Case Study: Designing a Satellite Bus — Closing the Mass Budget and Sizing the Structure

"An engineer can do for a dollar what any fool can do for two — and for a rocket, read 'a dollar' as 'a kilogram.'" — after a saying attributed to the engineer Arthur Wellington

Executive Summary

Where the first case study audited a finished vehicle, this one builds one. We will design the structure of a small $500\ \text{kg}$-class satellite bus from a blank sheet: lay out its mass budget with margin, size its primary structure against launch loads, size a propellant tank against internal pressure, run a material trade, and confirm the whole design closes under a launch-vehicle mass cap. Along the way we will hit the chapter's central surprise in a new form — that the primary structure is not limited by material strength at all, but by buckling and stiffness, so "use a stronger material" is the wrong lever. The result is a realistic first-cut structural design in which different materials do different jobs, and every number traces back to a section of this chapter.

Skills applied

• Building and closing a mass budget with margin against a launch cap (§23.5).
• Sizing primary structure for quasi-static launch loads and finding the real constraint (§23.1–23.2).
• Computing stress and applying a factor of safety; recognizing buckling vs. strength limits (§23.3–23.4).
• Running a specific-strength/stiffness material trade and picking per-part (§23.3).

Background

Our design target: a $500\ \text{kg}$ (dry) satellite bus for a mission that needs $\Delta v = 1.5\ \text{km/s}$ of onboard propulsion (say a GTO-to-GEO circularization plus station-keeping), using a bipropellant system at $I_{sp} = 320\ \text{s}$. It will ride as a rideshare with a launch-mass cap of $1{,}000\ \text{kg}$ to the drop orbit. The structure must survive a quasi-static launch load of $6\ \text{g}$ axial (per a representative launch-vehicle user's guide, §23.1) and provide enough stiffness that the bus does not resonate with the vehicle. All numbers here are Tier 3 (constructed for a plausible teaching design), chosen to be self-consistent and to land in realistic ranges.

Phase 1: Lay out the mass budget

Begin with the ledger of §23.5. Allocate the $500\ \text{kg}$ dry mass across subsystems using the representative shares as a first cut, then carry a $20\%$ margin (this is a maturing design, past the earliest concept, so a little below the $25$–$30\%$ of a first sketch):

Now add the two things the rocket equation forces on us: a mass margin, and the propellant implied by the delta-v budget. The mass ratio for $1.5\ \text{km/s}$ at $I_{sp} = 320\ \text{s}$ is $e^{\Delta v/v_e}$ with $v_e = 320\times9.80665 = 3{,}138\ \text{m/s}$:

import math
dry = 500.0
allocated = dry * 1.20                       # 20% mass margin
R = math.exp(1500 / (320 * 9.80665))         # mass ratio for 1.5 km/s
wet = allocated * R                           # launch (wet) mass
cap = 1000.0
status = "closes" if wet < cap else "OVER"
print(f"dry {dry:.0f}, +margin {allocated:.0f}, wet {wet:.0f} kg, "
      f"cap {cap:.0f} -> {status} ({cap - wet:.0f} kg spare)")
# Expected output:
# dry 500, +margin 600, wet 968 kg, cap 1000 -> closes (32 kg spare)

The design closes — but only by $32\ \text{kg}$, about $3\%$ of the cap. That is a tight budget, exactly the knife-edge a real spacecraft lives on, and it means every subsystem, structure first, is under pressure to hold or cut mass. Sanity check: propellant is $\text{wet} - \text{allocated} = 968 - 600 = 368\ \text{kg}$, a propellant fraction of $368/968 \approx 38\%$ — reasonable for $1.5\ \text{km/s}$ of chemical propulsion.

Phase 2: Size the primary structure

The bus's primary structure is a central thrust tube — a cylinder that carries the whole satellite's inertial load from its base (bolted to the launch adapter) up through the spacecraft during ascent. Size it for the worst-case axial load. Conservatively, take the tube as carrying the full launch mass at $6\ \text{g}$:

$$ F_{\text{axial}} = m\,n\,g_0 = 968\ \text{kg}\times6\times9.81\ \text{m/s}^2 \approx 5.70\times10^{4}\ \text{N} = 57\ \text{kN}. $$

Let the tube be $0.6\ \text{m}$ in diameter (radius $r = 0.3\ \text{m}$) with a wall thickness $t = 2\ \text{mm}$. The load-bearing cross-section is the thin ring, area $A = 2\pi r\,t = 2\pi(0.3)(0.002) = 3.77\times10^{-3}\ \text{m}^2$. The compressive stress is

$$ \sigma = \frac{F}{A} = \frac{5.70\times10^{4}}{3.77\times10^{-3}} \approx 1.51\times10^{7}\ \text{Pa} = 15.1\ \text{MPa}. $$

Now compare against an aluminum yield of $\sim 350\ \text{MPa}$: the factor of safety on yield is $350/15.1 \approx 23$. That is not a lean aerospace margin (§23.4) — it is enormous. Something is off with our instinct: the tube is absurdly overbuilt against strength. The resolution is the point of §23.3.

🚪 Threshold Concept: the thin tube fails by buckling, not by yielding. A stress of $15\ \text{MPa}$ is nowhere near aluminum's $350\ \text{MPa}$ yield, yet we cannot simply keep thinning the wall, because long before the metal yields the thin shell will buckle — fold up like a crushed can under compression. The classical critical buckling stress of a thin cylinder is $\sigma_{\text{cr}} \approx 0.6\,E\,(t/r)$, and real shells reach only a fraction of it (a "knockdown factor" near $0.25$ for an unstiffened cylinder). With $E = 70\ \text{GPa}$ and $t/r = 0.002/0.3 = 0.0067$: $$\sigma_{\text{cr}} \approx 0.25\times0.6\times70\times10^{9}\times0.0067 \approx 7.0\times10^{7}\ \text{Pa} = 70\ \text{MPa}.$$ So the buckling allowable is $\sim 70\ \text{MPa}$, giving a buckling margin of $70/15.1 \approx 4.6$ — a factor of five smaller than the yield margin of $23$. Buckling, not strength, sizes the tube. And buckling resistance depends on $E$ and geometry, not on strength — which is exactly why, per §23.3, a stronger material would not help, but a stiffer or cleverer geometry (or internal pressure) would. (The knockdown factor is illustrative — real shell-buckling analysis is subtle; flagged.)

Two consequences follow, both straight from the chapter. First, since strength is not the limit, there is no reason to pay for a high-strength material here — cheap aluminum is as good as titanium for a buckling-limited tube (recall $E/\rho$ is nearly identical for the metals, §23.3). Second, the final wall thickness will likely be set by an even softer requirement: stiffness, so that the bus's first natural frequency clears the launch vehicle's requirement (commonly $> 35\ \text{Hz}$ axial) and it does not resonate with the vehicle during the vibration environment of §23.1. We would tune $t$ and add ring stiffeners to hit that frequency — not to survive a static stress it already beats by 20×.

Phase 3: Size a propellant tank

The onboard bipropellant system needs tanks. Size a spherical fuel tank of radius $r = 0.35\ \text{m}$ held at a pressurant pressure $p = 20\ \text{bar} = 2.0\times10^{6}\ \text{Pa}$. For a sphere, the wall stress is the same in every direction, $\sigma = pr/(2t)$ (half the cylindrical hoop value), so the wall thickness needed for an allowable stress $\sigma_{\text{allow}}$ (with the factor of safety already applied) is $t = pr/(2\sigma_{\text{allow}})$. Titanium is the classic propellant-tank metal — high specific strength and compatible with hydrazine/NTO — so take $\sigma_{\text{allow}} = 500\ \text{MPa}$:

$$ t = \frac{p\,r}{2\,\sigma_{\text{allow}}} = \frac{2.0\times10^{6}\times0.35}{2\times5.0\times10^{8}} = \frac{7.0\times10^{5}}{1.0\times10^{9}} = 7.0\times10^{-4}\ \text{m} = 0.7\ \text{mm}. $$

The tank mass is the wall volume times density, $m = \rho\,(4\pi r^2)\,t$, with $\rho_{\text{Ti}} = 4{,}430\ \text{kg/m}^3$:

$$ m = 4{,}430\times(4\pi\times0.35^2)\times7.0\times10^{-4} = 4{,}430\times1.539\times7.0\times10^{-4} \approx 4.8\ \text{kg}. $$

A pair of tanks (fuel and oxidizer) is then about $9.5\ \text{kg}$ — comfortably inside the $55\ \text{kg}$ propulsion-dry allocation from Phase 1. Sanity check the thickness: $0.7\ \text{mm}$ for a $0.35\ \text{m}$ titanium sphere at $20\ \text{bar}$ is thin but plausible; a lighter design would overwrap it with carbon fiber (a COPV), trading manufacturing complexity for mass — a trade we would take only if the budget were not closing.

Phase 4: The material trade — no single answer

Collect the decisions. The chapter's four materials each win a different part of this spacecraft, which is the realistic outcome — there is no one "best" spacecraft material, only the best material for each load path.

🔧 Engineering Reality: match the material to the load path. A novice picks one favorite material for the whole spacecraft; an experienced structures engineer picks per-part, because the driver differs from part to part — buckling here, pressure there, stiffness somewhere else — and §23.3's figures of merit rank the materials differently for each driver. The thrust tube wants stiffness-per-mass (all metals tie, so pick the cheap one); the pressure tank wants strength-per-mass and chemical compatibility (titanium); the dimensionally stable antenna wants stiffness-per-mass at its best (composite). The same table in §23.3, read down different columns, gives different winners.

Phase 5: Confirm closure and iterate

Return to the budget. Structure came in at $120\ \text{kg}$ (Phase 1), and our Phase 2–3 sizing suggests the primary tube and tanks are comfortably inside that, with the tube likely set by stiffness (so its final mass depends on the frequency requirement more than on the loads). The wet mass closed under the $1{,}000\ \text{kg}$ cap with $32\ \text{kg}$ to spare (Phase 1). If a later mass-growth event ate that $32\ \text{kg}$ — as §23.5 warns it will — we would recover it by the levers this chapter provides: thin the buckling-limited tube (it has 23× yield margin to give, limited only by stiffness), overwrap the tanks as COPVs, or, at the program level, buy back mass by accepting a slightly higher launch cost. What we would not do is add a fat factor of safety "to be safe" — that spends the very margin we are fighting to keep.

💡 Intuition: A spacecraft design is a negotiation between the mass budget and every subsystem, refereed by the rocket equation. Structure is usually where the negotiation is hardest, because structure is heavy, its loads are the launch loads (the worst of the mission), and its margins are already the thinnest in engineering. Closing the budget is not a one-time calculation but a loop: size, sum, check the cap, and when growth eats the margin, find the mass again — most often in the buckling-limited primary structure, where strength margins are deceptively huge and the real limit is stiffness.

Discussion Questions

1. The thrust tube had a yield factor of safety of $\sim 23$ but a buckling margin of only $\sim 4.6$. Explain why quoting the yield factor of safety would badly mislead a reviewer about how much mass the tube can give back.
2. We chose aluminum for the buckling-limited tube even though titanium is far stronger. Justify this using $E/\rho$ and cost, referencing §23.3.
3. The budget closed with only $32\ \text{kg}$ ($3\%$) of margin. Given §23.5's warning about mass growth, is this design healthy at this stage? What margin would you want, and where would you find more if needed?
4. For the propellant tank we used $\sigma = pr/(2t)$ (a sphere) rather than $pr/t$ (a cylinder). Why is a sphere the more mass-efficient pressure-vessel shape, and what non-structural reason might still push a designer toward cylindrical tanks?

Your Turn: Extensions

• Option A (analysis). Recompute the thrust-tube stress if the launch load factor is $8\ \text{g}$ instead of $6\ \text{g}$. Does the tube still beat yield by a wide margin? Is buckling still the limit? What does this say about which load case sizes buckling-limited structure?
• Option B (computation). Write tank_wall(p, r, sigma_allow, shape="sphere") returning the required wall thickness (use $pr/2\sigma$ for a sphere, $pr/\sigma$ for a cylinder), and tank_mass(rho, r, t, shape="sphere"). Reproduce the $0.7\ \text{mm}$, $4.8\ \text{kg}$ result and compare a cylindrical tank of the same radius. (Hand-trace; add # Expected output:.)
• Option C (design). Take your MDR spacecraft. Lay out its mass budget by subsystem, carry a margin, add propellant from your delta-v budget (Chapter 3), and check closure against a real launch vehicle's payload capacity (you will pick the vehicle in Chapter 30). Identify your single most buckling- or stiffness-critical structural element and note the material you would pick and why.

Key Takeaways

1. Close the budget first, then size. A mass budget with margin, checked against the launch cap, is the frame every structural decision fits inside; ours closed by only $3\%$, the realistic knife-edge.
2. Primary structure is often buckling-limited, not strength-limited. The thrust tube beat yield by 23× but buckling by only ~4.6×, so buckling (a stiffness/geometry problem) sizes it — and a stronger material would not help.
3. Pick materials per load path. Aluminum for the stiffness-limited tube, titanium for the pressure tanks, composite for stiffness-critical panels — the §23.3 table gives different winners read down different columns.
4. Find mass, don't add safety factor. When growth eats the margin, recover it by thinning over-margined structure or smarter design — never by padding the factor of safety, which spends the margin you are trying to protect.