55P6) was used for velocity measurement in the constant temperature (CTA) mode. Typical sampling frequency and the number of samples were khz and 6,384, respectively. For the present experiments the anemometer was calibrated in the velocity range approximately from.5 m/s to.5 m/s. 3 Results 3. Flow visualization Fig. shows a visualization of the main unforced jet at Re = U m D/ν =3, where U m =. m/s is the mean velocity at nozzle exit and ν is the kinematic viscosity. The picture was taken under the continual light. The jet flow is smooth, rather narrow and slowly spreading as the surrounding air is entrained into

3 TOPICAL PROBLEMS OF FLUID MECHANICS 7 Figure : Visualization of the main unforced jet at Re = 3. the jet. The initial laminar character is visible, further downstream the jet apparently undergoes the transition to turbulence. Fig. 3 shows the phase-locked visualization of the forced jet at several subsequent phases of the working cycle. At the beginning of the cycle the main flow is sharply deflected to the right by the action of the left actuator. At ϕ =.5 (t/t =.65) the extrusion part of the left actuator ends and the extrusion part of the right actuator begins. The jet emanating from the nozzle is visibly deflected to the left (see ϕ = 4 ). The resulting bifurcated structure created by the to opposed parts of the cycle moves further downstream causing considerable spreading of the jet in the nearfield. The extrusion of the right actuator ends at ϕ = 45 (t/t =.5). In the next part of the working cycle the jet smoothes and gradually becomes symmetrical again till the new cycle begins (not shown in Fig. ). 3. Velocity field As the initial step the investigation of the SJs without the main flow was performed. Fig. 4 shows the instantaneous velocity magnitude of the synthetic jets during three subsequent working cycles without the main continuous jet. The hot-wire probe was inserted into the nozzle, on its axis, approximately.5 mm upstream (i.e. x =, y = -.5 mm), in this particular case. Together with the velocity the input voltage signal of the actuators is also shown in Fig. 4. The pulse of the input signal causes double peak in velocity signal with maximum values around m/s. The first peak is caused by the extrusion of the left actuator, the second peak is caused by the extrusion of the right one. The probe cannot catch the suction part of any actuator because its position is out of the reach of the suction. The considerable lower third peak is visible in the instantaneous velocity waveform. This peak is probably caused by reflection of the jet flow from the walls of the nozzle exit channel. The phase-average velocity magnitude during one cycle is shown in Fig. 5a and the corresponding phaseaverage root-mean-square (RMS) velocity is shown in Fig. 5b. The probe was place on the axis (i.e. x =, y = -.5 mm), the main jet was turn off. Three cases were investigated: only the left actuator was turned on, only the right actuator was turned on and both actuators were turned on (i.e. the same case as in previous Fig. 4). The velocity magnitude is as expected: during the first half of the pulse the fluid is extruded from the cavity of the left actuator and during the second half of the cycle the fluid is extruded from the opposite actuator. Both velocity peaks are commensurate, the duration of the second peak is slightly prolonged, probably due to disruption caused by the previous extrusion from the left actuator. When both actuators are turned on the resulting velocity is an envelope of both single velocity signals. The graph of the RMS velocity shows more surprising results: the second peak is approximately 6-times higher than the first one. This behavior is connected with the above mentioned disruption of the fluid during the first half of the pulse. The next investigation is aimed at the control of the main continuous jet. The evolution of the axis velocity (x = ) of the main controlled jet in the nearfield is shown in Fig. 6. The influence is the most distinct in the nozzle exit (y = ). The maximum velocity (u = 6. m/s) is reached at t/t =.6 (ϕ ),

4 8 Prague, February -3, Figure 3: Phase-locked visualization of the forced main jet at Re = 3.

5 TOPICAL PROBLEMS OF FLUID MECHANICS 9 Figure 4: Example of the instantaneous velocity magnitude of the control synthetic jets and the input signal. i.e. shortly before the beginning of the second half of the pulse. Then the velocity drops to u = 3 m/s at t/t =. (ϕ = 36 ) and subsequently rises again u =. m/s at t/t =.4 (ϕ 5 ), i.e. shortly after the end of the input voltage pulse. At the rest part of the cycle the velocity at the nozzle exit drops again and remains almost unchanged around u =.5 m/s. The similar trend can be observed also further downstream. The range of the reached velocities is obviously smaller than is the range in the nozzle exit. With the increasing distance the amplitude of the velocity decreases, at the farfield (not shown in Fig. 6) the velocity almost does not oscillate; cf. the visualization in Fig. 3. Figure 7 shows the phase-average velocity waveforms during one working cycle in the selected points on the jet axis. Fig. 7a (y/d = ) confirms the observation from Fig. 6. Two distinct velocity peaks as the consequence of two control jets are visible at the beginning of the cycle, further on (from approximately t/t =., ϕ 79 ) the jet is calm and the velocity is kept approximately constant at u =.5 m/s. The significant influence of the control jets is still visible at y/d =,.5 and 5. The velocity changes considerably during the cycle as the information about the control action reach the particular position. At y/d = the main jet becomes calm and steady again (cf. visualization in Fig. 3). The velocity profiles very near the nozzle exit (y =.5 mm) are shown in Fig. 8. At the beginning of the working cycles the velocity profiles are smooth and flat. The influence of the control jets is then apparent; the velocity on the right side rises considerably during the extrusion stroke of the left actuator (t/t =.6, ϕ ) and then decreases slowly. During the extrusion stroke of the right actuator, the velocity peak moves to the left part of the profile. It can be seen in Fig. 8 that the velocity peak on the left hand side is lower than that on the right. The jet remains slightly shifted to the right hand side. It is caused by the fact that the fluid is disrupted after the first half of the control pulse and more energy would be needed to overcome the tendency of the jet to be oriented to the right, cf. the previous Fig. 5a and 5b and the corresponding commentary. Just note here, that the maximum velocity on the axis is lower and occurs later during the cycle (t/t =., ϕ 44 ) than the maximum velocity on the axis in Fig. 6; the reason is the shift of the probe position more downstream. After the subsiding of the control effect, the velocity profiles become flat and smooth again till the new cycle begins. The shown hot-wire anemometry results during the cycle are in agreement with the visualization in Fig. 3; the jet is intermittent in character, it is distinctly bifurcated by the control jets, especially in the nearfield, the local velocity rises considerably as the control jets are operated. The jet oscillates form right to left. After the influence of control jets fades away the main jet return back to the initial state, i.e. rather smooth continuous jet. This behaviour is repeated with every working cycle. The time-mean evolution of the velocity along the jet axis is shown in Fig. 9a and the corresponding timemean RMS velocity is shown in Fig. 9b. The controlled main jet is compared with the unforced jet. The

6 Prague, February -3, 8 Iu I (m/s) 5 Actuator L Actuator R Actuators L and R (a) t /T u RMS (m/s) Actuator L Actuator R Actuators L and R (b) t /T Figure 5: The phase-average velocity magnitude of the SJs during one cycle (a) and the corresponding phase-average root-mean-square (RMS) velocity (b) t/t y (mm) Figure 6: Axis velocity (x = ) of the main controlled jet in the nearfield.

7 TOPICAL PROBLEMS OF FLUID MECHANICS t /T t /T 3 (b) y/d = y/d =.5 (d) t /T y/d = (a) (c) (e) t /T.5 y/d = 5 y/d = t /T Figure 7: Phase-average velocity waveforms during one working cycle in the selected points on the jet axis. behavior of the unforced jet is typical for the jets issuing from the round nozzles; the velocity is constant (U =.5 m/s) till the end of the potential core (around y = 7 mm, i.e. y/d = 7), then gradually decreases with the slope U ~ y -.. The exponent is somewhat higher than the usually stated U ~ y - (see e.g. Blevins [6]). The reason is probably a very short exit channel of the nozzle. The RMS velocity also behaves as can be expected; it is approximately constant and as it is reaching the end of the potential core it gradually increases, reaches its maximum shortly after the end of the potential core and then drops. The control effect causes the changes both in velocity and RMS velocity. The time-mean velocity of the forced jet in the nozzle exit is higher than that of the unforced one (U = 3.4 m/s). After the emanating from the nozzle exit the velocity slowly decreases. No velocity plateau as in the unforced case is present here. The slow decrease turns to sharper one around y = mm (y/d = ). However, the slope of the velocity decrease is not so steep as in the unforced case, U ~ y -.6. The RMS velocity in the nozzle exit is 38 times higher than the one of the unforced jet, than slowly decreases to the values of the unforced jet. Figure presents the time-mean dimensionless velocity profiles (Fig. a) and time-mean dimensionless RMS velocity profiles (Fig. b) of forced jet and their comparison with the unforced case. The velocity profiles of the unforced jet are rather narrow and spread only slowly with the increasing distance. The existence of the potential core can be observed as region near the jet axis where the remains the same as is on the axis (visible till y/d = 5). The profile at the nozzle exit is flat with rather thin boundary layer. Due to the short exit channel of the nozzle the parabolic profile (which would correspond to the low Re) was not developed. The RMS velocity is low, the higher values occur only on the edges of

8 Prague, February -3, t/t r (mm) Figure 8: Velocity profiles near the nozzle exit (y =.5 mm).. U (m/s) (a).... no control control..... y (mm) U RMS (m/s) no control control (b) y (mm) Figure 9: Time-mean velocity (a) and RMS velocity (b) decay. the jet in shear layer. The control has a significant effect on the time-mean velocity profiles. In the very near of the nozzle exit the profiles becomes sharper instead of flat one of the unforced jet. The region of the constant velocity around the nozzle axis (potential core region) does not exist here. With the increasing distance the profiles markedly spread. The control effect is also visible on the RMS velocity profiles. The values are distinctly higher in the nearfield and reach the level of RMS velocity of the unforced jet around y/d =.

9 TOPICAL PROBLEMS OF FLUID MECHANICS 3 y/d = y/d = y/d = 5 y/d = 5 y/d =.5 y/d =.5 y/d = y/d = y/d = y/d = (a) (b) Figure : Time-mean velocity (a) and RMS velocity (b) profiles. 4 Conclusions The axisymmetric air jet was actively controlled by a pair of counter synthetic jets running in opposed direction. The actuators input signal was a pulse modulated sinusoidal signal where only one pulse was generated at the beginning of the working cycle. The behaviour of the forced jet was compared with the properties of the unforced jet of the same Reynolds number (Re = 3). The resulting controlled jet is

10 4 Prague, February -3, 8 intermittent in character. At the beginning of the working cycle the flow is bifurcated, its velocity and width increases, the RMS velocity increases sharply. The region of the constant velocity on the jet axis (potential core) does not exist in the jet, the time-mean velocity decreases gradually from the nozzle exist, however, the velocity decay is slower than that of the unforced jet behind the potential core region. Although the control jets directly operate only /8 of the working cycle and their influence is the most significant in the nearfield, the control action influences the main jet during the whole working cycle and even in the farfield; the velocity is lower and the time-mean velocity profiles are wider. The present experiment proved that the pulse modulated synthetic jets. i.e. jets with relatively low power consumption can be used as an active control tool. This paper aimed to be the initial study dealing with the present experimental setup, so only one combination of the main jet (with rather low Reynolds number) and control jets were chosen. The next research will be aimed at wider range of parameters (various combinations of the Reynolds number of the main jet and the frequency and amplitude of the control jets, i.e. various Strouhal numbers) and the usage of the controlled jet as an impinging jet in heat transfer experiments. Acknowledgment The support of the Grant Agency CR (project number S) is gratefully acknowledged. References [] Lepicovsky, J., Ahuja, K. K. & Burrin, R. H.: Tone Excited Jets, Part III: Flow Measurements, Journal of Sound and Vibrations, vol. (985) pp [] Zhang, S. & Turner, J.T.: On the Development of a Turbulent Jet Subjected to Aerodznamic Excitation in the Helical Mode, Experimental Thermal and Fluid Science, vol. 78 (6) pp [3] Trávníček, Z., Peszyński, K., Hošek, J. & Wawrzyniak, S.: Aerodynamic and Mass Transfer Characteristic of an Annular Bistable Impinging Jet with a Fluidic Flip-Flop Control, International Journal of Heat and Mass Transfer, vol. 46 (3) pp [4] Saminy, M., Kim, J.-H., Kastner, J., Adamovich, I. & Utkin, Y.: Active Control of High-Speed and High-Reynolds-Number Jets Using Plasma Actuators, Journal of Fluid Mechanics, vol. 578 (7) pp [5] Smith, B.L. & Glezer, A.: The Formation and Evolution of Synthetic Jets, Physics of Fluids, vol. (998) pp [6] Trávníček, Z. & Tesař, V.: Annular Synthetic Jet Used for Impinging Flow Mass Transfer, International Journal of Heat and Mass Transfer, vol. 46 (3) pp [7] Gillespie, M.B., Black, W.Z., Rinehart, C. & Glezer, A.: Local Convective Heat Transfer from a Constant Heat Flux Flat Plate Cooled by Synthetic Air Jets, Transactions ASME Journal of Heat Transfer, vol. 8 (6) pp. 99. [8] Valiorgue, P., Persoons, T., McGuinn, A. & Murray, D.B.: Heat Transfer Mechanisms in an Impinging Synthetic Jet for a Small Jet-To-Surface Spacing, Experimental Thermal and Fluid Science, vol. 33 (9) pp [9] Pack, L.G. & Seifert, A.: Periodic Excitation for Jet Vectoring and Enhanced Spreading, Journal of Aircraft, vol. 38 () pp [] Tensi, J., Boué, I., Paillé, F. & Dury, G.: Modification of the Wake behind a Circular Cylinder by Using Synthetic Jets, Journal of Visualization, vol. 5 () pp [] Chen, F. J. & Beeler, G.B.: Virtual Shaping of a Two-Dimensional NACA 5 Airfoil Using Synthetic Jet Actuator, AIAA Paper 373 (). [] Zhang, P.F. & Wang, J.J.: Novel Signal Wave Pattern for Efficient Synthetic Jet Generation, AIAA Journal, vol. 45(7) pp [3] Qayoum, A., Gupta, V., Panigrahi, P.K. & Muralidhar, K.: Influence of Amplitude and Frequency Modulation on Flow Created by a Synthetic Jet Actuator, Sensors and Actuators A: Physical, vol. 6 () pp [4] Amitay, M. & Glezer, A.: Flow Transients Induced on a D Airfoil by Pulse-Modulated Actuation, Experiments in Fluids, vol. 4 (6) pp [5] Trávníček, Z., Němcová, L., Kordík, J., Tesař, V. & Kopecký, V.: Axisymmetric Impinging Jet Excited by a Synthetic Jet System, International Journal of Heat and Mass Transfer, vol. 55 () pp [6] Blevins, R.D.: Applied Fluid Dynamics Handbook, Krieger Publ. Comp., Malabar, Florida (3).

DYNAMICS OF CONTROLLED BOUNDARY LAYER SEPARATION

p.1 DYNAMICS OF CONTROLLED BOUNDARY LAYER SEPARATION Václav Uruba, Martin Knob Institute of Thermomechanics, AS CR, v. v. i., Praha Abstract: The results of experimental study on a boundary layer separation