Author Topic: Falcon and Dragon reusability  (Read 22840 times)

Offline tnphysics

  • Regular
  • Full Member
  • ****
  • Posts: 1073
  • Liked: 0
  • Likes Given: 0
Falcon and Dragon reusability
« on: 01/01/2009 07:15 PM »
SpaceX is planning to reuse the first stage of the F1, and all of the F9 and (probably) F9 Heavy.  This would make the F9 the first RLV.  Dragon will be the first reusable capsule.  How will reusability work out?

Offline ChefPat

  • Full Member
  • ****
  • Posts: 1051
  • Earth, for now
  • Liked: 123
  • Likes Given: 1020
Re: Falcon and Dragon reusability
« Reply #1 on: 01/01/2009 08:16 PM »
SpaceX is planning to reuse the first stage of the F1, and all of the F9 and (probably) F9 Heavy.  This would make the F9 the first RLV.  Dragon will be the first reusable capsule.  How will reusability work out?

The Shuttle is an RLV. Gemini was designed with reusability in mind.
Playing Politics with Commercial Crew is Un-American!!!

Offline William Graham

  • Senior Member
  • *****
  • Posts: 4142
  • Liked: 128
  • Likes Given: 77
Re: Falcon and Dragon reusability
« Reply #2 on: 01/01/2009 09:46 PM »
SpaceX is planning to reuse the first stage of the F1, and all of the F9 and (probably) F9 Heavy.  This would make the F9 the first RLV.  Dragon will be the first reusable capsule.  How will reusability work out?

The Shuttle is an RLV. Gemini was designed with reusability in mind.

The Shuttle ET is not reusable. Titan II was not reusable, and in practice, only one Gemini capsule flew twice.
« Last Edit: 01/01/2009 09:46 PM by GW_Simulations »

Offline Lars_J

  • Senior Member
  • *****
  • Posts: 6161
  • California
  • Liked: 665
  • Likes Given: 195
Re: Falcon and Dragon reusability
« Reply #3 on: 01/01/2009 10:40 PM »
SpaceX is planning to reuse the first stage of the F1, and all of the F9 and (probably) F9 Heavy.  This would make the F9 the first RLV.  Dragon will be the first reusable capsule.  How will reusability work out?
I think SpaceX is concentrated on getting the F1/F9 to work reliably before seeing *IF* it can be reused. And *IF* it will make economic sense to refurbish recovered stages.

If it will be cheaper to refurbish/reuse, they will do it.

Offline ChefPat

  • Full Member
  • ****
  • Posts: 1051
  • Earth, for now
  • Liked: 123
  • Likes Given: 1020
Re: Falcon and Dragon reusability
« Reply #4 on: 01/01/2009 10:41 PM »
SpaceX is planning to reuse the first stage of the F1, and all of the F9 and (probably) F9 Heavy.  This would make the F9 the first RLV.  Dragon will be the first reusable capsule.  How will reusability work out?

The Shuttle is an RLV. Gemini was designed with reusability in mind.

The Shuttle ET is not reusable. Titan II was not reusable, and in practice, only one Gemini capsule flew twice.

Space Shuttle Basics
The space shuttle is the world's first reusable spacecraft
http://spaceflight.nasa.gov/shuttle/reference/basics/index.html

I'd read that only one Gemini was reused, but what I'd said was "Gemini was designed with reusability in mind."
Playing Politics with Commercial Crew is Un-American!!!

Offline tnphysics

  • Regular
  • Full Member
  • ****
  • Posts: 1073
  • Liked: 0
  • Likes Given: 0
Re: Falcon and Dragon reusability
« Reply #5 on: 01/02/2009 05:53 AM »
I said reusable "capsule" - the Orbiter has wings.

Offline mlorrey

  • Member
  • Senior Member
  • *****
  • Posts: 2133
  • International Spaceflight Museum
  • Grantham, NH
  • Liked: 10
  • Likes Given: 0
Re: Falcon and Dragon reusability
« Reply #6 on: 01/02/2009 09:01 AM »
SpaceX is planning to reuse the first stage of the F1, and all of the F9 and (probably) F9 Heavy.  This would make the F9 the first RLV.  Dragon will be the first reusable capsule.  How will reusability work out?

From what I read, the Chinese capsule was using a wetted wood heat shield with the idea being it being replaced each mission, not sure if they've actually reused it at all. I heard they'd replaced that type of shield with something more modern, at least partly out of concern of publicity at the idea of flying wooden spaceships.

Generally speaking, if a goal of your manned program is nationalist propaganda, it makes more sense to use each spacecraft once and then mount them in museums and outside government buildings. Ergo, reusability is only a factor when money matters more.
VP of International Spaceflight Museum - http://ismuseum.org
Founder, Lorrey Aerospace, B&T Holdings, ACE Exchange, and Hypersonic Systems. Currently I am a venture recruiter for Family Office Venture Capital.

Offline hop

  • Senior Member
  • *****
  • Posts: 3276
  • Liked: 428
  • Likes Given: 775
Re: Falcon and Dragon reusability
« Reply #7 on: 01/02/2009 11:12 PM »
Generally speaking, if a goal of your manned program is nationalist propaganda, it makes more sense to use each spacecraft once and then mount them in museums and outside government buildings. Ergo, reusability is only a factor when money matters more.
Do you honestly believe that having a souvenir to display is a significant factor in the non re-usability of past manned capsules ? Not something more prosaic, like say, maximizing usable payload fraction, or the high cost of refurbishing ?

ISTR China used wood on some early recoverable capsules, but  not on Shenzhou. The suggestion that this change was merely to appear more modern seems unreasonable: Consistency  and mass seem far more likely.

Offline mlorrey

  • Member
  • Senior Member
  • *****
  • Posts: 2133
  • International Spaceflight Museum
  • Grantham, NH
  • Liked: 10
  • Likes Given: 0
Re: Falcon and Dragon reusability
« Reply #8 on: 01/03/2009 04:00 PM »
Generally speaking, if a goal of your manned program is nationalist propaganda, it makes more sense to use each spacecraft once and then mount them in museums and outside government buildings. Ergo, reusability is only a factor when money matters more.
Do you honestly believe that having a souvenir to display is a significant factor in the non re-usability of past manned capsules ? Not something more prosaic, like say, maximizing usable payload fraction, or the high cost of refurbishing ?

ISTR China used wood on some early recoverable capsules, but  not on Shenzhou. The suggestion that this change was merely to appear more modern seems unreasonable: Consistency  and mass seem far more likely.

The question is, can you mass produce spacecraft for throwaways cheaper per unit than building a one off and refurbing for each mission.

And how important is preventing LOC to less than 1/1500 missions? Production rates implies a greater risk of any given vehicle having a manufacturing error significant enough to cause LOC, whereas building a 'perfect' vehicle that gets refurbed with care should give lower risk given sufficiently perfected design.

If you accept as a given the replacement of the TPS with each mission (as would typically happen with any capsule and which is generally seen even with the STS, with over half of refurb man-hours going to TPS maintenance with each mission) then one component of any vehicle design is to make it easy to remove and replace the TPS from and to the airframe, whether its a capsule, lifting body, launch stage, or winged vehicle.

If you cant accomplish this, then production line methods should be used with very high quality standards and an acceptance of risks from individual vehicles not being tested throughout their flight envelope prior to launch.
VP of International Spaceflight Museum - http://ismuseum.org
Founder, Lorrey Aerospace, B&T Holdings, ACE Exchange, and Hypersonic Systems. Currently I am a venture recruiter for Family Office Venture Capital.

Offline 2.71

  • Member
  • Posts: 51
  • Liked: 1
  • Likes Given: 2
Re: Falcon and Dragon reusability
« Reply #9 on: 01/03/2009 06:07 PM »
Huh?
"Your political views are so interesting to me!"

Offline pippin

  • Regular
  • Senior Member
  • *****
  • Posts: 2553
  • Liked: 289
  • Likes Given: 39
Re: Falcon and Dragon reusability
« Reply #10 on: 01/04/2009 10:51 PM »
Production rates implies a greater risk of any given vehicle having a manufacturing error significant enough to cause LOC, whereas building a 'perfect' vehicle that gets refurbed with care should give lower risk given sufficiently perfected design.

Where on earth do you get that from????
Stable, industrial processes are typically orders of magnitude more reliable than one-off builds and refurbishing things means trying anew each time.

I would love to see ONE example where your theory holds true, can give you dozens (cars, planes, computers,...) where it's the other way around.

Offline mlorrey

  • Member
  • Senior Member
  • *****
  • Posts: 2133
  • International Spaceflight Museum
  • Grantham, NH
  • Liked: 10
  • Likes Given: 0
Re: Falcon and Dragon reusability
« Reply #11 on: 01/05/2009 10:28 PM »
Production rates implies a greater risk of any given vehicle having a manufacturing error significant enough to cause LOC, whereas building a 'perfect' vehicle that gets refurbed with care should give lower risk given sufficiently perfected design.

Where on earth do you get that from????
Stable, industrial processes are typically orders of magnitude more reliable than one-off builds and refurbishing things means trying anew each time.

I would love to see ONE example where your theory holds true, can give you dozens (cars, planes, computers,...) where it's the other way around.

Industrial processes achieve reliability for MOST of a production run by using quality processes to toss out items that dont pass. Statistically speaking, in any production run, there will ALWAYS be items on the end of the bell curve that will fail. Doing full quality inspections on every item produced is also prohibitively expensive.

So you've seen better reliability with the entire fleet of Ford Escorts vs a one-off Rolls Royce or other hand built custom?

NDI technologies can be equally applied to both mass produced and one-offs, doing it for mass produced items is obviously cost prohibitive.
VP of International Spaceflight Museum - http://ismuseum.org
Founder, Lorrey Aerospace, B&T Holdings, ACE Exchange, and Hypersonic Systems. Currently I am a venture recruiter for Family Office Venture Capital.

Offline pippin

  • Regular
  • Senior Member
  • *****
  • Posts: 2553
  • Liked: 289
  • Likes Given: 39
Re: Falcon and Dragon reusability
« Reply #12 on: 01/05/2009 11:47 PM »
Production rates implies a greater risk of any given vehicle having a manufacturing error significant enough to cause LOC, whereas building a 'perfect' vehicle that gets refurbed with care should give lower risk given sufficiently perfected design.

Where on earth do you get that from????
Stable, industrial processes are typically orders of magnitude more reliable than one-off builds and refurbishing things means trying anew each time.

I would love to see ONE example where your theory holds true, can give you dozens (cars, planes, computers,...) where it's the other way around.

Industrial processes achieve reliability for MOST of a production run by using quality processes to toss out items that dont pass.
Industrial processes achieve reliability through application of a learning curve and elimination of errors. You cannot do that with just one unit.
Quote
Statistically speaking, in any production run, there will ALWAYS be items on the end of the bell curve that will fail. Doing full quality inspections on every item produced is also prohibitively expensive.
Yes. And statistically speaking that can as well be your single unit. Full inspections are absolutely common for a lot of items, even in the automotive industry; each computer chip undergoes extensive testing and here we are talking units that sell at single to three digit dollar values.
For expensive items, even doing full inspections after intermediate production steps is common place.
Quote
So you've seen better reliability with the entire fleet of Ford Escorts vs a one-off Rolls Royce or other hand built custom?
Definitely. Look up reliability of Rolls Royce when they were still mainly hand built and you'll be shocked. No comparison to normal mass produced cars. "Hand made in England" is a common synonym for bad quality in the auto industry (Sorry, Chris).
Rolls Royce compensated for this by applying excessive service but that's pretty much the "Space Shuttle" business model, and especially it does definitely NOT reduce cost. Believe me, TCO of a Rolls Royce will be much higher than for a Ford Focus, and your point was it's cheaper to use RLVs.
Quote
NDI technologies can be equally applied to both mass produced and one-offs, doing it for mass produced items is obviously cost prohibitive.
No, as I said above, it's commonplace.

What kind of quality processes (for which products) have you worked on?
« Last Edit: 01/05/2009 11:50 PM by pippin »

Offline mlorrey

  • Member
  • Senior Member
  • *****
  • Posts: 2133
  • International Spaceflight Museum
  • Grantham, NH
  • Liked: 10
  • Likes Given: 0
Re: Falcon and Dragon reusability
« Reply #13 on: 01/06/2009 12:18 AM »
Where on earth do you get that from????
Stable, industrial processes are typically orders of magnitude more reliable than one-off builds and refurbishing things means trying anew each time.

I would love to see ONE example where your theory holds true, can give you dozens (cars, planes, computers,...) where it's the other way around.

Industrial processes establish "reliability" by minimizing the failure rate, but still, 1 million devices with a 0.001 failure rate still means you have a thousand failures in a given time span. If that device is a one man space capsule, that means you accept 1000 deaths as an acceptable mortality rate.

You point at "oh, but we flew 999,000 people to space safely"...

Now, imagine manufacturing air liners to be single usage vehicles, you throw them away after each flight. You'd need to manufacture tens of millions of planes each year, and your accident rate would be far higher.

Quote
Industrial processes achieve reliability through application of a learning curve and elimination of errors. You cannot do that with just one unit.

I dont think thats something to be assumed. You eliminate errors with a full and thorough test program.

Quote
Yes. And statistically speaking that can as well be your single unit. Full inspections are absolutely common for a lot of items, even in the automotive industry; each computer chip undergoes extensive testing and here we are talking units that sell at single to three digit dollar values.
For expensive items, even doing full inspections after intermediate production steps is common place.
Definitely. Look up reliability of Rolls Royce when they were still mainly hand built and you'll be shocked. No comparison to normal mass produced cars. "Hand made in England" is a common synonym for bad quality in the auto industry (Sorry, Chris).
Rolls Royce compensated for this by applying excessive service but that's pretty much the "Space Shuttle" business model, and especially it does definitely NOT reduce cost. Believe me, TCO of a Rolls Royce will be much higher than for a Ford Focus, and your point was it's cheaper to use RLVs.
[
No, as I said above, it's commonplace.

What kind of quality processes (for which products) have you worked on?

Both military contracting and commercial product development/manufacturing. This is actually a good question to ask, because as I mentioned in a prior thread, I invented this retrofit kit for exit signs back in the early 1990's. It used electroluminescent technology, and we initially went with the best EL lamp maker in the business to build the whole kit for us. They were historically a military contractor, so they engineered it to the nines with all the bells and whistles as to what they felt was needed to pass UL Standard 924. What they built for us (totally different from the design we specced) they charged us $65 per unit for, which was the price we needed to retail it at.

My father was the sales engineer there and retired that year, he had no input into what was done to our product by the engineering department, he started his own company manufacturing EL lamps for us and the whole kit for us to our own design. Our design passed UL Standard 924 with flying colors but only cost me $19 each (that was dad's price to my company, he was making a 40% margin on it).

I had previously worked for the first manufacturer myself, dad and I developed EL formation lights for fighter planes and transports (F-15, F-4, F-117, KC-10, F-18, AV-8B, among others). Our formation lights were considered the best in the industry.

So I know a bit about production engineering and quality assurance in both the military and commercial sectors. I have a good idea about why government contractors tend to need 3000-5000% markup over the cost of materials. Absolutely every specification gets engineered from first principles and to too tight tolerance even when it obviously does not need to be so. Engineers like to maximize their billable hours just like attorneys do.

So, an example of the one-off perfect design: A customer came to me, he ran maintenance for a school district, said the kids kept jumping up and ripping exit signs off the walls and ceilings. So I took some 1/4" plate steel, cut, welded, and drilled them into some mounts. We sandwiched some of our EL exit sign lamps in fiberglass, autoclaved them. They were mounted to the steel structure of the buildings. No more problem. If we'd given the problem to those engineers at the first company, they'd have taken four months to produce a design that cost over $1000 a pop and was engineered to resist the vandalism of a 180 lb kid but not a 250 lb kid. 95% of the cost would have been engineering overhead. I got the job done for $100, and last I heard those suckers were still there 15 years later.

Did I need to spend time doing the calcs to engineer it for a maximum weight tolerance? Not really, knowing the established structural strength of the materials was all that was needed to be known. But I did built it so that it was not going to fail at any point within its usage regime.

Now, you can say, "but with spacecraft, you have to engineer to save weight". Yes you do, I had that experience working on the engineering for the formation lights for the A-12 program, which is classified so I can't talk about that here.
« Last Edit: 01/06/2009 12:27 AM by mlorrey »
VP of International Spaceflight Museum - http://ismuseum.org
Founder, Lorrey Aerospace, B&T Holdings, ACE Exchange, and Hypersonic Systems. Currently I am a venture recruiter for Family Office Venture Capital.

Offline 2.71

  • Member
  • Posts: 51
  • Liked: 1
  • Likes Given: 2
Re: Falcon and Dragon reusability
« Reply #14 on: 01/06/2009 06:51 PM »
************************
mlorrey Quote:
Industrial processes establish "reliability" by minimizing the failure rate, but still, 1 million devices with a 0.001 failure rate still means you have a thousand failures in a given time span. If that device is a one man space capsule, that means you accept 1000 deaths as an acceptable mortality rate.

You point at "oh, but we flew 999,000 people to space safely"...
*************************

So following this logic, if you build 2 one-off perfect vehicles, then they become less safe.

But of course that is not true. If a design failure rate is 1 in 1000, then your perfect one-off has a 1 in 1000 chance of failure. But if you build 1000, the failure rate is STILL 1 in 1000 PER VEHICLE.

The only plausible argument otherwise is the assumption that it is possible to 100% perfectly test a one-off design, but it is not if you build many.

Non sequitor.

Clearly, you can test every vehicle. Even if you build 2.

The argument that "fewer vehicles equals better safety" is true in one case. A unbuilt paper design is the safest.

2.71
"Your political views are so interesting to me!"

Offline mlorrey

  • Member
  • Senior Member
  • *****
  • Posts: 2133
  • International Spaceflight Museum
  • Grantham, NH
  • Liked: 10
  • Likes Given: 0
Re: Falcon and Dragon reusability
« Reply #15 on: 01/06/2009 09:34 PM »
************************
mlorrey Quote:
Industrial processes establish "reliability" by minimizing the failure rate, but still, 1 million devices with a 0.001 failure rate still means you have a thousand failures in a given time span. If that device is a one man space capsule, that means you accept 1000 deaths as an acceptable mortality rate.

You point at "oh, but we flew 999,000 people to space safely"...
*************************

So following this logic, if you build 2 one-off perfect vehicles, then they become less safe.


I realize it seems nonintuitive, but its the same sort of logic that says that when presented three choices and two chances to choose, the second chance you should change which of the two remaining choices you'd go with.

Ok lets say you expect a given spacecraft design to fly 1000 missions over its term of service.

You could build 10 vehicles to fly 100 times each, 100 vehicles to fly 10 times each, or 1000 vehicles to fly 1 time each.

You want the chance of mission failure to obviously be significantly greater than the total number of missions you plan to fly. i.e. if you build it to have a 1 in 2000 chance of failure, and you fly 999 successful missions, the last one has a 50% chance of failure (ya I know standard probabilities says every mission is a 0.0005 chance of failure, Murphy says otherwise), which you can cancel and have a perfect safety record.

Anyways, the power law says you achieve 80% of your goals for 20% of the cost of achieving 100% of your goals.

This means that for a 1000 ship production run, if it costs you a million USD per ship in manufacturing QA to guarantee odds of 1 in 1600, thats a billion dollar QA budget over the lifetime of the vehicle on a one mission per vehicle plan.

Simplistically, if you build just 10 vehicles, and spend 5 million USD per vehicle to guarantee 1 in 2000 odds for those 10 vehicles, your manufacturing QA budget would be 50 million.

Now, its not quite as simple as that, the manufacturing cost of a vehicle that can handle 100 missions is obviously going to be a lot more than one that can handle 1 mission. Being more robust it will likely have poorer performance payload wise. You will have additional QA costs to do with ongoing maintenance that a one-mission bird doesnt need to deal with.

Theres a huge engineering paradigm conflict between the single use per engineers and the reusables engineers. Its the difference between ammunition and airplanes. Imagine trying to run the Berlin Airlift with DC-3's that could only fly one mission. Not possible.

A 747 is a far more complex vehicle than a Saturn V: more systems, more moving parts. A 747 is a vehicle, a Saturn V is ammunition.

The ammunition paradigm wins for space launch only if you never plan on offering services to a mass market.

Compare, for instance, an SR-71 to a fictional intercontinental reconnaissance missile you used once. You could say "Ah but Corona was such". Every orbit of a spy sat is like a new flight sortie, the comparison between spysats and spyplanes is on of automation vs humans, not reusables vs single missions. Take an MX missile that would do a spy mission. You'd need thousands to fly the same lifetime that the SR-71 did, much more expensively.
VP of International Spaceflight Museum - http://ismuseum.org
Founder, Lorrey Aerospace, B&T Holdings, ACE Exchange, and Hypersonic Systems. Currently I am a venture recruiter for Family Office Venture Capital.

Offline 2.71

  • Member
  • Posts: 51
  • Liked: 1
  • Likes Given: 2
Re: Falcon and Dragon reusability
« Reply #16 on: 01/06/2009 11:05 PM »
1.  if you build it to have a 1 in 2000 chance of failure, and you fly 999 successful missions, the last one has a 50% chance of failure.

***That is plainly untrue.

2. (ya I know standard probabilities says every mission is a 0.0005 chance of failure, Murphy says otherwise), which you can cancel and have a perfect safety record.

***Great, then we agree. Murphy applies equally to both one-off and mass produced vehicles.

2. This means that for a 1000 ship production run, if it costs you a million USD per ship in manufacturing QA to guarantee odds of 1 in 1600, thats a billion dollar QA budget over the lifetime of the vehicle on a one mission per vehicle plan.

Simplistically, if you build just 10 vehicles, and spend 5 million USD per vehicle to guarantee 1 in 2000 odds for those 10 vehicles, your manufacturing QA budget would be 50 million.

***This is based off the faulty logic of 1.

2.71
"Your political views are so interesting to me!"

Offline stexer

  • Member
  • Posts: 17
  • Liked: 0
  • Likes Given: 0
Re: Falcon and Dragon reusability
« Reply #17 on: 01/06/2009 11:24 PM »
SpaceX is planning to reuse the first stage of the F1, and all of the F9 and (probably) F9 Heavy.  This would make the F9 the first RLV.  Dragon will be the first reusable capsule.  How will reusability work out?

You are correct. If they find it economically feasible and choose to reuse all the stages, then the F9 would be the first RLV, since the Space Shuttle is not a LV, it's an orbiter(spacecraft). The F9 is a RLV (Reusable Launch Vehicle). The Space Shuttle can't reach orbit on it's own power, it requires a launch vehicle which consists of an External Fuel Tank and two SRB's. The SRB's are recovered and the Ext. Fuel Tank is discarded, so it is not a RLV.

Offline Jim

  • Night Gator
  • Senior Member
  • *****
  • Posts: 32000
  • Cape Canaveral Spaceport
  • Liked: 10622
  • Likes Given: 317
Re: Falcon and Dragon reusability
« Reply #18 on: 01/07/2009 12:56 AM »

Compare, for instance, an SR-71 to a fictional intercontinental reconnaissance missile you used once. You could say "Ah but Corona was such". Every orbit of a spy sat is like a new flight sortie, the comparison between spysats and spyplanes is on of automation vs humans, not reusables vs single missions. Take an MX missile that would do a spy mission. You'd need thousands to fly the same lifetime that the SR-71 did, much more expensively.

Bad analogy as usual. 

The SR-71 was not able to fly over denied territory. It never went over the USSR.   An  MX lobbed payload could
« Last Edit: 01/07/2009 12:57 AM by Jim »

Offline pippin

  • Regular
  • Senior Member
  • *****
  • Posts: 2553
  • Liked: 289
  • Likes Given: 39
Re: Falcon and Dragon reusability
« Reply #19 on: 01/07/2009 01:33 AM »
Err, mlorrey, sorry, but you have no clou on statistics.
If you design a vehicle based on Murphy's law you know your single unit will be a failure so no need to proceed anyway.

Tags: