Overview
For Project Osiris, two injectors were designed and manufactured: one for the ablative chamber, and one for the regenerative chamber. Each injector was composed of three parts (baseplate, centerbody, pintle tip) but only the baseplate was different across the two designs. Since the goal of the project is the regenerative chamber (see Ablative Chamber for the reason behind the ablative), this page will focus on the regenerative injector variant (see bottom for Ablative info).
In these cross sections showing the regenerative injector assembly, the three components that make up the injector can be seen (baseplate, centerbody, pintle tip). The pintle tip threads into the centerbody via a 5 (1/2" 20 Thread) ORB fitting. The centerbody interfaces with the baseplate via a bolted taper, which we chose to mitigate pintle concentricity issues. However, we still shim the annulus during assembly to eyeball concentricity. Our annular gap is 0.01".
You can also see three ports; the left port is for probing fuel manifold temperature and pressure, and the right is for the igniter (see the ASI section). The fuel sensing port is a 1/8" NPT, and the ASI port is a 1/2-20 ORB. The top axial port routes N2O to the pintle tip. There is also a fourth port that senses chamber pressure. Our main chamber pressure sensing PT connects to the ASI, but we wanted an additional PT to capture main engine startup transients. We'll probably plug this port for our actual flights on the lander.The main injector for our regenerative TCA is also a pintle.
Fuel comes up from the regen channels and feeds into the injector manifold, and then through the annulus. A small part of the fuel (~15% total fuel mdot) is used as film cooling, and is discharged through the ring of 9 0.5mm holes.
Centerbody
The centerbody inlets N2O at around 600 psi with an ORB fitting (it's an NPT in the photo, but we switched to ORB because ORB superiority). Really though, you should try to minimize the number of NPT fittings going into your injector, as you're probably going to run into a lot of integration issues that will make you uninstall and reinstall injector fittings multiple times.
There is a tee fitting upstream of the centerbody that gives us nitrous manifold pressure. It's probably a somewhat rough estimate as the flow has to constrict a lot more before it exits the pintle orifices, but since our control loop controls nitrous mdot indirectly through pc measurement, it doesn't really need to be that accurate.
Pintle Tip
Purpose: Injection of N2O oxidizer into Osiris combustion chamber
ii. Nitrous Oxide Flow Regimes
Nitrous flow can exist in either the SPI or HEM regime, where SPI stands for Single-Phase Incompressible and assumes quality of zero, meaning the nitrous remains 100% liquid. HEM stands for Homogenous Equilibrium Model and accounts for flash vaporization at the orifices. Flow regime impacts the mass flow rate equation which can be described as
| LaTeX Math Block | ||||
|---|---|---|---|---|
| ||||
\dot{m}=C_d A \sqrt{2 \rho_1 \Delta P} |
and
| LaTeX Math Block | ||||
|---|---|---|---|---|
| ||||
\dot{m}=C_d A \rho_2 \sqrt{2 (h_1-h_2)} |
for SPI and HEM respectively, and is therefore critical to accurate characterization of orifice coefficient of discharge. As manifold to chamber pressure drop increases, the likelihood of flow existing in HEM regime increases, therefore the first design step was determining flow regime at max throttle.
iii. Dyer Model and Supercharge
Initial estimates were based on the Dyer Model, specifically the parameter k, which is a ratio of vapor bubble growth time to fluid residual time. Large k values imply that flow approaches SPI while small k values imply that flow approaches HEM. Assuming a manifold pressure of 600 psi, chamber pressure of 350 psi, and nitrous density of 876.2099466 kg/m^3 at max throttle,
| LaTeX Math Block | ||||
|---|---|---|---|---|
| ||||
K=(P_{initial}-P_{chamber})/(P_{vapor}-P_{chamber}=1.163 |
P_vapor is calculated with CoolProp by solving for nitrous temperature given manifold pressure and density, then using temperature and a quality of 0 to solve for saturation pressure. This result, combined with a low orifice length/diameter ratio implies that flow will remain SPI. The Dyer Model was chosen primarily because it was designed to account for supercharge, as opposed to purely self pressurized. On Osiris, nitrogen is used to supercharge the nitrous.
iv. Cold Flow Test Results
Cold flow testing was performed with the intent of characterizing orifice coefficient of discharge and confirming flow regime, however results were inconclusive. cd was calculated by integrating the SPI mass flow rate equation and assuming a pressure drop from an injector manifold pressure transducer to atmospheric. It should be noted that pressure transducers were reading -4 as atmospheric and results adjusted accordingly.
cd values were somewhat scattered. Lower values towards the end are likely a result of nitrogen running through the components tester. Additionally, between peaks 4 and 5 manifold pressure was not given the time to return to the atmosphere. Regardless, no clear trend can be seen between cd value and throttle state. In the HEM regime cd should decrease as a function of dP and in SPI it should remain relatively constant. In both cases an increase in dP correlates with an increase in cd.
v. CFD
Additional CFD was performed using ANSYS Fluent with the goal of analyzing flash vaporization and the orifices and cavitation within the pintle itself. Inlet pressure was set to 600 psi and outlet pressure to 350 psi, with an operating pressure of also 350 psi. The multiphase mixture model was enabled for liquid and vapor nitrous. Results indicated no cavitation.
vi. Design Optimization
From these results, the flow regime was assumed to be SPI and the coefficient of discharge 0.8. From these assumptions, the ideal orifice area could be calculated and used to determine orifice geometry. An additional orifice was added at the base of the pintle tip to prevent tip melting; if there is no hole there, the nitrous will stagnate and be trapped underneath the radial orifices, causing it to flash boil, which then lets the hot gas melt the tip. The hole was sized based on the difference between ideal and actual orifice area.
vii. Water Hammer
Water hammer concerns were raised by BlueShift engineers at the first hot fire attempt, suggesting threads might fail during hot fire. Hand calcs based on modified Hooke’s Law
| LaTeX Math Block | ||||
|---|---|---|---|---|
| ||||
C=\sqrt{\frac{1}{\rho(1/K+D/(Ee))}} |
and Joukowsky
| LaTeX Math Block | ||||
|---|---|---|---|---|
| ||||
P=\rho CU |
P=CU indicate 995 psi max pressure. SolidWorks FEA indicates 2.2 min factor of safety. https://www.fluidmechanics.co.uk/hydraulic-calculations/water-hammer-surge-analysis/
FOD Issues
Pintle injectors at the amateur scale are oftentimes plagued by FOD due to very small annular gap sizes and tolerances. Even a +/- 0.001 tolerance on a classic pintle annulus can vary mass flow by around 10 percent, which may be unacceptable depending on your requirements. Additionally, a small annular gap means that any contamination in the system can clog it, which would lead to uneven combustion.
A good amount of big boy aerospace industry people will claim that collegiate teams are not capable of having a cleaning standard high enough to prevent FOD clogging of a 10 thou gap annulus (and will thus advise against a pintle or suggest some weird variants of it), but we found that flushing the system with water multiple times (without it hooked up to the injector) has resulted in very clean water flows. At this point, we are more concerned about FOD generators (i.e. chattering check valves) but this hasn't caused a real problem for us yet.
Annular Gap Trades
Although I say that small annular gaps are fine, there is, of course, a limit to that. I've seen people do gaps as small as 0.003" successfully, but we tried our absolute hardest to make our gap as large as possible. To do that, we had to trade annular gap with pintle diameter, fuel manifold pressure, throttling, and stiffness. Our requirement is to stay above 15% dP/Pc for both manifolds, which becomes a challenge as you throttle down because the "dP" decreases faster than the "Pc" as you throttle. However, lower stiffness also means a larger annular gap for the same mdot, as your "A" term needs to increase to compensate for a lower "dP" term. As for pintle diameter, we scaled it down as much as we reasonably could. If we had scaled it down to be even smaller than it currently is, we would have gotten a huge pressure drop into the pintle tip, and the radial holes around the tip probably would not have fit. There is a rule of thumb from somewhere that says your chamber inner diameter should be 3-5x your pintle diameter...I'm pretty sure this is bogus. This trade ultimately had us converge on a manifold pressure of 500 psi at full throttle.
Face Seal
In the assembly, there is one face seal, which interfaces with the combustion chamber. We were a bit concerned that there would be plastic deformation here when torquing the bolts, as we sized the injector such that the face with the face seal on it contacts the chamber before the injector flange contacts the chamber flange. However, we were able to validate that the lip with the face seal on it would not snap or plastically deform with +/- 0.002" tolerances.
Ablative Injector
The injector was fully designed and spec'd for the regenerative chamber and then modified to fit the ablative system. The main modifications were to remove the radial holes for fuel passthrough and face seal flange. Instead, fuel enters through a port on the side of the chamber (think only one radial fuel through hole), and there are a pair of face seals (one for ablative liner one for chamber wall) in place of the face seal + radial seal on the regen. Since the ablative chamber is not designed to run at steady-state, this baseplate was machined out of steel to function as a high-temp heat sink. See CAD below:
Note:
For the actual ablative injector the NPT fitting at the top of the centerbody was replaced with an ORB due to that fitting being removed often. The centerbody was also made out of copper since the steel regen centerbody was dropped, and so the copper part intended for the regen injector was used instead (it is the same part dimensions exactly).
Additionally, the ablative chamber flange was too thin (somehow slipped under the radar through both PDR and CDR), and so a thick steel supporting plate was added during engine integration, sitting "on top" of the injector 0.5mm holes.