SUPERSONIC PARACHUTE AERODYNAMIC TESTING AND FLUID STRUCTURE INTERACTION SIMULATION

Transcription

1 SUPERSONIC PARACHUTE AERODYNAMIC TESTING AND FLUID STRUCTURE INTERACTION SIMULATION J. Stephen Lingard (1), John C. Underwood (1), Matt Darley (1), Arrun Saunders (1), Lionel Marraffa (2), Luca Ferracina (3) (1) Vorticity Ltd, Chalgrove, Oxfordshire, OX44 7RW, UK, (2) European Space Agency, Noordwijk, The Netherlands, (3) ATG Europe B.V. on behalf of the European Space Agency, Noordwijk, Netherlands, ABSTRACT The ESA Aerodynamics of Download Systems: Parachutes activity expands the knowledge of parachute inflation and flying characteristics in supersonic flows. Model Disk-Gap-Band (DGB) parachutes were tested in the CNRC 1.5m by 1.5m Tri-sonic wind tunnel in Ottawa, Canada. The tests gathered data on inflation, drag and general flight stability at Mach numbers of 1.6, 1.8, 2.0 and 2.25 in the wake of a probe of similar configuration to Stardust. The parachute models, with diameters of 1.5 and 3.0 times the probe diameter, were flown at trailing distances from 7.5 to 10.5 forebody diameters. Parachute forces were measured and visual data obtained using high speed video and Schlieren. The results of FSI simulations used to reconstruct one of the CNRC tests are presented. Finally, the results are compared to those obtained within a supersonic wind tunnel test campaign of the same parachute conducted at the 10 x 10 NASA Glenn tunnel.. 1. NOMENCLATURE Α C d D 0 M t Inf τ V V x/d subsonic reference inflation time drag coefficient nominal parachute diameter Mach number inflation time non-dimensional inflation time tunnel flow speed =trailing distance ratio in forebody diameters. 2. INTRODUCTION The objective of the ESA Aerodynamics of Download Systems: Parachutes activity within the ESA Technology Research Programme was to expand the knowledge of parachute inflation and flying characteristics in supersonic flows: conduct wind tunnel tests to extend the available database and increase understanding of parachute performance at supersonic velocities; provide validation cases for CFD/FSI; conduct CFD/FSI matching of a selected case; develop improved algorithms and database for supersonic parachute performance. Building on a review of the state-of-the-art in supersonic parachute testing and modelling, wind tunnel tests have been planned and conducted to extend the available databases and increase understanding of parachute performance at supersonic velocities, by varying parameters in a systematic way. Scaled model parachutes (sized to comply with blockage and forebody shock re-impingement restrictions) were tested in the CNRC 1.5 m by 1.5 m Trisonic wind tunnel in Ottawa, Canada. The tests gathered data on inflation, drag coefficient and general flight stability at Mach numbers of 1.6, 1.8, 2.0 and 2.25 in the wake of a probe of similar configuration to Stardust. Parachute forces were measured and visual data obtained using high speed, standard video and Schlieren technique. The model Disk-Gap-Band (DGB) parachutes had diameters of 1.5 and 3.0 times the probe diameter and were flown at trailing distances ranging from 7.5 to 10.5 forebody diameters. Supersonic parachutes have been flown in many space missions [1], yet few supersonic test campaigns have taken place. Many missions have relied on heritage data, in part or in full. The earliest wind tunnel testing of supersonic parachutes intended for planetary missions was carried out on ribbon parachutes in the 1950 s [2]. Both solid and flexible parachutes were tested between Mach 1.6 and 3.0 in the Unitary Plan wind tunnel at NASA Langley. The importance of parachute porosity in supersonic performance and the variation of performance with Mach number were identified. Further wind tunnel testing at Langley in the early 1960 s identified the link between parachute suspension line length and drag coefficient and investigated a new geometry of parachute, flat-roof, conical inlet, which had a significantly higher supersonic drag coefficient than the ribbon chutes tested previously. Following the early wind tunnel testing of candidate parachute systems, flight tests were carried out in order to confirm whether the findings of the earlier testing translated to full scale. NASA carried out three sets of tests in the 1960 s and 1970 s: the Planetary Entry

2 Parachute Program [3], the Supersonic Planetary Entry Decelerator Program and the Supersonic High-Altitude Parachute Experiment. A total of 16 tests were carried out: five with ringsail parachutes, three with cruciform parachutes and eight with disk-gap-band (DGB) parachutes. The results confirmed earlier wind tunnel tests and demonstrated that the DGB and ringsail parachutes were candidates for planetary missions. The cruciform parachute had good drag characteristics but its stability at supersonic speeds was unacceptable. Further wind tunnel testing of the DGB design took place in the 16T and 16S tunnels at the Arnold Engineering Development Complex (AEDC) in the late 1960 s [4]. Variations on the parachute geometric porosity, trailing distance behind the payload and payload size were investigated. After the DGB was selected for the Viking lander, an extensive qualification programme took place. Further supersonic wind tunnel tests were performed at AEDC 5. These, in addition to four flight tests, confirmed the suitability of the DGB for the Viking mission. Following the Viking mission in 1972, no further western probes, using a supersonic parachute, were launched until the late 1990 s. Some additional testing took place within NASA in the following years, generally at small scale. One study worthy of note took place in 1974 [5], when a number of different parachute designs were tested in the NASA Langley 4 x4 tunnel at Mach 1.80 and the effects of reefing were investigated. The ESA Huygens probe was developed in the early 1990 s. For the Huygens mission, the main parachute was much smaller in relation to the payload than the Viking system so it was considered prudent to test this configuration in supersonic flow [6]. No full-scale supersonic tests of the Huygens parachutes took place. No supersonic testing of the parachutes for Mars Pathfinder, Mars Polar Lander, Mars Exploration Rovers or Beagle2 took place. All these missions relied on subsonic testing and heritage from the NASA work in the 1960 s, despite some of the parachute designs being significantly different to the tested configurations. The Mars Science Laboratory (MSL) mission utilized a DGB parachute, which was significantly larger than anything flown before. Even with this parachute, the ballistic coefficient for the MSL was also greater than previous missions, resulting in a slower deceleration, longer supersonic flight and a corresponding increase in criticality of the supersonic flight phase. It was thus decided to carry out supersonic wind tunnel testing [7], [8] to quantify the risks involved. The objectives of the test were twofold: to measure the flying shape of the canopy as it changes throughout the flight; and to provide data to validate computational fluid dynamics (CFD) simulations of the test. Parachute deployment was not modelled and the inflation was initiated from the streamed condition rather than bag-deployment. No full-scale supersonic testing of the MSL parachute took place due to the associated cost. Fluid structure interaction simulations [9] were conducted prior to the test campaign presented in this paper. The aim of this work was to assess the performance of the disk-gap-band parachute proposed for the ExoMars mission at flight conditions between Mach 0.4 and 2.8. The numerical solver LS-DYNA was used to model the fluid flow about the aeroshell and parachute. They showed that parachute drag loss in the supersonic regime results from forebody wake interference with the parachute bow shock. The results demonstrated two mechanisms that cause cyclical inflation and collapse of a parachute s canopy during supersonic flight. The first derives from flying the parachute too close to the base of the forebody, as it allows coupling of the subsonic region of the forebody wake with the subsonic region in the parachute. The second mechanism is the resulting effect of high-energy flow causing large bow shock movements, which result in large pressure fluctuations in the parachute with consequential gross shape changes. The test results presented in this paper attempt to build on the work of all the previous testing and add the body of knowledge surrounding supersonic wind tunnel testing of parachutes. Therefore, the main objectives of this test were to measure the inflation dynamics (quantitative) and the drag coefficient (quantitative). A qualitative assessment of parachute stability was also attempted. 3. TEST SETUP 3.1. Test Articles Two sizes of DGB parachute were manufactured for this test. The larger parachutes had a nominal diameter of 340 mm and were sized to have a diameter ratio of 1:3 with respect to the 113 mm forebody; roughly the same ratio as used on Huygens. Smaller parachutes with nominal diameter of 170 mm were also manufactured in order to investigate the effects of scale. The parachutes were manufactured from the lightest materials possible in order to represent the stiffness of the full-scale parachute as closely as possible. The parachutes were manufactured from low porosity fabric in order to exclude the variation in fabric porosity due to Reynolds number. They were manufactured from Nylon broadloom fabric with Kevlar lines. The characteristics of the parachutes tested are given in Table 1. Each parachute assembly consisted of: a parachute canopy; suspension lines; riser; and deployment bag. The parachutes were designed for maximum inflation loads of 5.0 kn (D0 340 mm parachute) and 1.3 kn (D0 170 mm parachute).

3 The test articles included risers with different lengths, which allowed the investigation of the effects of trailing distance. There were three options of trailing distance, x/d = 7.5, 9.0 and Parameter D mm D mm Flying diameter m m Reference Area m m 2 Expected Drag coefficient 0.55 (max) 0.55 (max) Drag area m m 2 Test section blockage 2.20% 0.54% Geometric porosity 21.7% 21.8% Line length m m Streamed length m m Number of lines Disk height m m Vent height m m Gap height m m Band height m m Gore width m m Mass kg kg Table 1. CNRC test parachute design characteristics 3.2. Test Apparatus The parachutes were deployed at Mach numbers between 1.60 and The nominal flow conditions at each of these Mach numbers are shown below in Table 2. M q p T p S T S V ρ(kg/m 3 ) (kpa) (kpa) (kpa) (K) (m/s) Table 2. Wind tunnel flow conditions. The mounting structure was a double, forward-swept strut that was designed to minimize flow disturbance and minimize the probability of shock interaction with the parachute. The general layout is shown below in Figure 1. The test parachute was packed in a void behind the scaled 60 spherical-cone forebody, Figure 2. The parachute was deployed via a drag cone and release mechanism. A small Nylon drag cone was attached to the end of a Kevlar extraction lanyard. The lanyard was attached to the release mechanism mounted on the forebody within the disturbance wake. A second length of Kevlar was attached at one end to the extraction lanyard at the release mechanism location and the other end was attached to the parachute deployment bag within the forebody void. On command a small pyrotechnic device operated the release mechanism allowing the force from the drag cone to extract and deploy the packed parachute. The parachute drag force was measured by means of a load cell mounted within the forebody at a rate of 25 khz. The data acquisition system was able to record for two seconds during the deployment followed by two seconds of steady state flight for each test. The tests were recorded on video data using one low speed, sidemounted, camera and two high-speed (5,200 fps) mounted to the side and rear of the deployed parachute. The high speed frame rate allowed approximately 30 and 16 frames to be captured during the deployment of the larger and smaller parachutes respectively. Schlieren videography was also available for the last three tests of the campaign. Figure 1. General layout of the test apparatus 3.3. Test Matrix Figure 2. Detail of forebody The matrix of tests performed is presented in Figure 3. The test matrix allowed for an analysis of the variation of drag coefficient, inflation time and trailing distance with varying Mach number. Figure 3 Test matrix

4 3.4. Results & Discussion The tests conducted allow assessment of the parachute inflation, drag and stability. In this paper we focus on the measured drag performance. Owing to the nature of flight for supersonic parachutes, statistical parameters are generally used to effectively analyse drag characteristics. Throughout the test campaign a number of parachutes suffered damage during deployment or damage during the flight due to the very high dynamic pressure for the tests (69 kpa). An analysis of the spread of drag coefficient data for a parachute flight was performed, which an indication as to whether the parachute was flying with or without damage. This was done by sorting the drag coefficient data into bins with a width of A histogram could then be produced to aid understanding of the content of the drag coefficient data. Figure 4 shows a typical spread of drag coefficient data for a parachute (test 53880) that had an ideal deployment and subsequently flew with no damage. Figure 5 shows a typical histogram indicating that damage has been sustained. The data are left biased. Figure 6. Drag coefficient data histogram for test The drag coefficients for all the test runs are presented below in Table 3. Test D 0 Mach x/d Mean C d(1%) C d(50%) C d(99%) (mm) C d Table 3. Parachute drag performance The effect of Mach number on parachute drag is shown in Figure 7. Above Mach 1.8 drag coefficient steadily decreases. Figure 4. Drag coefficient data histogram for test Figure 7. Effect of Mach number on drag coefficient Figure 5. Drag coefficient data histogram for test There may also be left-biased data in cases when the trailing distance is small, x/d < 9.0, as there is a coupling between the stagnation region in the parachute and subsonic flow filaments created in the forebody wake. An example of this is shown in Figure 6 for test An accurate analysis of all the tests was made by truncating the data to the period when the drag coefficient distribution indicated no damage had been sustained. The effect of varying trailing distance is shown in Figure 8. Drag coefficient is low when trailing distance is too short. Once adequate trailing distance to avoid subsonic coupling is established drag coefficient does not change significantly. Figure 8 Drag coefficient statistics for varying trailing distance ratio

5 Figure 9 shows the spread of drag coefficient data from the tests with 0.34 m nominal diameter parachutes to those with 0.17 m nominal diameter. All the tests had the same trailing distance parameter of 9.0. The drag coefficient of the smaller parachutes is lower than that of the larger parachutes. In both cases drag coefficient reduces with Mach number. Figure 10. Fluid mesh for LS-DYNA simulation cutaway through axis showing velocity contours The results of the rebuild are shown in Figures 11 and 12. Figure 9. Drag coefficient data showing the variation of scale and Mach number 4. FLUID STRUCTURE INTERACTION REBUILD LS-DYNA was used to rebuild test since there was Schlieren data available. The test was at Mach 2.0 and had a x/d of 9.0. Vorticity has extensive experience with LS-DYNA and has used it to model successfully parachute performance over a range of conditions, from subsonic to supersonic flight, in different atmospheres, and with fully flexible, deforming and moving parachute and forebody models. The simulation was developed as three separate meshes: rigid shell mesh for the forebody and strut; deformable membrane and cable mesh for the parachute; hexahedral, Eulerian mesh for the fluid domain. The fluid mesh comprised 7,461,384 hexahedral cells. No deformation of the fluid cells is used within the LS-DYNA ALE simulation; instead the flow is advected through the cells at each time step, similar to a conventional CFD finite-volume solution but with calculation of intersecting boundaries between the Lagrangian elements of the forebody, strut and parachute performed at each time step. Figure 10 illustrates the interaction of the fluid flow with the structural elements during the simulation. The colours show contours of axial velocity, with the shocks from the forebody and parachute clearly visible and the low speed or reverse flow behind the forebody and parachute also apparent. Figure 11. Schlieren image compared with FSI simulation Figure 12. Test drag coefficient for case compared with FSI simulation The excellent match obtained is very encouraging. 5. EXOMARS TESTING The parachute for the ExoMars probe is a 12.0 m nominal diameter disk-gap-band (DGB). The maximum deployment Mach number for the mission is 2.1. In the frame of the ESA Huygens program, a DGB parachute of the same configuration was tested in the Arnold s Engineering Development Center 16T wind tunnel up to Mach 1.5. It is necessary to demonstrate parachute

6 inflation at up to the ExoMars maximum Mach number of 2.1. The test article was a 6.88% scale model with a diameter of m (32.5 ). Testing was conducted at the NASA Glenn 10 x 10 (3.05 m x 3.05 m) supersonic wind tunnel. The test was intended to verify inflation, stability and drag coefficient at Mach 2.1 and above in a representative probe wake. Parachute force was measured using a load cell and visual data obtained using high speed and standard video in addition to Schlieren videography. The model characteristics are shown in Table 4. Parameter DGB Reference diameter m Flying diameter m Reference area m 2 Number of lines 20 Line length m Trailing distance m Flying length (clevis to vent) m Riser length m Trailing distance ratio x/d Tunnel blockage 2.9% Table 4. NASA test parachute design characteristics A forebody representative of the ExoMars entry module, a 70 sphere cone, was incorporated in the strut in front of the load cell. The parachute was attached to the forebody/strut via a swivel clevis, which was in turn mounted on a load cell. This was done to remove the effects of axial rotation of the parachute from the measurements of inflation and drag loads. The 6.88% scaled forebody had a diameter of m Dynamic data and high and low-speed video data were collected during the deployment and during each tunnel flow regime. To measure the parachute inflation force a Honeywell Model 75 load cell was used. It had a load range of 3 kn with a 0.1% full-scale accuracy. The data were recorded at a rate of 10 khz, which gave a resolution of approximately 250 measurements during the initial inflation phase. The test was also recorded with video from multiple angles using a combination of high-speed (2000 fps) and standard (30 fps) video cameras. A high-speed (2000 fps) Schlieren camera was positioned perpendicular to the flow in order to record the parachute profile and bow shock behaviour. The matrix of tests performed is presented in Table 5. The test matrix allowed for two tests at each dynamic pressure level. Test M p S T S V (m/s) q Re (x10 6 /m) (kpa) (K) (kpa) Table 5. Test deployment conditions The aim of the tests was to deploy the parachute at Mach number of 2.1 and cycle through a series of Mach numbers: 2.1, 2.0, 2.2, 2.3, 2.4, 2.5 and 2.6. The measured drag coefficients are shown superimposed on the ExoMars aerodynamic database and compared to the data measured at CNRC in Fig. 13. They are designated WT04. Figure 13 ExoMars DGB parachute database with CNRC and NASA Glenn data It can be seen that the drag coefficient data from NASA Glenn are markedly lower than those measured at CNRC. Moreover, the drag coefficient data showed much greater variability as shown in Figures 14 and 15. Figure 14. Drag coefficient for NASA Glenn testing at Mach 2.04 Figure 15. Drag coefficient for CNRC testing at Mach 2.0 The unsteadiness in the NASA Glenn data originated from large area oscillations of the parachute mouth not observed in the CNRC testing. Analysis of the Schlieren data showed some quite different flow features between the two tests. The CNRC tests showed a steady bow shock slightly disturbed by the forebody wake. The NASA Glenn tests showed shocks originating from the parachute

7 suspension lines causing large movements of the bow shock as shown in Fig. 16. Figure 16. Image from NASA Glenn testing showing line shock. The mechanism for the line shocks appears to be associated with spiked flow. For a spike ahead of a bluff body in supersonic flow, if the boundary layer is not fully turbulent then the separation point from the spike is mobile whereas if the boundary layer is fully turbulent then the separation point is stationary. This is analogous to the parachute line shocks. At sufficiently low Reynolds number the boundary layer on the suspension lines is not fully turbulent and the separation moves up and down the lines: forming conical shocks off the lines. If boundary layer is fully turbulent the separation is fixed as in the CNRC tests. 6. CONCLUSIONS ESA Aerodynamics of Download Systems permitted a very successful supersonic parachute test campaign at the CNRC Tri-sonic wind tunnel in Canada. Good quality drag was obtained over a range of Mach numbers and trailing distances. Drag coefficients were typically 40% higher than measured in ExoMars WT testing It demonstrated importance of testing at the correct Reynolds number Constant wake coupling was observed for a trailing distance of 7.5 forebody diameters. Only very occasional coupling occurred at trailing a distance 9.0 forebody which was selected to avoid this phenomenon Excellent agreement was achieved in an FSI rebuild of a Mach 2.0 test. 7. REFERENCES 2. Maynard, J. D., Aerodynamic Characteristics of Parachutes at Mach Numbers 1.6 to 3.0, NASA TN D- 752, Murrow, H. N., and McFall Jr., J. C., Some Test Results from the NASA Planetary Entry Parachute Program, Journal of Spacecraft and Rockets, Vol. 6, No. 5, 1969, pp Reichenau, D. E. A., Aerodynamic Characteristics of Disk-Gap-and Parachutes in the Wake of Viking Entry Forebodies at Mach Numbers from 0.2 to 2.6, AEDC-TR-72-78, Couch, L. M., Drag and Stability Characteristics of a Variety of Reefed and Unreefed Parachute Configurations at Mach 1.80 with an Empirical Correlation for Supersonic Mach Numbers, NASA TR R-429, Underwood, J. C., Development Testing of Disk- Gap-Band Parachutes for the Huygens Probe, AIAA Sengupta, A., Steltzner, A., Witkowski, A., Candler, G., and Pantano, C., Findings from the Supersonic Qualification Program of the Mars Science Laboratory Parachute System, 20th AIAA Aerodynamic Decelerator Systems Technology Conference and Seminar, Seattle, WA, 2009, AIAA Sengupta, A., Supersonic Testing of 0.8 m Disk Gap Band Parachutes in the Wake of a Subscale MSL Entry Vehicle, 20th AIAA Aerodynamic Decelerator Systems Technology Conference and Seminar, Seattle, WA, 2009, AIAA Lingard, J. S., Darley, M. G., Underwood, J. C., Simulation of Mars Supersonic Parachute Performance and Dynamics, 19th AIAA Aerodynamic Decelerator Systems Technology Conference and Seminar, Williamsburg, VA, 2007, AIAA Cruz, J. R. and Lingard, J. S., Aerodynamic Decelerators for Planetary Exploration: Past, Present and Future, AIAA