Spacecraft economy of scale

[DarkenedOne]

Something I have been wondering recently is how much can the cost of sending people into space be reduced by using larger spacecraft.  Now lets set aside the problems of investment, cost, and etc. for now.

It seems that in practically every transport sector there exist for a number of both physics and economics reasons.  Planes, ships, cars, and etc. all get more cost effective as they get bigger (assuming full use).   


The Soyuz only carried 3 people, and able to deliver them to the ISS at about $30 million.  Now the Dragon and the CST-100 are being designed to carry 7 people.  As I understand the Dragon will be able to cost less than the $30 million price tag of the Soyuz. 

In the satellite launch sector we can see that effect of economy of scale in the fact that the cost going from the Falcon 1e to the F9H is an order of magnitude.

Personally I believe that an order of magnitude reduction in cost could come from increasing the payload (in humans) of spacecraft assuming no new technology. 

[Robotbeat]

Quantity is just as (or perhaps more) important. There is no physics holding reusable launch vehicles (and spacecraft) back, just quantity of flights.

Best of both worlds would be large reusable launch vehicles flying several times a day per vehicle with fleets all over the world.

[DarkenedOne]

Quantity is just as (or perhaps more) important. There is no physics holding reusable launch vehicles (and spacecraft) back, just quantity of flights.

Best of both worlds would be large reusable launch vehicles flying several times a day per vehicle with fleets all over the world.

Physics are even less of an impediment with large rockets.  Larger rockets have higher mass fractions and reduced drag per unit mass.

But you are right that it is only one aspect of many.  I am saying though that this one aspect can result in an order of magnitude cost reduction. 

[Robotbeat]

Quantity is just as (or perhaps more) important. There is no physics holding reusable launch vehicles (and spacecraft) back, just quantity of flights.

Best of both worlds would be large reusable launch vehicles flying several times a day per vehicle with fleets all over the world.

Physics are even less of an impediment with large rockets.  Larger rockets have higher mass fractions and reduced drag per unit mass.

But you are right that it is only one aspect of many.  I am saying though that this one aspect can result in an order of magnitude cost reduction. 

There are a host of other problems with large rockets, especially with large engines. People make a simplifying assumption that bigger is always better, when that is most certainly not always the case. For instance, the taller your rocket (same width, propellant density, pressure), then the wider your throat has to be for the same T/W ratio. At some point, it becomes ridiculous. Saturn V's engines had no choice but to be high pressure because of that. I've heard estimates that cost for building a rocket engine is proportional to the chamber pressure to the fourth power or something like that.

Plus, large rockets generally mean even larger ground infrastructure costs. This is true for both rockets and spacecraft. There is, for instance, no real ground infrastructure in place for processing payloads larger than EELV-class.

Somewhere, I have a very long list of pros and cons of scaling up or scaling down rockets. Another is that if you have a big engine, the bell nozzle ends up ridiculously huge, which has structural issues and can cause mass to increase disproportionately compared to smaller engines.

Something real important to keep in mind:
"Akin's Laws of Spacecraft Design"
http://spacecraft.ssl.umd.edu/akins_laws.html
"8. In nature, the optimum is almost always in the middle somewhere. Distrust assertions that the optimum is at an extreme point."

It also reminds me of one of Richard Feynmann's talks where he related a story about when he worked in a job which required him to pick out and buy a kind of gear. It turns out that the optimum number of gear teeth is somewhere in the middle of the catalog. Too few teeth, and the gear is more likely to break. Too many fine teeth are pretty expensive to produce, so the optimum is somewhere in the middle.

At some point with rockets (scaling either up or down), you get to a scaling relationship which just blows up on you, like the "cost proportional to the fourth power of pressure" relationship I showed earlier. That part of the equation is inconsequential at low pressures, is only moderately important at moderate pressures, but makes things very quickly unfeasible at higher pressures. Same thing if you try to scale down too far, where you run into "minimum gauge" issues.

One consequence of the issue of the chamber pressure and throat diameter limitations of big rockets is that they tend to have low T/W ratios, which means much higher gravity losses. So while, yes, drag losses are much less, that is often dwarfed by the increase in gravity losses. The optimum is somewhere in the middle.

[QuantumG]

I remember hearing Jeff Greason at ISDC dismiss building a vehicle as big as Space Ship Two because there's no proven market, and so much immaturity when it comes to RLV design. SS2 is tiny by your definition. Just like the aircraft industry, slow steady growth is the way to get to mature designs and high density seating.

[Robotbeat]

Quantity is just as (or perhaps more) important. There is no physics holding reusable launch vehicles (and spacecraft) back, just quantity of flights.

Best of both worlds would be large reusable launch vehicles flying several times a day per vehicle with fleets all over the world.

Physics are even less of an impediment with large rockets...

As a physicist that has read several spacecraft systems engineering and rocket science textbooks, I strongly disagree.

[go4mars]

the taller your rocket (same width, propellant density, pressure), then the wider your throat has to be for the same T/W ratio. At some point, it becomes ridiculous. Saturn V's engines had no choice but to be high pressure because of that. I've heard estimates that cost for building a rocket engine is proportional to the chamber pressure to the fourth power or something like that...  That part of the equation is inconsequential at low pressures, is only moderately important at moderate pressures, but makes things very quickly unfeasible at higher pressures. One consequence of the issue of the chamber pressure and throat diameter limitations of big rockets is that they tend to have low T/W ratios, which means much higher gravity losses. So while, yes, drag losses are much less, that is often dwarfed by the increase in gravity losses.

Aside from ground handling/infrastructure, this suggests to me that the ideal 1st stage would be larger diameter and shorter than what is common.  Is that correct? 

[Robotbeat]

the taller your rocket (same width, propellant density, pressure), then the wider your throat has to be for the same T/W ratio. At some point, it becomes ridiculous. Saturn V's engines had no choice but to be high pressure because of that. I've heard estimates that cost for building a rocket engine is proportional to the chamber pressure to the fourth power or something like that...  That part of the equation is inconsequential at low pressures, is only moderately important at moderate pressures, but makes things very quickly unfeasible at higher pressures. One consequence of the issue of the chamber pressure and throat diameter limitations of big rockets is that they tend to have low T/W ratios, which means much higher gravity losses. So while, yes, drag losses are much less, that is often dwarfed by the increase in gravity losses.

Aside from ground handling/infrastructure, this suggests to me that the ideal 1st stage would be larger diameter and shorter than what is common.  Is that correct? 

Not necessarily. What's "ideal" depends on an enormous number of systems engineering factors that need to be carefully weighed and evaluated.

I was just pointing out that at an extreme, if you keep chamber pressure constant, you have to start making your rocket fatter instead of just taller... Things get complicated.

[JohnFornaro]

Somewhere, I have a very long list of pros and cons of scaling up or scaling down rockets.

As is well known, SLS will scale to 130 tons to orbit.  Somewhere on the forum is a thread where they speculate on 200 to 400 tons to orbit.  Would that be at the middle of this theoretical range?  Or would 1000 tons be the middle?  About where is that middle range?

[DarkenedOne]

There are a host of other problems with large rockets, especially with large engines. People make a simplifying assumption that bigger is always better, when that is most certainly not always the case. For instance, the taller your rocket (same width, propellant density, pressure), then the wider your throat has to be for the same T/W ratio. At some point, it becomes ridiculous. Saturn V's engines had no choice but to be high pressure because of that. I've heard estimates that cost for building a rocket engine is proportional to the chamber pressure to the fourth power or something like that.

Never heard of this being a problem.  Your welcome to provide a link.

Plus there is nothing that stops you from using many small engines to propel the rocket. 

Plus, large rockets generally mean even larger ground infrastructure costs. This is true for both rockets and spacecraft. There is, for instance, no real ground infrastructure in place for processing payloads larger than EELV-class.

True in all transportation industries.  However infrastructure generally has high capital, but low recurring costs.

Somewhere, I have a very long list of pros and cons of scaling up or scaling down rockets. Another is that if you have a big engine, the bell nozzle ends up ridiculously huge, which has structural issues and can cause mass to increase disproportionately compared to smaller engines.

"increase disproportionately compared to smaller engines" - generally this is not the case with engines in general. 

Something real important to keep in mind:
"Akin's Laws of Spacecraft Design"
http://spacecraft.ssl.umd.edu/akins_laws.html
"8. In nature, the optimum is almost always in the middle somewhere. Distrust assertions that the optimum is at an extreme point."

It also reminds me of one of Richard Feynmann's talks where he related a story about when he worked in a job which required him to pick out and buy a kind of gear. It turns out that the optimum number of gear teeth is somewhere in the middle of the catalog. Too few teeth, and the gear is more likely to break. Too many fine teeth are pretty expensive to produce, so the optimum is somewhere in the middle.

At some point with rockets (scaling either up or down), you get to a scaling relationship which just blows up on you, like the "cost proportional to the fourth power of pressure" relationship I showed earlier. That part of the equation is inconsequential at low pressures, is only moderately important at moderate pressures, but makes things very quickly unfeasible at higher pressures. Same thing if you try to scale down too far, where you run into "minimum gauge" issues.

One consequence of the issue of the chamber pressure and throat diameter limitations of big rockets is that they tend to have low T/W ratios, which means much higher gravity losses. So while, yes, drag losses are much less, that is often dwarfed by the increase in gravity losses. The optimum is somewhere in the middle.

I highly doubt this is the case, and if it is it would be the first case in transportation I know of where bigger vehicles did not result in lower costs. 

Secondly you are clearly not evaluating the greatest contributor to cost, and that is labor.  An Airbus 380 carries at about 4 times as many people as the Boeing 737, yet both are crewed by 2 people.  In the same way launch teams do not scale linearly with the rocket.  I do not believe that the size of manufacturing teams scale linearly either.

[Robotbeat]

You are thinking about this wrong. To start: Why aren't there ships ten times the largest ship ever? Obviously, it'd be a lot cheaper according to your simplistic thinking, right? So why isn't it done? Are people just stupid?


Another HUGE issue is sound energy, which scales very quickly with rocket take-off thrust. You put the payload at serious risk of damage, not to mention the ground infrastructure and the rocket itself.

And no links are needed about the T/W ratio issue, because it's basic rocket physics. You've read your Sutton, right?

[Robotbeat]

...
I highly doubt this is the case, and if it is it would be the first case in transportation I know of where bigger vehicles did not result in lower costs.

Spruce Goose. 

Secondly you are clearly not evaluating the greatest contributor to cost, and that is labor.  An Airbus 380 carries at about 4 times as many people as the Boeing 737, yet both are crewed by 2 people.  In the same way launch teams do not scale linearly with the rocket.  I do not believe that the size of manufacturing teams scale linearly either.

Rockets are pilot-less.

Larger rockets tend to have a harder time getting a high launch rate. Fundamentally, this is for the same reason that motorcycle engines can operate at tens of thousands of RPMs while giant ship motors operate at hundreds of RPMs. Thus, the motorcycle engines can have a far higher specific power than the ship motors, because the same volume can be used many times more per second to produce power.

Maintaining the infrastructure is expensive, too. Perhaps if large and small vehicles could somehow get the same launch rate, you'd be right, but they can't and won't (partially because of physical limitations... large objects have a lower natural frequency... moving them around must be done at a slower cycle rate). Thus, because small launch vehicles are capable of a higher launch rate, you can get proportionately a lot more use out of that infrastructure than the big launch vehicles. Your capital investment gets proportionately a lot more use.

Heck, even moving the vehicles can take hours just for the crawler (or whatever) to move back and forth.

As I said, the equations of scale meet in the middle somewhere. They always do. Not too small, not too big.

Another issue is boarding (or payload integration) time... It takes a lot longer to board an A380 than it does a bunch of smaller vehicles, since you can do the smaller ones in parallel across multiple terminals.

[DarkenedOne]

You are thinking about this wrong. To start: Why aren't there ships ten times the largest ship ever? Obviously, it'd be a lot cheaper according to your simplistic thinking, right? So why isn't it done? Are people just stupid?

1.  The ship building industry is limited by the size of the shipyards, ports, and canals.  That why they use terms like Panamax class to designate ship size.  Same things with the Airline industry where the A380 is as big as most airports can handle.  Same thing with trucks which are limited in both size and weight by what the roads can handle.

2.  While physics allows bigger vehicles to be able to transport matter at a lower cost per unit, economics states they also cost more to develop and development is risky. 

3.  Lastly vehicles grow to the size they need to most efficiently serve demand.  It would make no sense to design a 30 person space capsule right now when the only destination in space can only handle 7.  That is why all the new human launch vehicles going to LEO are being designed to handle 7 people.

Another HUGE issue is sound energy, which scales very quickly with rocket take-off thrust. You put the payload at serious risk of damage, not to mention the ground infrastructure and the rocket itself.

I believe that problem has already been solved by Apollo and the Shuttle programs.

Look I am just trying to say I wonder what the economics for a space that could take 50 people into space at a time rather than the current 3. 

[Robotbeat]

Without really much more infrastructure, you can launch 7 times more often with the smaller vehicles to get the same passenger numbers.

The ideal spacecraft size, like you said, depends on the market size.

But an important thing to consider is that a reusable launch vehicle offers an enormous improvement in potential safety, since you can afford 10-30 times more test flights before operations. This is partly why you don't see enormous disposable airplanes... Well, that and it's a heck of a lot cheaper at higher flight rates (which is what you're talking about).

I'm convinced that in the 10-30 mT payload range and 400+ mT or more in LEO per year, adding reusability (starting with first stage) would have a FAR greater impact on reducing costs than increasing vehicle size (since you're already clear of the worst effects of minimum-gauge and large drag that affect small vehicles disproportionately). (Rule of thumb is partial reusability has no leg to stand on under 8 flights per year and full reusability needs at least 40-50 flights per year.)

Anything much larger would be more than what our existing spacecraft-related ground infrastructure can handle... You stop being able to fit under bridges, etc. Thus, there's a disproportionate increase in infrastructure cost if you scale up in that direction too much.

[go4mars]

Not necessarily. What's "ideal" depends on an enormous number of systems engineering factors that need to be carefully weighed and evaluated.

I was just pointing out that at an extreme, if you keep chamber pressure constant, you have to start making your rocket fatter instead of just taller... Things get complicated.

So Sea Dragon, UR-700M, and other "tall" rocket designs are unlikely compared to shorter, larger diameter ones (for payloads in the 500+ tonnes to LEO range)?  Is that a fair generalization?

Spruce Goose. 

It never got the follow-on contract.     
Shipping, cruise ships, and tankers are basically getting stopped from growing by constraints of average port size. 

And it seems to me that eventually, reusable "busses" to space would be cheaper per person than reusable "cars or motorcycles". 

[Blackjax]

I tend to subscribe to the launch volume approach to bringing down costs rather than the large vehicle approach.  Anything much larger than a Falcon Heavy is a dubious proposition IMHO for a variety of reasons.  That being said, I have been wondering whether you could improve the case for larger vehicles to some extent by colocating the manufacturing, test, and launch locations all in the same geographic area.  You don't need to worry about all the logistics hassles of shipping giant cores for example, if they always remain on the grounds of the same facility.  The same characteristics that make a site good for doing launches, also make it good for doing testing (nobody will complain about the noise for example).

Perhaps this is a route to making larger vehicles at least a little more practical? 

[WellingtonEast]

An issue that doesnt seem to be discussed here is materials.

As far as I know the construction materials strenght and other properties are constants.

Hence why we dont get insects as big as a house or why people cant get to 100ft tall because the construction materials properties dont allow it.

Hence I tend to agree with Robobeat

[gbaikie]

I tend to subscribe to the launch volume approach to bringing down costs rather than the large vehicle approach.  Anything much larger than a Falcon Heavy is a dubious proposition IMHO for a variety of reasons.  That being said, I have been wondering whether you could improve the case for larger vehicles to some extent by colocating the manufacturing, test, and launch locations all in the same geographic area.  You don't need to worry about all the logistics hassles of shipping giant cores for example, if they always remain on the grounds of the same facility.  The same characteristics that make a site good for doing launches, also make it good for doing testing (nobody will complain about the noise for example).

Perhaps this is a route to making larger vehicles at least a little more practical? 

You could a similar advantage by launching from the Ocean. Shipping gaint cores is not problem if shipping between ports. So if manufacturing at location that one directly transfer a rocket to ocean transport, one transfer to say Island with a port and perhaps have integration of rocket with satellite at island then ship it to a location on ocean [preferable at equator] for the launch. If said island had in addition an airport which large planes could land- satellites and other payloads could flown to the island and with a modest amount personnel could integrate the vehicles.

This allows rockets to made anywhere in world near port facilities, and satellites to made anywhere in the world- allowing one to be near a large trained workforce, and allowing specialize of component parts being made in any location with access to a port [and including rivers and canals which can ship freight].

Launch from the ocean would mean launching a large inhabited areas. Requiring little in terms managing local boat traffic [fishing boats or pleasure craft] and allowing any inclination of orbit.

If does it all on land, one still needs the infrastructure at some distance from the launch pad [say 3 miles or more]. Which means for a large launcher some sort of crawler. With an island and port simple barge can carry and move any conceivable mass of a rocket, and can exceed 1 mph.  If wanted serious one get a speed 20 or 30 mph- though 5 to 10 mph would more normal speed of a barge. And one could launch 3-30 miles away from dock in same time and less expense as could get with a crawler.

[JohnFornaro]

So nobody so far ventures a number as to how many tons that "optimum" might be?  I'm glad to see agreement that there is, in theory, such an optimum.

[malenfant]

...
If does it all on land, one still needs the infrastructure at some distance from the launch pad [say 3 miles or more]. Which means for a large launcher some sort of crawler....

What's your justification for that?  To mitigate the lost assets in the event of an explosion at the pad?

I once posted similar thoughts than Blackjax in the sea dragon thread.
Why not build a launch pad that also double as your assembly jig and never move the thing 'till launch day

It seems no more improbable than sea dragon.

[Robotbeat]

...
If does it all on land, one still needs the infrastructure at some distance from the launch pad [say 3 miles or more]. Which means for a large launcher some sort of crawler....

What's your justification for that?  To mitigate the lost assets in the event of an explosion at the pad?

I once posted similar thoughts than Blackjax in the sea dragon thread.
Why not build a launch pad that also double as your assembly jig and never move the thing 'till launch day

Launch pads tend to be partially destroyed when large launch vehicles launch from them. Often need to be re-painted between every launch, etc. For a VERY large launch vehicle, the sound energy itself can damage the structure. Launch pads for HLVs often have costs in the billions of dollars. And a launch failure can mean you need to build a new one. If that's also your assembling building, then the damage is even worse, and you're back to square one in case of a launch failure (which for HLVs like the Saturn V or N-1 or SLS or Energia can release the same amount of energy as a multi-kiloton nuclear explosion).

[DarkenedOne]

...
If does it all on land, one still needs the infrastructure at some distance from the launch pad [say 3 miles or more]. Which means for a large launcher some sort of crawler....

What's your justification for that?  To mitigate the lost assets in the event of an explosion at the pad?

I once posted similar thoughts than Blackjax in the sea dragon thread.
Why not build a launch pad that also double as your assembly jig and never move the thing 'till launch day

Launch pads tend to be partially destroyed when large launch vehicles launch from them. Often need to be re-painted between every launch, etc. For a VERY large launch vehicle, the sound energy itself can damage the structure. Launch pads for HLVs often have costs in the billions of dollars. And a launch failure can mean you need to build a new one. If that's also your assembling building, then the damage is even worse, and you're back to square one in case of a launch failure (which for HLVs like the Saturn V or N-1 or SLS or Energia can release the same amount of energy as a multi-kiloton nuclear explosion).

Well if we insist on repainting the pad then I suppose.  Also water is used to absorb the sound in large launch vehicles.   

I have yet to see any evidence that the launch pad for the shuttle cost more to maintain per kg of payload than smaller launch pads of that type. 

[Robotbeat]

I have yet to see any evidence that Shuttle's pad costs less per kg. Remember, you can't just factor in one single launch, but also what launch rate the pad can support.

[indaco1]

So nobody so far ventures a number as to how many tons that "optimum" might be?  I'm glad to see agreement that there is, in theory, such an optimum.

This is a very basic equation found in the microeconomics literature:

AC = (SC + RC)/Q + V

where:

AC = Average cost per unit
SC=Startup cost  (developement and infrastructure cost)
RC=Recurring Fixed Costs  (eg. infrastructure to mantain, staff to pay etc.).
V = Variable cost per unit
Q = Number of units produced
 
Suppose SC, RC and V are a function of Q and of the size S of the veichle (it's not so simple I know).

For any possible value of the request Q we have an optimal value of S to minimize AC.

The problem is that the functions SC(Q,S), RC(Q,S) and V(Q,S) are pretty unkonwn  :-(

But looking at the equation it's quite clear why big can be bad, when SC is huge, V not that better and Q not immense.

[Robotbeat]

V and RC are a function of the rate of Q, as well...

[93143]

I have yet to see any evidence that Shuttle's pad costs less per kg. Remember, you can't just factor in one single launch, but also what launch rate the pad can support.

I have yet to see any evidence that the STS pad setup would have been a limiting factor below the originally-targeted dozens of flights per year.  As I understood it, the holdup was the Shuttle itself, with its extensive turnaround requirements and stringent launch criteria.

[indaco1]

V and RC are a function of the rate of Q, as well...

You mean rate as units produced in relation to time?

There's no rate and there's no time in this very simplified model.

The equation means that to produce a certain total quantity of units (cars, pens or kg of payload to LEO) you have a certain total cost, and a certain average cost per unit produced. 

Assume you produce all the units instantaneously.

I modified a little one of the basic equations used for the Cost-Volume-Profit (CVP) model:
http://en.wikipedia.org/wiki/Cost%E2%80%93volume%E2%80%93profit_analysis
They use X instead of Q for quanty of units produced and average cost per unit AC is not derived in this page, but the equations are the same...

[Robotbeat]

The rates matter because larger launch vehicle infrastructure cannot support the same rate as smaller launch vehicle infrastructure, other things being equal.

[93143]

It seems to me that if the LC-39 infrastructure were used at or near its limit, SLS would be a very good use of resources, significantly more cost-effective than the EELVs even at the same upmass rate.  The complaints are mostly about the abysmal flight rate in the pre-planning documents.

As an aside, while it is almost certainly true that (all else being equal) larger vehicles can't launch as often, you have failed to demonstrate that the upmass rate is lower.  In other words, Q may be constrained at a lower maximum, but it isn't at all obvious that mQ is too.

[QuantumG]

Yeah, and if wishes were horses we'd all be eatin' steak.

[Robotbeat]

It seems to me that if the LC-39 infrastructure were used at or near its limit, SLS would be a very good use of resources, significantly more cost-effective than the EELVs even at the same upmass rate.  The complaints are mostly about the abysmal flight rate in the pre-planning documents.

Are you comparing your mythical-launch-rate SLS to current-launch-rate EELVs? Or are you comparing EELVs at an equivalent mass-to-LEO rate, at which point the various options for reuse become viable?

(several have been studied for a partially reusable EELV, and one effort is already underway to replace the first stage of EELVs with a flyback booster by 2025/2030... the same time frame that Mars missions become relevant)

You're making an unfair comparison if you compare at the same launch rate. You must compare at the same IMLEO rate (especially, but not just, for propellant).

[93143]

Yeah, and if wishes were horses we'd all be eatin' steak.

Knock off the drive-by snark and pay attention.  What does the infrastructure capacity matter if we never reach it?

@Robotbeat:  Read what you quoted again.  I said "at the same upmass rate".  Those exact words.

In fact, I probably edited the thing after you read it, so you may want to go back and check...

The point I'm making is, so what if the infrastructure limits us to a certain launch rate?  (A) we'll never reach it, and (B) if we got anywhere near it, the flight rate would be high enough that SLS' economics would look great.

...

Basically, you sounded like you were claiming that SLS is too big because (presumably among other things) the launch infrastructure for such a large vehicle inherently limits the flight rate.  This is simply not true; SLS has never, to my knowledge, been criticized for poor economics at its maximum flight rate - the main complaint seems to be that the plan (such as it is) doesn't currently involve flying it anywhere near that rate.

[Robotbeat]

...
The point I'm making is, so what if the infrastructure limits us to a certain launch rate?  (A) we'll never reach it, and (B) if we did, the flight rate would be high enough that SLS' economics would look great.

But then you're making a totally unfair comparison between a hypothetical SLS that runs at a high launch rate and the current EELV at its low launch rate. AT 400mT/year around 2025+, EELVs would be easily justifiably partially reusable (and there's always Blue Origin, XCOR, SpaceX), and that's still a very low SLS launch rate (~4/year).

(FWIW, Lockheed says that their flyback booster becomes superior cost-wise to current EELVs even at just 8 launches per year... 400mT/year would be around 30-40 launches per year, a launch rate which is not without precedent but also is well beyond the 8 flights per year Lockheed says is worth it.)

Another issue is that you ignore other launch vehicles which would also be used for such an effort, such as Falcon 9 and Falcon Heavy (to say nothing of a reusable Falcon 9).

At just 4 launches per year, SLS must be compared to a partially reusable EELV or at least to Falcon 9 and Falcon Heavy, not to current EELVs. In other words, in order to get to the point where SLS can defeat current launch vehicles, you have to go well past the point where the existing efforts for partially reusable launch vehicles are more than justified.

[indaco1]

The rates matter because larger launch vehicle infrastructure cannot support the same rate as smaller launch vehicle infrastructure, other things being equal.

I agree, this phenomenon is very important...  in my model it can be represented  inside the  RC(Q,S) function (recurring Fixed Costs in function of total payload and size of the veichle).   

In other words, once fixed Q the RC - S curve is very steep for high S values.

I'm sorry I haven't found a more expressive model.

[93143]

...
The point I'm making is, so what if the infrastructure limits us to a certain launch rate?  (A) we'll never reach it, and (B) if we did, the flight rate would be high enough that SLS' economics would look great.

But then you're making a totally unfair comparison between a hypothetical SLS that runs at a high launch rate and the current EELV at its low launch rate. AT 400mT/year around 2025+, EELVs would be easily justifiably partially reusable (and there's always Blue Origin, XCOR, SpaceX), and that's still a very low SLS launch rate (~4/year).

(FWIW, Lockheed says that their flyback booster becomes superior cost-wise to current EELVs even at just 8 launches per year... 400mT/year would be around 30-40 launches per year, a launch rate which is not without precedent but also is well beyond the 8 flights per year Lockheed says is worth it.)

Another issue is that you ignore other launch vehicles which would also be used for such an effort, such as Falcon 9 and Falcon Heavy (to say nothing of a reusable Falcon 9).

At just 4 launches per year, SLS must be compared to a partially reusable EELV or at least to Falcon 9 and Falcon Heavy, not to current EELVs. In other words, in order to get to the point where SLS can defeat current launch vehicles, you have to go well past the point where the existing efforts for partially reusable launch vehicles are more than justified.

First, just to get it out of the way, this isn't just about raw upmass.  Single-chunk mass and volume are important too.  So you need to include the hit to spacecraft cost (if there is any, and at that upmass rate there will be) in the cost of the smaller vehicles.

Second, READ MY POSTS.  You asked a question.  I answered it.  You then proceeded to accuse me of unfairness, in the evident belief that I had answered the question oppositely to the way I actually did.

Finally, this economic argument is getting a bit handwavy.  I see little point in continuing it.  Just don't start asserting stuff as facts when it's derived from a mix of qualitative analysis and opinion.

(And SLS still isn't that great at 4 flights per year.  Not bad, but not great.  Take it to 24 and watch the EELV unit costs eat your lunch, even with partial reusability [it occurs to me that at that flight rate, partial reusability for SLS would make sense too...].  And no, you don't get to compare it with a full RLV; that's not only apples-to-oranges but is miles off your original complaint.)

[93143]

I'll repeat my earlier final paragraph, so you don't miss it:

Basically, you sounded like you were claiming that SLS is too big because (presumably among other things) the launch infrastructure for such a large vehicle inherently limits the flight rate.  This is simply not true; SLS has never, to my knowledge, been criticized for poor economics at its maximum flight rate - the main complaint seems to be that the plan (such as it is) doesn't currently involve flying it anywhere near that rate.

[QuantumG]

That seems like a reasonable point.. Given any fixed amount of money, there is an optimum size of launch vehicle for economic operations.

[Robotbeat]

I'll repeat my earlier final paragraph, so you don't miss it:

Basically, you sounded like you were claiming that SLS is too big because (presumably among other things) the launch infrastructure for such a large vehicle inherently limits the flight rate.  This is simply not true; SLS has never, to my knowledge, been criticized for poor economics at its maximum flight rate - the main complaint seems to be that the plan (such as it is) doesn't currently involve flying it anywhere near that rate.

My original point was in the context of optimal rates under a theoretical context about if the crawler (for example) limited the launch rate, but you responded about SLS. As far as I'm aware, most domestic launch vehicles (inc. SLS and the EELVs) are operating at well under the rate the infrastructure can handle, which is to say the infrastructure is not optimal for the current launch rate. Adding SLS to the mix makes the situation worse, since the infrastructure's optimal rate is FAR higher than will ever happen. SLS is suboptimal and will always be so because it will never operate at a level which utilizes its infrastructure to the fullest.

But my second point still stands. SLS, flying at high flight rate so as to optimally use the infrastructure, would be less economical than the alternatives (which includes partially reusable launch vehicles) at the same IMLEO rate.

So consider this a criticism of SLS's economics at maximum flight rate (a complete hypothetical, mind you). I know several others have made the same criticism, that at the launch rates where HLVs start to make sense, RLVs make even more sense. (And as far as minimum sized piece, again, that gets back to the limitations of road infrastructure and size of clean rooms and thermal vacuum chambers, etc, which strongly favors EELV-sized payloads over HLV-sized ones, even if you DO decide to launch said payloads on SLS.)

[93143]

My original point was in the context of optimal rates under a theoretical context about if the crawler (for example) limited the launch rate, but you responded about SLS.

You were talking about the Shuttle pad setup specifically.  What HLV is going to be using Shuttle infrastructure other than SLS?

But my second point still stands. SLS, flying at high flight rate so as to optimally use the infrastructure, would be less economical than the alternatives (which includes partially reusable launch vehicles) at the same IMLEO rate.

You haven't even begun to prove this.  You've simply asserted it.

...

Also, you're ignoring the possibility of any level of reusability of the HLV.  From what you said about the EELVs, it sounds like flyback boosters (for instance) would make sense long before the LC-39 infrastructure hit its maximum launch rate.

Finally, you're pulling a bait-and-switch, or having trouble staying on message, or both.  Your claim is that due to basic physics, infrastructure will limit the launch rate of an HLV, and thus prevent its being competitive with an MLV, all else being equal.  You then drag in RLVs when it becomes evident that the basic claim is not true.

The size of the pieces is a question of weighing facilities cost against increased complexity and mass in space, with a view towards a concrete desired capability.  The judgment simply cannot be made in the abstract; one needs more context than we have at this point.

[Robotbeat]

But my second point still stands. SLS, flying at high flight rate so as to optimally use the infrastructure, would be less economical than the alternatives (which includes partially reusable launch vehicles) at the same IMLEO rate.

You haven't even begun to prove this.  You've simply asserted it.

At 400mt/year, you're comparing an expendable launch vehicle still flying at a suboptimal rate to a (mostly) reusable launch vehicle flying at an optimal rate. This very likely changes the situation dramatically. Do you (given my assumptions) disagree? I mean, if you take my assumptions, can you see how my conclusion follows?

(You are, of course, welcome to question my assumptions.)

[93143]

I'm somewhat surprised at you.  That's nothing but qualitative handwaving; no conclusions can be drawn.  The phrase "very likely" betrays the fact that you realize this.

And where does 400 mT/year come from?  It has nothing to do with LC-39's launch rate limits, which are several times higher at least.  I'm sure you recall that high upmass rates tilt the economic balance towards HLVs.  Is it enough to overcome the advantage of flyback first stages (versus perhaps flyback boosters on SLS)?  No way to tell without numbers.

[Robotbeat]

I'm somewhat surprised at you.  That's nothing but qualitative handwaving; no conclusions can be drawn.  The phrase "very likely" betrays the fact that you realize this.

And where does 400 mT/year come from?  It has nothing to do with LC-39's launch rate limits, which are several times higher at least.  I'm sure you recall that high upmass rates tilt the economic balance towards HLVs.  Is it enough to overcome the advantage of flyback first stages (versus perhaps flyback boosters on SLS)?  No way to tell without numbers.

400mT/year launch rate comes from the rate at which we can be reasonably certain that at very least partial reusability for EELV-class (8-25mT) is warranted. I've heard estimates as low as 8 flights per year for partial reusability (flyback EELV first stage currently being designed and tested at subscale by Lockheed Martin) but about 40 flights or more for full reusability.

This is the point where we can say that the transition from expendable to at very least partial reusable makes sense with some certainty. That's why I chose it. You do agree that reusability is required to get where we all want to go (which is becoming truly spacefaring), right? This is, after all, why we switched to Shuttle, even if it failed its goals of low cost per kg to orbit because of several reasons which are beyond the scope of this thread. Those pioneers which pushed for Shuttle in the early days were absolutely right that reusability is essential.

Right now, reusability wouldn't require the enormous program that Shuttle was. There are several concurrent efforts underway as we speak. The only thing that is likely to cause none of them to succeed is too low of launch demand.  But the odds are tilted in their favor at 400mT, especially if that's mostly propellant. The equation then changes, because you're talking about mostly reusable launch vehicles at a high launch rate  versus at least mostly expendable (i.e. SLS still at a relatively low 4/year). Comparing at higher launch rates only tilts the odds more in favor of reusability.

[Chris Bergin]

Was away, so missed the mod report, but everyone needs to keep this thread calm. Argue the point, not the person.

[93143]

There's a big difference between partial reusability (such as flyback boosters on EELVs or SLS) and full reusability (such as Skylon or Falcon 9R).  The latter is a massive paradigm shift if it can be done cheaply enough, and your simple physics-based objection to high launch rates of an HLV seems a bit apples-to-oranges if you're comparing an expendable HLV to a reusable MLV.

...

When you write that higher flight rates "tilt the odds" even more towards reusability, you seem to conflate favourable slope vs. the expendable EELV with favourable slope vs. the HLV.

Basically, you haven't demonstrated that a partially-reusable EELV would have a lower cost per kg floor than SLS at the latter's maximum upmass rate.  Remember how badly J-246 stomped the Atlas V at very high flight rates?  I know SLS is not Jupiter, but it should actually be more cost-effective per kg once development is over and if you assume you can fully utilize it, so there is probably still a significant gap to make up.  (Never mind that at SLS' maximum flight rate, flyback boosters almost certainly make sense for it too...  and perhaps an RS-25 recovery pod...)

No one has yet demonstrated that a fully-reusable launch vehicle can even be built at all, so that comparison is a bit facile.  Obviously if something like Skylon works out, servicing a LEO propellant depot with SLS would be silly.  Launching a very large and/or heavy module that needs to be ground-integrated, though, is something that Skylon flat-out can't do, so in a robust exploration program, an expendable HLV would probably still have a place unless and until a reusable one shows up...

...

Okay, returning to the original point:

I don't see how your original argument regarding LC-39's maximum launch rate (the key phrase was "other things being equal") proves anything useful.  Furthermore, while the theoretical max launch rate may be lower, you don't need it to be as high to achieve the same upmass, and you didn't demonstrate that the maximum upmass would be lower, and certainly not that the cost per kg floor would be higher.  That was all I was trying to say, and as the rest of the argument is getting nowhere fast, I should probably take my own advice and back off.

[Also, I think you might have missed a few edits, judging by how things have been going...]

[93143]

Excuse me - my profuse apologies; I appear to have brought the SLS vs. EELVs fight to the Advanced Concepts board.  My bad...

[Robotbeat]

Excuse me - my profuse apologies; I appear to have brought the SLS vs. EELVs fight to the Advanced Concepts board.  My bad...

And I contributed.

I actually think we clarified our respective positions pretty well and come to a better understanding of the other's point of view. And that's worthwhile, I think.

[Rhyshaelkan]

You had a point early on, Robotbeat, about engine throat size. I would think that might have bearing on how rockets can scale, or an optimum found.

Do smaller engines refurbish better than larger engines? Do smaller engines lend themselves better than larger engines where flyback or recovery is concerned?

We need to recover people from space after we launch them. An Orion or Dragon style vehicle is relatively restricted in size by the diameter of the rocket. Launching an asymmetrical cylinder with TPS on one half might allow more people to travel without increasing the diameter of the rocket. Instead of an exponentially larger diameter rocket and/or exponentially larger fairing to facilitate a larger Orion or Dragon. A half fairing to make the asymmetrical cylinder aerodynamic might be a better way to fly more people with today's rockets.

[indaco1]

Do smaller engines refurbish better than larger engines?

IMHO: a "small because low pressure" engine refurbish much better.

An expander cycle low pressure low temperature engine could be extremely durable but the thrust is limited by the square-cube rule, due the way cooling is achieved in the walls of the combustion chamber.

Furthermore a low thrust per total nozzle surface area exacerbate the problem noted by Rotobeat that make very big launchers have to be "fat", again due the  square-cube law.

Conclusion:  perhaps the way to make a veichle really reusable (many launches without servicing and long service life) is to use relatively low pressure engines and keep it relatively small because low pressure engines don't scale very well.

I know altitude compensation is difficult with low pressure engines, but this is another story.

But I'm not an expert, please correct if I'm wrong.

[JohnFornaro]

This is a very basic equation found in the microeconomics literature:

AC = (SC + RC)/Q + V

where:

AC = Average cost per unit
SC=Startup cost  (developement and infrastructure cost)
RC=Recurring Fixed Costs  (eg. infrastructure to mantain, staff to pay etc.).
V = Variable cost per unit
Q = Number of units produced

Basic to you maybe, but totally new to me!  But thanks for posting; simple equations I can handle...

What I'm reading here is that Q is the only variable which forum posters can posit with a small degree of certainty.  AC can never be found, since none of the other values can be set to a broad agreement.

Which is my way of saying: "The problem is that the functions SC(Q,S), RC(Q,S) and V(Q,S) are pretty unkonwn ..."

But is there another possible equation.  since nobody knows, with sufficient certainty, what anything costs, is there a mass/thrust optimum.  Given an RP-1 engine, say, at what point does the mass/thrust ratio start deteriorating?  And just leave cost alone for the moment.

It occurs to me that it's also possible to constrain cost somehow, based on previous history.  Then, what mass do you get?

The OP begs the question:  What could this optimal mass be?  So far the thread debates assumptions; the possible biggest smallest chunk of mass continues to be unknown.

[indaco1]

..
The OP begs the question:  What could this optimal mass be?  So far the thread debates assumptions; the possible biggest smallest chunk of mass continues to be unknown.

Some other notes: 

- for veichles that already exists SC is zero for any additional production.

- for veichles derived by existing veichles SC is low and commonality allow them to share infrastructure and staff, ie even RC is low

- scale economies achieved by a big factory mass producing medium lift launchers probably dwarf scale economies achieved by an unefficiently manifactured heavy lift. 

Thus I'm not certain that V will be lower with an HLLV than with a mass produced medium lift.  If so the cost equation say that there are no doubts this HLLV is a loser.

If there's a technological or industrial breakthrough all the above has no value.
Eg. AC (Price/kg) declared for Falcon Heavy, if confirmed, will probably make it a winner.

[93143]

Thus I'm not certain that V will be lower with an HLLV than with a mass produced medium lift.  If so the cost equation say that there are no doubts this HLLV is a loser.

I think you mean V/m, where m is the payload to orbit.  And it is almost certainly substantially lower for an HLV than for an MLV, all else being equal, and in the context of the widely-understood payload ranges for these terms.

Also, regarding "mass production", the production methods are similar, if not more advanced for the HLV.  It's not like you're comparing an all-robot line at Decatur with a bunch of blacksmiths at Michoud.  The main advantage of "mass production" in this context is minimization of (SC+RC)/Q.

Backing off now, so as to not stir the mud too much...

[peter-b]

For ballistic re-entries, how large can capsules get? Do they perform better or worse as they get larger? Is the idea of a 50 passenger re-entry capsule silly?

[Robotbeat]

For ballistic re-entries, how large can capsules get? Do they perform better or worse as they get larger? Is the idea of a 50 passenger re-entry capsule silly?

If you're trying to maximize drag and minimize terminal velocity, then capsules perform worse as they scale up linearly. But there's other things to take into account.

[Rhyshaelkan]

ESA's IXV seems to go in the direction I was thinking.

http://www.zdnet.co.uk/news/emerging-tech/2011/06/28/europe-set-to-build-wedge-shaped-spacecraft-40093235/5/

Elongate the vehicle to provide more passenger seating without changing the diameter of the rocket. As long as you stay within payload mass, seems like it would work. Completely from a non-engineer view.

[Robotbeat]

ESA's IXV seems to go in the direction I was thinking.

http://www.zdnet.co.uk/news/emerging-tech/2011/06/28/europe-set-to-build-wedge-shaped-spacecraft-40093235/5/

Elongate the vehicle to provide more passenger seating without changing the diameter of the rocket. As long as you stay within payload mass, seems like it would work. Completely from a non-engineer view.

Yeah, scaling might work better if you scale in that manner. Obviously, other complications as well. Also, the bending loads will increase the required structural mass disproportionately.

[peter-b]

For ballistic re-entries, how large can capsules get? Do they perform better or worse as they get larger? Is the idea of a 50 passenger re-entry capsule silly?

If you're trying to maximize drag and minimize terminal velocity, then capsules perform worse as they scale up linearly. But there's other things to take into account.

Looking at dimensions, let's say we want to keep the ballistic coefficient Cb constant.

Cb=M/(Cd × A)

where M is spacecraft mass, Cd is drag coefficient, and A is cross-sectional area. Let's also (possibly fallaciously) assume that we simply linearly scale up all dimensions the spacecraft, and that mass is proportional to volume, giving:

Cb = k × d³ / (Cd × d²) = k × d / Cd

where d is a reference length.

So capsule re-entry heatshield braking performance indeed degrades quite quickly with scale, if nothing can be done to increase the drag coefficient. I guess that answers my own question... does anyone know what Cd is for e.g. Dragon?

Elongate the vehicle to provide more passenger seating without changing the diameter of the rocket. As long as you stay within payload mass, seems like it would work. Completely from a non-engineer view.

Per equations above, this doesn't work because you'd be greatly increasing M without changing A, surely?

[Robotbeat]

I interpreted his remark as increasing the length and reentering on the 'side,' thus the Area:Mass ratio would remain roughly the same (again, not counting an increase in structural mass from the larger bending loads, etc).

[peter-b]

I interpreted his remark as increasing the length and reentering on the 'side,' thus the Area:Mass ratio would remain roughly the same (again, not counting an increase in structural mass from the larger bending loads, etc).

Fair enough, but that's not "scaling", strictly speaking. :-P

[Robotbeat]

I interpreted his remark as increasing the length and reentering on the 'side,' thus the Area:Mass ratio would remain roughly the same (again, not counting an increase in structural mass from the larger bending loads, etc).

Fair enough, but that's not "scaling", strictly speaking. :-P

Agreed.

[Rhyshaelkan]

Yes the original post is trying to find that 'sweet spot' of launching people into space. Rather than scale up rockets, I was trying to find alternatives to fly people to and from space. Without relying on hyperly expensive spaceplanes.

I think we have great rockets in the cupboards. Prices to fly are coming down with more competition. Rework what we have to do what we need.

[gbaikie]

For ballistic re-entries, how large can capsules get? Do they perform better or worse as they get larger? Is the idea of a 50 passenger re-entry capsule silly?

Apollo return capsule was about 4 meters in diameter. Soyuz SA is 2.17 meters. And the 7 crew Dragon is 3.7 meters. [wiki]

If launch a much larger diameter capsule 50 passenger return is
possible.
The Shuttle orbiter [not a capsule] could have put a passenger bus
in it's cargo bay.
Shuttle Orbiter Specifications:
Gross Liftoff Weight: 240,000 lb
Empty Weight: 172,000 lb
Maximum Landing Weight: 230,000 lb
Maximum Payload: 55,250 pounds (25,060 kg)
Payload Bay dimensions: 15 ft by 60 ft (4.6 m by 18.3 m)

If allow 400 lbs per passenger. 50 is 20,000 lb. Of course allowing for them to live in orbit for a day or two to weeks is a different question as is
the idea of how one entertains such a crowd.

I would guess the shuttle orbiter is about equal to capsule of 20 meters in diameter.
But total mass of shuttle orbiter [Maximum Landing Weight: 230,000 lb] gets you lot potential seats. So if get 20 meter diameter capsule into space, one could land hundreds of passengers.

I believe largest current fairing is around 5 meters in diameter.

"The standard payload fairing sizes are 4 or 5 meters in diameter and of various lengths, are made by RUAG Space. Fairings sizes as large as 7.2m in diameter and up to 32.3m in length have been considered."
http://en.wikipedia.org/wiki/Atlas_V
And if wanted it, one could probably get a fairing diameter as much as 10 meters in diameter. And btw, if you be serious about Mars exploration, you need such fairing sizes.
So if had 50 passengers to return, one could get 10 meter diameter fairing on any of the existing rockets and a capsule which was 10 meter in diameter could land 50 people. Perhaps even with 7.2 meter one could stack 50 people into it.
7.2 meter is about 53 square meter
And 3.7 meters [Dragon] is 14 square meter.

Probably easier with 10 meter diameter. One could land a 20,000 lb payload with 7.2 diameter- but with 50 people there probably great value in having a "calmer, more spacious descent".

 

[RanulfC]

Again for everyone's "edification" as it were I present the "Para-Shield" concpet:
http://www.nianet.org/rascal/forum2006/presentations/1010_umd_paper.pdf

http://spacecraft.ssl.umd.edu/publications/2010/SpaceOps2010ParaShieldx.pdf

http://www.planetaryprobe.org/SessionFiles/Session4/Papers/Rohrschneider_Inflat&Deploy-Paper.pdf

Enjoy...

Randy