SpaceX Falcon 9 FT - ORBCOMM-2 - Dec. 21, 2015 (Return To Flight) DISCUSSION

[matthewkantar]

It has amazing cryo properties.
As to flammability, do you realize that SpaceX is already embedding carbon fiber inside LOX tank?
Lot of mitigation factors can be considered.

Where?

I assume the reference is to the COPVs located in the oxidizer tanks.

Matthew

[joek]

SO why was he allowed to attempt a landing and SX is not?

You're comparing apples and oranges. SpaceX already had similar permission for their Grasshopper testbed program in Texas and was planning to do higher altitude hops similar to Blue's at Spaceport America, but their plans changed.

RTLS at a Government launch site is an entirely different situation, where the booster is flying back from hundreds of miles downrange with a huge radius of potential impact points if its guidance system were to go astray. New Shepard and Grasshopper were going straight up with limited propellants and thus limited ability to stray from their privately-owned test sites. And in Blue's case it helps if you own 290,000 acres of empty scrubland. There's not much out there to hit if things go wrong. Contrast that with Cape Canaveral/KSC/Cocoa Beach, where there are plenty of buildings and people relatively near the landing site. The risk equation is much different.

Some insight into the differences can be seen from the FAA maximum probable loss (MPL) and insurance requirements.  Some examples (excludes pre-flight):
- Blue Origin (EP 14-009): $6M liability
- SpaceX ORB-1 and ORB-2 (LLS 14-089): $26.25M liability + $100M government property.
- SpaceX CRS (LLS 14-087): $45M liability + $100M government property. [1]

There is no mention of RTLS in any SpaceX licenses, although a "flight" is defined as "commencing with ignition" and concluded with the "last exercise of control" over the vehicle, "including safing" (standard boilerplate in all licenses and permits). RTLS supplementary requirements, if any, are unknown.  (The FAA is often behind in publishing updates.)


[1] Does not include Dragon reentry which is licensed separately; $33M liability (unpublished).

[Craig_VG]

"SpaceX launch date for return to flight now 19 Dec.- proximate cause a payload issue. AF approves booster return to land, Optimism FAA ditto"

https://twitter.com/TheLurioReport/status/673270563742289920

[mme]

Here is an overview I created of the KSC / CCAFS assets in a satellite view.  Grey markers are inactive/demolished.  Blue markers are active, or at least leased.  Most information came from Wikipedia.  Feel free to PM me with suggested corrections, additions, links that should be included in details, etc.  Especially for the few buildings scattered between the Air Force Station and the pads.  I think one of them is an armory, so that could be important.

I ran out of attention, but plan to add more pads and assets.

http://tiny.cc/ksc-ccafs-assets-map

NOTE: The one mile polygon around LC-13 was just to help me with some sense of scale, it is in no way an official criteria for anything (accuracy, blast, etc.)

[The Roadie]

So far - one guy tweeted with unnamed sources. Not very compelling. Yet.

[obi-wan]

So far - one guy tweeted with unnamed sources. Not very compelling. Yet.

Charles Lurio has an extensive network of contacts throughout the field, and he's very careful about checking alternate sources. He can (like anyone) be wrong, but I wouldn't bet against him.

[ericspittle]

I don't know how reliable of a source this is, but:

https://twitter.com/TheLurioReport/status/673270563742289920

"SpaceX launch date for return to flight now 19 Dec.- proximate cause a payload issue. AF approves booster return to land, Optimism FAA ditto"

[sigil]

Here is an overview I created of the KSC / CCAFS assets in a satellite view.  Grey markers are inactive/demolished.  Blue markers are active, or at least leased.  Most information came from Wikipedia.  Feel free to PM me with suggested corrections, additions, links that should be included in details, etc.  Especially for the few buildings scattered between the Air Force Station and the pads.  I think one of them is an armory, so that could be important.

I ran out of attention, but plan to add more pads and assets.

http://tiny.cc/ksc-ccafs-assets-map

NOTE: The one mile polygon around LC-13 was just to help me with some sense of scale, it is in no way an official criteria for anything (accuracy, blast, etc.)

Ah thank you! I love this. I've been looking for a good maps of KSC & CCAFS. I'd love to see it extended to show logistical and integration support buildings and areas too.

[Comga]

Here is an overview I created of the KSC / CCAFS assets in a satellite view.  Grey markers are inactive/demolished.  Blue markers are active, or at least leased.  Most information came from Wikipedia.  Feel free to PM me with suggested corrections, additions, links that should be included in details, etc.  Especially for the few buildings scattered between the Air Force Station and the pads.  I think one of them is an armory, so that could be important.

I ran out of attention, but plan to add more pads and assets.

http://tiny.cc/ksc-ccafs-assets-map

NOTE: The one mile polygon around LC-13 was just to help me with some sense of scale, it is in no way an official criteria for anything (accuracy, blast, etc.)

Nice map, but if you are going to mark LC-39, which is way up north, you should mark the CCAFS buildings, which are much closer to the west.
There is one remote building with parked cars WSW of X1.  How far is that?

[The Roadie]

I've got SpaceX friends also. Doesn't mean I'd burn them posting a confidence.

[Karloss12]

One last thing wrt land landing with this RTF.

The last attempt was primarily foiled by a stuck valve causing the stage to overshoot, it tried to compensate...etc, boom. It's fair to say they have mitigated that potential issue as they have with Hydraulics, etc. But that valve response was off by all of 2 seconds or less. Landing this stage is an epic battle of precision. A battle they have not yet won. A battle, if I was the Range, would prefer to keep at sea until it was.

SpaceX have a good track record of making sea craters within 10 metres of precise pre-defined locations.

So as long as the land pad is a safe distance away from buildings and people than what is wrong with SpaceX making land craters at this precise pre-defined location?

[Jim]

Here is an overview I created of the KSC / CCAFS assets in a satellite view.  Grey markers are inactive/demolished.  Blue markers are active, or at least leased.  Most information came from Wikipedia.  Feel free to PM me with suggested corrections, additions, links that should be included in details, etc.  Especially for the few buildings scattered between the Air Force Station and the pads.  I think one of them is an armory, so that could be important.

I ran out of attention, but plan to add more pads and assets.

http://tiny.cc/ksc-ccafs-assets-map

NOTE: The one mile polygon around LC-13 was just to help me with some sense of scale, it is in no way an official criteria for anything (accuracy, blast, etc.)

http://www.oocities.org/rbt_kraisee/Map.html
http://www.robsv.com/cape/Images/CCASmap.jpg

[Coastal Ron]

Landing this stage is an epic battle of precision. A battle they have not yet won.

Well, they haven't won it at sea yet.  But they have demonstrated it on land numerous times with test vehicles.

I see a lot of analogies for what they are doing as when a pilot lands an aircraft.  Ideally when landing an airplane you want to have a stabilized flight path when landing, where only small corrections need to be made.  If you have a big landing area you're shooting for (i.e. a Cessna 172 landing on a runway built for large jets) you don't have critical corrections you have to make.  But if you're an F-18 pilot coming in on an aircraft carrier at night in stormy weather, then if you need to nail that first landing you are making continuous fine & large adjustments all the way down.

To me landing on the barge is like landing on the aircraft carrier, and landing on land is like landing at the large commercial airport.

I don't have too much concern about the land landing, assuming they have the fuel to get back.

[the_other_Doug]

Landing this stage is an epic battle of precision. A battle they have not yet won.

Well, they haven't won it at sea yet.  But they have demonstrated it on land numerous times with test vehicles.

I see a lot of analogies for what they are doing as when a pilot lands an aircraft.  Ideally when landing an airplane you want to have a stabilized flight path when landing, where only small corrections need to be made.  If you have a big landing area you're shooting for (i.e. a Cessna 172 landing on a runway built for large jets) you don't have critical corrections you have to make.  But if you're an F-18 pilot coming in on an aircraft carrier at night in stormy weather, then if you need to nail that first landing you are making continuous fine & large adjustments all the way down.

To me landing on the barge is like landing on the aircraft carrier, and landing on land is like landing at the large commercial airport.

I don't have too much concern about the land landing, assuming they have the fuel to get back.

I will take that analogy further -- this is more like the difference between trying to land a new carrier jet on the carrier deck, and landing it for the first time in initial trials on the dry lake bed at Edwards AFB.  You have tens of meters of dispersion absorption capability with the RTLS without endangering anything (or anyone) and thus have an enhanced capability of getting the thing onto the ground in one piece.

(And yes, I understand that a carrier jet is Navy and wouldn't necessarily have its initial trials at Edwards -- but you know what I mean, I'm just trying to make a point.  Forgive the logical inconsistency, please.  )

[vulture4]

Regarding the struts, the two mitigation steps I saw discussed were increasing the strength of the strut and possibly doing 100% testing. However the strut already had a more than adequate 5:1 load factor. The failure occurred because of poor control of the manufacturing process, not a design flaw. An unanticipated falure mode in the design would have been present in every strut and would in almost all cases have failed within the first few flights. Inconsistent manufacturing is best detected by selecting periodic samples and testing to ultimate strength. This cannot be done with 100% of the parts, nor is this necessary.As C. Edwards Deming pointed out years ago, if manufacturing process is under adequate control a quantitative parameter such as strength at failure in a sample of parts consistently falls into a normal distribution with a standard deviation that is only a fraction of the specified variability. If even a few measurements fall outside of the normal distribution then the process is not under control. Testing doesn't correct the problem. The corrective action is to identify the source of the variability in quality and correct the manufacturing process.  Of course NDT of various types can be done with 100% of output, but this doesn't provide a quantitative measure of the critical property, in this case strength.

[meekGee]

Regarding the struts, the two mitigation steps I saw discussed were increasing the strength of the strut and possibly doing 100% testing. However the strut already had a more than adequate 5:1 load factor. The failure occurred because of poor control of the manufacturing process, not a design flaw. An unanticipated falure mode in the design would have been present in every strut and would in almost all cases have failed within the first few flights. Inconsistent manufacturing is best detected by selecting periodic samples and testing to ultimate strength. This cannot be done with 100% of the parts, nor is this necessary.As C. Edwards Deming pointed out years ago, if manufacturing process is under adequate control a quantitative parameter such as strength at failure in a sample of parts consistently falls into a normal distribution with a standard deviation that is only a fraction of the specified variability. If even a few measurements fall outside of the normal distribution then the process is not under control. Testing doesn't correct the problem. The corrective action is to identify the source of the variability in quality and correct the manufacturing process.  Of course NDT of various types can be done with 100% of output, but this doesn't provide a quantitative measure of the critical property, in this case strength.

Normal distribution of faults only works under some failure mechanisms, and additionally requires large numbers.

If you're making 10,000,000 TV screen, and you have a fault rate of .01%, so are dealing with 1000 fault sets, then yes.

If you'e making 1000 struts, and the loss of a single one means a loss of a rocket,  it's a different ballgame.

Also, some failures are a result of a hidden normally-distributed random variable, which creates a failure when above a certain threshold.  That's predictable.

But if the random variable has a simple binary distribution, it can be a lot harder to predict, if at all possible.

SpaceX ended up testing to failure thousands of parts.  I wonder what they found out about those issues.  We there a wide distribution of fault levels, or were they all good except for rarely occurring completely bum parts.

[Coastal Ron]

Regarding the struts, the two mitigation steps I saw discussed were increasing the strength of the strut and possibly doing 100% testing. However the strut already had a more than adequate 5:1 load factor. The failure occurred because of poor control of the manufacturing process, not a design flaw.

As designed, and if manufactured properly, the strut would not have failed.  There were two failures outside of SpaceX that lead to the Falcon 9 failure:

1.  There was a defect in the manufacture of that particular strut.  We don't know why, or what though.

2.  SpaceX stated that they required certification of each strut.  Obviously the supplier processes that were supposed to catch defects failed, but we don't know if the process was inadequate (which SpaceX should be partially responsible for) or if the processes that were in place were ignored or missed.

Regardless, SpaceX has stated that they are no longer using that same supplier, and that they will 100% inspect the part in-house to ensure minimum strength.

Inconsistent manufacturing is best detected by selecting periodic samples and testing to ultimate strength. This cannot be done with 100% of the parts, nor is this necessary.As C. Edwards Deming pointed out years ago...

For some applications, where failure is not an option, 100% inspection and testing can and is done.  We used to do that for government contracts for electronic systems all the time.  Even in consumer electronics 100% testing is done, although maybe not enough to get past infant mortality.  So it's more a matter of when the testing is done, and by whom.

Testing doesn't correct the problem. The corrective action is to identify the source of the variability in quality and correct the manufacturing process.

Some testing of components is done as part of the customers manufacturing process.  If you're building smart phones, then you power up main computer board after the assembly process to see if it responds to the proper inputs.  However there is only so much testing you can do with a rocket, and if the G-forces that would cause a failure can't be duplicated during ground testing then your only alternative is to test the component parts themselves to validate their strength.

[Jim]

Regarding the struts, the two mitigation steps I saw discussed were increasing the strength of the strut and possibly doing 100% testing. However the strut already had a more than adequate 5:1 load factor. The failure occurred because of poor control of the manufacturing process, not a design flaw.

Wrong, it is a design flaw.  Wrong, material and manufacturing technique for the application

[cscott]

Jim might be overstating the case.  But SpaceX did change the strut material and manufacturing technique to avoid the potential of a similar flaw in the future.

[vulture4]

"For some applications, where failure is not an option, 100% inspection and testing can and is done."

Nondestructive methods can be applied to 100% of output, but the fact that a part can be tested and is functional doesn't mean it will work more than a few minutes before it fails. For electronics, failure rates are typically constant. Each part will have an ultimate strength, and the ultimate strength of parts resulting from a well controlled process will necessarily be normally distributed. A consistent normal distribution means the process is under control. Even a small percentage of parts outside the normal range (the control limits in classical QC) means that the process is not adequately controlled. Even if every part meets spec on a 100% nondestructive test, that does not mean the manufacturing process is under control, since the NDT is only a pass-fail criterion and does not establish whether material properties are normally distributed! The process is out of control, and some parts may have flaws that will lead to unanticipated failure, particularly in a reusable LV which may undergo hundreds of launch cycles.

Mechanical parts made of steel may last indefinitely if they are loaded below the fatigue limit, but aluminum generally has no fatigue limit, i.e. it is subject to cyclic failure regardless of stress level if enough cycles are applied. Time to failure is inversely related to cyclic stress level, so with heavily loaded aluminum alloy skins on many pressurized aircraft there is no choice but to periodically check the entire structure for cracks. If it flies long enough they will be there.

I don't claim to be an expert, or even competent in this field, but I know that just requiring higher safety factors or 100% nondestructive testing to an acceptance spec is to miss the intricate beauty of this fascinating field, which links statistics and engineering in a way that is concrete and quantitative, as opposed to the pulled-out-of-the-air red-yellow-and-green "estimates" of failure rate often found in NASA documents, or the designation of a person who watches others do a task and makes a check on a clipboard as "quality control".

End of rant. Apologies to all.