A crewed spacecraft runs two separate coolant loops: an internal water loop that picks up heat from the cabin and avionics, and an external ammonia loop that radiates that heat to space (water is never routed outside — a leak would be catastrophic; ammonia can't touch the cabin — it's toxic). The two loops never mix fluid. They are integrated only through an Interface Heat Exchanger (IFHX), and a three-way mixing valve decides what fraction κ of the internal flow is routed through the IFHX versus bypassing it — this is the actual temperature-control knob.
Heat exchanger performance uses the standard effectiveness–NTU method:
C_water = κ · ṁ_int · c_water C_nh3 = ṁ_ext · c_nh3
C_min = min(C_water, C_nh3) C_r = C_min / C_max
NTU = UA / C_min (counter-flow HX, fixed UA)
ε = [1 - e^(-NTU(1-C_r))] / [1 - C_r·e^(-NTU(1-C_r))] (C_r < 1)
ε = NTU / (1 + NTU) (C_r ≈ 1)
Q_IFHX = ε · C_min · (T_loop - T_ext,in)
The internal loop's bulk temperature then obeys a simple energy balance against its own thermal mass (loop water + accumulator, C_loop):
dT_loop/dt = (Q_load - Q_IFHX) / C_loop
- κ = 0% (full bypass) → no heat exchange at all: the loop heats up without limit — a real integration failure mode.
- κ = 100% (full IFHX flow) → maximum coolant-side capacitance through the exchanger, maximum rejection.
- Raising external flow or lowering radiator return temperature both increase C_min or ΔT, pulling more heat out per second.
- Time is accelerated ~20× so equilibrium is reachable in well under a minute of real time.
This is the same integration principle used on the ISS's Low Temperature Loop / Moderate Temperature Loop coupling to its external ammonia Thermal Control System via interface heat exchangers. This 2D sibling renders the same effectiveness-NTU model as a flat schematic (pannable/zoomable) instead of a 3D scene, plus a live strip-chart of the loop temperature.