- Email: [email protected]
High velocity penetrators used a potential means for attaining core sample for airless solar system objects
Acta Astronautica 137 (2017) 274–286
Contents lists available at ScienceDirect
Acta Astronautica journal homepage: www.elsevier.com/locate/actaastro
High velocity penetrators used a potential means for attaining core sample for airless solar system objects
MARK
⁎
R.M. Winglee , C. Truitt1, R. Shibata Department Earth and Space Sciences, University of Washington, Seattle, WA 98195-1310, United States
A R T I C L E I N F O
A BS T RAC T
Keywords: Penetrator Sample return Subsurface sample Asteroids
Sample return missions offer a greater science yield when compared to missions that only employ in situ or remote sensing observations. Such missions have high ΔV requirements, and the return yields to date have been typically only of a few grams for robotic missions. Planetary penetrators offer an alternative that significantly reduce a mission's ΔV, increase sample yields, and allow for the collection of subsurface materials. The following details the design, development, and testing of penetrator/sampler technology capable of surviving supersonic impact velocities that would enable the collection of a solid core of geologic materials, without the need for any drilling equipment,,thereby reducing the overall mass and propellant budget. It is shown through both modeling and field testing that penetrators at speeds between 300 and 600 m/s (~Mach 1–2) can penetrate into the ground to depths of 1–2 m with overall structural integrity maintained. The first flight tests demonstrated the potential for survivability at these speeds. The second flight series demonstrated core sample collection with partial ejection of the sample return canister. The 3rd flight series demonstrated self-ejection of the sample return system fully intact and with the core retaining the full stratigraphy of the rock bed. The recovered sample also shows the survivability of macro-organic structures. Possible mechanisms for the recovery of the ejected core sample are also discussed.
1. Introduction The National Research Council [1] advised that Discovery and New Frontier class missions should play a critical role over the next decade in the study of primitive bodies given their relatively low cost and applicability for destinations both in the inner solar system, and beyond the asteroid belt. These missions will provide vital contributions in understanding the basic building blocks that created our world, as well as assessing potential hazards impacting bodies represent to our biosphere. Sample return missions from primitive bodies are excellent candidates for NASA's Discovery and New Frontier Programs given their potential for a high science yield while requiring only a fraction of the investment typical of a Flagship mission, and could supply materials long demanded by the science community for furthering our study of the solar system. To date more than 150,000 asteroids have been identified in the main belt alone [2], but difficulties in collecting samples has resulted in returning limited material. Scott Sandford [3], a team member of the OSIRIS-REx mission speaking at an Exploration Science Forum, highlighted some of the advantages sample return missions have over in-situ and remote
⁎
1
sensing methods, including: an increase in the quality of the data produced through the application of technology that was not available during the spacecraft's development; returned samples become research resources for both present and future scientists; sample analysis is not limited by design constraints of the spacecraft. He [3] also spoke to their potential to reduce research limitations that result from poor assumptions saying, “…if you decide to measure A, and you go there with your A measuring machine… it is possible that the main thing that you will learn… is maybe you should have measured B, and now you need new spacecraft and another mission”. The Apollo program, most noted for being the first exploration series to land humans on another rocky body, has provided the largest quantity of returned sample to date. Between the summer of 1969 and winter of 1972, Apollo astronauts gathered and returned more than 300 kg of lunar material. While NASA employed human beings to collect material from the Moon, the Soviet's Luna program became the first automated system to return samples collected from a minor body, albeit in much smaller quantities. The manned Apollo missions were able to bring back a few tens of kg each missions while the robotic Luna programs only brought back a fraction of a kilogram each [4]. However,
Corresponding author. E-mail addresses: [email protected] (R.M. Winglee), [email protected] (C. Truitt). Current address: Jet Propulsion Laboratory, 4800 Oak Grove Drive M/S 352H, Pasadena, CA 91109, United States
http://dx.doi.org/10.1016/j.actaastro.2017.04.022 Received 27 December 2016; Accepted 19 April 2017 Available online 21 April 2017 0094-5765/ © 2017 IAA. Published by Elsevier Ltd. All rights reserved.
Acta Astronautica 137 (2017) 274–286
R.M. Winglee et al.
S SRC TAGSAM Vs α ΔV
Nomenclature A ANSYS D Ks N
penetrator cross-sectional area in m2 Analysis System computer modeling software depth of penetration in meters, scaling factor dependent on mass penetrator nose coefficient
empirical target constant sample Return Container touch-And-Go Sample Acquisition Mechanism impact velocity in m/s an empirical constant equal to 1.75×10−5 Change in velocity
subsurface sample will be undertaken by Hayabusa 2 [11] which was launched in 2014 and will arrive at asteroid 162173 Ryugu in 2018. Hayabusa 2 will deploy a 2.5 kg kinetic impactor propelled by a 4.5 kg shape charge to impact the surface of asteroid 1999 JU3 at hypervelocity speed of 2 km/s in 2019, creating an artificial crater and collecting samples from greater depths than its predecessor. Hypervelocity penetrators has been used in two experiments to eject subsurface material into space where it is then examined using a remote spacecraft. The first was Deep Impact which in 2005 used a 370 kg impactor with an incident speed of 10.5 km/s to generate ejecta comet Tempel 1 [12]. A large increase in organic material as well as carbon dioxide was measured in the ejecta [12]. The Lunar Crater Observation and Sensing Satellite (LCROSS) in 2009 used the upper stage booster to impact a crater in permanent shadow at a speed of about 2.5 km/s. LCROSS was then able to successfully detect water in the ejecta plume [13]. Low velocity penetrators at speeds less than about 30 m/s have been used in variety of applications [14]. These include the dropping of instrument in remote areas to bunker busters. Sample return from icy objects has been proposed using penetrators with impact speeds of about ~30 m/s [15]. In this paper we examined an intermediary system for sample return which utilizes high velocity (200–800 m/s) impactors or penetrators to generate a core sample of subsurface material that is selfejected from the impact region. The concept for operations is described in Section 2 with the design parameters of the penetrator given in Section 3. The principle is very similar to the ice penetrator of Lorenz et al. [15] but the hardness of the rock is much higher that the ice in their studies, so that much higher impact velocities are required, and the material response to the impact is very different. Because of these differences the design and testing of the system is significantly modifiedfrom that of an ice penetrator. The data from the first few flights are given in Section 4 and show results for impacts at angles away from perpendicular impacts. Behavior of the penetrator and sample collected from a Mach 2 impact into sandstone is given in Section 5. Section 6 shows the results for a self-ejected core sample from a near-final version of the system. It is shown that the sample attained retains the overall stratigraphy of the rock, and macroorganisms retain intact in the recovered sample. A summary of results is given in Section 7.
manned missions are vastly more expensive, and no humans have traveled beyond the orbit of the Moon so that the issue of making significant sample returns in cost effective manner still remains an important issue. The Apollo and Luna programs both employed soft-landing techniques, requiring the expenditure of considerable amounts of fuel to safely arrive on the Moon's surface, and additional propellant to ascend once collection efforts were completed. These maneuvers significantly increase a mission's ΔV budget, and require flawless execution to ensure the survival of sensitive instrumentation. NASA's first sample return efforts after Apollo 17 was the Genesis mission [5], designed to characterize and sample the solar wind using a halo orbit around Lagrange point 1. The spacecraft gathered samples using collector arrays from late 2001 to spring of 2004, but the failure of its parachutes to deploy during Earth re-entry in September 2004 caused the Sample Return Capsule (SRC) [6] to impact the landing zone at more than 86 m/s, resulting in the possible contamination of the sample. Despite this non-ideal landing Genesis was able to successfully capture noble gases and isolate important isotopes. In contrast, the Stardust spacecraft [7] sent to collect samples from Comet 81 P/Wild 2 enjoyed a much better success. Launched in early 1999, the Stardust mission first collected samples of interstellar dust in 2000, and again in 2002 following an Earth gravity assist trajectory. The mission's flight team performed a close flyby of asteroid 5535 Anne Frank, using the opportunity as an engineering test of ground and spacecraft operations prior to intercepting Comet 81 P/Wild 2 in 2004, where it flew through the halo of gases and dust at the head of the comet. The spacecraft used an Aerogel filled grid to collect materials thought to pre-date the birth of the Sun, and their successful return to Earth in 2006 has provided new insights into our solar system [8]. There was also some contamination of the sample but signatures of returned cometary organics could be distinguished from the terrestrial contamination though there were concerns that the hypervelocity collection may have modified some of the organics. Robotic sampling for the actual surface of an asteroid was achieved by JAXA's Hayabusa [9] using a touch-and-go approach on asteroid 25143 Itokawa in 2005. In this scenario, sample retrieval is conducted as the spacecraft briefly contacts the surface of the sampling target, collecting a few grams of surface regolith before moving on. The touchand-go method avoids the problems of attaching a spacecraft to an asteroid and helps to reduce a mission's ΔV budget. In the case of Hayabusa, a tantalum pellet is fired at 300 m/s to produce ejecta which is then collected inside its sampling mechanism. Hayabusa demonstrated that touch-and-go sampling is possible, despite having hardlanded on Itokawa during a sampling attempt and collecting less than 1 mg of material. NASA's OSIRIS-REx spacecraft [10] will employ a similar touchand-go approach in 2019 to gather materials from asteroid 101955 Bennu using its Touch-And-Go Sample Acquisition Mechanism (TAGSAM). During contact with the surface, the TAGSAM will use a burst of nitrogen to blow regolith through a collecting sieve, and lab testing indicates that the method is capable of gathering more than 60 g of material. These two approaches are only able to attain surface samples which have the potential to be strongly modified compared to subsurface material due to its interaction with the solar wind. An attempt for a
2. High velocity penetrators for core rock sampling The concept of sample return by a high velocity penetrator is based on the fact that there is typically a high velocity difference (ΔV) between the target and the approaching spacecraft. A penetrator system can thereby use this kinetic energy for the impact without requiring any propellant. The mission concept for the release of the penetrator is shown in Fig. 1. The term high-speed refers to the impact velocity range of between 0.3–0.9 km/s. This velocity range is below hypervelocity speeds where chemical alterations can be produced but is sufficiently high energy to cause the breakup of most rock types. Depending on the diameter and speed of the penetrator and the hardness of the rock, the penetrator is expected to punch into the asteroid to a depth of 1–2 m. The sample return canister (SRC) lies within the penetrator which has feed ports on its tip to allow material 275
Acta Astronautica 137 (2017) 274–286
R.M. Winglee et al.
much each easier. In both case, radar and optical location finders would be needed but since this equipment is often included on solar system encounters, mass additions would be minimal. The actual collection of the sample could be achieved for example by an electromagnet. The bottom line is that the recovery would not require any new technologies and our focused here is to demonstrate that a self-ejecting core sample can be achieved by a high speed penetrator. 3. The penetrator dynamics – empirical and numerical simulations The ability of high speed penetrators to punch into rock has been documented by an empirical formulation derived from defense applications of penetrators [16] and, for the conditions in space applications, is given by
Fig. 1. Artist impression of the use of a high-speed penetrator for the collection and ejection of sun-surface core rock sample.
to move up into the center collection tube which feeds into the sample return canister. The tip of the penetrator is sufficiently hard that it is able to break up variable rock/soil strata. While there is some possibility of chemical metamorphism at the interface the cutting edge of the feed port and the sample intact, the bulk of the sample is expected to be unmodified. This broken up material then flows into the feed ports and into the central flow channel and eventually into the sample return canister. This material is under pressure so that its momentum, and/or a spring ejection system mechanism, forces the sample return canister to be ejected back above the surface while the penetrator itself remains in the ground. After it is ejected from the impact site, the SRC has to be collected by the mother spacecraft. There are two potential options available depending on the size of the object. For very small solar system objects (i.e. small asteroids) where the escape velocity is less than a few tens of m/s, the ejection of the SRC can be designed so that it reaches escape velocity. In this case, the mothership would have to locate the SRC on approach to the asteroid, rendezvous and pickup up the sample. The ΔV requirements are probably the lowest for this configuration but the potential for losing the SRC is probably the highest. The alternative is to have the SRC fall back onto target where it would be collected by the mother spacecraft basically by a touch-and-go encounter. The ΔV requirements are higher and there is a larger risk to the spacecraft due to landing requirements, but location finding of the SRC would be
D~αKs SN(m / A)0.7Vs
(1)
where A = penetrator cross-sectional area in m2 D = depth of penetration in meters, Ks = scaling factor dependent on mass and for the proposed applications has the form Ks =0.46×(m)0.15, m is penetrator mass in kg, N is penetrator nose coefficient which for a spherical projectile is ~ 0.5 while for a pointed projectile (and the proposed system) N is ~ 1.2, S = empirical target constant equal to ~ 0.7 for medium strength rock to 1.4 for low strength rock, Vs = impact velocity in m/s. α= an empirical constant equal to 1.75×10−5. For the speeds of 0.3–0.9 km/s for the present system, and for a 10–15 cm diameter penetrators the expected depth of penetration is about 1–2 m, dependent on rock type. At this depth the system would be able to sample material that has not been modified by long term exposure to the solar wind. In order to validate the empirical formation, we also use detailed numerical simulations via ANSYS – Autodyn. This commercial product
Fig. 2. Ejecta patterns for four different angle relative to the surface. There is an increase in the ejecta angle but the central cone ejecta has approximately the same form for all for cases.
276
Acta Astronautica 137 (2017) 274–286
R.M. Winglee et al.
can still perform for oblique angles of impact. The reason is that the moment of inertia of the penetrator is large for the high speed impacts being considered here that it breaks the rock as opposed to being deflected by the rock. In addition at high speeds, the ground cannot move significantly before the full width of the penetrator passes into solid material. Once it does so, the penetrator performance is completely independent of the surface properties. These results are also consistent with other studies for penetrator performance as a function of impact angle [17]. By the same arguments, the penetrator performance is also independent of the roughness of the surface of the target. Specifically, if the surface has boulders much larger than the penetrator diameter then the penetrator will just see it as an incline surface as in Fig. 2 and cut through it. For pebbles much smaller than the penetrator, then their inertia is too small to affect the trajectory and again the penetrator will cut straight through this material. Only in the unlikely case of the penetrator striking a cobble of similar diameter of
is designed for simulating the response of materials to short duration severe loading from impacts and high pressure explosions and is an important total for aerospace and defense applications. The ANSYS package also allows for element destruction and separation from the primary model associated with cracking and breakup of the impacted material. During impact, the stresses on the internal components are also calculated for possible failure modes. For the present application, different penetrator designs and impact hardness can be easily incorporated using tabulated data for the shock properties, Johnson Cook strength and failure models for different materials. As an example, Fig. 2 shows the performance of a short hollow cylinder penetrator going into medium strength material at 4 different impact angles using a grid of ~106 grid points. The measured depth of penetrator for this 100 g projectile was ~ 25 cm for this example which is within 20% of the depth predicted by the empirical formulation. The other important result in Fig. 2 is that the interaction is not strongly dependent on the angle of impact and that the system
Fig. 3. (a) Initial flight system using inline rocket motors, (b) sample collection system, and (c) final flight system using external booster rockets.
277
Acta Astronautica 137 (2017) 274–286
R.M. Winglee et al.
SRC (Fig. 3b). To bring the penetrator up to speed, a two-stage rocket system was used. A booster rocket motor lifted the system to 1000–1500 m. The rocket was then allowed to free fall for about 9 s when it would become aerodynamically stable in a downward trajectory. At this time the booster section of the rocket is ejected and the sustainer motors lit, bringing the system up to Mach 1–2 (depending on the size of the motors used) just prior to impact with the ground. In the initial configuration the sustainer motors were in-line with the long axis of the penetrator. Unfortunately, the motor could not be ejected in time before the impact occurred. For this reason, the later system used a series of 8 outboard motors for the sustainer stage. These motors impacted with the penetrator but because they are on the outside they could not influence the SRC performance, particularly with the large diameter of the rocket body. The final issue for the design of the penetrator was the positioning on the feed ports. The three configurations examine for this effort are shown in Fig. 4. The flow through the different configurations was modelled using SolidWorks’ FloXpress Analysis Wizard. Because this software cannot model the flow of a solid through feed ports, water was selected as the medium. This modeling is meant to only show qualitatively the difference in flow patterns and the derived speeds should only be considered as relative measures and not absolute speeds. The first configuration (Fig. 4a) where the feed ports are to the side of nose cone has the potential advantage of deeper penetration into the target material but would deflect most of the surface material since the tip will come in contact with the surface before the feed ports make contact. If the surface material (perhaps modified by space weathering such as the solar wind) is not required for sample collection then this might not be a significant loss if depth is the primary objective. The flow of material into the feed chimney is the slowest of the three configurations which translate to a reduced energy for the sample ejection. The second configuration (Fig. 4b) which is called the hollow point has the feed point through the tip of the nose cone. The hollow point flow pattern shows substantial deceleration on the outside rim of the hollow point which contributes to mushrooming. If mushrooming occurs then the depth of penetration will be reduced due to the
the penetrator is there a potential issue. However, if the mass of the penetrator exceeds that of the cobble (as per design), then again the penetrator inertia will stop it from undertaking a significant deviation and again, the surface roughness would not be considered an important factor in the operation of the penetrator. Results to this effect were observed for example during the near horizontal impact described in the following. The design of the penetrator derived from these simulations is shown in Fig. 3. The length of the penetrator is set to be approximately the same as the expected depth of penetration at about 1.66 m. The diameter was set at 15 cm, mainly due to the rocket motor requirements to bring the system up to speed. In space applications, the modeling indicated that the diameter could easily be reduced to 10 cm with little change in performance from that described below. The nose cone was made initially of Al 6061 with a mass of 1–3 kg. These nose cones work well for soft material such as playa, which was encountered in the first tests. As the testing progressed into harder material A2 tool steel was used for the nose cone with a mass of 6–6.5 kg. This material was still relatively easy to machine and had the ability to be hardened to a Rockwell hardness ~62c. Feed ports through the nose cone enable material from the impact area to move into the central feed chimney where it would be collected within the sample return canister. The sample return canister (SRC) was initially made of 1.5 mm thick Al which proved to have insufficient strength to withstand the build-up of pressure from the upflowing rock material. In the later tests the Al SRC was replaced by 3 mm thick steel SRC and this system was able collect sample and be ejected with minimal damage. The diameter of the SRC was 4 cm with a total length of 40 cm. The feed chimney was supported by a honeycombed aluminum structures with carbon fiber strengtheners. This material could withstand forces of more than 1 kN/cm2 yet have a sufficiently low density that the mass of the chimney and body tube was less than the mass of the nose cone. This support structure prevents the chimney from collapsing during impact and at the same time prevents the rock in the impact site from fully compressing the sides of the penetrator and restricting the diameter of the feed chimney. In the later versions of the system a shocking spring was added to the support of the SRC which provided additional upward force after impact to ensure ejection of the
Fig. 4. Relative flow model for three feed port designs.
278
Acta Astronautica 137 (2017) 274–286
R.M. Winglee et al.
tion yields the highest flow rates into the feed chimney and is thus the most efficient for collecting rock sample. The negative aspect of this configuration is that it is the structurally weakest though as demonstrated below still able to remain intact at supersonic impacts into sandstone. Another aspect of this system is that there is potential for some mixing of material moving through the side ports with material moving through the tip port, which would make the stratigraphy collected a little more complex to interpret. Despite the above modeling there is still some certainty in the optimal nosecone shape as there are still several uncertainty factors. Low speed penetrators tend to have blunt nose cones, which yields greater stability as the penetrator moves through the rock. For the high speed penetrators this stability issue appears to be outweighed by the braking issue. Field testing of the blunt nose cone figuration showed that the deceleration appeared to be too fast and the back end of the penetrator tended to try and push pass the nose cone causing the penetrator to buckle. For the sharp nosecone used here, bending was also seen to occur when the nose cone was made of the same material (Al) as the rear of the rocket. This bending is basically due to the way the different components of the rocket are decelerating and whether the rock is moving out of the way of the penetrator or vice versa. Having the density of the nosecone higher than the rear of the penetrator means that side friction can slow the rear at the same rate as the front. Similarly, a high density nose cone forces the rock material to move in response to the impact as opposed to the nose cone being deflected by the rock. Thus, the final design for the penetrator consisted of a tool-hardened steel nose cone, and Al body. Under these conditions, minimum bending of the rocket during impacts was seen in the field testing below. Fig. 5. Penetrator impacted at 150 m/s into playa embedding to ~1.22 m at 30° from vertical.
4. Initial field testing – impact angle dependencies One of the unknowns for contact with solar system objects is the actual surface conditions the penetrator would encounter since in general most objects do not have high resolution surface data until the arrival of the spacecraft. For the present application this translates to whether the penetrator will perform as planned when the angle of impact is significantly different from vertical. The simulation results in Fig. 2 indicate that the angle of impact is not a significant factor, provided the tip impacts the target before the sides. Thus, the minimum impact angle appears to be approximately set by the nose cone angle. This result was confirmed in initial field testing of the penetrator. As an example of a non-perpendicular impact, Fig. 5 shows the results of one of the earlier design penetrators flown between 2013 that impacted playa at about 30° from perpendicular at 150 m/s. The depth of penetration was approximately 1.3 m and the system showed essentially no deformation. Fig. 6 shows the results from a flight in 2014 that struck clay/soil for a near tangential impact. Similar to the simulations in Fig. 2 there is increased ejecta downslope with the length embedded into the ground being about the same as a near normal impact. The system in these test shot remained intact as well except for a little bending of the nose cone. These results show that the system can handle arbitrary impact angle. The main parameter that controls performance is the requirement that the velocity vector be aligned with the long axis of the penetrator. This is achieved in our terrestrial applications using fins as passive stabilization. In Fig. 6 due to non-simultaneous lighting of the booster motors there is a misalignment between the two vectors by about 5°. Nevertheless, the system is able to handle this difference because of the internal reinforcements of the system to absorb the energy of the hard impact. In space applications, this passive stabilization can be achieved by spinning up the penetrator at the time of release to the same effect, and then have rods released just prior to impact to slow the spin rate.
Fig. 6. Near tangential impact of our penetrator with the last frame prior to impact (camera is level with the ground). There is a larger ejecta pattern but the system remains intact and penetration depth is about the same as in Fig. 5.
increase in cross-section increasing the drag as the nose cone moves through the rock. The flow speed is higher than the side port configuration. Moreover, all material that is encountered by the nose cone moves up the feed port to be collected by the SRC. The third configuration (Fig. 4c) is the hybrid feed port which has feed ports at both the nose cone tip and along it sides. The configura279
Acta Astronautica 137 (2017) 274–286
R.M. Winglee et al.
Fig. 7. Successful ignition, launch and ignition of the clustered two-stage penetrator.
Fig. 8. (a) Images of the supersonic penetrator impact site, (b) students marking some of the larger debris items, (c) map showing the positioning of some of the debris and (d)-€ show the recovered top of the SRC.
astray due to a misfire of a motor or fin damage, the rocket could not reach any structure. The launch team was 2500 ft from the launch area and the rocket fired at 15 degrees away from the fire control area to ensure safety of the ground crew. Launches were typically performed over a 1–2 day period with recovery occurring in subsequent days after launches were completed. The launch of this system and the downward powered flight is shown in Fig. 7. The first high speed impact of the steel tip penetrator that demonstrated survivability was achieved in March 2015. This system had the hybrid nose cone shown in Fig. 4c. It was mated to an
5. Field testing: sandstone impact With the rocket configuration of Fig. 3b, we were able to attain supersonic impacts starting in December 2014 on a large ranch near Ione CA. The test area on the ranch was centered such that the nearest building was 1.5 miles away from the target area. An FAA waiver was attained for the flights up to 5000 ft above ground level, with appropriate notification going out prior to and after the launch period to relevant air route traffic control centers. The rocket boost system had a predicted maximum altitude of 3000 ft. Thus, even if the rocket went 280
Acta Astronautica 137 (2017) 274–286
R.M. Winglee et al.
Fig. 9. Trenching of the penetrator from the impact site in Fig. 7.
Fig. 10. Images of the recovered front section of the penetrator including nose cone and feed ports.
outwards. While not the expected outcome for the SRC, it nevertheless demonstrated that there is energy within the center feed system to produce self-ejection of material. The impact crater showed less surface disruption itself and was only about 1 m in radius (Fig. 9). This is an important result of the system in that only a small area is disturbed even for these high speed impacts. The reason for this is that the system is long and slender, and there are no explosives involved. There was actually no sign of the penetrator system in the impact crater, except for the red shock cord from the top of the SRC (Fig. 9a). The lack of penetrator components at the top of the crater was a little perplexing given that in all previous shots some part of the penetrator was evident in the impact crater. Once the top few inches of soil was removed (Fig. 9b), the underlying sandstone was revealed and the impact shaft contained a large amount of backfill. Loose backfill debris was removed revealing the top of the impact shaft, 10 cm in diameter, with clearly observable slickensides (Fig. 9b). The
aluminum extension resulting in an assembly to minimize the total mass of the penetrator. The penetrator was fitted with a full array of eight, outboard motors to provide the highest impact velocity following the post-apogee free fall. Slight ignition timing delays in the boost stage resulted in a launch angle of ~30°, reducing the ideal apogee altitude to ~1000 m. While this reduced the planned free fall time, the rocket remained stable during the ignition of the second stage and produced a very loud sonic boom, immediately followed by the sound of the impact. Impact velocity was estimated to be over 600 m/s, embedding to a depth of 1.6 m into solid sandstone. The impact was very energetic, and debris from the exterior cluster motors was spread across a broad swath of a few tens of meters (Fig. 8). The farthest piece of debris that was recovered was the top of the SRC which was found about 83 m from the impact site. It was coated in a film of top soil, which presumably was deposited during the impact, and the pressure within the tube blowing the top of the canister 281
Acta Astronautica 137 (2017) 274–286
R.M. Winglee et al.
Fig. 11. The core sample was exposed after removal of the encasing material.
digging out caused the nose cone to break into two separate pieces associated with the brittle fracturing. This material failure was likely the result of over-hardening of the steel (Rockwell hardness ~62c), instead of optimizing hardness and strength, resulting in a hard but brittle tip that still provided effective penetration and sample collection. The most important observation of the nose cone was that all feed ports were solidly packed with sandstone from the impact region. This sandstone is not fragmented but rather appears to have moved into the feed ports under cataclastic flow without fracturing. The cutaway of the penetrator in Fig. 11 shows that the sandstone moved up into the central feed chimney coherently for the first 60 cm from the tip of the nose cone. Note also that the diameter of the sandstone in the chimney increases with distance from the feed ports, which provides some indication of the pressures that developed on the sample from the impact. The encounter with the SRC (top left of Fig. 11) shows turbulent behavior, with the SRC significantly folded and distorted from its original cylindrical shape. This turbulent behavior is thought to arise from the loss of the top of the SRC which was found above the ground and the interaction of the cataclastic sandstone with the bottom baffle of the SRC. Without the back plate there is no buildup of pressure to decelerate the SRC. As a result, it is still moving forward when it encounters the cataclastic sandstone which then causes two-stream turbulent behavior to occur. The impact definitely produces physical metamorphism of the rock. The measured density of the undisturbed material was measured as 1.7 g/cm3. The density of the sample core was measured at 2.1 g/cm3. Thus, there is about a 25% increase in density due to the filling of voids in the rock from the compressive forces from the impact. Examination of thin sections of the recovered sample the bulk of the sample was not chemically, and we did not detect any signs of modifications to the minerals relative to the undisturbed materials. The main differences occurred along the contact surface of the cutting edge of the feed ports. There was some increased ion content along the surfaces that were in contact with the central feed port that appeared as a red ring in crosssection through the core sample. We were unable to determine whether this enhancement came from the steel feed port itself or from some chemical metamorphism from the impact.
Fig. 12. The epoxied SRC with its tip in front of the main nose cone and the penetrator system ready for flight.
top of the penetrator was not found until about 0.5 m of top material was removed (Fig. 9c). Earlier impact models suggested that material proximal to the impact shaft would experience varying levels of fracturing. There was some fracturing and examples of over-turned strata near the surface of the impact site, but these effects were limited to the first few inches and were not observed at greater depths. Instead, the rocket created a shaft with a diameter that matched the maximum diameter of the metal portions of the nose cone assembly of 10 cm (where as the main body of the penetrator had a 15 cm diameter). The extracted penetrator is shown in Fig. 10, and is basically intact except for the section holding the outboard motors, which was designed to strip away on impact. The penetrator shows no bending as it penetrated the sandstone, so that the energy absorbing section fulfilled its purpose of inhibiting compression along the axis of the penetrator; however, the diameter - particularly around the midsection - was reduced from its initial 15 cm to about 10 cm which was the diameter of the metal components of the nose cone. The nose cone itself experienced some brittle fracturing, particularly in the tip. The nose cone was intact in the crater, however stresses from the
6. Field testing: SRC ejection One of the important conclusions from this series of flight tests is that while material is able to move up the feed ports, it never reaches the height of the surface, which makes self-ejection problematic. In all cases it reaches only a height of about 0.3–0.5 of the penetration depth. 282
Acta Astronautica 137 (2017) 274–286
R.M. Winglee et al.
Fig. 13. Launch of the 3rd generation penetrator its impact and self-ejection of the SRC.
Fig. 14. A small low arc feature develops about 1 s after the impact, believed to be entrained smoke from the ejected SRC.
To overcome this problem for the final set of field tests, the penetrator was redesigned in an effort to ensure full ejection of the SRC. The SRC itself was fabricated from steel, and was moved fully forward in the penetrator so that it extended beyond the end of the machined nose cone (Fig. 12). In this forward position it would experience the greatest upward pressure but the steel construction would prevent the catastrophic failure seen in the Al SRC. A shock-absorber spring was also added to aid in the ejection of the sample. Moving the sample return canister forward increased the risk that it would bend on impact, but the weight and energy of the nose cone ensured that even if such bending were to occur the SRC would still be forced back into the
central feed column. Because of the loose fit of the SRC, it was epoxied in place and its feed port was covered for flight (Fig. 12); the epoxy was sufficiently robust enough to survive flight, but the contact joint would shatter on impact allowing the SRC to move through the center column of the penetrator without binding. The launch occurred in December 2015 at Ione CA and its flight and impact is shown in Fig. 13. During the field test, only four of the eight outboard cluster motor ignited (probably due to the igniters coming loose from vibration/g-forces during flight) so that the impact speed was only about 330 m/s. During flight the smoke grains that aid the visual tracking of the rocket were only weakly
283
Acta Astronautica 137 (2017) 274–286
R.M. Winglee et al.
Fig. 15. Trenching of the impact site in Figs. 13 and 14 and cutaway of the canister showing intact stratigraphy.
burning, but on impact they flared up to produce the enhanced smoke plume that last several seconds after impact. On inspection of the impact site (Fig. 13e), the SRC was found approximately 2 m from the impact site, thereby proving self-ejection of the sample is possible. Because of the slower speed, the booster motors are seen protruding from the impact side, though they are no longer attached to the main penetrator body. Close inspection of the video of the penetrator flight was unable to resolve the ejection of the SRC. However, the smoke trail as shown in Fig. 14 shows the development of a low (~3 m) hook feature that curves in the same direction as the location of the ejected SRC. The appearance of this feature is about 1 s after impact and fitting a parabolic trajectory along this feature yields an impact at the location of the SRC. It is believed that this smoke feature is associated with smoke being entrained with the SRC as it is ejected. Fitting a parabolic trajectory with a 1 s to apogee and 1 s back down to the ground at 2 m and a height above ground of 3 m and 1 m below ground, yields an ejection velocity of about 10 m/s at an angle of about 10° from vertical. At this speed, the SRC could be ejected to significant height above the surface of any small bodied asteroid where the gravitation forces are very much weaker. The trenching around the impact site is shown in Fig. 15a and shows top soil and clay to about 0.3 m. Below this material, there was a region of talc and other sedimentary material. As noted from the impact site the SRC was recovered intact as seen in Fig. 15b. The end of the SRC was deliberately slotted so as to fold inwards during the impact to produce self-sealing of the SRC. This folding was partially achieved as seen in Fig. 15c, though a couple of folds appear to have broken off and entered the SRC. Within the cutaway of the SRC (Fig. 15d) the ground stratigraphy is seen in to have been retained, with soil and clay near the top of the canister and talc (yellow material) at the bottom of the SRC. Close ups of the material in the top of the SRC are shown in Fig. 16. At the very top of the canister small pebbles are seen mixed in with top soil. The concentration of the pebbles is larger than seen in the top soil, indicating some loss of fine material during impact. This loss of material is due to holes in the SRC which is required to vent air trapped in the SRC during the impact. Since there is no air in space, vent holes will not be required in planetary applications and this loss of top material is not expected. Another important find within the SRC is the presence of macroscopic organic material associated with the grass in the impact region. This result is important as it demonstrates that organics can be preserved during the impact. This result also demonstrates that the heating produced by the impact is insufficient to destroy macroorganics and also means that chemical metamorphism is unlikely to occur during the impact, except for possibly the contact zone at the
cutting edge of the feed port The red ring of material that was seen in the Mach 2 impact discussed in the previous section was not see in the present sample. 7. Discussion Field testing results demonstrated the potential application of ground penetrators for sample return, but a number of parameters remain uncertain. Switching the SRC design to steel instead of the original aluminum and carbon fiber used in earlier evolutions, created a more robust container, and moving the SRC fully forward in the penetrator allowed the SRC ejection to begin before significant structural failures farther aft occurred in the penetrator during impact. This was demonstrated by the two tests discussed, but moving the SRC fully forward introduced a new limitation – sampling depth. In the earlier design, sampling began once the feed ports on the nose cone had opened after impact, allowing sampling of material from more than 1 m in depth. Moving the SRC forward resulted in sample collection from less than a meter in depth since the SRC was ejected before embedding was complete. Sampling depth might be improved by the development of a feed port cover that is progressively abraded away during the initial contact such that sampling does not begin at the moment of impact. Another large uncertainty is the potential for metamorphic effects on the sample materials during impact. Terrestrial metamorphic processes are due to chemical reactions, and extreme temperatures and pressures generally occurring over geologic timescales. Chemical and temperature metamorphism over the bulk of the collected sample is unlikely to occur because the time scale of the impact at a few milliseconds is insufficient to produce bulk heating of the sample. However, there is the possibility that there is some chance of chemical metamorphism at the surface which contacts the feed ports. Contact pressure at the first contact of the nosecone with the sandstone was estimated to be about 90 MPa, and about 40 MPa for the test during which the SRC was fully ejected, which are sufficient to produce cataclastic flow of soft rock. However, the sandstone core sample exhibited an unusual orange/red ring that concentrically ran throughout the length of the core. Thin section analysis of the sandstone samples are ongoing so no definitive identification of the orange ring has been made yet, but we speculate that the ring is either the result of steel in the machined tip being worn away during the impact, or from some type of chemical metamorphism along the contact surface. The deceleration profile during impact, and hence the resulting pressures created in the sample material, are governed by two major factors: impact velocity, and target density. Since impact velocity can be adjusted to minimize metamorphic potential, the largest uncertainty becomes the density of the targeted body. As material density 284
Acta Astronautica 137 (2017) 274–286
R.M. Winglee et al.
Fig. 16. Close up of the sample showing survivability of woody/grassy material with a thumb tack for scale.
8. Conclusions
increases, acceleration increases as the penetration depth decreases; with lower material density the penetration depth increases creating a longer acceleration profile and as such, lower pressures. The density of a primitive body selected for impact sampling becomes critical in optimizing the impact velocity of the penetrator so that a maximum embedding depth can be reached without the risk of serious metamorphic effects. This optimization will remain problematic for a time until a better characterization of the densities of primitive bodies has been made.
Sample return missions offer the greatest science return on investment given the wider array of analytical methods that can be applied in terrestrial laboratories. Human retrieval during the Apollo era has brought back the largest amount of non-terrestrial material but having a sustained human presence beyond low Earth orbit remains problematic. As a result, robotic missions are the present mainstay for sample return missions, and at this time these samples have been restricted to the surface of small solar system objects. In attaining these samples 285
Acta Astronautica 137 (2017) 274–286
R.M. Winglee et al.
[5] D.S. Burnett, et al., The genesis discovery mission: return of solar matter to Earth, Space Sci. Rev. 105 (2003) 509. [6] D.S. Burnett, K.M. McNamara, A. Jurewicz, D. Woolum, Contamination on anodized aluminum components on the Genesis Science Canister, Lunar Planetary Science Conference, 20050166923, 2005. [7] D.E. Brownlee, et al., Stardust: comet and interstellar dust sample return mission, J. Geophys Res. Planets 108 (2003). http://dx.doi.org/10.1019/2003JE002087. [8] J.E.P. Matzel, et al., Constraints on the formation age of cometary material from the NASA Stardust mission, Science 328 (2010) 483. http://dx.doi.org/10.1126/ science.1184741. [9] H. Yano, T. Kubota, H. Miyamoto, T. Okada, D. Scheeres, Y. Takagi, K. Yoshida, M. Abe, S. Abe, O. Barnouin-Jha, A. Fujiwara, S. Hasegawa, t. Hashimoto, M. Ishiguro, M. Kato, J. Kawaguchi, T. Mukai, J. Saito, S. Sasaki, M. Yoshikawa, Touchdown of the Hayabusa spacecraft at the Muses Sea on Itokawa, Sci. New Ser. 312 (5778) (2006) 1350–1353. [10] E. Beshore, et al., The OSIRIS-Rex asteroid sample return mission, IEEE Aerospace Conference doi: http://dx.doi.org/10.1109/AERO.2015.7118989, 2015. [11] Y. Tsuda, M. Yoshikawa, M. Abe, H. Minamino, S. Nakazama, System design of the Hayabusa 2 – asteroid sample return missions to 1999 Ju 3, Acta Astronaut. 91 356 (2013). [12] M.F. Hearn, et al., Deep impact: excavating Comet Tempel 1, Science 310 (2005) 258. http://dx.doi.org/10.1126/science.1118923. [13] A. Colaprete, R.C. Elphic, J. Heldmann, K. Ennico, An Overview of the Lunar Crater Observation and Sensing Satelitte (LCROSS), Space Sci. Rev. 167 (2012) 3. [14] R. Lorenz, Planetary Penetrators: their origin, history and future, Adv. Space Res. 48 (2011) 403. [15] R.D. Lorenz, W.V. Boyton, C.F. Turner, Demonstation of Comment Sample Acquisition by Penetrator, ESA Proceedings, IAA Conference on Low-Cost Planetary Exploration, SP-542, 387, 2003. [16] National Research Council. Effects of Nuclear Earth-penetrator and Other Weapons, 2005, The National Academies Press Washington, D.C 〈http://www.nap. edu/openbook.php?record_id=11282〉. [17] H. Shuraishi, S. Tanaka, M. Hayakawa, A. Fujimura, H. Mizutani, Dynamical characteristics of planetary penetrator: Effect of incident angle and attack angle at impact, Inst. Space Astronaut. Sci. Rep. 677 (2000) 22.
either soft-landing or touch-and-go methodologies have been employed. While collection of surface samples is important, these materials have been modified by space weathering, and more information about the origin of the solar system can be attained from subsurface samples. A soft-lander that deploys drilling equipment could attain such samples, but the costs and complexity remain a significant challenge for this approach. This paper addresses the concept of using high velocity ground penetrators is considered here which reduces the complexity by not requiring drilling equipment, and reduces cost through the use of the ΔV between the spacecraft and target during approach. A penetrator with a ΔV of 300–700 m/s can typically punch into rock to depth of 1– 2 m, depending on the hardness of the penetrator and the rock. In this approach the penetrator has feed ports in the nose cone, which enable rock material in the impact site to move up into a central feed column, and into the sample return canister. Reflected energy and pressure for the injected material is able to produce self-ejection of the canister with the extracted sample that captures most of the subsurface stratigraphy. Survivability of the penetrator and the ejection of the sample return canister is highly dependent on the hardness of the rock in the impact region, and on the material makeup of the penetrator. Nevertheless, it has been shown here that survivability at impact speed of ~600 m/s is possible and that rock material along with its stratigraphy can be collected and ejected away from the impact site. There is some physical metamorphism that occurs with voids in the rock being filled with compressed rock material so that the density of the recovered rock is higher than the undisturbed rock. However, the bulk of the material appears undamaged except possibly for a thin layer that is in contact with the feed ports. We also were able to demonstrate that organics can also survive the impact. The main uncertainty that remains in this technique is to optimize the collection of the sample return canister. The most viable mechanism at this time is to have radar reflectors on the canister and radar/ optical locators on the main spacecraft detect the location of the SRC. With sufficient propellant to allow maneuvering of the spacecraft to pick up the canister, the mother spacecraft would pick up the SRC using an electromagnet on a tether to reel the SRC into the mother craft. With an ejection speed of about 10 m/s, there are options as to whether the pickup point would be in space as the mother craft approaches the object or whether SRC lands back on the surface and is picked up by a touch-and-go encounter by the mother spacecraft.
Prof. Robert Winglee received his PhD in 1985, and after several postdoctoral positions joined the University of Washington in 1991. He is a professor in the Department of Earth and Space Sciences as well as the Director of the Washington NASA Space Grant Consortium and the Director of Northwest Earth and Space Sciences Pipeline. He has published over 130 papers on space sciences and technology.
Mr. Chad Truitt attained a BS in 2013 and an MS in 2016 in Earth and Space Sciences at the University of Washington in 2016. He is now working at the Jet Propulsion Laboratory, continues his pursuit of sample return through the development of new technologies for possible future missions to Mars.
Acknowledgments We would like to thank NASA's Innovative Advanced Concepts for providing funding under grants NNX12AR02G and NNX13AR37G and a research environment supportive of innovative concepts. We also wish to thank to all the volunteers, faculty, staff, and students from the University of Washington who have helped in everything from the fabrication of the penetrators, to their testing and recovery (including ditch digging) over the last three years.
Mr. Rick Shibata received his BS in Mechanical Engineering at the University of Washington, Seattle in 2016. His undergraduate research and internship experiences concentrated on the applications of computational analysis, and is presently applying to graduate schools for additional expertise.
References [1] National Research Council, Committee on the Planetary Science Decadal Survey, Vision and Voyages for Planetary Science in the Decade 2013–2022. National Academy of Sciences, Washington D, 2011. 〈www.nap.edu〉. (accessed 17 May 2015). [2] E. Asphaug, Growth and evolution of asteroids, Annu. Rev. Earth Planet. Sci. 39 (2009) 413–448. http://dx.doi.org/10.1146/annurev.earth.36.031207.1274214. [3] S. Sandford, et al., Assessment and control of organic and other contaminants associated with the Stardust sample return form comet 91P/Wild 2, Meteorit. Planet. Sci. 45 (2010). http://dx.doi.org/10.1111/j.1945-5100.2010.02031x. [4] C.R. Neal, The moon 35 years after Apollo: what's left to learn?, Chem. Erde 69 (2009) 3–43. http://dx.doi.org/10.1016/j.chemr.2008.07.002.
286
Contents lists available at ScienceDirect
Acta Astronautica journal homepage: www.elsevier.com/locate/actaastro
High velocity penetrators used a potential means for attaining core sample for airless solar system objects
MARK
⁎
R.M. Winglee , C. Truitt1, R. Shibata Department Earth and Space Sciences, University of Washington, Seattle, WA 98195-1310, United States
A R T I C L E I N F O
A BS T RAC T
Keywords: Penetrator Sample return Subsurface sample Asteroids
Sample return missions offer a greater science yield when compared to missions that only employ in situ or remote sensing observations. Such missions have high ΔV requirements, and the return yields to date have been typically only of a few grams for robotic missions. Planetary penetrators offer an alternative that significantly reduce a mission's ΔV, increase sample yields, and allow for the collection of subsurface materials. The following details the design, development, and testing of penetrator/sampler technology capable of surviving supersonic impact velocities that would enable the collection of a solid core of geologic materials, without the need for any drilling equipment,,thereby reducing the overall mass and propellant budget. It is shown through both modeling and field testing that penetrators at speeds between 300 and 600 m/s (~Mach 1–2) can penetrate into the ground to depths of 1–2 m with overall structural integrity maintained. The first flight tests demonstrated the potential for survivability at these speeds. The second flight series demonstrated core sample collection with partial ejection of the sample return canister. The 3rd flight series demonstrated self-ejection of the sample return system fully intact and with the core retaining the full stratigraphy of the rock bed. The recovered sample also shows the survivability of macro-organic structures. Possible mechanisms for the recovery of the ejected core sample are also discussed.
1. Introduction The National Research Council [1] advised that Discovery and New Frontier class missions should play a critical role over the next decade in the study of primitive bodies given their relatively low cost and applicability for destinations both in the inner solar system, and beyond the asteroid belt. These missions will provide vital contributions in understanding the basic building blocks that created our world, as well as assessing potential hazards impacting bodies represent to our biosphere. Sample return missions from primitive bodies are excellent candidates for NASA's Discovery and New Frontier Programs given their potential for a high science yield while requiring only a fraction of the investment typical of a Flagship mission, and could supply materials long demanded by the science community for furthering our study of the solar system. To date more than 150,000 asteroids have been identified in the main belt alone [2], but difficulties in collecting samples has resulted in returning limited material. Scott Sandford [3], a team member of the OSIRIS-REx mission speaking at an Exploration Science Forum, highlighted some of the advantages sample return missions have over in-situ and remote
⁎
1
sensing methods, including: an increase in the quality of the data produced through the application of technology that was not available during the spacecraft's development; returned samples become research resources for both present and future scientists; sample analysis is not limited by design constraints of the spacecraft. He [3] also spoke to their potential to reduce research limitations that result from poor assumptions saying, “…if you decide to measure A, and you go there with your A measuring machine… it is possible that the main thing that you will learn… is maybe you should have measured B, and now you need new spacecraft and another mission”. The Apollo program, most noted for being the first exploration series to land humans on another rocky body, has provided the largest quantity of returned sample to date. Between the summer of 1969 and winter of 1972, Apollo astronauts gathered and returned more than 300 kg of lunar material. While NASA employed human beings to collect material from the Moon, the Soviet's Luna program became the first automated system to return samples collected from a minor body, albeit in much smaller quantities. The manned Apollo missions were able to bring back a few tens of kg each missions while the robotic Luna programs only brought back a fraction of a kilogram each [4]. However,
Corresponding author. E-mail addresses: [email protected] (R.M. Winglee), [email protected] (C. Truitt). Current address: Jet Propulsion Laboratory, 4800 Oak Grove Drive M/S 352H, Pasadena, CA 91109, United States
http://dx.doi.org/10.1016/j.actaastro.2017.04.022 Received 27 December 2016; Accepted 19 April 2017 Available online 21 April 2017 0094-5765/ © 2017 IAA. Published by Elsevier Ltd. All rights reserved.
Acta Astronautica 137 (2017) 274–286
R.M. Winglee et al.
S SRC TAGSAM Vs α ΔV
Nomenclature A ANSYS D Ks N
penetrator cross-sectional area in m2 Analysis System computer modeling software depth of penetration in meters, scaling factor dependent on mass penetrator nose coefficient
empirical target constant sample Return Container touch-And-Go Sample Acquisition Mechanism impact velocity in m/s an empirical constant equal to 1.75×10−5 Change in velocity
subsurface sample will be undertaken by Hayabusa 2 [11] which was launched in 2014 and will arrive at asteroid 162173 Ryugu in 2018. Hayabusa 2 will deploy a 2.5 kg kinetic impactor propelled by a 4.5 kg shape charge to impact the surface of asteroid 1999 JU3 at hypervelocity speed of 2 km/s in 2019, creating an artificial crater and collecting samples from greater depths than its predecessor. Hypervelocity penetrators has been used in two experiments to eject subsurface material into space where it is then examined using a remote spacecraft. The first was Deep Impact which in 2005 used a 370 kg impactor with an incident speed of 10.5 km/s to generate ejecta comet Tempel 1 [12]. A large increase in organic material as well as carbon dioxide was measured in the ejecta [12]. The Lunar Crater Observation and Sensing Satellite (LCROSS) in 2009 used the upper stage booster to impact a crater in permanent shadow at a speed of about 2.5 km/s. LCROSS was then able to successfully detect water in the ejecta plume [13]. Low velocity penetrators at speeds less than about 30 m/s have been used in variety of applications [14]. These include the dropping of instrument in remote areas to bunker busters. Sample return from icy objects has been proposed using penetrators with impact speeds of about ~30 m/s [15]. In this paper we examined an intermediary system for sample return which utilizes high velocity (200–800 m/s) impactors or penetrators to generate a core sample of subsurface material that is selfejected from the impact region. The concept for operations is described in Section 2 with the design parameters of the penetrator given in Section 3. The principle is very similar to the ice penetrator of Lorenz et al. [15] but the hardness of the rock is much higher that the ice in their studies, so that much higher impact velocities are required, and the material response to the impact is very different. Because of these differences the design and testing of the system is significantly modifiedfrom that of an ice penetrator. The data from the first few flights are given in Section 4 and show results for impacts at angles away from perpendicular impacts. Behavior of the penetrator and sample collected from a Mach 2 impact into sandstone is given in Section 5. Section 6 shows the results for a self-ejected core sample from a near-final version of the system. It is shown that the sample attained retains the overall stratigraphy of the rock, and macroorganisms retain intact in the recovered sample. A summary of results is given in Section 7.
manned missions are vastly more expensive, and no humans have traveled beyond the orbit of the Moon so that the issue of making significant sample returns in cost effective manner still remains an important issue. The Apollo and Luna programs both employed soft-landing techniques, requiring the expenditure of considerable amounts of fuel to safely arrive on the Moon's surface, and additional propellant to ascend once collection efforts were completed. These maneuvers significantly increase a mission's ΔV budget, and require flawless execution to ensure the survival of sensitive instrumentation. NASA's first sample return efforts after Apollo 17 was the Genesis mission [5], designed to characterize and sample the solar wind using a halo orbit around Lagrange point 1. The spacecraft gathered samples using collector arrays from late 2001 to spring of 2004, but the failure of its parachutes to deploy during Earth re-entry in September 2004 caused the Sample Return Capsule (SRC) [6] to impact the landing zone at more than 86 m/s, resulting in the possible contamination of the sample. Despite this non-ideal landing Genesis was able to successfully capture noble gases and isolate important isotopes. In contrast, the Stardust spacecraft [7] sent to collect samples from Comet 81 P/Wild 2 enjoyed a much better success. Launched in early 1999, the Stardust mission first collected samples of interstellar dust in 2000, and again in 2002 following an Earth gravity assist trajectory. The mission's flight team performed a close flyby of asteroid 5535 Anne Frank, using the opportunity as an engineering test of ground and spacecraft operations prior to intercepting Comet 81 P/Wild 2 in 2004, where it flew through the halo of gases and dust at the head of the comet. The spacecraft used an Aerogel filled grid to collect materials thought to pre-date the birth of the Sun, and their successful return to Earth in 2006 has provided new insights into our solar system [8]. There was also some contamination of the sample but signatures of returned cometary organics could be distinguished from the terrestrial contamination though there were concerns that the hypervelocity collection may have modified some of the organics. Robotic sampling for the actual surface of an asteroid was achieved by JAXA's Hayabusa [9] using a touch-and-go approach on asteroid 25143 Itokawa in 2005. In this scenario, sample retrieval is conducted as the spacecraft briefly contacts the surface of the sampling target, collecting a few grams of surface regolith before moving on. The touchand-go method avoids the problems of attaching a spacecraft to an asteroid and helps to reduce a mission's ΔV budget. In the case of Hayabusa, a tantalum pellet is fired at 300 m/s to produce ejecta which is then collected inside its sampling mechanism. Hayabusa demonstrated that touch-and-go sampling is possible, despite having hardlanded on Itokawa during a sampling attempt and collecting less than 1 mg of material. NASA's OSIRIS-REx spacecraft [10] will employ a similar touchand-go approach in 2019 to gather materials from asteroid 101955 Bennu using its Touch-And-Go Sample Acquisition Mechanism (TAGSAM). During contact with the surface, the TAGSAM will use a burst of nitrogen to blow regolith through a collecting sieve, and lab testing indicates that the method is capable of gathering more than 60 g of material. These two approaches are only able to attain surface samples which have the potential to be strongly modified compared to subsurface material due to its interaction with the solar wind. An attempt for a
2. High velocity penetrators for core rock sampling The concept of sample return by a high velocity penetrator is based on the fact that there is typically a high velocity difference (ΔV) between the target and the approaching spacecraft. A penetrator system can thereby use this kinetic energy for the impact without requiring any propellant. The mission concept for the release of the penetrator is shown in Fig. 1. The term high-speed refers to the impact velocity range of between 0.3–0.9 km/s. This velocity range is below hypervelocity speeds where chemical alterations can be produced but is sufficiently high energy to cause the breakup of most rock types. Depending on the diameter and speed of the penetrator and the hardness of the rock, the penetrator is expected to punch into the asteroid to a depth of 1–2 m. The sample return canister (SRC) lies within the penetrator which has feed ports on its tip to allow material 275
Acta Astronautica 137 (2017) 274–286
R.M. Winglee et al.
much each easier. In both case, radar and optical location finders would be needed but since this equipment is often included on solar system encounters, mass additions would be minimal. The actual collection of the sample could be achieved for example by an electromagnet. The bottom line is that the recovery would not require any new technologies and our focused here is to demonstrate that a self-ejecting core sample can be achieved by a high speed penetrator. 3. The penetrator dynamics – empirical and numerical simulations The ability of high speed penetrators to punch into rock has been documented by an empirical formulation derived from defense applications of penetrators [16] and, for the conditions in space applications, is given by
Fig. 1. Artist impression of the use of a high-speed penetrator for the collection and ejection of sun-surface core rock sample.
to move up into the center collection tube which feeds into the sample return canister. The tip of the penetrator is sufficiently hard that it is able to break up variable rock/soil strata. While there is some possibility of chemical metamorphism at the interface the cutting edge of the feed port and the sample intact, the bulk of the sample is expected to be unmodified. This broken up material then flows into the feed ports and into the central flow channel and eventually into the sample return canister. This material is under pressure so that its momentum, and/or a spring ejection system mechanism, forces the sample return canister to be ejected back above the surface while the penetrator itself remains in the ground. After it is ejected from the impact site, the SRC has to be collected by the mother spacecraft. There are two potential options available depending on the size of the object. For very small solar system objects (i.e. small asteroids) where the escape velocity is less than a few tens of m/s, the ejection of the SRC can be designed so that it reaches escape velocity. In this case, the mothership would have to locate the SRC on approach to the asteroid, rendezvous and pickup up the sample. The ΔV requirements are probably the lowest for this configuration but the potential for losing the SRC is probably the highest. The alternative is to have the SRC fall back onto target where it would be collected by the mother spacecraft basically by a touch-and-go encounter. The ΔV requirements are higher and there is a larger risk to the spacecraft due to landing requirements, but location finding of the SRC would be
D~αKs SN(m / A)0.7Vs
(1)
where A = penetrator cross-sectional area in m2 D = depth of penetration in meters, Ks = scaling factor dependent on mass and for the proposed applications has the form Ks =0.46×(m)0.15, m is penetrator mass in kg, N is penetrator nose coefficient which for a spherical projectile is ~ 0.5 while for a pointed projectile (and the proposed system) N is ~ 1.2, S = empirical target constant equal to ~ 0.7 for medium strength rock to 1.4 for low strength rock, Vs = impact velocity in m/s. α= an empirical constant equal to 1.75×10−5. For the speeds of 0.3–0.9 km/s for the present system, and for a 10–15 cm diameter penetrators the expected depth of penetration is about 1–2 m, dependent on rock type. At this depth the system would be able to sample material that has not been modified by long term exposure to the solar wind. In order to validate the empirical formation, we also use detailed numerical simulations via ANSYS – Autodyn. This commercial product
Fig. 2. Ejecta patterns for four different angle relative to the surface. There is an increase in the ejecta angle but the central cone ejecta has approximately the same form for all for cases.
276
Acta Astronautica 137 (2017) 274–286
R.M. Winglee et al.
can still perform for oblique angles of impact. The reason is that the moment of inertia of the penetrator is large for the high speed impacts being considered here that it breaks the rock as opposed to being deflected by the rock. In addition at high speeds, the ground cannot move significantly before the full width of the penetrator passes into solid material. Once it does so, the penetrator performance is completely independent of the surface properties. These results are also consistent with other studies for penetrator performance as a function of impact angle [17]. By the same arguments, the penetrator performance is also independent of the roughness of the surface of the target. Specifically, if the surface has boulders much larger than the penetrator diameter then the penetrator will just see it as an incline surface as in Fig. 2 and cut through it. For pebbles much smaller than the penetrator, then their inertia is too small to affect the trajectory and again the penetrator will cut straight through this material. Only in the unlikely case of the penetrator striking a cobble of similar diameter of
is designed for simulating the response of materials to short duration severe loading from impacts and high pressure explosions and is an important total for aerospace and defense applications. The ANSYS package also allows for element destruction and separation from the primary model associated with cracking and breakup of the impacted material. During impact, the stresses on the internal components are also calculated for possible failure modes. For the present application, different penetrator designs and impact hardness can be easily incorporated using tabulated data for the shock properties, Johnson Cook strength and failure models for different materials. As an example, Fig. 2 shows the performance of a short hollow cylinder penetrator going into medium strength material at 4 different impact angles using a grid of ~106 grid points. The measured depth of penetrator for this 100 g projectile was ~ 25 cm for this example which is within 20% of the depth predicted by the empirical formulation. The other important result in Fig. 2 is that the interaction is not strongly dependent on the angle of impact and that the system
Fig. 3. (a) Initial flight system using inline rocket motors, (b) sample collection system, and (c) final flight system using external booster rockets.
277
Acta Astronautica 137 (2017) 274–286
R.M. Winglee et al.
SRC (Fig. 3b). To bring the penetrator up to speed, a two-stage rocket system was used. A booster rocket motor lifted the system to 1000–1500 m. The rocket was then allowed to free fall for about 9 s when it would become aerodynamically stable in a downward trajectory. At this time the booster section of the rocket is ejected and the sustainer motors lit, bringing the system up to Mach 1–2 (depending on the size of the motors used) just prior to impact with the ground. In the initial configuration the sustainer motors were in-line with the long axis of the penetrator. Unfortunately, the motor could not be ejected in time before the impact occurred. For this reason, the later system used a series of 8 outboard motors for the sustainer stage. These motors impacted with the penetrator but because they are on the outside they could not influence the SRC performance, particularly with the large diameter of the rocket body. The final issue for the design of the penetrator was the positioning on the feed ports. The three configurations examine for this effort are shown in Fig. 4. The flow through the different configurations was modelled using SolidWorks’ FloXpress Analysis Wizard. Because this software cannot model the flow of a solid through feed ports, water was selected as the medium. This modeling is meant to only show qualitatively the difference in flow patterns and the derived speeds should only be considered as relative measures and not absolute speeds. The first configuration (Fig. 4a) where the feed ports are to the side of nose cone has the potential advantage of deeper penetration into the target material but would deflect most of the surface material since the tip will come in contact with the surface before the feed ports make contact. If the surface material (perhaps modified by space weathering such as the solar wind) is not required for sample collection then this might not be a significant loss if depth is the primary objective. The flow of material into the feed chimney is the slowest of the three configurations which translate to a reduced energy for the sample ejection. The second configuration (Fig. 4b) which is called the hollow point has the feed point through the tip of the nose cone. The hollow point flow pattern shows substantial deceleration on the outside rim of the hollow point which contributes to mushrooming. If mushrooming occurs then the depth of penetration will be reduced due to the
the penetrator is there a potential issue. However, if the mass of the penetrator exceeds that of the cobble (as per design), then again the penetrator inertia will stop it from undertaking a significant deviation and again, the surface roughness would not be considered an important factor in the operation of the penetrator. Results to this effect were observed for example during the near horizontal impact described in the following. The design of the penetrator derived from these simulations is shown in Fig. 3. The length of the penetrator is set to be approximately the same as the expected depth of penetration at about 1.66 m. The diameter was set at 15 cm, mainly due to the rocket motor requirements to bring the system up to speed. In space applications, the modeling indicated that the diameter could easily be reduced to 10 cm with little change in performance from that described below. The nose cone was made initially of Al 6061 with a mass of 1–3 kg. These nose cones work well for soft material such as playa, which was encountered in the first tests. As the testing progressed into harder material A2 tool steel was used for the nose cone with a mass of 6–6.5 kg. This material was still relatively easy to machine and had the ability to be hardened to a Rockwell hardness ~62c. Feed ports through the nose cone enable material from the impact area to move into the central feed chimney where it would be collected within the sample return canister. The sample return canister (SRC) was initially made of 1.5 mm thick Al which proved to have insufficient strength to withstand the build-up of pressure from the upflowing rock material. In the later tests the Al SRC was replaced by 3 mm thick steel SRC and this system was able collect sample and be ejected with minimal damage. The diameter of the SRC was 4 cm with a total length of 40 cm. The feed chimney was supported by a honeycombed aluminum structures with carbon fiber strengtheners. This material could withstand forces of more than 1 kN/cm2 yet have a sufficiently low density that the mass of the chimney and body tube was less than the mass of the nose cone. This support structure prevents the chimney from collapsing during impact and at the same time prevents the rock in the impact site from fully compressing the sides of the penetrator and restricting the diameter of the feed chimney. In the later versions of the system a shocking spring was added to the support of the SRC which provided additional upward force after impact to ensure ejection of the
Fig. 4. Relative flow model for three feed port designs.
278
Acta Astronautica 137 (2017) 274–286
R.M. Winglee et al.
tion yields the highest flow rates into the feed chimney and is thus the most efficient for collecting rock sample. The negative aspect of this configuration is that it is the structurally weakest though as demonstrated below still able to remain intact at supersonic impacts into sandstone. Another aspect of this system is that there is potential for some mixing of material moving through the side ports with material moving through the tip port, which would make the stratigraphy collected a little more complex to interpret. Despite the above modeling there is still some certainty in the optimal nosecone shape as there are still several uncertainty factors. Low speed penetrators tend to have blunt nose cones, which yields greater stability as the penetrator moves through the rock. For the high speed penetrators this stability issue appears to be outweighed by the braking issue. Field testing of the blunt nose cone figuration showed that the deceleration appeared to be too fast and the back end of the penetrator tended to try and push pass the nose cone causing the penetrator to buckle. For the sharp nosecone used here, bending was also seen to occur when the nose cone was made of the same material (Al) as the rear of the rocket. This bending is basically due to the way the different components of the rocket are decelerating and whether the rock is moving out of the way of the penetrator or vice versa. Having the density of the nosecone higher than the rear of the penetrator means that side friction can slow the rear at the same rate as the front. Similarly, a high density nose cone forces the rock material to move in response to the impact as opposed to the nose cone being deflected by the rock. Thus, the final design for the penetrator consisted of a tool-hardened steel nose cone, and Al body. Under these conditions, minimum bending of the rocket during impacts was seen in the field testing below. Fig. 5. Penetrator impacted at 150 m/s into playa embedding to ~1.22 m at 30° from vertical.
4. Initial field testing – impact angle dependencies One of the unknowns for contact with solar system objects is the actual surface conditions the penetrator would encounter since in general most objects do not have high resolution surface data until the arrival of the spacecraft. For the present application this translates to whether the penetrator will perform as planned when the angle of impact is significantly different from vertical. The simulation results in Fig. 2 indicate that the angle of impact is not a significant factor, provided the tip impacts the target before the sides. Thus, the minimum impact angle appears to be approximately set by the nose cone angle. This result was confirmed in initial field testing of the penetrator. As an example of a non-perpendicular impact, Fig. 5 shows the results of one of the earlier design penetrators flown between 2013 that impacted playa at about 30° from perpendicular at 150 m/s. The depth of penetration was approximately 1.3 m and the system showed essentially no deformation. Fig. 6 shows the results from a flight in 2014 that struck clay/soil for a near tangential impact. Similar to the simulations in Fig. 2 there is increased ejecta downslope with the length embedded into the ground being about the same as a near normal impact. The system in these test shot remained intact as well except for a little bending of the nose cone. These results show that the system can handle arbitrary impact angle. The main parameter that controls performance is the requirement that the velocity vector be aligned with the long axis of the penetrator. This is achieved in our terrestrial applications using fins as passive stabilization. In Fig. 6 due to non-simultaneous lighting of the booster motors there is a misalignment between the two vectors by about 5°. Nevertheless, the system is able to handle this difference because of the internal reinforcements of the system to absorb the energy of the hard impact. In space applications, this passive stabilization can be achieved by spinning up the penetrator at the time of release to the same effect, and then have rods released just prior to impact to slow the spin rate.
Fig. 6. Near tangential impact of our penetrator with the last frame prior to impact (camera is level with the ground). There is a larger ejecta pattern but the system remains intact and penetration depth is about the same as in Fig. 5.
increase in cross-section increasing the drag as the nose cone moves through the rock. The flow speed is higher than the side port configuration. Moreover, all material that is encountered by the nose cone moves up the feed port to be collected by the SRC. The third configuration (Fig. 4c) is the hybrid feed port which has feed ports at both the nose cone tip and along it sides. The configura279
Acta Astronautica 137 (2017) 274–286
R.M. Winglee et al.
Fig. 7. Successful ignition, launch and ignition of the clustered two-stage penetrator.
Fig. 8. (a) Images of the supersonic penetrator impact site, (b) students marking some of the larger debris items, (c) map showing the positioning of some of the debris and (d)-€ show the recovered top of the SRC.
astray due to a misfire of a motor or fin damage, the rocket could not reach any structure. The launch team was 2500 ft from the launch area and the rocket fired at 15 degrees away from the fire control area to ensure safety of the ground crew. Launches were typically performed over a 1–2 day period with recovery occurring in subsequent days after launches were completed. The launch of this system and the downward powered flight is shown in Fig. 7. The first high speed impact of the steel tip penetrator that demonstrated survivability was achieved in March 2015. This system had the hybrid nose cone shown in Fig. 4c. It was mated to an
5. Field testing: sandstone impact With the rocket configuration of Fig. 3b, we were able to attain supersonic impacts starting in December 2014 on a large ranch near Ione CA. The test area on the ranch was centered such that the nearest building was 1.5 miles away from the target area. An FAA waiver was attained for the flights up to 5000 ft above ground level, with appropriate notification going out prior to and after the launch period to relevant air route traffic control centers. The rocket boost system had a predicted maximum altitude of 3000 ft. Thus, even if the rocket went 280
Acta Astronautica 137 (2017) 274–286
R.M. Winglee et al.
Fig. 9. Trenching of the penetrator from the impact site in Fig. 7.
Fig. 10. Images of the recovered front section of the penetrator including nose cone and feed ports.
outwards. While not the expected outcome for the SRC, it nevertheless demonstrated that there is energy within the center feed system to produce self-ejection of material. The impact crater showed less surface disruption itself and was only about 1 m in radius (Fig. 9). This is an important result of the system in that only a small area is disturbed even for these high speed impacts. The reason for this is that the system is long and slender, and there are no explosives involved. There was actually no sign of the penetrator system in the impact crater, except for the red shock cord from the top of the SRC (Fig. 9a). The lack of penetrator components at the top of the crater was a little perplexing given that in all previous shots some part of the penetrator was evident in the impact crater. Once the top few inches of soil was removed (Fig. 9b), the underlying sandstone was revealed and the impact shaft contained a large amount of backfill. Loose backfill debris was removed revealing the top of the impact shaft, 10 cm in diameter, with clearly observable slickensides (Fig. 9b). The
aluminum extension resulting in an assembly to minimize the total mass of the penetrator. The penetrator was fitted with a full array of eight, outboard motors to provide the highest impact velocity following the post-apogee free fall. Slight ignition timing delays in the boost stage resulted in a launch angle of ~30°, reducing the ideal apogee altitude to ~1000 m. While this reduced the planned free fall time, the rocket remained stable during the ignition of the second stage and produced a very loud sonic boom, immediately followed by the sound of the impact. Impact velocity was estimated to be over 600 m/s, embedding to a depth of 1.6 m into solid sandstone. The impact was very energetic, and debris from the exterior cluster motors was spread across a broad swath of a few tens of meters (Fig. 8). The farthest piece of debris that was recovered was the top of the SRC which was found about 83 m from the impact site. It was coated in a film of top soil, which presumably was deposited during the impact, and the pressure within the tube blowing the top of the canister 281
Acta Astronautica 137 (2017) 274–286
R.M. Winglee et al.
Fig. 11. The core sample was exposed after removal of the encasing material.
digging out caused the nose cone to break into two separate pieces associated with the brittle fracturing. This material failure was likely the result of over-hardening of the steel (Rockwell hardness ~62c), instead of optimizing hardness and strength, resulting in a hard but brittle tip that still provided effective penetration and sample collection. The most important observation of the nose cone was that all feed ports were solidly packed with sandstone from the impact region. This sandstone is not fragmented but rather appears to have moved into the feed ports under cataclastic flow without fracturing. The cutaway of the penetrator in Fig. 11 shows that the sandstone moved up into the central feed chimney coherently for the first 60 cm from the tip of the nose cone. Note also that the diameter of the sandstone in the chimney increases with distance from the feed ports, which provides some indication of the pressures that developed on the sample from the impact. The encounter with the SRC (top left of Fig. 11) shows turbulent behavior, with the SRC significantly folded and distorted from its original cylindrical shape. This turbulent behavior is thought to arise from the loss of the top of the SRC which was found above the ground and the interaction of the cataclastic sandstone with the bottom baffle of the SRC. Without the back plate there is no buildup of pressure to decelerate the SRC. As a result, it is still moving forward when it encounters the cataclastic sandstone which then causes two-stream turbulent behavior to occur. The impact definitely produces physical metamorphism of the rock. The measured density of the undisturbed material was measured as 1.7 g/cm3. The density of the sample core was measured at 2.1 g/cm3. Thus, there is about a 25% increase in density due to the filling of voids in the rock from the compressive forces from the impact. Examination of thin sections of the recovered sample the bulk of the sample was not chemically, and we did not detect any signs of modifications to the minerals relative to the undisturbed materials. The main differences occurred along the contact surface of the cutting edge of the feed ports. There was some increased ion content along the surfaces that were in contact with the central feed port that appeared as a red ring in crosssection through the core sample. We were unable to determine whether this enhancement came from the steel feed port itself or from some chemical metamorphism from the impact.
Fig. 12. The epoxied SRC with its tip in front of the main nose cone and the penetrator system ready for flight.
top of the penetrator was not found until about 0.5 m of top material was removed (Fig. 9c). Earlier impact models suggested that material proximal to the impact shaft would experience varying levels of fracturing. There was some fracturing and examples of over-turned strata near the surface of the impact site, but these effects were limited to the first few inches and were not observed at greater depths. Instead, the rocket created a shaft with a diameter that matched the maximum diameter of the metal portions of the nose cone assembly of 10 cm (where as the main body of the penetrator had a 15 cm diameter). The extracted penetrator is shown in Fig. 10, and is basically intact except for the section holding the outboard motors, which was designed to strip away on impact. The penetrator shows no bending as it penetrated the sandstone, so that the energy absorbing section fulfilled its purpose of inhibiting compression along the axis of the penetrator; however, the diameter - particularly around the midsection - was reduced from its initial 15 cm to about 10 cm which was the diameter of the metal components of the nose cone. The nose cone itself experienced some brittle fracturing, particularly in the tip. The nose cone was intact in the crater, however stresses from the
6. Field testing: SRC ejection One of the important conclusions from this series of flight tests is that while material is able to move up the feed ports, it never reaches the height of the surface, which makes self-ejection problematic. In all cases it reaches only a height of about 0.3–0.5 of the penetration depth. 282
Acta Astronautica 137 (2017) 274–286
R.M. Winglee et al.
Fig. 13. Launch of the 3rd generation penetrator its impact and self-ejection of the SRC.
Fig. 14. A small low arc feature develops about 1 s after the impact, believed to be entrained smoke from the ejected SRC.
To overcome this problem for the final set of field tests, the penetrator was redesigned in an effort to ensure full ejection of the SRC. The SRC itself was fabricated from steel, and was moved fully forward in the penetrator so that it extended beyond the end of the machined nose cone (Fig. 12). In this forward position it would experience the greatest upward pressure but the steel construction would prevent the catastrophic failure seen in the Al SRC. A shock-absorber spring was also added to aid in the ejection of the sample. Moving the sample return canister forward increased the risk that it would bend on impact, but the weight and energy of the nose cone ensured that even if such bending were to occur the SRC would still be forced back into the
central feed column. Because of the loose fit of the SRC, it was epoxied in place and its feed port was covered for flight (Fig. 12); the epoxy was sufficiently robust enough to survive flight, but the contact joint would shatter on impact allowing the SRC to move through the center column of the penetrator without binding. The launch occurred in December 2015 at Ione CA and its flight and impact is shown in Fig. 13. During the field test, only four of the eight outboard cluster motor ignited (probably due to the igniters coming loose from vibration/g-forces during flight) so that the impact speed was only about 330 m/s. During flight the smoke grains that aid the visual tracking of the rocket were only weakly
283
Acta Astronautica 137 (2017) 274–286
R.M. Winglee et al.
Fig. 15. Trenching of the impact site in Figs. 13 and 14 and cutaway of the canister showing intact stratigraphy.
burning, but on impact they flared up to produce the enhanced smoke plume that last several seconds after impact. On inspection of the impact site (Fig. 13e), the SRC was found approximately 2 m from the impact site, thereby proving self-ejection of the sample is possible. Because of the slower speed, the booster motors are seen protruding from the impact side, though they are no longer attached to the main penetrator body. Close inspection of the video of the penetrator flight was unable to resolve the ejection of the SRC. However, the smoke trail as shown in Fig. 14 shows the development of a low (~3 m) hook feature that curves in the same direction as the location of the ejected SRC. The appearance of this feature is about 1 s after impact and fitting a parabolic trajectory along this feature yields an impact at the location of the SRC. It is believed that this smoke feature is associated with smoke being entrained with the SRC as it is ejected. Fitting a parabolic trajectory with a 1 s to apogee and 1 s back down to the ground at 2 m and a height above ground of 3 m and 1 m below ground, yields an ejection velocity of about 10 m/s at an angle of about 10° from vertical. At this speed, the SRC could be ejected to significant height above the surface of any small bodied asteroid where the gravitation forces are very much weaker. The trenching around the impact site is shown in Fig. 15a and shows top soil and clay to about 0.3 m. Below this material, there was a region of talc and other sedimentary material. As noted from the impact site the SRC was recovered intact as seen in Fig. 15b. The end of the SRC was deliberately slotted so as to fold inwards during the impact to produce self-sealing of the SRC. This folding was partially achieved as seen in Fig. 15c, though a couple of folds appear to have broken off and entered the SRC. Within the cutaway of the SRC (Fig. 15d) the ground stratigraphy is seen in to have been retained, with soil and clay near the top of the canister and talc (yellow material) at the bottom of the SRC. Close ups of the material in the top of the SRC are shown in Fig. 16. At the very top of the canister small pebbles are seen mixed in with top soil. The concentration of the pebbles is larger than seen in the top soil, indicating some loss of fine material during impact. This loss of material is due to holes in the SRC which is required to vent air trapped in the SRC during the impact. Since there is no air in space, vent holes will not be required in planetary applications and this loss of top material is not expected. Another important find within the SRC is the presence of macroscopic organic material associated with the grass in the impact region. This result is important as it demonstrates that organics can be preserved during the impact. This result also demonstrates that the heating produced by the impact is insufficient to destroy macroorganics and also means that chemical metamorphism is unlikely to occur during the impact, except for possibly the contact zone at the
cutting edge of the feed port The red ring of material that was seen in the Mach 2 impact discussed in the previous section was not see in the present sample. 7. Discussion Field testing results demonstrated the potential application of ground penetrators for sample return, but a number of parameters remain uncertain. Switching the SRC design to steel instead of the original aluminum and carbon fiber used in earlier evolutions, created a more robust container, and moving the SRC fully forward in the penetrator allowed the SRC ejection to begin before significant structural failures farther aft occurred in the penetrator during impact. This was demonstrated by the two tests discussed, but moving the SRC fully forward introduced a new limitation – sampling depth. In the earlier design, sampling began once the feed ports on the nose cone had opened after impact, allowing sampling of material from more than 1 m in depth. Moving the SRC forward resulted in sample collection from less than a meter in depth since the SRC was ejected before embedding was complete. Sampling depth might be improved by the development of a feed port cover that is progressively abraded away during the initial contact such that sampling does not begin at the moment of impact. Another large uncertainty is the potential for metamorphic effects on the sample materials during impact. Terrestrial metamorphic processes are due to chemical reactions, and extreme temperatures and pressures generally occurring over geologic timescales. Chemical and temperature metamorphism over the bulk of the collected sample is unlikely to occur because the time scale of the impact at a few milliseconds is insufficient to produce bulk heating of the sample. However, there is the possibility that there is some chance of chemical metamorphism at the surface which contacts the feed ports. Contact pressure at the first contact of the nosecone with the sandstone was estimated to be about 90 MPa, and about 40 MPa for the test during which the SRC was fully ejected, which are sufficient to produce cataclastic flow of soft rock. However, the sandstone core sample exhibited an unusual orange/red ring that concentrically ran throughout the length of the core. Thin section analysis of the sandstone samples are ongoing so no definitive identification of the orange ring has been made yet, but we speculate that the ring is either the result of steel in the machined tip being worn away during the impact, or from some type of chemical metamorphism along the contact surface. The deceleration profile during impact, and hence the resulting pressures created in the sample material, are governed by two major factors: impact velocity, and target density. Since impact velocity can be adjusted to minimize metamorphic potential, the largest uncertainty becomes the density of the targeted body. As material density 284
Acta Astronautica 137 (2017) 274–286
R.M. Winglee et al.
Fig. 16. Close up of the sample showing survivability of woody/grassy material with a thumb tack for scale.
8. Conclusions
increases, acceleration increases as the penetration depth decreases; with lower material density the penetration depth increases creating a longer acceleration profile and as such, lower pressures. The density of a primitive body selected for impact sampling becomes critical in optimizing the impact velocity of the penetrator so that a maximum embedding depth can be reached without the risk of serious metamorphic effects. This optimization will remain problematic for a time until a better characterization of the densities of primitive bodies has been made.
Sample return missions offer the greatest science return on investment given the wider array of analytical methods that can be applied in terrestrial laboratories. Human retrieval during the Apollo era has brought back the largest amount of non-terrestrial material but having a sustained human presence beyond low Earth orbit remains problematic. As a result, robotic missions are the present mainstay for sample return missions, and at this time these samples have been restricted to the surface of small solar system objects. In attaining these samples 285
Acta Astronautica 137 (2017) 274–286
R.M. Winglee et al.
[5] D.S. Burnett, et al., The genesis discovery mission: return of solar matter to Earth, Space Sci. Rev. 105 (2003) 509. [6] D.S. Burnett, K.M. McNamara, A. Jurewicz, D. Woolum, Contamination on anodized aluminum components on the Genesis Science Canister, Lunar Planetary Science Conference, 20050166923, 2005. [7] D.E. Brownlee, et al., Stardust: comet and interstellar dust sample return mission, J. Geophys Res. Planets 108 (2003). http://dx.doi.org/10.1019/2003JE002087. [8] J.E.P. Matzel, et al., Constraints on the formation age of cometary material from the NASA Stardust mission, Science 328 (2010) 483. http://dx.doi.org/10.1126/ science.1184741. [9] H. Yano, T. Kubota, H. Miyamoto, T. Okada, D. Scheeres, Y. Takagi, K. Yoshida, M. Abe, S. Abe, O. Barnouin-Jha, A. Fujiwara, S. Hasegawa, t. Hashimoto, M. Ishiguro, M. Kato, J. Kawaguchi, T. Mukai, J. Saito, S. Sasaki, M. Yoshikawa, Touchdown of the Hayabusa spacecraft at the Muses Sea on Itokawa, Sci. New Ser. 312 (5778) (2006) 1350–1353. [10] E. Beshore, et al., The OSIRIS-Rex asteroid sample return mission, IEEE Aerospace Conference doi: http://dx.doi.org/10.1109/AERO.2015.7118989, 2015. [11] Y. Tsuda, M. Yoshikawa, M. Abe, H. Minamino, S. Nakazama, System design of the Hayabusa 2 – asteroid sample return missions to 1999 Ju 3, Acta Astronaut. 91 356 (2013). [12] M.F. Hearn, et al., Deep impact: excavating Comet Tempel 1, Science 310 (2005) 258. http://dx.doi.org/10.1126/science.1118923. [13] A. Colaprete, R.C. Elphic, J. Heldmann, K. Ennico, An Overview of the Lunar Crater Observation and Sensing Satelitte (LCROSS), Space Sci. Rev. 167 (2012) 3. [14] R. Lorenz, Planetary Penetrators: their origin, history and future, Adv. Space Res. 48 (2011) 403. [15] R.D. Lorenz, W.V. Boyton, C.F. Turner, Demonstation of Comment Sample Acquisition by Penetrator, ESA Proceedings, IAA Conference on Low-Cost Planetary Exploration, SP-542, 387, 2003. [16] National Research Council. Effects of Nuclear Earth-penetrator and Other Weapons, 2005, The National Academies Press Washington, D.C 〈http://www.nap. edu/openbook.php?record_id=11282〉. [17] H. Shuraishi, S. Tanaka, M. Hayakawa, A. Fujimura, H. Mizutani, Dynamical characteristics of planetary penetrator: Effect of incident angle and attack angle at impact, Inst. Space Astronaut. Sci. Rep. 677 (2000) 22.
either soft-landing or touch-and-go methodologies have been employed. While collection of surface samples is important, these materials have been modified by space weathering, and more information about the origin of the solar system can be attained from subsurface samples. A soft-lander that deploys drilling equipment could attain such samples, but the costs and complexity remain a significant challenge for this approach. This paper addresses the concept of using high velocity ground penetrators is considered here which reduces the complexity by not requiring drilling equipment, and reduces cost through the use of the ΔV between the spacecraft and target during approach. A penetrator with a ΔV of 300–700 m/s can typically punch into rock to depth of 1– 2 m, depending on the hardness of the penetrator and the rock. In this approach the penetrator has feed ports in the nose cone, which enable rock material in the impact site to move up into a central feed column, and into the sample return canister. Reflected energy and pressure for the injected material is able to produce self-ejection of the canister with the extracted sample that captures most of the subsurface stratigraphy. Survivability of the penetrator and the ejection of the sample return canister is highly dependent on the hardness of the rock in the impact region, and on the material makeup of the penetrator. Nevertheless, it has been shown here that survivability at impact speed of ~600 m/s is possible and that rock material along with its stratigraphy can be collected and ejected away from the impact site. There is some physical metamorphism that occurs with voids in the rock being filled with compressed rock material so that the density of the recovered rock is higher than the undisturbed rock. However, the bulk of the material appears undamaged except possibly for a thin layer that is in contact with the feed ports. We also were able to demonstrate that organics can also survive the impact. The main uncertainty that remains in this technique is to optimize the collection of the sample return canister. The most viable mechanism at this time is to have radar reflectors on the canister and radar/ optical locators on the main spacecraft detect the location of the SRC. With sufficient propellant to allow maneuvering of the spacecraft to pick up the canister, the mother spacecraft would pick up the SRC using an electromagnet on a tether to reel the SRC into the mother craft. With an ejection speed of about 10 m/s, there are options as to whether the pickup point would be in space as the mother craft approaches the object or whether SRC lands back on the surface and is picked up by a touch-and-go encounter by the mother spacecraft.
Prof. Robert Winglee received his PhD in 1985, and after several postdoctoral positions joined the University of Washington in 1991. He is a professor in the Department of Earth and Space Sciences as well as the Director of the Washington NASA Space Grant Consortium and the Director of Northwest Earth and Space Sciences Pipeline. He has published over 130 papers on space sciences and technology.
Mr. Chad Truitt attained a BS in 2013 and an MS in 2016 in Earth and Space Sciences at the University of Washington in 2016. He is now working at the Jet Propulsion Laboratory, continues his pursuit of sample return through the development of new technologies for possible future missions to Mars.
Acknowledgments We would like to thank NASA's Innovative Advanced Concepts for providing funding under grants NNX12AR02G and NNX13AR37G and a research environment supportive of innovative concepts. We also wish to thank to all the volunteers, faculty, staff, and students from the University of Washington who have helped in everything from the fabrication of the penetrators, to their testing and recovery (including ditch digging) over the last three years.
Mr. Rick Shibata received his BS in Mechanical Engineering at the University of Washington, Seattle in 2016. His undergraduate research and internship experiences concentrated on the applications of computational analysis, and is presently applying to graduate schools for additional expertise.
References [1] National Research Council, Committee on the Planetary Science Decadal Survey, Vision and Voyages for Planetary Science in the Decade 2013–2022. National Academy of Sciences, Washington D, 2011. 〈www.nap.edu〉. (accessed 17 May 2015). [2] E. Asphaug, Growth and evolution of asteroids, Annu. Rev. Earth Planet. Sci. 39 (2009) 413–448. http://dx.doi.org/10.1146/annurev.earth.36.031207.1274214. [3] S. Sandford, et al., Assessment and control of organic and other contaminants associated with the Stardust sample return form comet 91P/Wild 2, Meteorit. Planet. Sci. 45 (2010). http://dx.doi.org/10.1111/j.1945-5100.2010.02031x. [4] C.R. Neal, The moon 35 years after Apollo: what's left to learn?, Chem. Erde 69 (2009) 3–43. http://dx.doi.org/10.1016/j.chemr.2008.07.002.
286