Chapter 23: Structures and Materials

"Perfection is achieved, not when there is nothing more to add, but when there is nothing left to take away." — Antoine de Saint-Exupéry, Wind, Sand and Stars

Overview

For most of its life a spacecraft floats. In orbit it is weightless; nothing pushes on it harder than the faint pressure of sunlight and the whisper of a thruster. You might imagine, then, that a spacecraft's structure could be flimsy — a frame just stiff enough to hold its instruments in a line. And you would be exactly wrong, because before a spacecraft can float for a decade it must first survive about eight minutes of the most violent ride human engineering routinely inflicts on anything: it is bolted to the top of a controlled explosion, shaken until its teeth rattle, blasted with sound loud enough to be lethal to a human standing nearby, crushed under several times its own weight, and finally kicked free by small explosives. Everything structural about a spacecraft is designed for those eight minutes. The years of serene floating that follow ask almost nothing of it by comparison.

This is Part IV, where we leave the engine behind and build the rest of the vehicle — the structure, the thermal control, the power, the radios, the brains. Every one of these subsystems is a response to two pressures: the unforgiving space environment (theme #2), and the relentless tyranny of the rocket equation you met in Chapter 3, which taxes every kilogram exponentially. Structures is where those two pressures meet most directly. The job of a spacecraft's structure is to be strong enough to survive launch and light enough that the rocket equation forgives it — two demands in permanent tension, because strength usually comes from material and material is mass. The whole discipline is the art of resolving that tension: carrying the loads with the fewest possible kilograms. This chapter is where the book's fourth and most unglamorous theme — mass is the enemy — stops being a slogan and becomes a spreadsheet.

We proceed the way the book always does: physics first. We will find out what the launch environment actually is (§23.1), learn to tell the structure that matters from the structure that does not (§23.2), meet the four metals and composites that spacecraft are made of and the two numbers that decide between them (§23.3), put stress, strain, and the factor of safety on a firm footing (§23.4), and then assemble the master accounting tool of spacecraft engineering, the mass budget (§23.5). We close with a case study that ties it all together and advances our running Starship thread: why the most futuristic rocket ever built is made of the same stainless steel as a kitchen sink (§23.6).

In this chapter, you will learn to:

• Name and roughly quantify the four kinds of launch load, and say which one drives which part of a design.
• Trace a load path and tell primary structure (the vehicle dies without it) from secondary structure.
• Choose a material by specific strength and stiffness — strength per kilogram — not raw strength.
• Compute stress and strain, size a tank wall against internal pressure, and turn a factor of safety into a margin of safety.
• Build a mass budget with margin, and compute why 1 kg of structure can cost 5 kg of rocket.
• Explain the Starship steel decision as a whole-vehicle optimization, not a materials-science upset.

Learning Paths

🚀 Space Enthusiast: Read 23.1 (what launch does to a vehicle) and 23.6 (the Starship steel story), which are the two sections that will change how you watch a rocket and read a spec sheet. Skim the algebra in 23.4; enjoy the mass-budget logic in 23.5, which explains why engineers fight over grams.

📐 Engineering Student: Read everything. Sections 23.3 and 23.4 are the load-bearing wall of the chapter — specific strength/stiffness, hoop stress, and the factor of safety are tools you will reuse in every hardware chapter of Part IV. Do the ⭐⭐/⭐⭐⭐ exercises and both case studies.

🎮 KSP Player: You have watched a stack wobble, flex, and snap at max-Q or under too much thrust. §23.1 (loads) and §23.4 (stress and the factor of safety) are the physics behind the part stats and the "structural failure" you have earned the hard way. The mass-budget section is the game's whole optimization loop, made explicit.

🛰️ Industry Prep: This chapter's Mission Design Checkpoint hands you the tool the rest of a spacecraft program orbits around — the mass budget. Sections 23.4 (margins) and 23.5 (mass management) are the daily language of a structures or systems engineer. Read 23.6 as a case in how materials choices are really program choices.

23.1 Launch loads: the eight minutes that size everything

Start with the physical reality. A spacecraft on the pad is a delicate assembly of electronics, optics, and thin-walled tanks, sitting on top of a rocket whose entire purpose is to accelerate hard. The moment the engines light, that assembly is subjected to a battery of mechanical loads it will never see again in space. Understanding those loads — where they come from, how big they are, and which part of the vehicle each one threatens — is the starting point for every structural decision, so we catalog them first.

Definition (launch load). A launch load is any of the mechanical loads a launch vehicle and its payload experience during powered ascent (and the events bracketing it): the steady acceleration of the climb, the vibration shaken through the structure, the acoustic pressure of the exhaust and airflow, and the sharp shocks of stage and payload separation. For the great majority of spacecraft, the largest structural loads of the entire mission occur during these few minutes — which is why launch, not orbit, sizes the structure.

There are four kinds, and they threaten different things.

1. Quasi-static (steady) acceleration. This is the load you would guess: the rocket accelerates, and by Newton's second law (Chapter 2) every part of the payload feels a force $F = m\,a$ opposing that acceleration, as if gravity had been turned up. Engineers quote it as a load factor in multiples of $g_0 = 9.81\ \text{m/s}^2$: a "4 g" axial load means each kilogram of spacecraft presses back with four times its Earth weight. Peak axial acceleration for an expendable launcher is typically 3 to 6 g, and — this is the subtle part — the peak comes near the end of a stage's burn, not the start. At lift-off the vehicle is heavy with propellant and its thrust-to-weight is barely above 1; as it burns off hundreds of tonnes of propellant its mass plummets while thrust holds roughly constant, so $a = F/m$ climbs. Many vehicles must throttle down near burnout purely to keep the acceleration (and the force on the payload) within limits — the same throttle-for-loads logic you met at max-Q in Chapter 5, now driven by g instead of dynamic pressure. There is also a smaller lateral load factor (typically 1–2 g) from wind, gusts, engine gimbaling, and the $q\alpha$ bending of Chapter 5.

2. Vibration. The structure does not ride smoothly. Engine combustion, turbopumps, and the buffeting airflow of ascent inject a broadband random vibration that travels through the structure into every bolted-on component. It is described statistically — as a power spectral density in $\text{g}^2/\text{Hz}$ across roughly 20–2000 Hz, summarized by an overall level in $\text{g}_{\text{rms}}$ (root-mean-square g's), commonly 5–15 $\text{g}_{\text{rms}}$ at the component level. There is also lower-frequency sine vibration from the vehicle's own structural modes, including the dreaded longitudinal POGO oscillation. Vibration is what fatigues solder joints, loosens fasteners, and cracks brackets.

3. Acoustic loading. At lift-off, the exhaust of a large rocket is one of the most powerful sustained sound sources humans create. Sound reflected off the pad and generated by the turbulent exhaust plume bathes the vehicle in a pressure field that, inside the fairing, commonly reaches 140–145 dB — and approaches 180 dB near the engines. This matters enormously for large, light, floppy things — solar arrays, antenna dishes, insulation blankets — whose big surface area catches the sound-pressure fluctuations that structural vibration alone would not excite. Acoustic loading is worst at lift-off and again transonically.

Worked Example: turning "140 dB" into a real pressure. Sound-pressure level in decibels is $\text{SPL} = 20\log_{10}(p_{\text{rms}}/p_{\text{ref}})$, with reference $p_{\text{ref}} = 20\ \mu\text{Pa} = 2\times10^{-5}\ \text{Pa}$ (the threshold of human hearing). Invert it for the actual fluctuating pressure at 140 dB: $$p_{\text{rms}} = p_{\text{ref}}\times10^{\text{SPL}/20} = 2\times10^{-5}\times10^{140/20} = 2\times10^{-5}\times10^{7} = 200\ \text{Pa}.$$ Two hundred pascals is only about $0.2\%$ of sea-level atmospheric pressure — but it is fluctuating, tens to hundreds of times a second, across every square metre of the payload. On a $3\ \text{m}^2$ solar panel that is a shaking force on the order of $200\ \text{Pa}\times3\ \text{m}^2 = 600\ \text{N}$, reversing direction hundreds of times a second. Sanity check: 194 dB corresponds to a fluctuation of one whole atmosphere (the loudest a sinusoid can be in air before it would pull an absolute-zero vacuum on the trough), so our 140 dB sits far below that ceiling — plausible for inside a fairing. The reason acoustic loads menace light structures specifically: force is pressure times area, and a big thin panel has area to spare and little mass to resist.

4. Shock (pyroshock). Stage separation, fairing jettison, and spacecraft release are usually done with pyrotechnic or pneumatic devices that fire in milliseconds. The result is a shock — a very-high- frequency, very-high-peak (hundreds to thousands of g), but extremely brief transient. Its short duration means it barely moves heavy primary structure, but it is murder on brittle components: relays, crystals, ceramics, and anything with a resonance it can ring.

💡 Intuition: launch is a test you take once, at full difficulty. In space, loads are gentle and slow. During launch, they are large, fast, and all at once — steady g, random shaking, acoustic hammering, and shock, stacked in the same two-minute window on the lightest structure the rocket equation would let you build. Qualifying a spacecraft is largely a matter of proving, on the ground, that it will survive this one ride. "Test like you fly" (a mantra we return to in Chapter 32) means shaking, blasting, and stressing the flight article to launch levels before you trust it with a mission, because in an unforgiving environment (theme #2) there is no second attempt.

🔧 Engineering Reality: the payload user's guide. Every launch vehicle publishes a payload user's guide that specifies exactly these numbers — the design load factors, the random-vibration spectrum, the acoustic environment, the shock levels — that a payload must be qualified to survive. These are the contractual interface between spacecraft and rocket. When you select a launch vehicle in Chapter 30, matching your spacecraft's qualification levels against that guide is a go/no-go check. The Ascent Loads & Environment note you started in your Mission Design Review back in Chapter 5 is exactly the beginning of this comparison.

🔄 Check Your Understanding 1. Why does the peak axial acceleration usually occur near the end of a stage's burn rather than at lift-off? 2. A component survives the steady 5 g of launch easily but cracks during ground vibration testing. Which launch load does the vibration test represent, and why can it be more damaging than the steady g?

Answers

1. Thrust stays roughly constant while the vehicle sheds propellant mass, so $a = F/m$ rises as $m$ falls; near burnout the vehicle is lightest and the acceleration peaks (often forcing a throttle-down to protect the payload). 2. Random (and sine) vibration. A steady 5 g is a single, constant load the part can be sized for; vibration is an oscillating load that reverses thousands of times, so it drives fatigue — crack initiation and growth at stress concentrations — and it can hit a component's resonant frequency and amplify. A part strong enough for a static load can still fatigue-crack under vibration.

23.2 Primary versus secondary structure

Not all structure is equal. Some of it holds the vehicle together against the main launch loads; lose it and the vehicle is lost. Some of it merely holds a component in place; lose it and you lose one component. Telling these apart is the first thing a structures engineer does, because it decides where the analysis effort, the testing, and the margin go.

Definition (primary structure). The primary structure is the main load-bearing framework that carries the primary loads — the axial thrust and inertial loads of §23.1 — continuously from the engines, through the tanks and body, up to the payload. Its failure means loss of the vehicle. On most launch vehicles the primary structure is the propellant tank: the tank walls, the thrust structure that spreads the engine loads into them, the interstage, and the payload adapter.

Definition (secondary structure). Secondary structure carries only local loads — the weight and vibration of a single component and its own launch loads — and delivers them into the primary structure. Brackets, equipment panels, avionics mounts, and antenna booms are secondary structure. Its failure loses a component or a function, not the vehicle. (Some programs add a tertiary tier for non-structural items like harnesses and blankets.)

The idea that ties these together — and the single most useful concept in the whole chapter for reasoning about a structure — is the load path.

Definition (load path). A load path is the route a force takes through a structure, from where it is applied to where it is finally reacted (ultimately, on the pad or against the thrust of the engines). Every applied load must have a continuous, unbroken path to ground; structure exists precisely to provide that path. Where the path is interrupted — a cutout, a joint, a bracket — stress concentrates and the designer must pay attention.

💡 Intuition: follow the force like water downhill. Pick any component — a camera bolted inside a spacecraft — and ask: when the rocket pulls 5 g, where does the camera's inertial force go? Into its mounting bracket (secondary structure), which feeds it into an equipment panel (secondary), which is bolted to the spacecraft's central cylinder or thrust tube (primary), which passes it through the payload adapter into the launch vehicle's interstage (primary), down the tank walls (primary), through the thrust structure, and finally out through the engines as the reaction to thrust. That unbroken chain is the load path. A structure with no continuous load path for some load is not a light structure — it is a broken one.

Two features of rocket structure follow directly from the load path and from theme #4 (mass is the enemy), and both surprise newcomers.

First, the tank is the airframe. In an airplane, the fuselage is a structure and the fuel sits in separate tanks inside the wings. In a rocket, propellant is 85–95% of the whole vehicle (Chapter 3), so it would be madness to build a strong body and then put tanks inside it — you would be carrying two structures where one would do. Instead the propellant tank wall is the load-bearing skin. This stressed-skin or monocoque construction (from the French for "single shell") makes the tank do double duty: contain the propellant and carry the flight loads. It is the reason a rocket's body diameter is set by its tanks, and the reason structures and propulsion engineers cannot make decisions independently.

🚪 Threshold Concept: structure and propellant are the same object. Once you see that a rocket's primary structure and its propellant tank are one and the same wall, a great deal falls into place. It is why the internal pressure of the tank is a structural parameter — pressurizing a tank stiffens it against buckling, so some vehicles are structurally dependent on being pressurized (the extreme case, the Atlas "balloon tank," would collapse under its own weight if depressurized on the pad, and had to be kept inflated like a football at all times). It is why a dent in a tank is a structural defect, not cosmetic. And it is why the choice of material for the tank — the subject of §23.3 and the Starship case in §23.6 — is simultaneously a propulsion decision (does it tolerate cryogenic propellant?), a thermal decision (does it survive re-entry heat?), and a mass decision (how many kilograms of wall to hold the pressure?). The tank is where every subsystem argument in a rocket comes to be settled.

Second, primary structure gets the scrutiny. Because its failure is catastrophic, primary structure is analyzed exhaustively, tested to destruction on dedicated articles, and — reluctantly — given a factor of safety (§23.4). Secondary structure is designed more quickly and often more conservatively (a heavy bracket is cheap insurance if it is small). The engineering effort follows the consequences of failure, straight down the load path.

🔗 Connection: The distinction maps onto the reliability thinking of Chapter 32. Primary structure is, almost by definition, a single point of failure — there is one load path to the engines and no redundancy in a tank wall — so it is made reliable by analysis, testing, and margin rather than by a backup. You cannot carry a spare rocket body. This is why structural safety factors, though the lowest in engineering, are never zero.

🔄 Check Your Understanding 1. Classify each as primary or secondary structure: (a) a Falcon 9 propellant tank wall, (b) the bracket holding a star tracker, (c) the interstage between two stages, (d) a solar-array hinge. 2. In one sentence, why is a rocket's propellant tank usually also its primary structure, whereas an airliner's fuel tanks are not its fuselage?

Answers

1. (a) primary — it carries the main flight loads and its failure loses the vehicle; (b) secondary — it carries only the star tracker's local loads; (c) primary — it transmits thrust and inertial loads between stages; (d) secondary — it supports one array. 2. Because a rocket is ~90% propellant by mass, so making the tank itself the load-bearing skin (monocoque/stressed-skin) avoids carrying a separate body and separate tanks — one wall does both jobs — whereas an airliner's structure and its (much smaller) fuel load are better kept separate.

23.3 Materials: aluminum, composites, titanium, and steel

Now the question the whole chapter has been circling: what do you build the structure out of? There are four answers that dominate spaceflight, and choosing between them is not about which is "strongest." It is about which carries the load with the fewest kilograms — because, one more time, mass is the enemy. The right figure of merit is therefore not strength but strength per unit mass.

Definition (specific strength). The specific strength of a material is its strength divided by its density, $\sigma/\rho$ — the load-carrying capacity you get per kilogram. Its companion, specific stiffness $E/\rho$ (with $E$ the stiffness, or Young's modulus, of §23.4), is the deflection resistance per kilogram. In a mass-limited vehicle these two ratios — not the raw strength or stiffness — decide which material wins, because you can always add material to get more strength; what you cannot do is add mass for free.

Here is the landscape, at room temperature, in representative round numbers (Tier 2 — real alloys and layups vary; these are for comparison, not design). Strengths are ultimate tensile; densities in $\text{g/cm}^3$ so the ratios read cleanly.

(† Composite values are strongly layup-dependent: a unidirectional carbon-epoxy tape can exceed 1500 MPa and 130 GPa along the fibers but is weak across them, so a real quasi-isotropic laminate lands far lower. The single numbers here are illustrative laminate-level values — treat composite properties as directional and design-specific.)

Read the four materials off the table.

Aluminum alloys are the workhorse — the material most spacecraft and rocket stages are made of. Density around $2.7\ \text{g/cm}^3$, good specific strength, cheap, easy to machine and weld, and decades of design heritage. The 2xxx (aluminum-copper) and aluminum-lithium alloys like 2195 (used on the Space Shuttle external tank and SLS core) shave a few percent of density and add stiffness. When in doubt, a spacecraft is aluminum.

Carbon-fiber composites (carbon-fiber-reinforced polymer, CFRP) are the high-performance choice, with by far the best specific strength and stiffness in the table.

Definition (composite). A composite is a material made of two or more distinct constituents that remain physically separate in the finished part. In aerospace the dominant one is carbon-fiber-reinforced polymer: stiff, strong carbon fibers embedded in a polymer (usually epoxy) matrix. The fibers carry the load along their length; the matrix binds them, spaces them, and transfers load between them. Because the fibers can be aligned with the expected loads, a composite part can be tailored — strong exactly where and in the direction it needs to be — which is the source of both its superb specific strength and its engineering difficulty.

Composites do not corrode, do not fatigue the way metals do, and let designers put material only where the load path needs it. That is why fairings, interstages, many satellite structures, and high-pressure gas bottles (COPVs — composite-overwrapped pressure vessels) are composite. The costs are real, though: composites are expensive, labor-intensive, hard to inspect (damage can hide inside a laminate), and awkward to join (you bond or bolt rather than weld). They can microcrack under repeated cryogenic cycling and can be permeable to small molecules like hydrogen — problems that have defeated more than one composite-cryo-tank program.

Titanium (chiefly Ti-6Al-4V) is the specialist's metal: excellent specific strength, superb corrosion resistance, compatible with many propellants, and — crucially — it keeps its strength when hot, where aluminum has gone soft. It is expensive and hard to machine, so it is used where its properties are worth paying for: pressure vessels, high-load fittings, hot structures, and rotating engine parts.

Stainless steel is the outlier, and the table seems to condemn it: at room temperature its specific strength is the worst of the four (annealed, less than half of aluminum's), because it is nearly three times as dense. For decades that ended the conversation — steel was "too heavy for a rocket." Section 23.6 is the story of why that conventional wisdom was, for a fully reusable vehicle, wrong. Notice already the one column where steel is not penalized: specific stiffness.

🚪 Threshold Concept: for the metals, stiffness-per-kilogram is nearly the same. Look down the last column. Aluminum, titanium, and steel all have a specific stiffness $E/\rho$ of about the same value — roughly $2.5\times10^{7}\ \text{m}^2/\text{s}^2$ in SI ($E/\rho$: aluminum $72/2.7$, titanium $114/4.43$, steel $193/7.9$, all landing near 25). This is not a coincidence of the table; it is a deep fact about metallic bonding, and it has a startling consequence. A great deal of rocket structure fails not by the material yielding but by thin walls buckling — an elastic instability — and the resistance to buckling depends on stiffness and geometry, not on strength. The critical buckling stress of a thin cylinder scales as $\sigma_{\text{cr}} \propto E\,(t/r)$, with $t$ the wall thickness and $r$ the radius. Work through what that means for mass: to make a given cylinder resist buckling you need a certain $E\,t$, and the mass of the wall is proportional to $\rho\,t$, so the mass needed scales as $\rho/E$ — the inverse of specific stiffness. Because $E/\rho$ is nearly identical for aluminum, titanium, and steel, a buckling-limited rocket tank made of steel weighs about the same as one made of aluminum. Steel's density is offset almost exactly by its stiffness: a steel wall can be three times thinner. This single fact quietly demolishes the "steel is too heavy" objection for compression structure, and it is the hidden foundation under the Starship decision. Only composites, with their much higher $E/\rho$, escape the metals' shared ceiling — which is the other half of why they are prized.

🔧 Engineering Reality: pressure stabilizes structure. A thin cylindrical tank is far more prone to buckling (a stiffness problem) than to bursting (a strength problem) under launch's axial compression. Internal pressure fights buckling — it pre-tensions the wall — which is why launch vehicle tanks are pressurized not only to feed the engines but to stand up. This is the load-path point of §23.2 in numbers: the same wall carries hoop tension from pressure and axial compression from the stack, and the designer trades them against each other. It is also why "how strong is the material" is only half the question; "how stiff is it, and how does the geometry buckle" is the other half.

🔄 Check Your Understanding 1. Aluminum's raw strength (~470 MPa) is lower than titanium's (~950 MPa), yet aluminum is the more common spacecraft material. Give two reasons that are consistent with "mass is the enemy" and cost. 2. Using the table, a designer says "switching our buckling-critical tank from aluminum to steel will triple its mass, because steel is three times denser." Why is this wrong?

Answers

1. (i) By specific strength, aluminum (~174) is closer to titanium (~214) than the raw numbers suggest, so the mass penalty for using aluminum is modest; (ii) aluminum is far cheaper, easier to machine and weld, and has vast design heritage — for most parts its lower cost and easy manufacture outweigh titanium's performance edge. (Titanium is reserved for parts where its properties are genuinely needed.) 2. Because a buckling-limited wall's mass scales as $\rho/E$ (inverse specific stiffness), not as $\rho$ alone. Steel is ~3× denser and ~2.7× stiffer than aluminum, so a steel wall can be ~3× thinner and end up at nearly the same mass. $E/\rho$ is almost identical for the two metals, so the tripling never happens.

23.4 Structural analysis: stress, strain, and the factor of safety

To decide whether a structure survives its loads, we need to compare the load it carries against the load it can bear — and both must be expressed in the same currency. That currency is stress. This section makes stress and strain precise, then builds the factor of safety, the number that separates a structure that flies from one that fails.

Definition (stress). Stress is the internal force per unit area within a loaded material, $\sigma = F/A$, measured in pascals ($\text{Pa} = \text{N/m}^2$; structural stresses run to millions of pascals, so we use megapascals, $1\ \text{MPa} = 10^6\ \text{Pa}$). Pull a rod of cross-section $A$ with force $F$ and every internal surface perpendicular to the pull carries a normal stress $\sigma = F/A$ (tension if pulling, compression if pushing). Forces acting along a surface produce shear stress. Stress is what a material actually "feels," independent of the part's overall size.

Definition (strain). Strain is the fractional deformation a stress produces, $\varepsilon = \Delta L / L$ — the change in length divided by the original length. It is dimensionless (often quoted in "microstrain," $10^{-6}$). Strain is how much the material stretches; stress is how hard it is being pulled.

For most structural materials, over the working range, stress and strain are proportional — Hooke's law:

$$ \sigma = E\,\varepsilon, $$

where the constant $E$ is the Young's modulus (the stiffness of §23.3), with units of pascals. A stiff material (large $E$, like steel at 193 GPa) barely strains under stress; a compliant one (small $E$) stretches more. This linear, springy behavior holds up to the yield strength $\sigma_y$, beyond which the material deforms permanently (plastically); push further, to the ultimate strength $\sigma_u$, and it breaks. Structures are designed to stay comfortably below yield in normal operation.

The single most useful stress calculation in all of rocketry is the stress in the wall of a pressurized cylindrical tank — because, from §23.2, the tank is the structure. Internal pressure $p$ pushes outward on a cylinder of radius $r$ and wall thickness $t$, and the wall carries it as tension. There are two components. The hoop (circumferential) stress, which tries to split the cylinder along its length, is

$$ \sigma_{\text{hoop}} = \frac{p\,r}{t}, $$

and the longitudinal (axial) stress, which tries to pull the end caps off, is exactly half of it, $\sigma_{\text{long}} = p\,r/(2t)$. The hoop stress is twice the longitudinal, which is why a bursting pressure vessel splits lengthwise (along its seam), not around its circumference — a fact you can verify on any burst sausage.

Worked Example: sizing a rocket tank wall. Take a Falcon-9-class first-stage tank: diameter $3.66\ \text{m}$ so radius $r = 1.83\ \text{m}$, wall thickness $t = 4.0\ \text{mm} = 0.004\ \text{m}$, held at an internal (ullage) pressure $p = 3.4\ \text{bar} = 3.4\times10^{5}\ \text{Pa}$ (Tier 3 — representative). The hoop stress is $$\sigma_{\text{hoop}} = \frac{p\,r}{t} = \frac{3.4\times10^{5}\times1.83}{0.004} = \frac{6.22\times10^{5}}{0.004} \approx 1.56\times10^{8}\ \text{Pa} = 156\ \text{MPa}.$$ Compare that to the ~350 MPa yield strength of a 2xxx aluminum tank alloy: the wall is stressed to about $156/350 \approx 45\%$ of yield from pressure alone, leaving room for the additional axial compression and bending the wall carries during flight. Sanity check the arithmetic and the units: pascals times metres over metres gives pascals; $156\ \text{MPa}$ is a few percent of a metal's strength, exactly where you would want a working tank to sit — not near bursting, not wastefully thick. Halve the wall to $2\ \text{mm}$ and the stress doubles to $312\ \text{MPa}$, almost at yield with no margin left for flight loads — which is precisely the knife-edge a mass-optimized rocket lives on.

Now the number that decides whether such a wall is acceptable: the factor of safety.

Definition (factor of safety). A factor of safety (FoS, or safety factor) is the ratio by which a structure's demonstrated strength exceeds the maximum load it is designed to carry: $\text{FoS} = \sigma_{\text{allowable}}/\sigma_{\text{limit}}$, where $\sigma_{\text{limit}}$ is the limit load (the maximum expected in service) and $\sigma_{\text{allowable}}$ is the material's yield or ultimate strength. Equivalently, the design must survive an ultimate load equal to the limit load times a required factor. It is the margin against everything you cannot perfectly predict: material scatter, analysis error, manufacturing flaws, and loads slightly worse than expected.

Aerospace uses remarkably low factors of safety — typically 1.25 on yield and 1.4 on ultimate for metallic launch structures (higher for composites and for crewed vehicles, which may require 1.4 on yield and 2.0 on some components). A civil engineer designing a building uses 2 to 5; an elevator cable, 10 or more. The spacecraft engineer accepts a factor of 1.4 — a 40% margin — and no more, and the reason is theme #4 in one sentence: every increment of safety factor is an increment of mass, and mass is bought at the exponential exchange rate of the rocket equation. A building can afford to be five times overbuilt; a rocket that were even twice overbuilt would never reach orbit. So aerospace buys down the uncertainty instead — with better analysis, more testing, and tighter manufacturing — to justify the thinnest margins in engineering.

The bookkeeping tool that expresses "does it pass, and by how much" is the margin of safety.

Definition (margin of safety). The margin of safety is the fractional reserve a part has after applying the required factor of safety: $$\text{MS} = \frac{\sigma_{\text{allowable}}}{\text{FoS}\times\sigma_{\text{limit}}} - 1.$$ A positive MS means the part survives its factored load with margin to spare; $\text{MS} = 0$ means it passes exactly; a negative MS means it fails and must be strengthened (or the load reduced). Structures engineers hunt for parts with large positive margins — those are overweight and can be lightened.

Worked Example: does this strut pass, and is it overweight? A thrust-structure strut is made of aluminum 7075-T6 (yield $\sigma_y \approx 500\ \text{MPa}$, ultimate $\sigma_u \approx 570\ \text{MPa}$; Tier 2). Analysis says its worst-case flight (limit) load produces a stress $\sigma_{\text{limit}} = 360\ \text{MPa}$. Requirements: FoS $= 1.25$ on yield, $1.4$ on ultimate. Check both. $$\text{MS}_{\text{yield}} = \frac{500}{1.25\times360} - 1 = \frac{500}{450} - 1 = +0.11\ (+11\%),$$ $$\text{MS}_{\text{ultimate}} = \frac{570}{1.4\times360} - 1 = \frac{570}{504} - 1 = +0.13\ (+13\%).$$ Both margins are positive, so the strut passes — but only by about 11–13%. That is a good aerospace result: a lean, well-optimized part carrying just over its required load. Contrast a building strut, which might run a margin of +200% or more. If our strut had come out at +80%, the structures engineer would go back and thin it down, because on a rocket a large positive margin is not reassurance — it is wasted mass being dragged to orbit at the exponential price of Chapter 3.

Here is the whole check as code you can keep and point at any part:

def margin_of_safety(sigma_allow, sigma_limit, fos):
    """Margin of safety: fractional reserve after the required factor of safety.
    sigma_allow = material yield or ultimate strength (same units as sigma_limit).
    Positive => passes with margin; negative => fails."""
    return sigma_allow / (fos * sigma_limit) - 1.0

# Aluminum 7075-T6 strut, limit stress 360 MPa (see worked example above):
ms_yield = margin_of_safety(500, 360, 1.25)   # yield strength 500 MPa, FoS 1.25
ms_ult   = margin_of_safety(570, 360, 1.40)   # ultimate 570 MPa, FoS 1.40
print(f"MS(yield) = {ms_yield:+.2f},  MS(ultimate) = {ms_ult:+.2f}")
# Expected output:
# MS(yield) = +0.11,  MS(ultimate) = +0.13

🐛 Find the Error. An intern reports that a bracket is "safe with a factor of safety of 3.5." The lead engineer frowns and sends it back to be redesigned — lighter. Two puzzles: why is a factor of safety of 3.5 a problem on a spacecraft, and what has the intern likely confused about the difference between a factor of safety and a margin of safety?

Answer

A factor of safety of 3.5 means the part is 3.5× stronger than its worst load — on a spacecraft that is overbuilt by a factor of ~2.5 beyond the required ~1.4, i.e. it is dragging far more mass than it needs, which the rocket equation punishes all the way to orbit. The lead wants it thinned until it carries near the required 1.4 (a small positive margin of safety), not 3.5. The likely confusion: a factor of safety is a requirement you must meet (strength ÷ limit load ≥ 1.4); a margin of safety is how much you exceed that requirement (MS = allowable/(FoS×limit) − 1 ≥ 0). "FoS = 3.5" quoted as a virtue treats extra margin as free; on a mass-limited vehicle it is the opposite of free. The goal is FoS met with MS near zero, not FoS maximized.

23.5 Mass budgets and the enemy

We now assemble the master accounting document of spacecraft engineering. Just as a mission is governed by its delta-v budget (Chapter 3) — a ledger of every maneuver's velocity cost — a spacecraft is governed by its mass budget, a ledger of every kilogram. And where the delta-v budget is the currency of the trajectory, the mass budget is the currency of the vehicle. The two are locked together by the rocket equation, which is exactly why mass is the enemy.

Definition (mass budget). A mass budget is a running accounting of every contribution to a spacecraft's mass — structure, propulsion, power, thermal, avionics, payload, propellant, and a reserve called margin — summed and tracked against a hard cap set by the chosen launch vehicle and orbit. It is maintained from the first concept sketch to launch day, and "closing" the mass budget (getting the total under the cap, with margin intact) is one of the central obsessions of a spacecraft program.

A representative dry-mass breakdown for a spacecraft looks like the table below (Tier 2/3 — proportions vary widely by mission; a science probe, a comsat, and a crewed capsule differ enormously). "Dry" means without propellant.

Two disciplines make a mass budget real. The first is margin. Spacecraft mass always grows during design — components come in heavier than estimated, requirements are added, problems are solved by adding material. Historically, spacecraft gain 20–30% in mass from early concept to launch. So a well-run program reserves that growth from the start: a large mass margin at concept (often 25–30%), shrinking as the design matures and estimates firm up, toward a small margin (a few percent) at launch. A budget with no margin at the preliminary design review (Chapter 29) is not lean — it is a budget that has already failed and does not yet know it.

The second discipline is understanding why the enemy is so implacable — and this is where structures loops back to the rocket equation. A kilogram of structure is not merely a kilogram off the payload. It is a kilogram that every stage beneath it must accelerate to full speed, and by the rocket equation that costs propellant, and propellant needs tanks, which are more structure. The demands compound.

Worked Example: what one kilogram of structure really costs. Consider an upper stage that must deliver $\Delta v = 6.0\ \text{km/s}$ with an engine of specific impulse $I_{sp} = 348\ \text{s}$, so its exhaust velocity is $v_e = I_{sp}\,g_0 = 348\times9.807 = 3{,}413\ \text{m/s}$ (Chapter 3). The mass ratio it must achieve is $$\frac{m_0}{m_f} = e^{\Delta v/v_e} = e^{6000/3413} = e^{1.758} = 5.80.$$ Now add $1\ \text{kg}$ of dry structural mass to this stage and ask: to still deliver the same $6.0\ \text{km/s}$, how much more does the whole stage weigh? To keep the same mass ratio $R = 5.80$, if the final (dry) mass rises by $1\ \text{kg}$, the initial mass must rise by $R\times1 = 5.80\ \text{kg}$ — of which $1\ \text{kg}$ is the new structure and the other $R - 1 = 4.80\ \text{kg}$ is extra propellant to push it. So on this stage alone, 1 kg of structure costs 5.8 kg of launch mass (1 kg of itself plus 4.8 kg of propellant). And if this stage rides atop a booster, that 5.8 kg is payload to the booster, multiplied again by the booster's own mass ratio. A kilogram saved high in the stack can be worth ten or more kilograms at lift-off. This is the structural engineer's version of the tyranny of the rocket equation, and it is why a spacecraft team will spend a week of six people's time to save a hundred grams.

💡 Intuition: mass is compound interest, charged against you. Every kilogram you add is borrowed at the rocket equation's exponential interest rate, compounded once per stage. That is why "mass is the enemy" is not hyperbole and not mere thrift: on any other vehicle, adding a few percent mass costs a few percent performance; on a rocket it can cost the mission. The discipline this breeds — tracking every gram, celebrating a saved bracket, refusing a factor of safety above what is required — looks obsessive from outside and is merely rational from inside.

Here is a minimal mass-budget roll-up in code — the seed of the structures.py module you will grow in the Mission Design Checkpoint:

def dry_mass(subsystems):
    """Sum a dict of {subsystem: mass_kg} into a dry mass (kg)."""
    return sum(subsystems.values())

def with_margin(mass_kg, margin_frac):
    """Apply a mass-growth margin (e.g. 0.20 for 20%)."""
    return mass_kg * (1.0 + margin_frac)

sc = {"structure": 120, "power": 90, "payload": 140, "propulsion_dry": 55,
      "adcs_avionics": 35, "thermal": 15, "comms": 20, "harness": 25}
dry = dry_mass(sc)                     # 120+90+140+55+35+15+20+25 = 500
allocated = with_margin(dry, 0.20)     # 500 * 1.20 = 600
print(f"dry mass = {dry:.0f} kg, with 20% margin = {allocated:.0f} kg")
# Expected output:
# dry mass = 500 kg, with 20% margin = 600 kg

🔄 Check Your Understanding 1. A stage needs a mass ratio of $e^{\Delta v/v_e} = 4.0$. If you shave $2\ \text{kg}$ of structure off its dry mass, roughly how much lighter does the fully fueled stage become (same $\Delta v$)? 2. Why does a well-managed program carry a large mass margin early in design and a small one near launch, rather than a constant margin throughout?

Answers

1. To hold the mass ratio at 4.0, a $2\ \text{kg}$ cut in dry (final) mass allows the initial mass to drop by $4.0\times2 = 8\ \text{kg}$ — the $2\ \text{kg}$ of structure plus $(4.0-1)\times2 = 6\ \text{kg}$ of now-unneeded propellant. Saving structure saves propellant too. 2. Because mass estimates are uncertain early (concepts, not hardware) and firm up as parts are designed, built, and weighed. A large early margin covers the historical 20–30% growth from concept to launch; as that growth actually occurs and the numbers become real, the need for reserve shrinks, so the margin is consumed on purpose, reaching a few percent by launch. A constant margin would either be too small early (risking a budget that cannot close) or wastefully large late.

23.6 Case study: why Starship is stainless steel

We can now analyze the decision this chapter has been building toward, and advance our running Starship thread. When SpaceX designed Starship — the fully reusable, Mars-class vehicle that is the largest rocket ever built — the industry's every instinct said carbon-fiber composite: highest specific strength, flight-proven, the obvious choice for a mass-limited vehicle. SpaceX started down exactly that road, built tooling for a composite hull, and then, in 2018–2019, threw it out and switched to stainless steel — the material our own table in §23.3 flagged as having the worst specific strength of the four. It looked like heresy. It was, on closer analysis, a textbook systems-engineering decision: the right choice for the whole vehicle and program, even though it is the wrong choice for the tank in isolation. Walk through the reasoning and you will have used every idea in the chapter.

The objection, quantified. Steel is roughly three times denser than aluminum and far denser than carbon fiber. By the room-temperature specific-strength column, it is a terrible rocket material. If the tank's mass were set by room-temperature strength, steel would lose badly. Two facts rescue it.

Fact 1 — steel gets stronger in the cold, and the tanks are cold. Starship's propellants are liquid oxygen (about $90\ \text{K}$) and liquid methane (about $110\ \text{K}$), so its tank walls live at cryogenic temperature. The 300-series austenitic stainless steels have a remarkable property: they gain strength as they get colder, and stay tough (they have no brittle transition, unlike ordinary steel). A cold-worked 301 stainless that is ~1300 MPa at room temperature can approach ~1600–2000 MPa at liquid-oxygen temperature. Run the specific-strength number at cryogenic temperature and the picture changes:

$$ \left(\frac{\sigma}{\rho}\right)_{\text{steel, cryo}} \approx \frac{1600\ \text{MPa}}{7.9\ \text{g/cm}^3} \approx 203, \qquad \left(\frac{\sigma}{\rho}\right)_{\text{Al, cryo}} \approx \frac{550\ \text{MPa}}{2.7\ \text{g/cm}^3} \approx 204. $$

At the temperature the tank actually operates, steel's specific strength is essentially equal to aluminum's (both metals strengthen in the cold; steel gains more). The room-temperature table lied about the operating point. (These cryogenic strength figures are approximate and alloy/temper-dependent — Tier 2/3, flagged — but the direction and rough magnitude of the effect are well established and are what SpaceX cited.) And recall from §23.3 that for the buckling-limited part of the wall, specific stiffness governs, and steel's is already equal to aluminum's. On both counts — cryo strength and stiffness — the mass penalty that "common sense" assigned to steel largely evaporates.

Fact 2 — steel wins outright on the hot side. A reusable vehicle does not only launch; it re-enters, and re-entry is a thermal problem (Chapter 7). Here the materials diverge violently. Carbon-fiber composite is finished by ~150–200 °C; aluminum softens badly by ~150 °C and melts near 660 °C. Stainless steel keeps useful strength to ~800 °C and melts near 1450 °C. That difference changes the thermal protection system the vehicle needs. A composite or aluminum hull must be shielded everywhere it gets hot, with a heavy, fragile tile or ablator system (Chapter 24 takes up thermal control next). A steel hull can run red-hot over much of its surface and simply take it — so the leeward side may need no heat shield at all, and the windward side needs less. The heat-shield mass saved by a heat-tolerant skin can more than repay the tank mass spent on a denser metal. You do not optimize the tank; you optimize the tank-plus-heat-shield, and steel wins that sum.

Fact 3 — cost and manufacturability, for a vehicle meant to be mass-produced and reused. This is the systems argument, and for Starship it may be the decisive one. Stainless steel costs on the order of a few dollars per kilogram; aerospace carbon fiber costs on the order of $100+ per kilogram, before the labor of layup and autoclave cure and the scrap of trimming — SpaceX quoted a raw-material cost ratio on the order of fifty to one (Tier 2 — widely reported from SpaceX statements). Steel can be welded, quickly, in the open air, by a workforce building tanks like water towers, allowing the rapid iteration SpaceX's development style depends on. Composite hulls demand precise layup, expensive tooling, and long cures, and are hard to inspect and repair. For a vehicle intended to be built by the hundreds and flown again and again (theme #5, reusability is changing everything), the material that is cheap, weldable, tough, and forgiving beats the material that is a few percent lighter but slow and costly to build and fussy to reuse.

🔧 Engineering Reality: this is optimization at the right altitude. The steel decision is a lesson in what you optimize. If your objective is "minimum tank mass," carbon fiber wins and you build a beautiful, expensive, hard-to-reuse rocket. If your objective is "minimum cost per flight of a rapidly reusable vehicle over hundreds of flights" — the actual objective — then the trade includes heat-shield mass, material cost, build speed, robustness, and reusability, and steel wins. Choosing the wrong objective function gives you the wrong material with impeccable logic. Mass is still the enemy (theme #4) — steel's advocates had to prove the cryo-strength and heat-shield-savings arguments to show the mass penalty was small — but for a reusable vehicle, mass shares the throne with cost and turnaround. This is why §23.5 insisted the mass budget is tracked against a cap set by the mission, not minimized in a vacuum.

📜 From History: steel rockets are not new — the balloon returns. Stainless steel flew to orbit long before Starship. The Atlas missile and launch vehicle of the late 1950s used a stainless-steel balloon tank so thin it had to be kept pressurized at all times or it would collapse — the ultimate expression of "the tank is the structure and pressure holds it up" from §23.2. The Centaur upper stage, still flying, uses the same pressure-stabilized stainless balloon tank. The industry then spent decades moving to aluminum and composites in pursuit of minimum mass, and Starship's return to steel is not a step backward but a re-weighting of the same eternal trade — mass versus cost versus robustness — for the new era of reusability. The full Starship systems case study, where this decision sits alongside the Raptor engine (Chapters 17 and 18) and the catch-and-reuse architecture (Chapter 22), is the climax of Chapter 38.

🔄 Check Your Understanding 1. Room-temperature specific strength says steel is a poor rocket material. Give the two facts from this section that overturn that verdict for a cryogenic tank on a reusable vehicle. 2. Why is "minimize tank mass" the wrong objective function for the Starship decision, and what is the right one?

Answers

1. (i) 300-series stainless strengthens at cryogenic temperature, and the tanks operate cold (LOX ~90 K, methane ~110 K), so its operating specific strength roughly equals aluminum's — the room-temperature table is evaluated at the wrong temperature; (ii) steel tolerates re-entry heat (useful to ~800 °C), so it needs far less thermal-protection mass, and the heat-shield mass saved offsets the tank mass spent. (Bonus: for buckling, specific stiffness governs, and steel's already equals aluminum's.) 2. Because the vehicle is meant to be rapidly reusable and mass-produced, so the true objective is minimum cost per flight over many flights, which includes heat-shield mass, raw-material cost (~50× cheaper), weldability/build speed, robustness, and reuse — not the isolated mass of one tank. Optimizing tank mass alone picks carbon fiber with flawless logic and the wrong answer.

Mission Design Checkpoint: your mass budget and structures.py

This is a major checkpoint — you build the accounting document the rest of your mission design orbits around.

The design. Open your Mission Design Review and start a mass budget for your spacecraft (Track A comsat · B lunar lander · C Mars orbiter · D asteroid probe). Make a table with a row for each subsystem — structure, power, payload, propulsion (dry), attitude/avionics, thermal, comms, harness — and a first-pass mass estimate for each, using the representative shares in §23.5 as a starting sanity check (structure ~20–30% of dry mass, and so on). Sum them to a dry mass. Then add two things the rocket equation forces on you: (1) a mass margin — carry ~25–30% now, at this early stage, to cover the growth every spacecraft suffers; and (2) a note that your wet mass (dry + propellant) must come from your delta-v budget (Chapter 3) via the mass ratio, and must fit under your launch vehicle's payload capacity (Chapter 30). Your mass budget and your delta-v budget are the two halves of your vehicle, and from here on you will size everything against them.

The code. Create astrotools/structures.py and add the mass-budget tools of this chapter. Keep these signatures stable — later chapters (power in Chapter 25, the synthesis in Chapter 29) will call them:

"""astrotools.structures -- mass budgeting and simple structural checks (Chapter 23)."""

def dry_mass(subsystems):
    """Total dry mass (kg) from a dict {subsystem_name: mass_kg}."""
    return sum(subsystems.values())

def apply_margin(mass_kg, margin_frac):
    """Mass carried after a growth margin (e.g. 0.25 for 25%)."""
    return mass_kg * (1.0 + margin_frac)

def mass_budget(subsystems, margin_frac=0.25):
    """Return (dry_kg, allocated_kg) for a subsystem mass dict and a margin."""
    dry = dry_mass(subsystems)
    return dry, apply_margin(dry, margin_frac)

def margin_of_safety(sigma_allow, sigma_limit, fos):
    """Structural margin of safety; positive => passes with reserve."""
    return sigma_allow / (fos * sigma_limit) - 1.0

if __name__ == "__main__":
    sc = {"structure": 120, "power": 90, "payload": 140, "propulsion_dry": 55,
          "adcs_avionics": 35, "thermal": 15, "comms": 20, "harness": 25}
    dry, alloc = mass_budget(sc, margin_frac=0.25)
    print(f"dry mass = {dry:.0f} kg, with 25% margin = {alloc:.0f} kg")
    # Expected output:
    # dry mass = 500 kg, with 25% margin = 625 kg

That dry-mass-with-margin number is this chapter's contribution to your MDR, and it feeds directly into the vehicle sizing you will complete in Chapter 29: the mass ratio from your delta-v budget, applied to this dry mass, gives your propellant load and therefore your launch mass. Structure is where the paper spacecraft first acquires real kilograms — and the discipline of counting them, ruthlessly, is what the rest of the design depends on.

Summary

Structures is the art of surviving launch on the fewest possible kilograms. Carry these forward:

Numbers worth remembering: $g_0 = 9.81\ \text{m/s}^2$; aluminum $\rho \approx 2.7$, steel $\approx 7.9\ \text{g/cm}^3$; aerospace ultimate FoS $\approx 1.4$; $\sigma_{\text{hoop}} = pr/t$; and "1 kg of structure costs several kilograms at lift-off."

Spaced Review

Retrieval strengthens memory. Answer from memory before checking, then look back at the cited chapter.

1. (Ch. 3) A stage needs a mass ratio $m_0/m_f = e^{\Delta v/v_e} = 5.0$. Adding 1 kg of dry structure, at fixed $\Delta v$, raises the fully fueled stage mass by how much — and how much of that is extra propellant?
2. (Ch. 3) An engine has $I_{sp} = 330\ \text{s}$. What is its exhaust velocity $v_e$, roughly? (You'll need $g_0$.)
3. (Ch. 5) Max-Q puts a large aerodynamic force on a launch vehicle. Name the launch load in this chapter that this force belongs to, and one other load that also peaks around the same transonic period.
4. (Ch. 5) The $q\alpha$ bending load of Chapter 5 threatens a slender launch vehicle. Which is the vehicle's primary structure that must carry that bending, and why is holding $\alpha \approx 0$ a way of protecting it?

Answers

1. The fueled mass rises by $R\times1 = 5.0\ \text{kg}$ to hold the same mass ratio; of that, $1\ \text{kg}$ is the structure itself and $R-1 = 4.0\ \text{kg}$ is extra propellant to accelerate it. (One kilogram of structure, five kilograms of stage.) 2. $v_e = I_{sp}\,g_0 = 330\times9.81 \approx 3{,}240\ \text{m/s} \approx 3.2\ \text{km/s}$. 3. It is a quasi-static / lateral launch load (the aerodynamic bending contribution to the lateral load factor); acoustic and buffeting-driven vibration also peak transonically, near max-Q. 4. The tank/body — the monocoque primary structure — carries the bending as a distributed load along its length; holding $\alpha \approx 0$ (a zero-lift gravity turn) keeps the aerodynamic side force, and hence the $q\alpha$ bending moment on that thin tube, near zero, so the primary structure need not be built heavier to survive it — mass is the enemy, so you steer instead of stiffen.

What's Next

We have given the spacecraft a body: a structure strong enough to survive launch and light enough that the rocket equation forgives it, built of a material chosen by strength and stiffness per kilogram, sized against loads with the thinnest margins in engineering, and tracked in a mass budget that treats every gram as borrowed at compound interest. But we ended the chapter on a thermal note — steel was chosen partly because it shrugs off re-entry heat — and that is no accident, because the next enemy the structure must face is temperature. In space a vehicle is roasted on the sunlit side and frozen on the shadowed one, with no air to even out the difference; a tank of cryogenic propellant sits a few centimetres from a surface baking in raw sunlight. In Chapter 24 we take up thermal control: the radiative heat balance that sets a spacecraft's temperature, and the insulation, coatings, radiators, and heaters that keep every component alive between the extremes. The structure holds the spacecraft together; thermal control keeps it from cooking or freezing. Both are answers to the same unforgiving environment.