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

[Jim]

The ICE might be a lot easier to control. Or perhaps leave it pressed, what's the reason for depressing on coasts?

To prevent venting.

[Jim]

Depends - are they on a reusable stage? If not, then they probably are depressed...   

has no bearing on the matter

[envy887]

The ICE might be a lot easier to control. Or perhaps leave it pressed, what's the reason for depressing on coasts?

To prevent venting.

The reason they would have to vent during coast is excessive pressure from boiloff,  correct?

Isn't boiloff predictable enough to do a partial depress and time the boiloff pressure buildup to reach flight pressure right before restarting for the injection burn?

[Firehawk153]

Depends - are they on a reusable stage? If not, then they probably are depressed...   

has no bearing on the matter

Either Jim didn't get it (unlikely) or he didn't think it was funny (more likely).  FWIW it gave me a good chuckle.

[Oersted]

Does anyone have an idea what size buckling we are talking about? Fractions of millimeters? Millimeters? Centimeters?

No takers?

[Fred Bonyea]

To answer a few questions posed above, and flesh-out the argument over the last few days:

1) Buckling is inevitable. The expansion coefficient of the aluminum shell is quite high, and the linear expansion coefficient of carbon fibers is near zero. There is no way a thin aluminum bottle be wrapped with carbon fibers be not buckle, unless they have sandwiched a foam filler in-between, which they have not. It is also difficult to prevent fiber-wound bottles from buckling internally: To do so, the tension during winding must decrease with circumference in harmony with any viscosity changes in the resin - nobody can do that; and it must be a resin that does not shrink or expand during cure. I don't know where to find that, either. You can mitigate the buckling. (If you add more resin, but if the resin has a greater expansion coefficient than the fibers, you risk breaking the fibers anyway.)

2) The helium COPV is roughly the equivalent of a battery in an electronic circuit: It not only provides the pressure forcing LOX into the pre-combustion chamber, it also stabilizes the pressure of the O2 delivery system. You could heat the LOX by routing the lines near the exhaust gasses, but you must have initial pressure, and you must be able to meter the pressure during the burn: Gaseous He2 provides both the initial energy and required buffering.

3) Minuteman uses small titanium alloy helium tanks for nozzle vectoring. They are positioned near the nozzle. Larger versions would likely work well within a LOX environment, but qualification burns would be essential.

4) An entirely composite system will always result in tension between LOX and carbon-based structures. Fifty years ago, I watched a demonstration of a cigar soaked in LOX, and then ignited at one end. It  burned like a rocket, even after being quenched in sand.

5) Micro-voids formed during composite wrapping are unavoidable for the reasons I stated above. (I spent about ten years of my career investigating analytical methods that would reduce the risk of using composites.) Damage to composite materials during handling can be extremely difficult to  identify and track. One wrong step by a technician that is not acknowledged may be unidentifiable in any test matrix. It is a human factor that cannot be engineered around.

In my opinion, SpaceX's track record to-date demonstrates that they have excellent feedback mechanisms between their technical crews and management; but they must be very careful not to let budget concerns circumcise this dialogue.


 

[topo334]

One wonders how any circumcision would be accomplished.

[Robotbeat]

How is buckling "inevitable" except in the same sense as Akin's 25th Law? (a.k.a. Bowden's Law)

COPVs are in use all over, and SpaceX has been putting them in cryo tanks for years. They haven't all failed.

[Herb Schaltegger]

How is buckling "inevitable" except in the same sense as Akin's 25th Law? (a.k.a. Bowden's Law)

COPVs are in use all over, and SpaceX has been putting them in cryo tanks for years. They haven't all failed.

But SpaceX is the only organization using them filled with cold He while immersed in LOX cooled almost to the freezing point.

[Robotbeat]

How is buckling "inevitable" except in the same sense as Akin's 25th Law? (a.k.a. Bowden's Law)

COPVs are in use all over, and SpaceX has been putting them in cryo tanks for years. They haven't all failed.

But SpaceX is the only organization using them filled with cold He while immersed in LOX cooled almost to the freezing point.

That doesn't meet the definition of "inevitable," not by a long shot.

[Herb Schaltegger]

How is buckling "inevitable" except in the same sense as Akin's 25th Law? (a.k.a. Bowden's Law)

COPVs are in use all over, and SpaceX has been putting them in cryo tanks for years. They haven't all failed.

But SpaceX is the only organization using them filled with cold He while immersed in LOX cooled almost to the freezing point.

That doesn't meet the definition of "inevitable," not by a long shot.

Did you read Fred's post at all?

[envy887]

... The expansion coefficient of the aluminum shell is quite high, and the linear expansion coefficient of carbon fibers is near zero....  You can mitigate the buckling. (If you add more resin, but if the resin has a greater expansion coefficient than the fibers, you risk breaking the fibers anyway.)
...

By "expansion coefficient" do you mean coefficient of thermal expansion, or do you mean modulus? I can't see how thermal expansion could cause buckling in a cryo COPV liner, since buckling is a compression failure and the liner will be in tension... it always contracts much more than the wrap as the COPV is chilled. That contraction might cause voids, but not compressive buckles.

[Robotbeat]

How is buckling "inevitable" except in the same sense as Akin's 25th Law? (a.k.a. Bowden's Law)

COPVs are in use all over, and SpaceX has been putting them in cryo tanks for years. They haven't all failed.

But SpaceX is the only organization using them filled with cold He while immersed in LOX cooled almost to the freezing point.

That doesn't meet the definition of "inevitable," not by a long shot.

Did you read Fred's post at all?

Yes. And I'm well familiar with COPVs. I've made fixtures for winding COPVs. They didn't buckle.

It helps that you do autofrettage first.

And you CAN wind the COPV with varying tension.

Again, I don't see "inevitable" anywhere. I /can/ see how it could /possibly/ buckle but I fail to see how it's "inevitable."

[Roy_H]

Well, I hope this isn't too stupid as I have zero experience. But is the buckling caused by the tension of the carbon fiber during wrapping? If so could one prevent it by filling the tank with water before wrapping so the tank wouldn't compress easily?

[faramund]

I am completely naive about specifics of this - but I've done university level physics - FWTW. I've been reading this as - due to thermal expansion variation rates - buckles are inevitable (on at least a very small (microscopic) scale), and then under some conditions, those buckles can get so big that COPVs fail.

So I take spaceX's approach, as one that aims to minimise (but not eliminate) buckles. Presumably to some acceptable level.

If someone KNOWS that this is correct, I'm happy to be corrected.

[Pete]

I've made fixtures for winding COPVs. They didn't buckle.

It helps that you do autofrettage first.

Doing autofrettage absolutely **guarantees** buckling, that's what its purpose is.
To plastically deform the interior liner to match the exterior reinforcement.

It's any deformation that occurs later, due to stresses such as mismatching thermal expansions, or even the mere stress of releasing the autofrettage pressure load, that is of concern.
.
And very, *very* hard to accurately measure.

But saying you have worked with COPV's that didn't buckle at all, is like an engineer claiming that his bridge is perfectly rigid, and will never flex at all.

[jpo234]

IMHO the easiest solution would be to pressure the COPVs before submerging them into the LOX. The pressure would probably fix potential buckles. And in the unlikely case that there are still buckles present when the COPVs get submerged, there is no chance of trapping LOX because the liner is already expanded.

[Robotbeat]

Plastic yielding is not buckling.

[woods170]

How is buckling "inevitable" except in the same sense as Akin's 25th Law? (a.k.a. Bowden's Law)

COPVs are in use all over, and SpaceX has been putting them in cryo tanks for years. They haven't all failed.

But SpaceX is the only organization using them filled with cold He while immersed in LOX cooled almost to the freezing point.

That doesn't meet the definition of "inevitable," not by a long shot.

Did you read Fred's post at all?

Yes. And I'm well familiar with COPVs. I've made fixtures for winding COPVs. They didn't buckle.

It helps that you do autofrettage first.

And you CAN wind the COPV with varying tension.

Again, I don't see "inevitable" anywhere. I /can/ see how it could /possibly/ buckle but I fail to see how it's "inevitable."

There are several notable errors in Fred's post. Buckling for example is anything but unavoidable. The rest, well, you already touched those points.

[Johnnyhinbos]

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/