Well, I’ve noticed that my blogging has been getting rarer, snarkier, and of less substance than it used to be. Part of that is due to all the time we’ve been spending testing engines, killing the gremlins that like to inhabit our test trailer occasionally, and recovering from trips to our remote test site. Part of that is due to the work schedule, especially now that we’re getting structural pieces in for our vehicle. If any of you have ever had a kid, or nephew, or brother (or yourself) who got an Erector Set, Legos, Constructs, or anything similar for Christmas, you can imagine what our office looked like tonight. Just imagine that instead of three little kids, its three lovable, not-so-juvenile deliquents….
Part of my lack of bloggyness has also been due to getting sick several times (and having Tiff and little Jonny get sick too). Part of it has been due to the fact that I really, really don’t want to start blogging about politics or foreign policy, or crap like that. And part of it is because for some strange reason, I can’t seem to blog while the “dishwasher” is going….
Anyhow, with that irrelavent personal stuff out of the way, now for some good ol’ fashioned space punditry–because if you can’t actually report new information, at least you can try to look really smart by commenting on what everyone else has been commenting on, and put your own snarky spin on it.
I want to start by briefly mention something interesting that I think I’ve figured out about SpaceX’s pintle woes. I think that most of their engine related problems are interrelated, and that they have to do with the combination of using a high chamber pressure engine design with a pintle-injector and an ablatively cooled chamber wall. First I’ll give a couple of the reported facts, then a little explanation about why I think they fit together:
- Back about a year ago, SpaceX first mentioned that they had missed the performance goals for their Merlin-1 by a few percent, in spite of having very encouraging initial results.
- SpaceX runs with a fuel-centered pintle, and Elon mentioned that most of those performance losses were related to having to tune the pintle to get wall cooling right.
- Back in the fall (as discussed here and elsewhere), SpaceX had a chamber failure where the chamber lost structural integrity causing the nozzle to tear off.
One of the major drawbacks of a pintle injector design stems from the same thing that gives it most of its benefits–its unusual flow fields. Unlike in a normal rocket engine where the propellant mix, burn, and then accelerate in a nearly axial manner moving from the head-end to the nozzle, fluid flows in a pintle take a more complex path on their way out the thruster thingo. If you look at the picture, you can see that when the two propellant sheets mix, they form a cone shaped spray fan. That fan shape creates two donut-shaped vortexes, one between the fan and the head end, and the other between the fan and the throat. These two vortices have a lot to do with the stability and throttleability of the pintle engines.
Unfortunately, when you have this spray fan hitting the wall, by the time it gets there it is hot flamey stuff, which causes a localized area of higher-than-normal heat flux. Most rocket engines have only one area of really high heat-flux, the throat, like the figure on the right shows (I think this was scanned in by someone from Sutton’s Rocket Propulsion Elements).
There are good canonical ways for designing an ablative chamber to deal with the higher heat flux at the throat. Since you have a large reduction in chamber diameter at that point, you can go with a much thicker wall and eat some throat area erosion over time. Or you can put in a graphite or carbon-carbon insert in the actual throat section.
Dealing with the second high heat flux area is more problematic though, particularly for ablatively cooled engines. For a regen engine, especially a chamber-saddle-jacket or milled-wall type, you can just make the flow passages a little narrower in that section to up the amount of heat transfer rate into the cooling jacket. But for an ablative you have a bit of a quandry. The spray fan impingement point is often at a very inconvenient place–on part of the flat sections of the wall, or on the converging section of the nozzle. In both of those places it’s hard to put a graphite insert, but it’s also hard to deal with large erosion over time. It’s much easier to custom tailor the local heat transfer rates in a regen cooled engine than with an ablative engine. With an ablative engine, you really only can vary a few things: matrix material (ablator), reinforcement material, thickness, wrapping pattern, or adding inserts. And none of those are particularly easy to do, particularly away from the nozzle.
So, the canonical solution with pintles has been to have the fuel coming down the inside of the pintle (and shooting out radially), and to adjust the injection velocity to have some of the fuel punch all the way through the annular oxidizer sheet, impinge the wall, and create some fuel cooling. The problem is that any fuel that is used for wall cooling, by definition isn’t participating with the reaction, and causes a reduction in performance. Most of the previous pintle injectors were also used for engines operating at much lower pressures (100psi for the engines used on the LEM for example), so the amount of fuel film cooling needed was relatively small. I think though that SpaceX had to go with a lot higher amount of film cooling due to the much higher chamber pressures. Heat flux is related quite strongly to density which is related to pressure. Doubling the pressure usually doubles the heat flux.
I think the nozzle failure event was also related to this. According to Elon, this failure happened while they were doing some intentionally off-nominal tests (to verify that their engine could function at off-nominal mixture ratios and pressures), and I think what happened is that they were running the engine lean enough that the fuel jets were no longer reaching the wall well enough to keep the impingement zone fuel cooled. That would square pretty well with the failure seen.
So in summary, I think there’s a strong case that SpaceX’s engine problems haven’t so much been because ablative engines are inherently bad, or because pintle injectors are bad/overhyped, but because combining the two in a high pressure engine is a bad mix. Which is why Elon’s comment in a Q&A that Clark Lindsey linked to is so encouraging:
Ablative was chosen because we thought, incorrectly, that it would have a lower development cost. All future engines will be regen and we will be coming out with a version of Merlin 1 that is regen.
I think this will be a real win for them. Using a regen engine will allow them to get back quite a bit of performance on the Merlin-1, and should make the engine both more robust and more reusable.
Anyhow, that’s about all I can comment on this morning (as I have an industrial sized Erector Set that I’m getting payed to go play with, that’s calling my name), but I figured I’d share some of my insights with the rest of you.
Latest posts by Jonathan Goff (see all)
- Administrivia - July 17, 2018
- Research Papers I Wish I Could Con Someone Into Writing Part I: Lunar ISRU in the Age of RLVs - March 9, 2018
- Random Thoughts: A Now Rather Cold Take on BFR - February 5, 2018