Hypersonic Wind Tunnel Test of a Flare-type Membrane Aeroshell for Atmospheric Entry Capsules

Transcription

1 Trans. JSASS Aerospace Tech. Japan Vol. 8, No. ists27, pp. Pe_27-Pe_32, 2010 Original Paper Hypersonic Wind Tunnel Test of a Flare-type Membrane Aeroshell for Atmospheric Entry Capsules By Kazuhiko YAMADA 1), Masashi KOYAMA 2), Yusuke KIMURA 3), Kojiro SUZUKI 2), Takashi ABE 1) and A. Koichi HAYASHI 3) 1) Institute of Space and Astronautical Science, Japan Aerospace Exploration Agency, Sagamihara, Japan 2) Graduate School of Frontier Sciences, The University of Tokyo, Kashiwa, Japan 3) Graduate School of Science and Engineering, Aoyama Gakuin University, Sagamihara, Japan (Received July 14th, 2009) A flexible aeroshell for atmospheric entry vehicles has attracted attention as an innovative space transportation system. In this study, hypersonic wind tunnel tests were carried out to investigate the behavior, aerodynamic characteristics and aerodynamic heating environment in hypersonic flow for a previously developed capsule-type vehicle with a flare-type membrane aeroshell made of ZYLON textile sustained by a rigid torus frame. Two different models with different flare angles (45º and 60º) were tested to experimentally clarify the effect of flare angle. Results indicate that flare angle of aeroshell has significant and complicate effect on flow field and aerodynamic heating in hypersonic flow at Mach 9.45 and the flare angle is very important parameter for vehicle design with the flare-type membrane aeroshell. Key Words: Hypersonic Wind Tunnel Test, Flexible Aeroshell, Atmospheric Entry Nomenclature C D : drag coefficient C L : lift coefficient C PA : specific heat of air [=1006 J/kg/K] C PV : specific heat of Vespel [J/kg/K] C PZ : specific heat of ZYLON [J/kg/K] q int : heat flux just after model injection[w/m 2 ] R n : curvature radius at stagnation [m] S : cross-sectional area [m 2 ] T : temperature [K] T int : temperature before model injection [=300K] t : time [s] x : position [m] α : angle of attack [deg] σ : Stefan Boltzmann constant [= W/m 2 /K 4 ] ε : emissivity [=0.87] ρ V : density of Vespel [kg/m 3 ] ρ : area density of ZYLON filament textile h Z κ V 1. Introduction [=0.142 kg/m 2 ] : heat conduction coefficient of Vespel [W/K/m] A capsule-type vehicle with a large but low-mass, flexible aeroshell is promising as an innovative atmospheric entry system in the near future 1). The most appealing advantage of this system is reduced aerodynamic heating due to its low ballistic coefficient. An atmospheric entry vehicle with a flexible aeroshell can reach the surface of the Earth or other planets that do not have a severe high temperature environment. Such a large aeroshell can also be used for terminal deceleration of an atmospheric entry vehicle before touchdown, and a soft landing can be achieved without a conventional parachute and/or retro-jet. Various configurations and applications of flexible aeroshells for atmospheric entry systems have been researched and developed since ). Recently, NASA s Program to Advance Inflatable Decelerators for Atmospheric Entry (PAIDAE) started a program aimed at developing and implementing Inflatable Aerodynamic Decelerators (IAD) for both human and robotic use. To date, material development, mission and system design, wind-tunnel testing, and flight tests of deployable aerodynamic decelerators have been actively researched in this program 3,4). On the other hand, since 2000, our group has researched and developed a flexible aeroshell for atmospheric entry capsules 5-7), focusing on a flare-type aeroshell composed of a conical membrane and a ring-shaped outer frame attached to a blunt rigid capsule. Figure 1 shows overview of the mission concept assumed in this study and a schematic of a re-entry vehicle with this flare-type membrane aeroshell derived from a tension-shell structure 8). If an inflatable torus frame is adopted as an outer frame, the entire aeroshell consists only of a thin membrane structure, thus making fabrication of a large but low-mass aeroshell relatively easy. Three issues that need resolving in order to apply this aeroshell system to actual missions are currently being researched by our group: 1) To develop a large but low-mass, flexible aeroshell, including an inflatable structure. 2) To develop and evaluate flexible materials that are gas impermeable, highly heat resistant and sufficiently strong to withstand atmospheric entry conditions. 3) To determine the aerodynamic characteristics of an atmospheric entry vehicle with a flexible aeroshell over a wide range of Mach number. To resolve issues 2 and 3, in this study, hypersonic wind tunnel tests were done for a capsule-type model with our previously developed flexible aeroshell as a first step for the Copyright 2010 by the Japan Society for Aeronautical and Space Sciences and ISTS. All rights reserved. 1 Pe_27

2 Trans. JSASS Aerospace Tech. Japan Vol. 8, No. ists27 (2010) vehicle design in hypersonic flow. The main objectives of the tests were to establish a measurement method of the aerodynamic force and the aerodynamic heating of the flare-type membrane aeroshell in hypersonic wind tunnel. And we tried to clarify the performance of the flare-type membrane aeroshell as a decelerator in hypersonic flow and to investigate the aerodynamic heating environment around the membrane aeroshell for the development and evaluation of membrane materials. In this study, two different models with membranes of different flare angles (45º and 60º) were tested as reference configurations. thickness of ZYLON spun yarn textile was approximately 0.3 mm. Fig. 2. Schematic of the model used in the hypersonic wind tunnel test. Fig. 1. Schematic of reentry system with the flare-type membrane aeroshell sustained by an inflatable torus frame, and overview of the mission concept assumed in this study. 2. Experimental Setup 2.1 Experimental model Figure 2 shows a schematic of the capsule model with a flare-type flexible aeroshell tested in the hypersonic wind tunnel. The model consisted of a hemispherical head imitating a blunt capsule, a flare-type membrane aeroshell made of ZYLON 9) filament textile, and an aluminum torus frame imitating an inflatable torus. The ZYLON, which has high heat resistance and strength, is a candidate material for future actual atmospheric entry missions. The thickness of the textile used in this test was approximately 0.15mm. The hemispherical head was made of stainless steel or Dupont Vespel and was 20 mm in diameter. The stainless steel head is used for aerodynamic force measurement and the Vespel head was used for temperature measurements. The flare-type membrane aeroshell was 140 mm in diameter and the torus frame was 10mm in tube diameter. The torus frame was attached to the outer end of the flare-type aeroshell, and thus the maximum diameter of the aeroshell was 160 mm. Two different models with aeroshells of different flare angles were tested to determine the effect of flare angle on the flow field, the aerodynamic characteristics and the aerodynamic heating environment; 45º flare angle (F45-10 model) and 60º flare angle (F60-10 model). Figure 3 shows a photograph of the F45-10 model. The aeroshell was constructed by sewing together six trapezoidal petals made of ZYLON filament textile. The junction of the aeroshell with either the head or the frame was reinforced with ZYLON spun yarn textile stitched onto the backside of the aeroshell. The Fig. 3. Photo of the F45-10 model (flare-type membrane aeroshell made of ZYLON textile; 45º flare angle; 10mm tube diameter in a torus frame). 2.2 Experimental method The hypersonic wind tunnel tests were carried out using the JAXA φ1.27 m hypersonic wind tunnel (φ1.27 m nozzle exit diameter) located at Chofu Aerospace Center. Figure 4 shows the hypersonic nozzle and the test section. This facility can produce a uniform flow at a Mach number of 9.45 and Reynolds number of (1/m) when the reservoir pressure is set at 1.0 MPa and the temperature at 920 K. Table 1 lists the other uniform flow conditions. In this test, the aerodynamic force of the capsule model with the flare-type flexible aeroshell was determined by measuring the lift and drag coefficients (C L and C D, respectively) using a sting balance system in which the angle of attack (α) of the model was varied from -2.0º to 12º. The temperature distribution on the model surface was measured using infrared thermography 10). The flow field was visualized using the schlieren method and the deformation of the flexible aeroshell was also observed with video camera. Pe_28

3 K. YAMADA et al.: Hypersonic Wind Tunnel Test of a Flare-type Membrane Aeroshell for Atmospheric Entry Capsules model (upper set), a bow shock wave generated by the head, and a shock wave generated by the torus frame interacted at the middle of the aeroshell. Additionally, the shock wave in front of this model oscillated significantly and rapidly, although the membrane aeroshell was stable. These results show that (1) in terms of flow stability, the F60-10 model is superior to the F45-10 model, (2) the flare angle of the aeroshell has a significant yet complicated effect on the flow field, and (3) the flow field might change discontinuously between that for the 45º and 60º flare angles in the hypersonic flow at Mach Future research on the effect of flare angle on the flow field Mach number effect on flow stability will include numerical simulation. Fig. 4. Hypersonic nozzle and test section of JAXA φ1.27m hypersonic wind tunnel. Table 1. Uniform flow conditions in the hypersonic wind tunnel (Fig. 4). Reservoir pressure Reservoir temperature 1.0 MPa 920 K Mach number 9.45 Velocity 1330 m/s Static temperature 49 K Static pressure 33.3 Pa Static density kg/m 3 Dynamic pressure 2080 Pa Reynolds Number /m Heat flux at stagnation estimated by Sagnier s Equation 11) 137 kw/m 2 (R n=0.01m) 3. Results 3.1. Flow field and behavior of the aeroshell Figure 5 shows photos of the F45-10 and F60-10 models at 0º angle of attack in hypersonic flow during the wind tunnel tests. The aeroshell was deformed to the concave shape due to the aerodynamic force. Both models were quite stable without severe oscillation of the aeroshell. Fig. 5. Photos of the F45-10 (left) and F60-10 (right) models at 0º angle of attack α in hypersonic flow at Mach Figure 6 shows sequential schlieren photographs of the flow field (0.05-sec intervals) around the F45-10 and F60-10 models in hypersonic flow at Mach The shape of the bow shock in front of the model differs significantly between the F45-10 and F60-10 models. In the F60-10 model (lower set), the strong bow shock wave is similar to a nominal shock wave, and the flow field was quite stable. In contrast, in the F45-10 Fig. 6. Sequential schlieren photographs (0.05-sec intervals) visualizing the flow field around the F45-10 model (upper set) and F60-10 model (lower set) at 0º angle of attack α in hypersonic flow Aerodynamic coefficients Figure 7 shows the drag coefficient C D and lift coefficient C L of the F45-10 and F60-10 models. The frontal projected area was used as the reference area in the calculation of the aerodynamic coefficients. For the F45-10 model, C D of the F60-10 model slightly decreased and its variation was small (less than 3% in the range of angle of attack from -2 to 12 degrees). At α = 0º, C D of the F45-10 and F60-10 models was around 1.51 and 1.62, respectively. In terms of deceleration performance, the F60-10 model as a decelerator is therefore suitable because a large C D is desirable in hypersonic flow. The C L of the F45-10 model is less sensitive to α than is the F60-10 model, probably due to deformation of the membrane aeroshell against the flow. Figure 8 shows the inclination angles of the frame plotted as a function of α. If the aeroshell did not deform, then the inclination of the frame was equal to α. These results show that in the F60-10 model, the relative position of the model axis did not change with respect to the frame, whereas in the F45-10 model, the membrane aeroshell tended to force the frame to be perpendicular to the freestream direction. The effective α of the F45-10 model was therefore comparatively smaller than that for the F60-10 model, thus Pe_29

4 Trans. JSASS Aerospace Tech. Japan Vol. 8, No. ists27 (2010) explaining the difference between the F60-10 and F45-10 models in the sensitivity of their C L to α. 3.0 sec. The surface temperature distribution on the aeroshell differed between the two models. In the F45-10 model, the temperature near the outer edge of the aeroshell was higher than that near the head. Conversely, in the F45-10 model, the temperature near the head was higher than that near the outer edge. Fig. 7. Drag coefficient C D and lift coefficient C L of the F45-10 and F60-10 models as a function of the angle of attack α. Fig. 9. Surface temperature distribution of the F45-10 and F60-10 models measured (using an IR thermography) 1.0 sec after model injection into hypersonic flow. Fig. 8. Inclination of the frame for the F45-10 and F60-10 models as a function of the angle of attack α. 3.3 Surface heat flux q int The temperature distribution on the model surface in the hypersonic flow was measured using an IR thermography system 10), which obtained the surface temperature distribution as two-dimensional images at 0.3-sec intervals. Figure 9 shows the obtained images for both models 1.0 sec after the model was injected into the flow by the rapid model injection system. In this measurement, the emissivity of ZYLON filament textile (aeroshell material) was assumed to be the same as that of Vespel (head material), namely, For both models, the temperature on the aeroshell surface was higher than that on the head (Fig. 9) due to the small heat capacity of the thin membrane. The temperature on the aeroshell reached 300 ºC at 1.0 sec after model injection into the flow (Fig. 9), and was expected to reach nearly 500 ºC at The governing equation of the heat balance on the model surface was determined here to estimate the surface heat flux (q int ) from the time history of the surface temperature. In this analysis, the quasi-one-dimensional heat conduction equation as expressed in equation (1) was adopted for the model head, because the model head was a small hemisphere. The boundary condition of the heat conduction equation on the surface and center of the head is expressed in equation (2). The convective aerodynamic heating and radiative cooling that depend on the surface temperature were considered in the boundary condition on the model surface. dt T ρv SC pv = κ V S dt x x T 1 CpAT 4 4 T κv = qint εσ ( T Tint ), κv = 0 x surface 1 CpAT int x center On the membrane aeroshell, a simple equation that considers the heat balance between the convective aerodynamic heating and radiative cooling was adopted as expressed in equation (3). The heat convection in the thickness direction was ignored on the aeroshell because the membrane aeroshell was very thin. dt 1 CpAT 4 4 ρhc z pz = qint 2 εσ ( T Tint ) dt 1 C pat int (1) (2) (3) Pe_30

5 K. YAMADA et al.: Hypersonic Wind Tunnel Test of a Flare-type Membrane Aeroshell for Atmospheric Entry Capsules In these equations the specific heats of ZYLON and Vespel and the heat conduction coefficient of Vespel depend on the temperature. In this analysis, the relations between these properties and the temperature were determined according to the database in references 9) and 10). Both of these governing equations (2) and (3) can be solved, and the time history of the surface temperature can be calculated numerically if the initial heat flux (q int ) is given. The calculated time history of surface temperature must agree with the experimental data with the correct q int and the appropriate governing equation. In the analysis here, the time history was calculated using various q int in order to find the correct q int. Figure 11 shows a comparison of the history of surface temperature between the calculated results with various q int and the experimental data for the F45-10 model at two reference points, namely, stagnation point and at the middle of the aeroshell. The calculated results agree qualitatively with the experimental data, indicating that the governing equation adopted for the calculation was appropriate. The estimated q int at the stagnation point was about 137 kw/m 2 and that at the middle of the aeroshell was 43.5 kw/m 2. The total error in the estimated q int using IR thermography system was ±10% if the model was made of Vespel 10). On the aeroshell, the error was slightly larger than that on the head due to uncertainty in the emissivity of ZYLON filament textile assumed here. If the error in the emissivity of ZYLON filament textile is assumed to be ±10%, the total error in the estimated q int on the aeroshell is 14%. Quantitative analysis of q int on the torus frame was not carried out in this study, because the thermal structure of the frame that included the ZYLON filament textile, ZYLON span yarn textile and the hollow aluminum pipe was too complicated to derive the applicable governing equation for heat convection. direction. The q int on the aeroshell was averaged in the circumferential direction. Fig. 12. Heat flux q int distribution in the radial direction on the head and aeroshell of the F45-10 and F60-10 models estimated from the time history of the surface temperature. The q int at the stagnation point was 138.7kW/m 2 and 128.1kW/m 2 for the F45-10 and F60-10 models, respectively. In the F45-10 model, the q int value agrees with that predicted using Sagnier s equation (137kW/m 2 ). In the F60-10 model, q int was slightly lower than in the F45-10 model due to effect of the aeroshell. The q int distribution on the head was similar for both models, although q int of the F45-10 model was slightly higher than that of the F60-10 model. In contrast, the q int distribution on the aeroshell significantly differed between the two models. In the F60-10 model, q int on the aeroshell was maximum near the head and gradually decreased to the outer edge. In the F45-10 model, q int on the aeroshell was minimum near the head and sharply increased at the middle of the aeroshell, and reached a maximum at the outer edge. This trend for the F45-10 model can be explained by the shock-shock interaction at the middle of the aeroshell shown in Fig. 6. Thus, the F60-10 and F45-10 models showed opposite trends for the q int distribution on the aeroshell. In summary, the flare angle had a significant and complicated effect not only on the flow field but also the aerodynamic heating environment on the aeroshell. The flare angle is very important parameter for actual atmospheric entry vehicle design with flare-type membrane aeroshell. 4. Conclusions Fig. 11. Calculated and experimental time history of surface temperature at stagnation point and at the middle of the aeroshell of the F45-10 model. Figure 12 shows the q int distribution in the radial direction for both models just after model injection. The x-axis represents the distance from the body axis in the radial The behavior, aerodynamic characteristics and aerodynamic heating environment were investigated experimentally for a capsule-type vehicle with a flare-type flexible aeroshell sustained by a torus frame in hypersonic flow at Mach number of 9.45 in the JAXA φ1.27m hypersonic wind tunnel. In this test, the measurement method about the aerodynamic force and the aerodynamic heating of the flexible aeroshell was established. Two different models with membrane aeroshells of different flare angles (45º and 60º) were tested. The both Pe_31

6 Trans. JSASS Aerospace Tech. Japan Vol. 8, No. ists27 (2010) model were quite stable and produced the reasonable drag force. It is demonstrated that the flare-type membrane aeroshell works well as a decelerator in hypersonic flow. However, the flare angle of the aeroshell has a significant yet complicated effect on the flow field and aerodynamic heating condition and is very important parameter for the vehicle design. These results for the drag coefficient and flow field stability indicate that the aeroshell with 60º flare angle is superior under the test conditions used here. Our future research will include various parametric studies using CFD combined with wind tunnel tests. The design of an actual aerodynamic entry mission will be accelerated not only by the results of such parametric studies and by our results obtained in this study but also by factors such as aerodynamic stability and structural strength of the aeroshell against the aerodynamic forces. Acknowledgments The hypersonic wind tunnel tests were carried out in collaboration with the Wind Tunnel Technology Center in Aerospace Research and Development Directorate, Japan Aerospace Exploration Agency. Authors thank the staff of the JAXA φ1.27m hypersonic wind tunnel for their support and advice. References 1) Iannotta, B.: Down-to-earth: Transport for Space Cargo, Aerospace America, 38,7 (2000), pp ) Walberg, G. D.: A Survey of Aeroassisted Orbit Transfer, J.Spacecraft, 22 (1985), pp ) Clark, I. G., Hutchings, A. L., Tanner, C. L. and Braun, R. D.: Supersonic Inflatable Aerodynamic Decelerators for Use on Future Robotic Missions to Mars, J.Spacecraft and Rockets, 46 (2009), pp ) Clark, I. G., Cruz, J. R., Hughes, M. F., Ware, J. S. and Madlangbayan, A.: Aerodynamic and Aeroelastic Characteristics of a Tension Cone Inflatable Aerodynamic Decelerator, AIAA paper, , (2009). 5) Yamada, K., Suzuki, K. and Hongo, M.: Aerodynamic Characteristics of Frustum-Shaped Elastic Membrane Aeroshells in Supersonic Flow, J.Spacecraft and Rockets, 43 (2006), pp ) Kimura, Y., Yamada, K., Akita, D., Abe, T., Suzuki, K., Imamura, O., Koyama M. and Hayashi, A. K.: Study on Low-ballisticcoefficient Atmospheric-entry Technology using Flexible Aeroshell, ISTSpaper, 2008-g-18, (2008). 7) Yamada, K., Akita, D., Sato, E., Suzuki, K., Narumi, T. and Abe, T.: Flare-Type Membrane Aeroshell Flight Test at Free Drop from a Balloon, J.Spacecraft and Rockets, 46 (2009). 8) Anderson, M. S., Robinsion, J. C., Bush, H.G. and Fralich, R.W.: A Tension Shell Structure for Application to Entry Vehicles, NASA, TN, D-2675, ) Toyobo Co., Ltd. PBO FIBER ZYLON Technical Data Sheet ) Koyama, T., Tsuda, S., Hirabayashi, N., Seike, H., Hozumi, K. and Watari, M.: Measurement of Heat Transfer Distribution by Infrared Thermograph Technology, ISSN , JAXA-RR , (2007). 11) Sagnier, P. and Verant J.:Flow Characterization in the ONERA F4 High-Enthalpy Wind Tunnel, AIAA Journal, 25 (1998), pp Pe_32