SpaceX Raptor engine - General Thread 2

[dnavas]

...Most carbohydrate engines suffer ISP from running fuel rich, since unburnt fuel is heavier molecules which reach slower exhaust speeds.
...
The exception is Hydrolox. Hydrogen is lighter than either oxygen or steam. As such by running a hydrolox engine more fuel rich, you incrrase ISP (at reduced thrust) because you effectively just heat up hydrogen and exhaust it.
...
If the film starts reacting itself , it becomes as hot as the rest and you loose both properties...

A fuller explanation, including both CH4 and H2 can be found here: https://forum.nasaspaceflight.com/index.php?topic=35169.msg1227557#msg1227557  It's a great read, and I recommend it.  The exhaust isn't simple, and Isp is influenced by a number of complicating factors.

The main thing which prevents film cooling from helping Isp is that the fuel-rich film doesn't mix in the main chamber -- that's sort of the point.  Corvus pointed that out, but I thought I'd underline it.  Here's an interesting, short, but math-ful example: https://www.benjaminmunro.com/liquid-oxygen-methane-engine-development  If I'm scanning that doc correctly, the film cooling decreased ISP from ~268 to ~254.

[Lar]

Yes that is really a good post by Proponent. Thanks for resurfacing it.

[gongora]

Unless someone has any information about transpiration cooling having any connection to Raptor, the transpiration cooling talk should move to a Raptor Alternate Design Speculation Thread.

[Semmel]

...Most carbohydrate engines suffer ISP from running fuel rich, since unburnt fuel is heavier molecules which reach slower exhaust speeds.
...
The exception is Hydrolox. Hydrogen is lighter than either oxygen or steam. As such by running a hydrolox engine more fuel rich, you incrrase ISP (at reduced thrust) because you effectively just heat up hydrogen and exhaust it.
...
If the film starts reacting itself , it becomes as hot as the rest and you loose both properties...

The main thing which prevents film cooling from helping Isp is that the fuel-rich film doesn't mix in the main chamber -- that's sort of the point.  Corvus pointed that out, but I thought I'd underline it.  Here's an interesting, short, but math-ful example: https://www.benjaminmunro.com/liquid-oxygen-methane-engine-development  If I'm scanning that doc correctly, the film cooling decreased ISP from ~268 to ~254.

Very interesting read. They assume a mass flow rate for the film cooling of 0.3, I read that as 30%. Which is much more than I expected[1]. But then again, the engine they were designing was injecting liquid methane and liquid oxygen into the engine, which is very much different than the gas-gas design of Raptor. Maybe the rules for Raptor are completely different. Aslo, they dont motivate the 0.3 mass flow rate, so there is no way for me to guestimate how that would translate to Raptor.

[1]: I have no knowledge whatsoever on film cooling, just.. well, a film of propellant for cooling doesnt sound like a substantial portion of the fuel flow to me.

[WormPicker959]

...Most carbohydrate engines suffer ISP from running fuel rich, since unburnt fuel is heavier molecules which reach slower exhaust speeds.
...
The exception is Hydrolox. Hydrogen is lighter than either oxygen or steam. As such by running a hydrolox engine more fuel rich, you incrrase ISP (at reduced thrust) because you effectively just heat up hydrogen and exhaust it.
...
If the film starts reacting itself , it becomes as hot as the rest and you loose both properties...

The main thing which prevents film cooling from helping Isp is that the fuel-rich film doesn't mix in the main chamber -- that's sort of the point.  Corvus pointed that out, but I thought I'd underline it.  Here's an interesting, short, but math-ful example: https://www.benjaminmunro.com/liquid-oxygen-methane-engine-development  If I'm scanning that doc correctly, the film cooling decreased ISP from ~268 to ~254.

Very interesting read. They assume a mass flow rate for the film cooling of 0.3, I read that as 30%. Which is much more than I expected[1]. But then again, the engine they were designing was injecting liquid methane and liquid oxygen into the engine, which is very much different than the gas-gas design of Raptor. Maybe the rules for Raptor are completely different. Aslo, they dont motivate the 0.3 mass flow rate, so there is no way for me to guestimate how that would translate to Raptor.

[1]: I have no knowledge whatsoever on film cooling, just.. well, a film of propellant for cooling doesnt sound like a substantial portion of the fuel flow to me.

I too was curious about this, as I had assumed that the fraction of fuel used was much smaller. This document (PDF) on the morpheus test engine has some interesting tidbits, though:

All tests of this engine, however, were
performed with the film cooling valve at 100% open, resulting in ~32% of the engine methane flow being used for
film cooling.

So ~30% represents the "maximum". But did they test it at lower flow rates?

On test firing #7 a change to the injector orifices, and an
updated film cooling flowrate of ~19% resulted in a 6” long
“trench” erosion in the chamber wall extending into the
subsonic nozzle and through ~80% of the wall thickness in
the deepest location. Test stand time is valuable, though, so
to stay on schedule the film cooling was increased and testing
continued without further erosion. This was the second of
three such erosion incidents – each was weld repaired and the
combustion chamber/nozzle assembly is still flying.

So, going as low as 19% flow rate results in chamber erosion. So this is pretty old, did they test any more configurations? Here, from a later study:

the chamber/nozzle
is cooled with fuel film cooling (FFC), up to 30% of the total fuel flow. The high-FFC design traded engine
specific impulse efficiency for manufacturability and design costs.

It looks like they didn't, but this writeup points out something critical - since this engine is using only film cooling (no regenerative cooling in channels), all of the chamber/nozzle cooling is done with the fuel flow. So, perhaps any film cooling that would be done in the Raptor will be less that the ~30% shown here. How much, though?

[RobLynn]

Very interesting read. They assume a mass flow rate for the film cooling of 0.3, I read that as 30%. Which is much more than I expected.

It is a tiny engine of 5kN thrust, 0.25% of Raptor.  Flow for film cooling will be linked to wall surface area.  Raptor thrust chamber might have 10-15x surface area but 400x the thrust.  Raptor film cooling flow will only be a few % max. 

[spacenut]

Does anyone know when they will test the Raptor again?  Looking forward to final testing. 

[DigitalMan]

Does anyone know when they will test the Raptor again?  Looking forward to final testing.

Since SpaceX set up a foundry for rapid Raptor iterations I expect a lot of change over time similar to Merlin blocks.

[WormPicker959]

Very interesting read. They assume a mass flow rate for the film cooling of 0.3, I read that as 30%. Which is much more than I expected.

It is a tiny engine of 5kN thrust, 0.25% of Raptor.  Flow for film cooling will be linked to wall surface area.  Raptor thrust chamber might have 10-15x surface area but 400x the thrust.  Raptor film cooling flow will only be a few % max.

In addition, the HD series of engines used in the Morpheus project used only film-cooling, whereas Raptor also uses regenerative cooling, which should additionally reduce the need for film-cooling.

[Robotbeat]

It seems Tom Mueller is not longer a SpaceX employee:

https://twitter.com/lrocket/status/1099411086711746560

Not true, I am an advisor now. Elon and the Propulsion department are leading development of the SpaceX engines, particularly Raptor.  I offer my 2 cents to help from time to time

Do we know how long this has been? When I read that tweet I kinda had a sinking sensation.

https://www.linkedin.com/in/thomas-mueller-2094513b/

Hmmm... "Mars... Surface Power"

Glad they're starting to focus on that. It's one of the main challenges of their Mars architecture, along with surface ISRU.

[ZachF]

Something I worked on after doing the comparison a while back, and afterwards I think I can say that Raptor is objectively the best first stage engine ever made.

So, on the last post of the last Raptor thread I did a rough comparison of what Starship/SH would look like if it was powered by other engines to show the systemic superiority of Raptor vs. other designs and how thrust/$ and thrust/weight is not the end all be all, but one factor of many in first stage engine design.

So I set out to see how Raptor performs as a first stage vs. other engines in both re-usable and expendable forms.

The experiment is simple. The theoretical first stage will push a 100 unit mass to a dV +3,457 m/s (because that is roughlt what starship's first stage must perform), and for the reusable comparison it needs enough fuel left over to perform 3,348m/s of landing burns. (Derived from the 2016 ITS's dry mass, fuel mass, and requirement for 7% of fuel to land) I use TWR, fuel density, and mass to estimate dry mass.

I then come up with these figures.

-GTOW is the total mass of the stack in units; the second stage it is pushing has a mass of 100.
-Thrust is the thrust needed to lift this off (it's GTOW x 1.2)
-Stage mass is the total mass of the theoretical first stage in units
-Fuel mass is fuel mass
-Empty mass is empty mass
-Volume is the size of the first stage; it is basically fuel mass/fuel density.

Now for the reusable section I add rough fuel costs, while not important now, they may be more so in the future when they are fully and rapidly reusable.

The results are, in reusable form, Raptor cleans house in every category. It has the lowest relative GTOW, fuel mass, empty mass, required thrust, volume, and lowest fuel cost.

Raptor is so good in many of these metrics, it beats many expendable stages in terms of empty mass, volume, and required thrust with stage re-use!

In expendable form, Raptor is also overall the best first stage. It has the lowest first stage relative empty mass of all designs, and is among the lowest in fuel mass, GTOW, and even volume. Raptor is such a good first stage it nearly matches the Russian ORSC kerolox and hypergolic engines in volumetric terms!

In an interesting turn, ESA prometheus engine is actually a decent expendable option, but a terrible reusable one.

EDIT- Fixed table, and corrected some prometheus numebrs

[starsilk]

Something I worked on after doing the comparison a while back, and afterwards I think I can say that Raptor is objectively the best first stage engine ever made.

there seems to be a mistake with the ISP values you are using - you have Raptor vacuum ISP as 356 (reusable) and 358 (expendable). I assume that wasn't deliberate?

[Lar]

Maybe more film cooling when flown reusable? That would bump the isp down a bit?

[ZChris13]

Maybe more film cooling when flown reusable? That would bump the isp down a bit?

have we ever gotten numbers associated with that? speculation isn't numbers

[ZachF]

Something I worked on after doing the comparison a while back, and afterwards I think I can say that Raptor is objectively the best first stage engine ever made.

there seems to be a mistake with the ISP values you are using - you have Raptor vacuum ISP as 356 (reusable) and 358 (expendable). I assume that wasn't deliberate?

Whoops, yeah, You're right, but it doesn't change the numbers since it's a translational mistake. Simulation just used average ISP, which was 344 for both. It's very basic and doesn't take more complicated factors into it.

[Eventually Rises]

Does anyone know when they will test the Raptor again?  Looking forward to final testing.

Since SpaceX set up a foundry for rapid Raptor iterations I expect a lot of change over time similar to Merlin blocks.

Perhaps I was misled by the pace of the initial series of test-fires, but I had gotten the impression that the engine design was largely finalized and that they were just validating it;  it seemed as if construction of the second engine was imminent. 

What is the sentiment about the delay amongst those of you who have more experience?  Does the lack of a fully assembled engine test-firing actually mean anything at all?

[marsbase]

What is the sentiment about the delay amongst those of you who have more experience?  Does the lack of a fully assembled engine test-firing actually mean anything at all?

Delay?  This is the fastest engine development by anyone ever.  SpaceX is very good at modeling and learned a lot from the Merlin.  So SpaceX was able to produce an advanced engine based on a different fuel and with extraordinary thrust and ISP.   And it didn't blow up on initial firing.  The second engine is under construction with modifications based on the first set of test firings.  This is whirlwind progress.  Just watch!

[FinalFrontier]

The following is the rumor around McGregor. This is hearsay/second hand+ third hand so keep that in mind.
With that disclaimer noted:
We will use the final SN1 test which was the semi destructive test that was the end of the first campaign.
For this test:
Normal thrust: 195-230 Mt/f
Normal ISP: 240-290 secs (film cooling and preburner mix experimentation was going on in every firing)

For destructive 7.9 second test:
*steady state* thrust: 290 MT/f
Instantaneous max *pushed performance* thrust:  300-380 MT/f
Isp: 260-310 secs (exact value unknown)
Chamber pressure at this setting is unknown but is rumored to have been over 310 bar. LOPB pressure is unknown but would have been very high.

85% power rating thrust unknown


So goes the rumored numbers anyway. The team was experimenting with the mix ratios and other things the entire time. They adjusted the preburner performance of both prebuurners several times and there were several firings just on prebuurners only without full ignition. Film cooling was also being adjusted from test to test.

Took all of this with a large amount of salt when I heard it firsthand but the low ISP (right now) does not surprise me. The second engine should perform much better.

Given the BFH and BFS schedule I would expect to see the seoncd engine on the stand not too far down the road. I do not know of any "delays"

[Oli]

The results are, in reusable form, Raptor cleans house in every category. It has the lowest relative GTOW, fuel mass, empty mass, required thrust, volume, and lowest fuel cost.

Raptor is so good in many of these metrics, it beats many expendable stages in terms of empty mass, volume, and required thrust with stage re-use!

This is silly. Of course the highest performing engine will win if you look at those metrics. You would have to factor in engine cost/reusability, which is highly relevant for a booster stage.

[RobLynn]

The following is the rumor around McGregor. This is hearsay/second hand+ third hand so keep that in mind.
With that disclaimer noted:
We will use the final SN1 test which was the semi destructive test that was the end of the first campaign.
For this test:
Normal thrust: 195-230 Mt/f
Normal ISP: 240-290 secs (film cooling and preburner mix experimentation was going on in every firing)

How would it even be possible to push Isp that low at such high chamber pressures unless pushing a really lean mix (ie more oxygen less methane) or vast amounts of film cooling.  Really really bad combustion mixing would possibly also do it but that is exceedingly unlikely given years of development, subscale work etc.  And pushing over 300 tonnes force when from memory they only had ~190tonnes at ~25MPa is also exceedingly unlikely requiring chamber pressures over 40MPa etc.  Maybe there is some truth under there somewhere, but mostly not credible.