Case Study: Designing an End-of-Life Disposal System for an 800 km Constellation

"The cheapest satellite to remove is the one that removes itself, on schedule, every time."

Executive Summary

In the first case study we took a collision apart. Here we do the harder, forward-looking thing: we design a constellation so that its collision never happens — specifically, so that every one of its satellites leaves orbit on time when it dies. We are given a science requirement that forces a high, congested orbit ($800\ \text{km}$, sun-synchronous), and a disposal rule that orbit cannot satisfy on its own. We will size three candidate disposal strategies — natural decay, propulsive de-orbit, and a drag sail — price each in mass and delta-v, weigh their reliability, and recommend a complete end-of-life plan that a regulator would accept. This is exactly the analysis a real constellation operator must now win before it is granted a license, and the one you will run for your own mission in the Mission Design Checkpoint.

Skills applied

  • Computing natural-decay lifetime and testing it against disposal rules (§35.3).
  • Sizing de-orbit propellant by inverting the rocket equation (Chapter 3).
  • Sizing a drag device from the ballistic coefficient (Chapter 12, §35.4).
  • Estimating fleet collision risk (§35.2) and engineering for disposal reliability (Chapter 32).

Background

The requirement

  • Mission: a $600$-satellite Earth-observation constellation (Chapter 33), each satellite $m = 150\ \text{kg}$.
  • Orbit: a sun-synchronous orbit at $800\ \text{km}$ — chosen, as Chapter 12 showed, because $J_2$ precession there gives constant-lighting imagery. The altitude is a science requirement, not a free choice.
  • Disposal rule: the constellation is U.S.-licensed, so it must comply with the 5-year post-mission disposal rule (§35.3). Each satellite has a $5$-year operational life, then must be removed within $5$ years of retirement.
  • Each satellite presents a compact cross-section $C_D A \approx 1.5\ \text{m}^2$, giving a ballistic coefficient $\beta = m/(C_D A) = 150/1.5 = 100\ \text{kg/m}^2$.

The one hard constraint

The orbit is fixed at $800\ \text{km}$ by the science, and — as we are about to confirm — that is exactly the altitude where nature refuses to help. Every option below is a way of buying back the disposal that the required orbit denies us for free.

Phase 1: Does it decay on its own? (No — badly)

Apply the natural-lifetime estimate from §35.3, $\tau \approx H\beta/(\rho_0 a v)$. At $800\ \text{km}$: $a = 6{,}371 + 800 = 7{,}171\ \text{km} = 7.171\times10^{6}\ \text{m}$, and $$v = \sqrt{\mu/a} = \sqrt{\frac{3.986\times10^{14}}{7.171\times10^{6}}} = 7{,}456\ \text{m/s}.$$ With $\beta = 100\ \text{kg/m}^2$, scale height $H \approx 80\ \text{km} = 8.0\times10^{4}\ \text{m}$, and a thin $\rho_0 \approx 1\times10^{-14}\ \text{kg/m}^3$ (Tier 2, flagged — this swings by $10\times$ over the solar cycle): $$\tau \approx \frac{(8.0\times10^{4})(100)}{(1\times10^{-14})(7.171\times10^{6})(7{,}456)} = \frac{8.0\times10^{6}}{5.35\times10^{-4}} = 1.5\times10^{10}\ \text{s} \approx 474\ \text{years}.$$ Sanity check: several centuries, matching the Chapter 12 lifetime table. This fails the 5-year rule by a factor of roughly $95$, and even the old 25-year rule by a factor of nearly $20$. Left to decay, $600$ dead satellites would loiter in the most congested shell in space for longer than the entire history of spaceflight so far — twenty times over. Passive disposal is off the table. We must actively remove each satellite.

Phase 2: Option A — propulsive de-orbit

The direct fix: carry propellant to lower each satellite's perigee into the atmosphere at end of life. From $800\ \text{km}$ circular ($v_1 = 7{,}455\ \text{m/s}$), drop the perigee to $\sim 120\ \text{km}$ ($r_p = 6{,}491\ \text{km}$) so the orbit re-enters within weeks. The transfer ellipse has $a_t = (7{,}171 + 6{,}491)/2 = 6{,}831\ \text{km}$; by vis-viva its speed at the $800\ \text{km}$ apogee is $$v_{\text{apo}} = \sqrt{\mu\!\left(\frac{2}{7{,}171} - \frac{1}{6{,}831}\right)} = \sqrt{398{,}600 \times 1.325\times10^{-4}} = 7{,}267\ \text{m/s},$$ so the de-orbit burn is $$\Delta v = v_1 - v_{\text{apo}} = 7{,}455 - 7{,}267 = 188\ \text{m/s}.$$ Now size the propellant. With a small monopropellant thruster at $I_{sp} = 220\ \text{s}$ ($v_e = 220 \times 9.81 = 2{,}158\ \text{m/s}$), the rocket equation (Chapter 3) inverts to $$m_p = m\left(1 - e^{-\Delta v/v_e}\right) = 150\left(1 - e^{-188/2158}\right) = 150(1 - 0.9166) = 12.5\ \text{kg}.$$ So each satellite reserves about $13\ \text{kg}$ of disposal propellant — $8.3\%$ of its wet mass — and the fleet carries $600 \times 12.5 = 7{,}500\ \text{kg}$ of propellant whose only job is to leave orbit. That is a real cost (this is theme four, mass is the enemy: $13\ \text{kg}$ per satellite is $13\ \text{kg}$ that is not payload), but it is unavoidable at this altitude and it works: the satellite re-enters within weeks, cleanly compliant with the 5-year rule.

Phase 3: Option B — a drag sail

Can we avoid carrying propellant by lowering the ballistic coefficient instead? Deploy a drag sail at end of life to increase $C_D A$ and let drag finish the job (§35.4). Suppose we unfurl a $15\ \text{m}^2$ sail ($C_D \approx 2.2$), adding $2.2 \times 15 = 33\ \text{m}^2$ of $C_D A$ for a total of $34.5\ \text{m}^2$: $$\beta_{\text{new}} = \frac{150}{34.5} = 4.3\ \text{kg/m}^2.$$ Since $\tau \propto \beta$, the lifetime shrinks by the factor $4.3/100 = 0.043$, from $474$ years to $$\tau_{\text{sail}} \approx 474 \times 0.043 \approx 21\ \text{years}.$$ Sanity check and the catch: a picnic-blanket-sized sail turns a five-century derelict into a $21$-year one — a spectacular improvement, and enough to clear the old 25-year rule. But it still fails the 5-year rule by a factor of four. To reach $5$ years we would need $\beta \approx 1\ \text{kg/m}^2$, which means $C_D A \approx 150\ \text{m}^2$ — a sail of order $65\ \text{m}^2$ (about $8\times8\ \text{m}$) on a $150\ \text{kg}$ satellite, unwieldy to deploy and, worse, a sail that huge increases the collision cross-section by fortyfold for two decades on the way down, raising the very risk it was meant to reduce. At $800\ \text{km}$, drag augmentation is a partial fix at best; the higher you fly, the less a sail can do.

🔧 Engineering Reality: the disposal method must match the altitude. The two options are not interchangeable — the physics picks the winner by altitude. A drag sail is elegant and propellant-free where drag is already strong (below $\sim 600\ \text{km}$, a modest sail easily meets even the 5-year rule). At $800\ \text{km}$ it is marginal, and above $\sim 900\ \text{km}$ it is nearly useless. Propulsive de-orbit costs mass but works at any LEO altitude. Since our science orbit is fixed at $800\ \text{km}$, the sail cannot meet our rule and propulsion must. This is why you cannot design a disposal system in the abstract: you design it for the orbit the mission demands.

Phase 4: The risk while it flies, and the reliability of the fix

Two more numbers decide the design. First, how much debris risk does the fleet run during its life? Using the §35.2 flux model with $\Phi \approx 1\times10^{-5}\ \text{m}^{-2}\text{yr}^{-1}$ ($>1\ \text{cm}$ debris, Tier 3), each satellite of area $A \approx 3\ \text{m}^2$ over its $5$-year life has expected impacts $N = (10^{-5})(3)(5) = 1.5\times10^{-4}$. Across $600$ satellites, the expected number of catastrophic strikes is $$600 \times 1.5\times10^{-4} = 0.09,$$ so there is roughly a $9\%$ chance of losing at least one satellite to debris over a fleet-generation — and every such loss is a new derelict that our disposal system can no longer command. This reframes the whole design: disposal is not only about our own retirements, it is about not letting a debris strike convert one of our satellites into an uncontrollable hazard.

Which leads to the second number — reliability. A de-orbit that only works when the satellite is healthy is worthless, because satellites are disposed of because they have failed. From Chapter 32: if each satellite's de-orbit succeeds with probability $0.95$, then across $600$ satellites we expect $600 \times 0.05 = 30$ stranded derelicts — thirty new long-lived hazards in the worst shell in space, from a fleet that was "compliant" on paper. To keep the expected strandings low (say, below one or two), the per-satellite disposal reliability must be pushed toward $99.7\%$ or better — which means the de-orbit function must be a robust, independent, and tested subsystem, able to fire even after the payload and main bus have died.

Phase 5: The recommendation and the disposal plan

The physics has made the decision for us. At a science-mandated $800\ \text{km}$, natural decay fails by $95\times$ and a practical drag sail fails the 5-year rule by $4\times$; only propulsive de-orbit complies. The end-of-life disposal plan for the MDR:

  1. Method: active propulsive de-orbit. Reserve $\Delta v = 190\ \text{m/s}$ ($\approx 13\ \text{kg}$ propellant, $8.3\%$ of wet mass) per satellite, lowering perigee to $\sim 120\ \text{km}$ for re-entry within weeks, well inside the 5-year rule.
  2. Reliability: implement de-orbit as an independent, high-reliability subsystem (redundant valves, independent power, autonomous trigger on loss of contact) targeting $> 99.5\%$ disposal success, so that expected strandings across the fleet stay near zero (Chapter 32).
  3. Passivation: after the de-orbit burn, vent remaining propellant, safe the batteries, and de-spin, so a satellite that fails to re-enter promptly cannot later explode (§35.3).
  4. Collision avoidance while operational: subscribe to conjunction data, maneuver autonomously when $P_c > 10^{-4}$ (§35.3), and reserve a small station-keeping/avoidance propellant margin on top of the de-orbit reserve.
  5. Design for demise: ensure the $150\ \text{kg}$ satellites burn up fully on re-entry (casualty risk $< 1$ in $10{,}000$), especially dense reaction wheels and tanks (Chapter 7).

💡 Intuition: pay a little, always, instead of a lot, never. The $13\ \text{kg}$ of de-orbit propellant per satellite feels like a tax on payload — until you compare it to the alternative. Active removal of a stranded satellite (§35.4) would cost a dedicated spacecraft, a hazardous rendezvous with a tumbling target, and a legal negotiation, easily thousands of times more per object than the propellant that would have de-orbited it. Self-disposal is the overwhelmingly cheaper path, precisely because it happens while the satellite is still alive and controllable. The whole philosophy of sustainable design is to spend the small, certain cost up front rather than bequeath the large, uncertain one to everyone who flies later.

Discussion Questions

  1. The science required $800\ \text{km}$. If the mission could tolerate $550\ \text{km}$ instead, how would the disposal design change, and roughly what would natural decay give you there? (Estimate with the §35.3 formula.)
  2. A $15\ \text{m}^2$ drag sail met the old 25-year rule but not the 5-year rule. Argue whether the fivefold tightening of the rule was, in effect, a decision to mandate propulsion on high-altitude satellites.
  3. Phase 4 found a $9\%$ chance of a debris strike over the fleet's life. How does that risk change the reliability requirement on the de-orbit system, and why?
  4. Compare the total fleet disposal propellant ($7{,}500\ \text{kg}$) to the cost of actively removing even a handful of stranded satellites. Why is "self-disposal, reliably" the only economically sane policy at this scale?

Your Turn: Extensions

  • Option A (design). Redo the whole analysis for satellites operating at $1{,}000\ \text{km}$. Show that both natural decay and drag sails fail badly, and compute the de-orbit $\Delta v$ and propellant. Is there an altitude above which even propulsive de-orbit becomes prohibitively expensive relative to a graveyard?
  • Option B (computation). Write a Python tool disposal_report(alt_km, m, cda, isp) that prints the natural lifetime, the de-orbit $\Delta v$ and propellant, and a pass/fail against the 5-year rule. Reproduce this case study's numbers. (Do not run it; hand-trace and add # Expected output:.)
  • Option C (your mission). Apply this exact three-option analysis (decay / de-orbit / drag device, plus passivation and reliability) to your mission's disposal, and record the chosen plan and its delta-v in your MDR.

Key Takeaways

  1. The required orbit dictates the disposal method. At a science-mandated $800\ \text{km}$, natural decay ($\sim 470$ yr) and a practical drag sail ($\sim 21$ yr) both fail the 5-year rule; only propulsive de-orbit ($\sim 190\ \text{m/s}$, $\sim 13\ \text{kg}$/satellite) complies. Design disposal for the orbit, not in the abstract.
  2. Disposal is a mass line you reserve, not leftover you hope for. Set aside the de-orbit propellant explicitly ($8.3\%$ of wet mass here); a satellite that spends its last kilograms on operations cannot dispose of itself.
  3. Reliability is half the design. Satellites are disposed of because they have failed, so the de-orbit function must be independent and robust; at $95\%$ reliability a $600$-satellite fleet strands $\sim 30$ derelicts — nominal compliance, real disaster.
  4. Pay the small certain cost, not the large uncertain one. Self-disposal while alive is thousands of times cheaper than active removal after stranding — the economic heart of sustainable constellation design.