SpaceX Falcon 9 - AMOS-6 - (Pad Failure) - DISCUSSION THREAD (2)

[cscott]

S1 is bigger.  That alone could shift a resonance frequency, assuming that resonance-pumped cooling was the mechanism that pushed the LOX to the freezing point.

[hans_ober]

Also the position of the the Helium COPVs. They could be positioned differently in S1, in such a way that they get filled earlier or later than S2, and if the timings differ, they will experience different thermal conditions. And yeah, as mentioned, the size is a factor.

[Rocket Science]

Also the position of the the Helium COPVs. They could be positioned differently in S1, in such a way that they get filled earlier or later than S2, and if the timings differ, they will experience different thermal conditions. And yeah, as mentioned, the size is a factor.

IIRC I believe the size of the COPVs are the same, they just vary in the number of them installed in the LOX tanks...

[WHAP]

F9US "scares" right now for its performance trades. Centaur did not anticipate a "pusher" escape system for a capsule, as OSP's Boeing capsule then expected a puller. So now it needs a skirt.

Huh?  What does the skirt have to do with pusher vs. puller?  Maybe you can answer here https://forum.nasaspaceflight.com/index.php?topic=41440.0.

[Herb Schaltegger]

Also the position of the the Helium COPVs. They could be positioned differently in S1, in such a way that they get filled earlier or later than S2, and if the timings differ, they will experience different thermal conditions. And yeah, as mentioned, the size is a factor.

IIRC I believe the size of the COPVs are the same, they just vary in the number of them installed in the LOX tanks...

I think "relative size" might be a determinant; that is to say, the volume of the COPVs as compared to the size of the LOX tank in which they are immersed. That's the sort of thing that would play into an analysis of any potential resonance in the system during fill/drain operations.

[alang]

All sorts of questions:
Does oxygen in supposed composite voids have to be solid for ignition under pressure. Could contraction/expansion during loaded seal voids that had previously allowed seepage ?
Are the properties of oxygen crystals understood - any triboluminesce or piezoelectric effects possible? May not just be simple mechanical effects.

[Fred Bonyea]

Having read through the posts, I don't think anybody has yet clearly and explicitly stated in any detail what the theory or theories is/are here.  There seem to be two being discussed here in the context of Elon's reported comments at the NRO: 1. direct ignition, and 2. indirect ignition via COPV compromise.

Also, some people seem to be making comments without stating, as context, which of those two theories they are discussing.

For theory 1: Is there data explaining how solid oxygen compressed to high pressure in contact with carbon fibre composite might directly cause ignition at these low temperatures?  Even at hundreds of atmospheres oxygen is at < 100K when solid, right?  Does ignition really take place at those temperatures?  Is the idea that there would be local heating, before that causes melting of the oxygen?

'temperature' is the average kinetic energy within a given volume of gas or liquid. Gases will diffuse into micro-cracks and voids in an otherwise smooth surface. This 'breathing' is normal for virtually all materials; but if there is gas trapped behind a plug, and the gas is placed under more pressure, the kinetic energy per unit of volume can increase very rapidly. Suppose a fiber bottle has a void has a 3nm^2 inlet to a 3mm^3 void volume. Now plug that orifice with solid O2 and add pressure to the inner chamber, both increasing the diameter of bottle and the temperature within. The small void could collapse, boiling the LOX and creating a micro-cavity environment well above the combustion temperature of O2 and carbon fibers.  Since all of the surface organics are in a saturated O2 environment, there would be immediate unquenched burning.

For theory 2: In the case of the theory that solid oxygen caused compressive damage to fibres and subsequent COPV failure, how would a burst COPV cause ignition?  Frictional heating?  Again, data?

I think this is a more likely scenario, but one that is more difficult to resolve.  The COPV is at such a high pressure that this rapid release of high pressure helium into the supercooled O2 would raise the pressure within the O2 vessel so rapidly that it would burst, fragmenting the case and send shredding, sparking shrapnel through the common wall with the fuel tank.

[Rocket Science]

Also the position of the the Helium COPVs. They could be positioned differently in S1, in such a way that they get filled earlier or later than S2, and if the timings differ, they will experience different thermal conditions. And yeah, as mentioned, the size is a factor.

IIRC I believe the size of the COPVs are the same, they just vary in the number of them installed in the LOX tanks...

I think "relative size" might be a determinant; that is to say, the volume of the COPVs as compared to the size of the LOX tank in which they are immersed. That's the sort of thing that would play into an analysis of any potential resonance in the system during fill/drain operations.

Quite possibly, I would think that if LOX were tanked first it act as a "dampener" for any resonance to cover the COPVs, then fill the He. Now off the top of my head I'm not sure of slow fill/fast fill in the sequence for LOX...

[JamesH65]

Well very sad to hear my initial thinking was correct. Material failure due to intrusion/delam as a result of operational errors, more likely design failure due to overstress as a result of operational mistakes.


Basically the worst case for spacex this justifies most of the criticisms recently levied against them specifically that their lower cost access approach is resulting in unintended higher risks to payloads and flight rationale.

That said it's entirely fixable, but it will be harder than just changing a few parts out. Gotta change the methodology.

Nothing here is correct until SpaceX confirms it. Since all SpaceX have said is that they have some suspicions and are narrowing things down, it's much too early to be claiming anything, especially a design fault (since they have over 20 successful launches that did not show the issue and the design is the same)

[garidan]

The first problem was in flight, a bad strut on S2: in that case it could have been S2 has COPV filled with helium longer than S1, where it is obviously used as soon as from take off.

This one was during filling. It could be S1 needs to load much more LOX, so COPVs get submerged sooner than on S2. And on S2 lox could be colder from the beginning because recirculation happens quicker, beeing it shorter.

In any case, it could just be bad luck and it could have happened to S1 too. Only data can tell ...

[gospacex]

Well very sad to hear my initial thinking was correct. Material failure due to intrusion/delam as a result of operational errors, more likely design failure due to overstress as a result of operational mistakes.


Basically the worst case for spacex this justifies most of the criticisms recently levied against them specifically that their lower cost access approach is resulting in unintended higher risks to payloads and flight rationale.

Higher risk was not unintended.
SpaceX was trying to do things new way while "Old Space" was happy to not evolve.
When you do new things, you inevitably have more risk.

You try to guess where risks are and avoid making errors, but this is not a foolproof process.

SpaceX dodged a few bullets before. For example, they did not anticipate swirling of exhaust in tube channel Merlin nozzles, which made first launched F9 rotate. Luckily for them this rotation was rather small and did not affect the mission, but they did miss it.

[gospacex]

I'm surprised to hear that LOX is so difficult to protect against seeping into composites. I would think a relatively thin metal foil sheath on the outside should do it.

Oh goody, more layers...

Why "more layers" would be a big problem?

First you manufacture a COPV as usual.
Then you wrap it in a suitable metal foil with desired characteristics (zero porosity, LOX compat, needed CTE, corrugations if necessary to accommodate expansion, etc). Then you run the final 1-2 layers of fiber over it to keep it in place and some handling protection, this time not carbon fiber optimized for tensile loads, but a fiber designed for LOX compat.

[Stranger]

In MsGregor when the engine is burning the second stage of the test was carried out refueling tanks for sequence diagram as in Canaveral?

[mfck]

I'm surprised to hear that LOX is so difficult to protect against seeping into composites. I would think a relatively thin metal foil sheath on the outside should do it.

Oh goody, more layers...

Why "more layers" would be a big problem?

First you manufacture a COPV as usual.
Then you wrap it in a suitable metal foil with desired characteristics (zero porosity, LOX compat, needed CTE, corrugations if necessary to accommodate expansion, etc). Then you run the final 1-2 layers of fiber over it to keep it in place and some handling protection, this time not carbon fiber optimized for tensile loads, but a fiber designed for LOX compat.

That's not the industry trend though. It's been like metal vessel with metal overwrap --> solid metal --> composite o/w with metal liner --> composite o/w no-liner with excursions into plastics and ceramics. The trend seems to be generally dictated by material science.

Layers are problematic, because of the different CTEs, cryo temps and high pressure, a problem one cannot mitigate by just adding more layers, apparently. Maybe future advances in material science will allow cryo compatible foils with zero porosity and tunable CTE, but right now the search is about the matrix polymers. Corrugation does not seem be compatible with high pressures and reuse. With their design of submerged COPVs SX are trying to augment  for the material deficiencies, requiring titanium liner, by reducing one of the stress factors - the dT, which minimizes the CTE related stress while trying to save mass and, most probably, cost. Fancy layering is not on their roadmap, imho.

[Fred Bonyea]

...

Nothing here is correct until SpaceX confirms it. Since all SpaceX have said is that they have some suspicions and are narrowing things down, it's much too early to be claiming anything, especially a design fault (since they have over 20 successful launches that did not show the issue and the design is the same)

The first problem was in flight, a bad strut on S2: in that case it could have been S2 has COPV filled with helium longer than S1, where it is obviously used as soon as from take off.

This one was during filling. It could be S1 needs to load much more LOX, so COPVs get submerged sooner than on S2. And on S2 lox could be colder from the beginning because recirculation happens quicker, beeing it shorter.

In any case, it could just be bad luck and it could have happened to S1 too. Only data can tell ...

The best case scenario might be if the problem is thermal resonance and it can be duplicated.

If freezing O2 in composite voids is implicated; this could be a very difficult failure mode to duplicate and may require deliberately building bottles with high void volumes and cycle them through LOX conditioning cycles - a very expensive and slow process.

If none of the fault tree failure modes can be duplicated, painful choices will have to be made. It is possible, as others have suggested, that the bottle was damaged during handling and no one reported this. This would be very difficult to ascertain after-the-fact. There is no readily apparent resolution of this failure mode.

I was involved for several years in developing tests that could hopefully detect manufacturing flaws and/or fiber damage in composite systems. A lot of techniques that looked promising (flash thermography, eddy current field analysis, PCA of acoustic signatures) rarely work well on more than a thin layer of filaments in field testing.

Given their reliance upon composites, it would be interesting to know how much SpaceX invests in composite test method research.  I suspect they will be investing more. 

[CameronD]

I think "relative size" might be a determinant; that is to say, the volume of the COPVs as compared to the size of the LOX tank in which they are immersed. That's the sort of thing that would play into an analysis of any potential resonance in the system during fill/drain operations.

Quite possibly, I would think that if LOX were tanked first it act as a "dampener" for any resonance to cover the COPVs, then fill the He. Now off the top of my head I'm not sure of slow fill/fast fill in the sequence for LOX...

If it helps, I posted it again a few pages back:
https://forum.nasaspaceflight.com/index.php?topic=41252.msg1598684#msg1598684

To re-shape their procedures I imagine they'd need a stage or 2 to practice on.  Does anyone know if there just happens to be one set up at McGregor right now?

[dorkmo]

I'm surprised to hear that LOX is so difficult to protect against seeping into composites. I would think a relatively thin metal foil sheath on the outside should do it.

Oh goody, more layers...

Why "more layers" would be a big problem?

First you manufacture a COPV as usual.
Then you wrap it in a suitable metal foil with desired characteristics (zero porosity, LOX compat, needed CTE, corrugations if necessary to accommodate expansion, etc). Then you run the final 1-2 layers of fiber over it to keep it in place and some handling protection, this time not carbon fiber optimized for tensile loads, but a fiber designed for LOX compat.

That's not the industry trend though. It's been like metal vessel with metal overwrap --> solid metal --> composite o/w with metal liner --> composite o/w no-liner with excursions into plastics and ceramics. The trend seems to be generally dictated by material science.

Layers are problematic, because of the different CTEs, cryo temps and high pressure, a problem one cannot mitigate by just adding more layers, apparently. Maybe future advances in material science will allow cryo compatible foils with zero porosity and tunable CTE, but right now the search is about the matrix polymers. Corrugation does not seem be compatible with high pressures and reuse. With their design of submerged COPVs SX are trying to augment  for the material deficiencies, requiring titanium liner, by reducing one of the stress factors - the dT, which minimizes the CTE related stress while trying to save mass and, most probably, cost. Fancy layering is not on their roadmap, imho.

could they put the whole COPV inside of another aluminum cylinder with some nitrogen in the space between?

[CameronD]

could they put the whole COPV inside of another aluminum cylinder with some nitrogen in the space between?

To do this, you'd not only be increasing weight (the cylinder) but have issues with maintaining nitrogen supply and associated plumbing and then there's the possibility of the nitrogen freezing..

A vacuum bottle with an internal oxygen alarm sensor might be a better idea.

[RDoc]

could they put the whole COPV inside of another aluminum cylinder with some nitrogen in the space between?

To do this, you'd not only be increasing weight (the cylinder) but have issues with maintaining nitrogen supply and associated plumbing and then there's the possibility of the nitrogen freezing..

Nitrogen has a lower freezing point than Oxygen.

[CameronD]

Nitrogen has a lower freezing point than Oxygen.

-210degC is lower than -218degC?  Really??