Case Study: Designing a Nuclear-Electric Cargo Tug for the Outer Planets
"In the outer solar system, sunlight is a rumor. A serious cargo ship brings its own power."
Executive Summary
The first case study took a real nuclear engine apart. Here we do the harder thing: we design a vehicle on paper — a nuclear-electric cargo tug meant to haul supplies to the Jupiter system, where sunlight is too weak for solar arrays and patience is cheaper than speed. Given a cargo mass and a delta-v, we will size the whole vehicle: its thrust, its propellant, its reactor's power-system mass, and — the number that surprises everyone — the enormous radiator it needs to shed waste heat. Along the way we will see why a nuclear-electric spacecraft is, in the end, mostly a power plant with a payload bolted on, and why that is exactly the right design for the outer solar system. This is the kind of first-order sizing that opens every real advanced-propulsion study.
Skills applied
- Sizing an electric thruster's thrust from reactor power via $F = 2\eta P/v_e$ (§21.2, Ch. 16).
- Applying the rocket equation to a high-$I_{sp}$ vehicle that does carry reaction mass (Ch. 3).
- Sizing a power system from a specific mass $\alpha$, and a radiator from $Q = \varepsilon\sigma A T^4$ (§21.2, Ch. 24).
- Judging why nuclear-electric beats solar-electric in the outer solar system (Ch. 20, Ch. 25).
- Iterating a design and reading the thrust-versus-efficiency tradeoff in the numbers.
Background
The requirement
- Cargo (payload): $m_{\text{cargo}} = 10\ \text{t}$ of supplies for an outer-planet outpost.
- Destination: the Jupiter system (5.2 AU), where solar flux is ~$50\ \text{W/m}^2$ — about $1/27$ of its value at Earth.
- Delta-v: $\Delta v = 20\ \text{km/s}$. Low-thrust electric vehicles spiral slowly out of gravity wells and take an efficient but delta-v-hungry heliocentric route, so their delta-v budgets are large — but their sky-high $I_{sp}$ makes even 20 km/s affordable in propellant.
The design choices (our knobs)
- Electric power from the reactor: $P_{\text{elec}} = 2\ \text{MW}$.
- Thruster: ion, $I_{sp} = 6{,}000\ \text{s}$, overall efficiency $\eta = 0.55$.
- Power-system specific mass: $\alpha = 20\ \text{kg/kW}$ (an optimistic but stated goal for future reactor–converter–radiator systems).
- A fixed dry allowance: cargo + power system + tanks/thruster/structure, which we will assemble into a burnout mass $m_{\text{dry}}$.
Phase 1: Thrust and power-system mass
From §21.2, the thrust an electric thruster produces from electrical power is $F = 2\eta P/v_e$, with $v_e = I_{sp}\,g_0$. The power system's mass is simply its specific mass times its power. Let us compute both, along with the propellant and burn time, in one pass:
import math
g0 = 9.80665
# Design inputs
P_elec = 2.0e6 # reactor electrical power, W
eff = 0.55 # electrical -> jet efficiency
Isp = 6000.0 # s
alpha = 20.0 # power-system specific mass, kg per kW
m_dry = 60_000.0 # cargo + power system + structure (burnout mass), kg
dv = 20_000.0 # required delta-v, m/s
ve = Isp * g0
F = 2 * eff * P_elec / ve # thrust, N
m_power = alpha * (P_elec / 1000) # power-system mass, kg
R = math.exp(dv / ve) # mass ratio m0/mf
m_prop = m_dry * (R - 1) # propellant, kg
m0 = m_dry + m_prop # wet mass, kg
mdot = F / ve # propellant flow, kg/s
t_burn = m_prop / mdot # powered duration, s
print("thrust: ", round(F, 1), "N")
print("power mass: ", round(m_power/1000, 1), "t")
print("propellant: ", round(m_prop/1000, 1), "t")
print("wet mass: ", round(m0/1000, 1), "t")
print("burn time: ", round(t_burn/86400), "days")
# Expected output:
# thrust: 37.4 N
# power mass: 40.0 t
# propellant: 24.3 t
# wet mass: 84.3 t
# burn time: 442 days
Read the results. The 2-megawatt reactor buys just 37 newtons of thrust — about the force of holding up a 4-kilogram bag — because we spent our power on a very high exhaust velocity ($v_e = 58.8\ \text{km/s}$). And the power system alone masses 40 tonnes, four times the cargo. Already the design's personality is clear: this is a power plant that happens to carry freight.
Phase 2: The propellant, and a sanity check
The tug, unlike a solar sail, does carry reaction mass (xenon), so it obeys the rocket equation. Its mass ratio is $R = e^{\Delta v/v_e} = e^{20{,}000/58{,}840} = e^{0.340} = 1.405$, so it needs propellant equal to $m_{\text{dry}}(R-1) = 60\,\text{t} \times 0.405 \approx 24.3\ \text{t}$.
Sanity check. Does 24.3 t of propellant really deliver 20 km/s? Run the rocket equation forward: $\Delta v = v_e\ln R = 58{,}840 \times \ln(1.405) = 58{,}840 \times 0.340 \approx 20{,}000\ \text{m/s}$. Yes. Notice how little propellant a $6{,}000$-second engine needs for 20 km/s: a chemical stage at $I_{sp}=450$ s would need a mass ratio of $e^{20000/4414} = e^{4.53} \approx 93$ — over 90 tonnes of propellant for every tonne of burnout mass, which is physically impossible. High $I_{sp}$ is what makes a 20 km/s mission thinkable at all. This is the electric-propulsion payoff of Chapter 20, now powered by a reactor instead of the Sun.
Phase 3: Acceleration and how long it thrusts
The thrust is fixed at 37.4 N, but the vehicle lightens as it burns, so its acceleration rises from $a_0 = F/m_0 = 37.4/84{,}300 \approx 0.44\ \text{mm/s}^2$ at ignition to $F/m_{\text{dry}} = 37.4/60{,}000 \approx 0.62\ \text{mm/s}^2$ at burnout. To expel 24.3 t at a flow of $\dot m = F/v_e = 37.4/58{,}840 \approx 6.4\times10^{-4}\ \text{kg/s}$ takes about 442 days — nearly fifteen months of continuous thrusting. That is the trade nuclear-electric propulsion makes: it sips propellant, but it must fire for a year or more, and the trip to Jupiter (including the long coast) is longer still. Haste is not what this vehicle is for.
Phase 4: The radiator — the real spacecraft
Now the number that dominates the design and that beginners always forget. Our thrusters and converter turn only a fraction of the reactor's heat into electricity; the rest must be radiated away as infrared, because in vacuum there is no other way to lose heat (Ch. 24). Suppose the heat-to-electricity conversion is 25% efficient. Then to deliver $P_{\text{elec}} = 2\ \text{MW}$ the reactor must produce $P_{\text{thermal}} = 2/0.25 = 8\ \text{MW}$, and the waste heat to reject is
$$ Q_{\text{waste}} = 8 - 2 = 6\ \text{MW}. $$
A radiator sheds heat as $Q = \varepsilon\sigma A T^4$ (Ch. 24), with $\sigma = 5.67\times10^{-8}\ \text{W/m}^2\text{K}^4$. Take a radiator emissivity $\varepsilon = 0.85$ running at $T = 500\ \text{K}$. The heat rejected per square meter is
$$ \varepsilon\sigma T^4 = 0.85 \times 5.67\times10^{-8} \times (500)^4 = 0.85 \times 5.67\times10^{-8} \times 6.25\times10^{10} \approx 3{,}010\ \text{W/m}^2, $$
so the radiating area needed is
$$ A = \frac{Q_{\text{waste}}}{\varepsilon\sigma T^4} = \frac{6\times10^{6}}{3{,}010} \approx 2{,}000\ \text{m}^2. $$
Two thousand square meters — a panel roughly $30\ \text{m} \times 30\ \text{m}$ radiating from both faces. That radiator is by far the largest structure on the vehicle, dwarfing the reactor that feeds it.
🔧 Engineering Reality: The radiator area scales as $1/T^4$, so running the radiator hotter shrinks it dramatically — halving to $T$ would multiply the area by sixteen. This is why real nuclear-electric designs push radiator temperatures as high as materials allow, and why radiator technology (heat pipes, liquid-droplet radiators) is as central to nuclear-electric propulsion as the thruster itself. The reactor is not the hard part. Getting rid of the heat is.
Phase 5: Why not solar, and how to iterate
Two final judgments close the design.
Why a reactor and not solar arrays? At Jupiter's 5.2 AU the solar flux is only ~$50\ \text{W/m}^2$. To collect 2 MW of electricity at, say, 20% panel efficiency (~$10\ \text{W/m}^2$ delivered) would take $2\times10^{6}/10 = 2\times10^{5}\ \text{m}^2$ of solar array — 0.2 square kilometers, a sail of panels larger than any structure ever flown, and growing as $r^2$ the farther out you go. The reactor's 40-tonne mass suddenly looks like a bargain. This is the niche nuclear-electric propulsion was born for: high power, far from the Sun.
How to iterate. Suppose we raise the thruster to $I_{sp} = 9{,}000\ \text{s}$ to save propellant. Then $v_e = 88{,}300\ \text{m/s}$, and $F = 2(0.55)(2\times10^{6})/88{,}300 \approx 24.9\ \text{N}$ — less thrust from the same power, exactly as $F = 2\eta P/v_e$ demands. The mass ratio falls to $e^{20000/88300} = e^{0.227} = 1.25$, so propellant drops to $60\,(0.25) = 15\ \text{t}$ — a 9-tonne saving. But the lower thrust means the burn now takes about $15{,}000/(24.9/88{,}300) \approx 5.3\times10^{7}\ \text{s} \approx 620\ \text{days}$ — six months longer. Higher $I_{sp}$ bought less propellant at the cost of more time. There is no free lunch; there is only the trade you choose.
💡 Intuition: A nuclear-electric tug is a study in patience. It carries a power plant and a radiator that outmass its cargo several times over, and it thrusts with the gentleness of a resting hand for a year or more. What it gets for that patience is the ability to move serious mass through the dark outer solar system on a modest tank of xenon — a thing no chemical rocket and no solar array could do. It is the wrong vehicle for a hurry and the right one for a freight run to Jupiter.
Discussion Questions
- The power system (40 t) and radiator dominate the vehicle, while the cargo is only 10 t. Explain physically why a nuclear-electric vehicle is "mostly a power plant," referring to $F = 2\eta P/v_e$ and the specific mass $\alpha$.
- Radiator area scales as $1/T^4$. If a materials breakthrough let the radiator run at 700 K instead of 500 K, by roughly what factor would its area shrink? (Compute it.)
- We chose $\Delta v = 20\ \text{km/s}$. Why is a low-thrust electric mission's delta-v budget so much larger than a chemical mission's to the same destination, yet its propellant so much smaller?
- Under what change in requirements would you switch this cargo tug from nuclear-electric to nuclear- thermal propulsion? (Think about who or what is aboard, and how fast it must arrive.)
Your Turn: Extensions
- Option A (design). Resize the tug for a 5-MW reactor at the same $I_{sp}$ and $\alpha$. What happens to the thrust, the power-system mass, and the burn time? Is the vehicle better or worse for a fast cargo run? Explain.
- Option B (computation). Turn Phase 4 into a function
radiator_area(Q_waste, T, eps=0.85)and tabulate the area needed at $T = 400, 500, 600, 700\ \text{K}$. (Do not run it; hand-trace and add# Expected output:.) Comment on the $1/T^4$ scaling you see. - Option C (your mission). If your MDR mission (Track A/B/C/D) were relocated to the outer solar system or scaled up to require megawatt-class power, write two or three sentences on whether nuclear-electric propulsion would enter your trade, and what would dominate your vehicle's mass.
Key Takeaways
- A nuclear-electric vehicle is mostly a power plant. Power scales the thrust ($F=2\eta P/v_e$), and the power system's mass ($\alpha P$) plus its radiator dwarf the payload.
- High $I_{sp}$ makes big-$\Delta v$ missions affordable in propellant — 24 t of xenon for 20 km/s, where a chemical stage would need an impossible mass ratio of ~90 — but demands a year or more of continuous, gentle thrust.
- The radiator is the hidden giant. Rejecting megawatts of waste heat in vacuum takes ~2,000 m² of panel; its area scales as $1/T^4$, making radiator temperature a first-order design driver (Ch. 24).
- Nuclear-electric owns the outer solar system, where solar arrays would need square kilometers. Choose it for power far from the Sun — and choose nuclear-thermal instead when a crew needs speed.