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

[AS-503]

I think darn near any metal thick enough to contain that kind of pressure without significant overwrap will be prone to brittle fracturing modes under the thermal conditions described...it's tough requirements for ANY material to stand.

No, see the Saturn V

So what were the helium tanks in Saturn V made of? What was the thickness/weight. I cant find it.

IIRC Titanium spheres in (and outside) the second and third stages, aluminum cylinders in the first stage.
Also, some liquid super-critical He in the LEM.

[Proponent]

So what were the helium tanks in Saturn V made of? What was the thickness/weight. I cant find it.

According to the carefully researched paper attached to this post (see p. 5), it was Al.

[mikelepage]

If I'm understanding the current leading theory for the failure, the problem (solid Oxygen in the COPV liner overwrap) is something that they would only expect to occur with sub-cooled LOX, and not with regular LOX.  The solution is to use slightly warmer LOX? (with associated performance penalty).

Not sure of the actual order of loading, nor complexity of the tanking apparatus, but would it be possible/advantageous to have two LOX sources on the pad at different temperatures, and pre-fill the tank with the warmer LOX to saturate the COPV liners overwrap, then fill the rest with sub-cooled LOX so the performance hit isn't as great?

EDIT wrong terminology - I see I should have said overwrap (composite), not liner (aluminium)

[jpo234]

If I'm understanding the current leading theory for the failure, the problem (solid Oxygen in the COPV liner) is something that they would only expect to occur with sub-cooled LOX, and not with regular LOX.  The solution is to use slightly warmer LOX? (with associated performance penalty).

Not sure of the actual order of loading, nor complexity of the tanking apparatus, but would it be possible/advantageous to have two LOX sources on the pad at different temperatures, and pre-fill the tank with the warmer LOX to saturate the COPV liners, then fill the rest with sub-cooled LOX so the performance hit isn't as great?

The LOX had to be liquid to get where it is causing trouble. To me this implies, that the COPV was colder than expected and caused the LOX to freeze. So the solution will probably be a procedural change that makes sure that the COPV doesn't cool below the freezing point of LOX.

[Jet Black]

If I'm understanding the current leading theory for the failure, the problem (solid Oxygen in the COPV liner) is something that they would only expect to occur with sub-cooled LOX, and not with regular LOX.  The solution is to use slightly warmer LOX? (with associated performance penalty).

Not sure of the actual order of loading, nor complexity of the tanking apparatus, but would it be possible/advantageous to have two LOX sources on the pad at different temperatures, and pre-fill the tank with the warmer LOX to saturate the COPV liners, then fill the rest with sub-cooled LOX so the performance hit isn't as great?

let's say that some expansion in the helium (PV/T=K) caused a drop in temperature, that could still cool the LOX in the COPV to the point that it freezes. When they were talking of odd harmonics, I wondered if it might be something like this. I still wonder whether it might be something like knocking a part filled beer bottle.

[Proponent]

let's say that some expansion in the helium (PV/T=K) caused a drop in temperature, that could still cool the LOX in the COPV to the point that it freezes.

Under the applicable conditions, helium warms up as it expands.

[jpo234]

let's say that some expansion in the helium (PV/T=K) caused a drop in temperature, that could still cool the LOX in the COPV to the point that it freezes.

Under the applicable conditions, helium warms up as it expands.

And yet, it seems, the LOX froze.

[rsdavis9]

There is both expansion and compression of helium going on NEAR the copv's.
At first helium expands into the copv's..
At some point the dominate is the compression of the helium to flight pressure.
This assumes the helium lines are kept at high pressure up to some valve that controls the filling. (where?).
After the valve you have expansion and heating. In the tank compression and cooling.

[cscott]

let's say that some expansion in the helium (PV/T=K) caused a drop in temperature, that could still cool the LOX in the COPV to the point that it freezes.

Under the applicable conditions, helium warms up as it expands.

And yet, it seems, the LOX froze.

Exactly.  Hence the duration of the investigation.

[gospacex]

If I'm understanding the current leading theory for the failure, the problem (solid Oxygen in the COPV liner) is something that they would only expect to occur with sub-cooled LOX, and not with regular LOX.  The solution is to use slightly warmer LOX? (with associated performance penalty).

Not sure of the actual order of loading, nor complexity of the tanking apparatus, but would it be possible/advantageous to have two LOX sources on the pad at different temperatures, and pre-fill the tank with the warmer LOX to saturate the COPV liners, then fill the rest with sub-cooled LOX so the performance hit isn't as great?

The LOX had to be liquid to get where it is causing trouble. To me this implies, that the COPV was colder than expected and caused the LOX to freeze. So the solution will probably be a procedural change that makes sure that the COPV doesn't cool below the freezing point of LOX.

My reading of all available information is that COPVs were supposed to be impermeable by LOX, but rapid and uneven cooling of warm/hot COPV by splashing LOX during fueling caused delamination and LOX ingress. Regardless of phase (liquid/solid), that's already not a nominal condition. The fix would be to stop that from happening.

[HMXHMX]

So what were the helium tanks in Saturn V made of? What was the thickness/weight. I cant find it.

According to the carefully researched paper attached to this post (see p. 5), it was Al.

Only for the S-IC.  The S-VIB bottles shown in the photo above were titanium.

[HMXHMX]

If I'm understanding the current leading theory for the failure, the problem (solid Oxygen in the COPV liner) is something that they would only expect to occur with sub-cooled LOX, and not with regular LOX.  The solution is to use slightly warmer LOX? (with associated performance penalty).

Not sure of the actual order of loading, nor complexity of the tanking apparatus, but would it be possible/advantageous to have two LOX sources on the pad at different temperatures, and pre-fill the tank with the warmer LOX to saturate the COPV liners, then fill the rest with sub-cooled LOX so the performance hit isn't as great?

The liner (aluminum)can't be saturated with LOX as it is impermeable.  The over-wrap (carbon-urethane) might be.

[John Santos]

If I'm understanding the current leading theory for the failure, the problem (solid Oxygen in the COPV liner) is something that they would only expect to occur with sub-cooled LOX, and not with regular LOX.  The solution is to use slightly warmer LOX? (with associated performance penalty).

Not sure of the actual order of loading, nor complexity of the tanking apparatus, but would it be possible/advantageous to have two LOX sources on the pad at different temperatures, and pre-fill the tank with the warmer LOX to saturate the COPV liners, then fill the rest with sub-cooled LOX so the performance hit isn't as great?

The LOX had to be liquid to get where it is causing trouble. To me this implies, that the COPV was colder than expected and caused the LOX to freeze. So the solution will probably be a procedural change that makes sure that the COPV doesn't cool below the freezing point of LOX.

My reading of all available information is that COPVs were supposed to be impermeable by LOX, but rapid and uneven cooling of warm/hot COPV by splashing LOX during fueling caused delamination and LOX ingress. Regardless of phase (liquid/solid), that's already not a nominal condition. The fix would be to stop that from happening.

This is at odds with my understanding, which is that the overwrap (not the COPV itself) is and always was permeable to LOX.  Previously, the overwrap would absorb LOX (sort of like a sponge) and when the tank was fully pressurized and the liner expanded against the overwrap, the LOX got squeezed out.  Since it was liquid, this worked fine and didn't subject the fibers in the overwrap to any excessive forces.  However (possibly or probably due to the changed procedures and odd physics), this time, some of the LOX that had permeated the overwrap froze. 

When the tank expanded as it approached flight pressure, the frozen oxygen couldn't be displaced and either caused additional stress on the fibers or the sharp edges of the crystalline solid oxygen cut them or both.  Enough of the fibers broke that the overwrap delaminated, and the COPV liner burst.

Edit: Use the term "liner" instead of "core" to indicate the metal tank that is covered by the composite overwrap.  This seems to be the standard terminology, at least in this discussion.

[M_Puckett]

Could it have been caused by a hidden flaw in the bottle?  A crack in the overwrap caused by a bump either at the manufacturer or the Hawthorne assembly line?  I know composite SRM's have failed for similar reasons.

Would SpaceX have video of all the assembly and processing ops involving that bottle?

And the Saturn was where I got Titanium for the He bottles.  Of course, they weren't sitting in LOX either.

[M_Puckett]

If I'm understanding the current leading theory for the failure, the problem (solid Oxygen in the COPV liner) is something that they would only expect to occur with sub-cooled LOX, and not with regular LOX.  The solution is to use slightly warmer LOX? (with associated performance penalty).

Not sure of the actual order of loading, nor complexity of the tanking apparatus, but would it be possible/advantageous to have two LOX sources on the pad at different temperatures, and pre-fill the tank with the warmer LOX to saturate the COPV liners, then fill the rest with sub-cooled LOX so the performance hit isn't as great?

The LOX had to be liquid to get where it is causing trouble. To me this implies, that the COPV was colder than expected and caused the LOX to freeze. So the solution will probably be a procedural change that makes sure that the COPV doesn't cool below the freezing point of LOX.

My reading of all available information is that COPVs were supposed to be impermeable by LOX, but rapid and uneven cooling of warm/hot COPV by splashing LOX during fueling caused delamination and LOX ingress. Regardless of phase (liquid/solid), that's already not a nominal condition. The fix would be to stop that from happening.

This is at odds with my understanding, which is that the overwrap (not the COPV itself) is and always was permeable to LOX.  Previously, the overwrap would absorb LOX (sort of like a sponge) and when the tank was fully pressurized and expanded against the overwrap, the LOX got squeezed out.  Since it was liquid, this worked fine and didn't subject the fibers in the overwrap to any excessive forces.  However (possibly or probably due to the changed procedures and odd physics), this time, some of the LOX that had permeated the overwrap froze. 

When the tank expanded as it approached flight pressure, the frozen oxygen couldn't be displaced and either caused additional stress on the fibers or the sharp edges of the crystalline solid oxygen cut them or both.  Enough of the fibers broke that the overwrap delaminated, and the COPV core burst.

Has SpaceX said if they have managed to duplicate this failure mode in testing?

[HMXHMX]

If I'm understanding the current leading theory for the failure, the problem (solid Oxygen in the COPV liner) is something that they would only expect to occur with sub-cooled LOX, and not with regular LOX.  The solution is to use slightly warmer LOX? (with associated performance penalty).

Not sure of the actual order of loading, nor complexity of the tanking apparatus, but would it be possible/advantageous to have two LOX sources on the pad at different temperatures, and pre-fill the tank with the warmer LOX to saturate the COPV liners, then fill the rest with sub-cooled LOX so the performance hit isn't as great?

The LOX had to be liquid to get where it is causing trouble. To me this implies, that the COPV was colder than expected and caused the LOX to freeze. So the solution will probably be a procedural change that makes sure that the COPV doesn't cool below the freezing point of LOX.

My reading of all available information is that COPVs were supposed to be impermeable by LOX, but rapid and uneven cooling of warm/hot COPV by splashing LOX during fueling caused delamination and LOX ingress. Regardless of phase (liquid/solid), that's already not a nominal condition. The fix would be to stop that from happening.

This is at odds with my understanding, which is that the overwrap (not the COPV itself) is and always was permeable to LOX.  Previously, the overwrap would absorb LOX (sort of like a sponge) and when the tank was fully pressurized and the liner expanded against the overwrap, the LOX got squeezed out.  Since it was liquid, this worked fine and didn't subject the fibers in the overwrap to any excessive forces.  However (possibly or probably due to the changed procedures and odd physics), this time, some of the LOX that had permeated the overwrap froze. 

When the tank expanded as it approached flight pressure, the frozen oxygen couldn't be displaced and either caused additional stress on the fibers or the sharp edges of the crystalline solid oxygen cut them or both.  Enough of the fibers broke that the overwrap delaminated, and the COPV liner burst.

Edit: Use the term "liner" instead of "core" to indicate the metal tank that is covered by the composite overwrap.  This seems to be the standard terminology, at least in this discussion.

It is not given that the over-wrap need be permeable.  The whole point of tailoring the material properties of the wrap matrix material is to avoid permeability.

[MarekCyzio]

The liner (aluminum)can't be saturated with LOX as it is impermeable.  The over-wrap (carbon-urethane) might be.

Wouldn't this be dangerous? Possibility of pockets of LOX that get compressed between carbon fiber strands? Compression may cause LOX to become solid...

[gospacex]

If I'm understanding the current leading theory for the failure, the problem (solid Oxygen in the COPV liner) is something that they would only expect to occur with sub-cooled LOX, and not with regular LOX.  The solution is to use slightly warmer LOX? (with associated performance penalty).

Not sure of the actual order of loading, nor complexity of the tanking apparatus, but would it be possible/advantageous to have two LOX sources on the pad at different temperatures, and pre-fill the tank with the warmer LOX to saturate the COPV liners, then fill the rest with sub-cooled LOX so the performance hit isn't as great?

The LOX had to be liquid to get where it is causing trouble. To me this implies, that the COPV was colder than expected and caused the LOX to freeze. So the solution will probably be a procedural change that makes sure that the COPV doesn't cool below the freezing point of LOX.

My reading of all available information is that COPVs were supposed to be impermeable by LOX, but rapid and uneven cooling of warm/hot COPV by splashing LOX during fueling caused delamination and LOX ingress. Regardless of phase (liquid/solid), that's already not a nominal condition. The fix would be to stop that from happening.

This is at odds with my understanding, which is that the overwrap (not the COPV itself) is and always was permeable to LOX.  Previously, the overwrap would absorb LOX (sort of like a sponge)

Cimarron composites are said to build COPVs for SpaceX before. Watch this video from their site.



It's not scientific, but to me, the surface of COPVs tested looks like solid epoxy resin. Very much not like sponge.

[CJ]

There is both expansion and compression of helium going on NEAR the copv's.
At first helium expands into the copv's..
At some point the dominate is the compression of the helium to flight pressure.
This assumes the helium lines are kept at high pressure up to some valve that controls the filling. (where?).
After the valve you have expansion and heating. In the tank compression and cooling.

Bolding mine.

I can't shake one tidbit of info reported early on; that for this static fire, SpaceX was trying some new methods to alleviate the need launch or scrub (due to subcooled LOX temp issues taking away the ability to hold). If that's accurate, we may well be looking at procedures differing from those of the past. 

So, could one of them be relevant to the bit I bolded; a change that alters the temperature profile of the He loading (intentionally or not) and thus compression of the He cooling the COPV enough, coupled with colder LOX than previously, to allow O2 ice to form in the composite wrap?

[rsdavis9]

There is both expansion and compression of helium going on NEAR the copv's.
At first helium expands into the copv's..
At some point the dominate is the compression of the helium to flight pressure.
This assumes the helium lines are kept at high pressure up to some valve that controls the filling. (where?).
After the valve you have expansion and heating. In the tank compression and cooling.

Bolding mine.

I can't shake one tidbit of info reported early on; that for this static fire, SpaceX was trying some new methods to alleviate the need launch or scrub (due to subcooled LOX temp issues taking away the ability to hold). If that's accurate, we may well be looking at procedures differing from those of the past. 

So, could one of them be relevant to the bit I bolded; a change that alters the temperature profile of the He loading (intentionally or not) and thus compression of the He cooling the COPV enough, coupled with colder LOX than previously, to allow O2 ice to form in the composite wrap?

I believe this the case also. The procedures are what caused the lox ice to form.

Hopefully spacex has been successful in recreating the failure in their tanking tests at mcgregor.

The other possibility is that they haven't recreated the failure because of the low probability of the failure even when exacerbated by different tanking procedures. In which case they will file a accident report with associated probabilities of failure caused by such and such.