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

[Lumina]

Not exactly.

Apparently, if you load the helium tank just right wrong, there is a momentary drop in temperature that can freeze supercooled O2 into solid O2. This O2 is between the carbon overrap and the inner pressure shell of the tank.
Before this solid O2 can melt, helium pressures press the inner pressure shell into the carbon overwrap as designed. This squeezes out any liquid oxygen... but not solid oxygen.

As the pressure rises, Solid oxygen is pressed into elemental carbon until a spontaneous, catastrophic molecular exchange happens.

Great explanation. The sub-cooling change was a recent introduction of a new and relatively untested technology and on top of that, it was also a change which interacted with many other subsystems AND which forced a Conops change as Jim noted. From a systems perspective, this scared me from the first time they tried it. So, two months ago when I heard that the F9 blew on the pad during fueling, I immediately had the feeling that the novel sub-cooled LOX would somehow be a part of the root cause story.

My take from the Amos-6 episode and the way they developed the landing capability is that SpaceX are effectively learning by doing, which is probably unavoidable if you are trying to do things that the industry never really did, like propulsively landing your stages, or building spacecraft that can do return trips to anywhere in the solar system where there is water and carbon. The assumption of risk comes with the territory of learning by doing and it is also noble and admirable at the same time. Comparing SpaceX to other companies/organizations which focus on retiring all risk at the expense of the rate of advance of the technology really is an apples-and-oranges business, IMHO.

[yokem55]

I am under the impression that densification doesn't work well on S2 because of the location of the helium tanks.  I am also under then impression that it actually works on S1 since there have been no failures on S1 other than one engine was shut down early a couple of years ago.   Therefore should it be abandoned for S2 unless it is redesigned or flaws worked out?

By that logic, the S2 design was perfect in July 2016.

S2 is getting reworked anyways as they started dev on a new S2 replacement AFAIK at the end of last year.

Woah. Any details or L2 links that I've missed?

[Herb Schaltegger]

I am under the impression that densification doesn't work well on S2 because of the location of the helium tanks.  I am also under then impression that it actually works on S1 since there have been no failures on S1 other than one engine was shut down early a couple of years ago.   Therefore should it be abandoned for S2 unless it is redesigned or flaws worked out?

By that logic, the S2 design was perfect in July 2016.

S2 is getting reworked anyways as they started dev on a new S2 replacement AFAIK at the end of last year.

Woah. Any details or L2 links that I've missed?

it exists in other threads and came from Elon leading up to F9FT first flight.

Please provide sources. This "news" does not comport with any solid information I have seen posted anywhere.

[Proponent]

Image of wire copy in this tweet

Loren Grush ✔ @lorengrush
Sigh, I have a bad feeling I'm going to be covering a SpaceX launch during my annual Christmas party again
1:49 PM - 4 Nov 2016

Does anyone know why the quote mentions liquid helium? I was under the impression F9 used just cooled GHe? I can't listen to the CNBC clip myself. Thanks

For anyone who hasn't seen the relevant CNBC clip and can view it, it's here

Two guesses as to why liquid helium is mentioned.  Maybe its just sloppy terminology, and the helium is actually supercritical.  Or maybe SpaceX pumps a certain amount of honest-to-goodness liquid helium into the tank to chill things down faster, but by lift-off all of the liquid has evaporated (or undergone whatever you call the transition from liquid to supercritical).

[sdsds]

Two guesses as to why liquid helium is mentioned.  Maybe its just sloppy terminology, and the helium is actually supercritical.  Or maybe SpaceX pumps a certain amount of honest-to-goodness liquid helium into the tank to chill things down faster, but by lift-off all of the liquid has evaporated (or undergone whatever you call the transition from liquid to supercritical).

I'm struggling to understand this. Elon did say liquid helium; maybe he simply mis-spoke.

The critical point of helium is near 5.2 K at 0.23 MPa (roughly 2.2 atm). No amount of added pressure makes it a normal liquid until the temperature goes below 5.2 K, right? Does anyone believe SpaceX designed a propellant loading process that involves temperatures that low?

If not it must be a supercritical fluid and my money says Elon knew that. Potentially he didn't say the word "supercritical" because it has a different meaning (with negative connotations) in the field of nuclear fission.

EDIT to add diagram taken from http://bbradu.web.cern.ch/bbradu/cryogenics.php

[ChrisWilson68]

Two guesses as to why liquid helium is mentioned.  Maybe its just sloppy terminology, and the helium is actually supercritical.  Or maybe SpaceX pumps a certain amount of honest-to-goodness liquid helium into the tank to chill things down faster, but by lift-off all of the liquid has evaporated (or undergone whatever you call the transition from liquid to supercritical).

Or, maybe they didn't intend to put liquid helium in the tank at all but they made a mistake and ended up with some liquid helium.  For example, maybe there is very cold gaseous helium under high pressure in a feed line.  They open a valve to let the helium into the tank slowly.  On going from the high-pressure feed line to the lower-pressure tank, it expands and cools and turns liquid.  Then the liquid helium pools at the bottom of the tank and causes oxygen to solidify beneath it, between the tank liner and the carbon overwrap.  We know the oxygen solidified when the tank was at relatively low pressure, because that's the whole problem -- the O2 solidified, then the tank came up to pressure, squeezing the solid O2 between the metal inner layer of the tank and the outer carbon layer and then bang.

[john smith 19]

I'm struggling to understand this. Elon did say liquid helium; maybe he simply mis-spoke.

The critical point of helium is near 5.2 K at 0.23 MPa (roughly 2.2 atm). No amount of added pressure makes it a normal liquid until the temperature goes below 5.2 K, right? Does anyone believe SpaceX designed a propellant loading process that involves temperatures that low?

Given this is about 1/4 the boiling point of LH2, which SX have avoided like the plague precisely due to issues of cost and thermal management, using LHe makes no sense. 

If not it must be a supercritical fluid and my money says Elon knew that. Potentially he didn't say the word "supercritical" because it has a different meaning (with negative connotations) in the field of nuclear fission.

Now that I can believe.  Musk seems to have a pretty good feel for the average scientific literacy level of the US and its elected representatives.

[Fred Bonyea]

Interesting allotropes of solid Oxygen:
(quote=Wikipedia)
The metastable molecule tetraoxygen (O
4) was discovered in 2001,[38][39] and was assumed to exist in one of the six phases of solid oxygen. It was proven in 2006 that this phase, created by pressurizing O
2 to 20 GPa, is in fact a rhombohedral O
8 cluster.[40] This cluster has the potential to be a much more powerful oxidizer than either O
2 or O
3 and may therefore be used in rocket fuel.[38][39] A metallic phase was discovered in 1990 when solid oxygen is subjected to a pressure of above 96 GPa[41] and it was shown in 1998 that at very low temperatures, this phase becomes superconducting.[42](/quote)
Can you increase the reactivity of O2 by freezing it? Superconductivity can throw your heat transfer models out-of-the window. 

[AncientU]

Interesting allotropes of solid Oxygen:
<snip>

The allotropes of oxygen you mention both only occur at over several gigapascals.
This is considerably over the chamber pressure, never mind the tank pressure.
There is no conceivable way that any rocket would use tank pressures this high.

Doesn't have to be the bulk tank pressure.  If LOX is changing state within the over-wrap crevices, and then being compressed by differential CTE of the COPV materials, local pressure in the crevice could soar.

Differential CTE can create 'astronomical' stresses.

[Proponent]

Two guesses as to why liquid helium is mentioned.  Maybe its just sloppy terminology, and the helium is actually supercritical.  Or maybe SpaceX pumps a certain amount of honest-to-goodness liquid helium into the tank to chill things down faster, but by lift-off all of the liquid has evaporated (or undergone whatever you call the transition from liquid to supercritical).

Or, maybe they didn't intend to put liquid helium in the tank at all but they made a mistake and ended up with some liquid helium.  For example, maybe there is very cold gaseous helium under high pressure in a feed line.  They open a valve to let the helium into the tank slowly.  On going from the high-pressure feed line to the lower-pressure tank, it expands and cools and turns liquid....

It appears, though, that helium warms with expansion under the relevant conditions (its Joule-Thomson coefficient is negative).

[rsdavis9]

Two guesses as to why liquid helium is mentioned.  Maybe its just sloppy terminology, and the helium is actually supercritical.  Or maybe SpaceX pumps a certain amount of honest-to-goodness liquid helium into the tank to chill things down faster, but by lift-off all of the liquid has evaporated (or undergone whatever you call the transition from liquid to supercritical).

Or, maybe they didn't intend to put liquid helium in the tank at all but they made a mistake and ended up with some liquid helium.  For example, maybe there is very cold gaseous helium under high pressure in a feed line.  They open a valve to let the helium into the tank slowly.  On going from the high-pressure feed line to the lower-pressure tank, it expands and cools and turns liquid....

It appears, though, that helium warms with expansion under the relevant conditions (its Joule-Thomson coefficient is negative).

I agree about everybody getting the helium hot or cold wrong.

So I think as it expands into the tank it warms but as the total tank pressure goes up it cools. So maybe that means the net is zero.

So you could imagine warming the whole tank to some pressure, the lox cooling it down, and then a final pressure increase causing it to get colder.

Obviously there is a orifice or limiting pipe that the ground side pressure decreases across. (don't know where this is) After that it is just the whole tank pressure increasing which will cause a temperature drop.
(this is in the tank).
 

[cscott]

Back to the "thermoacoustic heat engine" theory?  Independent from normal Joule-Thompson processes, some resonance pushed the helium below 5K and the LOX below freezing.

https://en.wikipedia.org/wiki/Thermoacoustic_heat_engine

As I read it, all that's required is a standing acoustic (pressure) wave which pushes the pressure of the helium up enough to drop it's temperature below 5K.  If it's a standing wave, the now-liquid helium should basically sit still in that spot, which could freeze the LOX on the other side of the inlet tube/copv/whatever.

A limited number of discrete temperature/pressure sensors might find it difficult to resolve the actual character of a standing wave, so "normal" pressure/temperature readings could be misleading.  You might be able to detune the plumbing to avoid the resonance.  Another "cheap" solution might be to add additional sensors at narrower spacing to ensure that the frequency/amplitude of the standing wave can be reconstructed (Nyquist!) and the control software taught to control things based on the reconstructed Fourier amplitude, not just the literal measurements seen by the sensors (which could be nearer the node of the standing wave than the peak).  Or if the characteristics of the standing wave are well understood, the existing sensors could just be relocated to the peaks of the standing wave.

They could have gotten unlucky in triggering the resonance w/ the exact pipe dimensions and input pressures on the pad that day; reconstructing at MacGregor may have involved sweeping the pressures or temperatures until they got their test hardware to "sing" properly.

[Jet Black]

I'm curious, if something like the above did happen and the helium condensed, then what would happen to the COPV from a purely pressure point of view? Would the pressure from the LOX be enough to collapse it?

[ChrisWilson68]

I'm curious, if something like the above did happen and the helium condensed, then what would happen to the COPV from a purely pressure point of view? Would the pressure from the LOX be enough to collapse it?

Various media reports have been saying it wasn't pressure that collapsed the COPV, it was oxygen ice squeezed between the inner metal shell of the COPV and the carbon outer wrap that pushed oxygen and carbon together at high pressure causing a chemical reaction between the oxygen and carbon.

[cscott]

It woukd depend on exactly where the standing wave shows up, wouldn't it?  If the standing wave was in the tank itself, then the average pressure would be "as designed" but there would be peaks and troughs of high/low pressure.  Dunno what that would do to side wall loading.  Don't think it would "collapse" per se.

I'm sort of expecting that the resonance set up in a feed tube, and that the low temperature then got conducted somewhere "bad".  For example, standing wave in the inlet tube leading to lower temperatures around the mouth of the copv, causing some solid lox to form there, just where the dome structure and inlet make the fiber wrapping tricky.

[.Scott]

I agree about everybody getting the helium hot or cold wrong.

So I think as it expands into the tank it warms but as the total tank pressure goes up it cools. So maybe that means the net is zero.

So you could imagine warming the whole tank to some pressure, the lox cooling it down, and then a final pressure increase causing it to get colder.

Obviously there is a orifice or limiting pipe that the ground side pressure decreases across. (don't know where this is) After that it is just the whole tank pressure increasing which will cause a temperature drop.
(this is in the tank).
 

At first, when it expands into the tank, it will either stay the same temperature or cool - depending on its temperature.  This cooling effect is called the Joule-Thomson effect.  But it only affects Helium at temperatures below about 51K - already enough to freeze LOX.

[rsdavis9]

I agree about everybody getting the helium hot or cold wrong.

So I think as it expands into the tank it warms but as the total tank pressure goes up it cools. So maybe that means the net is zero.

So you could imagine warming the whole tank to some pressure, the lox cooling it down, and then a final pressure increase causing it to get colder.

Obviously there is a orifice or limiting pipe that the ground side pressure decreases across. (don't know where this is) After that it is just the whole tank pressure increasing which will cause a temperature drop.
(this is in the tank).
 

At first, when it expands into the tank, it will either stay the same temperature or cool - depending on its temperature.  This cooling effect is called the Joule-Thomson effect.  But it only affects Helium at temperatures below about 51K - already enough to freeze LOX.

You got it backwards. Above 50K it has the negative Joule-Thomson. Below 50K it behaves like most other gases in that expansion equals cooling and compression equals heating.

Helium has a negative Joule-Thomson coefficient at normal ambient temperatures, meaning it heats up when allowed to freely expand. Only below its Joule-Thomson inversion temperature (of about 32 to 50 K at 1 atmosphere) does it cool upon free expansion.[13] Once precooled below this temperature, helium can be liquefied through expansion cooling

[ChrisWilson68]

But it only affects Helium at temperatures below about 51K - already enough to freeze LOX.

Just because gaseous helium is below 51 K doesn't mean it will freeze any LOX on the outside of the COPV liner.  There's a layer of metal between the gaseous helium and the LOX, and the LOX, as a liquid, is likely to be better at conducting heat than the gaseous helium.  So, it could well be that even down to 5K the gaseous helium won't freeze any LOX, because the metal shell of the tank stays at roughly the temperature of the LOX, not the helium.  But then when the helium turns liquid suddenly it can cool the metal down a lot more by conducting heat out of the metal and a layer of LOX ice forms.

So, it's plausible the SpaceX design envisioned helium temps well below 51K and concluded that was not a problem.  The temperature of the helium coming into the tank might have been anything above 5K.

[cscott]

For the standing wave heat engine theory, it doesn't matter what sign the Joule-Thompson coefficient has: there are both high and low pressure peaks.  One of them is getting colder.  It would be more interesting if the coefficient sign change were somehow intimately involved, though.  What if the high pressure peak was above 50K and the low pressure peak dropped below 50K?  Then both half-cycles would be pulling heat from the environment.  Seems like this should be self-limiting once the high pressure region dropped below 50K, but that would be fun to model.

Liquid He is 5K, suggesting that (if Elon was speaking accurately) the effect, whatever it is, is happening well below the 50K transition temperature, but it's possible some of the "weirdness" of helium helped it unexpectedly drop that far.

[as58]

Just because gaseous helium is below 51 K doesn't mean it will freeze any LOX on the outside of the COPV liner.  There's a layer of metal between the gaseous helium and the LOX, and the LOX, as a liquid, is likely to be better at conducting heat than the gaseous helium. 

There doesn't appear to be a very large difference in thermal conductivity, only a factor of ~3. (The data are for gas at ~1 atmosphere, but at least for ideal gasses thermal conductivity is independent of pressure. Helium close to liquefaction is of course not exactly an ideal gas.)

https://www.bnl.gov/magnets/staff/gupta/cryogenic-data-handbook/subject.htm

edit: Below lambda point the thermal conductivity of liquid helium is huge, but it seems to me that getting helium to liquefy at all, let alone reaching the lambda point is quite a stretch.