1.2. BLUFF BODY AERODYNAMIC WAKE STRUCTURE CONTROL BY A HIGH FREQUENCY DIELECTRIC BARRIER DISCHARGE

Transcription

1 1.2. BLUFF BODY AERODYNAMIC WAKE STRUCTURE CONTROL BY A HIGH FREQUENCY DIELECTRIC BARRIER DISCHARGE Moralev I.A., Kazansky P.N., Chertov D.S., Klimov A.I., Bityurin V.A., Borisov I.A. 1 Joint Institute for High Temperatures RAS, Izhorskaya 13, bld..2, Moscow, , Russia ivmoralev@gmail.com Abstract. In recent work, flow control in two cases is studied: leading edge stall on the wing model at high attack angle and unsteady boundary layer separation on the circular cylinder. Discharge used for flow control is a dielectric barrier type capacity coupled HF discharge (CHFD), with typical parameters: HF frequency F HF ~350 khz, modulation frequency F M = Hz, mean HF power N HF < 20W/cm span. Airflow parameters are the followings: airflow velocity up to 140m/s, Re < 6x10 5. Wake structure behind models is studied with PIV, shadow methods, flow visualization and pressure measurements. Vortex shedding triggering is obtained when forcing at Sh~0.3 for a cylindrical model. Pressure measurements in the model s wake are performed. Drag decrease of about 40% is measured for a wing model at M~0.3, Re~3x10 5 and HF plasma on. This result is obtained at post-stall angles of attack (AoA). It is revealed, that HF plasma actuator effectiveness drops significantly at stall attack angle and Re > 4x10 5. This result should be connected with a boundary layer turbulisation near the leading edge of a wing model. This work follows our previous ones [1,2]. Nomenclature HF = high frequency DBD = dielectric barrier discharge SHFD = surface HF discharge f HF = HF frequency F M = modulation frequency T i = pulse duration I HF = HF electric discharge current U HF = HF electric discharge voltage N HF = pulse power input in plasma M =Mach number V af = airflow velocity P 0 = stagnation pressure P st = static pressure T g = gas temperature T R =rotation temperature Sh = Strouchal s number =attack angle of actuator I. Introduction Surface discharge flow separation control is a task of a great interest today. It is well known that efficiency of flow control near a body by a chord-wise dielectric barrier discharge (DBD) is decreased considerably at high airflow velocity (V af >10m/s). The reason is physical limitations on the DBD induced momentum, while pressure gradients in the external flow increases with velocity increase [3,4]. However, attempts are recently made to use the surface discharges for boundary layer control, referring to other mechanisms of discharge flow action. Several works were devoted to flow separation control at velocities up to 240 m/s by high-current nanosecond discharges [5,6]. Authors argue the discharge in these cases acts rather as a pulsing heat release source than a flow momentum generator. The problem of a boundary layer flow turbulization and generation of a longitudinal vorticity by means of DBD was studied in [7]. Surface high frequency discharge was proposed in [8,9] for separation control on a flat plate and cone drag reduction. It was shown, that at definite angles of attack and discharge regimes, significant change in pressure distribution behind the plate can be achieved. Main goal of our current work is to study the separation process and wake structure behind the bluff body with transversal SHFD, operating near the separation point. Cylinder was chosen for this study as an aerodynamic body with well-known flow structure in the wake for a variety of regimes [10]. Lowspeed flow around the cylinder with the DBD discharge actuators was studied in [11-13]. Wake structure at pulse-periodic discharge operation was studied in [11], actuator orientation effects in [12]. In [13], vortex shedding was synchronized along the length by means of co-flow DBD. However, all authors implies that it is ionic wind, induced by the discharge, that leads to effects obtained. In the works [14-16], a keen study of the cylinder wake excitation by a single DBD pulse initiated at different phases of natural shedding 22

2 cycle was performed. Measurements were performed by a high-speed PIV system, with simultaneous measurements of drag and lift force oscillations. Most dramatic changes were obtained for discharge initiation at the moment of separation point upstream motion. It was obtained, that vortex shedding can be disturbed for time period comparable to 6 cycles. This led to a 200% lift oscillations amplitude increase. Authors argued their results from the position of momentum addition, however, they obtained a closed separation region at the discharge location. Since the flow separation control around a wing is of significant practical importance, it was widely studied up to date. However, most authors, both working with conventional and with nanosecond DBD at velocities up to 20 meters per second, report an effective stall control at large attack angles at discharge positioned near the leading edge of the model. Several studies, concerning flap aerodynamics improvement by conventional DBD can be found in [17-19]. It is also known, that at Re<10 5, leading-edge stall mainly takes place due to laminar boundary layer separation. Boundary layer turbulization leads to separation delay due to a higher stability of the turbulent boundary layer. It is shown, that nanosecond discharge operation near the leading edge leads to significant (up to ) stall delay at high attack angles. Unsteady forcing seems to be an essential feature of any DBD, therefore several studies were focused on forcing frequency optimization. Starikovskii et. al. [20] reported a maximum efficiency of nanosecond discharge at Sh~1, Samimi et. al. [17] found frequency ranges for conventional discharge forcing. Furthermore, in vortex street formation along the shear layer behind the flap was obtained, synchronized to the actuator forcing. In spite of the wide use of the surface DBD discharges of any frequency range in the aerodynamic studies, up to the moment there is no clear understanding of the prevailing mechanism of the discharge action, especially at high oncoming flow velocities. Cylindrical model (fig.1a) was manufactured of a 40mm diameter quartz tube. Dielectric thickness was ~2mm. Discharge was organized in a spanwise direction. Electrodes were manufactured of aluminum foil 200 m thickness, the inner one was covered by epoxy or silicon resin to avoid arcing in the discharge gap. NACA airfoil model 8cm x 10cm (chord x span) was manufactured from Nylon-6. Model was positioned in the test section at a desirable attack angle using rotating windows. The electrodes were arranged near the model s leading edge (fig.1b), with wiring connected to the mounting points of the model. Flow blockage was up to 30% at 20 0 attack angle and 10% at 0 0 attack angle. Discharge was created by RF switch generator, loaded with the model through the aircore resonant transformer with resonance frequency ~350 khz. Output voltage measured was up to 20 kv, the peak discharge power- 1 kw. The generator was operating in the pulse-periodic regime, creating RF pulses with us duration and repetition frequency up to 10 khz. Typical RF pulse is shown on fig.2. Rise time was about 5-10 us. Pulse shape changed with discharge power due to resonant frequency shift during the pulse. These changes weren t controlled in the experiments. a II. Experimental setup Low-velocity (<20m/s) experiments were cared in the aerodynamic channel with 100x100x300 test section, operating in a continuous regime. Flow velocity can be changed by controlling the blower motor. Large-scale flow turbulence after the blower was quenched by honeycomb. However, turbulence level remained high enough (~several percents). For high-speed experiments (<140 m/s), aerodynamic system was supplied from high pressure tanks. Experiment duration was about 5-30s, depending on the airflow velocity. Fig.1. Model schemes. 1-ground lead, 2-exposed electrode, 3- quartz body, 4-discharge, 5- epoxy layer, 6- pressure tubes, 7- fastening bolt. Discharge voltage was measured with Tektronix P6014A HV probe, discharge current was measured in the ground lead with Tektronix AC current shunt. Stray current was negligible due to significant discharge capacity. Electrical power b 23

3 input was calculated by digital multiplication of current and voltage signals via TDS 2014B oscilloscope with ~10% error. Alternatively, energy release over the pulse was measured by an integration on the capacity Wake structure was studied by an array of mm Pitot tubes, positioned diameters downstream from the model, in the middle of its span. The pressure was measured by a 16-channel pressure scanner Esterline 9116 with 500 Hz time resolution. Shadow pictures of the flow were taken by excimer KrF laser (248nm) with 20ns pulse length. Direct shadow method with a divergent beam was used. Pictures were acquired by digital camera from the white-sheet paper used as wavelength convertor (UV to blue). PIV data were acquired by a LaVision PIV system, based on 200mJ NdYAG laser and a 2048x2048 pix camera. Frames were acquired at frame rate ~7Hz, with further averaging of images. Accuracy of a single measurement was about 2% (for 20 m/s), while averaging error was several times higher, velocity field resolution was about 1.5mm. Diagnostic equipment was synchronized with the discharge pulses through the frequency divider and delay generator, allowing phase-locked acquisition of data. Flow smoke visualization was also performed. Ammoniac smoke was injected in the rear point of the cylinder, tangentially to model s surface. III. Results and discussion A. Circular cylinder. As described earlier, SHFD ignition on the cylinder surface can impact the separation process on the cylinder surface, charge the properties of the flow in the wake and reduce aerodynamic drag [2,2123]. PIV images of the near wake behind the cylinder are presented on fig2. In this experiment discharge was operating on the frequency 130Hz. Phase-locked imaging was performed with a 1ms step. Discharge was located at ~80deg, with discharge channels propagating counter wise versa along the axis. Consequent stages of flow reaction on the discharge pulse can be described as follows. Fig.2. Velocity x-component (a) and vorticity (b) distributions in the wake behind a cylinder. V=20m/s, Re=60k, Sh=0.3 24

4 Fig.2 (Cont.). Velocity x-component (a) and vorticity (b) distributions in the wake behind a cylinder. V=20m/s, Re=60k, Sh=0.3 25

5 Fig.2 (Cont.). Velocity x-component (a) and vorticity (b) distributions in the wake behind a cylinder. V=20m/s, Re=60k, Sh=0.3 26

6 First, at t~1ms, a kink at the shear layer is formed. This kink cuts off the main vortex being shed from the source of vorticity. During this process, the separation point moves downstream and reaches the position ~140 deg at t~2ms from the pulse moment. At t~3ms, a new vortex is formed, and grows during the rest of a period. One can see smaller vortex structures in a shear layer, that further run into main one. Special study was performed in terms to separate the influence of the ions-induced velocity and thermal influence of the discharge. PIV data were acquired for three different discharge propagation directions: co-flow, counter-flow and spanwise. No qualitative difference in vortex shedding process was obtained for these cases. In both cases, the orientation of vorticity vector in the vortex core correspond to the one in the boundary layer. Flow visualization images are shown for comparison on fig.3. One can see large-scale vortices creation in the near wake, locked to the discharge pulse. discharge on. Drag coefficient based on a total pressure losses is plotted at fig.4 vs Re number. At Re~400k an abrupt decrease of the drag coefficient in a reference case is obtained. This should happen because of flow turbulization prior to separation point. Fig.4. Drag coefficient of the airfoil model vs flow velocity. IV. Conclusions PIV visualization images were obtained for the flow around a cylinder at Re~60k. Vortex triggering is obtained in the near wake. After the discharge, a separation point shift occurs, followed with the new vortex formation. Vortex formation length and diameter are reduced. SHFD at the wing leading edge is capable of stall control up to Re<400k, i.e. up to boundary layer transition. Fig3. Flow patterns at different delays after the discharge pulse 1-300us, 2-600us. V=20/c, F m =1kHz (Sh~2). B. Leading edge separation on a wing model Leading edge separation on a wing model was studied at V=20m/s in our previous work. However, flow velocities of a practical interest are significantly higher. Therefore, we extended our study to the range Re<4x10 4 and V<140m/s. Experiments were cared for two attack angles: 11 0 and 15. Drag decrease at both attack angles was obtained up to the oncoming flow velocities V<80m/s. Drag decrease due to separation point shift up to 10-50% was obtained at discharge on. At flow velocities 80<V<140m/s, model drag is decreased only at supercritical attack angles >*. At AoA=11 and V=140m/s the drag increases at References 1. V.A. Bityurin, A.N. Bocharov, A.I. Klimov, I.A. Moralev, and B.N. Tolkunov, Surface HF Plasma Actuator in Airflow, 40th AIAA Plasmadynamics and Lasers Conference June 2009, San Antonio, Texas, 2009, pp. AIAA A.I. Klimov, V.A. Bityurin, I.A. Moralev, B.N. Tolkunov, K. Zhirnov, and V. Kutlaliev, Surface HF Plasma Aerodynamic Actuator, 46th AIAA Aerospace Sciences Meeting and Exhibit 7-10 January 2008, Reno, Nevada, 2008, pp. AIAA A.V. Likhanskii, M.N. Shneider, D.F. Opaits, R.B. Miles, and S.O. Macheret, Limitations of the DBD effects on the external flow, 48th AIAA Aerospace Sciences Meeting Including the New Horizons Forum and Aerospace Exposition 4-7 January 2010, Orlando, Florida, D.F. Opaits, M.R. Edwards, S.H. Zaidi, M.N. 27

7 Shneider, R.B. Miles, and S.O. Macheret, Surface plasma induced wall jets, 48th AIAA Aerospace Sciences Meeting Including the New Horizons Forum and Aerospace Exposition 4-7 January 2010, Orlando, Florida, G. Correale, I.B. Popov, A.E. Rakitin, A.Y. Starikovskiy, S.J. Hulshoff, and L.L.M. Veldhuis, Flow Separation Control on Airfoil With Pulsed Nanosecond Discharge Actuator, 49th AIAA Aerospace Sciences Meeting including the New Horizons Forum and Aerospace Exposition. 4-7 January Orlando, Florida., D.V. Roupassov, a a Nikipelov, M.M. Nudnova, and a Y. Starikovskii, Flow Separation Control by Plasma Actuator with Nanosecond Pulsed- Periodic Discharge, AIAA Journal, vol. 47, Jan. 2009, pp R.E. Hanson, P. Lavoie, and A.M. Naguib, Effect of Plasma Actuator Excitation for Controlling Bypass Transition in Boundary Layers, 48th AIAA Aerospace Sciences Meeting Including the New Horizons Forum and Aerospace Exposition 4-7 January 2010, Orlando, Florida, 2010, pp. AIAA V.A. Bityurin, A.N. Bocharov, A.I. Klimov, I.A. Moralev, and B.N. Tolkunov, Surface HF Plasma Actuator in Airflow, 40th AIAA Plasmadynamics and Lasers Conference June 2009, San Antonio, Texas, 2009, pp. AIAA A.I. Klimov, V.A. Bityurin, I.A. Moralev, B.N. Tolkunov, K. Zhirnov, and V. Kutlaliev, Surface HF Plasma Aerodynamic Actuator, 46th AIAA Aerospace Sciences Meeting and Exhibit 7-10 January 2008, Reno, Nevada, 2008, pp. AIAA C.H.K. Williamson, VORTEX DYNAMICS IN THE CYLINDER WAKE, New York, F.O. Thomas, A. Kozlov, and T.C. Corke, Plasma Actuators for Cylinder Flow Control and Noise Reduction, AIAA Journal, vol. 46, Aug. 2008, pp S. Yamada, S. Koui, I. Hitoshi, H. Shinji, and M. Motosuke, Flow Behavior behind a Circular Cylinder by DBD Plasma Actuators in Low Reynolds Number, 48th AIAA Aerospace Sciences Meeting Including the New Horizons Forum and Aerospace Exposition 4-7 January 2010, Orlando, Florida, 2010, pp. AIAA T. McLaughlin, M. Munska, J. Vaeth, and T. Dauwalte, Plasma-Based Actuators for Cylinder Wake Vortex Control, 2nd AIAA Flow Control Conference, Portland, Oregon, June 28-1, 2004, 2004, pp. AIAA T.N. Jukes and K.-so Choi, Long Lasting Modifications to Vortex Shedding Using a Short Plasma Excitation, Physical Review Letters, vol. 102, 2009, p T.N. Jukes and K.-so Choi, Flow control around a circular cylinder using pulsed dielectric barrier discharge surface plasma, Physics of Fluids, vol. 21, 2009, p T.N. Jukes and K.-so Choi, Control of unsteady flow separation over a circular cylinder using dielectric-barrier-discharge surface plasma, Physics of Fluids, vol. 21, 2009, p J. Little and M. Samimy, High-Lift Airfoil Separation with Dielectric Barrier Discharge Plasma Actuation, AIAA Journal, vol. 48, Dec. 2010, pp J. Little and M. Samimy, Control of Separation from the Flap of a High-Lift Airfoil with DBD Plasma Actuation, 48th AIAA Aerospace Sciences Meeting Including the New Horizons Forum and Aerospace Exposition 4-7 January 2010, Orlando, Florida, 2010, pp. AIAA J. Little and M. Samimy, Control of Separation from the Flap of a High-Lift Airfoil with DBD Plasma Actuation, 48th AIAA Aerospace Sciences Meeting Including the New Horizons Forum and Aerospace Exposition 4-7 January 2010, Orlando, Florida, 2010, pp. AIAA D.V. Roupassov, a a Nikipelov, M.M. Nudnova, and a Y. Starikovskii, Flow Separation Control by Plasma Actuator with Nanosecond Pulsed- Periodic Discharge, 47th AIAA Aerospace Sciences Meeting Including The New Horizons Forum and Aerospace Exposition 5-8 January 2009, Orlando, Florida, 2009, pp A.I. Klimov, I.A. Moralev, V.A. Bityurin, P.N. Kazansky, and D.A. Chertov, Flow Around Wing Model with a Surface HF Discharge, 49th AIAA Aerospace Sciences Meeting including the New Horizons Forum and Aerospace Exposition. 4-7 January Orlando, Florida., I.A. Moralev, V.A. Bityurin, P.N. Kasansky, A.I. Klimov, and D.A. Chertov, Flow Control around Wing Model by HF DBD Discharge, The 9th Iternational Workshop on Magnetoplasma Aerodynamics, I.A. Moralev, V.A. Bityurin, P.N. Kasansky, A.I. Klimov, and D.A. Chertov, Flow Control around Cylinder by HF DBD Discharge, The 9th Iternational Workshop on Magneto-plasma Aerodynamics, 2010, pp