Designing Merlin M

[Ionmars]

This is a thread to discuss the feasibility of designing a simple engine like Merlin for a 12m heavy lift booster.

Suppose there is a billionaire from an alternative universe (Mr. B) who wants to build a city on  Mars in our universe. He observes the rapid progress of SpaceX toward this goal, but is still not satisfied that the cost of travel will be low enough. He views the Raptor engine as an excellent machine that may eventually achieve Isp=380, an efficiency level that would allow the conquest of space. Still, complexity of each engine and cost to install plumbing for  many engines for each booster are outside his comfort zone. Is there a simpler engine design and a simpler plumbing plan for the booster that would bring down the total cost of building a spaceship booster?

Perhaps the other SpaceX engine, Merlin, could offer a constructive example. It was designed to be simple and easy to build with a high T/W and is already in production. Merlin’s characteristics meet many of  the requirements for a spaceship booster, but it needs to use methane fuel rather than RP-1 for greater efficiency in combustion and for simple (same-fuel for both stages)  ground servicing at launch.

Mr. B has stated that he wants a new methane fueled engine. He would employ it in a new 12m booster to lift the spaceship toward orbit. He wants it to be cheap to produce and simple to plumb inside the booster, so as to lower the over-all cost of the spaceship. Given these requirements, how would engineers on NASA Spaceflight forum specify the design a new Merlin-like engine for Mr. B?

ISTM  the first specification would be the size of the engine. The physical dimensions of each engine could determine many potential cost savings in its production. It would also determine possible plumbing arrangements inside the booster. As suggested in the sketch below, size of the nozzles (presumably largest dimension) could range from 12m (Merlin-G for Gigantic) down to the present size of Merlin D or smaller. (Merlin D4 could be an engine similar to  Merlin D burning CH4.) The smallest engine size would require the largest number for each booster, and would  require the most complicated and massive plumbing system. Some middle compromise between these extremes could be “Merlin M.”

What could be the design specs (size, Isp, type of injector and others) for Merlin M?
Your reasons for them?

[2megs]

Let's step back and look at the entire system. Isp trades for mass fraction, and the high Isp of Raptor is part of why Starship can afford the penalty of recovery hardware in the second stage. (Conversely, the extremely good mass fraction of the F9 upper stage is the reason SpaceX can get away with GG kerolox there.)

So barring some other breakthrough on ultralight recovery hardware, your Merlin M doesn't just need to be a little bit cheaper, it would need to be so much cheaper that Mr. B's second stages can be profitably expended. And launching 6+ expendable tankers per mission is likely a non-starter.

SpaceX isn't just doing the things they do for bragging rights. The trade space for full reusability forces them down certain paths.

[Asteroza]

The closest reference to compare would be Sea Dragon's single monster pressure fed engine.

[hkultala]

Here are practically three major false assumptions that are related

1) That methane-based gg engine would give equally good thrust/weight and thrust/cost ratio than kerosine based,
2) and raptor is currently behind merlin in T/W because it used the more complicated FFSC cycle
3) than methane gives considerably better isp than kerosine.


But in reality, the direct isp advantage of methane over kerosine is very small. Raptor has much better isp mostly because merlin has much pressure and closed cycle. Look at isp of high-pressure kerosene-based closed cycle engine(RD-170/180/190) and see the differences. (*)

And, methane has clearly worse impulse density than kerosine. It needs thicker pipes, bigger pumps (and bigger, heavier fuel tank). This is the main reason why raptor currently lacks Merlin in T/W.

Most of the "complexity disadvantage" of FFSC for weight can be compensated by placing one of the pumps directly on top of the main chamber, which Raptor does.


In a simple gas generator engine, methane is not a very good fuel. Methane only becomes good fuel because it fits very well to reusable FFSC engine:

1) There is less coking than with kerosene.
2) The pump and pipe sizes needed are more similar for both sides of the FFSC which has some advantages
3) Methane is cheaper than kerosene. When everything is reusad, the fuel cost will start to have more effect on the launch price.
4) Something else I don't remember


(*) Note that RD-170/RD-180/RD-190 does not place the pump on the top of the chamber, which is major contributor to the extra weight (bad T/W ratio compared to raptor or Merlin). It needs longer pipes of hot oxygen-rich high-pressure preburner exhaust.

[Ionmars]

Let's step back and look at the entire system. Isp trades for mass fraction, and the high Isp of Raptor is part of why Starship can afford the penalty of recovery hardware in the second stage. (Conversely, the extremely good mass fraction of the F9 upper stage is the reason SpaceX can get away with GG kerolox there.)

Yes, high Isp is definitely required in the overall system. This is why the Merlin approach would only apply to the booster, where heavy lifting is the main goal. The second stage should employ Raptor 2 or any more advanced version because outer space is where Isp is King.

[Ionmars]

Here are practically three major false assumptions that are related

1) That methane-based gg engine would give equally good thrust/weight and thrust/cost ratio than kerosine based,
2) and raptor is currently behind merlin in T/W because it used the more complicated FFSC cycle
3) than methane gives considerably better isp than kerosine.
...
...

I agree that we shouldn't assume that a switch to CH4 would automatically yield high T/W. AIUI T/W is partly related to size of engine, which is one of the reasons for carefully selecting a middle-of-the road engine size. We are looking for simplicity and lower cost, but T/W is also a part of the tradeoff.

Correction: According to Livingjw in Reply#10 below:  "In practice, the thrust to weight of rocket engines tend to be constant over a broad range of design thrusts. There is probably less than a 5% change in thrust to weight in a rocket designed from 100 - 1000 klbs thrust range. Everything else being equal."

[Ionmars]

...
In a simple gas generator engine, methane is not a very good fuel. Methane only becomes good fuel because it fits very well to reusable FFSC engine:

1) There is less coking than with kerosene.
2) The pump and pipe sizes needed are more similar for both sides of the FFSC which has some advantages
3) Methane is cheaper than kerosene. When everything is reusad, the fuel cost will start to have more effect on the launch price.
4) Something else I don't remember

If we can actually achieve the simplicity and low cost of Merlin from an advanced Raptor engine then we should go for it. If not, we may have to consider the lower cost of a Merlin approach for a booster.

Some people may consider this "Merlin" engine to be a retreat from higher tech. Please remember that SpX has already carried out one such retreat from "high tech" by rethinking the hull material and choosing inexpensive stainless steel over advanced composites.

[gin455res]

Will green (or blue) ammonia ever compete on price with methane in the next 10-40 years? 


It would have the advantage of not drawing the ire of the environmentalists.

[DanClemmensen]

Will green (or blue) ammonia ever compete on price with methane in the next 10-40 years? 


It would have the advantage of not drawing the ire of the environmentalists.

This will depend on the price of natural gas, which will in turn depend on how harshly natural gas and oil are regulated and on how rigorously those regulations are enforced. Today, NG and crude oil production and distribution releases a truly enormous amount of methane into the atmosphere, and it is becoming much easier to identify the sources. One Starship full of monitoring satellites would pretty much pinpoint all sources in near real time. Fixing the leaks will be very expensive.  This also affect the price of "blue" ammonia, so you really need to look at green ammonia. On the other hand we have "green" CH₄, which can be used responsibly for specialty applications like methalox rockets. Use of "green" CH₄ to replace NG for general use is not a good idea because the ageing distribution system is too leaky, all the way from the big pipelines down to the neighborhood distribution pipes.

"Green" ammonia is cheap where electricity is cheap, and the price of electricity is much more stable than the price of NG. The ammonia plants can run on even cheaper "excess" electricity and shut off when the grid asks them to via dynamic demand-response signalling. 

[Adriano]

My flying machines professor Gabrielli, designer of Ww2 fighters, NATO G91, and G222, a transport still in use, started his course on flying machines with a chart of all animals. Practically all insects fly, but elephants for sure don’t and whales can’t even walk. He explained that lifting power of wings go with wing area, proportional to the square of wing span, while weight goes with the cube of wing span. We see that doubling the area of a rocket engine nozzle (square of nozzle diameter) maintaining chamber pressure, will double the thrust, but the weight will increase with the cube of the nozzle diameter. Hence a large number of small engines instead of a single big engine is a better design (if we can solve the more complex piping design from tank to engine). The challenges then are the cost of building smaller engines (hard to have costs proportional to the nozzle area) and cost, weight and complexity of piping from tanks to engines (it seems Spacex Superheavy design leaves room for improvement).

[livingjw]

My flying machines professor Gabrielli, designer of Ww2 fighters, NATO G91, and G222, a transport still in use, started his course on flying machines with a chart of all animals. Practically all insects fly, but elephants for sure don’t and whales can’t even walk. He explained that lifting power of wings go with wing area, proportional to the square of wing span, while weight goes with the cube of wing span. We see that doubling the area of a rocket engine nozzle (square of nozzle diameter) maintaining chamber pressure, will double the thrust, but the weight will increase with the cube of the nozzle diameter. Hence a large number of small engines instead of a single big engine is a better design (if we can solve the more complex piping design from tank to engine). The challenges then are the cost of building smaller engines (hard to have costs proportional to the nozzle area) and cost, weight and complexity of piping from tanks to engines (it seems Spacex Superheavy design leaves room for improvement).

- It is not completely correct that the mass of the combustion chamber and nozzle grows proportional to its volume.

- Material to resist pressure is proportional to volume as stated. This applies to the chamber, nozzle and manifolds.

- But, the combustion chamber gets proportionally smaller for larger engines since it is driven by dwell time. This makes the combustor and injector scale with area, not volume.

- Finally, every part of the combustor and nozzle is regenerative cooling passages which also scale with area.

- In practice, the thrust to weight of rocket engines tend to be constant over a broad range of design thrusts. There is probably less than a 5% change in thrust to weight in a rocket designed from 100 - 1000 klbs thrust range. Everything else being equal.

John

[gin455res]

How hard would it be to convert the existing merlin to run on propane?


Then a single stick 7m falcon heavy might benefit from having cryogenic stainless steel strength on both propellant tanks.


Perhaps eventually some water or power boost fluid injection into the turbopump could be added to reduce the coking.

[DanClemmensen]

How hard would it be to convert the existing merlin to run on propane?


Then a single stick 7m falcon heavy might benefit from having cryogenic stainless steel strength on both propellant tanks.


Perhaps eventually some water or power boost fluid injection into the turbopump could be added to reduce the coking.

There would need to be a market for this product. It would have to be superior to a solution based on Raptor engines, which have already been developed, and the Starship manufacturing approach. Even if you decided (why?) to build an F9-class launcher instead of a super-heavy, why pick propane? Operational advantages, if any, will need to outweigh the development costs.  SpaceX appears to think that the F9/FH are mature products that will become obsolete when Starship is operational. I would think that they are in the best position to evaluate the viability of an F9 replacement, but I could be wrong.

As a non-expert, my personal opinion is probably worthless, but I think basically all existing and planned launchers will become uneconomic when Starship is operational. Only non-economic considerations such as national pride and institutional inertia will keep them flying. Any true competitor must be fully, rapidly, and inexpensively reusable. The only real problem with Starship is the TPS. If a competitor can create a cheaper and less labor-intensive TPS, Starship will have real competition.

[gin455res]

How hard would it be to convert the existing merlin to run on propane?


Then a single stick 7m falcon heavy might benefit from having cryogenic stainless steel strength on both propellant tanks.


Perhaps eventually some water or power boost fluid injection into the turbopump could be added to reduce the coking.

There would need to be a market for this product. It would have to be superior to a solution based on Raptor engines, which have already been developed, and the Starship manufacturing approach. Even if you decided (why?) to build an F9-class launcher instead of a super-heavy, why pick propane? Operational advantages, if any, will need to outweigh the development costs.  SpaceX appears to think that the F9/FH are mature products that will become obsolete when Starship is operational. I would think that they are in the best position to evaluate the viability of an F9 replacement, but I could be wrong.

As a non-expert, my personal opinion is probably worthless, but I think basically all existing and planned launchers will become uneconomic when Starship is operational. Only non-economic considerations such as national pride and institutional inertia will keep them flying. Any true competitor must be fully, rapidly, and inexpensively reusable. The only real problem with Starship is the TPS. If a competitor can create a cheaper and less labor-intensive TPS, Starship will have real competition.

Why pick propane? It can be stored at lox temperatures. At lox temperatures stainless steel is very strong. This enables a merlin-like engine to be combined with  Starship style tankage. Propane has similar mixture ratio (by volume) to kerosene so (perhaps - IANARS-either) allows for merlin conversion. Merlin is a far simpler engine than Raptor, so perhaps a derivative could be mass produced cheaper per unit force than a Raptor. However, the coking is an issue for re-usability.  A 7m merlin-derived booster is not an F9 class launcher. It is a falcon heavy class booster with fewer volume constraints, and a much more heat resistant structure.


Musk has recently intimated that the next engine doesn't need to be bleeding edge, rather it needs to be 10 x cheaper to produce than raptor (whether this is Raptor 2, or current Raptor is not clear to me). Merlin is simpler than Raptor it only fails ease of re-usability due to coking.

[TomH]

...propane....can be stored at lox temperatures.

Following the ideal gas law, propane needs 836 kPa of pressure to remain liquid at 20°C (ambient room temperature). Propane would not need to be chilled at all and one advantage could be less solar shading required.

Musk wants propellant cost to be the most expensive component of SH/SS cost. The average price of propane today was $2.34/gal. The cost for Musk to liquify methane coming from his own well is significantly lower than that. Propane is derived primarily from petroleum, whose price is more volatile than cost of natural gas. Over the last 5 years, wholesale propane prices have fluctuated between $4.00 and $0.75 per gallon. Elon may not want to deal with something that unpredictable.

Liquid propane has 10% more energy than liquid methane for equal volume, however propane is higher density. When you measure the energy based on equal mass, ISP is about equal. It is possible to deep cryo propane for greater density; SpaceX obviously already has experience with super-chilling propellants.

For propane, it is harder to remove sulphur, something you want to avoid having in a rocket engine. Propane usually has added butane, something that evaporates from the mix over time. IDK if prototype propene rocket engines have used pure propane or the standard household mix including butane. And then you have to consider Butane's ISP and coking properties.

Propane production on Mars is likely a bit more complex than methane. While both are alkanes, propane is heavier. Coking is almost non-existent with methane; propane can coke to some degree, something you don't really want in an engine you intend to use a thousand times.

https://www.pioneerlng.com/lng-vs-lpg/

[Pete]

But in reality, the direct isp advantage of methane over kerosine is very small. Raptor has much better isp mostly because merlin has much pressure and closed cycle. Look at isp of high-pressure kerosene-based closed cycle engine(RD-170/180/190) and see the differences. (*)

I really, **really** do not understand what you are saying here.
The RD-180 (kerosine)has a vacuum ISP of 338
The  Raptor (Methane) has a vacuum ISP of 380

Yet you say this makes for a " isp advantage of methane over kerosine is very small"

380, compared to 338, is not a "very small" improvement.
It means that to get 10km/s, you need 40% more fuel.   That's about whats needed to get to orbit.
But to carry the 40% more fuel at liftoff, you need more and/or bigger engines. About the same tankage size, due to better kerosine density. But your liftoff mass ends up being 60% more than with the methane-fuelled comparible rocket. Also more expensive, because more engine needed.

That does not make for a "very small" difference.

[Gliderflyer]

But in reality, the direct isp advantage of methane over kerosine is very small. Raptor has much better isp mostly because merlin has much pressure and closed cycle. Look at isp of high-pressure kerosene-based closed cycle engine(RD-170/180/190) and see the differences. (*)

I really, **really** do not understand what you are saying here.
The RD-180 (kerosine)has a vacuum ISP of 338
The  Raptor (Methane) has a vacuum ISP of 380

Yet you say this makes for a " isp advantage of methane over kerosine is very small"

380, compared to 338, is not a "very small" improvement.
It means that to get 10km/s, you need 40% more fuel.   That's about whats needed to get to orbit.
But to carry the 40% more fuel at liftoff, you need more and/or bigger engines. About the same tankage size, due to better kerosine density. But your liftoff mass ends up being 60% more than with the methane-fuelled comparible rocket. Also more expensive, because more engine needed.

That does not make for a "very small" difference.

The RD-180 has a much lower expansion ratio. A better comparison would be C*, not Isp. However, the density Isp of methane and kerosene are fairly similar from I recall, which could matter more for launch vehicles (depending on your assumptions and what you are optimizing for of course).

[DanClemmensen]

Why pick propane? It can be stored at lox temperatures. [...]

When I take my propane tank for my gas grill down the the refill station, they don't fill it at LOX temperatures. propane (a.k.a. LPG or liquefied petroleum gas) is generally stored and transported at room temperature. Are we talking about something else?

[Pete]

But in reality, the direct isp advantage of methane over kerosine is very small. Raptor has much better isp mostly because merlin has much pressure and closed cycle. Look at isp of high-pressure kerosene-based closed cycle engine(RD-170/180/190) and see the differences. (*)

I really, **really** do not understand what you are saying here.
The RD-180 (kerosine)has a vacuum ISP of 338
The  Raptor (Methane) has a vacuum ISP of 380

Yet you say this makes for a " isp advantage of methane over kerosine is very small"

380, compared to 338, is not a "very small" improvement.
It means that to get 10km/s, you need 40% more fuel.   That's about whats needed to get to orbit.
But to carry the 40% more fuel at liftoff, you need more and/or bigger engines. About the same tankage size, due to better kerosine density. But your liftoff mass ends up being 60% more than with the methane-fuelled comparible rocket. Also more expensive, because more engine needed.

That does not make for a "very small" difference.

The RD-180 has a much lower expansion ratio. A better comparison would be C*, not Isp. However, the density Isp of methane and kerosene are fairly similar from I recall, which could matter more for launch vehicles (depending on your assumptions and what you are optimizing for of course).

nono.
hkultala explicity asks for and compares the ISP, so forget about your C*
hkultala  also asks us to look at the RD-180, to see how its ISP (using kerosine) has only "very small" difference compared to Raptor.
I compare the two and I see a large difference, not a "very small difference"

[Gliderflyer]

I couldn't quite figure out what he was saying (Merlin isn't closed cycle), so I figured there was a mistype somewhere and I went more general. Closed cycle kerosene engines can get close to methane performance.