A cryogenic propellant depot constantly leaks heat in from sunlight and the surrounding structure. That heat boils a small amount of liquid propellant into vapor, which must be vented (boil-off) — unless an on-site ISRU plant (electrolysing water ice into propellant) is producing new liquid faster than it's being lost.
Q_leak = Q_base · shield_factor / (1 + 0.09·N_MLI) [W]
m_boil = max(0, Q_leak − P_cryo) / L_v [kg/s]
m_net = m_ISRU − m_boil [kg/day]
tank(t+dt) = clamp(tank(t) + m_net·dt, 0, capacity)
- MLI layers — each extra multi-layer-insulation blanket layer adds a radiation barrier; effective heat leak falls with diminishing returns, per the empirical layer-density curves used on real cryogenic stages.
- Sun shield — a deployable parasol that blocks ~85% of direct solar heating when facing the sun; stowing it (for maneuvering or launch) exposes the tank to full solar flux.
- Cryocooler power — active refrigeration intercepts heat before it boils propellant; modeled here at unit efficiency for clarity (real cryocoolers have a coefficient of performance well below 1, so this is an optimistic simplification).
- ISRU production rate — propellant manufactured locally (e.g. electrolysing lunar or asteroid water ice into LOX/LH2, or the Sabatier reaction for methane) instead of launched from Earth. This is the resource-utilization lever: a high enough production rate can outrun boil-off entirely, letting the depot fill even with a leaky tank.
- L_v (latent heat of vaporization) differs by propellant — LH2 boils off fastest per watt of heat leak (lowest L_v), which is why hydrogen depots need the most aggressive insulation and cooling.
Real-world relevance: this is the core trade study behind every proposed cislunar or Mars propellant depot — NASA and commercial architectures all live or die on whether local resource production can keep ahead of unavoidable cryogenic boil-off.