What's better, more low thrust engines, or fewer high thrust engines?

[Halo_003]

I'm curious what the big advantages of multiple engines (clusters) are over singular big engines. Let's say for example, Falcon 9 uses 9 Merlin D engines, Delta IV uses a single RS-68A. SpaceX explains that more give them the engine out capability, plus it lets them land. However, it introduces a lot more points of failure over the singular RS-68A in the Delta IV. What do you guys thing is better in general? Single big engine, or a cluster of smaller ones?

Edited February 11, 2016 by Halo_003

[Bill Phil]

Depends.

If you want engine out, more is the way to go. If you want efficiency, bigger engines are your friend.

Also, sometimes one big engine can be cheaper than a bunch of small ones. 

[DerekL1963]

11 minutes ago, Halo_003 said:

SpaceX explains that more give them the engine out capability, plus it lets them land.

It's also a bit of making lemonade from lemons - the Merlin was originally designed for the much smaller Falcon I.   (A booster that went nowhere because, despite it being an article of faith in the space community for decades, there turned out to not be much of a market down in that payload size.)  Plus, engine-out capability has always struck me as a bit of red herring...  modern engines fail very rarely indeed.

It's also worth noting that smaller engines are somewhat cheaper to develop and build.

 

4 minutes ago, Bill Phil said:

If you want engine out, more is the way to go. If you want efficiency, bigger engines are your friend.

Also, sometimes one big engine can be cheaper than a bunch of small ones. 

Also one big engine simplifies the design of your thrust structure, makes your flight control software moderately simpler, etc... etc...

So, one isn't really "better" than the other.   As Bill said, it depends on your goals.

[Temstar]

Worth pointing out that when you make a large number of a given engine, the price for each engine decreases due to mass production.

Falcon 9 uses 10 Merlin engines per rocket which is a lot, yes the engine on the 2nd stage is modified into an upper stage engine but its still a Merlin at its core.

Having a lot of engines also makes throttling down easier - you just shut some of those engines off instead of having to make your one big engine deep throttlable.

[Nibb31]

Economies of scale.

[swjr-swis]

I do not understand the choice. Cluster of Mammoths all the way.

[Streetwind]

4 hours ago, DerekL1963 said:

Plus, engine-out capability has always struck me as a bit of red herring...  modern engines fail very rarely indeed.

It's still a good thing that SpaceX has that capability. Because one Falcon 9 was indeed saved by it already...

Falcon 9 v1.0 flight 4 (CRS flight 1) suffered a loss of one Merlin 1C about one third of the way into the first stage burn. Instead of terminating the flight, ascent continued with the remaining eight engines, and orbit was eventually achieved after a fully nominal second stage burn. Due to the change in the ascent path, the orbit was lower than planned, and phased slightly differently, but the Dragon spacecraft was capable of making up for this and reached the ISS just fine.

SpaceX also requested permission from the primary customer (NASA) to relight the second stage engine once in order to raise the orbit further, so the secondary payload (an Orbcomm OG2 pilot satellite) could be deployed into its intended orbit. NASA evaluated and ultimately declined the request, citing ISS proximity concerns. The secondary payload was released into the lower orbit, where it decayed rapidly and reentered a few days later.

[RizzoTheRat]

Given that a fuel pump for a Saturn V's F1 engine needed 55,000 BHP*, presumably as you get bigger you get more and more problems.  Yes you could use multiple smaller fuel pumps that need less power, and therefore lighter components, but by the time you've fitted multiple duplicates of all the ancillaries you're not at many less parts than just using multiple engines.

 

* Rocket fuel pumps need crazy power, even the relatively small 30kN rocket in Bloodhound SSC has a 550 BHP supercharged jaguar V8 to drive the pump.

[Hannu2]

It is difficult to justify from here that solutions of space companies are completely wrong. However, it seems to be extremely difficult and expensive to develop and build huge engines to the point that their scale economics begin to benefit. There are few historic examples and they have had large problems. F1 was a propaganda machine far too expensive for all commercial use. Four chambered RD170 had not much users even smaller 2 chamber versions were more popular. It seems that also SpaceX has severe difficulties to plan large Raptor because engine shrinks in every new rumor.

9 engines gives also benefit in return. They use only one of them and it is "too powerful". It can not be throttled so low that stage can hover. Landing with such engine must be extremely accurate and well controlled.

[shynung]

45 minutes ago, Hannu2 said:

9 engines gives also benefit in return. They use only one of them and it is "too powerful". It can not be throttled so low that stage can hover. Landing with such engine must be extremely accurate and well controlled.

In other words, when SpaceX recover their boosters, they do it by suicide-burning.

[Frozen_Heart]

I'm partial to engine clustering rather than a single large engine. Smaller engines are cheaper to design and build and benefit from mass production.

Engine out capability is always nice to have as well.

[kunok]

5 hours ago, Temstar said:

Worth pointing out that when you make a large number of a given engine, the price for each engine decreases due to mass production.

If you get the reuse of the stages, you won't have "mass production" so doesn't makes sense, and I'm not sure how much benefits this scale of economy, for me that is still far from mass production levels, example, you get the benefits for changing the technology making the parts example changing from machining to forging, or using machine tools specially designed for the task, that doesn't really match that level of production. And yes the tooling cost is divided between the number of units but it won't be that big.

IIRC the said they used 3d printing with some of the parts and that's the total opposite of mass production machinery, it's prototyping machinery, yeah it have benefits but it's slow and very expensive, it will benefit more to the big engine than 9 little engines.

The huge production investment of Space X is contradictory with the alleged goal of reusing the stages.

Guys as a manufacturing engineer I see a lot of puzzle pieces but I don't see how they match at all, something feels wrong and contradictory. But, hey, it's their problem and maybe I'm wrong.

I think they doesn't have a bigger engine because the development cost AND times, nothing to do with the manufacturing.

If I said something that doesn't make sense, it will be because I'm not used to talk about manufacturing in English.

[Nibb31]

41 minutes ago, kunok said:

If you get the reuse of the stages, you won't have "mass production" so doesn't makes sense, and I'm not sure how much benefits this scale of economy, for me that is still far from mass production levels, example, you get the benefits for changing the technology making the parts example changing from machining to forging, or using machine tools specially designed for the task, that doesn't really match that level of production. And yes the tooling cost is divided between the number of units but it won't be that big.

SpaceX has scaled its factory to produce 400 engines per year. They are currently at around 100, which counts as mass production in the aerospace industry. By comparison, SNECMA produces less than a dozen Vulcain engines per year.

But you're right, the benefit of reuse goes against the benefit of mass production. That is one of the factors that goes into the balance when people discuss reusable stages. Reusability means that your production volume decreases and the unit cost increases.

41 minutes ago, kunok said:

IIRC the said they used 3d printing with some of the parts and that's the total opposite of mass production machinery, it's prototyping machinery, yeah it have benefits but it's slow and very expensive, it will benefit more to the big engine than 9 little engines.

SpaceX uses 3D printing for prototyping. In many respects, Space is old-school when it comes to actual design work. They tend to rely a lot on tinkering, prototyping, and blowing stuff up. Traditionally, most aerospace companies do their designs and simulations on computer models and only cut metal when the design is 99% complete. It's a very different approach.

41 minutes ago, kunok said:

The huge production investment of Space X is contradictory with the alleged goal of reusing the stages.

Yes, that's correct, and it's a problem they will have to deal with.

41 minutes ago, kunok said:

Guys as a manufacturing engineer I see a lot of puzzle pieces but I don't see how they match at all, something feels wrong and contradictory. But, hey, it's their problem and maybe I'm wrong.

I think they doesn't have a bigger engine because the development cost AND times, nothing to do with the manufacturing.

There's also the fact the engine that you have is cheaper than the engine that you don't have yet. The Merlin engine was based on NASA's Fastrac reference design.

 

[Bill Phil]

7 hours ago, Nibb31 said:

Economies of scale.

... Only work of you actually can lower the cost per engine. In some cases with many engines that couldn't be done. You'd havevto reduce the amount of time to build one, and eventually, you hit human limits, and there aren't many assembly lines for rockets.

Need an example? H-1. 150 built over 19 rockets. Very little cost reduction. And if we want to get technical there were many similar engines used on missiles... And yet very little cost reduction. Economies of scale only work to an extent.

[kunok]

44 minutes ago, Nibb31 said:

SpaceX has scaled its factory to produce 400 engines per year. They are currently at around 100, which counts as mass production in the aerospace industry. By comparison, SNECMA produces less than a dozen Vulcain engines per year.

400 engines? they expected to sell almost 40 falcon 9 every year?

I know that this level is considered mass production, I was pointing more that aerospace level of mass production don't have the reduction cost level of the huge reduction other industries have, it relevant but not that big. I need to improve my speech in English

1 hour ago, Nibb31 said:

SpaceX uses 3D printing for prototyping. In many respects, Space is old-school when it comes to actual design work. They tend to rely a lot on tinkering, prototyping, and blowing stuff up. Traditionally, most aerospace companies do their designs and simulations on computer models and only cut metal when the design is 99% complete. It's a very different approach.

Actually I was talking about this, http://www.spacex.com/news/2014/07/31/spacex-launches-3d-printed-part-space-creates-printed-engine-chamber-crewed they seem to start using 3d printing to make the main oxygen valve for the production models. They said it's for the mechanical properties you can't get casting, but you can get this sort of properties with high pressure injection molding (I'm not sure how it's said in English) with it's a mass production process. And 3d printing doesn't have that good properties.

Off topic, they are 3d printing inconel for test, inconel718 I bet, uhhh, that's a pain in the ass to machine, and it's even worse to 3d print  (And for your info, in the metal industries you talk about additive manufacturing not 3d printing)

[AngelLestat]

9 hours ago, Temstar said:

Worth pointing out that when you make a large number of a given engine, the price for each engine decreases due to mass production.

Falcon 9 uses 10 Merlin engines per rocket which is a lot, yes the engine on the 2nd stage is modified into an upper stage engine but its still a Merlin at its core.

Having a lot of engines also makes throttling down easier - you just shut some of those engines off instead of having to make your one big engine deep throttlable.

This up here..  plus redundancy..  you can have some engines fails that it will reach orbit anyway.

[Nibb31]

1 hour ago, kunok said:

400 engines? they expected to sell almost 40 falcon 9 every year?

That's what the factory is scaled for, although I don't think it's realistic in the foreseeable future, although a Falcon Heavy uses 28 engines in one go.

With 11 F9s and 1 FH per year, they would need ~140 per year, which is still closer to aircraft production volumes than what's usual in the space industry.

 

[tsotha]

11 hours ago, Nibb31 said:

That's what the factory is scaled for, although I don't think it's realistic in the foreseeable future, although a Falcon Heavy uses 28 engines in one go.

With 11 F9s and 1 FH per year, they would need ~140 per year, which is still closer to aircraft production volumes than what's usual in the space industry.

 

Twelve launches is probably too low - they have a two-year backlog and are expecting more orders.  They claim to be on track for a launch "every couple weeks".  Is that real?  Who knows.

[fredinno]

 

On February 11, 2016 at 5:54 AM, Bill Phil said:

... Only work of you actually can lower the cost per engine. In some cases with many engines that couldn't be done. You'd havevto reduce the amount of time to build one, and eventually, you hit human limits, and there aren't many assembly lines for rockets.

Need an example? H-1. 150 built over 19 rockets. Very little cost reduction. And if we want to get technical there were many similar engines used on missiles... And yet very little cost reduction. Economies of scale only work to an extent.

That's since the SIB was man-rated, a very early rocket design, AND used multi-block tanks of now-retired missiles.

It also didn't really launch enough- it was killed very quickly.

14 hours ago, tsotha said:

Twelve launches is probably too low - they have a two-year backlog and are expecting more orders.  They claim to be on track for a launch "every couple weeks".  Is that real?  Who knows.

Realistic- unless you have a rocket launch failure. Then that goes compleely out of wack. And very high production rates and low prices can cause this due to lower safety standards.

[Bill Phil]

24 minutes ago, fredinno said:

 

That's since the SIB was man-rated, a very early rocket design, AND used multi-block tanks of now-retired missiles.

It also didn't really launch enough- it was killed very quickly.

Realistic- unless you have a rocket launch failure. Then that goes compleely out of wack. And very high production rates and low prices can cause this due to lower safety standards.

Even so, economies of scale work to an optimum, and eventually diseconomies of scale takes over, increasing costs.

There were other man-rated rockets that weren't nearly as expensive per kg.

I wouldn't say quickly, design started in the late 50s, had launches into the mid 70s. Pretty good, actually.

[mikegarrison]

It all depends. Let's look at airplanes. Commercial jets used to have three or four engines by preference. This was partly due to the available amount of takeoff thrust (look at the B-52 with eight engines) and partly due to engine unreliability. If you lose an engine on takeoff you need to worry about still being able to climb. If you lose an engine away from an airport, you need to worry about getting to the nearest alternate. And any time you lose an engine you need to worry about controlling the non-symmetric thrust.

Over time, two things happened. Engines became much more reliable, and efficiency became tied to fan size and engine thrust. So the bigger fan engines with higher thrust also were more efficient. Once reliability got to be high enough, it became safer to only have two engines than to have four. The big concern was what would happen if you lost both engines at the same time, but with reliability so good this was nearly impossible (unless you run out of fuel, but then having four engines wouldn't help either). The chances of having any engine go out were cut in half, and the chances of having more than one go out were so small they could be discounted completely. And having two high thrust engines was more efficient than four low thrust engines.

Anyway, what that means for rockets is that really you have to consider what would happen if the most critical engine would fail, what the chances are that multiple engines would fail on the same flight, and what the relative efficiencies of the engines are.

[Beowolf]

On 2/11/2016 at 3:56 AM, Streetwind said:

SpaceX also requested permission from the primary customer (NASA) to relight the second stage engine once in order to raise the orbit further, so the secondary payload (an Orbcomm OG2 pilot satellite) could be deployed into its intended orbit. NASA evaluated and ultimately declined the request, citing ISS proximity concerns. The secondary payload was released into the lower orbit, where it decayed rapidly and reentered a few days later.

What, you mean NASA doesn't finish rendezvous with a 100 m/s main engine burn a few meters from the station? And I thought I was running a realistic space program!

[wizzlebippi]

Aircraft aren't the best example.  Engines have become more reliable mostly because fuel management and scheduling have been taken away from human hands (why some older airliners required a flight engineer).  The odds of a mechanical failure really haven't changed.  This is partly because engines are becoming more complex, and partly because they largely reuse the same core, give it a larger fan, and spin it faster.  The real reason why aircraft like the 747 exist is because before ETOPS (Extended-range Twin engine Operations Performance Standards), the airlines were required to have aircraft with 3+ engines for direct routes.  Basically, the aircraft was never allowed to be more than 1 hour away from an airport with a suitable runway unless it had more than 2 engines.  Now, the FAA allows a twin engine aircraft to certify to whatever time it can sustain single engine flight after making it half way to its intended destination with appropriate reserves.  There may still be a few routes inaccessible to a twin engine aircraft, but there can't be many left.

For a rocket, none of this applies.  Having more, smaller engines increases the odds of an engine failure, and any engine failure reduces TWR and can affect the rocket's ability to make the desired orbit.  From that standpoint, you want the fewest number of engines possible.  As previously mentioned, the only reason Space X's Falcon 9 has 10 Merlin 1D engines is it's cheaper to manufacture 1 engine with the ability to fit a longer engine bell. 

 

[SurrealShock]

.

Edited April 25 by SurrealShock

[Dusty9261]

I think it's worth chipping in that the N1 moon rocket failed 4 times [its only flights too] BECAUSE it used so many engines, which made for complex and fragile plumbing/wiring/etc.

Versus, of course, the Saturn V, and I think we all know how that went.

Not saying many small engines are bad, but I feel the need to point out that one can only have so many before it becomes a problem in and of itself.

Edited February 13, 2016 by Dusty9261