MAE 5540 Spacecraft Propulsion, Cornell University

Apollo CSM Propulsion Redesign Trade Study

From helium pressure-fed to turbopump: decoupling chamber pressure from tank pressure

Role: Team of five — I led the feed-system redesign, parametric analysis, and turbopump CAD
Status: Completed

Apollo CSM Propulsion Redesign Trade Study

Headline results

−24.6%

Required propellant mass at 5 MPa chamber pressure

5.5 MPa

Design chamber pressure, up from 0.69 MPa

+6.1%

Specific impulse gain — real, but not the main effect

\varepsilon \approx 62

Nozzle area ratio for ideal expansion

Rocket PropulsionSolidWorksMATLABTrajectory SimulationNozzle Design

The baseline and what it costs

The Apollo Command and Service Module's Service Propulsion System used the Aerojet AJ10 — a bipropellant engine burning $\mathrm{N_2O_4}$ and Aerozine-50, rated at 20,500 lbf, responsible for lunar orbit insertion and departure. It was designed to be extremely reliable, and it achieved that reliability partly through a helium pressure-fed architecture.

Pressure-fed systems, though, push tank pressures upward to achieve acceptable chamber pressure, and tank pressure drives tank mass. The pressure budget is unforgiving: tank pressure has to support chamber pressure plus injector and line losses.

P_\text{tank} \gtrsim P_c + \Delta P_\text{inj} + \Delta P_\text{lines}

\begin{aligned} t &\approx \frac{P_\text{tank}\, r}{2\sigma_\text{allow}} \\[2pt] m_\text{tank} &\propto \frac{P_\text{tank}\, r^3}{\sigma_\text{allow}} \end{aligned}

Tank mass grows with pressure and with size, and the helium system adds further dry mass through high-pressure bottles, regulators, and plumbing. Raising chamber pressure to gain performance therefore costs mass twice over in a pressure-fed architecture.

The turbopump concept

A turbopump feed system decouples chamber pressure from tank pressure. Tanks can be kept at relatively low pressure — chosen for inlet conditioning and cavitation margin rather than for chamber pressure — while the pumps raise propellant pressure to what the injector and chamber require.

\begin{aligned} P_\text{out} &\approx P_c + \Delta P_\text{inj} + \Delta P_\text{lines} \\ \Delta P_\text{pump} &\approx P_\text{out} - P_\text{tank} \\[2pt] H &= \frac{\Delta P_\text{pump}}{\rho g_0} \end{aligned}

Parametric results

For a fixed mission $\Delta v$ , required propellant mass depends on both specific impulse and dry mass. Working through the rocket equation across candidate chamber pressures separates those two contributions.

\begin{aligned} \Delta v &= g_0 I_\text{sp} \ln\!\left(\frac{m_0}{m_f}\right) \\[2pt] m_p &= m_f\left[ \exp\!\left(\frac{\Delta v}{g_0 I_\text{sp}}\right) - 1 \right] \end{aligned}

Design	$P_c$	$I_\text{sp}$ (s)	% $\Delta u_e$	$m_f$ (kg)	$m_p$ (kg)	% $\Delta m_p$
Baseline (pressure-fed)	0.69 MPa	314	0.0%	6,100	9,044	0.0%
Turbopump + low tank pressure	1 MPa	320	1.9%	5,025	7,239	−19.9%
Turbopump + low tank pressure	3 MPa	330	5.1%	5,025	6,912	−23.6%
Turbopump + low tank pressure	5 MPa	333	6.1%	5,025	6,819	−24.6%
Turbopump + low tank pressure	6 MPa	335	6.7%	5,025	6,759	−25.2%

The result worth reading carefully is the gap between the two percentage columns. Raising chamber pressure by nearly an order of magnitude buys only a few percent in exhaust velocity — that is real but modest. The large propellant saving comes from the dry-mass reduction that low tank pressure enables, dropping final mass from 6,100 kg to 5,025 kg. Attributing the whole 25% to specific impulse would be the easy mistake, and it would be wrong.

We designed to 5.5 MPa chamber pressure, which sits at the knee of the curve: nearly all of the available propellant saving, without pushing pump and chamber requirements further for the last fraction of a percent. The matching nozzle for ideal expansion at that chamber pressure sizes to an area ratio of about 62.

CAD and routing

The bipropellant system CAD was built jointly with a teammate: I did the turbopump inclusion and propellant-line routing, and my teammate handled tank resizing. The routing is where the architecture becomes concrete — the pump inlets, the discharge lines to the injector, and the preburner gas path that drives the turbine all have to physically fit within the Service Module envelope.

Figure 1: Internal and external side views of the bipropellant system CAD.

Figure 2: Top view of the bipropellant system CAD.