A regeneratively-cooled thrust chamber routes the fuel (or fuel/oxidizer) through narrow channels machined into the outer wall before it reaches the injector, so the propellant itself absorbs the heat that would otherwise melt the liner. Heat crosses three resistances in series at every station along the nozzle:
q = (T_gas − T_coolant) / (1/h_g + t_wall/k_wall + 1/h_c)
- Hot-gas film, h_g — scaled with a Bartz-style correlation, h_g ∝ P_c^0.8·(R_t/R(x))^1.8, so it peaks sharply at the throat where the flow area (and radius) is smallest.
- Wall conduction — a thin copper-alloy liner (k ≈ 350 W/m·K), a small resistance next to the two film terms.
- Coolant film, h_c — turbulent internal flow, Dittus–Boelter: Nu = 0.023·Re^0.8·Pr^0.4, h_c = Nu·k/D_h. A narrower channel raises velocity (and h_c) for the same mass flow.
The coolant enters at the nozzle exit and flows upstream toward the injector (countercurrent cooling, as in most real engines), warming station by station as it absorbs the local heat flux:
ṁ·c_p·dT_coolant/dx = q(x)·(2πR(x))
This 2D view marches the identical axial energy balance as the 3D version, but renders it as a flat side-profile schematic (top) colored by local hot-wall temperature, plus a live axial temperature-profile chart (bottom) plotting coolant, cold-wall and hot-wall temperature against position — the two panels are two different ways of reading the same 1D solution rather than a flattened 3D render. Push mass flow too low or channels too wide and the peak hot-wall temperature climbs past what a copper liner survives — that trade-off is the whole design problem.
Simplifications: constant gas recovery temperature along the nozzle, uniform channel geometry (no taper), no boiling/two-phase heat transfer, steady-state only.