CostEffective Management of Orbital Debris CODER Debris Workshop
Cost-Effective Management of Orbital Debris CODER Debris Workshop University of Maryland November 15 -17, 2016 Joe Carroll Tether Applications, Inc. 619 -421 -2100; tether@cox. net 2016 CODER Workshop Tether Applications, Inc. Nov 2016, pg. 1
Useful Terminology for Orbital Debris Cars (~3000) Intact objects, mostly ton-class; <1% of all lethal LEO objects 98% of target area & 99% of mass for debris-creating impacts! Easy to track & avoid, but the source of hubcaps & shrapnel Hubcaps Tracked fragments, mostly >10 cm, <1 kg: ~ 2% of lethal objects (~9, 000) Hubcaps dominate tracking costs; most are too light to shred cars 44% are from just 2 collisions: Fengyun/A-sat + Cosmos/Iridium Shrapnel Lethal untracked fragments, ~1 gm: >97% of all lethal objects? (~500, 000? ) Too small to track & avoid (now), but too heavy to shield against This is the expensive (but invisible) direct threat to assets! We worry mostly about a cascade of hubcap/car collisions in low earth orbit making more & more hubcaps. We should worry mostly about lethal shrapnel, and the accidental car/car collisions that will create most of it! 2016 CODER Workshop Tether Applications, Inc. Nov 2016, pg. 2
Orbital Debris 101 1. We must estimate debris costs to make rational debris policy decisions. 2. We know even less about debris costs than about the cost of most fixes! 3. The value of actionable warnings is the average net money they save. 4. Debris can shred ~1, 000 X its mass, or disable ~1, 000 X its mass. 5. We focus on >10 cm hubcaps, but ~1 cm shrapnel causes most losses. 6. Car + car = shrapnel will force action before a hubcap + car cascade does. 7. ~2/3 of mass in LEO is Russian; 12. 5% US; 10. 5% PRC; 10. 5% other. 8. LEO altitude drives orbit life, target count, & shrapnel life. So fly low! 2016 CODER Workshop Tether Applications, Inc. Nov 2016, pg. 3
How Much Lethal Shrapnel Is Now in LEO? Nobody knows! (the 2 best documented models differ by a lot) Ø Horizontal lines below (from ODQN 2015: 2) are factors of 100, not 10 Ø Fluxes >1 cm differ by 5 X at ISS (415 km), and by 50% at 850 km Ø Fluxes >1 mm differ by 30 X at ISS, and by 220 X at 850 km 2016 CODER Workshop Tether Applications, Inc. Nov 2016, pg. 4
Cars and Owners at Congested Altitudes in LEO 1. Larger objects have higher collision risk, and their high mass raises collision yield. 2. Tight altitude clustering of massive large objects further raises their collision risk. 3. Most Iridium satellites can (& will? ) maneuver, & may deboost when replaced. 4. Russian rocket bodies are most of the mass in most of the crowded altitude bands. 5. They aren’t the only issue, but they will be the source of most collisional shrapnel. Tons/Km Mass at 450 -1050 km in April 2016 (93% of future shrapnel!) 18 Zenit + 20 Tselina 2, 71 o Kosmos-3 M 2. 4 x 6 m, 1434 kg; 282 at 600 -1600 km Kosmos-3 M: 138 at 83 o +13 at 66 o & 74 o 73 Iridium Kosmos-3 M Zenit 3. 9 x 11 m 8300 kg, 71 o Iridium ~2 x 4 x 8 m, 550 kg, 86 o 2016 CODER Workshop Tether Applications, Inc. Nov 2016, pg. 5
Altitude Congestion Drives Shrapnel Creation 1. As industry practices improve, most breakup types will become less frequent. 2. But collisions of large objects are rising, roughly with the square of LEO mass. 3. There were no car/car breakups until 2009, but chance is now 7%/year and rising. 4. ~93% of accidental-collision LEO shrapnel will be made at 460 -1020 km altitude. Mean Expected New >1 Gram Collisional Shrapnel/Year, Per Km Creation Altitude, April 2016 ~93% 2016 CODER Workshop Tether Applications, Inc. Nov 2016, pg. 6
Relative Shrapnel Creation by Cars & Cubesats Compare cubesat/car and car/car collisions at the same altitude: 1. Cubesat/car has only 1/2 the collision radius, so only 1/4 the area. (Good) 2. And cubesat/car collisions involve only 1/2 the total mass. (Also good) 3. And cubesat/car may create less shrapnel / kg. (But JSC doesn’t think so) Ø So cubesat/car shrapnel creation is ~1/4 * ~1/2 * <1, or ~10% of car/car. Ø But that is from cubesats weighing only 0. 1 -0. 5% as much as typical cars. Ø Do cubesats have a right to create 20 -100 X as much shrapnel per kg as cars? Ø Note that typical collision-made hubcaps are ~0. 1 kg, and cannot shred cars. Now compare cubesats in lower orbits (as most are): 1. Over the 350 -750 km range, circular orbit life scales with ~Altitude 9. 2. And the mean orbit life of any shrapnel created also scales with ~Altitude 9. 3. Non-ISS “car” mass scales with ~Alt 5. 5 at 250 -570 km; ~Alt 1 at 570 -1000 km. Ø Over 350 -750 km range, expected shrapnel-years scales with ~Altitude 21. Ø Flying 20% lower cuts life by a factor of 7. 5, but mean shrapnel-yrs by ~100! 2016 CODER Workshop Tether Applications, Inc. Nov 2016, pg. 7
How Can We Disrupt the Shrapnel Life Cycle? Options for individual operators: 1. Don’t worry about it, but do accept a growing risk of asset loss from it. 2. Armor new assets, to increase the mass threshold for lethal impact. 3. Use lower altitudes (even below ISS!), and reboost to stay in orbit. Options needing better debris detection & tracking (probably optical): 4. Detect and track shrapnel precisely, so assets can affordably dodge it. 5. Use pulsed laser ablation to nudge cars & hubcaps, to prevent collisions. Options involving wholesale removal or collection: (all “A-sat-like”!) 6. Melt and vaporize shrapnel (& hubcaps? ) w/continuous laser heating. 7. Deorbit shrapnel & hubcaps (even cars? ) using pulsed laser ablation. 8. Capture & deorbit large debris, or collect for later deorbit or recycling. I think these 5 options will all be used. 2016 CODER Workshop Tether Applications, Inc. Nov 2016, pg. 8
Can Conjunction Warnings Be a Viable Business? The value of warnings is simply the net money they save operators: 1. Warnings are not warnings of impact, but of tiny chances of impact! 2. The net value differs for each satellite, each impactor, & each warning. 3. Dodging hubcaps improves public safety and may reduce future liability. 4. But most value will come from dodging now untracked 1 -3 cm shrapnel. 5. The more lethal threats you warn against, the more valuable the service. How to sell LEO conjunction warnings: 1. Find & track all the “probably lethal” impactors (>5 -10 mm) you can. 2. Get good at predicting their orbits (including variations in drag area!) 3. Maybe provide free preliminary warnings, so operators buy updates. Ø But you must avoid the reputation of the “boy who cried wolf” 2016 CODER Workshop Tether Applications, Inc. Nov 2016, pg. 9
Conjunction Error Causes, Effects, and Cure Conjunction errors come from observation + prediction errors: 1. Radar observations are precise in range, but not in transverse directions. Ø In contrast, telescopes are precise in 2 -D, but give no direct range info. 2. Along-track errors in LEO come mostly from drag prediction errors: Ø Air density is inferred and predicted using the USAF HASDM model. Ø Debris attitude dynamics are now ignored but do affect future drag area. 3. Conjunction errors include the effects of errors for both objects. Conjunction errors are orders of magnitude larger than the objects: 1. Each day gives many warnings, but each decade has only 1 -2 collisions. Ø Warnings do not predict collisions, but rather tiny chances of collisions. 2. Smaller errors reduce the number of warnings, and also avoidance DVs! To greatly reduce warning errors (& hence operator costs), consider: 1. A world-wide network of telescopes (or fewer sites, with laser lighting). 2. Far better drag data and modeling, including debris attitude motion. 2016 CODER Workshop Tether Applications, Inc. Nov 2016, pg. 10
What Is the Best Way to Track LEO Objects? 1. Suitable telescopes may be both far cheaper & far better than radar: Ø Allow quick 2 D direction fixes accurate within meters (against stars). Ø But LEO objects are visible only near dusk & dawn, & only when clear. Ø So don’t buy dedicated scopes; time-share suitable networkable scopes! 2. Agile computer-controlled direct-drive mounts seem required: Ø Each telescope may make ~1, 000 slews/year to track LEO objects. Ø This will quickly wear out geared drives, but not direct-drive mounts. 3. One option: Plane. Wave 700 mm f/6. 5 CDK ($210 K): Ø Ø Ø Telescope is fully automated, to allow remote operation Has direct-drive alt-az mount and dual Nasmyth focus Over 20 are already in use; most may be time-shareable May allow tracking of most ~1 cm shrapnel <1000 km But detection & IOD of ~1 cm shrapnel is a challenge. 2016 CODER Workshop Tether Applications, Inc. Nov 2016, pg. 11
Self-Protection of Satellite Constellations 1. Constellation satellites can have small shrapnel-detecting cameras: Ø Iridium, Global. Star, One. Web, & GEO birds can monitor near their altitude. Ø Even small cameras can detect nearby shrapnel having low relative velocity. Ø Once found, shrapnel might be tracked by TDI imaging or ground telescopes. 2. Ion-beam shepherding can move dead satellites: Ø Push on & despin tumbling objects without contact. Ø Reentry is not needed: move to less crowded altitudes. 3. Will smart self-interest reliably lead to “good enough” practices? Ø Rates of random collision scale with Sqr(total # of non-maneuvered objects). Ø Iridium now has ~15 dead birds at its altitude, + 6 other intact dead objects/km. Ø How many will One. Web eventually have, if it has ~10 X as many satellites? My main interest is LEO, but the above applies to GEO as well as LEO. 2016 CODER Workshop Tether Applications, Inc. Nov 2016, pg. 12
Can “Laser Nudging” Prevent Debris Collisions? This combines parts of two prior laser concepts: 1. Orion (Phipps) uses ablation to deorbit all debris, & must also find small debris. Ø We want to slightly nudge tracked objects, days before worrisome conjunctions. 2. Light. Force (NASA Ames) uses light pressure to nudge debris, to avoid collisions. Ø This requires extremely good debris orbit predictions, since light pressure is so weak. Ø Ames focused on debris <100 kg, but most shrapnel will come from debris >1000 kg. But ablation at long range requires expensive large lasers & telescopes! 1. Ablation gives >10, 000 X the impulse of light pressure, but only if intense enough. 2. So find minimum laser + telescope sizes that can prevent most shrapnel creation. 3. Then figure out what else might be done by such a system (possibly enhanced). 4. Adaptive Optics + “Lucky Image” timing (AO + LI) can greatly improve focus. 2016 CODER Workshop Tether Applications, Inc. Nov 2016, pg. 13
We Also Need an Exit Plan! Nudging forever = paying forever; an exit plan is needed! 1. Reentry imposes strict liability; wholesale reentry may not be politically viable. 2. But “ventilating” large objects can greatly reduce how much survives reentry. 3. Collecting much of the mass appears feasible, due to its inclination clustering. 4. Tumbling debris is a challenge, but ion shepherds or nudging can stop tumbling. Consider “tethered orbiting scrapyards” 1. >80% of mass is in 6 narrow inclination bands + 1 wider band (sun-synch). 2. Capture with EDDE or ion-engine tug when scrapyard nodes align w/debris. 3. Deliver to the (briefly coplanar) scrapyard, kept at a less crowded altitude. 4. Maneuver scrapyards to dodge loose debris. 5. Decide later on reentry vs recycling in orbit. Nodal regression rates vary w/altitude & inclination 2016 CODER Workshop Tether Applications, Inc. Nov 2016, pg. 14
Collect Spent Stages; Recycle and/or Deorbit Later 1 2 3 4 5 1. Deliver “scrapyard tethers” to 74º & 82º orbits at 650 -750 or ~1050 km. 2. EDDE or ion rockets capture stages at nodal co-incidence: 3. Then they drag the stages to the scrapyards and hand them over. 4. One can add rockets later, to target scrapyard deorbit from low altitude. 5. Or add devices to recycle most of the mass and ventilate the rest. 6. Retrieve the scrapyard tether & cut up the stored stages, one at a time. 7. EDDE takes stacked “shingles” to customers; ventilated stages reenter. 2016 CODER Workshop Tether Applications, Inc. Nov 2016, pg. 15
How Should We Decide How to Manage Debris? 1. Estimate the current costs and cost trends of debris in LEO and GEO. Ø We know even less about the cost of “non-management” than about many fixes. Ø The cost of not managing debris is driven by satellite losses to shrapnel impact. Ø We must learn a lot more about current and likely future shrapnel costs! 2. Develop models of costs & benefits of different intervention options. Ø Regulatory agencies have procedures for approving models that affect policies. Ø Industry has a chance to participate and challenge key model assumptions. 3. Instead of license yes/no, let debris costs set Pigovian fees/bounties: Ø Pigovian fees penalize private acts that impose substantial public costs. Ø Pigovian bounties reward private acts that provide substantial public benefit. Ø Parking fees can fund debris removal bounties that null out public cost increases. If LEO and GEO users won’t pay parking fees for new net additions to debris costs, it seems unjust for general funds to pay for any removals! 2016 CODER Workshop Tether Applications, Inc. Nov 2016, pg. 16
The Cost of Debris and Debris Management 1. Surprisingly, little effort has been put into estimating the cost of debris. Ø We must estimate debris costs & trends to justify real $ investments in fixes! 2. How can we “run an honest horse race” involving different fix options? Ø Different options have mostly different targets, limitations, and idiosyncrasies! Ø If a timely nudge prevents a collision, it may be more valuable than a full removal! Ø And how valuable is it to move debris to a less crowded altitude, but not deorbit it? Ø A “prize” for the first removal only makes sense with a real follow-on market. Ø And how do you adjust payments when there are unintended side-effects? 3. Consider a continuing bounty program that pays for net benefit on future costs: Ø This requires an accepted means of estimating the value of different changes. Ø This is needed anyway to pay fairly for nudging, removal, relocation, etc. Ø Several countries might contribute if their businesses can collect bounties. Ø If LEO users pay fair parking fees, general funds might pay for net reduction. 2016 CODER Workshop Tether Applications, Inc. Nov 2016, pg. 17
Parking Fees Must Offset New Costs Added Debris cost changes form a continuum: + Fragment by impact, explosion, or botched removal Put new non-maneuvering object into crowded orbit Launch new object into a low or uncrowded orbit 0 No change—eg, benign failure of launch or removal Nudge an object to preclude a potential collision Move an object from crowded to uncrowded altitude - Deboost an object into a short-lived orbit or reentry We expect free parking in the suburbs, but not downtown (LEO). If those adding costs don’t pay for removal, why should anybody else? So to get $ for net reduction, LEO users should null out risk additions. 2016 CODER Workshop Tether Applications, Inc. Nov 2016, pg. 18
Conclusions 1. Decision makers need $-based figures of merit for debris. 2. The main direct debris cost is from shrapnel disabling satellites. We don’t yet know that direct cost, or many of the indirect costs. 3. Better detection & tracking can both determine & reduce costs. 4. But if LEO users won’t pay to null out their own cost additions, allocation of “general funds” for net removal seems unjustified. 5. Non-technical issues will affect selection of technical options: debris management is not just a technical puzzle; it is wrapped in an political/economic mystery, inside a diplomatic enigma. And as Winston Churchhill went on to say, back in 1939: “Perhaps there is a key. That key is Russian national interest. ” 2016 CODER Workshop Tether Applications, Inc. Nov 2016, pg. 19
Debris Overviews Bibliography UN Treaties on space: www. oosa. unvienna. org/oosa/en/Space. Law/treaties. html. US Space Policy: www. whitehouse. gov/sites/default/files/national_space_policy_6 -28 -10. pdf NASA & Aerospace Corp websites: http: //orbitaldebris. jsc. nasa. gov/, www. aerospace. org/cords/ ESA debris website: www. esa. int/Our_Activities/Operations/Space_Debris Heiner Klinkrad, Space Debris: Models and Risk Analysis, Springer/Praxis, 2006. Secure World Foundation (focuses on sustainability of space usage): http: //swfound. org/ Debris Collision, Costs, and “Commons” Analyses of Cosmos/Iridium impact & other impact issues: http: //wangting. org/pages/ Analysis of BLITS impact event: www. celestrak. com/publications/AMOS/2013/AMOS-2013 a. pdf Bill Ailor et al. , “Effects of Space Debris on the Cost of Space Operations, ” paper IAC-10. A 6. 2. 10 Papers by Levin & Carroll on debris costs, at www. star-tech-inc. com/id 27. html Garrett Hardin, “Tragedy of the Commons”: www. sciencemag. org/content/162/3859/1243. full see also the extensive discussion at http: //en. wikipedia. org/wiki/Tragedy_of_the_commons EDDE See papers and presentations at www. star-tech-inc. com/id 27. html Optical Tracking and Laser Debris Removal Pulsed laser ablation nudging paper IAC-14, A 6. P, 52 x 24670 by Carroll; please request from tether@cox. net AMOS Conference proceedings (optical tracking, etc. ): http: //amostech. com/? page_id=20 Laser light pressure for nudging: www. amostech. com/Technical. Papers/2013/Orbital_Debris/STUPL. pdf Laser ablation for debris removal: http: //arxiv. org/ftp/arxiv/papers/1110. 3835. pdf www. sciencedirect. com/science/article/pii/S 0094576513002749 2016 CODER Workshop Tether Applications, Inc. Nov 2016, pg. 20
Backup Slides on Managing Orbital Debris 2016 CODER Workshop Tether Applications, Inc. Nov 2016, pg. 21
Debris and the 1972 UN Liability Convention Key details in 1972 UN Convention on Liability for Space Objects: 1. Damage from reentry poses full liability; in space, it is “for fault” (undefined!). 2. If A launches B’s payload from C’s territory, all 3 have launching state liability. 3. Selling and/or re-registering an object do not transfer the launching state liability. 4. Damaging another state’s space object makes you share in its future liabilities. 5. But the convention lets signatories agree “off-line” on reimbursement terms for different cases, for losses suffered due to liability payouts under the convention. Possible implications: 1. Nudging US + Russian debris below 1020 km can greatly reduce shrapnel creation. 2. This prevents collisions now, while letting actual debris removal take more time. 3. A bilateral US/Russia agreement may be both necessary and sufficient; no UN or other multi-lateral agreement seems needed to deal with most current LEO debris. 2016 CODER Workshop Tether Applications, Inc. Nov 2016, pg. 22
What Might a Laser Nudging System Look Like? Nudging either of 2 debris objects can prevent most predicted collisions: 1. For 1 km/day shift, we need 4 mm/sec DV (~1/50, 000 of a typical deorbit DV!). 2. More lead time & better orbit predictions reduce required nudge size & number. 3. With perfect predictions we would need only 1 tiny nudge, every 5 -10 years! Ø So improve tracking and prediction before designing a laser nudging system! What might typical laser + telescope performance requirements be? 1. The hot-spot flux varies with both range and angle of incidence: Ø Flux at zenith on a surface ~45 o to beam = flux 30 o from zenith, angled 25 o to beam. 2. Consider 4 mm/sec along-track DV on a 1434 kg Kosmos-3 M, at 45 o angle error: Ø This needs 8 newton-sec impulse. If Cm=20 dyn-s/J, we need 40 k. J delivered energy. Ø This might involve ~8 pulses each with ~5 k. J in the hot spot (~10 k. J at the laser). • A Zenit may need ~50 pulses, over several passes if the laser’s heat capacity is low. Ø Assume Il√t is >5 W(√sec)/cm to ensure ablation, with l=1. 06 E-4 cm & t=1 E-8 sec. Ø Then the hot spot needs 5 J/cm 2 in 1000 cm 2: 30 cm dia at 45 o. Ø At 1000 km, w/SRS & imperfect AO + LI: ~8 m mirror dia? 2016 CODER Workshop Tether Applications, Inc. Nov 2016, pg. 23
A-sat Issues 1. All effective debris management options have some A-sat functionality! 2. Doing nothing lets limited A-sat use hide behind failures due to shrapnel. 3. International acceptance of laser ablation of space objects may require: a) b) c) d) Advance public listing of planned targets On-site means to verify that no unplanned objects are lased Recognition that most objects needing lasing will be Russian Accepting that several countries will want their own lasers… Zenit 3. 9 x 11 m 8300 kg, 71 o 2016 CODER Workshop ? One. Web, OTB, etc. Iridium ~2 x 4 x 8 m, 550 kg, 86 o Kosmos-3 M 2. 4 x 6 m, 1434 kg, 83 o Tether Applications, Inc. Nov 2016, pg. 24
What About GEO? 1. LEO has higher collision rates, energy, & chance of lethal damage. 2. It is more feasible to quantify, detect, & dodge shrapnel in LEO. 3. Many of the management options for LEO & GEO are different. 4. Commercial GEO operators may get proactive about their interests. 5. This may lead to regular laser or ion-beam nudging of GEO debris. 6. Current GEO SSA surveys may improve awareness & practices. 2016 CODER Workshop Tether Applications, Inc. Nov 2016, pg. 25
Abstract for CODER 2016 Workshop Better practices have greatly reduced the rate at which we are adding to orbital debris. But future large constellations will eventually include enough failed satellites that their collisions alone will cause a costly growth of debris fragments too small to track but too heavy to shield against. It is time to start proactively managing orbital debris in LEO, and planning for it in GEO. Current management options are costly. Surprisingly, we know even less about what debris management is worth than how much it might cost. We do not even know how much small untracked debris has already cost in satellite failures. A cost-effective approach seems to require all of the elements listed below, because anything less will neglect important parts of the problem: 1. Learn more about the amount and likely costs of lethal orbital debris, in both LEO and GEO. 2. Develop good ways to estimate the costs of future changes, including new constellations. 3. Fund technology developments likely to best reduce the costs of debris + debris management. 4. Create a regulated commercial bounty program that pays for actual reductions in future costs. 5. Collect “parking fees” that fund enough bounties to null out the net effects of future launches. 6. Work with Russia, since it owns 2/3 of the mass now at congested altitudes (450 -1050 km). 2016 CODER Workshop Tether Applications, Inc. Nov 2016, pg. 26
- Slides: 26