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

[woods170]

I think it merits pointing out something posted earlier on this thread.

Buckling of the mettalic liner can occur during the depressurization phase of the proof test because it has been plastically stressed during the pressurization phase. Therefore it is in a state of residual compression. Local progressive debonding between the composite overwrap and the metallic liner may then occur, allowing the liner to buckle inward.

https://shellbuckling.com/presentations/otherTopics/pages/page_180.html



I maintain that the point where the overwrap circles the polar boss is an ingress point for LOX if there is a buckle in the liner at this region. And the paper flatly states this is a proven failure mode that can happen after autofreggage.

I don't have access to the actual scholarly article but here's a link to it:

https://www.tib.eu/en/search/id/BLCP%3ACN085735249/Analysis-of-Debonding-of-Filament-Wound-Composite/

Yes, can happen. Not necessarily will happen. Fred's post stated that buckling is unavoidable. As in: will always happen. And that is not the case.
Trouble is: Fred is mixing up buckling and plastic yielding. The two things are different in the sense that buckling can result from plastic yielding.

[meekGee]

Buckling is not only elastic.  Elastic buckling conditions are simply the simplest case,, which is first studied in statics 101....

Especially here where the failure mechanism did not involve the progression of a buckling deformation into a large macro deformation, but rather created a condition for fluid seepage, phase change, bulk stress, and conditions for starting a chemical reaction...

So I am not sure exactly what they mean by the word buckling, and whether said buckling, without the SOX issue, is otherwise a problem.

[alang]

Dumb question: I understand from that much non destructive testing is possible in important engineering applications ultrasound, *boroscope and maybe even x-raying components.
Presumably this can't be easily be used during destructive testing designed to reproduce a problem, due to the conseqent destruction of expensive test equipment.
Is there any realistic way to do remotely do x-ray movies in real time of components undergoing destructive testing?
This assumes a desperate state of ignorance about a problem and a very urgent need to solve it.
Yes, I am being lazy, there is a lot to read on industrial radiography.

*Edited for stupidity.

[Fred Bonyea]

I think it merits pointing out something posted earlier on this thread.

Buckling of the mettalic liner can occur during the depressurization phase of the proof test because it has been plastically stressed during the pressurization phase. Therefore it is in a state of residual compression. Local progressive debonding between the composite overwrap and the metallic liner may then occur, allowing the liner to buckle inward.

https://shellbuckling.com/presentations/otherTopics/pages/page_180.html



I maintain that the point where the overwrap circles the polar boss is an ingress point for LOX if there is a buckle in the liner at this region. And the paper flatly states this is a proven failure mode that can happen after autofreggage.

I don't have access to the actual scholarly article but here's a link to it:

https://www.tib.eu/en/search/id/BLCP%3ACN085735249/Analysis-of-Debonding-of-Filament-Wound-Composite/

Yes, can happen. Not necessarily will happen. Fred's post stated that buckling is unavoidable. As in: will always happen. And that is not the case.
Trouble is: Fred is mixing up buckling and plastic yielding. The two things are different in the sense that buckling can result from plastic yielding.

Yes I was - fiber displacement occurs on many scales, and the term buckling seems to be reserved for gross, multilayer void formation, and that is not inevitable. On the small scale, there will be microvoids, warping and unfavorable stress planes. These are very difficult to detect and quantify in situ.
 It is my understanding that the suspected failure mode involves microstructure - small veins or cavities that saturate with LOX the same way a diver saturates their joints with N2. If the pressure changes too rapidly in the bottle as a whole, trapped voids saturated with O2 may undergo a diesel explosion....and there she went. 

[jpo234]

What about letting the helium pressure do the unbuckling and submerge the bottles when they are fully pressurized?

[Fred Bonyea]

Dumb question: I understand from that much non destructive testing is possible in important engineering applications ultrasound, horoscope and maybe even x-raying components.
Presumably this can't be easily be used during destructive testing designed to reproduce a problem, due to the conseqent destruction of expensive test equipment.
Is there any realistic way to do remotely do x-ray movies in real time of components undergoing destructive testing?
This assumes a desperate state of ignorance about a problem and a very urgent need to solve it.
Yes, I am being lazy, there is a lot to read on industrial radiography.

Not a dumb question. Nondestructive testing of composites is an extremely challenging field. X-ray is almost useless, as the materials are too low and uniform in density. It can be useful in the examination of composite/metal bosses or attachment point. 

I spent a good part of my career trying to develop reliable nondestructive analytical tests using scanners equipped with ultrasonic, eddy current, and a variety of spectral sensors, including single-sided NMR and flash themography.   

In flash thermography, you hit the opposite (or same) side of the item with a quick, intense infrared heat source, then take rapid scanning measurements of the heat energy emitted from the opposite side or reflected from the same side.  In pass-through mode, voids will develop as a dark patch (insulating) and a resin rich region may transfer heat more rapidly. (In reflective mode, it is just opposite, with heat sinking locations conducting more heat away from the surface than voids.) In my opinion, this was the most promising technique for detecting damaged case structure - although generally if damage showed up in testing it was also identifiable in visual inspections.

Likewise, ultrasonic measurements reduced using principle component analysis worked well when buckled sections were deliberately wound into the bottles, but failed to pick up small buckles and voids found during dissections.

I have been out of the industry for ten years, so I am sure the techniques have been improved upon; but since COPVs continue to fail on rare occasions and in unpredictable ways, there is work to be done.   

[Robotbeat]

What about letting the helium pressure do the unbuckling and submerge the bottles when they are fully pressurized?

I believe that's the procedure now. But fully pressurized doesn't mean full. If they then submerge the COPVs in LOx, the pressure will drop as the helium cools and they can keep adding more helium to compensate. From what I understand they were deeply subcooling the helium to speed this process up, perhaps even deeper than the LOx temperature. Very cold helium means you can fit more in the same COPV at the same temperature as well as fill faster as you can get a higher mass flow rate in the same tubes. That's why they've gone back to using 4 COPVs in the LOx tank instead of just 3.

[matthewkantar]

Dumb question: I understand from that much non destructive testing is possible in important engineering applications ultrasound, horoscope and maybe even x-raying components.

I hope they have not turned to using the Horoscope, not repeatable in my opinion.

Matthew

[Fred Bonyea]

What about letting the helium pressure do the unbuckling and submerge the bottles when they are fully pressurized?

If we are talking about buckled fiber bundles that are cured into the composite system; since they can relax slightly under pressure (as in the drawings), this puts more strain on the unbuckled regions, lower the burst capacity of the bottle. This failure mode is usually identified in proof testing.

Gases, or LOX trapped in buckled layers and other voids are suspected of being the root culprits in the Amos 6 pad failure. It sounds like they are (still) unavoidable evils; and why mitigating redesign efforts are underway.

[alang]

Dumb question: I understand from that much non destructive testing is possible in important engineering applications ultrasound, horoscope and maybe even x-raying components.

I hope they have not turned to using the Horoscope, not repeatable in my opinion.

Matthew

Oops - corrected. I remember laughing the other day when someone's incident report came through as an indecent report...

[Herb Schaltegger]

Buckling is not only elastic.  Elastic buckling conditions are simply the simplest case,, which is first studied in statics 101....

Especially here where the failure mechanism did not involve the progression of a buckling deformation into a large macro deformation, but rather created a condition for fluid seepage, phase change, bulk stress, and conditions for starting a chemical reaction...

So I am not sure exactly what they mean by the word buckling, and whether said buckling, without the SOX issue, is otherwise a problem.

This. This exactly. Thank you, meekGee, for phrasing this so well.

[TomH]

How will this issue affect long term reusability? Even if the plasticity is minimized, if a stage is re-flown scores of times, with each flight including static fire tests, the COPVs will wind up being pressurized and depressurized perhaps hundreds of times. I am reminded of the de Havilland Comet which suffered catastrophic failure due to pressurization metal fatigue. The company lost its lead in commercial jet production due to this single issue.

Even with the eventual COPV redesign, how will they be able totally to prevent any delamination/buckling over hundreds of periods of tanking/pressurizing-detanking/depressurizing? Will COPVs need to be moved outside the prop tanks so they can be more easily inspected and replaced? Can new materials fabrication eliminate the plasticity issue entirely, or at least enough that it will not compromise the vessel?

[gongora]

How will this issue affect long term reusability? Even if the plasticity is minimized, if a stage is re-flown scores of times, with each flight including static fire tests, the COPVs will wind up being pressurized and depressurized perhaps hundreds of times. I am reminded of the de Havilland Comet which suffered catastrophic failure due to pressurization metal fatigue. The company lost its lead in commercial jet production due to this single issue.

Even with the eventual COPV redesign, how will they be able totally to prevent any delimitation/buckling over hundreds of periods of tanking/pressurizing-detanking/depressurizing? Will COPVs need to be moved outside the prop tanks so they can be more easily inspected and replaced? Can new materials fabrication eliminate the plasticity issue entirely, or at least enough that it will not compromise the vessel?

The COPVs could be replaced after a certain number of cycles.  Part of figuring out the reuse for the boosters is finding out how long they can safely use different pieces.  The boosters will need maintenance from time to time.

[Kaputnik]

How will this issue affect long term reusability? Even if the plasticity is minimized, if a stage is re-flown scores of times, with each flight including static fire tests, the COPVs will wind up being pressurized and depressurized perhaps hundreds of times. I am reminded of the de Havilland Comet which suffered catastrophic failure due to pressurization metal fatigue. The company lost its lead in commercial jet production due to this single issue.

Even with the eventual COPV redesign, how will they be able totally to prevent any delimitation/buckling over hundreds of periods of tanking/pressurizing-detanking/depressurizing? Will COPVs need to be moved outside the prop tanks so they can be more easily inspected and replaced? Can new materials fabrication eliminate the plasticity issue entirely, or at least enough that it will not compromise the vessel?

The COPVs could be replaced after a certain number of cycles.  Part of figuring out the reuse for the boosters is finding out how long they can safely use different pieces.  The boosters will need maintenance from time to time.

Wouldn't that involve cutting the tanks open? Or do they fit out of the 'manhole'?

[envy887]

This was part of the reason they are subjecting the JCSAT-14 booster to 10 full-duration test fires. Everything gets cycled repeatedly, including COPVs.

[Darga]

How will this issue affect long term reusability? Even if the plasticity is minimized, if a stage is re-flown scores of times, with each flight including static fire tests, the COPVs will wind up being pressurized and depressurized perhaps hundreds of times. I am reminded of the de Havilland Comet which suffered catastrophic failure due to pressurization metal fatigue. The company lost its lead in commercial jet production due to this single issue.

Even with the eventual COPV redesign, how will they be able totally to prevent any delimitation/buckling over hundreds of periods of tanking/pressurizing-detanking/depressurizing? Will COPVs need to be moved outside the prop tanks so they can be more easily inspected and replaced? Can new materials fabrication eliminate the plasticity issue entirely, or at least enough that it will not compromise the vessel?

The COPVs could be replaced after a certain number of cycles.  Part of figuring out the reuse for the boosters is finding out how long they can safely use different pieces.  The boosters will need maintenance from time to time.

Wouldn't that involve cutting the tanks open? Or do they fit out of the 'manhole'?

They can fit out of the hole.

[alang]

Dumb question: I understand from that much non destructive testing is possible in important engineering applications ultrasound, horoscope and maybe even x-raying components.
Presumably this can't be easily be used during destructive testing designed to reproduce a problem, due to the conseqent destruction of expensive test equipment.
Is there any realistic way to do remotely do x-ray movies in real time of components undergoing destructive testing?
This assumes a desperate state of ignorance about a problem and a very urgent need to solve it.
Yes, I am being lazy, there is a lot to read on industrial radiography.

Not a dumb question. Nondestructive testing of composites is an extremely challenging field. X-ray is almost useless, as the materials are too low and uniform in density. It can be useful in the examination of composite/metal bosses or attachment point. 

I spent a good part of my career trying to develop reliable nondestructive analytical tests using scanners equipped with ultrasonic, eddy current, and a variety of spectral sensors, including single-sided NMR and flash themography.   

In flash thermography, you hit the opposite (or same) side of the item with a quick, intense infrared heat source, then take rapid scanning measurements of the heat energy emitted from the opposite side or reflected from the same side.  In pass-through mode, voids will develop as a dark patch (insulating) and a resin rich region may transfer heat more rapidly. (In reflective mode, it is just opposite, with heat sinking locations conducting more heat away from the surface than voids.) In my opinion, this was the most promising technique for detecting damaged case structure - although generally if damage showed up in testing it was also identifiable in visual inspections.

Likewise, ultrasonic measurements reduced using principle component analysis worked well when buckled sections were deliberately wound into the bottles, but failed to pick up small buckles and voids found during dissections.

I have been out of the industry for ten years, so I am sure the techniques have been improved upon; but since COPVs continue to fail on rare occasions and in unpredictable ways, there is work to be done.   

Thanks for the very detailed answer. The point of my question is that I get very uneasy when very clever people who know a lot of theory and are good at modelling don't have much of an empirical foundation. Sometimes there is a temptation to build a castle in the air based on foundations of, well, air.
I'm not saying that is what is happening here because I don't have the background but I have experienced the psychology of problem solving in other things.

[Reflectiv]

So I am not sure exactly what they mean by the word buckling, and whether said buckling, without the SOX issue, is otherwise a problem.

From the anomaly update:
"When pressurized, oxygen pooled in this buckle can become trapped; in turn, breaking fibers or friction can ignite the oxygen in the overwrap, causing the COPV to fail. In addition, investigators determined that the loading temperature of the helium was cold enough to create solid oxygen (SOX), which exacerbates the possibility ..."

That sounds to me like the SOX is not necessary?

[meekGee]

So I am not sure exactly what they mean by the word buckling, and whether said buckling, without the SOX issue, is otherwise a problem.

From the anomaly update:
"When pressurized, oxygen pooled in this buckle can become trapped; in turn, breaking fibers or friction can ignite the oxygen in the overwrap, causing the COPV to fail. In addition, investigators determined that the loading temperature of the helium was cold enough to create solid oxygen (SOX), which exacerbates the possibility ..."

That sounds to me like the SOX is not necessary?

Agreed, but it seems a pretty significant detail, which reinforces "I don't understand the explanation"...

[Patchouli]

So I am not sure exactly what they mean by the word buckling, and whether said buckling, without the SOX issue, is otherwise a problem.

From the anomaly update:
"When pressurized, oxygen pooled in this buckle can become trapped; in turn, breaking fibers or friction can ignite the oxygen in the overwrap, causing the COPV to fail. In addition, investigators determined that the loading temperature of the helium was cold enough to create solid oxygen (SOX), which exacerbates the possibility ..."

That sounds to me like the SOX is not necessary?

Possible though I wonder could a covering of a material similar to Xcor's Nonburnite along with changed loading procedures be used to prevent LOX infiltration of the composite overwrap?

Or is Nonburnite strong enough to completely replace the the carbon overwrap and eliminate the failure mode or could fiberglass be used instead of carbon fiber?